Intentionality in Sellars: A Transcendental Account of Finite Knowledge [1 ed.] 9781032114934, 9781003221364, 9781032117584

This book argues that Sellars’ theory of intentionality can be understood as an advancement of a transcendental philosop

126 81 7MB

English Pages 230 [253] Year 2021

Report DMCA / Copyright

DOWNLOAD PDF FILE

Recommend Papers

Intentionality in Sellars: A Transcendental Account of Finite Knowledge [1 ed.]
 9781032114934, 9781003221364, 9781032117584

  • 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

Routledge Studies in American Philosophy

INTENTIONALITY IN SELLARS A TRANSCENDENTAL ACCOUNT OF FINITE KNOWLEDGE Luz Christopher Seiberth

It has long been accepted that Sellars is a Kantian philosopher, but until now, we lacked a detailed understanding of what that meant and why it mattered. Seiberth’s monograph demonstrates in detail that Sellars was not just a Kantian philosopher but also a first-rate Kant scholar. Seiberth shows that Sellars’s claims about Kant can withstand intense critical scrutiny and meticulous comparison with Kant’s texts and arguments. With hermeneutic sensitivity and analytic acumen Seiberth shows that Sellars’s engagement with Kant is inseparable from his philosophy. He also shows that contemporary Kant scholars cannot afford to ignore Sellars. Carl B. Sachs, Marymount University, USA In this serious, in-depth book, the author underlines the strong Kantian streak in Sellars’s system and emphasizes the crucial work of the theory of picturing, two fundamental yet often neglected aspects of Sellars’s theory of intentionality. Patrice Philie, University of Ottawa, Canada Sellars is now recognized as a philosopher who defended the broadly Kantian view that the contentfulness of our thought and intentionality is a matter of the largely implicit and social prescription of rules or laws that are of the understanding’s own making, in the ‘space of reasons’. This normative account has generally been detached from the more naturalistic side of Sellars’ story that emphasizes our environmental embeddedness, and in particular the nature of the systematic causal hookups and representational mappings between mind and world that Sellars called ‘picturing’. Seiberth’s book on Sellars is the first to explain in systematic depth the Kantian transcendental method that constitutes Sellars’ attempt to bring both of those sides of the story together in their necessary and dynamic interdependence. The convincing result is that the normative story of our intentionality cannot be told without the scientific naturalist story, and vice versa. James O’Shea, University College Dublin, Ireland

Intentionality in Sellars

This book argues that Sellars’ theory of intentionality can be understood as an advancement of a transcendental philosophical approach. It shows how Sellars develops his theory of intentionality through his engagement with the theoretical philosophy of Immanuel Kant. The book delivers a provocative reinterpretation of one of the most problematic and controversial concepts of Sellars’ philosophy: the picturing-­ relation. Sellars’ theory of intentionality addresses the question of how to reconcile two aspects that seem opposed: the non-relational theory of intellectual and linguistic content and a causal-transcendental theory of representation inspired by the philosophy of the early Wittgenstein. The author explains how both parts cohere in a transcendental account of finite knowledge. He claims that this can only be achieved by reading Sellars as committed to a transcendental methodology inspired by Kant. In a final step, he brings his interpretation to bear on the contemporary metaphilosophical debate on pragmatism and expressivism. Intentionality in Sellars will be of interest to scholars of Sellars and Kant, as well as researchers working in philosophy of mind, epistemology, and the history of nineteenth- and twentieth-century philosophy. Luz Christopher Seiberth is a lecturer at the chair for theoretical philosophy at the University of Potsdam where received his Ph.D. in philosophy. He is the co-editor of Fraught with Ought: Writings from Wilfrid Sellars on Mind, Meaning, and Metaphysics (Oxford University Press, forthcoming) and the organiser of the International Sellars Colloquium (www. wilfridsellars.org/isc).

Routledge Studies in American Philosophy Edited by Willem deVries, University of New Hampshire, USA and Henry Jackman, York University, Canada

Challenging the New Atheism Pragmatic Confrontations in the Philosophy of Religion Aaron Pratt Shepherd John Dewey’s Ethical Theory The 1932 Ethics Edited by Roberto Frega and Steven Levine The Philosophy of Ralph Waldo Emmerson Joseph Urbas Pragmatism and Social Philosophy Exploring a Stream of Ideas from America to Europe Edited by Michael G. Festl C. I. Lewis The A Priori and the Given Edited by Quentin Kammer, Jean-Philippe Narboux, and Henri Wagner Charles Peirce on Ethics, Esthetics and the Normative Sciences James Jakób Liszka The Ethics, Epistemology, and Politics of Richard Rorty Edited by Giancarlo Marchetti Intentionality in Sellars A Transcendental Account of Finite Knowledge Luz Christopher Seiberth For more information about this series, please visit: https://www.­ routledge.com/Routledge-Studies-in-American-Philosophy/book-series/ RSAP

Intentionality in Sellars A Transcendental Account of Finite Knowledge

Luz Christopher Seiberth

First published 2022 by Routledge 605 Third Avenue, New York, NY 10158 and by Routledge 2 Park Square, Milton Park, Abingdon, Oxon, OX14 4RN Routledge is an imprint of the Taylor & Francis Group, an informa business © 2022 Luz Christopher Seiberth The right of Luz Christopher Seiberth to be identified as author of this work has been asserted 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: Seiberth, Luz Christopher, author. Title: Intentionality in Sellars: a transcendental account of finite knowledge/Luz Christopher Seiberth. Description: New York, NY: Routledge, [2021] | Series: Routledge studies in American philosophy | Includes bibliographical references and index. Identifiers: LCCN 2021032728 (print) | LCCN 2021032729 (ebook) | ISBN 9781032114934 (hbk) | ISBN 9781003221364 (ebk) | ISBN 97810032117584 (pbk) Subjects: LCSH: Sellars, Wilfrid. | Intentionality (Philosophy) Classification: LCC B945.S444 S445 2021 (print) | LCC B945. S444 (ebook) | DDC 128/.2–dc23 LC record available at https://lccn.loc.gov/2021032728 LC ebook record available at https://lccn.loc.gov/2021032729 ISBN: 978-1-032-11493-4 (hbk) ISBN: 978-1-032-11758-4 (pbk) ISBN: 978-1-003-22136-4 (ebk) DOI: 10.4324/9781003221364 Typeset in Sabon by KnowledgeWorks Global Ltd.

Contents

Abbreviationsx Foreword by Ray Brassierxi Acknowledgementsxxi

Introduction1 The Concept of Intentionality

1 Transcendental Methodology9 1.1

Transcendental Methodology and the Role of Analytic Truths 9 1.2 Presuppositional Analysis 17 1.3 Levels of Abstraction 19 2 Transcendental Psychology27 2.1 The Meta-Classification Thesis 27 2.1.1 Categories as Meta-Classificatory Rules 28 2.1.2 Objection 1: Categories as Highest Kinds 30 2.1.3 Objection 2: Categories as Arbitrary Constructs 31 2.2 The Concept of an Object of Experience in Kant 35 2.2.1 Categories for Mental Acts 36 2.2.2 Contents We Think and Contents We Experience 39 2.2.3 The Role of Synthesis 46 2.2.4 Successiveness, Subjective, and Objective 52 3 Perceptual Experience69 3.1 The in Itself and One World 71 3.2 Sheer Manifold of Sense 75 3.3 Sensations, Sense Impressions, and Intuitions 77

viii  Contents 3.4

3.5 3.6 3.7 3.8 3.9 3.10

From Image-Models to Intuitions 78 3.4.1 Empirical Schemata 82 3.4.2 Objects 85 Objectivity and Image-Models 86 Intuitions in the Context of Judgements 87 Three Forms of Direct Reference 90 The Instability Thesis 91 Agreements and Disagreements with Kantian Commitments 93 The Status of the Concept of Nature 94

4 Non-Relationality102 4.1 4.2 4.3 4.4

The Expectation 103 The Non-Relationality Thesis 104 Objections from Relationalism 108 Dissatisfaction 113

5 Transcendental Phenomenalism121 5.1 The Argument from Double Nature 122 5.2 The Argument from Immanence 128 5.2.1 Kantian Immanence 131 5.2.2 Sellarsian Immanence 137 5.3 The Argument from Counterparts 142 5.4 Picturing 144 6 Objections and Consequences155 6.1 Objection from Pragmatism 155 6.1.1 Rortyian Immanence 159 6.1.2 Meeting Rorty’s Challenges 160 6.1.3 The Transcendental Function of the Concept of Picturing 169 6.2 Objection from Global Expressivism 176 6.2.1 Truth as Picture and as S-Assertibility 181 6.2.2 The Distinction of i-Representation and e-Representation 182 6.2.3 Discussing Global Expressivism 186 6.3 Consequences for Scientific Realism 190

Contents ix Conclusion Retrospect Outlook

201 201 205

References207 Works by Wilfrid Sellars 207 Works by Other Authors 211 Index

223

Abbreviations

The following text uses the standard abbreviations for Sellars’ works found at http://www.ditext.com/sellars/bib-s.html and in O’Shea (ed.) 2016a. This facilitates citing Sellars’ works independently of the venue in which they are published. In most cases, numbers in citations refer to paragraphs, e.g. KTI 53 = Kant’s Transcendental Idealism §53. However, if the article does not have numbered paragraphs, page numbers are referred to instead, e.g. TC 216 = Truth and Correspondence page 216. Works comprising more than one chapter are cited in the following manner: e.g. Science and Metaphysics is cited as SM, paragraphs inside individual chapters are distinguished as follows: SM I §12 refers to chapter one, paragraph 12 and SM V §49 to chapter five paragraph 49. Unless indicated otherwise, all additions in square brackets are mine. Luz Christopher Seiberth

Foreword by Ray Brassier

It is one of the ironies of philosophy’s history that over the course of his long career (from his first appointment at Iowa in 1938 until his death in 1989), Wilfrid Sellars came to occupy an increasingly marginal position within the philosophical tradition he helped establish. Sellars was one of the first generation of Anglophone philosophers to self-identify as a practitioner of ‘analytic’ philosophy. Yet despite the academic distinction he achieved (including professorships at Yale and Pittsburgh), Sellars’ standing within the analytic tradition is sharply disputed. Unlike his contemporary W. V. O. Quine, widely acknowledged as a pivotal figure even by those who lament his influence, Sellars figures barely or not at all in some histories of this tradition.1 There is no doubt that a reputation for obscurity, coupled with what even sympathetic observers describe as the “sheer drearines” (Crane 2008, 33–35) of his writing (certainly when compared to Quine’s scintillating prose), must have winnowed Sellars’ readership during his lifetime. Sellars’ alleged obscurity is generated by the dialectical cast of his writing: instead of proceeding linearly from perspicuous premises to clear conclusions, his texts coil and uncoil in labyrinthine and sometimes disorienting trajectories. They do so because their aim is not merely to refute but to unravel the tangle of distinct confusions motivating the errors they seek to diagnose. Likewise, the lustrelessness of Sellars’ style has been exaggerated. While his early essays are certainly convoluted – and would probably be unpublishable today – his mature style is austere yet elegant and frequently leavened by pithy epigrams. Robert Brandom suggests a more telling reason for Sellars’ contested status within the analytic tradition. During the 1940s and 1950s, Sellars was at the vanguard of the analytic movement, producing difficult but groundbreaking work that culminated with 1956’s Empiricism and the Philosophy of Mind (henceforth EPM), which Tyler Burge, certainly no disciple of Sellars, called “the most influential paper of the period” (Burge 1992, 33). Sellars’ animating ambition was, as he put it, to usher analytic philosophy from its Humean into its Kantian phase (Brandom 2015, 4–5) – an ambition signalled in the subtitle Variations on Kantian

xii  Foreword by Ray Brassier Themes, which he appended to Science and Metaphysics, published in 1968. But as Brandom points out, in the decade between 1960, which saw the publication of Quine’s seminal Word and Object, and 1970, when Kripke delivered the influential lectures subsequently published as Naming and Necessity, “The Kantian Kehre of analytic philosophy was not taking place – indeed, it barely registered as a possibility.” (Brandom 2015, 20). It was during this decade that Sellars fell radically out of step with the analytic mainstream. Analytic philosophy underwent a shift – but it was a shift from empiricism towards rationalist metaphysics, rather than from Hume to Kant. Sellars’ professional eclipse persisted until his (alcohol-induced) death in 1989. But he continued to have some prominent admirers, one of whom was Richard Rorty. In his 1970 review of Science and Metaphysics, Rorty described Sellars as “the most original and far-sighted systematic philosopher now writing in English” (Rorty 1970, 66). Then, in 1979s seminal Philosophy and the Mirror of Nature, Rorty enlisted Sellars’ critique of the given in his campaign against foundationalism (together with its avatars, which include essentialism, representationalism, epistemology, metaphysics, and ultimately philosophy itself). 2 Rorty’s subsequent renown kept Sellars’ name alive in Anglo-American philosophy. But, however beneficent, Rorty’s endorsement skewed ensuing interpretations of Sellars’ achievement. His distinction between ‘left’ and ‘right’ Sellarsianism suggested that the systematicity of Sellars’ thought is vitiated by a basic, easily identifiable inconsistency between commitment to the autonomy of the normative (the progressive option championed by Rorty) and commitment to its naturalisation (the reactionary option defended by metaphysical backsliders). This encouraged the idea that any reckoning with Sellars requires opting either for one or the other. But to approach Sellars’ work in terms of this alternative is to embrace the very dichotomy he sought to overcome. As James O’Shea (2007, 2010, 2011) has pointed out, what is most confounding in Sellars is his insistence that normativity is at once autonomous (irreducible) and heteronomous (reducible).3 The gambit that this may not be mere incoherence on Sellars’ part has inspired a new generation of scholars to pursue the cohesiveness of Sellars’ project beyond the Rortyan divide between left normativism and right scientism.4 Luz Christopher Seiberth’s Intentionality in Sellars makes a fundamental contribution in this regard. It re-emphasises Sellars’ Kantianism precisely in order to illuminate what is most radical about his naturalism. Rorty disregarded Sellars’ debt to Kant because he viewed transcendental philosophy as another iteration of metaphysics, which is to say, foundationalism. Yet Sellars famously rejected the alternative between the tortoise of foundationalism and the serpent of coherentism.5 The transcendental component of Sellarsian naturalism consists in the claim that the mind’s knowing of nature is conditioned by the nature it knows. What is the nature of this ‘conditioning’? Once again,

Foreword by Ray Brassier xiii we seem forced to choose: either causal foundation or justificatory coherence. Sellars’ ingenious solution, as Seiberth demonstrates, is that founding is justified conceptually insofar as justification is founded causally. Fleshing out this solution requires Sellars to amend Kant by distinguishing the receptivity of sense, which is not conceptual, from the guidedness of intuition, which is (SM I §40). Where sensations are conceptually intuited conscious states, sense-impressions are transcendentally postulated un-conscious states of organisms; states whose properties are analogous to, but not identical with, those of their typical physical causes.6 They are the non-­conceptual representings that condition the system of conceptual representings, mediating causation and justification. On the basis of this transcendental postulate, Sellars believes he can establish the ‘Janus-faced’ character of languagings as items bridging the normative and causal realms. This is the crux of his theory of picturing, connecting natural-linguistic objects to non-­linguistic objects. Seiberth shows why picturing cannot be properly understood until we have reconstructed the need for it from a transcendental perspective. Picturing anchors intentionality in the causal order. But anchoring is not grounding. To demonstrate this, Seiberth recodes the Kantian distinction between transcendental immanence and metaphysical transcendence by distinguishing between the horizontal and the vertical dimensions of intentionality. Intentionality has been traditionally conceived as a vertical relation connecting thoughts to things (a word–world relation). As is well known, Sellars proposes that it be reconceived as a horizontal relation among thoughts construed as languagings (a word– word relation). Inference is the basic transition from thought to thought. And while perception need not always issue directly from inferentially articulated judgment, it is always indirectly conditioned by it. This is why, as Sellars puts it, the logical dimension in which perception founds judgment is complemented by one in which judgement founds perception.7 What Sellars seeks is the point where this horizontal dimension of logical justification, which contains perception, judgement, and action, intersects with the vertical dimension of efficient causation, which contains none of these. Identifying the point of origin from whence logical immanence and causal transcendence diverge is the goal of Sellars’ transcendental analysis of intentionality. What makes it possible to anchor intentionality in nature is the fact that the transitions from perception to thought and from thought to action form behavioural circuits that can be situated within the causal realm. Thus, while inferences are internally coordinated by relations of justification, their links to perception and action form patterns of behaviour that are externally and causally coordinated with the environment. It is this junction between rule-­governed rational activity and pattern-governed behavioural regularity that Sellars pinpoints in what O’Shea calls his ‘norm-nature meta-­principle’: “Espousal of principles is reflected in uniformities of performance.”

xiv  Foreword by Ray Brassier (TC 216) To explain what this reflection consists in, as well as its properly transcendental import, Seiberth underlines the import of Sellars’ distinction between logical and empirical form. Concepts have logical form by virtue of their inferential roles. Since these inferential roles are justificatory, logical form is equivalent to epistemic power.8 But since Sellars rejects the idea that thoughts as inner episodes are metaphysically distinct from their linguistic expression (whether overt or not) and insists that an utterance of ‘p’ is not just an expression but is itself a thinking that-p, concepts are indissociable from linguistic episodes and thereby possess empirical as well as logical form.9 Their empirical form consists of their ‘matter-of-factual’ properties such as shape, size, colour, and internal structure.10 Linguistic performance reflects normative principle at the level of empirical form. But linguistic performance also stands in a determinate relation to the object(s) by which it is elicited. As Seiberth puts it: “The empirical form of linguistic utterance stands in a determinate relation to the empirical form of those objects by which they are (transcendentally taken to be) evoked” (123). Thus, there is a correspondence between the empirical form of objects and that of the utterances they elicit. However, this is not a first-order correspondence between the empirical properties of utterances and those of the objects causing them, but rather a second-order isomorphism between relations among the properties of objects and relations among the properties of the utterances they cause. It is second order, Seiberth makes clear because the isomorphism “depends for its structure on the method of projection inherent to the conceptual scheme in which it is produced” (144). In other words, we determine the isomorphism between the relations among object properties and the relations among utterance properties. This isomorphism is the picturing relation, understood as a non-­intentional correlation between the matter-of-factual properties of utterances and those of non-linguistic objects. But as Seiberth points out, the transcendental dimension of Sellars’ naturalism follows from his claim that the picturing relation is not only the consequence but also the condition of normative activity. In other words, there can be no espousals of principle without uniformities of performance: “The isomorphism is at once a result of linguistic activity and its transcendental precondition” (123). Consequently, there is an important sense in which logical form depends on empirical form. Seiberth puts this point as follows: [T]he empirical form, allowing a function to be realized or reflected in a uniformity of behaviour, sustains the logical form. And this line of thought helps to address the question of how Sellars can fuse the horizontal dimension of interrelated linguistic functions that govern and inform linguistic activity with the vertical dimension. The solution lies in connecting the claim that logical form needs to be functionally realized in some way or another to the claim that the

Foreword by Ray Brassier xv logical form of linguistic statements is embodied in the materially rich linguistic activities of language users. (Seiberth 125) This has significant consequences for the Rortyan attempt to pit left against right Sellarsianism. Seiberth cites a revealing passage in which Rorty dismisses picturing as the forlorn appeal to a transcendent dimension of truth: Picturing is for Sellars what disclosedness is for Heidegger. It is the extra dimension which relates social practices to something beyond themselves and thus recaptures the Greek problematic of humanity’s relation to the non-human (of nomos vs. physis). In Sellars’ case, this non-human something is ‘the world’. In Heidegger’s case, it is ‘Being’. (Rorty 1988, 216) But as Seiberth demonstrates, the picturing postulate – properly understood, as a matter-of-factual relation in rerum natura between linguistic and non-linguistic objects – is not supposed to index the world’s transcendence vis-à-vis the immanence of human practices, but rather the point at which the immanent logical powers of concepts in the intentional order intersect with the matter-of-factual causal properties of objects and events in the real order. The only sense in which the latter could be said to ‘transcend’ the former is the sense in which what Sellars calls ‘the real’ (or ‘nature’) is devoid of logical form (which belongs only in the conceptual order).11 Lacking logical form, the real cannot ground anything. Thus, as Seiberth points out, picturing is not supposed to provide a metaphysical ground for semantic assertibility, which is truth coextensive with the resources of justification. Nor is it intended to establish a correspondence between assertion and fact since, on Sellars’ account, facts and assertions ultimately coincide.12 The picturing correspondence is not between assertions and facts but between the matter-of-factual properties of acts of assertion and those of the objects eliciting those acts. Semantic assertibility cannot be ‘read off’ picturing activity, and in this sense, writes Seiberth, “there cannot be any semantic usage of the concept of picturing something as something.” (162). Nevertheless, because the picturing isomorphism is also a projection from within our conceptual scheme, “we can consider the items doing the picturing from the intentional stance point of view, thereby allowing us to make contentful claims about them” (147). Seiberth’s point here is that while picturing is a non-intentional correlation between natural objects, we can still establish conceptual correlations among the matter-of-factual properties of these objects. This is one of the aims of science. The correlation of empirical forms is not conceptual. There is a categorial difference

xvi  Foreword by Ray Brassier between the matter-of-factual properties of a musical score and those of the vinyl recording in which it is inscribed since music scores and vinyl discs are different kinds of objects. Yet, there is a causal connection between the matter-of-factual properties of musical structures and those of the devices we use to record them. And this causal connection between the matter-of-factual properties of categorially distinct objects (music scores and vinyl discs) can be described and explained. In doing so, we establish new conceptual connections between the factual properties of categorially distinct objects. It is in this sense, Seiberth suggests, that the picturing postulate is required to account for the advancement of empirical knowledge. Seiberth’s moral here is that the Rortyan rejection of picturing follows from mistaking a transcendental postulate for a metaphysical foundation. What is transcendental is the postulate of the interdependence between reasons and causes, or rules and regularities. Logical power is irreducible to, but also inextricable from, empirical form. This interdependence cannot be dissolved without either inflating the conceptual to the point where it swallows the empirical (reinstating the serpent) or collapsing reasons back into causes (reintroducing the tortoise). Yet Rorty believes he can avoid both without having to resort to picturing. Like Sellars, Rorty wishes to maintain the irreducibility of the normative and the cognitive supremacy of natural science (Sachs 2013, 693). But he insists this can be done without having to invoke (as he thinks Sellars does) “a luminous, self-justifying, self-sufficient synoptic vision” (Rorty 1999b, 20). This insistence follows from the claim that the difference between the normative and the factual, or justifying and describing, boils down to a difference in social practices. While justifying and describing are distinct practices, they are also indissociable, and once this is understood, according to Rorty, the invocation of a set of empirical properties capable of exhibiting language’s – and thereby knowledge’s – relation to reality is not only implausible but unnecessary. It is not facts that ground the normative space of reasons, but social practices. Rorty’s appeal to practice presupposes that “a philosophical account of our practices need not take the form of descriptions of our relation to something not ourselves, but need merely describe our practices.” Indeed, Rorty continues, “the desired ‘relation to the world’ which representationalists fear may be lacking is, Sellars was implying, built into the fact that these are our practices – the practices of real live human beings engaged in causal interaction with the rest of nature” (Rorty 1988, 222). Yet, our causal interactions with ‘the rest of nature’ are mediated by our social interactions with our fellow human beings. If describing these social interactions falls within the cognitive purview of natural science, Rorty turns out to be a reductionist after all. If it does not, it must be because they can only be described ‘from the inside’, so to

Foreword by Ray Brassier xvii speak, or subjectively. And indeed, this seems to be Rorty’s position. He wants Enlightenment secularism – the rejection of transcendence – without Enlightenment rationalism – the desire to see through appearances. Because they are ours, Rorty seems to think that the difference between essence and appearance does not apply to practices. They cannot be turned into objects of knowledge because they are woven together into an intricate meshwork, which we inhabit and cannot step outside. Their plurality indexes the irreducible diversity of human motives, interests, and concerns. It is this diversity that underwrites the difference between prescribing and explaining, or justification and causation. Since we cannot objectify them, the only proper attitude to adopt towards our practices is not knowledge but hope.13 But if practices, in their irreducible plurality, can only be subjectively described and comprehended, Rorty reinstates at the collective level the self-intimating character which he, following Sellars, denies to mind at the individual level. In other words, the transparency Rorty denies when it comes to mind’s self-relation or its relation to nature re-emerges in his claim that our relation to the world is ‘built into’ our relation to our own social practices. Thus Rorty’s appeal to social practice regenerates the transcendence he wishes to avoid. What he rejects in Sellars is the suggestion that some third thing is required to ground the connection between the normative and the factual. But the third thing pops up regardless in Rorty’s account, despite his insistence on its redundancy. It is capitalist liberal democracy as the set of institutions, norms, conventions, and values shaping our practices. These constitute a system of social forms which are perpetuated by our collective practices but whose institution we did not collectively adjudicate. This is simply to say that human practices generate social forms that they do not intend. Such forms are opaque to the practices that generate them; they transcend the subjectivity of those engaging in the practices they shape. What is ignored in Rorty’s claim that our practices do the work Sellars ascribes to picturing is the possibility that we do not always know what we are doing; or as Marx pointed out, that among our habitual everyday practices are things we do without knowing what we are doing – in Marx’s example, buying and selling commodities – and this for the simple reason that we cannot subjectively comprehend the nexus of social relations implicated in commodities and money. This is to say that the subjectivity of practices is objectively mediated by social forms that shape practices even as they are shaped by them. These forms are objective rather than intersubjective because they are not conceptual in nature; their structure and function are not intersubjectively instituted. Indeed, they shape intersubjective space without being shaped by it. It is this objective mediation of subjective practice by social form that has to be accounted for by anyone who proclaims, as Rorty does, that social practice is the bedrock upon which the epistemological edifice founders.

xviii  Foreword by Ray Brassier Philosophers sympathetic to Rortyan pragmatism have suggested replacing his distinction between right and left Sellarsianism with the distinction between ‘object naturalism’ and ‘subject naturalism’. Huw Price defines object naturalism as the ontological view according to which “all there is the world studied by science, together with the epistemological view that “all genuine knowledge is scientific knowledge” (Price 2013, 5). By way of contrast, subject naturalism, in Price’s definition, “begins with what science tells us about ourselves. Science tells us that we humans are natural creatures, and if the claims and ambitions of philosophy conflict with this view, then philosophy needs to give way. This is naturalism in the sense of Hume, then, and arguably Nietzsche” (ibid.). The contrast seems simple enough. Object naturalism says that empirical science tells us everything there is to know about the world. But the accompanying definition of subject naturalism is much more vague. What is it exactly that science tells us about ourselves? We know we are physical entities and biological organisms, constrained by physical law and biological necessity. But this leaves much undetermined. The whole question is whether physics and biology tell us much about our nature as social and cultural beings. And in this regard, Hume’s claims about the primacy of custom and habit, or Nietzsche’s claims about the primacy of power and domination, sound more like speculative extrapolations from fragments of science than scientifically informed hypotheses. This is why, as Price subsequently makes clear, the distinction between subject and object naturalism only really acquires traction in a semantic register. Robert Brandom helpfully reformulates it in semantic terms. Object naturalism is naturalism about the objects and properties that a vocabulary allows us to think and talk about; how it represents the objective world as being. By way of contrast, writes Brandom, “Subject naturalism is a pragmatic naturalism, rather than a representational semantic naturalism. The subject naturalist makes no assumptions about whether the target vocabulary admits of a properly representational semantics [….] What the subject naturalist wants is a naturalistic account of the discursive practices of using the target vocabulary as meaningful in the way it is meaningful” (Brandom 2015, 91). In emphasising the primacy of discursive practice over representational content, subject naturalism subordinates semantic inferences (what is implied by what is said) to the pragmatic inferences (what is implied by saying) that are implicit in discursive practice. Pragmatic inferences are the practical codification of rule-governed patterns of behaviour. These behaviours are enforced by a complex system of social norms. Thus, pragmatic inferences embody social norms. But if so, then when the subject naturalist says she wants to give a naturalistic account of the discursive practice of using a vocabulary, there is a danger that this naturalistic account of inferential practice on its own terms (“as meaningful in the way it is meaningful”) extends to a

Foreword by Ray Brassier xix naturalisation of pragmatic inference which also naturalises the social norms enveloping those inferences. Thus, where object naturalism disregards the mediating role played by social relations in our relationship to nature, subject naturalism disregards it in our relation to our own practices. Its one-sided emphasis on the subjectivity of practice leads subject naturalism to ignore its objective dimension, which cannot be subjectively comprehended. But whether prioritising subjective practices or object properties, both subject and object naturalism ignore the way in which practices and properties are articulated by social forms. What mediates objective fact and subjective practice is a system of social forms that cannot be grasped from either side of the divide. No doubt, the subject naturalist will readily admit that these forms are man-made, not God-given. The question remains whether she is willing to admit not just that humans do not always understand what they have made but that what they have made can turn against them. Raymond Brassier, American University of Beirut

Notes 1. See for instance Scott Soames’ Philosophical Analysis in the Twentieth Century, two volumes, Princeton University Press, 2003 and 2005, as well as Soames’ The Analytic Tradition in Philosophy, two volumes, Princeton University Press, 2014 and 2017. There is not a single mention of Sellars in any of Soames’ four volumes. This is remarkable given their otherwise extraordinary comprehensiveness: even Schelling is granted one mention. Sellars’ name is mentioned four times in the introductory chapter of Soames’ Analytic Philosophy in America (Princeton University Press, 2015). But his work is never discussed, either there or anywhere else in the book. By way of contrast, Quine, David Lewis and Saul Kripke are each allotted separate chapter. Williamson 2014, 7–37 mentions Sellars once in passing as an influence on Rorty, about whom he is politely scathing. For both Soames and Williamson, Kripke is the towering figure of the second half of twentieth century analytic philosophy: in their narratives, admiration for Kripke is inversely proportional to regard for Sellars and the philosophers he influenced. 2. For a penetrating analysis of Rorty’s debt to Sellars, see Sachs 2013, 682–707. 3. See James O’Shea, Wilfrid Sellars: Naturalism with a Normative Turn (Polity, 2007). 4. They include Aude Bandini, Anke Breunig, Boris Brandhoff, Luca Corti, Dionysias Christias, Stefanie Dach, Kevin Fink, Fabio Gironi, Michael Hicks, Bryce Huebner, Jeremy Koons, David Landy, Vera Lyubenova, Catherine Legg, Peter Olen, David Pereplyotchik, Patrick J. Reider, Carl Sachs, Daniel Sacilotto, Lionel Shapiro, Preston Stovall, and Peter Wolfendale.

xx  Foreword by Ray Brassier

5. “One seems forced to choose between the picture of an elephant which rests on a tortoise (What supports the tortoise?) and the picture of a great Hegelian serpent of knowledge with its tail in its mouth (Where does it begin?). Neither will do. For empirical knowledge, like its sophisticated extension, science, is rational, not because it has a foundation but because it is a self-correcting enterprise which can put any claim in jeopardy, though not all at once” (EPM §38). 6. The existence of sense-impressions, writes Sellars, “is postulated on general epistemological or, as Kant would say, transcendental grounds, after reflection on the concept of human knowledge as based on, though not constituted by, the impact of independent reality. It is postulated rather than ‘found’ by careful and discriminating attention” (SM I §22). 7. “[I]f there is a logical dimension in which other empirical propositions rest on observation reports, there is another logical dimension in which the latter rest on the former” (EPM §38). 8. “Kantian ‘categories’ are concepts of logical form where ‘logical’ is to be taken in a broad sense, roughly equivalent to ‘epistemic’. To say of a judging that it has a certain logical form is to classify it and its constituents with respect to their epistemic powers” (KTE §23). 9. “An uttering of ‘p’ which is a primary expression of a belief that-p is not merely an expression of a thinking that-p, but is itself a thinking, i.e., a thinking-out-loud that-p.” (LTC 519). This means that even silently thinking to oneself that p is an instance of languaging for Sellars. 10. “[L]inguistic episodes have not only logical powers but also, and necessarily, matter-of-factual characteristics, e.g., shape, size, color, internal structure, and that they exhibit empirical uniformities both among themselves and in relation to the environment in which they occur” (KTE §27). 11. “[P]ropositional form belongs only in the linguistic and conceptual orders” (NAO III §76). 12. “Comparing a judging with a state of affairs could only be comparing a judging with another judging of the same specific kind, and this would no more be a verification than would checking one copy of today’s Times by reading another” (KTE §26). 13. Thus Rorty is sanguine about the link between liberal democracy and social justice, but choleric about the suggestion that social practices may be significantly shaped by forces and relations of production: “I see this preference for knowledge over hope as repeating the move made by leftist intellectuals who, earlier in the century, got their Hegelianism from Marx rather than Dewey. Marx thought we should be scientific rather than merely utopian—that we should interpret the historical events of our day within a larger theory. Dewey did not. He thought one had to view these events as the protocols of social experiments whose outcomes are unpredictable” (Rorty 1998, 37).

Acknowledgements

I thank Johannes Haag, Dominik Perler, Pirmin Stekeler-Weithofer, and Huw Price for their unconditional and never ending support. It is with your encouragement that this book can now see the light of day – as the revised version of my PhD thesis, defended at the University of Potsdam on 21 November 2018. I also want to express my gratitude to Allie Simmons and Andrew Weckenmann for their guidance in preparing the book. I am indebted in equal measure to the members of the International Sellars Colloquium: Alma Buholzer, Anke Breunig, Christian Barth, Bill deVries, Catherine Legg, David Beisecker, Luca Corti, Dionysis Christias, Fabio Gironi, Michael Hicks, Griffin Klemick, James Hutton, Jeremy Koons, David Landy, Jim O’Shea, Johannes Haag, Marc Joseph, Jonathan Knowles, Kyle Mitchell, David Pereplyotchik, Mahdi Ranaee, Lionel Shapiro, Peter Olen, Carl Sachs, Preston Stovall, Raymond Brassier, Robert Kraut, Ronald Loeffler, Ryan Simonelli, Stefanie Dach, Taka Matsui – I am looking at you. Thank you all for being such excellent interlocutors over the past years. Special thanks to Huw Price and his Cambridge lab group, Annika Böddeling, Georgie Statham, Kyle Mitchell, Mat Simpson, Shyane Siri­ wardena, and to Hugh Mellor for having me in his Serious Metaphysics Group. Markus Wild and Ruth Millikan, our workshop at the University of Basel gave me ample occasion to rethink and defend my views. In particular, regarding the problem that although ‘that which comes earlier may provide a framework for making sense of that which comes later’, the order of being has to accord with if not flow out of the order of understanding. While working on this book I have benefited immensely from countless conversations, comments, and discussions. In particular, I wish to thank Achim Wamßler, Adrian Moore, Alexander Buchinski, Alison Fernandes, Amie Thomasson, Anil Gomes, Anton Friedrich Koch, David Lauer, David Löwenstein, Dina Emundts, Georg Spoo, Gustav Melichar, James Conant, Jan-Philipp Kruse, Karen Koch, Larissa Wallner, Lilja Walliser,

xxii  Acknowledgements Marie Merscher, Nadja El Kassar, Philipp Blum, Robert Pfeiffer, Simon Schüz, Stefan Hagemann, Tal Giladi, Thomas Jussuf Spiegel, and Tobias Rosefeldt. I have also profited greatly from discussions of parts of this book with audiences at the University of Cambridge, the Free University of Berlin, the Humboldt University of Berlin, University of Johannesburg, Université Catholique de Louvain, Università di Padova, the University of Potsdam and The Czech Academy of Sciences, Prague. I thank James DiFrisco for being an invaluably good philosophical sport, always ready to explore uncharted territory with me. And then I am indebted to Dörte Köhler-Seiberth and Hermann Seiberth, Alma Seiberth, Friderike Ahrens, Lea Merk, Milena Alexandra Kraft, Pia Humbert, Sebastien Warshaw, Elena Baltuta and Sergiu Sava – I am grateful for having you around. Luz Christopher Seiberth University of Potsdam, October 2021

Introduction

The Concept of Intentionality This book is about intentionality. It investigates the capacity of consciousness to be about objects. This could also be described as a consciousness’ power to present and represent states of affairs, properties, or abstract contents such as things we might desire or things we remember. The book is organised around the leading question of how the American philosopher Wilfrid Sellars understands the concept of intentionality and the role it plays in his philosophy. Now, since Sellars develops his account of intentionality in large part by interpreting the theoretical philosophy of Immanuel Kant, we can also say that it is a book about the tradition of thought inspired by Kant’s critical philosophy and theory of experience. The core phenomenon this book sets out to explain belongs to the canon of philosophical puzzles. It is a puzzle in the sense that we are, at times, too close to it to even notice it, and if we were to try to explain it, we might easily run out of words because we find it to be so fundamental. The following then introduces a context in which one could quite naturally come to ask questions relevant to the phenomenon of intentionality. On some mornings, when waking up, one finds oneself in a peculiar state in which the state of things seems unclear. This occurs when, just before opening one’s eyes, one is not only unsure where one is or when one is but perhaps also who one is or how old one is. There might be snippets of dialogue, images, or vivid impressions of colours or tastes. And there might be even more complex emotions and impressions, which, once we are fully awake and under the shower, are either gone or whose significance continues to elude us. As soon as we find ourselves orientated in our daily routine, we know where to turn, what to do next. Or we can, at least, rely on our instincts to steer us towards a cup of tea. Looking back, it seems almost ridiculous to think that all those questions were open for even one second. But if we allow ourselves to take that initial moment seriously, we can use it to elucidate a basic and central capacity of consciousness. DOI: 10.4324/9781003221364-1

2  Introduction This is the capacity to represent an object, e.g. the content of a memory, as existing. Such a content might be an object, an event, a result, or a procedure. The possibilities seem boundless. Philosophers distinguish between acts of thinking about existing objects and acts that are about inexistent objects. But they also distinguish between the power of a consciousness to be about something, its intentionality, and its power to intend something, e.g. form a particular intention, have goal or rational desire. One might develop the intention to go and see the Pyramid of Cheops or Khufu, the oldest and largest of the Giza pyramids. And upon arriving there, one might find it to be even larger than one imagined. But in order to even form the intention of finding out about the pyramid’s actual size, one needs to be able to represent it to oneself. In this sense, the phenomenon of intentionality is more fundamental than its etymological sibling ‘intention’. This is why one can think about the lost second book of Aristotle’s Poetics, even if there is no way to actually read it. Here too our ability to have a thought about it is prior to forming the intention about it, i.e. the volition to read it. There is one strand in theories about intentionality that deals with this particular question of intentional inexistence. Theorists wonder how consciousness is capable of presenting a content before itself as objective – e.g. the lost second book of Aristotle’s Poetics or Pegasus – even though these contents clearly do not exist among the observable things of the world we live in. The strand of this book is concerned with and addresses a more fundamental question: How is it possible for a consciousness to be about an object at all, presupposing nothing about the object’s existence or inexistence? Like many canonical philosophical topics, the concept of intentionality has a history. The word itself is of medieval scholastic origin (Jacob 2014; Perler 2004) and was brought to the contemporary philosophical agenda by Franz Brentano’s (1874) Psychology from an Empirical Standpoint. Brentano’s analysis has shown that ‘intentionality’ is inseparable from ‘possibility’, and that questions about our intentional relation to objects presuppose our thinking of their possibility. Sellars, however, takes his cue not from Brentano but rather from Kant. In Kant, he finds a methodology for approaching the question of intentional purport that proceeds by way of abstracting – a way that can be characterised with the help of distinguishing descriptive projects from normative or justificatory projects. Kant does not simply describe our intentional relation to objects but asks a higher-order question, which, according to Sellars, is in the spirit of a presuppositional analysis. The question is: How can one justify the relation between our representations and that which they are about? Sellars thus places his own project in the tradition of transcendental philosophy, which can be understood as a justificatory or normative project. It is characterised by asking how it is possible that the object of thought and the object of experience are related to one another. How does

Introduction 3 it come to be that there is a difference between the two, and how might one give a theory that places these two relata in an intentional relation? The project is justificatory in the sense that it asks: Is there something discoverable that guarantees that these two relata fit together? Since this undertaking involves reflecting on necessary preconditions and presuppositions of our intentional directedness, one might wonder how one even gets it started. The necessary starting point for Sellars is the world of common sense, the world we live our lives in. It is from the midst of our mundane concerns that we abstract and reflect on the common and necessary elements underlying those concerns. Sellars here draws a distinction between the order of being and the order of explanation (cf. BBK §5). An evolutionary biologist interested in the genesis of a biological attribute or character throughout a specific period of time works on reconstructing a reasonable sequence in the order of being (e.g. in the tradition of Tomasello 2014). The philosophical methodology in which Sellars develops his account of the concept of intentionality, however, reflects on what is prior, i.e. constitutive, in the order of explanation. Transcendental philosophy is interested in uncovering the conditions of possibility for intentional directness. In this sense, it is not restricted to a specific domain of philosophical topics. Rather, the problem of intentionality derives its importance from the fact that it is located at the intersection of epistemology, philosophy of language, metaphysics, and, one might also say, pragmatism and the metaphilosophical question of philosophical methodology. Some would further claim that the problem of intentionality is not just the mark of the mental (Crane 1998) but actually the central problem of philosophy, or, at least, that it is central to our understanding of consciousness (Siewert 2017). In any case, what stands out is that philosophers tend to develop a theory of intentionality not by working only in one domain of philosophy but also by defending interrelated views in several overlapping areas of philosophical concern. This is particularly true of the philosophical works of Sellars. Of all the areas in Sellars’ writings one could turn to, this book preserves its unity by mainly concentrating on works that read Sellars as a transcendental philosopher. This interpretive approach is novel in the current state of research on Sellars. A systematic inspection of the history of Sellars-scholarship reveals three waves of reception. The first wave comprises of Sellars’ contemporaries and colleagues, such as Richard Rorty, Willard Van Orman Quine, Daniel Dennett, John Haugeland, Jay Rosenberg, and certainly Roderick Chisholm, Graham Harman, Ausonion Marras – maybe ‘direct interlocutor’ is a different category than ‘first-wave interpreter’? – to name but a few philosophers who have engaged with Sellars’ work during his lifetime. Also, in the 1970s a number of volumes and issues begin to appear that are specifically devoted to themes from the philosophy of Sellars,

4  Introduction e.g. Noûs 7/2 (1973), Castañeda (1975), Delaney et al. (eds.) (1977), and Pitt (1978). The second wave comprises what one might call the generations of students and those who influenced them, e.g. the writings of Ruth Garrett Millikan, Patricia and Paul Churchland, Mark Lance, Robert Kraut, Rebecca Kukla, William A. Rottschaefer, Michael Williams, and Johanna Seibt. These authors seek new ways of engaging with Sellars’ thought and produce work of their own that is formulated in discussion and dissociation from central Sellarsian ideas. Most influential here are, certainly, Robert Brandom (1994, 2000, 2008, 2015) and John McDowell (1994, 2013).1 But it is only another ten years after that a series of monographs starts to appear, each of which, for the first time, aiming to display Sellars’ philosophy in its own right and from an encompassing birds-eye-view. The beginning is marked by the monographs of deVries’ (2005) study, followed by O’Shea (2007), Rosenberg (2007), and Haag (2007). The three latter authors stress the Kantian features of Sellars’ philosophical system. And it is in these works that the methodological dimension of Sellars’ approach to intentionality is, for the first time, explicitly traced back to Kant in a way that has previously been wanting, in e.g. Bremer (1997) and Wolf & Lance (eds.) (2006). It is fair to say that, partly due to these monographs, the third wave in Sellars-research has taken shape and reached us. That is, another decade or so down the road, a new generation of readers has undertaken all sorts of attempts to connect and contrast Sellars’ writings with related fields inside philosophy. Some of these third-wavers, i.e. Raymond Brassier, Dionysis Christias, Luca Corti, Stefanie Dach, Michael Hicks, Griffin Klemick, Jeremy Koons, David Landy, Peter Olen, Carl Sachs, Lionel Shapiro, Preston Stovall, and Ted Parent are collected in Between Two Images. The Manifest and Scientific Conceptions of the Human Being, 50 Years On edited by Gabbani (2012), Sellars and his Legacy edited by O’Shea (2016a), Wilfrid Sellars, Idealism, and Realism: Understanding Psychological Nominalism edited by Reider (2016), Sellars and Contemporary Philosophy edited by Pereplyotchik & Barnbaum (2016), and, more recently, The Legacy of Kant in Sellars and Meillassoux: Analytic and Continental Kantianism edited by Gironi (2017) as well as Sellars and the History of Modern Philosophy edited by Corti & Nunziante (2018), Wilfrid Sellars and TwentiethCentury Philosophy edited by Brandt & Breunig (2019). And more than that, one has begun reading Sellars as a modern classic (as is reflected in the canonical way of citing his works, see Sicha (ed.) (2002, 565 ff.) and O’Shea (ed.) (2016a)). As Wanderer & Levine (2013) maintain, it is time to interpret Sellars outside of the confines of his membership in the ‘Pittsburgh School of Philosophy’, i.e. in other ways than Maher (2012) and Sachs (2019b).

Introduction 5 From this overview, it is apparent that three areas of Sellars scholarship are underrepresented. First, what is missing is a study that clarifies the relationship between Sellars’ conception of philosophical methodology and his account of intentionality. Second, new Sellars-critical developments in the field of philosophical expressivism, most notably by Huw Price (2011, 2013, 2015, 2017), but also Rouse (2015), have occurred that necessitate a re-evaluation of the relationship between Sellars’ nominalistic, expressivist leanings and his commitments to a scientific form of realism. Third, and what can be seen, lastly, as an additional requirement, none of the studies up to the present day discuss the relationship between Sellars’ transcendental methodology and the dominant forms of philosophical method (but see Sachs 2015; Christias 2018; Dach 2018; Hicks 2020, and with regard to Sellars’ early work, Olen 2016). This requirement is ever more urgent in light of the fact that the Kantian brand of philosophical methodology is largely absent from recent metaphilosophical monographs on philosophical methodologies – edited by Daly (2010), Haug (2013), Schönwälder-Kuntze (2015), Cappelen et al. (2016), and D’Oro & Overgaard (2017). This book responds to the desiderata outlined above by pursuing the following programme. Chapter 1 prepares the ground for understanding Sellars’ exegesis of Kant’s theoretical philosophy by substantiating the hypothesis that Sellars’ conception of intentionality is at heart based on a Kantian transcendental method. This involves distinguishing three levels of abstractions, delineating the role of presuppositional analysis, and identifying necessary preconditions and presuppositions of empirical knowledge. The Kantian insight that objects of empirical knowledge have to conform to logically synthetic universal principles is presented in a fresh way in contrast to traditional conceptions of descriptive metaphysics. As a decidedly normative project, Sellars’ transcendental methodology is thereby put on the map of received philosophical methods. Chapter 2 develops this idea in more detail by a reconstructive discussion of Sellars’ Kant exegesis. One result of this discussion is that Sellars’ engagement with Kant’s thought evidences a dissatisfaction with the solutions Kant has to offer on the question of how far we are in contact with the world. That is the case because Kant’s transcendental idealism states that we are, ultimately, only in contact with appearances. In Anglophone philosophy, this ‘claim’ is not understood as a truism, as the thesis that all our knowledge claims refer to possible experience. Rather, it is taken to be a dissatisfying result, one which leads Sellars to articulate his own conception of intentionality in the course of delineating the details of perceptual experience as a theory that builds on Kant’s thought, but also aims to extend it, amend it, and partially reject it. Chapter 3 presents an encompassing interpretation of Sellars’ unified theory of perceptual experience. This includes addressing the question of how we are related to a world of obtainable represented contents.

6  Introduction Answering this question amounts to an account of the sense in which we can say that the world is given to us, but at the same time is structured by our taking or perceiving it. It involves clarifying Sellars’ notion of nature and understanding what it means to claim, transcendentally speaking, that nature is given to us ‘in thought’ as well as ‘in experience’ (with ‘nature’ here being used synonymously with ‘the world’ or ‘objects’). One aim is to clarify in what sense nature can be said to be disclosed ‘through our actions’, and, in the larger sense, as ‘emerging throughout the ongoing scientific enterprise’, an enterprise that self-corrects its conjectures about the underlying structure of nature. Although the role of conscious, intentional action plays an important and constitutive role in epistemic contexts, it is shown to be nonetheless subordinate to the question of how the concept of intentionality is to be understood as operating on top of unconscious relations between objects and response-tokens they occasion in us. A thesis central to this chapter is that intentional agency presupposes the idea that in forming and carrying out an intention, the relevant intentional content or object has to be already available in a qualified sense. The chapter is brought to a close with a brief presentation of Sellars’ reasons for rejecting the Kantian theory of experience as a framework within which we can explain the status of our present and future epistemological situation. The issue here concerns the challenge to specify what it is that Sellars is saying that Kant could not also say. The response to this challenge is codified in the thesis of the instability of the Manifest Image. Chapter 4 develops an interpretation of Sellars’ own positive account of intentionality. It has two steps. A first step is to clarify the methodological reasons Sellars has for developing a non-relational account of the way in which acts of representing are related to their contents. A second step is to reconstruct and develop the motives for Sellars’ project of articulating a transcendental phenomenalism, as it will be called in Chapter 5. The aim here is to understand Sellars’ projective account of why we can be said to be in direct contact with the world, despite the non-relational dimension of linguistic intentionality. As it will turn out, the semantic dimension of intentional ways in which we speak about the world is a matter of functionally classifying individual contents. Chapter 5 critically discusses the understanding of a relational or vertical conception of intentionality that is ingrained in the account Sellars gives of our languagings’ double nature. In addition to the horizontal way of understanding how we can explain the non-relationality of our knowledge claims about the world, there’s a further vertical dimension to be accounted for. This is a claim about a direction of fit which rests on the two-part thesis (1) that our claims about the world are themselves part of the world and (2) that our perceptual responses when they are considered as natural-linguistic objects directly relate to objects in the world.

Introduction 7 The thrust of this line of thought is captured in the argument from double nature (Chapter 5.1) which aims at reconstructing Sellars’ reasons for claiming that the utterances (or languagings) we produce in the course of perceptually responding to external impingements possess not only semantical form but also empirical form. This involves showing how Sellars’ non-relational conception of intentionality can be read as allowing him to give an account of the place of thoughts and mindedness in nature. To that effect, a second argument is reconstructed, the argument from immanence (Chapter 5.2). This involves discussing whether Sellars can have his theory live up to the requirement that it is articulated from a standpoint that itself can be said to be meaningfully part of the world it claims to develop a theory of knowledge for. This prepares the ground for his claim that languagings, when considered with regard to their empirical form, can be taken to form pictures of the objects they are isomorphically related to. This amounts to a bifurcation in the concept of truth. Sellars does not propose to replace the concept of truth with that of warranted assertibility, according to which all questions about the truth of a statement are settled by what is, either at a personal or at a communal level, taken to be warrantably assertible. What position does he propose instead? According to Sellars there is (1) a conception of truth internal to the conceptual structure we currently live in, and (2) a conception of truth in terms of adequacy. The internal conception, internal to a non-relational conception of intentionality (understood in terms of semantic assertibility) is to be supplemented with a dynamic conception of truth that can do justice to the fact that our conceptual system as a whole evolves and is just one in a sequence of such systems throughout time. To account for the change that our conceptual system as a whole undergoes, and to explain how that forces us to think of a stratum of interrelated processes underlying such a conceptual change, Sellars introduces the notion of theoretically postulated ‘counterpart objects’. This is reconstructed as the argument from counterparts (Chapter 5.3). Its aim is to establish that we can, with the help of the methodology of analogical concept formation, gain a more and more adequate understanding of these counterpart-objects. Relevant here is that the notion of an increase of adequacy presupposes a non-­semantical picturing relation and the double nature of ‘languagings’ serving as an interface to our epistemic engagements with the world. The thesis defended throughout this book is that Sellars’ commitment to a transcendental method of analysis is the overarching explanatory principle for understanding his account of intentional reference to a world we are a part of. Here it matters that Sellars’ engagement with the Kantian finding that our knowledge (Erkenntnis) always pertains to appearances does not consist in a criticism of this Kantian result in a sense that would turn into either dogmatic idealism or a naive and thus dogmatic realism. Rather, it consists in an attempt to ask in what sense

8  Introduction this can be extended by a perspective that takes the forms of knowledge production in theoretical sciences to add something important to our knowledge of our shared life-world. The methodology Sellars finds in Kant leads him to endorse what I will term a ‘transcendental phenomenalism’ that is by and large Kantian. This is a phenomenalism that at heart is an idealism, though it presupposes and is interdependent with a transcendentally motivated realism. This transcendental phenomenalism underlies not only Sellars’ account of the concept of ‘objects in nature’ but also his understanding of the question of how we can take ourselves to be adequately picturing them better and better, as we progress in refining our development of theories of nature. Chapter 6 addresses the tension inherent in this conception head on and makes it the topic of an extended engagement with objections from pragmatism and expressivism. This involves reconstructing and assessing longstanding as well as more recent objections from Richard Rorty and Huw Price, respectively. One focus of the objections is to challenge Sellars’ commitment to scientific realism in light of seemingly incompatible commitments to a non-reductive phenomenalism. A subsequent discussion evaluates the objections and their consequences for an interpretation of Sellars’ overall position. The Conclusion reflects on the course of the argument of the book and ties it back into the initial question of how Sellars understands the concept of intentionality. It then ventures into a quick survey of further desiderata of research.

Note 1. The reader might wonder why this book does not engage more closely with the work of these authors on the philosophy of Sellars. There are a number of reasons for this. The first is that a recent study has undertaken just that, see Maher (2012). A second reason is that neither of them discusses the notion of ‘picturing’ which is so central to the interpretation of Sellars developed in the following. Brandom (2015, 16) does not see in any way how to constructively engage with it. A third reason is, in the case of McDowell, that his interpretation of Sellars has undergone substantial revisions over the last 20 years. To track those would have meant to write a book with an altogether different focus. However, for a ‘compare and contrast’ discussion, see Haag (2014, 2017), and more recently Haag (2019) who defends the Sellarsian account against an alleged blind-spot. But see also Rosenberg (2007, Chapter 13) and Redding (2012). A fourth reason is that both Brandom and McDowell consider Sellars to operate with a scientific prejudice, McDowell (2000, 2013) perhaps more so than Brandom. According to this view, Sellars’ motivation for the claim that the Scientific will supersede the Manifest Image framework rests on a dogma. My discussion of the status of Sellars’ commitment to scientific realism in Chapters 3.8, 6.1.3, and 6.3 can, in part, be read as a response to this allegation. Cf. Shapiro (2013) and Haag (2019) for nuanced responses.

1

Transcendental Methodology

1.1 Transcendental Methodology and the Role of Analytic Truths This chapter introduces the framework in terms of which Sellars develops his account of intentionality. As we have seen in the Introduction, the problem of the nature of the intentionality of reference to objects can be put, as an assumption that there are objects, as a question about the constitution of the framework in which we encounter them. Only at the end of such an inquiry can we turn to the question of what specific objects there are. In this sense, our ability to address the ontological question depends on the methodological work to be carried out in specifying the necessary structural elements of a theory of intentionality. So what is the methodology behind Sellars’ project in terms of which he accounts for intentionality? It is a method that advances by suspending or bracketing the commonsensical conception of objects. The unity of the transcendental method lies in considering this concept of the purport – which our intuitive representings aspire to have towards empirical objects – as itself an open question.1 Objective reference or reference to objects – intuitions with a content that is ‘about’ the objects in the real – is thus the fundamental systematic problem that the transcendental method is designed to answer. 2 While a descriptive approach to the intentionality inherent in perceptual experience may begin by assuming that the concepts of meaning, truth, and reference are already innocuously in play, transcendental philosophy seeks to demonstrate our entitlement to their employment. One can characterise the methodology generally in terms of this aim: To show how they should figure in our concept of an ‘object of experience’. 3 To this end, we begin by looking at the opening passages of Sellars’ paper ›Kant’s Theory of Experience‹ which paradigmatically display a transcendental reflection on the activity of developing a theory of knowledge. At the centre of such an investigation is the concept of judging that an empirical state of affairs obtains. The first evidence for such a reading DOI: 10.4324/9781003221364-2

10  Transcendental Methodology is to be found in KTE. It gives the core program and is one of Sellars’ earlier papers on Kant. Sellars begins by way of a commentary on the statement that ‘an empirical judging is true if the state of affairs it represents is one which did, does or will obtain, i.e. was, is or will be actual’. This scrutiny must aim at clarifying the concepts of an empirical judging, of truth, of a state of affairs, and of what it is for a state of affairs to obtain or be actual. This Kant proceeds to do with important, indeed, dramatic results. […] [F]rom the Kantian point of view, the above [classes of] concepts pair up in an interesting way: judging with state of affairs, and truth with actuality. Indeed to say that they pair up is to understate the closeness of their relationships. For, Kant argues, in effect, that the pairs turn out, on close examination, to be identities. (KTE §6, §8) Sellars’ interest in Kant is arguably abstract. The focus is not on a discussion of the nature of specific knowledge claims. Nor is it on the nature of one particular theory, say, from a particular time in the history of philosophy. Rather, the productive philosophical work takes place on a prior level. Sellars treats Kant as a philosopher with a generic aim: To show the essential elements that any theory of knowledge has to comprise and account for. In other words, Sellars reads Kant as a transcendental philosopher inquiring into the necessary and elementary structure of any account of intentionality. In addition to interpreting Kant this way, Sellars clearly endorses the fundamentals and defends them against alternatives. So to Sellars, Kant formulates a theory of knowledge that does not simply take the common sense framework as a given but carries out a presuppositional analysis with regard to its constitution. But where does such an analysis begin? With regard to the question of where to start that process, Sellars joins the Kantian outlook of turning to the subject of epistemically significant states. A telling passage illustrates this self-reflexive dimension in the meta-theoretical approach to a theory of intentionality. Such a theory is about the constitution of that to which an intentional relation points. But then the epistemological categories enjoy a special kind of explanatory priority over the ontological ones. In fact: The core of Kant’s “epistemological turn” is the claim that the distinction between epistemic and ontological categories is an illusion. All so called ontological categories are in fact epistemic. (KTE §9) A theory of intentionality as a theory of empirical knowledge has to say something about what there is, how it can be known, and from where

Transcendental Methodology 11 to approach it. So, transcendental philosophy starts by reflecting on a central component of consciousness, the activity of judging: The central theme of the Analytic is that unless one is clear about what it is to judge, one is doomed to remain in the labyrinth of traditional metaphysics. On the other hand, to be clear about what it is to judge is to have Ariadne’s thread in one’s hand. (KTE §7) A manifest common sense conception of knowledge is constructed in terms of tables and everyday objects. But this starting point is unavailable to the reflexive transcendental analysis of how a conception of knowledge can be devised. In a specific sense, it would already presuppose the subject aspiring to know something to be in contact with objects. So, when Sellars reads the Kantian turn as concerning the status of the ontological categories, he underlines his agreement with an approach that starts with the subject. The starting point might still be ordinary objects, but Sellars’ involvement with this Kantian claim aims beyond that framework. It does not merely invoke that starting point but marks a re-interpretation. The ontological categories, telling us what things have ‘being’, are, according to the Kantian perspective, secondary to the processes of knowing. Epistemic access is first in this perspective. Sellars paraphrases the conclusion as an insight for which he aims to give a particular interpretation: He reads the categories as items we can interpret with regard to their functioning in epistemology. Categories thus become abstract rules for the classification of acts of judgement. The specifics of the interpretation Sellars gives to the basic Kantian idea of attributing categories a central position come out in the reformulation of epistemology as transcendental linguistics. How does Sellars develop that theme of the epistemological turn? To Sellars, Kant’s operating on this initial statement consists in investigating and laying out the concept of empirical knowledge: [The epistemic categories] are “unified” by the concept of empirical knowledge because they are simply constituent moments of this one complex concept […]. (KTE §9) To Sellars, the operation of articulating our conception of what it is to know something of and reliably about the world consists in developing our concept of an object of empirical knowledge. It is in this sense that transcendental philosophy takes place on a meta(-conceptual) level.4 The main concern is with the elements that go into the concept of knowledge. And the concept of ›knowledge‹ has complexity in a specific sense.

12  Transcendental Methodology Unlike cases where we may want to know specific historical matter-­ of-factual details, e.g. ’Are there apples in the fridge now?‘, the transcendental conception of the concept of knowledge concerns requirements of knowledge in a very abstract and generalisable sense. To Kant, 5 and here Sellars joins him, transcendental knowledge is a specific kind of knowledge about knowledge. And since the starting point is the subject in its local epistemic aspirations, the meta-reflection on the activities of the subject itself concerns the subject’s abilities and the subject’s consciousness. The elements of theory-building are to be found in the transcendental philosopher’s reflection on the elements of mental activity. With regard to the Kantian origin of that kind of a reflection, Sellars thus can write: The task of “transcendental logic” is to explicate the concept of a mind that gains knowledge of the world of which it is a part. (KTE §13) And to have a conception of such a concept of a mind or consciousness gaining knowledge of its world is to construe – in modern terms – a theory of intentionality by reflecting on the necessary elements any such theory has to comprise. The first central elements coming into view, at the centre of the specific articulation given by the transcendental method, are meta-classificatory acts. Such acts consist in interpreting the elementary activities of consciousness in search of a coherent theory of knowledge. The project of formulating such a theory by way of reflecting basic mental activities qualifies as open in the following sense. All viable alternative construals coming into consideration are tested, discussed, and – where applicable – endorsed or deflected. To be clear about this specification of the status to which transcendental philosophy thereby aspires: A transcendental demonstration of the coherence of a conception of intentionality does not yield an eternally fixed, a time-invariant theory. The idea that the approach transcendentally investigates the presuppositions of all viable theories just means to say that unless a convincing alternative is brought into view, the conception is treated as sound as it stands.6 The transcendental method entails that we can generate synthetic truths by an analysis of our epistemic capacities. The type of knowledge the transcendental philosopher is after can be termed, with a nod to Kant, as knowledge in the domain of ‘transcendental logic’. It consists of ›analytic knowledge about synthetic knowledge‹ (KTE §10 n2). What kind of truth does such knowledge comprise? Sellars writes: It is obvious to the beginning student that the truths of “transcendental logic” cannot themselves be “synthetic a priori.” If they were, then any transcendental demonstration that objects of empirical

Transcendental Methodology 13 knowledge conform to synthetic universal principles in the modality of necessity would be question-begging. (KTE §10) In considering what meta-theoretical facts the transcendental analysis of our conception of knowledge brings out, Sellars remarks that those, on pain of regress, cannot have the status of synthetic a priori truths. In other words, such a conception would start by presupposing a structure from which a demonstration would follow that our empirical knowledge conforms to that structure. How are we to understand the status of these truths? It must in a tough sense be an analytic truth that objects of empirical knowledge conform to logically synthetic universal principles. It must, however, also be an illuminating analytic truth, far removed from the trivialities established by the unpacking of ‘body’ into ‘extended substance’ and ‘brother’ into ‘male sibling’.7 (KTE §10) This disanalogy highlights the difference between the methodology of classical conceptual analysis and the meta-theoretical reflection on the elements of theory-construction characteristic of the transcendental method.8 In what sense do they differ? Conceptual analysis uses (in the sense that it assumes it as ‘already given’) the framework of common sense of philosophical (objectual) categories operating on an unspecified conception of reference to an object. In contrast to this, the transcendental methodology that Sellars advocates may begin with local ordinary descriptive knowledge claims but abstracts from them to raise questions about the very idea of reference überhaupt.9 It is concerned with the pure accusative of reference, reference to something, some x. The transcendental project of inquiry is, systematically speaking, prior in the sense of analysing what grounds the notion of representational purport as such. Another way of characterising this particular interest, as Sellars remarks in a footnote to the quoted passage, is to see it as ›analytic knowledge about synthetic knowledge‹. The idea here is that something general can be brought out about all synthetic forms of knowing, something general enough to be analytically true of all kinds of knowledge of matter of fact. Two points about the relevant kind of analytic truth: First, we need to exclude a misleading interpretation of them. Analytic truths as they figure in transcendental reflections on the aspects that go into the formulation of a theory of intentionality are not supposedly eternal truths. Rather the notion of such analytic truths goes together with the notion of dialectically defendable theories. At the centre of attention, there are truths that are treated as invariant without being eternal. And while

14  Transcendental Methodology their specification would amount to a class of sentences – for theories when stated explicitly are sets of sentences –, they are different in nature from analytic truths that are specified with regard to their content, such as ‘2 + 2 = 4’. Analytic truths in the context of formulating a theory of intentionality can, in principle, be open-endedly reformulated. That is, there might be veritable alternatives to the shape in which we currently cash them out. But whenever one given formulation can, in fact, be discounted, it shows that this particular version is either no real alternative or merely a further form of saying the same, in disguised forms.10 So the notion of analytic truths pertains to a conception of meta-theoretically robustness: A form of analytically derivable knowledge (something true) that is true in the sense of expressing a precondition of frameworks that can yield first-order empirical knowledge. When Sellars works on discarding alternatives, this amounts to discounting attempts at delineating conceptions of knowledge which differ in their structure. Such are the skeptic and the idealist positions. Second, what is being analysed when we address the question of what analytic truths are? For a start, in the passage, Sellars tells us that he agrees with Kant that the status of analytic truths is that of a specific kind of meta-knowledge. What could that mean? I interpret the point of the disanalogy to be that of pointing to a difference in levels. Where conceptual analysis targets the relations and nature of the relata within a given conceptual scheme, the transcendental reflection targets analytic truths, enabling the construction of any coherent conceptual scheme. So Sellars maintains: It must in a tough sense be an analytic truth that objects of empirical knowledge conform to logically synthetic universal principles. (KTE §10) The transcendental analysis of consciousness’ activity brings out second-­ order knowledge. If this is what analytic knowledge about synthetic knowledge consists in, the item under consideration at that second level is not about any specific matter-of-factual object. This is important to register. Second-order knowing paradigmatically concerns presuppositions or structural facts. Knowledge about them amounts to truths that we can show by way of analysis. We can bring it out not just once, incidentally, or historically contingently. And it isn’t the case that just one person can bring it out. Rather, it is a kind of truth that anyone will bring out and follow, whoever undertakes the relevant transcendental reflection on the meta-theoretical elementary processes of construing a theory of intentionality. The kind of truths on that construal bear the characteristics of being binding for any finite rational being gaining knowledge of a world

Transcendental Methodology 15 of which it is a part. All infinite beings, in Sellars’ view, are analogically extensions from our starting point.11 But what makes the analytic truths be truths in a tough sense? It is the sense that they make their way into any formulation of a theory of empirical knowledge. So they are neither historically contingent in the motley way the French language developed out of the Latin language. Nor are they ‘eternal’.12 That is to say, they are non-eternal in the sense that what they ‘really’ are, and how their status is to be accounted for, is perhaps disputable, but only in so much as the possibility of a more coherent construal is something that can in principle be excluded. If the status of the relevant analytic truths considered here is meta-­semantical, it concerns the conditions of possibility of the obtaining of relations among further conceptual relations, truths about the status of conceptual dependencies, and presuppositions. [1] Kant is not seeking to prove that there is empirical knowledge, but only to show that the concept is a coherent one and [2] that it is such as to rule out the possibility that there could be empirical knowledge not implicitly of the form [:] ‘such and such a state of affairs belongs to a coherent system of states of affairs of which my perceptual experiences are a part’. (KTE §11) According to Sellars’ construal, the aim of the Kantian programme is to show that we need an understanding of a system of state of affairs as one that also comprises the knowing agent. Where knowledge is to be a mental state of matter-of-factual affairs, the knowing subject itself has to be part of that world. Two key findings that will come into view on the basis of this in the next chapter are examples of such analytic truths: The indispensability of (1) demonstrative reference and of (2) a categorial frame, such that any conceptual system that yields empirical knowledge will necessarily comprise these elements (cf. Chapter 3). What Kant articulates, according to Sellars, is a close connection between the concept of empirical knowledge and the category of actuality, that is, the classificatory rules we operate on when articulating our claims about what specific states of affairs obtain. In a related passage, in a slightly later paper on Kant’s epistemology, Sellars unfolds this specific thought in greater detail: In epistemology, which, as concerned with good thinking in its various modes, is a fortiori concerned with thinking as such, this general principle becomes the epistemic principle that any true content of thought, e.g. that Socrates is wise, must, in principle, be an element in a certain kind of large context, e.g. an I thinks the true thought of a world in which Socrates is wise.

16  Transcendental Methodology Roughly, the form of empirical knowledge is: an I thinking (however schematically) the thought of a [spatio-]temporal system of states of affairs, to which any actual system of affairs belongs. (I §9) The reflection on the construal of a theory of contact with a world (of which a thinking being is a part in the attempt of gaining knowledge of it) brings out the transcendental dimensions in the concept of a knowledge-­ gaining agent. The reflection concentrates on the I, on the activity of consciousness. Any plausible theory of empirical knowledge has to take into account that combining, synthesising, activity of the thinking self. Sellars writes accordingly: It is not always realized that the philosophical, or higher order, claim that experience requires synthetic necessary principles is, for Kant, an analytic truth arrived at by what we would call ‘conceptual’ analysis, but which Kant calls ‘Erklaerung’ [sic]. (SM II §53) Sellars launches an alternative criticism for which he finds support in the Kantian text. The idea is that the subject of the transcendental account has to do the constructive work first, the work of combining items that seemingly belong together independently of the subject’s activity. But if there is any zest to the idea that the transcendental activity is prior to any world-directed activity, it is here, in the combination of aspects into a coherent view, that this shows: Thus, in the Transcendental Analytics, the above unrestricted principle about thinking provides the clue to the form of the phenomenal world. This world is a represented world, and Space and Time, and the Categories are its “forms”. … They are but moments (to use Hegel’s term) in its larger form: An I thinking a complex spatial-temporal-causal system of states of affairs — including say, α and β (the synthetic unity of apperception). From this synthetic unity, as Kant points out, it follows analytically that The I which thinks α is identical with the I which thinks β (the analytic unity of apperception). (I §10) What is the ‘form’ that Sellars sees Kant as bringing out here? It seems that it has to do with a performance, with the act of representing a content as spatio-temporally ordered, as falling into a specific class or category, as being of a specific functional use. The main idea behind all

Transcendental Methodology 17 this is that any successful account of the notion of ‘empirical knowledge of objects of experience’ will have to incorporate this combining activity (i.e. self-ascription). Thus, to the transcendental philosopher, the concept of an object of experience and the concept of knowledge of such an object already involves an activity on the part of the subject to which such a knowledge is (self-)ascribed. With this, we have a first outline of how transcendental analysis concerns the nature of the very possibility of knowledge.

1.2  Presuppositional Analysis So one aspect of the transcendental project of inquiry is that it is not only concerned with a meta-perspective on first-order knowledge claims but with questions of the following type: What makes possible the reference to objects? And it is these how possible questions that lead transcendental philosophy to elaborate further fundamental constraints any conception of knowledge has to comply with. This brings out three distinguishing features. The first feature that sufficiently characterises transcendental knowledge concerns its claim to an extreme generality: The analysis aims to bring out constraints that are binding for all finite cognisant beings alike. The second feature concerns the kind of status such an analysis lays claim to; it aims to uncover necessary constraints in such a way that their validity is without possible alternatives. The third feature is that the transcendental analysis uncovers how the target notion, the concept of empirical knowledge, is actually constrained by being knowledge as finite rational beings are said to possess. Traditionally, the metaphorical limiting concept of God’s knowledge comes in here. Why so? Because God’s infinite powers allow begetting objects by thinking alone. And here, the intentional relation is trivially satisfied because God, in producing the objects, is surely also in touch with them. Not so in the case of finite rational beings. Beings like us can come to know only objects that are determinate enough to be identified by empirical means. Thus, our intentionality has to be explained also with reference to the forms of our sensibility and the forms of our intuition. We do not have the power to be in touch with objects through thinking alone. We also require sensible input, and somehow we need to bring our conceptual faculties to bear on that input in a way that allows us to qualify it in important respects. It is this notion that goes hand in hand with the idea that our finitude amounts to an extra constraint in the articulation of the aspects of our intentionality. One further way of specifying that constraint shows up when we look at the fact that, from the highest point of abstraction, it is surely the case that the general logic is binding for any rational being, finite and infinite alike. That is so because at that point of abstraction, there

18  Transcendental Methodology is no content at all. It is only when we take further steps in specifying how far a content can be given to consciousness that the constraints come into view. And while we don’t have any clear conception of what the intentionality of infinite beings must be structured like (save in some analogical construal of such consciousness), we can explicate what constraints our own thinking is bound by. And this third constraint ushers in a sensible condition our thoughts are grounded in. This is codified in the Kantian idea that our sensibility exercises a transcendental function in that the forms of time and space constrain our reference to objects. All objects we experience, we experience as somewhere and at some time. It is characteristic of transcendental philosophy that it strives to go beyond a descriptive approach to our knowledge claims. A presuppositional analysis of the most general and abstract constraints which underlie any conception of knowing has, therefore, its echo in the fact that such an approach is prepared to reflexively call into question any presupposition used in mere descriptions. So what kind of necessary presuppositions are on the table here? In a bottom-up construal that gradually abstracts from the specifics of the epistemological scenario, we have a subject putting forward a knowledge claim. In ordinary knowledge claims, an object is taken to be present and a subject is taken to be equipped with perceptual capacities in a way that yields knowledge. The reflection on the latter presuppositions amounts, in Kantian terminology, to an investigation into the transcendental psychology of the subject. This explains why Sellars describes his projects as defending a version of Kant’s transcendental psychology (KTE §39–44). In the terms our analysis endorsed so far, there is nothing peculiar about the project of analytically bringing out the most fundamental constraints that the concept of consciousness gaining knowledge is bound by. It is only where the program of uncovering such constraints is misconstrued as mere speculation that Sellars’ endorsement of this fundamental Kantian commitment looks outdated and dogmatic.13 The first piece of evidence for the hypothesis that Sellars aims to defend a version of transcendental psychology in Kant’s sense is to be found in the following passage. [O]nce it is appreciated that Kant’s account of the conceptual structures involved in experience can be given a linguistic turn and, purged of the commitment to innateness to which given his historical setting he was inevitably led, his theory can be seen to add essential elements to an analytic account of the resources a language must have to be the bearer of empirical meaning, empirical truth, and, to bring things to their proper focus, empirical knowledge. (OAPK §31)

Transcendental Methodology 19 Here Sellars openly characterises his own project as a linguistic version of the Kantian one. The aim of Sellars’ project is to bring out the necessary resources for languages to yield empirical knowledge. Importantly, Sellars here does not target any particular language, but the presuppositions binding for any conceptual system that functions as yielding empirical knowledge of matters of fact.14 One observation is that Sellars characterises his project as ›a linguistic version of Kant’s position‹ (KTE §39). By ‘linguistic’ Sellars means that most of what Kant was after can be cashed out in terms of an explication that reformulates the Kantian ideas in a post-linguistic turn fashion.15 To construe the concepts of meaning, truth, and knowledge as metalinguistic concepts pertaining to linguistic behaviour (and dispositions to behave) involves construing [… them] as governed by ought-to-bes which are actualized as uniformities by the training that transmits language from generation to generation. (KTE §40) The presuppositional analysis, according to this formulation, aims to bring out the most abstract aspects that any language has to comprise to yield meaning, truth, and knowledge. But how are we to understand this precisely? To reformulate the nature and status of the constraints that are the target here, we will continue our development of the concept of a transcendental method by invoking the notion of levels of abstraction.

1.3  Levels of Abstraction When Sellars comments on KTE, he thereby joins the intellectual tradition of German Idealism, which considers the concept of intentionality as analysable and not as primitive. This tradition contrasts with more contemporary (perhaps naturalistically inclined) positions that find intentionality to be a basic feature of thought, the mark of the mental, or of intelligent organism’s mindedness – about which there is little or nothing to say. Such positions maintain that intentionality is a concept about which we cannot really say anything insightful without marching in a circle (e.g. Rosen 2001). A good way to introduce the motivation behind distinguishing levels of abstraction is to return to the difference between descriptive approaches to knowledge and justificatory approaches. For the latter, to give a theory of knowledge amounts to legitimising it by articulating necessary constraints that are binding for all finite beings. It is a distinguishing feature of their methodology that two more features come to bear in its application: (1) A claim to extreme generality and (2) the fact that the unique feature of establishing constraints aspire to be without alternatives.

20  Transcendental Methodology A philosopher with conservative inklings might oppose that to posit the existence of a meta-level itself doesn’t bring much gain, as any next theory may comment on an opposing view without being in any sense ‘higher or prior’ to the first. To avert such an objection, it is important that the second-order meta-position concerns the presuppositions of the first-order claim. And that it shows that the presuppositions themselves are not adjustable at will. So if the transcendental meta-knowledge is to carry any weight, it is to yield knowledge that is binding downstream for all opposing theories alike (cf. Haag 2017, 26). In that sense, transcendental knowledge concerns constraints that are binding for all theories alike, explaining the intentionality of finite rational beings in their aspiration towards empirical knowledge. So the relevant kind of knowledge that we call transcendental has the structure: Each type of knowledge necessarily has the following form … Each type of meaning necessarily has the following form … Each type of truth necessarily has the following form … And it is decisive that the kind of transcendental philosophy Sellars advances operates with the aspiration to be the only theory that articulates the right kind of form. But if we try to characterise the status such a claim of transcendental philosophy is intended to have, we need to distinguish three levels on which one can make such claims. The first level is that of empirical matter-of-factual statements. Let that be level O, or the object-level. The second level, M1, abstracts from features constituting this first level. In a sense, this is the level most philosophical discussions or commentaries take place on. On level M1, we may make statements about such statements as ‘the table is brown’. And we say it ascribes a property to a substance. Here the existence of tables and their properties are assumed to be in play. And we take the referential framework in which we make such first-order knowledge claims for granted. It is only on a second meta-level, M2 that we come to investigate, in the sense of calling into question, the fundamental assumptions that are already operative on M1. The distinctions are reflected in the notion of an object. M2 does not presuppose anything about the object of concern but takes the object to be a mere placeholder, an empty accusative.16 In light of this distinction, it becomes clear that the appropriate target for a theory of intentionality is to clarify the downstream dependency inherent in this layering.17 On what grounds should we subscribe to a methodology that distinguishes levels of abstraction? The strongest argument in favour of such a methodology holds that it allows us to avoid empty or pseudo-debates (cf. Haag 2012b, 2016). And it allows drawing important distinctions between kinds of concepts that any sound theory of intentionality will have to include. It helps distinguishing first-level, descriptive, concepts

Transcendental Methodology 21 and the notion of demonstrative reference built into them from second-­ order (M1) concepts such as ‘category’, ‘means’, ‘object’, or ‘process’. The undertone in defence of this is: Without the distinctions in play, so the claim, grasp of the area of concern remains confused, and the dispute starts from wrong assumptions. The general scheme defending the method of distinguishing levels of abstraction then consists in dividing at least these two classes of concepts. The paradigm of what jobs concepts can have then no longer needs to be descriptive usage,18 but further uses come into view as well. What kind of uses? For instance, concepts for coordination and subordination of other concepts have a different function. Their functional task may be explicitly meta-linguistic in that it concerns the articulation or explication of the appropriate usage and functions of language. Every grammar for a natural language requires such meta-linguistic terms. A thesis central to Sellars’ view on language is that a number of concepts are covertly meta-linguistic (cf. SM III, LTC, and SRLG). That is, their job is to regulate our uptake and communicate an aspect of their function without being descriptive in any direct sense. Now the idea is that the philosophically curious and interesting concepts are often meta-linguistic concepts, in that their job is not to describe an aspect of the world, but to allow us to highlight and understand a specific function (e.g. of concepts such as ‘thought’ or ‘language’). According to Sellars’ views of language, the concept ‘means’ itself is meta-linguistic in that it conveys a functional dimension of concepts that we connect by using the concept ‘means’ between them: The claim “rot’ means rouge’ on that account is covertly meta-linguistic and really concerns the functioning of both terms, the single scare quoted and the italicised one (cf. Brandom 2016; and Chapter 4). A further reason for applying the method of distinguishing levels of abstraction is that a number of contemporary interpreters of Sellars develop their views on intentionality on both a meta-level and on a metameta-level. This highlights the point that, while it is one thing to distinguish methodologically different levels of claims (O, M1, and M2), it is yet another to interpret the relevance, status, and significance of claims made on the respective levels.19 In the following, we will elucidate the way in which Sellars’ Kant exegesis can fruitfully be interpreted with the help of this distinction of levels (Chapter 2). And we will articulate the justificatory dimensions that hold between them. The core idea guiding our discussion is that what is established as a requirement on M2 will influence how things can be on levels M1 and O, respectively. The same point can be expressed by addressing the question of how to explain the relation between the method of transcendental analysis and the notion of exploring a conceptual framework – the latter being a summary term for various methodological performances undertaken in philosophical sub-disciplines. For in philosophical analyses such as carried

22  Transcendental Methodology out (on M1) by phenomenology or conceptual analysis, we explore the logic or structure of a particular conceptual framework (PH 92). By placing philosophical methodologies on M1, we have a way of conceiving of different strands of work done by philosophical methodologies in Sellars’ thinking over against the method of distinguishing levels of abstraction. But what about the relation between the method of transcendental analysis and Sellars’ other central methodological commitment? Namely that philosophical problems – especially including the problem of perception and intentionality – must be addressed in terms of how we articulate ‘conceptual frameworks’ (cf. PH, SM V, TE, SSIS). 20 Analysing the logic of perceptual and epistemological notions in terms of their conceptual framework seems, at least in part, different and alternative with respect to the method of transcendental analysis. That is, e.g. in M1 level philosophies, we find rival accounts addressing the problem of perception – and the nature of perceptual vocabulary – by identifying different theories of perception, each belonging to a specific (manifest or scientific) framework. And indeed, Sellars seems to think that there are various theories of perception that can be developed within the Manifest Image framework, each using a distinct set of concepts (‘sensation’ is one of them, ‘proper function’ and maybe also ‘thought’). On the other hand, within the Scientific Image, there will be different theories of perception and perceptual consciousness (‘sensa’ and non-particulate ‘pure processes’ cf. PSIM VI, FFMP III, SSIS), which are methodologically and conceptually developed within that rival framework. A good amount of Sellarsian reflections are aimed precisely at understanding how these different theories of perception (in different frameworks) can be said to be ad odds or chime with each other (cf. Dach 2018, Corti 2021a). 21 Thus it is essential to note that in my conception of the Scientific Image the framework of the basic science replaces the frameworks of the “other levels”. It will, therefore, contain a successor concept [i.e. sensa] which replaces the concept of a sense impression which belongs to the Manifest Image. (SSIS 413) By these lights, the notion of ‘sense impression’ will be disposed precisely because it belongs to a theory of perception that is typical of the Manifest Image. Pressure is generated from an inter-framework point of view. However, the method of distinguishing levels of abstraction emphasises a different methodological route in Sellars’ analysis of perception: The relevant notions like ‘sensation’ or ‘thoughts’ are themselves not addressed as part of a theory that is developed within a certain framework (and to be contextualised to that framework). Rather, the

Transcendental Methodology 23 move is to look at perceptual concepts as being justified ‘transcendentally’, insofar as they manifest conditions of possibility for intentionality – namely as a condition for the applicability of a certain framework in the first place (according to a complex sense of ‘applicability’ that I will specify throughout the book; cf. Chapter 5 ff.). 22 So what is the place of the notion of a ‘conceptual framework’ within the reconstruction of Sellars’ argument unfolding across the subsequent chapters? The transcendental method carried out at the M2 level aims at identifying the necessary conditions for a contentful conceptual framework to be intentional überhaupt: As a consequence, such transcendental consideration aims to apply to all possible frameworks. And more than that, a transcendental analysis aims at articulating what is involved in understanding something as an experience by grasping it in terms of a framework (cf. Chapter 5.2 below and my rendition of this idea in terms of an immanence criterion). Thus, the method of differentiating between levels of abstraction is required in order to even be able to gain a position from which distinctions among (M1) points of view can be made, and conditions (downstream from M2) binding for all frameworks as such be spelled out; this includes higher-level claims and related analyses of claims as claims situated on a certain level. The transcendental method behaves inclusively and neutrally in relation to methodologies deployed downstream, e.g. on a lower level (M1) in explorative analyses and explications of framework features. 23 If this neutrality is deviated from at all, then only in order to emphasise core-framework features – e.g. demonstrative reference, categoriality, and trans framework features (such as are address under the heading of immanence criterion or cross-framework counterparts, Chapters 5 and 6). Taken together, these are precisely those features which allow ascribing representationality to a framework and make it one that we can treat as meaningful about the world as part of which we understand ourselves. It is in this sense that the transcendental method is the genericum within which it is possible to locate and assess the “descriptive phenomenology of Husserl and the conceptual analysis of the developing phase of Oxford philosophy”, which Sellars goes on to call “halting approximations” (IRH §2). This approximation to ‘categorial knowledge’ I understand as follows. When it comes to determining core features of frameworks (an M2 operation), levels of abstraction and e.g. meta-­ classification serve to individuate frameworks as such, e.g. as what philosophers (on M1) deploy in articulating the explanatory merits of rivalling theories of perception for (level O) framework furniture, e.g. objects, states, properties, etc. Despite Sellars’ commitment to a scientific realism (cf. Chapter 6), the meta-philosophical distinction between Scientific and Manifest Images itself does not represent a higher-level framework but a tool for reflection (here I follow Dach 2018).

24  Transcendental Methodology The claim inherent in this approach to reading Sellars as deploying a transcendental method of analysis distinguishing levels is this: The analysis provides the registry in which methods such as conceptual analysis, teleological theories of mental content and functions, Strawson-style transcendental arguments, Husserlian phenomenology, etc. can all be characterised and assessed for their conceptual merits.

Notes 1. As Kant formulates, it in the February 1772 letter to Marcus Herz (10:130–131), and as Sellars takes up in SM, KTE, IKTE, and in KTI. 2. Cf. Westphal (2004, Chapters 1 and 7) for an alternative reconstruction of Kant’s method. 3. The spirit of this enterprise is akin to the debate about the nature of transcendental arguments cf. Cassam (1987), Callanan (2006), Robert Stern (2003, 2020) and future research will have to bring out respective similarities and differences. 4. Here one might find that the meta-conceptuality of this activity marks a necessary but not yet sufficient requirement for the activity to qualify as ‘transcendental’. This is a point to which we return in discussing the special status of analytic truths. 5. Cf. [N]ot every a priori cognition must be called transcendental, but only that by means of which we cognize that and how certain representations (intuitions or concepts) are applied entirely a priori, or are possible (i.e. the possibility of cognition or its use a priori). Hence neither space nor any geometrical determination of it a priori is a transcendental representation, but only the cognition that these representations are not of empirical origin at all and the possibility that they can nevertheless be related a priori to objects of experience can be called transcendental. Likewise, the use of space about all objects in general would also be transcendental; but if it is restricted solely to objects of the senses, then it is called empirical. The difference between the transcendental and the empirical therefore belongs only to the critique of cognitions and does not concern their relation to their object (B81). 6. The discussion of objections from pragmatism and expressivism can be read to present just such challenges, see Chapter 6. 7. Cf. Haag (2007, 6 and fn 47), and likewise Rosenberg (2005, 123) who offers a commentary on B135. 8. Cf. Westphal (2017, Chapter 20) for an expansive discussion and defence of the role of synthetic necessary truths in Kant, Sellars and contemporaries. Westphal (2017, 439) finds ›Sellars was right to highlight such ‘synthetic necessary truths’, that such truths require – and are required for – robust pragmatic realism‹. Critically assessing the meta-linguistic nature of Sellars’ analysis of synthetic necessary truths, Westphal indicates the central aspect to be on his reading (ibid.): ›All of these [reflections] converge in one simple point with profound ramifications for epistemology: there are simply no non-modal descriptive terms of any sort, no matter how putatively simple‹. 9. Cf. Förster (2012, Chapter 1) and Haag (2012c, 2017, 18–26) from whom I take my inspiration for reading Sellars in this light. 10. This is a point Rorty (1988) and subsequently Price (2013) will take issue with. One of the most interesting questions with respect to their criticism of Sellars’ program will turn out to be: What status can the positions claim

Transcendental Methodology 25 for themselves, out of which Rorty and Price launch their attack? As we will show in Chapter 6, the neo-pragmatists are themselves committed to a transcendental position that requires a notion of robustness of reference to ‘objects’ in the very same sense that Sellars develops explicitly. Cf. also Raymond Brassier’s Foreword to this book. 11. It is with regard to points like these that the merit of distinguishing levels of abstraction will become evident. On their basis we can specify the status of the reflection expressed in the claim that ›in principle we, rather than God alone, can pride the cash‹, since the concept of God shows up as an analogically extended notion of our finite capacities. Cf. Chapter 5.4. 12. On Stekeler-Weithofer’s (2009, 197–198) view, the status of such truths is best conceived of as situation-transcendent default rules of inference. As such they are neither analytic nor empirical, which, while we treat them as a priori binding, are at best generic rules in that they are never or seldom without exceptions. 13. Cf. Westphal (2020, 79–88) for a congenial assessment of this point, by way of a crisp appraisal detailing the programme and ambition of Sellars’ Science and Metaphysics. 14. A reader who may wonder about the exact nature of Sellars’ understanding of conceptual structures in relation to Davidson’s criticism of conceptual schemes can find a discussion of this in Chapter 6.1.3. See also Peregrin (2018) and Shapiro (2019) for a recent take on this question. 15. I will touch later on the point made in (RM), (ME), and (EPM) which states that transcendental psychology is not reducible to transcendental linguistics. Sellars’ argument here is that not all thoughts are linguistic but some thoughts are genuine thoughts aside of all language. 16. In the mid-1970s, Sellars reports how in his Oxford days he realised he ›would have to work out a whole new way of looking at the conceptual order. The situation was roughly the following. I had already broken with traditional empiricism by my realistic approach to the logical, causal, and deontological modalities. What was needed was a functional theory of concepts which would make their role in reasoning, rather than a supposed origin in experience, their primary feature. The influence of Kant was to play a decisive role‹ (AR §29). This much is clear from SM’s subtitle alone. Three years after these autobiographical reflections he publishes IKTE which develops the idea of levels of abstraction with regard to the Kantian categories.   Such that a ‘new way of looking at the conceptual order’ amounts to articulating the idea that reflections on intentional purport are nested. His theory of dot-quotations, meaning as functional classification and cross framework identification of counterpart objects rest on a general ‘functionalist’ understanding of concepts. It is precisely by reflecting on the nested status meta-classificatory concepts play in reasoning (on M1 and M2) that Sellars integrates his conceptualist leanings, the rejection of the myth, and his functionalism within the methodology of distinguishing levels of abstraction. 17. In view of this layering one objection could be raised that asks: Do we not run into a regress of further levels? To this the appropriate reply would be that the extreme generality of the reference relation at M2 does not allow for a further meta-level. In this sense, M2 is without an alternative. And where there really is an alternative to the kind of knowledge formulated at that level, the claim towards uniqueness of the previous articulations of M2 wasn’t met. But that does not install a level M3.

26  Transcendental Methodology 18. Cf. CDCM 79: ›Once the tautology ‘The world is described by descriptive concepts’ is freed from the idea that the business of all non-logical concepts is to describe, the way is clear to an ungrudging recognition that many expressions which empiricists have relegated to second-class citizenship. . . are not inferior, just different‹. 19. Cf. KTE, IKTE, TTC, and especially the debate between Brandom and Price in Price (2013); cf. also Christias (2014) and more recently Dach (2018) and Hicks (2020). 20. I am indebted to Luca Corti for raising this point to me. 21. This is most explicitly expressed in SSIS 395 when making a point against Cornman: ›[H]e consistently underestimates the role in my philosophy of the contrast between inter-framework and intra-framework relationships, and, in particular, of the concept of the replaceability of one framework by another, a special (if global) case of which would be the replaceability of the “Manifest” by the “Scientific” Image of Man-in-the-World‹. 22. According to Sellars, on my reading, the relation between conceptual frameworks and what they are about is not a mere question of ‘application’ of the former to the latter. This will come to light in the discussion of the very concept of object-hood (Chapter 2), demonstrative reference (Chapter 3), non-relationality (Chapter 4), and the role of counterpart objects in making cross-framework identity claims (Chapter 5). 23. Consider Sellars’ reference to phenomenology in discussing American common-sense realist Everett Hall (IRH §1–2): ›Hall offers instead [of Hume] a conception of philosophy as “neither a priori nor empirical” [Hall 1961, 5]. By ‘empirical’ he has in mind […] inductive methods of the empirical and theoretical sciences. He argues for “a third kind of knowledge” which he calls “categorial” (p. 6). The test of claims falling within this “third enterprise” is to be found “in the forms of everyday thought about everyday matters in so far as these reveal commitment in some tacit way to a view or perhaps several views about how the world is made up, about its basic ‘dimensions’” (p. 6). “We find,” he continues, “these forms of everyday thought chiefly in the grammatical structures (in a broad sense) of daily speech, in what may be called the resources of ordinary language, although they are also present in the ways in which we personally experience things.” “... the latter,” [Everett] adds, “reflect, to a great extent, the formative influence of our mother tongue,” (p. 6)‹. And, Sellars continues: ›This characterization of the philosophical enterprise illustrates once again the catholicity, i.e., the universal sweep, of Everett Hall’s philosophy, for, in my opinion, this conception of philosophy is the truth to which both the descriptive phenomenology of Husserl and the conceptual analysis of the developing phase of Oxford philosophy are halting approximations‹.

2

Transcendental Psychology

2.1  The Meta-Classification Thesis In the last chapter, we introduced the aim of presuppositional analysis. This aim can be summed up as that of establishing transcendental constraints. That is, constraints on the concept of our intentional relation as finite beings who lay claim to knowledge of representable objects in nature. And it was here that the advantages of reading Sellars as subscribing to the method of distinguishing levels of abstraction began to show up. By taking an abstractive standpoint towards meta-level knowledge claims, the constraints on those claims to meta-knowledge came into view. In this chapter, we discuss Sellars’ interpretation of the Kantian categories as undertaken from a standpoint on M2, a standpoint that applies the method of abstraction. This can be expressed in the form of a thesis. The meta-classification thesis holds that Kantian categories are recipes for functional meta-classification of mental acts of judging. We will develop this thesis in three steps. In a first step, we reconstruct Sellars’ interpretation of the Kantian categories. What are they and how do they figure in our intentionality? It is one thing to reconstruct Sellars’ specific reading of Kant’s account. It is another to say how Sellars’ own account is supposed to work. Our aim is to arrive at an answer to this more specific task by the end of this section. Then we consider objections to Sellars’ reading of the categories that considers them as arbitrary postulates or as the highest genera of objects in the world. The aim of refuting these objections is to show that Sellars defends a version of Kant’s transcendental psychology. Finally, we tie the results of this section back to the idea of a presuppositional analysis that was our concern in the previous chapter. This is done by showing how Sellars’ take on the categories is part of an overarching idea. The method of analysing intentionality by differentiating levels of abstraction ultimately yields a robust structural a priori knowledge. That is, analytic knowledge – won by way of an analysis – about synthetic knowledge. DOI: 10.4324/9781003221364-3

28  Transcendental Psychology 2.1.1  Categories as Meta-Classificatory Rules Since Sellars’ interpretation of the categories surfaces in a number of different places, it merits a detailed exegetical analysis. We begin by formulating a preliminary hypothesis: Sellars interprets the categories as required at all levels of abstraction. To confirm their continuity (as doubly pure, un-schematized and as schematized categories) is to justify their role in the development of the concept of an empirical object. A first explanation of that hypothesis can be given in the form of a metaphor. Whereas the presuppositional analysis takes us to the highest point of abstraction, that is, to the discipline of general logic, it is by way of a constructive synthesis – in which the categories figure as the rules of such syntheses – that we get back to the concept of empirical objects in space and time. We need to climb back, as it were, down the levels we came up in abstracting away from our place of departure in common sense conceptions of being in touch with objects around us. There the account took the form of unwinding the thread of our intentional reach such that it would leave us only with its purest form, the empty accusative. For our intentionality to reach out to the world is for it to consist in the claim to an objective purport to some Gegenstand überhaupt. And we now, from that highest point of general logic, have to synthetically combine all the elements, add them to the thread of the main account.1 And that winding up of all the elements consists of the demonstration of the categories’ necessary and constitutive role for our finite human capacities. It is those capacities and their functions which feature prominently in the explanation of intentionality. Sellars expresses that task at one point as that of demonstrating how the categories are unified by the concept of empirical knowledge because they are simply constituent moments of this one complex concept […]. (KTE §9) The project of a presuppositional analysis thus has its corresponding part in a constructive account revealing the necessary elements and the form that our concept of empirical knowledge and the concept of a knowable object in nature must have. And it is in that second, complementary task that the theory of the Kantian categories has its home. 2 The account explains how categories show up in their role at the highest point of abstraction. There they are but pure forms of judgement. Their introduction is tied to acts of thinking. They are the structures available to the understanding in thinking and forming judgements generally. The categories are in first instance simply identical with the forms of judgment […]. These forms of thought would be involved in

Transcendental Psychology 29 thinking about any subject matter from perceptual objects to metaphysics and mathematics. (IKTE §44) The account of categories takes its start from a characterisation of the act of thinking a content. Categories accompany and structure the thinking of contents of a range of subject matters, be they abstract, as metaphysical or mathematical ‘objects’ would be, or be they less abstract, as objects in perceptual context would be. Categories in their doubly pure instance pertain to acts of thinking any contents whatsoever. At the level of general logic, we are only concerned with pure form, that is, only with the structure that acts of thinking conform to. That is why Sellars can maintain that ›categories and concepts, all are features of acts of thought‹ (KPT XVII §19). The relevance of this approach stands out when we consider that we are on our way back to items in nature. We do not know yet what these are. To establish what items in nature are, we first need to be clear about what components enter into the qualification of the cognitive state that we can then characterise as being a state of knowledge of something in nature. At this point, the transcendental philosopher brings out that the primary elements we are in contact with are thoughts, or more specifically, acts of thinking. Acts that have to have determinate features, such as to be about something specific. Here, it matters that properties of acts of thinking can be very different from properties as they are represented by such acts. One might entertain a thought, which we can express as the act of judging that ›Tom is tall.‹: A judgement, wherein the content ‘tall’ is related suitably to the content ‘Tom’. I might come to form that judgement quickly yet with certainty. But the qualifications ‘quickly’ and ‘with certainty’ pertain to the act of judging, not to the content: That state of affairs that Tom is tall itself exemplifies neither quickness nor psychological certainty. Here it also matters that the seeming infinite number of acts of judging I might bring forth are compatible with being grouped in a finite number of types of judgement. In the case at hand, the judgement is one of the ‘Tom is tall’ kind. And when we consider the structure of that judgement in more detail, it turns out to be of a subject-predicate form. The following passage makes it clear in what sense Sellars interprets the Kantian categories as concepts for the activity of classifying acts of thinkings. Thus, a judging that Tom is tall would, in its generic character, be a judging of the subject-predicate form. It is a judging that a certain substance has a certain attribute. (These two ways of putting it are equivalent.) If we focus our attention on the predicate we can characterize the judging more specifically as a judging that [ascribes to] a certain substance […] the attribute [of being] tall.

30  Transcendental Psychology Thus, just as to say that a judging is a judging that a certain substance has a certain attribute is to say that the judging is of a certain generic kind (i.e., has certain generic logical powers); so to say that a judging is a judging that a certain substance is tall is to classify the judging as one of the such and such is tall kind, i.e., to classify it in a way that ascribes to it the more specific conceptual powers distinctive of the concept of being tall. Indeed, for the judging to "contain the concept of being tall", is nothing more nor less than for it to have these specific powers. (KTE §25) In this, we find evidence that Sellars’ account is devised from a standpoint that investigates the shape and structure of the most abstract features of thought.3 With regard to the notion of presuppositional analysis, this implies a reference to the method of abstraction and to the level M2. It is at this level that Sellars’ interpretation of the categories as functional meta-classifications begins to make sense.4 To ascribe to an act of judging a categorial form, based on the structure of the act, is to classify the act of thinking with regard to its expressing a conceptual power. In the present example, it is the conceptual power or attribute of ‘being tall’. By way of elucidation, we can say, the category of a predicate allows us to group a range of judgings that qualify further judgings of the substance kind. The very idea of a category is functionally specified as a meta-­ conceptual notion. From viewpoint of M2, categories are characterised as functional meta-classifications of acts of judgement which ascribe epistemically relevant powers to judgeable contents.. 2.1.2  Objection 1: Categories as Highest Kinds We can further detail this way of interpreting categories by considering an opposing view of categories that understands them as highest kinds of objects. Sellars rejects such a (perhaps traditional)5 approach by arguing that this would lead into metaphysical absurdities. Categories cannot be the most generic sortal concepts that can occur in statements about the objects around us. (KTE §28) For that would lead to a hierarchical view, in which categorial statements were to be analysed as qualifying natural objects. That this cannot be right, can be gleaned from the fact that ascending the metaphysical hierarchy would force us to assume the existence of quality-individuals, relation-individuals

Transcendental Psychology 31 as the subjects of statements of the form … is a quality … is a relation.

(KTE §27)

This is so because the view that categories characterise objects forces us to hold that there must be such things as would figure, grammatically speaking, in the subject position of such statements. And, sure, in an ordinary language sense of what categories are, we should expect to find the explanation that they are what ‘gives us the nature’ of the objects around us. Or, from a more pragmatic point of view, one can say: Categories are what allows us to group and subdivide what’s there. However, here too, a bit of semantic self-consciousness comes in handy. Sellars writes accordingly: Medieval logicians began the process of reinterpreting the categories that culminated in Kant’s Critique, by recognizing that certain statements (thus ‘Man is a species’) which seem to be about queer entities in the world are actually statements that classify constituents of conceptual acts. (KTE §29, cf. PR §11) In a first step, we have thus pinned down the relevance of this reading of the categories by placing it in the context of presuppositional analysis. To see it there makes it actually easier to appreciate the functional dimension of categorial classifications. Acts of thinking are grouped according to the function they have within the overall epistemology of the thinking agent. The relevant epistemic powers of individual acts of judging are collocated, such as to allow an agent to be expressing determinate judgements about any matter of fact. We are not yet in a position to discuss the schematisation of the categories to allow judgements about matter-of-factual affairs (cf. Chapters 2.2.4 and 4.2). But we have reconstructed Sellars’ interpretation of the categories as according to them a transcendental and therein epistemologically constitutive function. Without categorial qualifications of acts of judging, there is no determinate content to thought. 2.1.3  Objection 2: Categories as Arbitrary Constructs To what extent can Sellars count as also defending a version of Kant’s transcendental psychology and why is this important to consider? The Kantian account of experience in comprising categories as necessary components was not always met with cordiality. Perhaps because their role has always proven to notoriously difficult to pin down, the more abstract feature of Kant’s account and the categories especially have seemed more like dubious elements, postulated on meagre grounds.

32  Transcendental Psychology It is often said today that Kant’s Critique consists of important insights into the logical geography of our conceptual structure which are obscured almost to the point of invisibility by a tedious and fictitious “transcendental psychology.” Kant is said to postulate a mechanism consisting of empirically inaccessible mental processes which “constructs” the world of experience out of sense impressions. (KTE §39) Since the categories are prime candidates for figuring in unconscious mental processes that operate on sense impressions, so constructing our phenomenal world, the objection holds that their introduction suffers from arbitrariness.6 Worse, the entire philosophical psychology of the epistemic agent’s make-up, envisaged by transcendental philosophy, threatens to be fictitious. It is this objection that Sellars aims to take seriously. But not only in the hope of vindicating Kant as a sound philosopher; the ambitious project is to defend Kant’s theory of experience and give a modernised account of the structure of his transcendental psychology. Sellars writes accordingly: If my argument is correct, this criticism is misdirected. The true situation can be seen by assessing the validity of a corresponding criticism directed against a linguistic version of Kant’s position. (KTE §39) The defence Sellars announces takes an indirect shape. It suggests to construe a linguistic version of Kant’s project and test whether the criticism still applies there. If it does not, Sellars can take himself to have done two things: First, to have defended Kant’s transcendental psychology, and second, to have given a contemporary picture that demonstrates the validity of Kant’s thinking for our present philosophical concerns. It is in this line of thought that the interpretation of the categories belongs. The idea is that to be able to present a functional meta-­ classificatory account of what categories are there for is to give us an account of knowable items that justifies the necessary features any system of knowledge is characterised by. In Kant, this account constitutes transcendental psychology and explicates the central capacities and conditions binding for epistemic subjects. In Sellars, the account takes the shape of transcendental linguistic, reflecting on the necessary features of a language that is to constitute a conceptual framework for yielding knowledge of the world it is used in. What are the steps of Sellars’ defence in turn? From what we have learned about the categories up to this point, it has become evident that they are precisely not simply postulated as additional elements to a thin conception of knowledge.

Transcendental Psychology 33 The presuppositional analysis took us to the point of general logic by way of the method of abstracting away from a common sense view of objects and their knowledge. At each stage, the investigation consisted of an analysis of the presuppositions implicitly underlying the respective claim to knowledge. This procedure was repeated at the next level, until we were left with the pure forms of thought. Here the insight is that any valid perspective on knowledge whatsoever rests on these presuppositions and it is this point which rules out the objection from arbitrariness. The pure forms of thoughts are not introduced as hypothetical elements about which we cannot know anything in everyday life. They are what remains at the bottom of the pot, once all the water has been boiled away. Their fundamentality consists in the fact that they structure all presumed reference to objects in all domains we aspire to gain knowledge of: ›These forms of thought would be involved in thinking about any subject matter from perceptual objects to metaphysics and mathematics‹ (IKTE §44). As we saw, this shows that the essential forms are discovered by way of an abstraction as forms that necessarily are constitutive of thinking a Gegenstand überhaupt. This strongly implies that they are binding for any being, finite and infinite being alike. Infinite beings also have to form their judgements by way of, say, ascribing properties to substances. There’s no meaningful thinking of structured thoughts that aspire to be objectively valid (in a sense to be explained) that is not structured by these Kantian forms of judgement.7 So when Sellars introduces the notion of a category as something that we win by noticing that we necessarily apply these forms of thought in every act of articulating more specific thinkable contents, the categories are shown to qualify individual acts of thinking.8 And the closer we get to the notion of a knowable item in nature, the more schematisation to spatio-temporal structures is required to specify these forms, providing the specific differences to the genera of forms they belong to. Kant not only rediscovered these insights [by medieval logicians], but extended them in such a way as to connect categories not only with the logical forms in the narrowest sense (roughly, syntactical powers) studied by formal logicians, but, to an extent not always recognized, with the logical powers in a broader sense which are essential to a conceptual framework the employment of which generates knowledge of matter of fact. Thus he thinks of the categories as together constituting the concept of an object of empirical knowledge. (KTE §29) Accordingly, the general act of judging that ‘S is p’ is at the level of general logic. General logic involves making judgements of that “general act” type.

34  Transcendental Psychology The more specific thought that 5 is an uneven number is at the level of transcendental logic and ranges over a variety of abstract fields of knowledge. While the thought that this red pyramid here in the front of me is at the level of spatio-temporally identifiable objects. The idea of a gradual schematisation of the pure forms of thought, down to specific matter-of-factual judgements show the categories to be continuous in their presence as organising thinkable contents. In the concept of them, we make explicit what it is that allows us to classify different acts of judging according to their function. But they are not arbitrarily so continuous. Categories necessarily structure acts of thinking at all stages of abstraction. They are what allows us to functionally weigh judgings against one another with regard to their relevance and their powers: Kantian “categories” are concepts of logical form where ‘logical’ is to be taken in a broad sense, roughly equivalent to ‘epistemic’. To say of a judging that it has a certain logical form is to classify it and its constituents with respect to their epistemic powers. (KTE §23) Here one might imagine a critic saying ›I grant for argument’s sake that the categories are always present in our acts of judging at various stages of abstraction. What does that add to the project of giving us an account of a necessary form that any account of knowledge has to have?‹ In response to this challenge, we need to make explicit what specifically Sellars introduces into his account of the categories when he argues that categories meta-classify according to the functional dimension of a judging’s epistemic power. In the modernised language, Sellars advocates this means: The conception of the categories as the most general classifications of the logical powers that a conceptual system must have in order to generate empirical knowledge is the heart of the Kantian revolution. (KTE §29) With this, it becomes clear that any account of knowledge of matter of fact, that is, knowledge of items in nature, has to accord categorial qualifications of the contents known some positive function or other. Central here is the notion of ‘acts of judging’ because categories give form to judgings. If acts of judging are to be about something more and more specific, (i.e. determinate) then these judgements need to be qualified with regard to their categorial structure. And it is in that way that the qualification gives their content its epistemic importance.

Transcendental Psychology 35 We are not yet at a stage concerned with spatio-temporal objects. So transcendental logic only tells us that to think of objects überhaupt is to think of them in terms of acts of judging the contents of which can be classified with respect to their epistemic powers. This way of qualifying categories dismisses the objection that their introduction be arbitrary. And it also brings out in what sense their involvement in accounting for the structure of our factual intentionality has to have a necessary dimension. One that can be expressed in terms of lawfulness or by saying that categories give form to judgings.9 Having reconstructed Sellars’ interpretation of the role of the Kantian categories, we now need to turn to his treatment of fundamental Cartesian distinctions. On Sellars view, they are fundamental enough to be called categories which form the background to Kant’s understanding of how we are to frame the concept of an object of experience.10

2.2  The Concept of an Object of Experience in Kant One might wonder why Sellars stresses the Cartesian heritage in Kant’s thought. After all, Descartes is just one rationalist predecessor. What makes Descartes so important for understanding the overall argument of Kant’s theory of experience? Sellars has his focus on a very specific trait of Descartes thought. It is mainly the epistemological and ontological categories applied to mental acts that resurface throughout Sellars’ papers on Kant and his Science and Metaphysics (KTI §3, 20–28, 53; SM II §2–3, 8, 30, 36, 49, 55, 73). Sellars argues that the epistemological and ontological categories change. Since categories are the notions through which we understand the nature of mental acts and their contents, it matters how they are interpreted and in what way the forms of their application can change. This has serious implications for Kant’s project of a Transcendental Psychology. Either the categories are explained with regard to their necessary features as fixed, or we are to give up the project altogether and conceive of them as subject to conceptual change, as we do with regard to empirical concepts. So what can it reasonably mean to claim that they ‘change’? First of all, it means that they are not given to us, but that we win them, extract them, formulate them and receive them from our dialogue with the positions in the history of philosophy. They change in so far as we apply further distinctions in our attempts of interpreting their role in our thinking. If we consider them according to a transcendental method of analysis not on the paradigm of objects but on the functional paradigm of acts of classification they give a form to, we can see that they can figure in a great variety of explanatory projects. As Sellars demonstrates in SM I and II, this concerns, for instance, the notion of ‘existing in a representation’.

36  Transcendental Psychology 2.2.1  Categories for Mental Acts What categories we make use of in articulating our understanding of the metaphysics of intentionality, is therefore a matter of the arguments and distinctions we bring to the task. And here, the importance Sellars places on the methodology of a transcendental analysis stands out most clearly. It lies precisely in the fact that we necessarily have to re-interpret the categories to use them meaningfully. And in our meta-theory (on M2) we can formulate a way such theory-producing activity advances as follows: It is exactly Sellars’ point to show (on M2) that Kant’s re-interpretation of the Cartesian categories (on M1) for mental acts and their contents is the background to his transcendental idealism.11 This can be taken as the reason for Sellars’ dwelling on the Cartesian background of Kant’s understanding of the nature of representations (Vorstellungen) in all their classes and sub-classes. To this, Sellars adds: The importance of the categories Descartes applies to mental acts lies […] in the fact that they can be seen to be less sophisticated counterparts of distinctions which are drawn with more or less rigour in those contemporary philosophies of mind which have been influenced by formal semantics. (SM II §3) Thereby highlighting the continuous importance of the distinctions, we find in Descartes for what we try to achieve in contemporary philosophy of mind. But the most important reason for returning again and again to the subtleties inherent in thinking about the act/content distinction and what it is to exist inside and outside mental acts is that: [T]he theory of mental acts or representations […] underlies Kant’s distinction between existence as appearance and existence in itself. (SM vii) Just how important this point is, one might easily overlook.12 But to make it visible, one only need rehash how, from the perspective of the project of a transcendental analysis, approaching the world, and its affecting us has to be carried out in terms of an account that says how mental acts and what they are acts of present us with knowledge of the world. And in as much as that amounts to an articulation of the necessary feature that any system has to have to generate knowledge of empirical matter of fact, this is still in line with the task of defending Kant’s transcendental psychology.13 And for Kant to say what exists means working with the distinction between existing as the content of an appearance and existing outside or independently of being an

Transcendental Psychology 37 appearance for us. On this note, Sellars specifies the nature of his program in the following way: My primary concern, however, is with the ontological aspects of Kant’s idealism, and only incidentally (and by implication) with epistemological issues concerning synthetic a priori knowledge. I note in passing, however, that since the ontological aspects of Kant’s idealism concern in large part the ontology of mental states, and since, although epistemology is not psychology, it is mental states which are the proper subjects of epistemic appraisal, Kant’s ontology of mental states is directly relevant to his epistemology and, consequently, to his transcendental philosophy. (KTI §3) The root notion is the theory of mental acts. This is not based on psychological reasons, but as explained in the first chapter, on the methodological reason that they are systematically the ground on which Kant develops his Idealism. Here we have the reason for why Descartes’ understanding of mental acts is such an essential part of the project of Kant’s defence of transcendental psychology. This leads to the further question of how Kant’s theory of the categories for mental acts hangs together with the categories that are the pure concepts of the understanding? This is a question we can address from within the strictures of our transcendental analysis. The notion of a world has to be developed by explaining how something shows up as the content of our acts of representing. Here, two directions of fit present themselves to us. Either we generate this content, or the content has its place of origin outside of ourselves. Both directions, of course, are to be elaborated in more detail. The idea central here is that we are to deal with two kinds of acts of representings that ultimately have to come together to form empirical cognitions proper: Conceptual representings and intuitive representings. So we can say that our knowledge of the world has to encompass two modes of existence. The first one is that of ‘existing in thoughts or as a feature of thought’. It is a mode we encountered when articulating the nature of categories and of conceptual items. Their way of being is within or as features of conceptual acts of representings (cf. KPT XVII §19). The categories give internal structure to conceptual acts of thought such that they can be about specific items, which, when schematised and thereby specified to our forms of sensibility can be about spatio-­ temporal items. Categories, although being the most abstract features of acts of thought, are subordinate to mental acts as the highest point of having content before one at all. (In the order of explanation here, the -ing precedes the -ed, acts of representing are prior to items represented.) The second mode of existing is that one of ‘existing as sensible items in

38  Transcendental Psychology intuitive acts of representing’. It follows that any theory that is to do justice to the idea that our engagement with the world comprises our passivity with regard to it has to find a place for acts of representings that deliver us material to be taken up into mental acts of perceiving. If we were to generate the contents of all of our engagements with the world, the world would be at no point an independent factor in the generation of our knowledge of it. And to avoid that consequence the impact on our sensibility has to be interpreted in a way that gives it the transcendental function of providing friction with the world, such that our conceptualisations of it can be wrong. This is the systematic reason for Sellars’ to endorse the Kantian two stems. As our acts of representing fall into roughly two classes, conceptual and intuitive representings, this mirrors the two stems: The part played by the understanding and the part played by sensibility. Our transcendental reflections on the articulation of the concept of knowledge have brought out that both can be read as forms to a material that they structure. The difficulty of how to account for the friction of our mental acts with the world is not off the table yet. Rather, we need to understand the depth of its thrust to see how it continues to plague the prospect of arriving at a world known. This is even more pressing where Kant’s account steers into an idealism about the world given to us only ever in appearances. Such that the question of how Kant handles the requirement of having a friction for our conceptual content (such as to not spin in the void, in the sense McDowell 1994 envisaged) becomes the question of how transcendental idealism can be made compatible with the concept of there being an independent world at all. Sellars reconstructs the essentials of what he takes to be Kant’s transcendental idealism by connecting our theme of ‘necessity’ with the question of what it means to say of an item that it has actual (wirkliche) existence. For our programme, this entails saying something about the difference between a traditional sense of being actual and a critical sense of being actual. And it entails addressing the question in what way Sellars’ reconstruction of Kant’s program of a transcendental idealism finds Sellars’ approval. So we will ask (1) Is Sellars a transcendental idealist in Kant’s sense? (Chapter 2.2) Here the answer will be positive with regard to the analysis of the mode of existence Kant gives for manifest empirical objects of experience. And we will ask (2) Is Sellars a transcendental realist in Kant’s sense of the term? (Chapter 2.2.4) Here, the answer will be negative since Sellars agrees with Kant in holding that the in itself as a theoretical notion cannot be spatio-temporal (SM II §14, 43). The long answer will also have to say something about the way in which the world is to be perceived from the vantage point of the Scientific Image (cf. Chapters 5.3 and 5.4). And this will lead us to consider in depth (3) the points of disagreement Sellars finds with Kant (cf. also Chapters 3.9 and 3.10).

Transcendental Psychology 39 2.2.2  Contents We Think and Contents We Experience What argumentative support is there in Kant for the claim that any reference to objects has to be rule-based? The significance of this claim lies in the implication that if that is true, it is a transcendental requirement on the way in which both our capacities, on our receptive synthetic activity and on the (meta-)classificatory activity of our spontaneity figure in a successful account of intentionality. So it comes as no surprise that the two topics of (1) a Gegenstand überhaupt and (2) what it takes to understand the nature of the presupposition of getting at it continues to be at the centre of the discussion around the status of represented contents. In this section, we will prepare the ground for the line of thought that adds (3) the notion of rules as another necessary component. If in the last section we have learned something about the necessary forms the understanding brings to judgeable contents, it is now time to look at the corresponding forms on the side of our receptivity. This means addressing the question in what way Sellars accords sensibility a transcendental function in his account of intentionality. In so far as Sellars endorses Kant’s two stems as central to any account of intentionality, he thereby also is bound to endorse the hylomorphism inherent in how they are to merge. At the core then lies the thesis that no representable contents are given to our cognition without being formed in one way or another by the work of the understanding and sensibility respectively. Thus our task in this chapter will be to reconstruct Sellars’ reading of Kant’s account of what it is to be an object of acts of representing. We will suggest that one can understand the Kantian commitment to hylomorphism as showing most clearly in the working together of the distinctions of act and content and the properties that they have. As will emerge, Kant’s account of objects takes the form of a transcendental idealism, the gist of which Sellars underwrites. A related task we turn to towards the end of this section will be to identify the points of departure that Sellars secures for himself in pursuing his transcendental methodology. If the theme introducing Kant’s idealism is that of the relation between form and matter, it will be helpful to recall what in the last sections we have learned about the categories. The account spelled out the ways in which categories enable functional meta-classifications of the contents of representations with regard to their epistemic powers. That way it explained one form of necessity constitutive of our having epistemic contact with anything at all (that is not the self, i.e. that is independent of it). Now by turning to the material side of the nature of contents that get taken up into the processes that operate their formation on them, we encounter another form of necessity. This reflection on the material side is inherently connected to the discussion of the categories by

40  Transcendental Psychology being a precursor to it. Sellars utilises a passage in A104 to fend off an interpretation of Kant’s theory of experience which he considers ›a mistake‹ (KTI §36). Though one that understandably occurs when one fails to distinguish acts from contents. The interpretation reads Kant as arguing that material objects are […] logical constructions out of mental states which, though not spatial, are representations of spatial items — i.e. of which spatial items are the contents or immanent objects. (KTI §36) In KTI Sellars presents a reconstruction that is at once able to avoid such a constructivist reading while at the same time allowing for enough leeway as to show how one could come to think that this could be a plausible reading of Kant’s idealism. So how should we understand the overarching topic of Sellars’ engagement with Kant’s idealism? It is best to read Sellars as preparing the way for his own account of what objects of experiences are. Sellars prepares this by carefully reconstructing Kant’s account of perceptual experiences. Let us trace where he begins to depart from Kant’s account. Our reading of Sellars Kant-interpretation will focus on the methodological aspects of his constructive analysis and lay aside numerous controversial issues in contemporary Kant scholarship. Up to this point, we have encountered the notion of an item at the centre of a representation only in a very thinned out, bare version. It was an object only in the negative sense that it served as the point of referral for consciousness’ directedness to something about which knowledge claims are to be made. This has led us by way of a presuppositional analysis to consider the engagement of the understanding in dealing with representational contents by way of forming judgements about them. So now it is time to consider the second stem, which concerns the transcendental role played by the import of our sensibility’s receptivity.14 Here Sellars puts stress on the historical dimension of Kant’s theory of the representations of sensibility. That is, the first larger parts of KTI are devoted to introducing the frameworks of thoughts as conceived of in the Cartesian tradition. By laying stress on the methodological dimension of Sellars’ thinking, we can illustrate how Sellars’ engagement with Kant’s transcendental idealism can be read.15 This shows how Sellars works out his own theory of objects of experience by way of reconstructing Kant’s discussion of the status of the ›ontological categories which the Cartesian tradition […] had applied to thoughts‹ (KTI §12). This explains why Sellars never tires of saying that Kant construed the status of our engagement with the world as one we only come to terms with when we see that he applies, at basis, Cartesian terminology to it. This concerns, foremost, the distinctions between formal

Transcendental Psychology 41 and objective reality and existing inside and outside thought. So when Sellars follows Kant closely in addressing the concept of an object of representation, the Cartesian roots make themselves felt where, methodologically speaking, the direction of inquiry points from the inside of the content of a representation (a Vorstellung) to its (causal) place of origin, rather than the other way round. That is, the Kantian way of putting the question, ‘how do we have and what is the status of our claim to knowledge of objects?’ does not start with objects simpliciter. That there are objects, to the Cartesian, is a fact we do not start with, but we arrive at: Problematic idealism regards the claim that material things and processes exist per se as a coherent one, but one which can be established only by an inference from our perceptual states; an inference from effect to cause. (KTI §20) And it is this direction of inquiry then that Sellars finds in Kant’s too, where first we have to get clear about representings of sensibility to then be in a position to formulate what its objects are. The crux of the matter as Kant clearly saw is his account of what it is to be an object of acts of representing. He formulates this account schematically in a passage (A 104) which prepares the way for the first edition Transcendental Deduction, and develops it in a more full-­ bodied way in a key passage in the Second Analogy which occurs in both editions. (KTI §37) As Sellars does not tell us explicitly which passage he has in mind, I propose it be the schematic formulation in A104 that draws heavily on the distinctions between appearance and in itself and between existing inside and existing outside acts of representings: And here then it is necessary to make understood what is meant by the expression “an object of representations.” We have said above that appearances themselves are nothing but sensible representations, which must not be regarded in themselves, in the same way, as objects (outside the power of representation). 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 some thing in general = X, since outside of our cognition we have nothing that we could set over against this cognition as corresponding to it. (A104)

42  Transcendental Psychology The passage is schematic in that it relates two theoretical notions, a representation and its purported object – the thing at which it is directed. It pushes the question of what it is to articulate an account of an object’s existence outside the bounds of a representation. The thought seems to be that when we try to think of something that corresponds to and yet is distinct from our cognition it needs to be ’something in general‘, an idea that Kant marks with the letter ‘X’. We can interpret the reason Kant finds for this – since outside of our cognition we have nothing that we could set over against this cognition as corresponding to it – as requiring the following line of thought: If we are to merely think of an object outside of a cognition, we cannot think of it as determinate, as that would mean ‘setting it over against something else’ or contrasting it with the determinations given to it in the cognition of it. All that remains is that there is a thing that is merely general in being negatively characterised as outside the grasp of our faculties enabling cognition. While this first part of the passage makes it seem as if there’s very little to be gained from stipulating something to be outside our thought, the subsequent passage makes it clear that the transcendental reflection on the theoretical idea of there being something that is a thing for us, that enables our cognition of it, introduces an element of necessity. Note in the following how it is a notion of necessity that is arrived at by way of the method of presuppositional analysis in the sense we encountered before (Chapter 1.2). Kant writes: 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 in so far 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) There are two key ideas in the passage. First, in order for there to be unity among our cognitions, the concept of an object has to be regarded as unified in them. And second, the notions of unity that show up in this account of object cognition are such that they carry an element of necessity. A further important idea we need to highlight concerns the fact that cognitions are characterised such that they purport to be about an object. The notion of object cognition, therefore, contains something like a direction, a demand, a claim to something, a Denknotwendigkeit. And it is that notion that will stay with us for a little longer.16 Let’s unfold the considerations entailed by this one sentence in more detail. I understand the aim of that sentence to be that of articulating the interdependence of object and thinking: An interdependence elucidated

Transcendental Psychology 43 via the centring notion of unity. So the task will be to explicate how it is that in order for the cognitions to be able to be all about one and the same object, they need to necessarily share a unity among themselves, which as a unity thereby also is taken to correspondingly be in the object. When Kant writes that ›our thought of the relation of all cognition to its object carries something of necessity with it‹ (A104), we can understand him as analysing the notion of a cognition’s intentional relation to an ‘object’ of thought. And it is in this analysis that Kant can say he ‘finds … something of necessity’. It is the notion of necessity we introduced in the first chapter under the heading ‘analytic truth in the tough sense’. Kant is not presupposing anything specific about objects or their uptake; rather, he is analytically uncovering what presuppositions go into our concept of an object überhaupt. This means, from within the framework we have used to characterise Sellars’ agreement with Kant’s method, we can say Kant here operates on M2 with regard to the concept of an object that is to help us secure a contact or friction with the world. And as the world is to show up as the content of cognitions concerning objects in nature, it is all the more important to understand the requirements on cognising objects. Let’s reconstruct the reason Kant gives for bringing up the notion of necessity appealed to. What does he mean by it? According to the wording of the sentence, we are told, we ‘find that the thought of how all our cognitions are to be related to objects carry something of necessity with it’. since namely the latter [i.e. object] is [so] regarded as that which is opposed to our cognitions being determined at pleasure or arbitrarily rather than being determined a priori (A104) The reason Kant gives for introducing the notion of necessity at this M2 level is that otherwise, our cognitions would be arbitrary. And the opposite conception, at this level of generality, to being arbitrary is ›being determined a priori‹. Above (Chapter 2.1.1), we have encountered the claim that all thought of any subject matter has to be given to us through acts of classification in order to be something for us at all. The most explicit form we can invoke here would be the logical forms of judgement structuring thoughts about any subject matter whatsoever. I take it that this appeal to the content-determining function of categorial classification is what Sellars has in mind when he suggests that A104 ›prepares the way for the first edition Transcendental Deduction‹ (KTI §37). With the difference being that here the content we are dealing with is even less specific, as we are yet to understand the relation between cognitions and their objects. The key idea in the substantiation Kant provides in that sub-clause can be put this way: In order for there to be coherence and structure

44  Transcendental Psychology among our cognitions, we need the concept of something corresponding to them, something that gives us occasion to consider an array of cognitions as inherently similar and thus to be ‘of the same content’.17 And while this formulates a requirement on our cognitions, in the next sub-clause, Kant moves to the implications this has for the concept of an object: 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 the cognitions are to be related [sich beziehen sollen] to an object, they need to be in agreement with one another and exhibit a unity in them that makes up, constitutes, represents, or amounts to the concept of an object. The closing sentence of this passage presents in detail the analysis of what it is achieved in the concept of an object: It is to be that, which gives unity to various representations. We have chosen this more general formulation ‘representation’ of which cognitions are a special subclass, to be able to highlight a continuity of the elaboration given here with the one in the Second Analogy, of which Sellars claims that it ›develops it [i.e. the concept of an object] in a more full-bodied way in a key passage […] which occurs in both editions‹ (KTI §37). That key passage states: [T]he object is that in the appearance which contains the condition of this necessary rule of apprehension (A191/B236) Note how in taking up the theme of answering ‘what it is to be an object of acts of representing’ Kant has ‘appearance’ instead of ‘cognition’ and articulates the notion of ‘necessity’ in terms of ‘rules of apprehending’. Irrespective of these terminological differences, we can summarise the key assertion contained in both passages (A104 and A191) as follows: All reference to objects necessarily contains a notion of unity which is generated from within our purport (Anspruch) to cognise objects as that which all cognitions have to presuppose such that they can share it. I have chosen this formulation to highlight the interdependence of the respective notions of unities. For I read the analysis to imply a rejection of two alternative readings: ( a) The object’s unity constitutes unity among cognitions. (b) The unity among cognitions constructs objects.

Transcendental Psychology 45 By solely focusing on the closing of the A104 sentence, one could arrive at (b): ›they must have that unity that constitutes the concept of an object‹ (A104). Whereas a too selective and one-sided reading of the directly preceding sentence can lead to (a) ›insofar as they are to relate to an object our cognitions must also necessarily agree with each other in relation to it‹. We can obviate both by placing emphasis on the methodological reflections it contains. Cognitions can be seen as characterised with regard to their ‘purporting to be directed at something as their shared content’. And this shows when we take into consideration how this passage wants us to consider what is necessary for explicating the ‘thought of the relation of all cognitions to its object’. In other words, we are at the level adding detail to the pure accusative, the mere X about which we attempt to gain insight. This means, long before we can even try to make claims about any direction of determination, we need to introduce the further concomitant idea of there being something that enables corresponding structural presuppositions. The cognitions have to have features that are such that they can be said to mirror the features of that, of which they are to be cognitions. And when Kant begins to fill in the notion of an object as ‘that which is opposed’ (was dawider ist) to our cognitions, he elaborates the idea that a content of such kind has to be radically distinct from a cognition. We can imagine the following objection: Granted that we can have meta-cognitions, representations of representations – but with regard to their content, their aboutness or their of-ness, these do not effect that first order representations thereby get to have the requisite independence built into them, such that their content could be said to be ‘opposed’ to them. Sellars here is concerned with the question of how such a content can in this sense be ‘opposed to’ cognition while also be taken up into cognitions and necessarily exhibiting a structure therein. The systematic aim here is, on its flip side, to articulate the concept of an object of experience. Here we have been given only the first half of the most abstract consideration of what necessarily has to enter into such a concept. A further aim is to qualify the notion of an object with regard to its metaphysical status as an entity bearing the relevant kind of independence from us. It is this requirement which ushers in a focus on rules of synthesis. In what follows, we will show how Sellars not only deals with A104 but also covers the main ideas inherent in the subsequent paragraphs of A105, which address the question of what it is for a content of manifold to be taken up into rule-guided acts of representing in such a way as to yield objective contents.

46  Transcendental Psychology 2.2.3  The Role of Synthesis The aim of this subchapter is to reconstruct Sellars’ interpretation of the role of synthesis in Kant’s account of the concept of an object of experience. This involves clarifying the respective roles of conceptual and sensory components and how they interact in the theory of empirical synthesis. Sellars’ Kant exegesis takes passages from A104 and A191 to be central. I look at them independently (Chapter 2.2.3) before reconstructing what Sellars makes of them (Chapter 2.2.4). This involves saying something about the relation between two concepts: Firstly, the concept of a unity of which one can be conscious; and secondly, the concept of a sequence of acts of representing. According to our reconstruction, Sellars invokes the Cartesian act-content distinction to argue for the thesis that the status of acts of perceiving (their actuality) and the status of the object perceived are interdependent. This has consequences for how we are to understand the ontological status of an object of experience. According to Sellars’ account of perceptual experience, Kant’s doctrine of synthesis (Synthesislehre) necessarily plays an important role in any correct account of the status of objects of experience. The thesis that Sellars defends can be reconstructed as follows: We cannot talk about objects nor about their metaphysical status independently of previously having clarified exactly how the manifold and our perceptual acts hang together. That is what the doctrine of synthesis is designed to do. From a slightly different angle, we could also explicate the idea thus: While in A104, we learn something about the respective notions of unity among cognitions and their object, the articulation of this interdependence is given in the abstract. As soon as we attempt to have our account do justice to the specifics of perceptual experience in every day scenarios, we will need to say more about their dynamical nature. Every single one of my perceptions is a perception extended in time. It has a beginning and an ending. The objects of my perception may be partly occluded; I gain more information about them by and by. The details in my perceiving may gradually become more precise, and so forth. So it is one thing to establish the thesis that there needs to be a notion of unity in our cognitions and their contents. It is another, and one requiring further elaboration, to establish how such a unity is brought about. The term Kant has on offer here is synthesis as an activity that brings isolated items into a coherent unit. Objects do not simply pop up as units in our cognitions. Rather, we need an account of how the material on which our cognitions operate becomes material about which we can form judgements.

Transcendental Psychology 47 If we interpret the role of Kant’s doctrine of synthesis from this angle, we follow the order of inquiry correctly as prescribed by the transcendental method. After all, we want to know what it is to be an object of experience and putting that question first implies that we cannot introduce any unwarranted assumptions about objects. What we can introduce is the notion of our passivity with regard to our receptivity capacities. For an account of our fundamental epistemological situation will have to explain sooner or later that we receive impressions or minimally construed ‘impingements’ from the world. Although we cannot make any claims about its determinate nature at this point, what we can say is that we can consider it a manifold.18 And it is this manifold that ends up being the material on which our cognitive capacities come to operate. Kant’s way of explicating the details of such an operation is to delineate a theory of synthesis. To support our reading of Sellars’ endorsement of Kant’s account about the role of synthesis activity, we will look into Kant’s text before discussing the interpretation given in KTI. In the passage directly following A104, Kant connects the aforementioned notion of unity and necessity with elements coming from two opposite directions, with the manifold of sense and with the formal unity of consciousness i.e. transcendental apperception. He writes: It is clear, however, that since we have to do only with the manifold of our representations, and that X which corresponds to them (the object), because it should be something distinct from all of our representations, is nothing for us, the unity that the object makes necessary can be nothing other than the formal unity of the consciousness in the synthesis of the manifold of the representations. (A105) With regard to our direct epistemic access, Kant points out; we are in touch only with the manifold of our representations. That there also be an object, the X, which corresponds to them was a thought we introduced as the underlying transcendental presupposition. The first concerns our phenomenology, the latter the methodological point that for there to be a cognition it needs to be a cognition of something. According to the Kantian line of thinking, as Sellars reads him here, the manifold of our representations eventually adds up to having only the status of appearances. In a first approximation, appearances are the contents of acts of representing. As such, they are what we can come to be in conscious contact with, prior to any further evidence guaranteeing that something matter-of-factual corresponds to them. It does not so much matter that we are confronted first and foremost with appearances because all that really matters is that the material we are to take up in our acts of synthesis has the function of providing

48  Transcendental Psychology the ground for cognitions that can acquire the status of being ‘world-­ directed’.19 To what extent they succeed in being that will be of further concern to us later. 20 For now, it suffices to say: To cognise an object that is more than merely thought of or imagined is to think something that is unified in the cognition of it. And Kant conceives of a cognition as something to which a synthetic process leads up to. That is why he writes: It is clear, however, that […] the unity that the object makes necessary can be nothing other than the formal unity of the consciousness in the synthesis of the manifold of the representations. (A105) To be in possession of the notion of an object is to have brought about formal unity in our consciousness of the synthetic process which takes up the manifold and works it into a unit: A unit constructed in an orderly synthesis. Consequently, Kant can write: Hence we say that we cognize the object if we have effected synthetic unity in the manifold of intuition. But this is impossible if the intuition could not have been produced through a function of synthesis in accordance with a rule that makes the reproduction of the manifold necessary a priori and a concept in which this manifold is united possible. (A105) Kant tells us that synthetic activity ultimately has to result in an intuition. Therefore such an intuition depends on having been brought about under the reign of a rule-guided process of reproduction. Here, Kant refers to the three stages of Synthesis that make up part of the so-called A-deduction. Any singular item needs to be taken up, needs to be reproduced in order to be such that it can be present next to a further item, and the result of these operations needs to be thought in a concept.21 To give this abstract formulation a vivid example Kant adds the following qualification: Thus we think of a triangle as an object by being conscious of the composition of three straight lines in accordance with a rule according to which such an intuition can always be exhibited. Now this unity of rule determines every manifold, and limits it to conditions that make the unity of apperception possible, and the concept of this unity is the representation of the object = X, which I think through those predicates of a triangle. (A105) The analysis of the unity in our consciousness of an intuition of a triangle exhibits all the aforementioned characteristics. It is, as consciousness,

Transcendental Psychology 49 concerned with holding together three separate lines and needs to reproduce each one of them in order to arrive at the superordinate content of a triangle. But in order to have this synthetic activity end up in the intuition of a triangle, Kant tells us, we need a rule guiding the synthesis in such and such a specific, necessary way, lest we intuit a U-shaped figure instead of a triangle. All cognition requires a concept, however imperfect or obscure it may be; but as far as its form is concerned the latter is always something general, and something that serves as a rule. Thus the concept of body serves as the rule for our cognition of outer appearances by means of the unity of the manifold that is thought through it.22 (A106) But it is not just the interplay of determinate singular items constituting the manifold we come to intuit and the general concept serving as the rule for its apprehension and reproduction. It is also the point that this entire operation implies a reference to the activity of the transcendental self because the entire activity’s outcome has to be something that eventually can become a content for the epistemic agent. That is why Kant relates the aforementioned conceptions of necessity back to the highest point of abstraction, the transcendental apperception. Every necessity has a transcendental condition as its ground. A transcendental ground must therefore be found for the unity of the consciousness in the synthesis of the manifold of all our intuitions, hence also of the concepts of objects in general, consequently also of all objects of experience, without which it would be impossible to think of any object for our intuitions; for the latter is nothing more than the some thing for which the concept expresses such a necessity of synthesis. Now this original and transcendental condition is nothing other than the transcendental apperception. (A106–7) Kant here approaches the notion of an object once more from the perspective of the transcendental self, or the I which does the thinking and perceiving – an agent whose synthetic activity has to be accompanied by the very possibility of being able to ascribe such an activity to itself, as the source of the very activity. The unity of consciousness therefore means the very idea that whatever comes to be present in front of me as the content of my acts of apprehensive and reproductive acts of representing needs to be such that I can consider it as mine. In that spirit, Kant can write: Now no cognitions can occur in us, no connection and unity among them, without that unity of consciousness that precedes all data of

50  Transcendental Psychology the intuitions, and in relation to which all representation of objects is alone possible. (A107) At this point, Kant can be seen to have established that all aspired purported representation of objects outside of us requires an apprehensive synthetic activity which involves in its last, recognitive, stage the contribution of general concepts. About this role of concepts in the fundamental synthetic activity, Kant adds: This pure, original, unchanging consciousness I will now name transcendental apperception. That it deserves this name is already obvious from this, that even the purest objective unity, namely that of the a priori concepts (space and time) is possible only through the relation of the intuitions to it. The numerical unity of this apperception therefore grounds all concepts a priori, just as the manifoldness of space and time grounds the intuitions of sensibility. (A107) What exactly does Kant mean when he claims the transcendental apperception grounds all concepts a priori? He elucidates this particular claim by correlating it with the way in which “the manifoldness of space and time grounds the intuitions of sensibility”. By this, Kant argues that we can explicate the way in which the manifold is constitutive of our spatio-­ temporal sensible intuitions. I read his argument as follows: Without there being a formal aspect that provides so much as the framework within which we can be conscious of our intuitions, they would be impossible. Since, in our case, intuitions are always intuitions of spatio-­ temporal items, we can additionally read it to mean that the transcendental ideality of space and time a fortiori carries over to the status of items we perceive as being in space and time. This would imply that they (the objects) too are necessarily transcendentally ideal. This implication is important for it, allows Kant to draw a distinction between that which is transcendentally speaking ideal while retaining a sense in which it is empirically real. Now transcendental apperception is said to be in a similar way constitutive of all our concepts a priori. Why is the grounding relation similar? Because ‘the numerical unity of this apperception’ is the highest point we can think at all, which matters when we remember that the pure concepts of the understanding are concepts for thinking unities. Here we are told that also the a priori concepts of space and time depend on this highest point. Why? Because the transcendental apperception is the first and highest form of ‘being a unit’, a unity of self-consciousness. In other words, if one is to think a content, one has to start from this highest point. There is no point beyond or above it, such that one could

Transcendental Psychology 51 meaningfully think a content (and not just a formal thought, e.g. S is p), for to think something is to think of it as reflected in this numerical apperceptive unity. And this notion of unity is distinct and prior to the category of unity we later on learn about in the table of categories. It is a notion requisite for thinking more determinate contents with spatio-temporal coordinates. Another way of formulating the reflections on the conditions of possibility Kant carries out here is to say that the transcendental ground is the origin of unity in the following sense: All my mental acts, be they concerned with conceptual items or the transformation of sensory manifold material into classifiable contents, need to be such that they can be thought of as mine. This line of thought finds Sellars’ wholehearted agreement, an agreement which he expresses by saying that as a theme it adds an essential detail to an encompassing account of synthetic activity, such that we can understand ›Kant’s conception of the transcendental unity of apperception as a necessary correlate of the intuition of objects‹ (KTI §39 fn18). In view of this qualification, we can turn to Sellars’ treatment of this synthetic activity, looking for cues as to how he sees this correlate as built into it. Do we find evidence in Sellars that brings the respective roles of synthesis and transcendental apperception into one context? We do when we look at the place where Sellars reflects on the details of the contribution made by rules to synthesis activity in the context of perceptual experience: Our primary concern is with perceptual acts or takings. But in the first passage referred to above [A104], Kant makes his key point in a way which abstracts, as the [A190-191] passage in the Second Analogy does not, from essential aspects of perceptual takings. Nevertheless, the concepts for which Kant is preparing the way is that of rules for generating perceptual takings. (KTI §38) Here Sellars gives his interpretation of the significance of A104, in contrast to A190, as concerned with the question of how rules figure in how we take in the world. And according to his reading, Kant aims to explicate the said necessity inherent in the thought of the relation of our cognition to its object as requiring a rule bringing order into our acts of representing. And it is here, in a comment on the problematic nature of such synthesis-guiding rules, that Sellars connects the theme of synthetic activity with transcendental apperception: The term ‘rule’ is a dangerous one, for it suggests deliberate activity or, at least, activity which would be deliberate if it weren’t so

52  Transcendental Psychology hasty and, in the ordinary sense, thoughtless. Actually the most useful concept is that of a sequence of acts of representing which can reflectively be classified as conforming to a rule which is (at least in principle) graspable by thought. The rules in question must […] be available, if one is to recognize that one’s acts of representing belong together as an intelligible sequence. (KTI §39) If we are to understand the contribution made by such rules properly, Sellars argues, they cannot be construed on the paradigm of rules for action (LTC 508). 23 Rather their explanation is best given with respect to the standpoint of the transcendental apperception from where we can reflectively classify such a sequence and recognise it as mine. This is how I understand the claim that the transcendental apperception is a necessary correlate to the intuition of objects. Sellars’ commentary on how rules guide the constitution of our representings’ content concerns very basic acts. They are so basic that the conscious availability of rules for them is an open question. The point of the comment is to make explicit their constitutive role for acts of representing as belonging together in an intelligible sequence. Here Sellars addresses the notion of necessity Kant invokes in A104. Our representations considered as acts of representing can be brought into an intelligible sequence. The idea of such a sequence in which acts of representing can be said to necessarily agree with each other is constituted by their purporting to be about the same content. That is why Kant continues by saying: ›Now this unity of rule determines every manifold and limits it to conditions that make the unity of apperception possible‹ (A105). Such a unity of rules guides my synthetic acts of apprehending a manifold. I need the concept of a unity prior to perceiving a unity because I need to bring order into the manifold I am presented with. With regard to the triangle example Kant can therefore say ›the concept of this unity is the representation of the [transcendental] object = X, which I think through those predicates of a triangle‹ (A105). The transcendental object is a unity-presupposition on my behalf, which I, in a transcendental apperceptive act, have to bring in, such that it enables and guides the synthesis acts of perceptual content. Both unities can be identified as transcendental presuppositions: The unity definitive of my apperception and the unity definitive of a transcendental object. Both ground the dynamic synthetic process of combining a manifold in a sequence of individual acts of representing (cf. A129–130). 2.2.4  Successiveness, Subjective, and Objective With these considerations about the transcendental function of the concepts of unity and necessity at the most abstract level of object constitution in place, we now arrive at the pending question of how

Transcendental Psychology 53 this makes object-perception possible. How can there be independent objects and in what sense can we claim them to be objectively out there, if all that we can ever be concerned with are sequences of acts of representings? It is this worry that Sellars addresses when he writes: Now it might be thought that by introducing the concept of a rule-­ governed sequence of perceptual representings, i.e. acts of perceptual taking, I am giving hostages to the view that material objects consist of rule-conforming sequences of perceptual takings. (KTI §40) Sellars here aims to avert a constructivist reading of what he feels one might read into Kant. At the centre of such an interpretation would lie the claim that the reality of material objects is nothing else but to be rule-conforming sequences of perceptual episodes. But is Sellars putting forward such a bad phenomenalism in his analysis of Kant’s theory of perceptual experiences? Sellars is telling us that to argue in this way is precisely not the function of Kant’s doctrine of synthesis. So he quickly replies: That I am not is implied by the fact (which I hope to make clear) that even the most tough-minded transcendental realist grants that veridical perceptual takings have the coherence which Kant is attempting to clarify by the concept of rule-conforming sequences. Roughly, Kant’s transcendental realist thinks of the perceiver as deriving these rules by induction from experience, whereas Kant thinks that induction itself presupposes an antecedent grasp of these rules. (KTI §40) The relevant claim Sellars wants to make is a different one. But if his point is not to argue for the claim that objects consist really of nothing else but rule-conforming acts of perceiving, what is his point? It is to argue that Kant’s account of what it is to be an object of acts of representing (KTI §37) develops a conception of actuality as a mode of existence that is defined by lying between the two extremes of (a) being merely imagined or constructed and (b) existing per se or in itself. That is why the passage quoted addresses the constructivism inherent in the assumption that objects are nothing but logical constructs out of mental states (cf. KTI §36) and opposes it to a direct realism. This would be a transcendental realism that claims our perceptual acts put us in touch with objects and allow us to inductively ‘read off of them’ the rules for their synthesis (cf. Rosenberg 2007, 230–242; and Westphal 2004, 250 ff.). How then can Sellars read Kant as falling into neither position?

54  Transcendental Psychology In view of these alternatives, we can return to a detail of a passage in A105 where Kant writes: Thus we think of a triangle as an object by being conscious of the composition of three straight lines in accordance with a rule according to which such an intuition can always be exhibited. (A105) In his commentary on this passage, Sellars refines his interpretation of Kant’s position: He asks us to consider the intuitive representation of a triangle. Here the rich implications of the concept of a perceptual taking are laid aside, for the moment, and we are given an explanation which could concern a construction in pure geometry. For the essential point he wants to make is that while the object of the intuitive representing is indeed a triangle, the triangle is not an existent per se, and that although the content triangle specifies sequences of representing which count as coming to represent a triangle, the object of the representing of a triangle is not the sequence of representings which culminate in the representing of the triangle. A triangle is neither a mental act of representing a triangle, nor is it a sequence of mental acts each of which represents a part of a triangle. (KTI §41) The notion of an object is to be differentiated from sequences of acts of representings, from existents per se, and from individual mental acts. I read all of this as Sellars’ way of preparing the ground for Kant’s discussion in the Second Analogy. There the question begins to turn on the appropriate interpretation of subjective and objective successiveness. It is on this way that we arrive at the following question, which I take to be at the centre in A190. How is an account of ‘object perception’ possible that brings in rules as making a necessary contribution to our understanding of it as being ‘objective’? What is the contribution made by rule-conforming sequences if we are not to suppose that objects are reducible to mere subjective successiveness? On this Kant writes: The apprehension of the manifold of appearance is always successive. The representations of the [triangle’s] parts succeed one another. Whether they also succeed in the object is a second point for reflection, which is not contained in the first. (A190/B235) From the points made in A105 we have learned that Kant holds that all intentional directedness at objects requires us to undergo successive acts

Transcendental Psychology 55 of synthesis apprehending the manifold of appearances. The question now is what is involved in the thought of something as existing beyond this subjective successiveness on the part of our acts of apprehending. Before Kant ties the notion of successiveness back to the rules which we can reflectively appeal to in order to account for our acts’ conformity, he needs to add a vital detail to the concept of an ‘object’: Now one can, to be sure, call everything, and even every representation, insofar as one is conscious of it, an object only what this word is to mean in the case of appearances, not insofar as they are (as representations) objects, but rather only insofar as they designate an object requires a deeper investigation. (A190/B235) On the basis of our previous discussion (Chapter 2.1), we can interpret Kant here as applying the Cartesian distinctions of existing inside and outside of mental acts, of being an object of thought objective or formaliter. The content of an act of apperception can be called an object. It would be an object of thought in the traditional sense, a mere content of a mental act (cf. Chapter 2.2.2). That is why we have to qualify such way of talking by a commentary that makes explicit how endorsing this way of talking about mental content still only equals committing oneself to the object being nothing over and above a representational content. That is why Kant clarifies the status that objects of appearances as representeds have by adding: Insofar as they are, merely as representations, at the same time objects of consciousness, they do not differ from their apprehension, i.e., from their being taken up into the synthesis of the imagination, and one must therefore say that the manifold of appearances is always successively generated in the mind. (A190/B235) But how do we then come to have the concept of an object as something that is more than merely a construct out of our subjective acts of representing it? Kant addresses this question by arguing that we cannot simply help ourselves to the notion of ‘existing outside of our acts of representing’ by appealing to things in themselves. If appearances were things in themselves, then no human being would be able to assess from the succession of representations how the manifold is combined in the object. 24 For we have to do only with our representations; how things in themselves may be (without regard to representations through which they affect us) is entirely beyond our cognitive sphere. (A190/B235)

56  Transcendental Psychology Why is it of no help to appeal to things in themselves at this stage of the exposition of our positive account of object cognition? Because our program is to elucidate how we can arrive at a notion of mind-independence that proceeds through the means of the internal characterisations we have, i.e. the materials our theory so far restricts us to. To simply try to jump to things in themselves would be to violate these terms. The very distinction holds because of our reflection on the nature of our cognitive capacities up to this point (cf. SM I §22 ff.). Our acts of representing give us contents as represented contents. To suppose that such acts relate directly to objects out there is to abandon the transcendental project of securing a robust notion of intentionality. In view of these reminders about the restrictions, we operate under our question becomes: Can we attain a coherent concept of ‘mind-­ independence’ and ‘objectively being out there’ from within if all we have got is the manifold as it appears to us?25 Now although the appearances are not things in themselves, and nevertheless are the only thing that can be given to us for cognition, I still have to show [so soll ich anzeigen] what sort of combination in time pertains to the manifold in the appearances itself even though the representation of it in apprehension is always successive. (A190/B235, cf. also A129) In effect Kant argues for the conclusion that, in our theory production, all we can, in fact, work with are appearances. They are the only material we have to elaborate our cognitions’ claim to being successfully world-directed. But be that as it may, we still face the obligation to explain the status of the manifold that ends up in my subjectively successive acts of apprehending. That is why he persists in asking about the content of our appearances ›what sort of combination in time pertains to the manifold […] itself‹. Which we can render as the question: How do we explicate the notion of objective successiveness? How can we construe there to be a successiveness in the material which presented itself to us in our apprehensions? This move implies an abstraction away from how things appear to us, an abstraction that we thereby direct at something we conceive of as our antecedent to our appearances. In other words, the question is, how can we make room for the radical difference between what we have to work with and what we aim to entitle ourselves to? Our attempt of saying something about our intentional directedness points us in its purport outside the confines of mere appearances. It is the fact that the act of perceiving a large house, a skyscraper, in front of me takes time which alerts me to this difference. As it is too big to be seen at one glance, the sequence of my acts of perceiving it is successive – and in time – in a way in which their integrated content is not. And yet, I hold my acts of taking in the house to be about one and the

Transcendental Psychology 57 same thing. While this formulates the direction of inquiry, we also need to keep in mind the requirements we operate under. So when Kant issues a reminder about our fundamental epistemic situation Now, however, as soon as I raise my concept of an object to transcendental significance, the house is not a thing in itself at all but only an appearance, i.e., a representation, the transcendental object of which is unknown. (A191/B236) he does so in order to say with more precision what the question is, he tries to address: [T]herefore what do I understand by the question, how the manifold may be combined in the [house-]appearance itself (which is yet nothing in itself)? (A191/B236) Sellars comments on this particular moment in Kant’s analysis. We can now rephrase what Sellars is saying there: By arguing that the house is not to be construed as a thing in itself, Kant averts a transcendental realist reading. Likewise, he aims to avert a dogmatic26 idealism or constructivist reading by arguing the following: It is true that the content we represent under the title house allows us to specify sequences of acts of representing which count as coming to represent a house. Yet, the core notion we presuppose as underlying all our acts of representing a house is not that of a sequence of representings. Those are two different things. A house is more than the sequence of representings constituting the content of a mental act. Sellars concludes: ›A house is neither a mental act of representing a house, nor is it a sequence of mental acts each of which represents a part of a house‹ (cf. KTI §41). If a house is neither a mental act nor a sequence but something distinct from both, then we need to ask: how do these elements stand to the object house? Here it becomes apparent that Kant is trying to answer the question of how we get from an explanatory state (1) where we only have our elements (acts, sequences of apprehending a manifold into an appearance, represented contents) to (2) achieving a result that is more than these elements. To get from one to two, Kant needs a conception of object that steers clear between the constructivism and the transcendental realism. So what are we doing when we consider the manifold as combined in the appearance itself (in der Erscheinung selbst)? Kant elucidates how he wants us to analyse such a thought: Here that which lies in the successive apprehension is considered as representation, but the appearance that is given to me, in spite

58  Transcendental Psychology of the fact that it is nothing more than a sum [Inbegriff] of these representations, is considered as their object, with which my concept, which I draw from the representations of apprehension, is to agree. (A191/B236) Here Kant is walking us through the resulting architecture of the interacting elements in a way that allows us to comprehend the conceptual space he wants to secure for his account of objects. From a transcendental vantage point, that which lies in the successive apprehension is nothing but a represented content. But what is the something laying therein? Whatever it is, for a start, it presents itself to us through our acts of representing. It is considered as a represented, according to our analysis of our basic epistemic situation, which is that we have to do with represented contents of various sorts, and only with representeds, not with the in itself. But in so far as the sequence of acts of representings can be taken to be about one and the same content, we can also call such a content an appearance. Kant writes: ›[An] appearance is nothing more than a sum of these representations‹ (A191/B236). When we abstract from this characterisation we come to see that ›the appearance is considered as the object‹ common to a sequence of acts of representings.27 Therefore it cannot be identified with them. After all, Kant writes that from these apprehensive acts of representing, I draw my concept. For in the concept, I re-cognise i.e. think the individual represented contents as of one and the same thing. This implies that I need the notion of a thing and the corresponding concept. 28 In so far as my acts of representing are to be true of the object, my concept needs to be made to agree with the object. According to this analysis, Kant develops the concept ‘object’ out of the interplay between two types of representeds (a) successive acts of apprehending and (b) appearances. We see the method of abstraction at work in Kant’s analysis of object cognition. Our intentional directedness towards objects hinges on the assumption that we can meaningfully distinguish the representations of apprehension (i.e. our individual acts of apprehending parts of the house) from the object (house as that which gives rise to them). Kant articulates this requirement as follows: One quickly sees that […]29 appearance, in contradistinction to the representations of apprehension, can thereby only be represented as the object that is distinct from them if it [the apprehension] stands under a rule that distinguishes it [as a specific object-apprehension] from every other apprehension, and makes one way of combining the manifold necessary. (A191/B236)

Transcendental Psychology 59 Kant argues that if we are to successfully draw a distinction between the representations of apprehensions and an appearance we can do so only by considering this appearance’s determinate apprehension as the result of a rule. An appearance has to be apprehended in a rule-guided way. For only if it stands under a rule which selectively combines the features of my apprehensive representations into one can it thereby individuate a determinate representation. To pick up the example from above: A skyscraper presents itself to me in my appearance of it. There is a distinction to be made between my individual apprehensive representations and the resulting appearance. Yet this distinction I can only draw on the condition of (2) representing the ›appearance … as the object that is distinct from them if‹ the apprehension stands under a rule. And on the prior condition (1) that the rule makes one way of apprehending necessary. The second condition brings in the theme of treating what is inside of me, my appearance, as something that is actual outside of me. The first condition brings in the claim that such a re-locating is licensed by a rule. Kant concludes this line of reconstructing what goes into the concept of an object of experience by writing: That in the appearance which contains the condition of this necessary rule of apprehension is the object. (A191/B236) The objectivity we need, Kant tells us, is to be explained with regard to a necessary rule of apprehension. This rule has its origin (so we have to presuppose) not in ourselves but in the object thus synthesised. Why do we have to presuppose it to be contained in the object, and what makes it a condition on our apprehending? I understand the thought entailed in this concluding sentence of Kant’s analysis to amount to the following: If I were to synthesise arbitrarily, there could not be object-constancy.30 But if I suppose my synthesis to stand under a rule of apprehension, a rule that I consider to be a matter of the object’s being so and so, I thereby gain the thought of my apprehensions’ being determinately structured. As a rule-guided act of apprehending which yields a specific content, it can be made to agree with other such apprehendings and on that basis be about the same thing. But how do we deal with the constraint that the same content we thereby aim at cannot be thought of as a thing in itself? We see from the following passage dealing with this constraint that Sellars is on the same page and tries to articulate how one can read Kant as formulating a coherent thought about this very problematic here: Kant is answering the question ‘In what sense is the house the object of successive acts of apprehension?’ His negative answer is that it is not qua house in itself. The actuality of the house qua object is not its existence per se. What is his positive answer? As I see it, he is telling

60  Transcendental Psychology us that the house qua object is that aspect of the content of the perceptual takings which explains (together with certain other factors) the belonging together, as state of the perceiver, of certain perceptual takings (apprehendings). But that aspect of the content of these perceptual takings is simply the content house which they share, thus house over there left-front-edgewise to me house over there facing me left side of house over there facing me etc., etc.

(KTI §48)

To which a paragraph later he adds: Thus, the concept of a house as a perceptible object essentially involves a reference to perceptual acts, i.e. to the perceptual takings of a perceiver. (KTI §49) Sellars has two points about Kant’s positive answer. First, the notion of an object is defined as the common content of various apprehendings. A portion of the apprehendings is considered as their central aspect. And it is this aspect in them that allows them to share the same content. In our analysis, this depended on the condition that the object thus construed was the source of rules for putting together various acts of apprehending. Why do they need to be put together? It is because of the second point Sellars has, which states that a sequence of apprehensive acts is always necessarily point-of-viewish. Each of my acts of perceiving something is an act that takes place from a certain point of view, a view which differs minimally to any subsequent act. And it is the notion of a rule on the basis of which we can make the move from a sequence of pointof-viewish apprehendings to something that as their common content is conceived to be non-point-of-viewish, which amounts to an object that can be confronted from various perspectives. Sellars aims to summarise this train of thought in the following way: The object of a perceptual representing of a house is the non-­ perspectival content house; yet as the sort of item that can be the object of a perceptual representing, it must provide rules for explaining (together with other factors) why such and such sequences of perceptual takings with perspectival contents were necessary. (KTI §50)

Transcendental Psychology 61 And such a move is to do with ‘other factors’ that have to come to play a role in our resulting conception of perceptual experience. Sellars prepares for this specification by arguing that a mere geometrical construal does not suffice. And it is this specification that accounts for the shift in examples we can retrospectively notice from A104 to A190. In A104ff., we considered a geometrical construction of a triangle by way of abstracting from the specifics of our perceptual takings. But such takings that are richer in sensory content, as the takings of objects, are acts of the structure ›house in front of my sightful eyes‹, as Sellars stresses (KTI §51). And if we attempt to do justice to this fundamental perspectival nature of our purported uptake of objects, we need to say something like the following: House by itself can generate no explanation of the occurrence of a sequence of perceptual takings. It is only house in such and such relations to a perceiver[‘s body and sense organs] which can do this. […] In my argument I have thinned out this mutual involvement of object, circumstances and embodied perceiver into a ghostly ‘object from a point of view.’ But Kant took seriously the fact that perceivers are embedded in a spatio-temporal system of interacting substances. (KTI §51) The core idea is this: Only when we acknowledge this mutual involvement (of object, circumstances, and embodied perceiver) we understand the fundamental material structure of our own epistemic situation with regard to empirical knowledge. Objects present themselves to us first and foremost as the content of sequences of perceptual acts. This content is always perspectival. Yet, we do not normally think of our perceptual takings of objects as perspectively structured. We take ourselves to directly perceive the objects our perspectives have in view. However, we only come to conceive of the shared content of a sequence of our takings as a non-perspectival object by presupposing that they are about the same object. This implies the claim that it is the object whose robust non-­perspectival existence gives rise to a necessary rule guiding our sequences of apprehending it as an object (A191/B236). If I walk around a house, first from the left and then from the right, I can expect to undergo a similar but reversed sequence of perceptual takings. Here the notion of the non-perspectival, invariant object helps explaining why such and such sequences of perceptual takings with perspectival contents were necessary. So how does this elucidation help us to understand the legitimacy of the move from contents to their object? There is something of a necessity in this presupposition on our behalf. Prior to having the concept of an object, Kant is telling us, we need to consider our rule-guided synthetic

62  Transcendental Psychology acts of apprehending as accountable to the object they are about. Yet, the object is never given to us as it is but only through our synthesis of it. In assessing the legitimacy of Kant’s analysis, Sellars concludes Kant, in fact, steers clear between the transcendental realist and the constructivist. But let us pay close attention to how Sellars cashes out Kant’s position: Kant denies that material things and processes exist per se [an sich], but he holds that in the critical sense they can be actual as contents which make an essential contribution to the explanation of the patterns in which perceptual experiences occur. (KTI §52) With regard to A191, we can reformulate the characterisation given here as follows: Although Kant’s analysis of our epistemic situation only allows for representations and not for things in themselves to be given to us, we can derive a meaningful concept of an ‘object’ from the notion of a specific class of representations, for we have learned that we can consider appearances as units which systematically hang together with perceptual takings. And it is the notion of a rule-guided synthesis that entitles us to the claim that our sequence of subjective perspectival perceptual acts has its origin in a non-perspectival object. This object thereby stands at the end of our analysis in the order of knowing while we presuppose it to be at its beginning in the order of being. In this sense, the concept of an object (e.g. a house) is neither a thing in itself nor a mere construct but an actual content. The house is that which makes an essential contribution to how we are to explain the perspectival structure of our perceptual experiences. The first conclusion that Sellars draws from this is that Kant’s idealism with regard to the status of objects of experience is a coherent position. It is coherent in the sense we encountered above. For it can cash in a realist intuition which even the most tough-minded transcendental realist wants to be in a position to explain, i.e.: The fact that veridical perceptual takings have the coherence which Kant is attempting to clarify by the concept of rule-conforming sequences. (KTI §40) Think of how my perceptions of a house allow me to continue perceiving the same house, no matter which direction I choose in walking around it. What allows them to be veridical perceptual takings is the claim that they are takings of the same content. A content about which Kant tells us that it is in a critical sense actual.

Transcendental Psychology 63 According to Sellars, Kant’s analysis of the concept of the perceptual takings’ intentional relation to their objects has a further implication: But the deeper thrust of Kant’s transcendental idealism is the thesis that the core of the knowable self is the self as perceiver of material things and events. (KTI §52) The implication stated here is fundamental in the following sense. The subject of Kant’s theory of experience, the epistemic subject we take to have perceptual takings, is as transcendentally ideal as the object it is oriented towards. How does this follow from the thesis that ›the core of the knowable self is the self as perceiver of material things and events‹? It follows when we translate the abstract talk of ‘perceptual takings’ back into the perceptual acts of an embodied perceiver. This means the status we assign to objects carries over to the perceiver. How does Sellars argue for this? And if it is relatively easy to see that the distinction between actual and non-actual material things and events is tied to the concept of an actual sequence of perceptual takings, it has (until recently) proved less easy to see that the distinction between actual and non-actual sequences of perceptual takings, i.e. between perceptual takings which are correctly and those which are incorrectly taken to have occurred in one’s mental history, is tied to the concept of actual material things and events. (KTI §52) The argument rests on an interdependence between our perceptual takings and the objects they are about. Because in justifying our acts as correct acts, we already need to appeal to actual objects. There’s no other way for me to argue for the claim that my perception of the house is veridical unless I can also make use of the concept of a house that exists independently of my conceiving of it. Yet, the presuppositional analysis has had the result that we can only think objects as actual on the condition that they are the ground of our representations. This is a presupposition on our behalf, as we conceive of them as occasioning our synthesis of them by way of our rule guided perceptual takings. It is in view of this analysis of the interdependence of objects and perceptual acts that we can now make sense of the prima facie rather cryptic concluding paragraph of KTI: Kant saw that the concept of an object of perception contains a reference to the perceptual takings which are the criteria for its [the object’s] actuality. He also saw that the concept of a perceptual taking, as the taking of an object, contains a reference to material

64  Transcendental Psychology things and events which, if actual, would imply its [the taking’s] own actuality. The actuality of perceptual takings and the actuality of material things and processes are not logically independent. (KTI §53) This conclusion is about the following insight. From the vantage point of a transcendental analysis, we understand that the self as the agent of perceptual experiences in flesh and blood is as transcendentally ideal as are the objects it aspires to gain knowledge about.31 Because, only by being embodied and thereby part of the world of which one aims to gain knowledge, one can be said to be subject to the impingements of its objects. So it is the deeper sense of the idealism that the self and the world are both equally transcendentally ideal and share the same actuality. And that they are, is something that we can establish by way of a transcendental analysis of our concept of an object of experience. In light of this reconstruction of Sellars’ engagement with Kant’s theory of experience, one conclusion to draw is that Sellars’ understanding of the ontological status of objects is deeply Kantian. It is Kantian in the sense that he endorses the claim that perceiving something as an object requires an account of the role of rule guided synthesis: An account that explains the difference between subjective and objective aspects of empirical synthesis in terms of how rules govern synthetic activity.32 Sellars also agrees with Kant in holding that the objects we are confronted with in perceptual experience can neither be thought of as things in themselves nor as merely imagined phenomenal sense contents. Rather, their status is to be understood in terms of their actuality. And for Sellars, the task of clarifying the actuality of the objects perceived goes hand in hand with clarifying the actuality of our own acts of perceiving them. This interdependence-thesis about the explication of self and world will concern us in the next chapter, where we address a number of consequences that will draw Sellars beyond the metaphysical confines of Kant’s theory of perceptual experience.

Notes 1. This characterises the arc of our task until the completion of Chapter 5.4. 2. Another way to express this thought with regard to the First Critique: We know from the Aesthetic but mainly on the basis of the Metaphysical Deduction that the Pure Concepts of the Understanding are involved in thinking objects in nature, so the Transcendental Deduction serves as demonstration of our entitlement to the coherence of such knowledge. Sellars expresses this in the passage we already encountered in Chapter 1: ›It is […] obvious on reflection, that Kant is not seeking to prove that there is empirical knowledge, but only to show that the concept is a coherent one‹ (KTE, §11). 3. O’Shea (2011, 336) brings out how this project of delineating the transcendental psychology goes hand in hand with articulating invariant features across the development of frameworks, stressing how ›the ontology

Transcendental Psychology 65











of thoughts, in the sense of accounting for what it is to be a thought and what it is to be a thinker, remains the same account across changing conceptual frameworks, and will remain the same in the ideal scientific image of thinkers-in-the-world. Changes in the conceptual roles that constitute the contents of thinkable thoughts, however radical such changes might be in the course of scientific theory-succession, do not involve the replacement of the conception of thoughts as normatively rule-governed inferential role-players.‹ 4. Parallel writings (e.g. MFC, NI, SM IV, I) unfold the idea of functional meta-classification into a theory of meaning and conceptual change. 5. Concerning the criticism directed at an ‘Aristotelian’ category-theory: On the one hand, Sellars here does not seem too interested in doing justice to that tradition, on the other hand, that entire discussion might be of minor significance and Sellars does not aim for exegetical accuracy. Thus I leave the question aside whether a positive account of Aristotle’s categories can avert Sellars’ criticism. Cf. Thomasson (2018), Cumpa (2020) and O’Shea (2021b.). 6. The reference is probably directed at Strawson, who in The Bounds of Sense (1966, 32) criticises “the imaginary subject of transcendental psychology” — that is, one year prior to the publication of KTE. 7. For further discussion of this particular claim see the contributions in Miguens (2020). On Kant’s rendition, God’s intellectual intuitions except him from having to use the categories of the understanding. Cf. Kreines (2007), Förster (2012) and Haag (2012a) on this point. 8. The second half of the B-deduction can be read as designed to demonstrate our entitlement to apply the categories to objects we experience. Cf. Conant, in O’Shea (ed.) (2017, 120–139). 9. This interpretation has its parallel in a passage in TTC characterising the status of truths expressed in law-like statements: ›What Kant takes himself to have proven is that the concept of empirical knowledge involves the concept of inferability in accordance with laws of nature. To grant that there is knowledge of the here and now is, he argues, to grant that there are general truths of the sort captured by lawlike statements. As far as specifically human knowledge is concerned, he was convinced that the idea that knowable objects are located in Space and Time carries with it certain general commitments as to the form of these laws. These commitments could, he thought, be known a priori or noninductively. Thus, the transcendental knowledge that spatio-temporal objects of knowledge must conform to certain generalizations which are themselves logically synthetic is itself, according to Kant, analytic‹ (TTC §54). 10. For an extended treatment of Cartesian influence on Kant and for a reconstruction of Sellars’ later day view on colours, see Rosenberg’s (2007, Chapter 9), roadmap to the place of colours in the scheme of things. 11. Hoverer, here we address only the methodological question of why Kant subscribes to the Cartesian framework. For now, we bracket the further worry what metaphysical consequences Kant is justified in deriving from them. But see Chapters 5 and 6. 12. On this theme of overlooking Sellars writes: ›I have spoken of actuality where Descartes speaks of ‘formal reality.’ In Kant’s terminology what Descartes means by ‘formal reality’, and which Cartesians would equate with actuality, is ‘existence in itself’ or, to use the latinate term of the Prolegomena, existence ‘per se’ Kant clearly accepts the Cartesian contrast between ‘existence in thought’ and ‘existence per se’, – so much so that he takes it for granted, as did most of his predecessors‹ (KTI §15).

66  Transcendental Psychology 13. See Brassier (2011) for a related encompassing discussion, pushing this thought to its critical boundaries. 14. Any substantial reconstructive work of this dimension in Sellars’ reading of Kant has to say something about how his account changed over the years from SM to IKTE. A good place to look is Rosenberg’s (2007, Chapter 13), who sees this change in the move from a two staged to a three stages account of perceptual consciousness. The first co0ncerns just intuitions, the second also image-models. I will skip over the discussion of the historical intricacies of Sellars’ intellectual development and instead devote the energy to a reconstruction of the transcendental role of image-models. Cf. also Wunsch (2007, 2014); Horstmann (2018) for a recent exegetical discussion of the productive imagination’s status in Kant scholarship. 15. Cf. Also the reason Sellars gives for choosing A104 and A191/B236 for his discussion of Kant’s transcendental idealism as a way out between the problematic idealism of Descartes and the dogmatic idealism of Berkely in KTI §6: ›I shall not … discuss the second edition refutation [of idealism] as such, since I hope to convince you that its familiar claims are indeed contained in the first edition, indeed in the first edition refutation itself. For in spite of its ‘subjectivist’ flavor, the latter contains a reference to and, indeed, a summary of the answer to the question “What is an object of representations?” raised first in the introductory passages of the first edition Transcendental Deduction [A104-107], answered in highly schematic form, raised again is the Second Analogy [A190-191], and this time given a fleshed-out answer which is repeated in the text of the second edition. And it is Kant’s answer to this question which is central to the contrast he draws between his idealism and the idealisms he calls ‘dogmatic’ and ‘problematic’‹. 16. It will resurface and be explicated in Chapters 3 and 5, where we turn to reconstructing Sellars’ articulation of the concept of ‘objects’ within the framework of ‘transcendental phenomenalism’. 17. This analysis also contains an implicit reference to the activity of the transcendental apperception, whose task it is to produce unities (cf. KTI §39 and fn 18). 18. In SM II and KPT XVII Sellars comments at length on Kant’s alternatingly talking about ‘manifolds’ (pluralities) and ‘the in itself’ (a singular) highlighting their grammatical status, for according to Kant, as Sellars reads him, we are in no position to make claims about the numerical nature of the logical subject we intend to speak about when we theorise about the causal antecedent to our sense impressions. According to Kant, we just do not get to say whether the in itself is one or many. 19. Here, one could discuss the question whether Kant’s doctrine of synthesis stands on its own, or whether it depends on an interpretation of his unknowability thesis about the in itself (cf. Chapters 3.1, 3.8–3.10). Instead of doing that, however, I want to highlight that at the current stage of our explication of the concept of an object of experience, we are in no position to simply appeal to a form of metaphysical realism. All we can say is that there is something out there that gets taken up into our acts of processing it (becoming sensations and later material for cognitions). As Sellars will endorse the Kantian conception of an ‘appearance’ in the form of his ‘sense impressions’ and talk of ‘sensory states’ nothing hinges on this point. 20. Cf. Chapter 5 on Sellars’ transcendental phenomenalism and Chapters 3.1–3.10 passim, and Chapter 5.2.2 especially, for a reconstruction of the reasons Sellars has for rejecting Kant’s thesis about the unknowability of the in itself.

Transcendental Psychology 67 21. Cf. TTC. The categories are such ‘concepts’ which serve as the most basic forms in light of which we classify acts of judgements in which we cognise contents. 22. In giving more detail to the kind of rules Kant has in mind, he adds: ›However, it can be a rule of intuitions only if it represents the necessary reproduction of the manifold of given intuitions, hence the synthetic unity in the consciousness of them. Thus in the case of the perception of something outside of us the concept of body makes necessary the representation of extension, and with it that of impenetrability, of shape, etc.‹ (A106) The notion of materiality introduced with these aspects gets taken up in Sellars’ argument (Chapter 2.2.4). There Sellars argues for an interdependence of the materiality of the object intuited and the materiality of the act of intuiting. 23. Sellars’ argument for the claim that Kant cannot have rules of action in mind here runs as follows: To suppose that rules of action were in play here implies the assumption that one could use such a rule to tell oneself: ‘Now I look at the bottom line of the house and move upward along its left side, take the roof into consideration and come down on the right side’. With regard to the rule regress problems familiar from Wittgenstein’s Philosophical Investigations it is clear that an application of such a rule would invite the question whether there needs to be a further rule on the account of which one is capable of following that first rule. That Sellars is well aware of the dangers of such an approach is apparent from his reluctant use of the term ‘rule’ and the specific usage he ascribes Kant as making of it. It is also apparent from his distinctions of rules into rules of action and rules of criticism. See SM III, SRLG and MFC, cf. Hübner (2015, 141–143). 24. In other words, a direct realism about appearances would have to hold that our perceptual states directly represent things as they are, this option however is precluded by the Kantian conception of unknowability. However, the passage could also be given the following reading: If our epistemic position were construed such that our appearances presented us directly with how things are, then there would be no intermediate stage of successive synthesis. There would be no difference between a manifold and an object because the content of the appearance would have to be interpreted as directly giving us the object. Against this, Kant holds that then we could not assess how the manifold gets combined. There would not be any synthesis. Our apprehension on such terms would not be apprehension, it would be direct confrontation. Against this Kant has to argue that it is implausible. According to his analysis of the structure of our perceptual states we are confronted with our acts of representing. And as such acts they are different from their content. And our perspectival relation to the antecedent of such contents implies that we combine various different perspectival uptakes into one. Thus, without an account of synthesis, no combining of the manifold in the concept of an object. 25. That Kant aims to answer this question in the affirmative is apparent from the formulation he uses to reject the realist approach. The transcendental realist takes appearances as things in themselves. But if we were to perceive those we would no longer be in any position to ›assess from [our internal perspective on] the succession of representations how the manifold [… can be thought to be objectively] combined in the [concept of an] object‹ (A190/B235). 26. Cf. KTE §36 remarks on the illegitimate moves of such a position and how Sellars understands Kant as arguing against them.

68  Transcendental Psychology 27. This is the place where Kant affirms that we do the supposing here. We consider something as something else. One could put this by saying that this is the place where transcendental idealism does its magic. What appears as empirically real is in fact transcendentally ideal insofar as it is nothing over and above our taking it to be the sum of our representations. 28. How Kant can meaningfully entitle himself to the notion of an empirical concept at this point is a different inquiry, but see Schlösser (2013). 29. Kant here writes: ›since the agreement of cognition with the object is truth, [here] only the formal conditions of empirical truth can be inquired after‹ (A191 / B236).   About which Sellars writes: ›And since, for Kant, the concept of matter-­ of-factual truth concerns the agreement of what we represent with what is, in the critical sense, actual, rather than, as traditionally with what exists per se, he can pay his respects to what he calls “the nominal definition of truth” while giving it a radically new interpretation‹ (KTI §53).   The ‘element of respect’ Sellars hints at shows in the fact that Kant preserves the formulae of making the two things stand in a relation of agreement. The ‘radically new interpretation’ consists in claiming that truth can also be asserted about the agreement between two contents that are represented. Kant thereby achieves his middle way of transcendental idealism, since both are in a critical sense ‘actual’, but neither ‘in itself’ nor ‘merely constructed’. 30. Cf. for a discussion of this requirement see Stroud (2002, 2013). 31. In what sense this limits Kantian Idealism to a merely formal definition of the conditions of empirical truth was touched on the discussion of the significance of KTI §53. 32. Landy (2015), and more recently (2018) discusses this idea fruitfully by way of bringing Hume, Kant, and Sellars in conversation.

3

Perceptual Experience

The aim of this chapter is twofold. On the basis of the previous two chapters, the first aim is to reconstruct an encompassing account of Sellars’ understanding of the structure and elements of perceptual experience that he finds in Kant. The second aim is to investigate the reasons that lead Sellars to go beyond the conception of intentionality inherent in the Kantian account of experience. With regard to the results from the previous chapter, the first aim can be introduced in light of the following question: How exactly are we to assess Sellars’ conclusion that the deeper thrust of Kant’s idealism is to be seen as a thesis about both the self and its world? For which domain does this conclusion hold? According to Sellars’ analysis, this is an argument about the structure of our entitlement to knowledge of empirical affairs. With regard to common sense objects, we are empirical realists in as much as we can meaningfully treat the apples and houses, pyramids, and sunsets to be really out there. But with regard to the transcendental analysis of our epistemic situation, as embodied perceivers we ourselves and a fortiori, these objects are both transcendentally ideal. And there is an additional thought in the background to this conclusion. In the form of a question, it can be stated as follows: Aren’t the considerations about the transcendental object at the other side of our intentional directness to something outside and ‘without’ of us formal methodological considerations antecedent to metaphysical conclusions about the nature of the objects and ourselves? The answer to this question will be affirmative, and in an important sense so. This will propel our investigation of what Sellars is getting out of Kant. For a while Sellars (in his engagement with Kant’s idealism in KTI) may appear to be endorsing Kant’s considerations, he does so only partially, and only to have a base on which to build his own further elaborations about the knowability of objects of empirical knowledge. We can prepare the ground for a detailed discussion of Sellars’ own advancement of Kantian ideas and theses by refining our reconstruction of his take on Kant’s theory of perceptual experience. So far, we have mainly dealt with the details and overtones of the concept of an object DOI: 10.4324/9781003221364-4

70  Perceptual Experience of experience. And we have looked at the sense in which at the heart of Kant’s conception of intentionality lies the concept of an object überhaupt. The wider context in which the analysis of the concept of an object and the concept of successiveness belongs is that of perceptual experience. A justificatory word about the perspective from within which we will address it in the following. We have been investigating the details of Kant’s transcendental psychology as a theory of the fundamentals and of the framework for empirical knowledge of a world of which we are a part. This analysis moved from the distinction of act and content, the reflection on the classifications we undertake with regard to such contents to an investigation of the articulation of the concept of an object we successively can come to experience as independent of ourselves. The systematic question we now face, when we move from the discussion of successiveness to the articulation of perceptual experience, is a question about how to hinge the elements of experience to the admittedly abstract notion of an object as we construed it (Chapter 2.2.4). The reason for proceeding in this way is still the methodological one that the transcendental perspective requires us to give an account such as to end up with a concept of an objective world, not to simply presuppose it. And as that gives us the direction for our investigation, we can render the question we need to address next more precisely by seeing it as question of how to connect two theoretical domains. If our analysis so far concerned (1) the requirements on thinking objects of experience as the contents of classifiable acts of representation and as existing outside such acts, the next step in our analysis concerns (2) the requirements on perceiving and ultimately knowing these objects to be of a specific nature. To get from the first to the second, we need to ask: What is the relation between the concept of a transcendental object and the concept of the in itself? Both concepts play an important role in Kant’s overall framework and it is the task of the transcendental philosopher to bring out the reasons behind the architecture of the resulting theory-building. Our reflection on the concept of a transcendental object provides the context for understanding Sellars’ interpretation of the role of further concepts which come to play a heavy weight epistemological and metaphysics-critical role in Kant’s resulting conception of perceptual experience. Those comprise the concept of the in itself, the concept of a bodily and mental state, a sensation, an impression and ultimately also that of a conscious intuition of an object. We begin by looking at the way in which Sellars reads Kant as introducing the notion of the in itself into his theory. From this, it will become clear that the advancement of the theory operates by way of a transcendental postulate, i.e. a methodological feature which the concept of the in itself shares with the concept of a transcendental object. A fruitful way to approach this

Perceptual Experience 71 topic is to ask what is involved in conceiving of the intentional relation to a transcendental object? How can it serve as the root notion for further explanations of epistemically efficacious concepts? Such are concepts we need to fill in the thought of something as opposite to us, with the power to impinge on our senses, i.e. to which we are perceptually receptive.

3.1  The in Itself and One World What is the systematic idea behind introducing the concept of an in itself? And how can it be traced back to the concept of the transcendental object? We approached the concept of such an object by means of an analysis of the presuppositions that underlie all aspirations to successful world-directed reference. At this point, we can characterise this abstractly identified reference point with regard to the faculty we employ when directing our intentionality towards it. But if a transcendental object is an object for intellectual intuition, it is an object only in the sense that we take it to be the content of a pure intuiting, which, more or less, is an activity powered by the mind, by our spontaneity. In other words, being concerned purely reflectively, transcendentally identified object is, in the first instance, only an intellectual affair. Thus, receptivity has no say in it. So when we develop the line of thought that leads from the transcendental object to the in itself, we do so by introducing the notion of receptivity. And the fact that we are passive and receptive is a fundamental ingredient to an encompassing account of intentionality. So, how are we to understand the role of this concept of passivity in our being receptive? On the one hand, we have aspects that concern the architecture of our mental faculties and the interplay among them. On the other hand, we need to enrich the account of intentional directedness with a more structured and informative characterisation of the items that correspond to our receptivity. In as much as we are bound by the transcendental method of analysing intentionality from within, we are barred from jumping to the notion of manifest objects. But what we can in fact do, by way of theory construction, is tracing the argumentative path designed to help us grasp the very core notion of something independent from us that works on our receptivity. And here we are doing two things at once, when we engage in such tracing: (1) We are going backward from something that shows up due to our being sensible to something that we hold responsible for it. And (2) we are conceptualising the direction of impact forward from something that has the power to affect us to it appearing to us. Why is it important to be aware of the two directions we thereby cover? Because transcendentally speaking, what is given to us as the content of an appearance is the first point of entry into this account about affection; it is the

72  Perceptual Experience moment from which we work our way backwards. Meaning, from a meta-perspective we can say by way of a commentary that introducing the notion of something causally impinging on us is predicated after reflection on one specific moment in an entire sequence of steps. What is prior in the order of our explaining here, the appearance, comes later in the order of being and affecting; it comes because it is conceived to be brought about by something that is different from it and is construed as having the powers to effect such appearing in us. Does this line of thinking get us to the in itself? One element in his [Kant’s] strategy hinges on the fact that the epistemological tools he brings to his task include [1] not only his version of the Cartesian distinction between formal and objective reality but [2] the idea that the human (or any finite) mind is passive with respect to the impressions which initiate its conceptual representation of a world of which it is a part. (SM II §30) In this passage, Sellars reports the by now familiar observation that the in itself has as its root notion the Cartesian distinction between existing inside and outside of a thought. This amounts to the claim that we are to understand the genesis of the concept of the in itself as the result of an inference that carries us from something inside of a representing to something outside of and antecedent to it. How can this inference be made explicit? Let us look at Sellars discussing this question: Kant thinks it to be a necessary truth that if there are appearances there are things in themselves. In one sense, although a limited one, he is obviously right, and the point is more interesting than is the superficial one that “appearance” implies “something appears,” though Kant does sum up his view in this manner. The point is rather that it is an analytic truth that if there are representeds [appearances] there must be representings [acts of representing an appearance]. (SM II §27) What does this imply for our knowledge of the in itself? And how does Sellars interpret Kant’s reasoning here? According to Sellars: It must be granted that this argument establishes at most that if there are representeds there must be representings which exists simpliciter or an sich. It does not establish that there are non-­representings an sich […]. Nor does the argument establish that what exists simpliciter is in any sense unknowable, nor that represented individuals can only

Perceptual Experience 73 be “appearances of,” rather than identical with, items which exist in themselves. But Kant, of course, thinks he can establish these additional points. (SM II §29) In Kantian terminology, the only thing we are certain to be in touch with are appearances, showing up as the contents of our acts of representing them. For Sellars, the point Kant takes himself to have established, i.e. that there is an in itself, actually only yields that there are acts of representings, i.e. our intuitional acts. This leaves open the possibility that what gives rise to the contents we come to represent has an existence outside of our representings, although Sellars grants that we are bound to represent such contents in ways restricted by our receptivity. If this Kantian argument does not establish anything certain about the in itself, why is it of interest to us? Because it demonstrates the transcendental character of Sellars’ Kant exegesis and ties it to the distinctions of act and content, which we saw to be so basic for an articulation of the resulting notion of intentionality. In other words, the passages lend support to the claim about the importance of the Cartesian categories for thinking contents formaliter and objective (i.e. as thought of). So if Sellars brings up this argument for introducing the concept of the in itself but dilutes its conclusion so drastically to establish only that there are acts of representings, aren’t we back at the point where we started, stranded with appearances, i.e. acts of representing something as appearing to us? In a sense, we are, but now we are in the position to see that Sellars reads his rejection of Kant’s reasoning to have made room for his disagreement with Kant’s unknowability thesis regarding the in itself.1 What role does the in itself play in an account of our perceptual experience? If the argument for introducing the in itself does not proceed from appearances and treats them as appearances of an in it itself, what is it? In the context of discussing what it takes to move from subjective to objective successiveness we already learned that we require the very idea of there being a cause to the structure of our acts of representing. In other words, the argument is not just that ›if there are appearances [then] there are things in themselves‹. The argument rather has to be that our acts of representing something do not just happen in the void. That is, if they are to happen in nature, be about nature and ultimately lead to knowledge about a world of which we, with our perceptual facilities as embodied perceivers, are a part, then the relation between the in itself and the appearances has transcendentally to be conceived of as an affective relation, involving our somatically realised sensible receptivity – which is what we are going to focus on in the following, while acknowledging that we also act spontaneously.

74  Perceptual Experience The point is that although all we are in fact in touch with are representings, these representings need to be seen as the result of an external impact on our senses. If knowledge involves the cooperation of spontaneity and receptivity, here, our sensibility has to be integrated into the architecture of our theory as yielding the material on which our understanding can then be said to operate. And this is the path we embarked on when above we asked the question of how the transcendental object relates to the in itself. We partially answered that question by saying that receptivity and passivity had to be integrated into an encompassing account but we lacked the details. Now we can address these details by asking further: What does it mean to hold that we are passive with regard to impingements from outside, affecting our senses? It means, in the first place, that we have to work our way backwards from full blown representations, i.e. intuitions, which already involve the understanding to more brute, and prior stages where the impingements hit right on our senses. What can be said about this moment in the genesis of perceptual material? If we go backwards in this sense, we have to move from intuitions, which are representations with an ‘x as y’ structure backward to – what is conceptualised as – their cause. Doing that amounts to reflectively identifying the causal origin of our appearances. Remember that ‘appearance’ is Kant’s term for characterising the epistemic status of conscious acts of representing, such as intuitions. So methodologically what Sellars argues is the following: We move from intuitions (appearances) to the in itself when we consider that the thought of being affected by something outside of us can be carried out in the phenomenal realm and in the noumenal realm. In preparing the ground for this distinction, Sellars writes: I have been arguing that of the items Kant calls “intuitions,” those which are representations of a manifold as a manifold constitute a special class of [conscious and conceptually structured] representations of the understanding. They belong, as such, to spontaneity. (SM I §21) According to Sellars, we start with full-blown intuitions and from them move to something which must be prior to them. Something which – in a sense to be explained – must be simpler than them. The class of causally efficacious items fulfilling these requirements is a manifold of representations. Here we might ask, in what sense are they ‘representations’? The short answer is: they are, on the epistemic subject’s side of this equation, the result of impingements from outside of the subject. They are the very first point of contact we can theoretically single out. [What feeds into intuitions, their] “receptivity” is a matter of the understanding having to cope with a manifold of representations

Perceptual Experience 75 characterized by “receptivity” in a more radical sense, as providing the “brute fact” or constraining element of perceptual experience. (SM I §21) So the class of representations Kant calls intuitions presupposes2 a manifold of representations impressing themselves on us (cf. Chapters 2.1.1, 2.2.2). This notion imports the idea of grounding and restricting the material of our acts of intuiting in an important, guiding sense.

3.2  Sheer Manifold of Sense How are we to understand this manifold of representations? In the following passage, Sellars most explicitly reflects on the methodological operation that underlies its introduction into the resulting theory: The latter manifold has the interesting feature that its existence is postulated on general epistemological or, as Kant would say, transcendental grounds, after reflection on the concept of human knowledge as based on, though not constituted by, the impact of independent reality. It is postulated rather than “found” by careful and discriminating attention. The concept of such a manifold is, in contemporary terms, a theoretical construct. Let us, following Kant (and a long tradition which goes back at least to Aristotle), call the items which make up such a manifold “sense impressions,” and let us explore independently the idea that it is reasonable to postulate the existence of sense impressions. (SM I §22) Now, if the manifold is a theoretical construct we come to postulate as brought about by ‘sense impressions’, we can ask further if we can know anything about the nature of that which causes such impressions on our senses.3 Sellars’ answer is that we can, in fact, know something analogically. The methodological operation of postulating an origin to our affective states reiterates at this point, such that we can say ‘we can know about an in itself’ by way of analogy.4 We transpose (1) the causal relation of affections we see among physical objects and bodily states to be similar to (2) the causal relation of affection we postulate to hold between things in themselves and outer sense. About the latter notion of affection, Sellars writes: [The idea of such noumenal] causation is the impact of things in themselves on outer sense which Kant calls ‘affizieren’. (I §43 fn17)

76  Perceptual Experience This leaves us with the question of how to square the two forms of causation mentioned. The analogical transposition behind the construction of the two senses of ‘cause’ can be laid out as follows. According to our reading of how Sellars interprets Kant, it is the method of analogically transposing the notion of cause from the phenomenal to the noumenal that is central here. Evidence for this reading can be found in the following passage in which Sellars walks us backwards from (2) phenomenally perceivable cases of being affected to (1) transcendentally postulated cases we conceive of by way of analogy: The “double affection” theory [says] [2] that sensory states of the empirical self as part of Nature are brought about by physical stimulation of the sense organs (mediated by neuro-physiological states), but [1] that this “bringing about” and the factors it involves are themselves appearances represented by minds under the impact of things in themselves. (SM II §41 fn20) Again, we proceed backward in the order of explanation (to account for a forward direction of fit (and affection), we thereby try to reconstruct as happening in the order of being (from 1 to 2). If we try to invert this way of displaying the elements and bring them into one line of thought, we come to read it in the following way. Sellars writes: The doctrine of “double affection” is an essential feature of Kant’s thought. Correctly understood, it simply tells us that [1] the transcendentally conceived non-spatial, non-temporal action of the nonego on human receptivity, generating the manifold of sense (which action is required to explain how the esse of the experienced world can be concipi and yet non-arbitrary and intersubjective) has as its counterpart in the represented world [2] the action of material things on our sense organs and, through them, on the sensory faculties of the empirical self. (SM II §57) How does this amount to a positive answer5 about the in itself? We have learned in what way the manifold of sense is a transcendental postulate, which in turn is brought about by something that impinges on our senses, i.e. the in itself.6 Both these elements of our theory of perceptual experience are theoretical constructs. In other words, of the in itself we only get a relational characterisation from within our claims of knowledge. The following addition traces the introduction of the in itself back to the Cartesian framework of existence inside and outside of acts of representing. We get one line of thought when we consider noumenal affection to cause phenomenal affection. That is, as Sellars reads Kant,

Perceptual Experience 77 we can bring the two affections into a coherent sequence when we pay close attention to the strictures Kant places on the elements of his theory of experience: Thus, of the two “affections” one has representative being only, while the other has both representative being “in” transcendental thought and also being simpliciter or “in itself.” This can be put by saying that one “affection” is the appearance of the other, but unless this is supplemented by a properly Kantian analysis of appearance it acquires the air of intolerable paradox (a “twoworld” theory) […]. (SM II §57) From this, it is clear that Sellars rejects a two-world reading. Let us try to make explicit the reasons he has for this. The argument presumably involves the following steps: (1) On the basis of analogically transposing phenomenally available affections to noumenally postulated affection, we can invert the directing of thought and argue that (2) things in themselves bring about in us states to which we respond with the unconscious formation of sense impressions.7 (C) Sense impressions constitute a manifold which we can take up in acts of intuiting as a manifold, i.e. as an intuitional content with determinate structure.8 In this sense of reading the account, it is one world we come to be conscious of. To Sellars, then, there are no two worlds.9 Rather, there are several stages to our perceptual experience of one world. The first stage is such that it can only transcendentally be identified ›after reflection on the concept of human knowledge as based on, though not constituted by, the impact of independent reality‹.

3.3  Sensations, Sense Impressions, and Intuitions The gist of adding the qualification that human knowledge is ›not simply‹constituted by such an impact lies in the further constraint that not all acts of representing something are conscious. That is, the distinction between conceptual and non-conceptual states of consciousness calls for an explanation of the here relevant transformations. From the simple to the complex, from the unconscious to the conscious, from bodily states to mentally represented contents. The point of entry here is the claim: ›Sense impressions are non-conceptual states of consciousness‹ (SM I §25). And a further thesis central to Sellars’ understanding of

78  Perceptual Experience sense impressions is that they are extended and in space. He holds this against Kant: Kant’s treatment of sensation is notoriously inadequate and inept. From the premise that sense impressions as mental states are neither literally extended nor in physical space, he infers that they are in no sense spatial, i.e., that they in no way have a structure which conforms to a geometrical axiomatics. The idea that sensations are “purely intensive magnitudes” has always made it difficult to understand how sense impressions could have a meaningful connection with physical states of affairs. (KTE §3 fn1) But if sense impressions are something with regard to which we are passive, we have to ask what their relation is to intuitions.10 For Intuitions, according to Sellars, invariably involve an active element, which is why he characterises them as ›perceptual takings‹ (SM III §33). This very fundamental principle is also reflected in Sellars’ rejection of the Myth of the Given, in the sense that what is ‘given’ actually is always something that is taken to be in a certain way and of a certain kind. Since we have dealt with the formal or categorial structure of intuitions in the previous chapter, it is now time to expand our investigation to cover the synthetic processes that operates on the material side of sensible acts of experiencing.

3.4  From Image-Models to Intuitions Sellars’ view is by and large Kantian in that he defends a conception of the interplay of sensibility and spontaneity as central to the intentionality of intuitions. (IKTE §27) What are intuitions and how to analyse them? The characterisation by which Sellars advances his answer to this basic question is functional: The role of an intuition is a basic and important one. It is to bring particular objects before the mind for its consideration. (IKTE §48) Intuitions as a sub-class of representations do not concern general but particular objects. It is in intuitive representings that objects can come to show up, through the involvement of another faculty, the understanding, here picked out as ‘the mind’, i.e. consciousness elaborating on them. In order to reconstruct and systematically discuss the qualifications Sellars adds to this basic, functional view of intuitions, we will now turn to his papers ›Kant’s Theory of Experience‹, and ›The Role of Imagination in Kant’s Theory of Experience‹.

Perceptual Experience 79 Before we begin, a remark is in order about what could appear as a methodological break. Up to this point, our discussion centred on methodological reflections that were by and large transcendental in that they concerned the necessary features any system of knowledge has to have to yield knowledge of the world of which it is a part. In the paper on ›The Role of the Imagination‹, it appears, Sellars jumps back to square one, investigating what’ is involved in basic sense perception. He describes the method underlying his investigation here as a form of phenomenological reflection (IKTE §2–4). And there certainly is room for discussion about the relation between results established by way of transcendental reflection and results established by phenomenological reflections. One could, for instance, argue that conclusions of transcendental arguments always deserve priority over conclusions arrived at by way of phenomenological investigations. While one can agree that this is indeed a contentious question, one can also hold that there’s a smooth path through the obstacles that operates with the figure of specification. And this is the sense in which we will read Sellars’ aims in these papers. The transcendental requirements he established in SM and KTE are being filled, in this later paper, IKTE, with a particular kind of detail.11 This time, it is not coming downstream from the higher points of abstraction about the general and necessary constraints on any theory of epistemology, but coming upstream from what is essential to the phenomenal character of our everyday experience. The considerations Sellars develops in IKTE then fit the picture insofar as he adds detail which saves the appearances we live by.12 As Sellars states right at the beginning of IKTE his aim […] is to give a sympathetic account of Kant’s theory of the role played by what he calls the productive imagination in perceptual experience. (IKTE §1) Where and in what way does the productive imagination exactly come into play? The scenario Sellars construes is as simple as can be. He suggests us to consider the act of perceiving an apple. Red on the outside, three-­ dimensional, in such and such a relation to the perceiver’s body, and, as is the nature of apples, opaque. Now the productive imagination comes into play when we address the question of how it is that we conceive of the apple as containing a white inside, a certain taste, weight and smell, a backside: How can a volume of white apple flesh be present as actuality in the visual experience if it is not seen? The answer should be obvious. It is present by virtue of being imagined. (IKTE §16)

80  Perceptual Experience The actuality of the white, as Sellars goes on to elaborate, must for its existence depend on the perceiver, and as such, it must be, so he concludes, ›in‹ the perceiver. The point about the white, it should be clear, also goes for the backside of objects, as well as for all those qualities we do not directly see but see of [that], what we see (IKTE §15). They all are present by virtue of being imagined, and as such ›features … present in the object of perception as actualities‹ (IKTE §21). The same goes for causal properties, which we do not see, although we see objects as having them.13 (IKTE §22) From these ingredients, obtained through phenomenological reflection on acts of perceiving, Sellars draws the following conclusion that ties the qualifications together: Thus, imagining a cool juicy red apple (as a cool juicy red apple) is a matter of (a) imaging a unified structure containing as aspects images of a volume of white, surrounded by red, and of mutually pervading volumes of juiciness and coolth, (b) conceptualizing this unified image-structure as a cool juicy red apple. Notice that the proper and common sensible features enter in both by virtue of being actual features of the image and by virtue of being items thought of or conceptualized. The applehood enters in only by virtue of being thought of (intentional in-existence). (IKTE §23) The productive imagination, on this account, generates the content of our perceptual acts by combining the activities of (a) unifying actual phenomenal features present in one’s experience with the ‘image’ of inner structure (i.e. sensing-cum-imaging) and of (b) conceptualising it under one heading, i.e. in the concept. The act of imaging thereby contributes the unified image encompassing the hidden sensible features we would encounter when turning the apple around or opening it. The act of conceptualising, Sellars tells us, brings in categorial features which belong to the resulting image by virtue of being thought of. This indicates that the relevant categorial features do not easily come in a sensible form but require the involvement of conceptual representations.14 What does that leave us with? Sellars concludes from this: The upshot … is that perceptual consciousness involves the construction of sense-image models of external objects. (IKTE §25) So Sellars wants to be able to say that the productive imagination construes image-models as intermediate theoretical entities we need to introduce in

Perceptual Experience 81 accordance with the phenomenological reflection carried out here. In order to entitle himself to this move, he adds: Notice once again that although the objects of which we are directly aware in perceptual consciousness are image-models, we are not aware of them as image-models. (IKTE §27) This should make clear the theoretical status of image-models as forming an additional, though unconscious, intermediate level in the process by which we ultimately want to account for the nature of our intentional relation to objects of experience. What makes image-models so interesting is that they occupy a very important position on the battlefield of Kant’s theory of experience. They not only unify visible-sensible and invisible-categorial aspects of our perceptual consciousness into a coherent one but, in this way, also involve our sensible and conceptual faculties. This becomes clear when we consider the case of recognising ›this particular apple as the one we bought yesterday‹. It is by virtue of these cases that Sellars characterises image-models as complexes in which understanding and sensibility blend: The most significant fact is that the construction is a unified process guided by a combination of sensory input on the one hand and background beliefs, memories, and expectations on the other. The complex of abilities included in this process is what Kant calls the “productive” as contrasted with the “reproductive” imagination. The former, as we shall see, by virtue of its kinship with both sensibility and understanding unifies into one experiencing the distinctive contributions of these two faculties. (IKTE §25–26) At this point, Sellars has not revealed yet, how exactly he interprets the two faculties to come together. The first step is to reconstruct the reasons for what Sellars calls the ideality of the image-model world. In a second step, we address the further question of how sensibility and understanding are thought to merge in the conception of an image model.15 If our perceptual consciousness is based on the production of image-­ models, this implies that we do not experience objects as they are right in front of us but that we first form sense-image models of external objects. How does Sellars support this claim? Sellars’ main argument for this indirect conception of perceptual experience is meant to be backed up by the phenomenal fact that all object-perception happens from a point of view. The perspectival character of the image model is one of its most pervasive and distinctive features. It constitutes a compelling reason for

82  Perceptual Experience the thesis of the transcendental ideality of the image-model world. Image-models are “phenomenal objects.” Their esse is to be representatives or proxies. Their being is that of being complex patterns of sensory states constructed by the productive imagination. (IKTE §28) Clearly, we can make the distinction between our perspectival take on objects and their really being non-perspectival three-dimensional items. But Sellars, at this stage of walking through the process of the act of perceiving, is talking about the nature of image-models as image-models which are perspectival in relation to me, the perceiver too, who also has to figure somewhere in them. And, being essentially mental items that appear to us, while we construct them, they are representings through which we form our grasp of perceivable objects. Note that since this is all that is available to us at this point in theory-construction, there’s also no robust outside against which we could contrast and assess the correctness of our image-models. Say you attend a guided tour through Rome and come to learn of its architectural intricacies. The detail you learn of blends with the sensory input, your background beliefs and eventually becomes memorised as aspects out there. Historical details about a house, for instance, since they are aspects one only represents conceptually, have their being, as aspects, only ›in‹ you, the representer. When one engages in envisaging anew the sensuous aspects, perhaps a particular colour and surface texture, one engages the reproductive imagination, which would also allow me to associate my own future in such a house. From the standpoint of the phenomenologist, our contact with what we come to perceive as objects thus is always already governed by the operations of the productive imagination, in the sense that we never directly encounter objects as such, because what shows itself to us as an object comes to consciousness as the perspectival content of an image model. If that was all there was to say about our basic epistemological situation, a number of strong idealist consequences would ensue. For instance, we would all look onto the world through our own singular and subjective windows, with no option available to ascertain that we ever shared a common viewpoint. This is one interpretation of Sellars’ thesis of ›the transcendental ideality of the image-model world‹ we have arrived at. Yet the way he puts it makes it look as though there’s still a difference to be made out between the image-model and the real world. 3.4.1  Empirical Schemata Sellars interprets Kant’s notion of empirical16 schemata as essential to accounting for the concept of a world that is at once the origin and result

Perceptual Experience 83 of our perceptual experiences. In a passage that he signals as comprising all the elements into a first complete account of perceptual experience he writes: [1] In the first place, the productive imagination is a unique blend of a capacity to form images in accordance with a recipe, and a capacity to conceive of objects in a way which supplies the relevant recipes. [2] Kant distinguished between the concept of a dog and the schema of a dog. The former together with the concept of a perceiver capable of changing his relation to his environment implies a family of recipes for constructing image models of perceiver-­­confronting-dog. (IKTE §31, numbers added) From the quote we learn that he explains image-construction as a twofold operation. If we invert the order, and recount its components, we get the following sequence: ( a) A conceiving of objects that supplies recipes (b) A constructing of images following these recipes This inverted order of Sellars’ presentation marks a difficulty. It lies in the fact that the notion of an object is the explanandum of this account and thus cannot be invoked in the explanans. But we can read Sellars as trying to address this issue head on. In what might prima facie look like an evolutionary, ontogenetic remark, Sellars states that even a ›child [which] does not yet have the conceptual framework of dogs, houses, books, etc., [nonetheless] does, according to Kant, have an innate conceptual framework.‹ (IKTE §32). And what in Kant might be taken to be a commitment to innatism, in Sellars as we saw in the previous two chapters, is the result of our analytic reflections on presuppositions, on the constraints on theory-building. Any account of world-directed knowledge has to endorse, Sellars adds: a proto-theory, so to speak, of spatio-temporal physical objects capable of interacting with each other; objects — this is the crux of the matter — which are [causally powerful and hence] capable of generating visual inputs which vary in systematic ways with their relation to the body of a perceiver. (IKTE §32) This is the relevant notion of a capacity appealed to in (a) above. A capacity to conceive of objects, however thin; and then most certainly not under the heading ‘object’, but as something out there that is more than just inside my thoughts. In other words, we seem to begin with a thin notion of object as the centre of a proto-theory and only later

84  Perceptual Experience come to develop a thick notion of objects as they figure in full blown conceptual-­activity. How then does the ‘systematic varying in the generation of visual input’ add up to a ‘supplying of recipes’? ‘Recipes’ here seems to be Sellars’ term for rules. And rules are what the productive imagination accords with in constructing its image-models. At this point it is helpful to draw a distinction between, on the one hand, the attempt to write the history of concept formation in the sense of accounting for there being empirical concepts at all. Something, perhaps done in the anthropology of evolutionary cognition (cf. Tomasello 2014; Garfield 2012). And on the other hand, the attempt to account for how it is that individual thinkers, in full possession of their capacities, a conceptual framework, language, and sociality, have perceptual experiences that involve the activity of the productive imagination. It is in the latter sense that we can read Sellars here. This involves interpreting him as suggesting that repeated exposure to the ‘systematic varying in the generated input’ supplies one with recipes not all by itself – no man is an island – but through one’s participating in a practice of language, of using words for the things around us. But are we given recipes or rules for synthesising in educative scenarios that teach us to tell things apart or see them as ones? It seems that we derive recipes from empirical concepts.17 That would mean that in a sense, empirical concepts just are the very recipes we appeal to when demarcating the boundaries of objects – a fact artists may play with when blurring the boundaries of what we see and what we see something as.18 In this way, we can account for the apparent jump from talk about image-model construction guiding recipes to a more full blown exemplary scenario that ties the elements from steps (1) and (2) together anew. This finally puts us in a position to reconstruct Sellars’ understanding of what empirical schemata are. For this, we’re told to consider what happens as a perceiver walks around a pyramid, taking it in: The [empirical] concept of a red pyramid standing in various relations to a perceiver entails a family of concepts pertaining to sequences of perspectival image-models of oneself-confronting-a-pyramid. This family can be called the schema of the concept of a pyramid. (IKTE §33) The bigger unit now under close scrutiny in the example is the array of points of view that all concern the same immobile content but each from a slightly different visual angle. We can so much as say that pyramid-contents from viewpoints 1, 2, 3 … n would not amount to one coherent whole were it not for the uniting function of the schema that ties the various view points on pyramids, the family, as Sellars names it, into one. Empirical schemata thus can be characterised in regard to their function of giving us the method of constructing image-models in unison with empirical concepts.

Perceptual Experience 85 But how does the notion of an empirical schema help to advance the prospects of arriving at a solid conception of an object? For that we best consider the distinctions developed in its wake. The pyramid schema arises as what in our perspectival perception connects our first-personal point of view with a series of contents (manifolds) as the same content. And while the sequences are individual, perspectival image-models, each a slice of my view on the pyramid, the resulting concept of that very pyramid at their core is a concept of a non-­perspectival object. Reflecting on the benefits of these working distinctions Sellars writes: It is in terms of these considerations that Kant’s distinction [in the schematism A141] between (a) the concept of an object, (b) the schema of the concept, and (c) an image of the object, as well as his explication [in the Second Analogy A191] of the distinction between a house as object and the successive manifold in the apprehension of a house is to be understood. (IKTE §35) If we go through the progression of the argument, we see the distinctions at work. We started with a thin notion of object, as we learned, that matters in the common sense scenario of looking at an apple, wondering about its flip-side. This gave us the import of the image of the object, as well as its occurrence in sequences of individual perspectives. Then we were told that the notion of an image as it appears in the schematism really is the place where the productive imagination takes in the sensory manifold and works on it, by way of (rule- or) recipe-guided constructing of image-models in accord with empirical concepts. From the perspectival nature of the image-model, we abstract in an ultimate step to arrive at the concept of an individual, non-perspectival object. 3.4.2 Objects It is in view of this battery of distinctions that Sellars turns to citing the following passage from the Second Analogy, where Kant argues for the claim that the synthesis of apprehension proceeds according to the necessary rules that lay in the notion of its object. Appearances, therefore, have to be rule-governed in order for them to conceivably be about objects. While the appearance as such presents to us its content in a sequence, what it shows to us in a successive manner has to be considered as outside of such successiveness. The object is that in the appearance which contains the condition of this necessary rule of apprehension. (A191)

86  Perceptual Experience One might wonder how an object can be said to contain a rule. But asking such a question is equal to stepping outside the bounds of the level of consideration that Sellars wants us to work with him on. We best envision that level by seeing him as asking the question: What is it at all to think of the content of a representation to be objective, to be more, over and above being merely represented? Sellars is really on a par here with Kant, which we glean best from looking at a telling preceding passage of the same paragraph. Kant addresses the question of how to understand that ›the manifold may be [objectively] combined in the appearance itself (which [since it is an appearance] is yet nothing in itself)‹? To this Kant offers: Here that which lies in the successive apprehension is considered as representation [Vorstellung], but the appearance that is given to me, in spite of the fact that it is nothing more than a sum [der Inbegriff] of these representations, is considered as their object, with which my concept, which I draw from the representations of apprehension, is to agree [zusammenstimmen soll]. (A191) Insofar as we are dealing with the content of successive apprehension, we have to say that an appropriate analysis of our concept of an object is best predicated from a transcendental point of view which takes the metaphysics of mental acts seriously. An object is what lies in the successive apprehension and is by being an appearance ›nothing more than the sum‹, the Inbegriff, of a representation. Sellars agrees with this: ›The principle of such construction [of synthetically apprehending] is not happenstance but conformity to recipe – schemata derived from concepts‹ (IKTE §36). And this is how empirical concepts inform the schemata, as the methods or recipes for constructing image models out of sensible material through which we view objects as that in the appearance which systematically varies in relation to our perspectives on it.

3.5  Objectivity and Image-Models It is exactly this inclination to treat the objects distilled out of perspectival images as distinct from us, that is at the bottom of our conception of objectivity. If Kant’s treatment of the categories is one of a schematisation to the specifics of our spatio-temporal object-world, there is a further question left open after the possibility of the synthesis of the manifold. According to Kant that synthesis is always already guided by the productive imagination, in its transcendental role (A120). Such that the productive imagination informs the process of the construction of the material out of which the sense-image-models are built. And through the processes of synthesis of apprehension and recognition

Perceptual Experience 87 they are turned into intuitions. This means that the categories are part of the ‘thinking’ of intuitional content without having to be literally perceptible as part of the content. That is, the mathematical categories of quantity for instance already constitute the categorial features of image-models without being in the image-models as an image. The reason for this is that we cannot have ‘images’ of the categories. Sellars and Kant here are very clear on that. And it is this move of the restrictive power that the conceptual (qua productive imagination) plays on the sensory which we turn to next.

3.6  Intuitions in the Context of Judgements Up to this point then, our reconstruction has shown that Sellars endorses a Kantian idealism with regard to the resulting conception of manifest objects of experience. It arose from the discussion of image-models whose nature it is to be representatives or proxies. If Sellars were to leave it at that, one could infer that he was by and large a transcendental idealist. Our next aim is to show that this is not the case. Rather, by discussing his understanding of intuitions, the nature of the act of intuiting in relation to the content thus intuited (KTI §50–53), he makes room for a conception of transcendental realism. At the core of this realism, we will come to examine, lay two claims: (1) The mutual interdependence of the actuality of the perceiver and the perceived and (2) the attainability of an account of what objects are in-themselves. We can introduce the elements of this position by first addressing the question of how intuitions figure in Sellars’ account of our perceptual experience of objects considered so far. For this, we return to his phenomenologically derived distinction between what we perceive the object as and what we perceive of the object. The latter, so he tells us, is a matter of the ›conceptual content of a complex demonstrative thought‹ (IKTE §38, cf. SM II, SSIS, SSOP and SSPC). Intuitions put us directly in contact with objects by referring to them demonstratively. The linguistic version of the content of an intuition of a pyramid comes in the phrase ›this yellowish pyramid from a certain point of view‹. But what exactly is the relation between image-models and intuitions? Before trying to answer this question, we need to summarise why Sellars places such an importance on image models. In their additive function, they play into our manifest phenomenology by supplying the inside and flip side of objects we visually confront. They do so by operating on the material that impinges on us. The ought-to-be rules guided operating happens in accord with empirical schemata. And in that sense, we can conclude that the overarching function of image-models is to stand as proxies between the things-in-themselves and our conscious representations of manifest objects. How do they do this? The action of transforming the sensory material into image-model contents is as

88  Perceptual Experience an operation a complex meeting ground for both faculties, the understanding and sensibility. The transformation marks a shift in the epistemological account transcendental philosophy reconstructs here, a shift from the properties of unconscious sensory material to the conscious representation of image-model contents.19 The fact that we do never perceive image-models as image-models underlines their being proxies. The fact that the transformation is carried out by the productive imagination working with empirical schemata means that Sellars builds an active module into a decisive place. Under one description it is the place where the world’s impingements on the subject’s passivity are to result in knowledge. Under a more refined description we can say: If the world was – within this part of the theory – simply to impress its structure on the subjects’ consciousness, this would be a clear case of the idea the Myth of the Given. Since the structure is not given by the world but only – in a placeholder sense – the material we operate on the Myth can be said to be averted. Sellars here underlines Kant’s focus on the forms of intuition as always already shaping what we get to have in front of our eyes. But Sellars also reads Kant as holding that the forms of sensibility already transform the world’s impingements into impingements ‘for us’ so to speak (cf. SM II). This is to demarcate that at the very first stage, in the sequence of stages that make up the epistemological account (leading from the things-inthemselves to perceptual experience of full blown objects) our senses effect an unconscious formation of the material. The formation through our senses means that the impingements that result in unconscious20 sensory states of the body are already heavily distorted (cf. SM I). This allows us to observe in as early as this stage, the earliest possible, how Sellars prepares his farewell to a naïve realism as a thesis to the effect that perception puts us directly, i.e. without mediation, in touch with things as they are in themselves. This seems to furthermore imply that since we never get to have the things independently of our distorting them, there’s no point in supposing that we could judge one sensing to be more accurate than another. Why? Because no independent third to compare them with is ever available. A further point about the transformation that is of importance concerns the move from something external into something internal. If an object in my tactile vicinity makes an impression on my senses, the properties of the objects as I sense them undergo a transformation from (a) being whatever they are to begin with, to (b) being that in virtue of which my bodily state is systematically, though unconsciously altered, to (c) being the conscious content of an image model, e.g. ›This perspective onto a pyramid in front of me from this particular angle‹. Again, with the last step (c) being such that I do not conceive of it as an image model but as a pyramid presenting itself to me. It is only by way of the transcendental reflexion that we distinguish the image-model of the pyramid as one stage

Perceptual Experience 89 in the formation process of visual experience, from its origin in something that brings it about. 21 Whereas within a naïve construal of our basic epistemological scenario we just confront a pyramid, full stop. What is the importance of this transformation matter and what is the reason for underlining the nature of the image-model as a representing in the function of a proxy figuring in between the world and our empirical knowledge of it? For Sellars, they constitute one compelling reason for the ‘transcendental ideality of the image-model world’ as he puts it (IKTE §28). So far we have understood one dimension in that claim that is to do with the shift from unconscious impacts into conscious representeds or contents of representings. About such represented contents, about the inside of image-models, Sellars writes that they do not present anything beyond common and proper sensible and that this be the reason for withholding full categoriality from them. We do not see of an object its being a substance. The mode of presence categorial features (the transcendental content) have enters into the account of perceptual experience at a different stage. Here one might contest that in order to bring something into an image, in order for an image-model content to be something at all, it has to be at least informed by the conceptual capacities that make up the mathematical categories. Which is compatible with the claim that the productive imagination at this point does not use the full-blown conceptual capacities as the understanding, later on, does when furnishing the intuitions of objects as intuitions with the help of the dynamic categories. In other words, there’s a reading available that takes the passage (from IKTE §39) to be compatible with the claim that image-model contents are constructed by the productive imagination. The productive imagination operates with the mathematical categories of Quantity and Quality. Such a reading would not have to argue against Sellars claim that we do not see the categories. Sellars could be read as trying to underline the Kantian idea that we do not get the categories through the forms of sensibility but as arguing that they have to come in through the understanding as its forms. With this, we can map how image-models are formed and how they feed into intuitions: ( 1) (2) (3) (4)

Sensations are taken up into image-model contents. Image-model contents stand under the mathematical categories. Mathematical categories allow for units or quanta with qualities. Qualified quanta serve as transcendentally ideal internal objects to which the demonstrative reference of intuition points.

(C) The content of intuitions results from a two stage process: Internal constitution of the content on the basis of our receptivity (1–3), uptake of material as two complex components of one resulting intuited content (4).

90  Perceptual Experience With these steps in view, we can interpret Sellars’ claim that image-­ models are distinct from intuitions and that their content enters into them. This explains how merely represented properties qua imagemodel come to be taken to be properties of an object outside of us. While all of this rests on the claim that intuitions are nonetheless guided by impingements. For an impression of sense or sensory state, comes about in a perceiver due to an outside item, doing the affecting, or impinging of the perceiver. The item thus is prior to its transformation ‘into’ a state, and the state is prior to its projection onto internal objects, and this in turn is prior to projection, qua intuiting, onto something outside of the perceiver. The difference between image-models and intuitions thus has to do with the fact that image models do not contain visual images of categorial features. ›We do not perceive of the object its character as a substance having attributes‹ (IKTE §39). Conceptually loaded content has to come into the resulting notion of an object from somewhere else. 22 The relevant claim here is that we can extract such ‘transcendental content’ from intuitions. Sellars aims to make this point with regard to the difference in logical structure between image-models and intuitions. ›Image-models do not have grammatical structure‹. Whereas ›an act of intuition [is] a demonstrative thought — a Mentalese “this”‹ (IKTE §47).

3.7  Three Forms of Direct Reference With our reconstruction of Sellars’ exegesis of Kant’s epistemological framework in place, we can now draw a tripartite distinction alongside aspects in the concept of ‘direct reference’. According to Sellars’ analysis we have: The direct form of reference characteristic of unconscious, immediate sense-impressions (SM I §37). 23 They are transcendental identified by virtue of their role of effecting specific sensory states in us. 24 Their direct referential character plays an important role because on the basis of this theoretical posit, Sellars can argue that while the conceptualisation of that impingement may be subject to change, the impingement as such is invariant (SM I §21 ff.). The direct form of reference characteristic of intuitions. They are conceptual states which bring both the mathematical and the dynamic categories to bear in the generation of a representation. They do that in presenting us with a conceptually structured content that is spatio-temporally schematised. That they can count as direct referential is a feature of (1), the sense that although two perceivers may bring different terms to the explanation of their states, the perceptual state is brought about by similar sense-impressions. On the basis of the linguistic rules of the language two speakers share, they can then triangulate their responses

Perceptual Experience 91 and learn to identify what the other speaker and perceiver respond to by the use of which term. Thus, they count as direct referential within a given conceptual structure (cf. Chapter 5.3). Although from a transcendental point of view, the reference of intuitions points only back to determinate sensory states, i.e. to internal objects of image-models, not to the objects that brought them about. In this regard, the subject of common sense states of perceiving mistakes the features of his own sensory states to be features of external objects. For the transcendental philosopher, the systematic mistaking is a fundamental feature of our epistemic situation that needs to be accounted for and averts at once (a) a commitment to naive realism or (b) epistemic foundationalism and (c) a lapsing back into the Myth of the Given. And lastly, the third form of direct reference takes the form of an isomorphism between that which effects a perceptual state in us and the state itself. This further form of a relation between the item known and the state knowing it; however, strictly speaking does not belong into the Kantian frame of investigating our epistemological situation. Sellars calls it a picturing relation. The subsequent chapters are devoted to unfolding how Sellars conceives of this relation in rerum natura (Chapter 5) and how it stands to the first two forms of reference (Chapter 6.1.3). In the form of a preview, we can say: Sellars’ analysis of this isomorphism takes it to be a causal, non-intentional, and non-semantic notion.

3.8  The Instability Thesis We have seen Sellars agreeing with Kant that we can characterise our epistemological situation by saying that the common sense objects and the qualities we take them to possess are transcendentally speaking ideal, and empirically speaking real. But according to Sellars, this conclusion confines Kant to a view one which the categorial structure is taken to be invariant, to be once and for all fixed and the categories to be schematised only to space and time. Against this Sellars maintains that we can conceive of the theoretical structures generated by the sciences25 as schematising the categories in new ways that might well go beyond our current articulations of the concepts of time and space. The thesis I wish to defend, but not ascribe to Kant, though it is very much a ‘phenomenalism’ in the Kantian (rather than Berkeleyian) sense, is that although the world we conceptually represent in experience exists only in actual and obtainable representings of it, we can say, from a transcendental point of view, not only that ­existence-in-itself accounts for this obtainability by virtue of having

92  Perceptual Experience a certain analogy with the world we represent but also that in principle we, rather than God alone, can provide the cash. For, as I see it, the use of analogy in theoretical science, unlike that in theology, generates new determinate concepts. It does not merely indirectly specify certain unknown attributes by an ‘analogy of proportion.’ One might put this by saying that the conceptual structures of theoretical science give us new ways of schematizing categories. (SM II §49) In light of these concepts of counterpart qualities, we can think the thought of developing schemes (cf. Chapter 5.3). And this means going beyond Kant. If we need to go beyond Kant it is, among other things, because the Manifest Image conception of the world is unstable. In what follows, we will abstain from defending Sellars’ argument for this instability thesis and instead refer the reader to O’Shea (2007), van Fraassen (1999) (Chapters 1 and 2) and O’Shea (2021b), to Rosenberg 2007 (Chapters 2 and 9), deVries 2005 (Chapter 6) and Wilson (2017, 44–45, 204 n4). These authors discuss in detail Sellars’ commitment to the claim that all our language-entry-moves are in principle replaceable by the language of micro-physics (SM V §90, but see also SSMB §66, PH 101, SII 189).26 The aim of the instability thesis as we will attribute it to Sellars is to make room for advancing beyond Kant’s understanding of intentionality. An ultimate aim of Sellars therein is to show that one can give up Kant’s thesis about the unknowability of the in itself by replacing the in itself with theoretical concepts we can control because we introduced them (e.g. SK I §4). Even though we will not engage in defending this thesis, we can state it in an argument form since it will be of importance in a later stage of our overall argument (Chapters 6.3.1 and 6.3).27 Instability thesis: (1) From a transcendental standpoint one can challenge the MI and its status as a correct framework for encountering the world and explaining our place in it. (2) This involves challenging its basic categories (colour, extension, ­person- and object-hood). (3) Earlier conceptual structures (CSi, CSj) have been unstable. (4) There is reason to believe that there will be rival interpretations of the categorial status of the basic items at the bottom of our current conceptual structure (CSO). 28 (C) The Manifest Image conception of objects and their properties is unstable. With the argument from instability Sellars aims to establish that the by and large Kantian commitment to the empirical reality of the objects of

Perceptual Experience 93 the Manifest Image is under stress and therefore unstable. 29 Ultimately it will have to give way to reformulations of its contents (cf. Chapters 5.3 and 6.3). The argument has its place in the vicinity of Sellars’ approval of two results he takes from his analysis of Kant’s conception of perceptual experience. Sellars agrees with Kant’s contention that two dimensions are vital for any account of our epistemological situation: (K1) The categorial frame and the (K 2) direct reference of intuitions, or in Sellars’ reformulation, the reference we can express in demonstrative phrases of the form ‘this-such’. Also, the analysis of the role of image-models has resulted in an account of the constitutive function which empirical schemata play. It is because of them that we can unite individual perceptual snippets into a coherent concept of an object into a full-blown cognition. With Kant, we can think the thought that the emergence of new empirical concepts will allow us to re-conceive, or to conceptualise in different terms the same intuitional input. That is, we can think the thought that our intuitions will be schematised differently on the basis of new empirical concepts. Sellars agrees with this but argues further that Kant is lacking the standpoint from which to think not only empirical concepts, to evoke not just new empirical schematisations, but also categorial concepts to be schematised in new ways. The relevant point here is that the conceptual structure of the Manifest Image generates aporias of its own (cf. deVries 2005, Chapters 2 and 9-10, 2020a SEP). It generates perceptual puzzles which require explanations that bring us to give up the categorial frame definitive of the Manifest Image and endorse rival conceptions of its contents.30

3.9 Agreements and Disagreements with Kantian Commitments By way of summarising, we can formulate the points of Sellars’ agreements with Kant as follows. (1) We can establish a priori transcendental knowledge of the conditions of possibilities for our knowledge of matter of fact. (2) In as much as the world is accessible to us through the contents of our image-models it is transcendentally ideal. And it is the result of a systematic mis-taking in that we mistake the features of our states to be features of objects outside of us (SSOP §§93–4, cf. deVries 2005, 233, see also Haag 2007, 432). (3) Sellars shares with Kant an understanding of the categories for mental acts, as well as for existing inside and outside a representation. (4) Sellars also shares a version of Kant’s phenomenalism, a version he defends although he does not to attribute it to Kant (cf. SM II 49, SM V §101–102). (5) Sellars also shares Kant’s distinction of the forms of thought and forms of intuitions. (6) And he agrees with Kant that transcendental apperception and object cognition are coeval. (7) Lastly, Sellars agrees with Kant’s analysis of laws as

94  Perceptual Experience complex statements that should be given a metalinguistic interpretation (cf. TE, TC, CIL, and OAFP and TTC). On the basis of what we have established throughout this chapter, there are also a number of points on which Sellars disagrees with Kant. The main point turns on the aforementioned aim of Sellars to replace the notion of the in itself with theoretical concepts we can scientifically control and develop. He formulates this in the following way: Kant’s valiant attempt to clarify the dependence [of appearances on being brought about by things in themselves] by means of a ‘transcendental’ psychology [,] which abstracted form the specific features of human experience as we represent it in the conceptual framework which is natural to us[,] was rejected as involving fictitious mental machinery.31 Kant’s agnosticism is, indeed, to be rejected; but this corrective lies not in dialectical attacks [as offered by the neo-­Hegelian interpretation of Kant’s phenomenalism] on the concept of an ‘unknowable’ but in the development of more subtle theories of concept formation. (SM II, §55 fn1) Sellars here utters a promissory note about how his own interpretation of Kant underwrites his phenomenalism in almost all regards with only some important divergences. The two key ones are (a) the status of colour and shape and (b) the unknowability of things in themselves, for the latter, he claims we can develop a more refined analogical concept formation theory which gets us to how they are. Thus, while we may embrace Kant’s brand of idealism with regard to the phenomenal world, there’s room for a transcendentally motivated realism with regard to the world as it is addressed by the sciences (SM II 49).

3.10  The Status of the Concept of Nature ‘Nature’ or ‘world’ in Sellars are terms of reflexion. We need them to articulate our understanding of causal structures responsible for our perceptual responses. In his critical approach to what could be called his metaphysics of epistemology (cf. ME 61–119), the concept of nature or of a world figures as a placeholder, especially in his characterisation of the natural world as it is understood in the Manifest Image. But this is in tension with the reductive tendencies in Sellars’ nominalistic leanings. According to this nominalism and reductionism, all normative, ethical, religious, moral, and in a related sense language game internal phenomena are reducible to physical objects and their properties. Thus, although Sellars acknowledges the reality and the normative dimension in our common sense usage of the concept of ‘world’, his explanatory aim in the long run is, at least in part, reductive. If the concept of nature can be

Perceptual Experience 95 taken to be the sum of our presuppositional and postulational efforts in the moral and natural sciences, then we are responsible for its content and structure, for its role and significance. Yet, there is a further role for the concept of nature which Sellars’ analysis of Kant’s concept of experience is intended to make room for. It is the idea that our acts of perceptually responding to our environment are themselves part of the environment. The ampliative dimension lies in the claim that their being part of this environment always outruns our specification of them. The focus here is on the acts sustaining conceptual responses to the world, not on the content. They form patterns which are of interest in Sellars’ account of how our cognitions can be taken to be related to the world (Chapter 5, and cf. Brassier’s Foreword above). Our passivity with regard to the sensory input suggests that our conceptualisations of it can always be refined. The realm of nature thus is the realm of what we continuously aspire to grasp. But on account of our confinement to phenomena, to the content of schematised intuitions we can have, we are never directly related to the causal origin, to that which evokes in us determinate perceptual responses. Here it might seem as if Sellars’ use of the concept of nature were ambiguous. On the one hand, nature seems to be a concept of a totality, such that nature ‘is everything’, similar to the concept of ‘world’. It is what is phenomenologically available to us in perceiving the world. On the other hand, nature seems to be that, which can be explained by the methods of the natural sciences. If Sellars wanted to defend the idea that the two coincide, he would have to show that they are identical on account of the thesis that the realm of possible cognition coincides with the realm over which the natural sciences have dominion. But here, it is arguably controversial whether such a project can be carried out on the basis of the methodologies the natural sciences operate with. 32 If Sellars is not a scientific realist in the sense of defending a reductive naturalism, what status does the concept of nature have in Sellars’ account of the concept of intentionality? The reading proposed in the following interprets Sellars to argue for a distinction between (a) the natural world as it is given to us in our perceptual responses and (b) a realm of causes that we take to be responsible for initiating these responses in us. In the following (Chapter 4), we will turn to elucidating the kind of intentionality possible within the former, and then turn to discuss what Sellars has in mind in taking the realm of causes to have a postulational but also a real status (Chapter 5). The guiding question is what form our relation to their respective contents can take. The world we live in and that has meaning and modalities is the world we have in mind in using the concept ‘world’ in common sense discourse. The concept of a natural world that Sellars develops by way of abstraction

96  Perceptual Experience is a very thin concept of that which underlies our conceptualisations, which ignites in us the responses we live by. Although it is true that our perceptual responses to the world are structured by the conceptual materials, rules, and the epistemic norms we bring to the task, there is a sense in which our responses are distinguishable from that which causes us to respond in determinate ways. In other words, there is a realm of causal origins outside of the domain of the conceptual about which we aim to know more, but which also always outruns our specifications of it. However, the natural world as the world of causes is a world that we need to articulate always from within our conceptualisations. And yet our conceptualisations of it form a part in it by being reflected in uniform patterns of linguistic conduct (cf. Chapter 5.1). Sellars thinks that the realm we can be perceptually related to is the space of reasons, while the realm that occasions us to respond to the sensory states we happen to be in is in an important sense larger than the conceptualisations we develop of it. Its ‘contents’ are something we continually aspire to be intentionally related to. It is the aim of articulating the contents of our perceptual responses in functional terms, such that they can figure in our explanations of our perceptual states. The distinction between that which appears and that which causes us to have appearances Sellars takes over from Kant. However, as we saw, he also aims to develop a theoretical framework that allows him to go beyond Kant’s unknowability thesis. What in Kant are the things in themselves in Sellars become the targets and contents of postulational theoretical structures. The interpretive challenge at this point is to clarify in what sense Sellars claims that the theoretical structures definitively and conclusively decide over what exists and what does not. As it is an open question how to understand the structure of our intentional relation to the contents we can make claims about and how to understand our relation to those causal origins which we merely postulate to be in place because we find ourselves experiencing determinate sensory states. Sellars endorses Kant’s transcendental psychology with regard to the resulting conception of objects given to us and the objects merely thought to be responsible for the representations we, in fact, have (cf. KTI §53). But Sellars rejects the thesis that this once and for all confines us to appearances only. Against this, he contends we can move beyond appearances. To see what this amounts to and how he thinks he can support this claim requires specifying what it means to be related to the intersubjectively available contents of the world we live in.

Notes 1. The reasons Sellars has for diverting from Kant at this point will concern us in Chapter 3.6–3.10. And in Chapter 5, we will discuss Sellars’ transcendentally motivated realism in the context of his transcendental

Perceptual Experience 97 linguistics. See Westphal (2020, 330–334) for an alternative articulation of the context a critical understanding of transcendental linguistics belongs in, and passim for a discussion of the contemporary relevance of Kant’s transcendental methodology. 2. For Kant the manifold is transcendentally ‘presupposed’. Articulating his own position, Sellars remarks about Kant that the ›manifold has the interesting feature that its existence is postulated on general epistemological or, as Kant would say, transcendental grounds‹ (SM I §22). The transcendental methodology Sellars adheres to (cf. Chapters 1.1–1.3) shows how on his analysis he can treat ‘positing’ and ‘presupposing’ as functionally equivalent. This he can do because all the method of ‘reflection on the concept of human knowledge’ is to do is show that the concept of knowledge is ›based on, though not constituted by, the impact of independent reality‹ (ibid.). 3. Sellars registers how Kant postulates the existence of a manifold on transcendental grounds but goes on to ›explore independently the idea that it is reasonable to postulate the existence of sense impressions‹. My reconstruction characterises the overall methodology governing this consideration as transcendental because I read Sellars as aiming to provide an account from the in itself to object perception proper. This includes introducing the notion of sense impressions on explanatory grounds, as while working inside the Manifest Image framework (cf. EPM VI, where sense impressions explain cases of hallucinations). This is compatible with holding Sellars to be a committed to a transcendental method throughout his writings (i.e. from EPM to SM, SSIS, FMPP and SSOP). The pressure generated by the impact of an independent source carries the concept of passivity over into the scientific framework and the elaboration of the placeholder notion of ‘sensa’ therein. Reflections on the ontology of perception are tied to the framework they are carried out in. 4. Cf. ›Kant is as willing as most theologians to make use of “analogy”. But, as has often been pointed out, when the granted negative analogy is of a certain character, only the word remains‹ (I §85, fn29). 5. The negative answer Sellars attributes to Kant concerns the thesis about the unknowability of the in itself: ›What he [Kant] does deny […] is that the relations we conceptually represent are the relations which the in itself exemplifies‹ (SM I §76). 6. Here one might have a different argument for sense-impressions in mind, according to which the manifold-of-sense is postulated by the theory as caused by ‘physical objects’ in the Manifest Image sense of the term ‘physical object’. In line with Chapter 1.3 above, this much can be conceded while adding the following point. In the order of explanation, we start within the framework of physical things. And by invoking relevant analogies, we introduce the framework of sense impressions as playing a role in our perceptions of these things.   However, it is only from a higher standpoint of distinguishing frameworks as such (cf. PH VII) that Sellars can maintain that ›the interpretation of the framework of sense impressions as a theoretical framework suggests that the analogy between the attributes of impressions and the perceptible attributes of physical objects is but another case of the role of analogy in theoretical concept formation‹ (SM, I, §53). So what links the introduction of the notion of ‘sense impression’ and ‘manifold of sense’ is the search for a causal story giving the order of coming into being in which our perceptual episodes are evoked in us. The notion at the root of this story is the Kantian ‘in itself’ viewed as a placeholder concept to

98  Perceptual Experience









be substituted in a later conceptual frame. Cf. WSNDL 201–202: ›Kant’s ding[e] an sich in my view become the scientific objects of theoretical science. Using that metaphor what we have is science as giving us insight into that which it is which appears to us in the conceptual framework which we learn […] use in acting and suffering and thinking‹. 7. In that sense impressions help to explain the occurrence of certain conceptual representations – a fact for which they, as unconscious responses, need not be consciously perceived (cf. SM I §22, 28, 42, 44). 8. The important distinction here is that although sense impressions count as representations, they precisely do not present a manifold of sense as a manifold of sense. Sellars underlines this by arguing that sense impressions are not of anything complex (SM I §19, 21). 9. But see Willaschek (1998), Allais (2004) and Rosefeldt (2015) as influential contenders on this issue. See Stang (2016) for a recent discussion of reasons for why one might suppose that there are two distinct worlds. 10. Macbeth (2010, 204) nicely captures the constraints of this stance: ›At the level of sense impressions we just do – mechanically, as it were – correctly picture how things really are. But such pictures are not true because sensing is not knowing, not an episode in the space of reasons. What is required for knowing is that we have concepts of such impressions (or concepts of whatever is the counterpart of such impressions in the final theory), and we can come to have such concepts, Sellars thinks, in virtue of our capacity to develop theories involving analogical concepts that will not only explain why ordinary perceptible objects obey the lawful generalities they do as far as they do, but also (because they are analogical) ground the idea that it is those very sense impressions that our talk is about‹. 11. And then there’s a further way of elucidating the relationship between the two approaches that on a surface level appear to be methodological equals. The analysis carried out in IKTE has its focus on the phenomenal features of experience and advances in a descriptive manner. But in SM, KTE and elsewhere Sellars takes a transcendental stance towards the very idea of a theory of experience. Which is why he there speaks of postulation and addresses the notion of the ‘in itself’ in a manner reflective of the presuppositional structure requisite for referentiality and intentional directedness. Above I provided an account of how the transcendental method of analysis and the analysis of frameworks are integrated (Chapter 1.3). The phenomenological considerations put forth in IKTE presume the transcendental considerations of SM – i.e. ›contain the gist of the Kantian scheme‹ (IKTE §1) — in precisely the sense that the impact of the ‘in itself’ can be described as a general feature any theory of perception will have to introduce. In a second step, this feature can be characterised from the internal standpoint of our current conceptual frame by carrying out a phenomenological analysis articulating the ›structure of perceptual experience‹ (§3) basic to this frame. 12. Chapter 6.3 discusses in more detail this tension between the phenomenological reflection and phenomenalism of our Manifest Image reality and the realism Sellars defends from a transcendental point of view. 13. Cf. Haag (2015) an extended discussion reconstruction of the functions the productive imagination has in Kant. 14. This will later on lead to the question whether image-models already contain categories, and if so which categories. Since Sellars is willing (1) to take intuitions to ‘contain’ categories, and since (2) the content of image-models enters into intuitions, it is at least an open question whether Sellars should

Perceptual Experience 99 have claimed that (3) the mathematical categories of quality and quantity already inhere in image-modes. After all, we conceive of image-models as individual contents with determinate qualities. 15. The resulting ‘unity’ allows for two readings. On a first, additive, reading ‘sensing cum image’ can obtain independently of conceptualisation since the form of sensing-cum-imaging merely guides the application of a certain concept ‘from the below’. On the second, ‘transformative’ reading ‘sensing cum image’ is not separable from the unity given to it by the concept. On this reading the schema-guided unifying process constitutes the ‘sensing cum image’ in the first place. The components are not independent from the act of unification, and is rather the concept that guides the formation of the image-model. Given Sellars’ account of empirical schemata (elaborated below) I am inclined to endorse the second reading. If I give a chronological rendition of the elements – first of sensing, then of sensing cum imaging and finally of conceptualising – I do so in a manner of analytically reconstructing them before highlighting their transformative role in the unifying process (IKTE §26) bringing about object-perception. 16. Sellars’ account of transcendental schemata is distinct from his account from empirical schemata. Cf. IKTE §41 where he underlines Kant’s use of transcendental schemata by way of interpreting their function as that of specialising categories to the confines of our spatio-temporal intuiting. The contrast class here being God’s pure-intuiting unbounded by time and space. Cf. O’Shea (2021b) and Cumpa (2020). 17. The IKTE-slides in KTM suggest that even the perspectival nature of one’s experience is a function of the combination of the concept of an object (pyramid) and the concept of oneself as its perceiver. I am indebted to Luca Corti for this hint. 18. Cf. e.g. as are familiar from the drawings of M.C. Escher or manifest in the copper works of Fred Eerdekens, www.fred-eerdekens.be/files/­downloads/ light-fredeerdekens-svh.pdf (last access 13.09.2021). 19. This material cannot be identical with but feeds into phenomenal properties. To clarify, an episode of ‘sensation’ (sensing red-ly) but also phenomenologically identifiable aspects of experience (this expanse of red in my experience), as well as the concept of ‘red’, are downstream from that shift. That is what it means for unconscious sensory material to feed into conscious representations. 20. It’s an interesting question in how far the bodily states be better conceived as pre-conscious than as unconscious. After all, we do feel our body at various times in various places alerting us to its affections. These are affection, most of which do not represents anything beyond the nerve-signal that they carry. Yet it is also clear that one cannot be conscious all at once of the various signals that together compose a tactile impression. This would mean that each and every signal-component would be accompanied by a conscious representation which itself would have to be independent of a neural basis. However the point of giving conscious representations a neural or sensory basis was exactly to be able to say that the productive imagination picks and chooses, out of the myriad signals that there presumably are, only those that in accordance with an empirical recipe allow it to construe the image-model content of the core components which result in an intuition. It is in this Leibnizian sense that we can conclude: There are always more pre- or unconscious sensory states, petites perceptions, than end up in image-model contents of which we in effect are conscious.

100  Perceptual Experience 21. Transcendentally speaking “raw materials” feeding into this process are an abstraction from the whole that is always already unified. Such wholes, however, do only exist from the perspective of one framework or other, in our case, the framework of physical things. The point of delineating the operation in a chronological, “additive” way is to show what activities are working on materials, the conceptualisations of which will develop in accord with the frameworks and conceptual systems we bring to the task. Cf. Chapters 5.2.2 and 6. 22. Cf. Landy (2009), and more extensively Landy (2015), who has argued for an inferentialist understanding of object representation and perception. While I am generally in favour of Landy’s line of impact, especially of the way the Deductions are evoked to substantiate this point, I like to think that it lacks the distinction of levels of abstractions that transcendental Sellars’ project is based on. 23. Why call it reference? Only for the sake of aligning and distinguishing between the three forms. On a thinner characterisation the notion of directness captured under this heading is but a simple form of causal connection (postulated by the theory) between an object and an impression. From a transcendental point of view, this point matters for the following reason. While ever new frameworks for the interpretation of (the origins of) our experiences can be developed, this minimal notion of a causal connection has to be retained throughout. Thus, while the notion of ‘sense impression’ will be disposed (SSIS 413) precisely because it belongs to a theory of perception that is typical of the Manifest Image, it is the causal theme inherent this thin notion of directness that forces us to find counterpart concepts – sensa as Sellars argues. 24. Expanding on Sachs (2011, 2015), Corti (2021a, 3–4) distinguishes transcendental from explanatory arguments for the introduction of sensations. I am open the idea that there are contexts in which the psychology of perception requires positing sensations in an ‘explanatory’ sense. However, since sensations or sense impressions are neither conceptual nor do they play an epistemic role, such that ›the subject of sensations is not aware of them (they are not apperceived), … epistemically inert‹ (ibid. 7), I read Sellars’ treatment of sensations as argumentatively in line with his overall transcendental methodology. 25. In this passage, Sellars leaves it open which of the sciences he has in mind. As I see it, nothing prevents Sellars from accepting an interpretation according to which it is not just the hard sciences but also the moral sciences. Along these lines Brandom’s modal-Kant-Sellars thesis can be read as an attempt to establish the role of subjunctive reasoning as basic for the use of empirical predicates. Cf. Brandom (2008, 2015). 26. See Haag (2007, Chapters 9 and 10) and Haag (2016) for a defence of Kant against Sellars’ criticism. For Haag the status of colours and extension play a central role in a way in which they do not in my argument. 27. Every argument that forces us to give up the Manifest Image framework as a framework for explaining the nature of objects of (possible) experience has to be a transcendental argument. 28. For instance, it is clear that in psychiatric contexts we have replaced the concept of a ‘soul’ as an explanatory concept in favour of ‘neurological states’. And where appropriate, in our psychological medicinal practices we defer to explanations that invoke biological and bio-chemical processes. 29. See the final discussion printed at the end of Pereplyotchik & Barnbaum (2016) for an overview of current positions towards this thesis.

Perceptual Experience 101 30. As O’Shea (2007, 163) puts it: ›We cannot know in advance, of course, what the specific categorial ontology of the ideal Scientific Image would look like … but it is encumbent on the transcendental philosopher to envision what its most general categorial features would have to be like ‹. Cf. also Bremer (1997). 31. Cf. KTE §39 where Sellars makes the same point. 32. Cf. deVries (2016c, 219 ff.) and Hicks (2020) for a recent contribution to this discussion, also see the Conclusion.

4

Non-Relationality

This chapter deals with the following question: What does Sellars’ non-relational account of intentionality consist in? This involves addressing why Sellars develops an account of intentionality of such a peculiar form, according to which there are no intentional relations between items in the conceptual order and items in the real order. The significance of this latter question can be seen when we recall from the previous chapter that Sellars endorses the consequence of Kant’s analysis of objects of experience that what we represent as the content of our appearances is transcendentally ideal. Sellars arrives at this strong result with Kant and not against him (in SM II, KTE, and KTI). The result is that the world, according to Kant, exists only as the contents of acts of representings. And it is this result which is behind the need for a theory of intentionality. Such a theory, accordingly, needs to be able to accommodate the lesson learned from Kant but also needs to address the further question: How can we nonetheless be said to live in a world in which our states and utterances have meaning? The threat giving urgency to this worry is that such a theory could be taken to imply a strong coherentism. This would imply that all the claims we think are contentful by being about the world are really spinning in the void without having any friction with a realm outside the conceptual.1 On the surface, it seems, empirical claims stand in a relation to matter-of-factual affairs, while on account of the non-relational deep grammar of content-ascriptions, that would never be the case. 2 From this follows a further question we need to address: How is this analysis of intentionality compatible with Sellars’ commitment to a transcendentally motivated realism? This last question provides the target point for which this chapter prepares the ground. Thus Sellars’ non-­relational account of intentionality is presented here, while the larger context in which it systematically belongs becomes clear in Chapters 5 and 6. The following discussion develops the reason for being dissatisfied with the kind of answer Sellars gives in SM III and related writings (mainly MFC, LTC, NI, TTC, and NAO IV–V). It presents the non-relational account of intentionality as consistent with Sellars’ transcendental methodology. A DOI: 10.4324/9781003221364-5

Non-Relationality 103 proper understanding of the reasons Sellars has for developing this conception of aboutness and meaning requires making this methodological background explicit.

4.1  The Expectation When we reconstructed Sellars’ exegetical appreciation of Kant’s theory of experience as a transcendental form of linguistics, we saw him arguing that Kant’s account of our epistemological situation can be made coherent (Chapter 2). There he argued that the in itself should be accepted as a transcendentally postulated source of our affections (Chapter 3). In light of this, one can take Sellars’ account of intentionality to emerge out of his engagement with Kant’s transcendental idealism. It becomes clear when we pay attention to the question with which Sellars opens his own account of intentionality in SM III, where he asks: ›What, after all, does it mean to say that a content exists ‘in’ representings?‹ (SM III §4). Following up on this question, Sellars brings out that the Kantian theory of objects of experience concerns individual and determinate contents. But an encompassing theory of intentional phenomena populating our lifeworld would not only need, among many other things, to accommodate individual and determinate contents but also general and abstract contents.3 How can these, as contents, be shared by one thinker across time and among different thinkers (cf. SM IV §1)? That is, how can we explain that one content is shared by many? (cf. SM III §5). It is this demand for a theory that explains the intersubjective availability or the public character of what we conceptually represent as real which constitutes the systematic aim for Sellars in developing an account of intentionality. The abstract formulation of the aim of such a theory of intentionality – as serving to explain how one content can be shared by many – can easily conceal the wide-ranging importance that goes with this aim. Thus it becomes necessary to mark the distinction between explicating what goes into possessing the concept of an object of experience and explicating how we can live in a shared world of norms, values, standards, and shared intentions. A theory of our relation to the world will surely want to accommodate these at some point in its articulation. Kant’s theoretical framework may allow us to understand what requirements all thinkers are subject to when considered formally as epistemological agents. But it seems to be a separate task to explain how our reference to a world we did not make allows us to furthermore be in touch with abstract and general objects and be shaped by linguistic and epistemic norms in perceiving things to be a certain way. In other words, an encompassing theory will have to be able to include these kinds of contents. But the more general and explanatorily prior aim is to understand intentional directedness to concrete individual contents.

104  Non-Relationality If we consider Sellars’ program as answering to these requirements, what we arrive at is the kind of expectation a reader of SM will bring to bear in its third chapter. To flesh this expectation out a little more: One may well come to a first reading of SM III with the following expectation: We wanted to know how we can be meaningfully related to the world of which we are a part. However, the exegesis of the Kantian concepts of Sensibility and Under­ standing in SM I and of Appearances and Things in Themselves in SM II left us stranded with a dissatisfying conception of ‘the world’. The conception of an object of experience dissolved into mere appearances and the ‘in itself’ in which they were supposedly grounded is analysed as mere transcendental postulate. A Kantian may manage to feel at home in such a world, but a metaphysically more ambitious reader will hope to get beyond appearances. So it is this expectation with which one may come to read Sellars as proposing a theory of the aboutness of our mental acts with the aim to get us further than Kant did. And there is surely reason for taking Sellars, writing in the mid-1960s, to be articulating such a theory in the spirit of the linguistic turn. After all, the central idea SM III revolves around is that of ‘meaning’ (cf. SM ix). The expectation raised can be specified further, in the words of Quine, as the hope that Sellars, in providing a theory of meaning will give an account of our terms’ reference to the world (cf. Quine 1964, Chapter 3). And this hope is not entirely unjustified, since Sellars does claim to develop an analysis or semantics for ‘means talk’ (cf. EPM VIII, SM III, IV), and to provide corresponding analyses of the related notions ‘refers’, ‘represents’, ‘denotes’, and also ‘is true’ (which he does in IM, SRLG, LTC, MFC, NI, CIL, MFC, TTC, TTP). In other words, it is a plausible expectation that Sellars’ theory will provide an account of what it takes for our expressions to put us in touch with the world. The reader may have developed this expectation in the previous chapter already with regard to the intentionality of acts of representing and may have initially been disappointed there. Now a similar form of disappointment threatens Sellars’ program. The source of this worry is the counter-intuitive idea of the non-relational nature of conceptually shaped intentionality. Yet the fact that Sellars develops a unified account for an entire array of technical terms concerning modes of aboutness seems to promise an encompassing answer.

4.2  The Non-Relationality Thesis Sellars sees his articulation of this conception of intentionality as standing in line with his previous work. In the preface, he writes about his undertaking in SM III: I develop in detail the account of ‘intentionality’ which was sketched in ‘Empiricism and the Philosophy of Mind’ […]. Among other

Non-Relationality 105 things, this chapter provides, as I see it, the cash for a long standing promissory note concerning the non-relational character of ‘meaning’ and ‘aboutness’, a thesis I have long felt to be the key to a correct understanding of the place of mind in nature. (SM ix) There are many regards in which one can connect EPM to SM. Sellars refers to two main points inherently connected with the transcendental methodology. The first point criticises the claim that we are directly or, as we will go on to say, ‘vertically’ connected to the world. This claim finds its expression in the rejection of the Myth of the Given (cf. O’Shea 2021a). The second point targets the claim that we get vertically connected to items in the world when we are introduced to the logic of ‘means talk’. These two points are essential to the rebuttal of direct relationalism and come into play when Sellars develops the horizontal antithesis, i.e. his non-relational conception of aboutness. From EPM (1956) to FMPP (1981), Sellars criticises conceptions of our epistemological relation to the world that fall for the Myth of the Given, repudiating following claim: If a person is directly aware of an item which has categorial status C, then the person is aware of it as having that categorial status C. (FMPP I §44) The world does not imprint itself on consciousness as a seal imposes an image on melted wax (FMPP I §45).4 An item’s categorial status does not reveal itself to us simply in virtue of affecting us. In an extended reading of that rejection, this applies to all items one can be conscious of. For being conscious of something always entails being conscious of it as something. 5 And it is this qualification of the content which Sellars highlights when he points out that the very determination of something as something is inherent in every act of cognition. It is so fundamental that it structures every act of taking something to be the content of a thought, or the content of a conscious act of thinking. It is this thesis he is after when he argues that it is always a matter of classifying a content with regard to its features as of a certain kind (cf. SM III §43, §67 and SM IV § 8–9). If the point holds that this analysis pertains to items of any sort, it carries over to mental items too. And this, in turn, entails a rejection of the very idea that the contents of our thoughts could be transparent to us from sideways on. If mental items do not simply reveal their status to us, this means that even the most simple act of coming to be conscious of something is an act of coming to be conscious of it as of a certain character. And Sellars’ thesis is that this is a matter of classifying it with regard to the function it plays (cf. Chapter 2.1). We may be

106  Non-Relationality occasioned to represent a mental item, but this representation does not get its status accorded to it by way of being the result of an affection. This fundamental idea constitutes a cornerstone of Sellars’ philosophy of mind.6 The transcendental line of thought already inherent in EPM then is this: The very idea that something can be directly given to us without also being embedded in a larger context is to be rejected at all levels of inquiry (cf. EPM sect. VII). The argument here is that all representings are necessarily embedded in inferential practices. And this larger context, according to Sellars, is not just formed by ‘being in a solitary act of classification’, but by the idea that such acts need to be learned in intersubjective interactions. One is inculcated and educated into being a member of the community of functionally classifying one’s states in accord with the epistemic, conceptual, and discursive norms of one’s peers.7 From this functional point of view, intentional vocabulary serves the purpose of allowing us to draw distinctions and exert control in making and having and sharing experiences. This is the larger context in which acts of classification make sense. If we are to bring our theorising about the intentionality of our states and claims up to speed then, how are we to understand the intentionality inherent in being conscious of something, in ‘having’ something as a content thought of, or in perceiving something to be the case? Sellars puts the non-relationality thesis about intentionality in a way that demarcates their fundamentally holistic nature as conceptual items: Non-relationality thesis. All aboutness is to be understood on the model of the attribution (in meaning statements) constituted by functional classification and does not involve relating a linguistic to a non-linguistic item. If no item can be given to us directly, and if being conscious of an item is a matter of its functional characteristics, then all items have to play a function that in one way or another is related to other items embodying further functions (cf. SM III §57, NAO V §1–3). In light of these considerations, we can turn to the announcement Sellars makes in writing that SM III ›provides […] the cash for a long-standing promissory note concerning the non-relational character of ‘meaning’ and ‘aboutness‹ (SM ix). The qualification ‘non-relational’ concerns the part of his thesis that all items, i.e. content-bearers qualify as such (i.e. as meaningful items) only because they are suitably related by being functionally classified, not by being about an item that is directly given to us in a direct perceptual act (cf. Chapter 2.2).8 To be clear, ‘aboutness’ and ‘meaning’ may be properties of terms or claims or characteristics of mental states, and while one can agree with all of this, one might still object the following: Is it not the most natural

Non-Relationality 107 thing in the world to consider my thought about a pyramid in front of me to be about that very pyramid? In light of the Kantian transcendental analysis of what is involved in being in a perceptual state, we can reply to this objection from common sense as follows: The very starting constellation we take to be a natural assumption – i.e. that there is a pyramid, and that we can form thoughts about it – is already the result of a sequence of synthetic and therefore classificatory operations. Starting this way, the objection already presupposes a relational conception of reference, while on Sellars’ analysis, the relevant sequence involves a transformative synthesis and classification of elements according to categories and implicit empirical concept-formation rules (Chapter 2.2). Sellars does not reject aboutness claims as such, but only relational analyses of them. With regard to the Cartesian distinctions we encountered as fundamental also to the Kantian transcendental psychology, we can say: The very categories of existing inside or outside of a thought are at work in even grasping a content (cf. Chapter 2.2). It is in this sense that the EPMtheme of rejecting the Myth of the Given in all its guises carries over into the analysis of aboutness in its various forms. No item presents itself to us directly, with its categorial status imprinted on it (FMPP I §§43–45). And if existence inside or outside thought is never simply given, the only means to explain how aboutness can be non-relational is to articulate the constructive involvement of the thinker and perceiver in the process of relating to items she comes to grasp as existing independently of her. Since the Myth is to be rejected in that most fundamental of forms, we are never in a position to directly compare an item in the world with an item in mind, i.e. consciousness. For both are only given to us in light of being contents of our acts of representing, contents we classify as being of a certain kind. This is the argumentative background to the thesis that aboutness has a non-relational character. Sellars substantiates this non-relationality thesis by interpreting meaning statements as metalinguistic. That is, he analyses statements giving the meaning of a term and argues that they are, in a concealed way, really about linguistic functions, not about the objects in the world. Meaning statements occur, for instance when learning another language, when we learn that ‘rot’ in German means what red means in the English language. The positive claim holds that they tell us something about the linguistic function of an expression by way of classifying the item functionally, an idea for which Sellars invents the convention of dot-quoting the relevant role (SM III, NI, MFC). Expressions such as ‘rot’, ‘rosso’, and ‘rouge’ are • red • s which function like ‘red’ so we can learn about a new expression by taking it to be an item of the same category, as having the same function. The negative claim Sellars makes about meaning statements is that they do not tell us how a determinate meaning gets attached to a certain object.9

108  Non-Relationality A spatial metaphor may serve to elucidate the general sense of the two claims. (1) Sellars’ positive claim amounts to the thesis that in meaning statements, we get pointed in a horizontal direction. This is most readily visible with syncategoremata, words such as ‘all’, ‘and’, ‘if’, or ‘not’. We do not learn these words due to their referential powers. To understand them, we need to learn about the way they function in bigger, i.e. sentential and inferential contexts. However, the way in which meaning statements draw a horizontal connection between the linguistic functioning and role of words is less obvious when we consider terms10 that seem to have a clear extra-linguistic referent, such as a pyramid or an apple. (2) But Sellars also wants to deny meaning statements about these latter expressions a vertical power of attaching to objects. This includes denying that meaning statements themselves attach to objects and that they describe the words as attaching to object. He grants that it may be that we have occasion to learn a word such as ‘pyramid’ when confronted with a pyramid (cf. LTC 511, 513; MFC 421–423.), but, – pace the Myth of the Given – it is not the case that ‘reference’ is a relation between linguistic and non-linguistic items, nor is it the case that a word gets its meaning imbued onto it in virtue of the pyramid presenting itself as the term’s referent.11

4.3  Objections from Relationalism How plausible is it even to suggest that our language is not vertically about the world but its aboutness only a matter of functional classification? Does Sellars intend to argue that we are never in touch with the world when we speak about it, that we are always spinning in the void with our words? To come to terms with these questions, it is helpful to reconsider what Sellars in effect argues (cf. MFC 421–423, NI §§11–25). His train of thought can be stated as follows: A statement that gives you the meaning of a term you have not learned yet has the job of conveying a meaning. The statement does that job by transporting you to other terms and expressions, not by establishing a direct connection to a non-linguistic entity (cf. SM IV §8 ff.).12 What if the scenario is changed, for example, such that one confronts an unknown object and is told that around here, one refers to such a thing by uttering ›It is a pyramid‹. This objection challenges Sellars’ claim that this is a non-relational affair. Yet, the objection also changes the setting quite a bit. Sellars aims to make his point by dissecting meaning statements as having the metalinguistic job of conveying the linguistic functions of expressions mentioned in them. To move to a concept-­acquisition scenario requires a different commentary, which would rest on an account of how speakers can be trained to identify objects and use expressions according to those functions. How then could such commentary move anyone having inclinations towards a relational conception of meaning to even consider a

Non-Relationality 109 non-relational, horizontal view of a term’s ‘meaning’? To respond to such an objection Sellars’ commentary would have to establish a number of points about the way in which thought, perception, and the linguistic expression of a perceptual episode hang together.13 Of interest for us, i.e. from the methodological point of view we have been unfolding, is the fact that this objection from relationalism slips in an unqualified sense of ‘object’. What is illegitimate from Sellars’ point of view is that assuming an entitlement to such an unqualified sense of ‘object’ entails the claim that we can, in effect, be perceptually conscious of an object as a bare something (a bare particular). In other words, the entitlement is illegitimate because of the entailment. This would imply being conscious of it as something for which the epistemic subject merely happens to have no words, as of yet. Against this, Sellars holds, even the most primitive of perceptual episodes always already has the form ‘this-such’ and is never a bare ‘this’ (SM I §§11–15).14 This objection from relationalism is the concept-acquisition version introduced above. The reply to this objection now targets the very idea that an item is taken to be directly available to a perceiver, even prior to her having words for it (cf. GEC §16, 44, MGEC §18). For according to Sellars’ rejection of the Myth of the Given, such direct and innocuous availability was never an option.15 It was not an option with regard to items presented to us by outer sense nor with regard to items we come to be conscious of through inner sense. This allows Sellars to counter that to be conscious of an item is to have already responded to it by way of classifying it. With this established, a next step is to make explicit that the objection from relationalism assumes the validity of the premise that we can be unproblematically and directly be conscious of an item. To show the problematic structure of that premise, we can make it explicit as the assumption holding that: One can be conscious of something as something without already having qualified it as of a certain kind. The reply Sellars gives in defence to this objection is conceptualist through and through. A perceptual response, according to Sellars’ reading of Kant’s theory of experience, is never unmediated. But that should not come as a surprise, given the role that meta-classificatory judgments played above (cf. Chapter 2.1) where we articulated Sellars’ endorsement of Kant’s analysis of what it takes to be an object of experience for us. But are we justified to drag this analysis of how concepts and objects hang together in Kant into a defence of non-relationality about meaning-­ statements? According to our reading, this is justified and required. For when Sellars rehearses Kant’s view on objects intuited, he states that ›to be an item one can be conscious of is already to be a conceptual response‹ (cf. KTE §42). This justification derives from the premises (a) that all our perceptual responses to the world are already conceptually shaped and from the claim (b) that relating one’s conception of the item to a further expression is a meta-linguistic doing. Lastly, (c) it also derives from the

110  Non-Relationality premise that perceptual responses systematically involve mis-takings of the features of one’s perceptual states as the features of external objects (cf. Chapters 3.7–3.9). Most likely, a hardened relationalist will not quite be convinced by this line of argument. A further objection could go like this: To say that objects can only be something for us when they are taken up in perceptual acts and thereby necessarily become conceptually structured does not preclude a relational conception of intentionality.16 After all, we are impinged upon by objects, and this fact does not change even though we need a holistic web of words to say meaningful things about them. The gist of this objection can also be clothed in the spatial metaphor introduced above: Why should we give up the vertical dimension (of our words’ relating us to their referents or their truth-makers) just because we need the horizontal dimension to explain the semantic significance of any given single word? And does Sellars want to give up the vertical dimension altogether? As this objection and its concomitant question are more serious than its prima facie looks, it persists and is addressed again in Chapters 5.2 and 6.1. For in fact Sellars does not want to deny the ‘causal theme’ inherent in the idea that we are affected by the world and, in a sense, thereby prompted to come up with conceptual episodes and words expressing what we take ourselves to have experienced.17 And Sellars can concede that objects prompt us and are part of language-entry uniformities, without holding that such encounters alone suffice for objectterms to have the meaning they do. But the deeper problem for Sellars’ account concerns the more general fact that his alignment with the Kantian theory of experience requires him to treat the inner and the outer world as necessarily subject to acts of classification and interpretation.18 As a transcendental philosopher, Sellars takes the assumption to be wrong that the world is directly given to us. But he also argues that the sheer manifold plays a substantial part in the way in which we come to have unconscious and conscious representations of it, which we take up into qualifications that accord them epistemic significances (SM I §10).19 For Sellars wants to retain the idea that affection is prior to classification (in the order of being)20 and also underwrite the claim that classification is prior to being in a meaningful perceptual state (in the order of understanding). 21 The solution proposed here lies in methodologically distinguishing the horizontal and the vertical aspect of the difficulty at hand. 22 Sellars’ thesis that meaning-statements and their particular aboutness and role are to be interpreted along horizontal lines as non-relational thus concerns the question of how words, mental states, and items with an inferential articulation more generally – that is, all items embodying a functional role – can have an intentional purport which points beyond them. Sellars answers this question by arguing that such items are best

Non-Relationality 111 identified, and their aboutness be best understood as articulated by ascription of the functional roles they realise.23 He explains his strategy for substantiating this thesis as follows: My ultimate aim is to argue that extensions are limiting cases of intensions and cannot be understood apart from them. Thus, classes, in the logistic sense, cannot be understood apart from properties, nor truth apart from propositions. As I see it, Quine’s attempt to by-pass intensions simply misses the point. He has looked for intensions in Plato’s beard, but, like the bluebird of happiness, they have always been in his own backyard. (SM III §43) The relevant argument here is that a Quinean outlook aiming to do away with intensionality altogether (Quine 1960, Chapter 6) is to be rejected on the ground that our grasp of senses (contents conceived of as intensions) is conceptually prior to that of extension or denotation (SM IV §86, SM III §67). Senses, on this account, can only be understood from a standpoint skilled in navigating the logical space of reasons (EMP §38). And it is the semantic rules structuring this logical space which are constitutive of the identity of the senses. They are what governs the expressions and allow us to reflexively sanction each others’ use of language (SM III §39–41). For ‘understanding’, an expression amounts to being able to deal with the role a linguistic or conceptual item plays for us by way of suitably relating it to other such items (IM 337–338). The real mistake lies in assuming prior acquaintance with extra-­ linguistic objects or items in the real, an acquaintance which would have to take place in the total absence of normative specifications of their role in our lives. To Sellars, this is inconsistent, as we have seen in tracing this line of thought back to his argument against epistemic foundationalism in EPM (cf. SK and SSIS, SSOP and PH). But this, as such, does not exclude their vertical aspect altogether. The explication of the concept of intentional relatedness to a world we are a part of does not come to a final halt here. We still need to accommodate the fact that our perceptually takings, our acts of responding to the environment ‘as being a certain way’, are occasioned vertically by something outside of the sphere of the conceptual.24 And we need to account for the fact that they, as acts, also have a certain materiality to them. Notwithstanding that our responses can only be meaningful to us by being conceptually shaped through and through. In presenting the encompassing account of the non-relationality of aboutness and meaning (and of denotes, and refers), Sellars also has a further aim which was left aside so far. We can introduce this aim with reference to the idea that our responses to items inside and outside of us are learned responses. They are learned in the sense that we have been

112  Non-Relationality trained to respond to our environments, to respond in systematic ways. This training began shaping how we take in what we see long before we were in any position to say what was going on there. In fact, learning to say what is going on is learning to comply with norms, implicit rules the linguistic community trains its new members to internalise. 25 The idea that our conceptual-inferential, as well as our non-inferential direct responses exhibit a certain uniformity is what allows Sellars to introduce the claim that functional roles, reflected in pattern-governed, inferentially articulated uniformities, can be shared among different members of a language. 26 According to this characterisation, ‘sharing a meaning’ is an immanent activity (cf. Chapter 5.2.2). What is more, the tie between language and the world Sellars thereby introduces is indirect in the following sense. In reflecting on what is entailed in treating a statement as expressing a matter-of-factual truth, Sellars writes, we move in accordance with a principle of inference that licences moving from ‘that green a is true’ (in our language) to ‘green a’ (in our language) (NS 246). The importance of interpreting truth and meaning talk with regard to inferential moves lies in the following: [I]t is by virtue of the fact that we draw such inferences that meaning and truth talk gets its connection with the world. In this sense, the connection is done rather than talked about (NS 246) The connection is done in the sense that we are performing truth moves. We treat certain assertions and claims as claimable and act in accordance with inference rules we consider each other to be bound by (i.e. rules involving use alongside mention of semantic vocabulary, i.e. meaning and truth talk). 27 This, however, averts a coherentist picture of a frictionless realm of meanings. For our engagement in a materially rich linguistic practice is grounded in a rule-governed exchange of linguistic uniformities, utterances and overt behaviour among speakers. Although the role of the ‘truth move’ is central also to Sellars’ argument of how to understand the relation of thought and language, it was treated as subordinate here. 28 Still, ‘meaning’ is not be equated with ‘reference’, for Sellars argues that ‘meaning talk’ applies even in the case of discourse about fictional entities (cf. SM III). For us it suffices to understand Sellars’ claim that we are trained to interpret our perceptual experiences and our affective states as being of a certain character. And that we therein come to be able to critically assess the correctness of others’ and our own responses in light of more or less explicit knowledge of the norms and rules for such criticism. 29 What stands out from this is that we are bound by the rules governing our use of the conceptual material we have available as a linguistic

Non-Relationality 113 community of speakers. And ‘material’ here means the resources of norms and standards, of rules and instructions, trainings, and inculcations. Those guide and enable our attempts at making sense of our experiences infer what to think next and come up with practical conclusions on which to act. But the radical conclusion remains that semantic expressions or (linguistic representatives) and statements composed of them giving the meaning of terms, and statements reporting what a mental state is about, never get us to objects in the world. And Sellars can draw this conclusion because the idea that our conceptual responses are vertically occasioned has no epistemic significance, on the assumption that classification is in fact prior in the order of understanding to states of being affected. Sellars’ commitment to the thesis ‘affection first’ will be picked up below (cf. Chapter 5), when the theoretical elements are in place that are requisite to spell out his matured position. What matters here is that horizontal non-relationality rests on the claim that we train each other to respond in accordance with the linguistic and perceptual norms to exhibit the uniformity of patterns we can all agree on as being correct.

4.4 Dissatisfaction If we take seriously the transcendental background of Sellars’ position on meaning and aboutness as non-relational, this has strong implications for the project of understanding our knowledge of objects and for the attempt of clarifying our own position in the scheme of things. We followed Sellars’ Kant exegesis to a point which left us with appearances and with higher-order knowledge about the necessary structure the knowable world has to have to become knowable to us. But then Sellars tells us that all statements in which we can express the meaning of our perceptual states and the aboutness of our words are never vertically about the world. The reasons Sellars lays the focus specifically on statements about words and perceptual states is that on his account, these are a special case of our talk about the world but in their speciality the most fundamental cases for the application of a non-relational account. From here, one can generalise, so to speak. This non-relational intentionality thesis implies a strong anti-representationalism. Transcendentally speaking, we never perceptually represent the world the way it is. For the phrase, ‘the way the world is’ to us only has meaning on the basis of our classificatory activity, never independently of it. Or rather, since all perception involves systematic mis-taking (Chapter 3.8) and is necessarily conceptually mediated, a representation is never a simple, direct relation to an item in the non-linguistic environment (cf. MGEC §18). The push towards this conclusion gains momentum the very moment we try to presuppose an innocuous sense of the world and find ourselves having to spell out this term ‘world’ in terms of discrete objects

114  Non-Relationality of knowledge (cf. Chapters 3.10, 5.2.2, and KTE 1–4). For ‘world’, on a Kantian reading is but a term of reflection (Reflexionsbegriff) and not a sortal among other sortals. Rather, it is the presuppositional ground we need to meaningfully perceive objects ‘in it’, make inference involving them and act with regard to them. The dissatisfaction has a further ground in the fact that Sellars holds on to the view that there is an ‘in itself’ that can be taken to be responsible for the conceptual representations we have (Chapters 2.2.2-3.2). The tension rises because he denies our representations the kind of aboutness one would naturally expect, given this causal commitment to our acts of perceiving the world being the results of being affected by it. But we also saw that the distance grew between a transcendentally posited object and our full-fledged acts of intuiting. Here we can add that Sellars’ commitment to the transcendental ideality of the image-model world (cf. Chapter 3.5) added a great deal of distortion to the picture. If we are in large part responsible for furnishing the objects with concealed qualities, spatial structures, their flip-sides, and dispositions, how can we even claim to have knowledge of independent objects? While this may serve as a reminder about conclusions Sellars derives from his engagement with Kant’s theory of experience (cf. Chapters 2 and 3), this is only the background motivating the conception of intentionality he begins developing in IM and EPM and elaborates in SRLG, SM III, LTC, NI, and MFC. We could also read Sellars as aiming to establish independent arguments for the non-relationality thesis, independent from the results of his Kant exegesis. But even if we assume for a second that the objects of common sense are not only empirically but also transcendentally real, Sellars argues, the meaning of the words we use to characterise our practical and epistemic engagement with them does not spring from these objects directly. This point hangs on the underlying functional conception of what conceptually articulated contents are and what it takes for them to possess logical powers (a point we developed in Chapter 2.1 and pick up again in Chapter 5). The non-relationality thesis then takes the following shape: Our words do not immediately relate us to the material world qua meaning. And contrary to appearance, their vertical referential powers are really nothing over and above their horizontal inferential potential to enable language-­ exit moves, i.e. actions, and to evoke further meanings in us. It is a process involving language-entry moves, the formation of image-models, and the act of taking an object of reference to be in front of one (cf. Chapter 2.3). This gives rise to the remaining sense of dissatisfaction we are left with in SM III – vertical access to the world seems well lost. If this were the outlook Sellars were to leave us with, a further consequence would be that we could never investigate or prove the representational correctness of our scientific endeavours. For the non-relationality thesis also implies a strong internalism, in the sense that the (meta-)semantic

Non-Relationality 115 rules we have for assessing meaning statements and reference claims are meaningful only inside our conceptual system, inside the current use of language we live by (cf. Chapter 5.2). And if there is no meaningful place outside that conceptual system, we are bound to articulate our conceptions of what the world is like, beyond our concepts and terms, or as the apparent content of our conceptual engagements, always from within (cf. Chapter 6.3).30 Over the years, Sellars has worked out interrelated details of the non-relationality thesis. SM III provides a metalinguistic analysis of the logic of meaning-statements and of how technical terms such as ‘aboutness’ and ‘meaning’ function. By contrast, NI and NAO IV, SM IV analyse ‘denote’, ‘refers’ and ‘is true’. 31 These analyses have a feature in common. Sellars’ overall account rejects a direct realist reading of the referential dimension of these terms. To be precise, Sellars’ use of ‘referential’ picks out what he calls the dimension of linguistic representation. And while he thinks that we can successfully refer to linguistic representations of, say, ‘Mont Blanc’ as a term for a pen, or a mountain, their ‘of-ness’ is analysed in non-relational terms. To be able to explicitly address meanings and references of linguistic signs as such, their relationship to the ‘world’, i.e. ‘mountains’ and ‘pens’ does not matter. This does not prevent this relationship, for example, as the correctness of statements qua ‘images’ – isomorphic relations between non-­semantic objects – from playing a role in the definition of the correctness of certain statements (namely precisely the factual statements). He would grant that the mountain or pen, an object in the real can be involved in language entry- and language-exit moves, and as such they do, of course, exist. But Sellars’ theory for explaining the very operation of grasping the meaning of their linguistic representations highlights their non-­relational dimension. And with this thesis, one can be dissatisfied. All referential terminology that commonly is taken to articulate a word-world relation – terms such as denotes, refers, is true of – in Sellars’ hands become terms whose real function is argued to be that of licensing horizontal, non-relational, functionally classificatory, i.e. truth moves.32 To say that a statement is true, on the non-relational reading, is analysed into the act of authorising a ‘truth move’. The statement contained, accordingly, is authorised by the qualification ‘is true’, as a statement to be asserted. Thus, the terminology for which Sellars develops the non-relational reading is translated with regard to its functional dimensions as terminology which explains how ‘means’-statements convey functional roles of claimable contents. So what we are left with at the end of SM III is an understanding according to which we are always inside our conceptual scheme and the aboutness of our mental states is only a matter of functionally qualifiable contents, never of matter-of-factual qualities and objects. This goes for claims that are overtly meta-linguistic, as when we talk about language

116  Non-Relationality use, and for claims that Sellars argues, we should analyse as covertly meta-linguistic, as in the cases of statements of law-like empirical regularities, mathematical provability, moral, and practical reasoning. 33 These considerations explain the dissatisfaction stemming from a realist intuition, the expectation for a theory of intentionality. But there is also another more recent expectation that gives further cause for feeling uneasy. In the initial characterisation of SM III’s aim, Sellars announced a naturalistic account of the mind’s place in nature, of how the non-­ relational character of ‘meaning’ and ‘aboutness’, provides ›key to a correct understanding of the place of mind in nature‹ (SM ix). And in a related passage, he states: The position which I shall ultimately delineate and defend […] can […] be characterized as […] a naturalistic interpretation of the intentionality of conceptual acts. (LTC 1) How are we to understand this? Is the internalism meant to be an idealism? Is what Sellars calls nature just nature so-called? If nature does not show itself, since we are never directly, vertically related to it, how can the non-relational character help us see the right place of consciousness in it? Up to this point there is, in fact, very little ‘naturalistic’ about this conception of intentionality. And although Sellars allows for a causal dimension in our receptivity towards impingements from outside, this only serves as a placeholder for a more encompassing treatment of this vertical dimension and the elements causally active in it. To address these dissatisfactions, we need an understanding of the concept of nature from a transcendental point of view. This would be a standpoint that succeeds in making sense of the idea that our acts of classifying (considered as tokenings) can be part of nature, and that also succeeds in specifying it from within (cf. Chapters 6.2.3 and 6.3). Which is to say, Sellars’ solution to these dissatisfactions needs to clarify how the horizontal dimension in his non-relational account of aboutness can be reconciled with a ‘vertical’ dimension of relatedness between tokenings of acts of classifying and the causal source they originate from. This is a task to which we turn in the next chapter.

Notes 1. Cf. McDowell (1994, 67 ff.) and the criticism this position receives from his therapeutic perspective. 2. Here one can distinguish the question (1) whether claims about intentionality are relational claims from the question (2) whether empirical statements are to be explained in terms of their intentional or semantic relation to empirical facts. The strategy followed in this chapter is to show that Sellars’ answer to the first question gives the framework for understanding

Non-Relationality 117















the nature of the relation at issue in the second question. In other words, Sellars analyses meaning-ascriptions as non-relational and generalises the results of this analysis to explain the aboutness of empirical claims. 3. By such phenomena I mean the ‘things’ Sellars mentions in PSIM §1 among which he counts not only ›‘cabbages and kings’, but numbers and duties, possibilities and finger snaps, aesthetic experience and death‹. 4. From this it is evident that Sellars rejects Quine’s conception of experience and of ›[s]cientific method […] as a matter of being guided by sensory stimuli‹ (Quine 1960, 22). See Rosenberg (2007, 33–46) for a contrastive discussion of the comparative merit of arguments advanced by Quine and Sellars. 5. This is what sets human conceptually shaped cognition aside from animal perception. For a discussion of the relationship between (a) primitive representational systems that allow animals to navigate their surroundings and (b) human conscious map-making see Rosenberg (2007, 108–113) and Millikan’s (2012, 276–280) reply to objections from deVries (2012a). 6. Cf. Brandom’s study guide to EPM. And also cf. I §9 where Sellars discusses the related but more general point he finds in Kant that conceptual items embodying thought can only be given a functional characterisation. Which implies that their having an epistemic significance at all rests on their feature of playing a functional role. 7. Cf. SM III §37–41 and cf. also SM IV §61 for the explicit statement of the semantical uniformities requisite for linguistic behaviour. Cf. current research on the use of ‘emotion category labels’ e.g. in Barrett (2006); Azari et al. (2020). 8. Cf. SM IV §9 where Sellars recapitulates SM III as defending the ›hypothesis that the relation of representings to their contents is to be construed on the model of the relation of linguistic expressions to their senses, thus (representing) of (content) on the model of (expression) stands for (sense) […] followed by an analysis according to which to say what an expression stands for is to classify it‹. That is, to someone who would object in claiming that, say, stones are available to people who don’t functionally classify them, Sellars would respond what is functionally classified are linguistic or mental items, not stones. But since being conscious of stones is a matter of functionally classifying a mental content as ‘of’ or ‘about’ a stone, the objection holds no water. Being conscious of a stone is a matter of having a mental act that would be correctly functionally classified this way. But one needn’t actually be classifying one’s mental acts in order to be thinking about a stone. That would lead to a regress Sellars wants to avoid. Thinking about a stone doesn’t require thinking about one’s act of thinking about a stone. I am indebted to Lionel Shapiro for helpful comments on this. 9. Cf. NAO IV §64 where Sellars states this thesis: ›According to this analysis, meaning is not a relation for the very simple reason that ‘means’ is a specialized form of the copula. Again, the meaning of an expression is its ‘use’ (in the sense of function) in that to say what an expression means is to classify it by means of an illustrating functional sortal‹. 10. As nothing depends on it, ‘term’ and ‘word’ are used interchangeably. 11. A deeper concern lies in the fact that the Kantian analysis of how intuitions are brought about by the in itself is relevant for Sellars’ conception of counterpart objects, and of how the in itself can be conceived of as guiding our concept formations from ‘outside’ of our current conceptual scheme. Chapter 5 deals with the challenge Sellars has to address here; by way of foreshadowing that discussion we can say: We need to be able to conceive

118  Non-Relationality













of intuitions as being occasioned by an object in nature (some thing in itself), even though that thing is never present to me directly (cf. the forms of direct reference distinguished in Chapter 3.7). On the assumption that our ability to picture objects adequately progresses, the thought becomes pertinent that our acts of intuiting in a future conceptual scheme will be more adequately related to their causal source (Chapter 5.4). Note how this is conceived of as a non-semantic isomorphic relation between two items in the real, i.e. not as a more and more fine grained representational, semantical relation between a mental state and its object (Chapter 5.3). 12. Sellars commentary about the logical structure of meaning statements should not be read to imply that a word cannot ever stand in a relation to an extra-linguistic item. Rather, it is exactly the job of the technical notion of a language-entry transition to qualify linguistic responses as standing in just such a relation. Sellars non-relationality thesis, thus, is a thesis about where conceptual items get their conceptual articulation from, not a thesis to the effect that there is absolutely no connection between items in our environment and our linguistic responses to them. 13. Cf. IM, SRLG, MFC and LTC for subsequent attempts at doing that in detail. 14. For the case where the object of interest is a cube, Sellars spells this out as follows: ›The hyphenated phrase ‘this-cube’ expresses a representing of something as a cube in a way which is conceptually prior to cube as a general or universal representation; that is, in a way which is conceptually prior to predication or judgment. The strength of the position lies in the fact that the individual represented in perception is never represented as a mere this, but always, to use the classical schema, a this-such‹ (SM I §15). 15. Cf. FMPP I §49: ›[…] we will be construing the concept of seeing as ab initio cognitive‹. But see the range of positions on this matter represented in Schulting (2016). 16. Cf. the discussion in Haag (2014) of the notion of ›guidance from outside‹ (of the realm of the conceptual) which Sellars wants to secure as central to the development of our conceptual encounter with the world (in SM I §39). Cf. also McDowell 1998b, and see Shapiro (2013) defending Sellars against the charge that the reason he demands a sideways-on view of language or thought is that he can’t make sense of intentional relations. Shapiro agrees with McDowell that Sellars demands a sideways-on- view, but, Shapiro argues, not for the reason McDowell cites. Shapiro’s Sellars can make sense of intentional relations. 17. Cf. SM I §7, §49, and §50: ›[S]ince the causal powers of an object are not themselves causes, the description of impressions tends to be purged of all expressions which logically imply causal properties: [e.g.] an impression of a pink ice cube tends, in view of the causal properties implied by ‘ice’, to become an impression of a pink cube.‹ 18. Still, one can be more or less optimistic about the implications this has for our grasp of the source of our cognitive responses to the world. Chapters 5.2.1–5.2.2 spell out the differences between a by and large Kantian reading and a Sellarsian reading of this. 19. How exactly to make sense of this idea of a guidance by the sheer manifold is a deeply contentious issue. Responses range from amiable defence, as in Sachs (2015) or Christias (2015), engaging reconstructions, as in Haag (2009, 2014, 2016), to outright hostility, as in McDowell (1988a and 1998b). McDowell (2013, 62) argues that we should rather take “the objects themselves to supply the external constraint to our conceptual goings-on”. I will pick up the question in what sense Sellars can be read

Non-Relationality 119













as holding that objects are responsible for our perceptions of them when I come to discuss the role of counterpart objects to the objects as we conceive of them in the present conceptual structure or Manifest Image. 20. In NAO V §63 Sellars puts this point the following way: ›That languagings are evoked (in contexts) by happenings of certain kinds is a causal fact which is nevertheless essential to their conceptual character. This causal aspect of perceptual takings, introspective consciousness, inferences, and volitions accounts for the selecting of one world story rather than another and connects the ‘is’ of this selecting with the rule-governed or ‘ought to be’ character of the language. The ‘presence’ of this unique story at each stage in the development of the language makes possible the referential framework of names, descriptions and demonstratives and, by so doing, makes possible the exploratory activity which lead to the story’s enrichment and revision‹. From this we see that Sellars does not want to give up entirely on the notion of direct reference. The pressing question therefor becomes what sense of ‘directness’ he can retain. Sellars argues that the referential framework alone doesn’t suffice to generate the holistic-­f unctional character of linguistic expressions. The causal aspect, i.e. the vertical fact that in important regards we are passive in our being affected by something that is other than and prior to how we conceptualise it in our current conceptual system, plays its role in the gradual revision of our world-story. But this is a point concerning the conceptual change of our conceptual system as a whole – and how to understand the transcendental presupposition that something must underly that change – and not the genesis of meaning of a single expression, nor the logic of meaning-statements. 21. This point will surface again, when in Chapter 5, we discuss the succession of conceptual frameworks (and what this implies for the status of the objects we intuit in our current, and successor conceptual schemes). For there we will have to show how Sellars can give an account of the status of counterpart objects as they are conceived (i.e. perceived and conceptually classified) in a future conceptual scheme. As we will aim to show there, Sellars is committed to arguing that one and the same in itself is responsible for our current and for future object perceptions – even though our conceptualisations of what these affections mean will undergo severe changes. 22. In a way we have already encountered a version of this interpretive solution when we looked (Chapter 2.3.3) at Sellars’ claim that sense-impressions are radically non-conceptual and unconscious (vertical impingements) while intuitions are conceptually loaded responses to individuals (this-suches) that are always already taken in (i.e. horizontally) as of a certain kind. 23. The functional role given in dot-quotes, i.e. • triangular •, can for instance be realised in Euclidian geometry or Riemannian vector spaces, in a children’s drawing and in a declarative utterance about a pyramid in front of one. As different as these cases may seem, they can be grouped on account of their realising one and the same functional role in different, specified, ways. 24. In MGEC §15 Sellars affirms this causal dimension, arguing that an encompassing ›formulation of a theory of intentionality‹ requires doing ›justice to both the logical and the causal dimensions of discourse about mental acts‹. 25. This account is given in detail in MFC and LTC, and reconstructed in deVries (2005, Chapter 2); O’Shea (2007, Chapters 3–4), Rosenberg (2007, Chapters 6–7), and in Haag (2007, Chapter 9.7).

120  Non-Relationality 26. Cf. NAO V §62: […] A language, in its primary mode of being, simply is the pattern of beliefs, inferences and intentions of whatever logical form or conceptual level which linguistic behavior (in the full-blooded sense) embodies. It is the instrumentalist conception of language which misconstrues it as a medium which is neutral as between the alternatives which are ‘expressible’ in it‹. 27. Cf. Lionel Shapiro’s reading of Sellars, who interprets Sellars as arguing that our intentional relatedness is implicit in the (non-)inferential moves constitutive of our rule-governed practice of asserting matter-of-factual claims. As he puts it: ›According to Sellars, we do not use semantic vocabulary to describe language-world relations. Rather, our taking language to relate to the world is implicit in the moves (inferential or non-inferential) licensed by our semantic assertions‹ (Shapiro 2014b, 792, 802). What stands out from Shapiro’s formulation is this: No matter what we in turn interpret the significance of a claim to amount to (from a normative perspective on language), the act of claiming relates the content to the world by virtue of being realised in the inferential move. This matter-of-factual realisation of a claiming, as uniform performance will be of crucial importance (cf. Chapter 5.1). 28. But see Shapiro (2014b), who traces the changes in the framework Sellars develops to account for the function of semantic vocabulary. Among other things Shapiro defends an interdependence thesis according to which the roles of truth-talk and meaning-talk are intertwined and cannot be separated. Thus, Shapiro argues Sellars does not think we could have meaning-talk without truth-talk, even if there are cases where the former is applicable and the latter isn’t. Insofar as Shapiro is interested in the ‘meaning of meaning attribution’ (Shapiro 2014b, 797), one can take him to ask a meta meta-semantic question in the sense of Burgess & Sherman (2014, 1–16). In contrast to this, our focus lies on the question what understanding of the role of Sellars’ account of meaning and truth talk contributes to explaining the concept of intentionality at different levels of abstraction (cf. Chapters 1.3 and 6.3). On our reading, the notion of an intentional relation is more encompassing, than is cashed out by the analysis of the non-relational nature of meaning and truth-ascriptions. One reason is that one can be intentionally related to contents about which questions of reference, existence, and truth are uninteresting or subordinate, as is the case, e.g. with fictional entities. 29. One can read Robert Brandom’s work since the late 1980s to be inspired by this. He takes up this thought in 1994 and subsequent writings (2001, 2002, 2008, 2015) and unfolds it out into a theory of non-relational intentionality. 30. Maher’s (2012) The Pittsburgh School of Philosophy in building on McDowell (1994) and Brandom (1994) can be read as a reconstruction of the joint project which aims to show us how we can be comfortably and worry-free at home inside our (current) conceptual scheme, independently of such a standpoint outside. Cf. also Redding (2020) tracing the lineage of this thought to Hegel. 31. For a discussion of the details, similarities and differences among these analysis see Shapiro (2014b) and deVries (2005, Chapters 2 and 4), O’Shea (2007, Chapters 3–5). 32. Cf. The inclusive summary Sellars gives in TTC 45. 33. In short, for all cases that Price (1997) subsumes under the heading ‘M-Worlds’. The delicacy of this point is that already here we can see that Sellars does not make a distinction between a merely internal kind of representation and an external, environment tracking kind of representation, which Huw Price wants to read into CDCM and ITM (cf. Chapter 6.2).

5

Transcendental Phenomenalism

This chapter defends a reading of Sellars according to which his position amounts to a transcendental phenomenalism.1 This is a position about our epistemological relation to the objects we encounter in the conceptual structure we live by. It is not a thesis about the nature of these objects. It is a thesis about the transcendental ideality of the objects of common sense, a thesis that is designed to be compatible with the results of transcendental methodology (Chapters 1.1–1.3). This entails that the negative results about what these objects are not can be supplemented by a positive account of how else we can conceive of them. A subsequent outlook discusses this positive sense and argues that for Sellars questions about their ontological home remain a matter of our pragmatics. To that end, three arguments from Sellars’ writings are reconstructed to explicate the cornerstones of his overall position. Each of these arguments serves to show how essential the concept of a linguistic picture is to his account of intentionality. 2 Sellars takes the conception of ‘picturing’ from Wittgenstein’s Tractatus logico-philosophicus (1921), however, not without altering it in important regards (SM ix, V §8, 26, 56, 61). Where Wittgenstein defended a theory of representation in terms of a ‘picture theory of the proposition’, Sellars rejects its central claim that it is facts that picture facts.3 According to Sellars, it is objects that picture objects (NAO V §92).4 More precisely, ‘picturing’ is a feature of basic matter-of-factual statements when considered as tokens of utterances or inscriptions. The definitive claim is that such statements can themselves be viewed as ‘natural-­linguistic-objects’ which, on the basis of a complex method of projection, have a picturing capacity, or, more tacitly put, as standing in a special relation to the objects by which they are brought about. 5 The aim of this chapter is also to reconstruct the argumentative resources Sellars thinks he has for holding that such a relation underlies all conceptual structures.6 This includes clarifying why he mobilises the concept of picturing to resolve the difficulties and dissatisfaction we retain from his account of his non-relational account of intentionality (Chapter 4.4). And it entails demonstrating that this concept bears DOI: 10.4324/9781003221364-6

122  Transcendental Phenomenalism a systematic relationship to the instability thesis about nature of the objects of the Manifest Image (Chapter 3.8). It is the argumentative weight of this thesis that forces Sellars to introduce the conception of picturing. In light of these considerations, we will unfold the significance of the concept of picturing in Sellars’ overall account of intentionality by reconstructing three arguments (Chapters 5.1–5.3). In a subsequent chapter, the presentation is rounded up by an extended discussion of two objections, the objection from pragmatism (Chapter 6.1) and the objection from global expressivism (Chapter 6.2). Transcendental phenomenalism then emerges as a thesis not just about the fundamentals of our epistemic situation. It also amounts to a thesis about the status of the framework as a whole, in terms of which we are to understand these fundamental elements as elements of the concept of intentionality. The aim pursued here is to show that it is by way of reflexively demonstrating the necessary components of this concept that Sellars can give an account of our intentional relation to an evolving life-world of which we, as perceivers, speakers, and agents, as well as our sensory-perceptual, linguistic and intentional acts, are a part (Chapter 6.3). Chapter 4 reconstructed the non-relational character of mental states and conceptually articulated intentionality. This involved discussing the thesis that sensory states are unconscious responses evoked in perceivers that bear a causal relation to the domain of transcendentally postulated things in themselves. However, it was shown that the realist expectation of Sellars’ conception of intentionality remained unfulfilled; it is still on the agenda. This expectation is further reinforced by the fact that Sellars’ commitment to a version of realism has always been central to his philosophical project.7 This chapter deals with the question: What does Sellars’ realism consist in and how does it address the desiderata that emerged in his non-relational account of intentionality? This mandates discussing if Sellars’ conception of realism can accommodate the demand for a ‘vertical’, or relational account of intentionality.8 To unfold in what way Sellars picks up this demand, we need to trace how he brings the results from the reflections on the dependency claim – the thesis about the relation between logical and empirical form (cf. Chapter 2.3) – to bear on his conception of meaning and linguistic activity.

5.1  The Argument from Double Nature The following interpretation is a reconstruction of three arguments supporting the claim that Sellars is committed to a transcendental phenomenalism. The aim of the first argument reconstructed in this section, the argument from double nature, is to establish that the non-relationality thesis is compatible with a larger concept of nature, according to which the linguistic performances (languagings) constitutive of language-­games are themselves qualifiable in more than logical,

Transcendental Phenomenalism 123 semantic or normative regards. The main thesis is that utterances or languagings have a double nature, a Janus-face, that comes out when one abstracts from their normative features and looks at their causal underside. (1) Acts of perceiving find overt expression in linguistic utterances which belong in the larger context of semantic and epistemic norms (i.e. a method of projection; NAO V §85, 114, SM V §10, 56). (2) On account of epistemic, semantic, and behavioural norms, we individuate objects in the environment epistemically relevant to us (NAO V §69) and respond to them by producing languagings. (3) As principles are reflected in uniformities of responses, linguistic utterances are systematically produced in uniformities of responses to objects impinging on us (TC 216/NAO V §95). (4) By being reflected in uniform performances, every languagings not only has a logical, semantic and syntactic form but as an utterance, also an empirical form (NAO V §64). (5) The empirical form of objects allows them to be the causal ground of affections in us, and of our norm-governed linguistic responses (languagings) to them (NAO V §63). (6) The empirical form of languagings stands in a determinate relation to the empirical form of those objects by which they are (transcendentally taken to be) evoked (NAO V §59, 63–66). (7) The determinate relation is a second-order isomorphism in accordance with the method of projection entailed in the norms of the language, not a first-order ‘similarity’ (BBK §41, 52, NAO V §105, SM V §58–59). (C) The double nature of languagings means they can viewed semantically, and then they are about their contents (1–2). But when they are considered from an engineering standpoint (3–4), they are systematically isomorphic (5–6) to the objects by which they are brought about (7). The isomorphism is at once a result of linguistic activity and its transcendental precondition. To get the right angle on how Sellars takes this isomorphism to support his version of phenomenalism, it is important to keep in mind that it is introduced as a theoretical notion on the basis of an abstraction: It is a matter of abstracting from the logical form, not a matter of reducing it to the empirical form. Sellars considers his non-relational account of intentionality to be central to the correct account of the place of consciousness in nature. However, as we have seen so far, the explication of the horizontal character of the functional classificatory activity has not given away anything about its place in nature. So how are we to understand this claim? And why does Sellars claim that

124  Transcendental Phenomenalism non-relationality provides the decisive clue? To set the stage for the argument Sellars gives, we turn to a key passage in his essay on Kant’s theory of experience (KTE 22–27) in which he elaborates his reading of what Kantian acts of judgment are and in what relations they stand to their contents. The passage prepares the ground for the argument that notwithstanding its non-relational character conceptual activity is part of nature: The logical form of a stating is clearly not the empirical configuration of the sentence it illustrates, though having an appropriate empirical configuration is a necessary condition of the stating’s having the logical form it does, in the language to which it belongs. (KTE 22) A matter-of-factual statement, e.g. the sentence that this pyramid over there is red, has a logical form, which Sellars denies its relational connection with the extra-linguistic object ‘red pyramid’. However, once the sentence comes into view as itself an object in nature, this passage makes a claim about two relata and the special relationship between them. The relata come into view when we distinguish two ways of viewing judgements and statements in which they are expressed. The first relatum consists in the familiar thought that judgements have a special logical form, one that distinguishes them from other mental states such as ‘believing that’, perceiving or sensing, intuiting, and being unconsciously affected (cf. Chapters 2.1 and 2.2.3). So when Sellars, in this context, speaks of ‘form’ he has the conceptual structure in mind which we can appeal to in characterising the judgment as of a special type. But what, we may ask, is the significance of the additional information that a judgment’s logical form ›is clearly not the empirical configuration‹ of the sentence to which it stands in an illustration relation? What is such an empirical configuration in the first place? We can understand this disjunction by focusing on sentences. My subject-predicate judgment, considered as a vocal sequence, has an empirical configuration when I record my uttering the sentence ‘Tom is tall’ and have it pressed on vinyl. It also possesses these configurations as a mere utterance, for instance, by way of making the air quiver. The example of grooves on the record serves to elucidate what it takes to manifest an empirical configuration. As plastic grooves, they are just one material illustration of my overt act of judging. There may be countless others; one may write it down by hand, or use Morse code, etc. What matters is that there is some relationship between judging as a conceptual episode and its empirical material realisation. In the way the example is construed, it seems that the configuration of the vinyl depends on my prior act of stating the judgement into a microphone.

Transcendental Phenomenalism 125 But that is not how Sellars has it. He inverts the order of dependence and claims that having an appropriate empirical configuration is a necessary condition of the stating’s having the logical form it does, in the language to which it belongs. (KTE 22) The first conclusion we can draw from this train of thought is that utterances, statings, or languagings have a ’double nature‘.9 Or, as one can put it, they have a causal underside.10 Since Sellars is clearly not arguing that the vinyl exists prior to my utterance, what does he have in mind by arguing that the empirical form is a necessary condition of the logical form? One way to approach this riddle is to consider how chess figures as material objects embody functional roles. If we construe a spoken language as an interrelated system of utterances (cf. NAO V §62), we can throw light on the sense in which the vocalisation stating ›Tom is tall‹ is a necessary condition for the logical form of the judgment it realises (in the sense of ‘making it actual’). The functional or conceptual role of interest to us when we investigate the meaning or aboutness of a mental state does not reside in the void; it has to have some realisation. The fact that conceptual ‘pieces’ or ‘role-players’ must have determinate factual character, even though we don’t know what that character is, save in the most general way, is the hidden strength of the view that identifies mental acts with neurophysiological episodes. (NI §25) In this sense, the empirical form, allowing a function to be realised or reflected in a uniformity of behaviour, sustains the logical form. And this line of thought helps to address the question how Sellars can fuse the horizontal dimension of interrelated linguistic functions that govern and inform linguistic activity with the vertical dimension. The solution lies in connecting the claim that logical form needs to be functionally realised in some way or another to the claim that the logical form of linguistic statements is embodied in the materially rich linguistic activities of language users. Given Sellars’ view of functional roles, we can cash out the functional role of a term across languages as follows: Recall how ‘rouge’, ‘rot’ and ‘rosso’ are all cases of the functional role expressed in the English language by the word red, which is captured by saying they all embody the function • rot • in their respective languages (cf. NAO IV, SM III-IV). Each language realises the functional role we can meta-linguistically characterise as the • rot • in its respective uniformity or pattern. In Sellars’ characterisation to possess knowledge of how to use empirical predicates

126  Transcendental Phenomenalism is to espouse or endorse linguistic rules or principles.11 Such a practical endorsement, implicit in one’s use being characterisable in terms of such rules, is not a matter of conscious approval; it is a matter of growing up into a community of language users. And as a user of a language one’s performances are always already shaped by the linguistic norms which sustain the interconnectedness of perception, inference, and action in the community. It is against the backdrop consideration as to how the use of a language rests on an implicit endorsement of its constitutive principles that Sellars formulates what O’Shea (2007) has termed the norm-nature meta-principle: Espousal of principles is reflected in uniformities of performance. (TC 216) The uniformities of linguistic behaviour make up the base-level, object-­ language performances conveying the meaning, e.g. that something is red. Accordingly Sellars writes: It should be borne in mind that linguistic episodes have not only logical powers but also, and necessarily, matter-of-factual characteristics, e.g., shape, size, color, internal structure, and that they exhibit empirical uniformities both among themselves and in relation to the environment in which they occur. (KTE §27) So we can look at an act of judging from within by considering it only with regard to its abstract qualification of having a specific logical form. But in order to get clear about the sense in which a judgings’ logical form is part of the world it is a judgment about, we have to take away that abstraction and investigate how such a logical form needs to be realised in any given language to have the significance it does in that very language. And then we see that acts of stating, qua being acts persons carry out in an interrelated system of linguistic behaviour, always have determinate empirical configurations or empirical form. With this, we have the first shot at an explication of why Sellars makes this peculiar claim about the correlations between the two ways of characterising the forms of a judging. And therein, we see how the argument from the double nature of languagings lends support to Sellars’ transcendental attitude towards nature. It establishes that thoughts and utterances are part of nature and cashes in the promissory note that the non-relational character of intentionality provides the key to the correct place of mind in nature (SM ix). The reason we can see them as part of nature lies in the fact that although we normally specify them horizontally with regard to their functions and thus non-relationally, we can also see how

Transcendental Phenomenalism 127 thoughts find their expression in linguistic episodes or languagings. And thus, the argument goes, to exist, thoughts need to be realised in uniform linguistic performances, in patterns governed by epistemic norms which exhibit empirical form – independently of how we characterise them. If we grant Sellars this much, we see that the argument from double nature helps us understand how the non-relational intentionality of thought and conceptual activity can be accorded a place in nature in its overt realisation as a languaging. Can this double-nature treatment of intentionality be reconciled with Sellars’ commitment to a realism in his transcendental conception of nature? Speaking in the metaphor from the previous chapter, we can see the worry returning that this argument alone does not afford us intentional reference in a vertical sense. We have not yet accommodated the thesis that our thoughts and utterances about the world are matter-of-factual. And that they need to be matter-­ of-factual is a requirement Sellars explicitly endorses: An essential requirement of the transmission of a language from generation to generation is that its mature users be able to identify both extra-linguistic items and the utterances that are correct responses to them. This mobilizes the familiar fact, stressed in [KTE 27], that, in addition to their logical powers, linguistic expressions have an empirical character as items in the world. We can ascertain, for example, that a person does in point of fact respond as he ought to red objects in sunlight by uttering or being disposed to utter ‘this is red’. Again, we can ascertain that, other things being equal, he is not disposed to enlarge, as he ought not, utterances of ‘it is raining’ into ‘it is raining and it is not raining’. (KTE §42) In light of this passage, it should only be natural that a reader of Sellars becomes even more dissatisfied with the current state of his account of intentional reference. We need to be able to identify ‘extra-linguistic items’ to evaluate the conceptual responses of our fellow speakers and perceivers. But if our conceptual responses are only meaningful horizontally, there is no such activity as directly identifying an item in nature, given the restrictions placed on the notion of ‘nature’ laid out above (Chapter 3.10). It is this sense in which the rejection of the Myth of the Given in EPM is still with us. We can summarise the positive result of the double nature argument as follows. We seem to have found a way of placing conceptual activity qua its empirical form in nature without having to say what this nature is, ›save in the most general way‹, i.e. an item ›necessarily‹ has to have ›matter-of-factual characteristics, e.g. shape, size, colour, internal structure‹ (NI §25). So despite the double nature of judgings, possessing

128  Transcendental Phenomenalism empirical and logical form, we face a subsequent question. A question which in a transformed sense presents the demand for an account of our intentional reference in a vertical sense: What is the relation of empirical acts of judging to matter-of-factual states of affairs?

5.2  The Argument from Immanence The challenge we carry over from the argument of double nature is this: According to the double nature argument, we can see languagings as necessarily having an empirical form, but we are in no position to establish what exactly this form is. To hold that one could assume a position outside the series of our conceptual episodes in order to determine such a form would be equivalent to claiming that one could avail oneself to a view from sideways on, i.e. from outside our conceptual structure. But that would be an inconsistent position, a fact of which Sellars is well aware (SM V §48). But why is it so difficult to attain proper empirical knowledge about the causal underside of our languagings? To answer this question, this chapter follows a line of thinking that connects Sellars’ non-relationality analysis of intentionality and Kant’s analysis of our knowledge of objects of experience. It will be called the argument from immanence or the immanence criterion because supports another cluster of core commitments at the heart of Sellars’ transcendental realism in his conception of nature: That our epistemic engagements with the world (1) have to be and in fact (2) can be part of the world of which they purport to yield knowledge. The difficulty for this argument lies in the fact that our epistemic access is barred from direct intentional reference. In other words, this argument commits us to rejecting direct, i.e. ‘vertical’, intentional reference. However, as will come out, this does not lead to the vertical dimension being excised altogether. Rather, the argument from immanence is required to put this dimension in its proper place within an account of intentionality. And in Chapters 2 and 3, we learned that Sellars agrees with Kant’s analysis of nature as the content of a system of representables (KTE §1–5). Accordingly, the context of the argument from immanence is given by the related question of how the horizontal dimension of inferential significance can be squared with the vertical dimension of our passivity, i.e. the receptivity of our senses. We might add that verticality is not always passive, e.g. we name things, we look for things, we build things. To address the question of the relation of our empirical judgements to matter-of-factual states of affairs appropriately, requires detailing restrictions binding us to an immanent (meta-)perspective on the intentionality of our claims to knowledge. The argument from immanence then has to demonstrate that the elements are part of the world which our intentionality purports

Transcendental Phenomenalism 129 to be about. An essential aspect of this argument is to demonstrate this without confining the conception of intentional reference to Kantian appearances or relapsing into the Myth of the epistemically Given. If we assemble the passages in which Sellars implicitly argues for an immanent conception of intentionality and try to give this argument an explicit reconstruction, it takes the following shape: (1) Receptivity entails that we respond to impingements from outside with modification in our sensory and bodily states. (2) In acts of perceiving, we systematically mistake the features of our sensory and bodily states to be features of the items (objects) affecting us. (3) Acts of perceiving (takings or acts of perceiving) are governed by (epistemic) semantic assertibility norms. (4) Acts of perceiving, when considered with regard to their double nature, have an empirical form which is shaped by uniform patterns of (linguistic) performances (e.g. acts of perceiving or uttering). (5) In virtue of being reflected in uniformities of performances, acts of perceiving and their overt expression in linguistic utterances form natural-linguistic-objects. (C) How we conceptualise nature is reflected in an isomorphism within nature. The determinate isomorphism between natural-­linguisticobjects and objects co-varies with the S-assertibility norms governing the conceptual structure (method of projection). In what follows, we will discuss if Sellars succeeds in showing that his transcendentally motivated phenomenalism cashes in the requirement on transcendental linguistics: that the delineation of the most fundamental features amounts to an account that underwrites that our claims to knowledge can be formulated from within, from an immanent position.12 The argument from immanence integrates intentionality in a nature as conceived of by a transcendental phenomenalism with an underlying realism, by applying the results from the argument of double nature to acts of judgements. It works as a specification of the double nature argument. It explicates the role of matter-of-factual judgements by showing that there is a (double nature) way of looking at them such that they can be understood as being part of the world of states of affairs they aim to be judgements about, and that they can do that, even though we read such judgements as meaningful judgements only from the internal perspective familiar, e.g. from an intentional stance. We have already learned how central the concept of judging is when interpreting Sellars’ reconstruction of Kantian categories, analysed as meta-classificatory acts of judging (Chapter 2). Now our task is more

130  Transcendental Phenomenalism specific as we aim to understand the overall role of empirical judgements, judgings that something matter-of-factual is the case. In empirical judgements we articulate our conception of items in nature. Hence these are judgements that interest us the most with regard to both their logical and their empirical form.13 The opening passages of KTE underline just how important an appropriate grasp of the status of the concept of an empirical judging is for the task of formulating a theory of intentional reference: [It] must aim at clarifying the concepts of an empirical judging, of truth, of a state of affairs, and of what it is for a state of affairs to obtain or be actual. […] The central theme of the Analytic is that unless one is clear about what it is to judge, one is doomed to remain in the labyrinth of traditional metaphysics. On the other hand, to be clear about what it is to judge is to have Ariadne’s thread in one’s hand. (KTE §6–7) Sellars follows this up with a remark about the relation between the above-mentioned pair of concepts mentioned, which sets the stage for how we are to understand the thesis that our knowledge of the objective world can be immanent knowledge: Now from the Kantian point of view, the above concepts pair up in an interesting way: judging with state of affairs, and truth with actuality. Indeed to say that they pair up is to understate the closeness of their relationships. For, Kant argues, in effect, that the pairs turn out, on close examination, to be identities. (KTE §8) This passage contains two difficulties: What licenses one to take the pairs to be identities? And why should one consider the status of the concept of an act of judging and the concept of a state of affairs to be identical? This requires taking into account what structure Sellars thinks a framework needs to have such that it allows us to articulate and critically reflect on the ontological commitments expressible by these concepts. In as much as the Kantian categories were analysed to be meta-classificatory concepts of functional roles in the conceptual activity, so are of concern object-level concepts of something being matter-of-factual as central to thinking the concept of a world of objective contents. The bare essentials of an ontological framework to which any theory of knowledge, in line with the argument from immanence, has to conform are as follows:: Roughly, the form of empirical knowledge is: an I thinking (however schematically) the thought of a temporal system of states of affairs to which any actual state of affairs belongs.14 (I §9)

Transcendental Phenomenalism 131 And, as we can make explicit, such a system would need to also include the state of affairs of thinking this very thought. This general characterisation of the form of empirical knowledge (cf. Chapter 1.3) was chosen above as the starting point because downhill from here, there are two ways of looking at immanence, a Kantian and a Sellarsian way. The following discussion aims at disentangling them, thereby showing to what degree Sellars follows Kant’s analysis of empirical judgings, and where exactly he criticises and departs from Kant. This chapter takes the form of contrasting Sellars’ reading of Kantian immanence with Sellars’ own version. We need to look at both to see the contrast between strong and moderate immanence. This latter form of immanence is required for a proper understanding of the double nature argument and, on that basis, of the transcendental function of the concept of picturing. (5.4) It is with these distinctions at hand that we can return to the leading question of how Sellars can account for the vertical dimension of our intentional reference to a world of which we are a part. 5.2.1  Kantian Immanence Just how Sellars interprets Kant’s concept of something matter-of-factual, or more generally an empirical content, hinges on the question of whether Kant’s account of empirical content meets the aforementioned criterion: That our epistemic engagements with the world are part of the world of which they purport to yield knowledge. Here it matters to ask: What sense of ‘part of the world’ do we arrive at with Sellars’ Kant? A first clue is provided by a passage elaborating the relation between an empirical thought and its content: Kant, of course, grants that thought has ‘content’ as well as ‘form’ — but the content consists of concepts, for example empirical concepts — and these in their turn are ‘functions’. When we think of a shape, e.g. of a triangle, our thought, needless to say is not triangular — it contains the concept of a triangle, itself a rule or function by which the mind can generate representations of triangles (A141; B180). (I §21 fn8) In this passage, Sellars applies his meta-conceptual, non-relational reading of the intentionality of conceptual acts. He applies it in explicating how he takes Kant to view the status of the concept of an empirical content. Note how radical Kant’s position on this reading becomes: The contents of judgings about something matter-of-factual, a specific triangle, for instance, consist of concepts (conceived of as rules). Recall how we covered the qualification that concepts serve as rules for generating representations, e.g. of triangles above (Chapter 2.3). Sellars finds his non-relational reading of intentionality and content is already present

132  Transcendental Phenomenalism in Kant’s conception of how we are to analyse empirical representations. On his reading, Kant already rejects a relational conception of the aboutness of judgings: If judgings qua conceptual acts have ‘form’, they also have ‘content’. Of all the metaphors that philosophers have employed, this is one of the most dangerous, and few have used it without to some extent being taken in by it. The temptation is to think of the ‘content’ of an act as an entity that is ‘contained’ by it. But if the [logical] ‘form’ of a judging is the structure by virtue of which it is possessed of certain generic logical or epistemic powers, surely the content must be the character by virtue of which the act has specific modes of these generic logical or epistemic powers. (KTE §24) Sellars’ main premise is as follows: The conceptual content ‘state of affairs’ is ‘that of which a judgment may be true’. And such content is derivative of the judgment’s specific logical powers. We can read this as an affirmation of the horizontal nature of the aboutness of acts of judging. The form and content of an act of judging are a matter of a genus and species. The generic logical power of a judging consists in being a judging of a certain categorial type, say a subject-predicate judgment. The specific mode qualifying this genus gives the judging token its epistemic significance, in case e.g. the judging concerns an individual substance and ascribes to it a specific predicate. It is in the frame of this analysis of what judgings are that Sellars reads Kant as providing an answer to the question of how empirical judgings are taken to be about something matter-of-factual. In other words, we can read him as discussing the question of how to understand the ‘referential’ power of judgements. Against this backdrop, we can see why Sellars highlights the metaphorical nature of the characterisation ‘being the content of a judgment’. According to this Kantian analysis, the very idea of being a content is nothing over and above being a conceptual content with a specific logical significance or power. This non-relational reading of ‘judgment contents’ serves to explain the status ›being an actual state of affairs‹. Thus, a judging that Tom is tall would, in its generic character, be a judging of the subject-predicate form. It is a judging that a certain substance has a certain attribute. (These two ways of putting it are equivalent.) If we focus our attention on the predicate we can characterize the judging more specifically as a judging that a certain substance has the attribute tall. Thus, just as to say that a judging is a judging that a certain substance has a certain attribute is to say that the judging is of a certain generic kind (i.e., has certain generic

Transcendental Phenomenalism 133 logical powers); so to say that a judging is a judging that a certain substance is tall is to classify the judging as one of the such and such is tall kind, i.e., to classify it in a way that ascribes to it the more specific conceptual powers distinctive of the concept of being tall. Indeed, for the judging to “contain the concept of being tall” is nothing more nor less than for it to have these specific powers. (KTE §25) For Sellars, this Kantian analysis of the contents of judgings amounts to an argument from immanence. In judging something to be out there, one is still doing nothing over, and above making horizontal moves, i.e. one is functionally classifying contents as correctly or incorrectly asserted or judged. And one bases one’s endorsement or rejection of such contents on the logical or epistemic significance one takes, we all take, these contents to possess.15 Sellars accordingly writes: Kant correctly concludes from the above that there is no such thing as comparing a judging with an actual state of affairs and finding the judging to be ‘correct’ or ‘justified’. For, according to the above analysis, an ‘actual state of affairs’, since it has judgmental form, is simply a true species of judging, i.e., to use Peircean terminology, a judging-type that it would be (epistemically) correct to token.‹ [in the fn:] ›Put in linguistic terms, an ‘actual state of affairs’ is a true species of stating, i.e., a stating-type that it would be epistemically correct to token. (KTE §26, fn6) From which Sellars concludes: Thus ‘comparing a judging with a state of affairs’ could only be comparing a judging with another judging of the same specific kind, and this would no more be a verification than would checking one copy of today’s Times by reading another. (KTE §26) In the terminology we developed, this means that what Sellars finds in Kant is a strong version of immanence. In a related passage in SM II Sellars articulates what he takes this Kantian line of thinking to amount to: Kant remains […] a “phenomenalist” […]. Kant’s phenomenalism can be put […] by saying that physical objects and events exist only “in” certain actual and obtainable conceptual representings, the intuitive representings synthesized by the productive imagination in response to the impressions of sense. (SM II §46)

134  Transcendental Phenomenalism But here, it is important to pay attention to the dialectical treatment Sellars gives to his engagement with Kant’s text. Sellars takes Kant’s analysis quite seriously because it reflects an important insight about the non-relational character of intentionality when considered only from inside the perspective of a speaker. That is, Sellars endorses Kant’s argument about the status of our conceptual responses. But he wants to reject the idea that this is all we can say about the fundamental nature of our epistemic situation with regard to extra-linguistic items. If there is no such thing as comparing a judgment with an actual state of affairs, because all one’s conception of such a state really amounts to is just another judgeable content, then we do not have a determinate concept of how matters of fact are in themselves. This means, Kant’s analysis of our epistemological situation entails not only a phenomenalism but also an agnosticism about the ontological status of matter-of-factual items in nature. From the analytic perspective of the transcendental philosopher tracing how we can explicate a conception of our knowledge that meets the immanence criterion, Kant’s account invokes a constellation that implies a strong immanence, a conception of immanence that will turn out as too strong. We are so caught up in our concept of an object in nature that we never actually get to evaluate our concept applications from outside: Kant’s agnosticism, however, if taken seriously — i.e., construed as the view that we have no determinate concepts of how things are in themselves — means that no conceptual response can be evaluated, in the above manner, as correct or incorrect. Rules of the form (Ceteris paribus) one ought to respond to φ items with conceptual acts of kind C. could never be rules in accordance with which people criticize conceptual responses; for, on his official view, the esse of any item to which any empirical predicate applies is already to be a conceptual response, not something that is responded to. To put it bluntly, only God could envisage the ought-to-be’s in terms of which our conceptual responses are to be criticized. (KTE §43) From this, we can see why Sellars takes Kant’s analysis to be consistent on the assumption that God plays the role of the final judge of our epistemic engagement.16 In other words, while Sellars agrees with Kant’s analysis of our concept of an object of experience as underlying all the qualifications we covered in Chapter 2, he does not agree with Kant’s epistemological background setting. On Sellars’ interpretation, Kant’s agnosticism rests on a model according to which God’s intellectual intuitions function as limiting concepts of our own flawed intuitions. What does this entail for our epistemological situation?

Transcendental Phenomenalism 135 [I]n our attempt to give an account of how our intuitive representings [judgings about matter of facts] might be Erkenntnisse without being literally true, we are limited to making use of such abstract concepts as existence-in-itself, existence ‘in’ representings, receptivity, form of intuition, judgment, etc. One might formulate it [Kant’s position] as follows: Reality is such that finite minds non-arbitrarily, in accordance with their forms of receptivity, and their conceptual frameworks, represent this-suches [determinate individuals] and make judgments about them. Only God, however, knows how reality is. (SM II §50) But to say what it means that God knows how reality is, we are bound to negatively extrapolate from our own epistemological capacities to intuit and know things as they are and to attribute a pure and more efficient version of them to God.17 Sellars, in effect, rejects this background model: Kant’s account implies indeed that certain counterparts of our intuitive representations, namely God’s intellectual intuitions, are literally true; but these literal truths can only be indirectly and abstractly represented by finite minds, and there is an impassible gulf between our Erkenntnisse and Divine Truth. (SM II §51) In other words, Kant needs the limiting concept of God’s intellectual intuitions as a contrast operating in the background of his delineation of the limitations that restrict our intuitions. From this, we can see why Sellars reads Kant as putting forward an agnostic phenomenalism: For Kant, then, an act of intuiting a manifold is a thinking of a thissuch in space and/or time. The this-such is something that exists ‘in’ the act. The problem with which Kant is dealing can be characterized […] as that of whether individuals in space and/or time also have existence per se. Kant’s answer […] is that these intuited items exist only ‘in’ acts of intuition. That is, no items in space and/or time exists per se. He will nevertheless insist that some items which exist in acts of intuition are actual. This obviously requires a distinction between actuality and existence per se, which were conflated by his predecessors. (KTI §17) In our discussion of KTI above (Chapter 2), we saw that Sellars gives a benevolent reading of the notion of actuality which Kant develops as the immanent way for represented contents to exist. From that angle,

136  Transcendental Phenomenalism it is clear that Sellars endorses at least one aspect of Kant’s analysis of acts of intuiting. Intuitions, as conceived of in an everyday, common sensical setting, put us in touch only with actual contents. With regard to the status of objects of the Manifest Image, Sellars agrees with Kant (cf. Chapter 5.3). More importantly, however, Sellars disagrees with Kant’s agnosticism concerning the knowability of the things in themselves (cf. SM V §102). The shape of the Kantian conception of empirical knowledge is an immanent conception only in the following sense. We are affected by the in itself, and we form intuitions because of the passivity that is characteristic of our receptivity, and yet our world does not contain things in themselves. Sellars reconstructs Kant’s reason as follows: Thus, while Kant undoubtedly thinks that there are features of the in-itself which are, in some sense, the counterparts of the plurality of physical appearances, he finds this notion empty in that, as he sees it, we can have no determinate conception of this plurality. All determinate conception, as far as human minds are concerned, involves spatio-temporal schematization, and, as we have seen, he regards the concepts of Space and Time as unambiguous in a way which entails that if Space and Time are transcendentally ideal, anything we determinately conceive of in spatial or temporal terms must be transcendentally ideal. (SM II §33) God’s intellectual intuitions are not subject to the spatio-temporal schematisation of the Kantian categories; they are acts of the intellect. Our intuitings, however, are barred from reaching behind that barrier, and for that reason, the concept of a thing in itself remains empty. It cannot be given a determinate characterisation, only a transcendental identification within the scope of the chain of reasoning taking into account that our intuitive acts of representing are caused. With this, we see that this is all strong immanence can do on the score of accounting for verticality. This is the meaning behind the restriction contained in the tongue-incheek remark (KTE §43) that ›only God could envisage the epistemic norms of correctness in terms of which our conceptually articulated intuitive responses are to be criticized‹. The very thought, however, that God can have a determinate conception of how the in itself affects us implies a static conception of Divine truth. It is static because it is definitive and ultimate, and, as we can say from a transcendental point of view, because it functions as a limiting concept generated in abstraction from our humane fallibility. That is, there is the notion of an impassible gulf between our cognitions (Erkenntnisse) and Divine intuitions in the background of the way in which the Kantian account of our knowledge is strongly immanent.

Transcendental Phenomenalism 137 Now, since Sellars does not think this is the last word on the question of how the structure of empirical knowledge can be given an immanent reading, we need to consider his argument from counterparts which opens up a path of going beyond Kant. Before we turn to an extended discussion of that argument, however, we first have to see what moderate immanence amounts to. 5.2.2  Sellarsian Immanence What form does Sellars’ argument from immanence have? Can Sellars ensure that the resulting conception of our intentional reference amounts to a theory that, as ›a theory is part of the world of which it purports give knowledge‹ (KTE 41)? The double nature argument is presupposed here because the elements of the theory, languagings in which empirical matter-of-factual statements are uttered, have a double nature. So this is in the background when Sellars takes to reject the static conception of Divine Truth he finds in the background of Kant’s unknowability thesis: If […] we replace the static concept of Divine Truth with a Peircean conception of truth as the “ideal outcome of scientific inquiry,” the gulf between appearances and things-in-themselves, though a genuine one, can in principle be bridged. (SM II §51) It is helpful to first sketch the overall strategy this rejection of ‘Divine Truth’ fits into. Carrying out this program entails replacing (a) the role of God in the overall picture as the super-perceiver with the analogical concept-forming work of the overall scientific community. It also entails replacing (b) the concept of intellectual intuitions as the instrument for gaining knowledge about the in itself with conceptual structures generated by the sciences to understand the nature of counterpart objects. And finally, it entails (c) considering statements made in these conceptual structures through the lens of the double nature argument as standing in increasingly more adequate isomorphic relations to the objects they are (horizontally conceived) about. These are causally as well as spontaneously evoked relations in which languagings, in so far as they are considered as natural-linguistic-objects, picture the objects they are brought about by.18 Sellars’ account of our immanent knowledge is best approached by addressing this replacement thesis stating: The static concept of Divine Truth can be replaced by a Peircean conception of truth as the ideal outcome of scientific inquiry. Our interpretive hypothesis leading the following discussion can be put this way: To perform this replacement in effect amounts to altering the underlying notion of our epistemic relation to the in itself. That is, Sellars needs to reject Kant’s unknowability

138  Transcendental Phenomenalism thesis about the in itself, and thereby to make room for alternative ways of schematisation. Thus, the function of object-concepts generating conceptual structures of the theoretical sciences is to be scientific objects in their role of a knowable in itself. This entails slanting Kant’s concept of the in itself. In effect, it amounts to claiming that the empirical world and its objects do not only exist as the contents of our acts of representing it (SM V §101–102). Put positively this means, Sellars has to argue that the ‘world’ and its objects exist also meaningfully outside of our acts of representing it. For this he has to claim that they do so in ways which we can approach by way of analogically extending our current concepts to capture counterpart qualities we attribute them to have (cf. Chapter 5.3).19 Do we have evidence for this kind of optimism in Sellars? The following passage from SM VI shows that we do. According to Sellars, there is a viable way of being a realist in the required sense. In the context of Sellars’ aim of rejecting the Divine concept of truth, we can interpret why he thinks we should be realists about the source of our representings. He argues that not only God but our own scientific endeavours can account for the structure of our acts of intuiting: As I see it, in any case, a consistent scientific realist must hold that the world of everyday experience is a phenomenal world in the Kantian sense, existing only as the contents of actual and obtainable conceptual representings, the obtainability of which is explained not, as for Kant, by things in themselves known only to God, but by scientific objects about which, barring catastrophe, we shall know more and more as the years go by. (SM VI §61) This includes the endorsement of a certain type of phenomenalism on Sellars’ behalf (SM II §49; cf. Chapter 5.3). He is careful not to claim that this is Kant’s position (SM V §102). If we focus on what shape he gives to the knowability thesis, we see that in this passage, he claims that (1) if we substitute the Kantian conception of the unknowable things in themselves with concepts of scientific object, we (2) can come to know gradually more and more about them. The form of an argument about the concept of ‘knowability’ here is strictly transcendental. It is the concepts of the theoretical sciences in terms of which we are to articulate our understanding of the material nature of the objects we take to be knowable. To think something as knowable is to think it as an element in a realm for the constitution of which we sign as responsible. This gives us the essentials of Sellars’ version of the argument from immanence. As epistemic agents, we can obtain knowledge of the world we are a part of. This amounts to a transformation of the conception of

Transcendental Phenomenalism 139 our passivity. Now we can at least think ourselves as passive with regard to objects impinging on us, and about which we can come to know more as time moves on – not of a world of objects that is in any sense ‘behind’ our world, or transcendent to it. 20 The difficulty in understanding the specific build of Sellars’ theoretical stance lies in the fact that he approves of one dimension of Kantian immanence. To put it in a short characterisation: Sellars endorses Kant’s phenomenalism with regard to the ontological status of the objects of the Manifest Image, but he rejects Kant’s account of their status from the standpoint of the Scientific Image. This is the meaning of the thesis that Sellars endorses a transcendental phenomenalism but is also committed to a transcendentally motivated realism about the objects we come to conceptualise in new ways. What exactly makes it a transcendentally demanded realism is the following line of thought: If nature is not to be equated with the in itself, nor what common sense says it is, then nature needs to be conceived of as a content we can, in fact, develop and approximate as the content of the conceptual structures we generate in pursuing scientific inquiries. Here is a passage lending support to this interpretation, a passage that articulates this idea from a level of abstraction that leaves it open which sciences are included. The thesis I wish to defend, but not ascribe to Kant, though it is very much a ‘phenomenalism’ in the Kantian (rather than Berkeleyian) sense, is that although the world we conceptually represent in experience exists only in actual and obtainable representings of it, we can say, from a transcendental point of view, not only that existence-in-itself accounts for this obtainability by virtue of having a certain analogy with the world we represent but also that in principle we, rather than God alone, can provide the cash. (SM II §49) This shows that Sellars’ knowability thesis rests on the claim that things in themselves can be construed as, in fact, having a determinate structure we can aspire to eventually conceptualise. And more than that, the idea that they have a structure already presupposes or makes use of a positive analogy with objects we have successfully conceptualised. That is what it means to say that we can provide ‘the cash’. Our claims about matters of fact are not only subject to God’s assessment. The idea that our claims are correctly semantically assertible raises the question: assertible by whom? (cf. SM V §48) In reply, we can formulate a dynamic conception of truth that accounts for conceptual changes inside and across conceptual frameworks. Such a conception of truth, however, would have to implement the role of analogical concept formation as substantial. Then our analogically formed conceptual structures could be generating new

140  Transcendental Phenomenalism determinate explanations of the causal structures effecting our conceptual responses to them:21 [T]he use of analogy in theoretical science, unlike that in theology, generates new determinate concepts. It does not merely indirectly specify certain unknown attributes by an “analogy of proportion.” [Instead, it specifies the positive role of analogical concept in theoretical science] One might put this by saying that the conceptual structures of theoretical science give us new ways of schematizing categories. (SM II §49) With this, we have glimpsed in what way Sellars wants our knowledge to be knowledge brought about by and directed at objects which are in an encompassing way part of our horizon. 22 But while all of this is formulated in a promissory and optimistic way, what arguments can Sellars offer to evade the claim that we are confined to strong Kantian immanence? How does Sellars think we can get a foot in the door leading out of that confinement? In KTE Sellars parsed the claim contained in the unknowability thesis namely, that only God would know which of our claims to knowledge would ever be correct, as the claim that we would be confined to the internalist fact-checking analogy according to which ‘comparing a judging with a state of affairs’ could only be comparing a judging with another judging of the same specific kind, and this would no more be a verification than would checking one copy of today’s Times by reading another. (KTE §26) To elucidate this fact-checking analogy Sellars utilises the argument from double nature: In evaluating the significance of this point, it should be borne in mind that linguistic episodes have not only logical powers but also, and necessarily, matter-of-factual characteristics, e.g., shape, size, color, internal structure, and that they exhibit empirical uniformities both among themselves and in relation to the environment in which they occur. They can be compared as objects in nature with other objects in nature with respect to their matter-of-factual characteristics. I mention this, because the fact that we tend to think of conceptual acts as having only logical form, as lacking matter-of-factual characteristics, i.e., as, to use Moore’s expression, diaphanous, makes it difficult to appreciate that the ultimate point of all the logical powers

Transcendental Phenomenalism 141 pertaining to conceptual activity in its epistemic orientation is to generate conceptual structures which as objects in nature stand in certain matter-of-factual relations to other objects in nature. (KTE §27) In the footnote, Sellars articulates why he takes this to amount to the point that necessarily leads beyond the Kantian form of strong immanence while also opening up the space for a conception of nature of which we, as thinking and perceiving agents, can be understood to be a part of: The basic flaw in the Kantian system (as in that of Peirce) is in its inability to do justice to this fact. The insight that logical form belongs only to conceptual acts (i.e., belongs to “thoughts” rather than to “things”) must be supplemented by the insight that “thoughts” as well as “things” must have empirical form if they are to mesh with each other in that way which is essential to empirical knowledge. I have developed this point in [TC] and [SM V]. (KTE §27 fn7) The relevant thesis here is that it is due to their double nature, then that conceptual elements, not only linguistic ones, stand in a causal, isomorphic relation to the items or objects, to which they normally are only conceived to be semantically related. From a logical point of view, one can distinguish between a claim to the effect that (a) a causal connection and (b) an isomorphism obtains between a system of languagings and some structure of the world. The very idea of an isomorphism as such does not entail a causal connection, since one can have two domains of things be isomorphic but not causally connected. Picturing, however, covers both cases – as it entails the claim that languagings are causally evoked by happenings of certain kinds, e.g. in language-entry moves, and the claim that languagings can be brought about freely in empirical uses (NAO V §63). This adds a further element to the claim that our knowledge about the world is modestly immanent. Knowledge-claims, understood as conceptual or linguistic activity, are always realised in tokenings of utterances or inscriptions. And the argument there is that a conceptual structure, qua double nature, embodies not only knowledge of matter of fact but, by necessarily having an empirical form, is itself part of nature. Sellarsian immanence then takes the following form: (1) judging-­ tokenings about empirical matters of fact (2) themselves can be construed to have empirical or matter-of-factual form, and (3) therefore stand in a matter-of-factual relation to one another. In the wording of NAO V, this becomes the thesis that (1) acts of judgings concerning matters of empirical facts are languagings that themselves can be construed as (2) natural

142  Transcendental Phenomenalism linguistic objects which (3) are systematically related to other objects in nature. The conclusion Sellars is driving at can be stated as follows: the matter-­ of-factual relation of picturing is a transcendental requirement on any language. But it is also a conclusion worth driving at because it secures that a language that – in a sense to be explained – pictures the world can legitimately be said to be part of this very world, thereby meeting the immanence criterion. However, to unfold the full range of functions, the notion of picturing comes to play in Sellars transcendentally motivating a realism (about the causal origin of our perceptual responses) while endorsing a transcendental phenomenalism (about our epistemic access to the impingements on us generated by these structures), we need to take into consideration a further argument about the role of counterparts to the ‘objects’ we treat as real here and now.

5.3  The Argument from Counterparts The aim of this argument is to create conceptual space for the idea that we can think grades of truth in the development of conceptual structures as a whole (cf. SM V §54–102). (1) The concepts we use in our present conceptual structure to individuate objects will have successor concepts in later conceptual structure. Successor concepts are derived from our concepts and analogically build on them (cf. ‘gas’ 50 BCE and ‘gas’ now). (2) If later conceptual structures comprise counterpart concepts, they also comprise counterpart objects individuated on account of them. The objects they individuate are counterpart objects to the objects we individuate in our current conceptual structure. (3) The non-conceptual component (demonstrative core) in the direct reference of intuitions (covert languagings) is invariant across conceptual structures, even though its conceptualisation changes according to the semantic norms of the respective conceptual structure. (4) New concepts can be generated to account for the same, invariant impingements, new languagings be produced in response to the same causal input. (5) If languagings in a later conceptual structure will have a logical form that is more conducive to our epistemic ends (allowing us to predict more successfully), on account of their double nature, this will be reflected in their empirical form. (6) Insofar as they form a more adequate picture than languagings in earlier conceptual structures did, languagings produced by a later conceptual structure are more conducive to our epistemic ends qua empirical form.

Transcendental Phenomenalism 143 (7) This makes possible the thought that different, later conceptual structures – when considered with regard to their empirical form – produce languagings that are more adequately isomorphic to the objects they are (from the viewpoint of their respective S-assertibility norms) about. (C) The isomorphism between objects and languagings of new conceptual structures (as a whole across time) makes possible the thought of grades of matter-of-factual truth in the sense of increasing degrees of adequacy of empirical form across the development of conceptual structures. To give up on the agnostic idea that we can never say anything about the items causing us to have perceptions of determinate individuals in space and time is to move to a position according to which we conceive of such items in terms of counterpart qualities. And to do that, in turn, is to analogically extend the concepts we already possess and to systematically form new concepts (cf. SM I §68–72). These would be counterpart concepts to the concepts we already know how to work with. If, per impossible, Kant had developed the idea of the manifold of sense as characterized by analogical counterparts of the perceptible qualities and relations of physical things and events he could have given an explicit account of the ability of the impressions of receptivity to guide minds, endowed with the conceptual framework he takes us to have, to form the conceptual representations we do of individual physical objects and events in Space and Time. He could thus have argued that when on a certain occasion we come to have an intuitive conceptual representation that this green square adjoins that red square, we do so by virtue of having a complex of non-conceptual representations which, although non-spatial and without colour, have characteristics which are the counterparts of square, red, green and adjoining, and which make them such as to account for the fact that we have this conceptual representation rather than that of there being a purple pentagon above an orange ellipse. (SM I §78) With the help of counterpart objects, we are come to be in a position to account for systematic co-variances between representational episodes and their causes, without stipulating that we come to perceive the objects as they are in themselves. The dimension of the conceptual continues to play its constitutive role, a role it never loses in favour of a direct account of perception. In this regard, Sellars remains a critic of the Myth of the Given 23 – even where he introduces the notion of later conceptual

144  Transcendental Phenomenalism structures in which we can be taken to respond with different yet (pictorially) more adequate conceptualisations to the same stimuli we are subject to in our present conceptual structure. Yet, the crucial thesis for Sellars’ advancement of the Kantian frame holds that languagings generated by theoretical sciences are characterised by different degrees of pictorial adequacy. This, in turn, is not an epistemic first-order claim about languagings in future conceptual structures but a claim about what is entailed in thinking the thought of an increase in epistemic success. In the next section, we turn to an elaboration of what this transcendental line of thought amounts to and to answering the question in what sense this allows for the thought of a non-semantic grading of truth qua picturing capacity of languagings.

5.4 Picturing The concept of picturing allows Sellars to articulate how the vertical dimension is built into intentionality. In this regard, the picturing postulate resembles the point zero, where the axes of the dimensions come to meet. If a natural linguistic object does its picturing work by being isomorphically related to an object in its vicinity, the burden seems to shift to the concept of an isomorphic relation. Sellars clarifies that it is a second-order isomorphism he has in mind here (TC 217–219). A first-order isomorphism would hold between the sole of my shoes and the footprints I leave in the sand. But since neither my thought about a red pyramid in front of me is red, nor my uttering ‘there is a red pyramid in front of me’, the relation between these items in the real is a second-order isomorphism. 24. It is second order in the sense that it depends for its structure on the method of projection inherent to the conceptual scheme in which it is produced. As Sellars writes, alluding to Wittgenstein’s Tractatus: Picturing […] is […] a relation between two relational structures. (SM V §56) It is central to Sellars’ version of immanence that this picturing relation is non-semantic, i.e. that it does not operate because it serves representational, or intentional relational purposes. 25 Rather, it serves as the causal underside of our linguistic activity, of our conceptual activities we are normally bound to interpret and read from within the intentional stance towards the linguistic activity of fellow interpreters and speakers, as we can say, appealing to Dennett’s (1987) classic formulation. However, it takes the perspective of transcendental philosophy to see the transcendental function of the picturing as contributing to the epistemic significance of language as a tool for generating knowledge. The relevant claim here is that picturing it is not just non-semantic; rather, that it is the precondition of there being empirically contentful

Transcendental Phenomenalism 145 statements at all. In this sense, picturing is the missing vertical transcendental prerequisite for the non-relational intentionality of our conceptual system as a whole. This interpretation of the transcendental significance of picturing we can read from the claim that the ultimate point of all the logical powers pertaining to conceptual activity in its epistemic orientation is to generate conceptual structures which as objects in nature stand in certain matter-of-factual relations to other objects in nature. (KTE §23) This claim echoes and substantiates the dependency claim at the heart of the double nature argument (cf. Chapter 5.1): The logical form of a stating is clearly not the empirical configuration of the sentence it illustrates, though having an appropriate empirical configuration is a necessary condition of the stating’s having the logical form it does, in the language to which it belongs. (KTE §22) In light of these considerations, the Sellarsian version of immanence can take it as a matter of indispensability that our knowledge about the world is part of the world it claims to be knowledge of, because the very knowledge claims are just another type of natural-linguistic-objects in the very same nature they purport to be knowledge claims of. 26 The difficulty here lies in the fact that we can look at knowledge claims from the intentional stance and from the engineering stance. 27 One time considering them with regard to their logical form, another with regard to their empirical form. When we view knowledge claims as conceptually contentful, we put focus on their logical form. But the aboutness of conceptually articulated knowledge claims is a matter of their functional roles, not of their direct intentional reference to empirical matters of fact. Regarding the latter, the idea of picturing plays its core role since Sellars claims that the correctness of assertions is to be defined in terms of the non-semantic adequacy of the underlying picturing relation and not vice versa. Sellars is most explicit about this in the opening of the chapter on picturing: I speak of the distinctive functions of first-level matter-of-factual discourse; for even within this level essential distinctions must be drawn if we are to grasp the difference between the primary concept of factual truth (truth as correct 28 picture), which makes intelligible all the other modes of factual truth, and the generic concept of truth as S-assertibility. (SM V §9)

146  Transcendental Phenomenalism According to this qualification, the picturing relation is basic and primary because it needs to be in place such that claims about the semantic assertibility of statements can be made. 29 In other words, the concept of picturing comes to play the role of factual truth, enabling the horizontal dimension where truth is a matter of the semantic assertibility of statements.30 Why is ‘being true’ also a matter-of-factual relation among objects in nature? It is a matter-of-factual relation in the sense that (a) conceptual items qua empirical form are natural linguistic objects standing in an isomorphic relation to (b) other objects in nature. For only when thoughts and utterances have empirical form can they ‘mesh’ with things (KTE §27). And with this, we have a first way of replying to the question we raised above (at the beginning of Chapter 5.2): Does Sellars’ transcendental methodology afford us a direct intentional reference to the world? This also means finally replying to the dissatisfaction that Sellars’ non-­ relationally oriented elucidations of the concept intentionality have not afforded us direct intentional reference (Chapter 4.4). According to Sellars, we are directly, though interpretation-­dependent, related to the contents of our matter-of-factual judgements by the fact that these, in turn, are conceptually articulated contents possessed by logical powers.. And we are directly related to objects in nature by the fact that our languagings themselves, considered as natural linguistic objects are part of nature. Just what this relation, metaphysically speaking is, we have not yet developed, but that it holds is part and parcel of Sellars’ transcendental perspective. 31 According to Sellars, wanting to be semantically related to something outside of such a conceptual uptake is wanting the world to imprint itself, its categorial status on us. It is, in other words, one of the many variations of the Myth of the Given. Does this imply that Sellars rests content with a horizontal version of intentionality? In a sense yes, since the argument from immanence obligates him to conceiving of intentionality in the order of the conceptual as a horizontal affair through and through. This was apparent in his endorsement of the transcendental ideality of the image-model world and in his agreement with Kant’s analysis of intuitions or language-­ entry moves as always already possessed of categorial structures (i.e. with regard to B143 and B165). Or, as we can now say, in light of the terminology discussed above, as possessed of logical powers. But the transcendental analysis of how to give consciousness a place in nature does not stop there.32 It does not stop there because Sellars is committed to an idea of nature that does not dissolve into our estimations of it. Instead, we fare better when we read Sellars’ argument from immanence as a transcendental consideration leading him to the following conclusion: We can approach nature always only from within our conceptual scheme. And that this is so follows from the non-relational character of

Transcendental Phenomenalism 147 our mental states and claims. With regard to their intentional directedness, all articulations of the content-fulness of our thoughts and claims about the world amount to functional classifications in which we express our endorsements of actual contents as being of a certain kind (Chapter 4.2). But that never takes us outside the realm of the conceptual, which is a view fully exploited by Hegel’. And yet Sellars’ conception of nature is larger than the one he finds in Kant. To see that, we need to explicitly take up the position of the transcendental philosopher (considering from M2 the construction of a theory which itself can be said to be part of the world it is set to be about). It is one thing to find that our grasp of empirical contents is nothing over and above a conceptually articulated judging content. It is another to grant that such an act, qua being the act of a person, needs to be also conceived of such that it is part of the world. And this is a line of thought, according to which our language (considered as a conceptual structure as a whole) stands to ‘nature’ then viewed as a system of representables, in matter-of-factual relations.33 These are relations to which Sellars is committed on the basis of the following thought that brings the two arguments from double nature and from immanence on one track: If we make explicit that the argument from immanence is a specification of the argument from double nature applied to matter-of-factual judgements, then we are able to take matter-of-factual judgements to be related to something non-linguistic by being natural-linguistic objects picturing objects in nature. And because such a matter-of-factual relation holds between the items pictured and the items doing the picturing, we can consider the items doing the picturing from the intentional stance point of view, thereby allowing us to make contentful claims about them. Sellars holds that our linguistic activity is reflected in causal uniformities, the structural features exhibited by the utterances we produce (NAO V). Their isomorphic relation is never epistemically accessible from within linguistic practices.34 Rather, it is external in the sense of being the manifestation of linguistic practices in manifest uniform patterns. We can never make use of this relation, say, to further epistemic ends. It is the causal shadow of our languagings. It is never available and given to us in semantic terms. To claim that it is only a transcendental requirement on our linguistic practices being efficacious is to claim that something like a picturing isomorphism has to be in place in order for our linguistic practices to advance through time in a multiscalarity of engagements (Cobley 2019, 703). Yet, we cannot step outside, i.e. go sideways, to look at the picturing capacity of any single statement. Even though, as one might ask, in doing cognitive science, are we not studying picturing capacities? Yes, we do, however, always parasitically on our current method of projection, i.e. the framework of concepts, rules and (inferential- and bio-­ functional) norms.

148  Transcendental Phenomenalism With these qualifications in place, we now need to consider two objections raised from a pragmatist perspective. They will allow us to detail the specific role that the concept of picturing plays for Sellars’ fully developed conception of intentionality. The argument pursued here has it that this specific role can only be explicated appropriately when the transcendental methodology is seen as providing the frame. It is the frame for articulating how we are to understand the relation between our passivity and our spontaneity in the context of generating (scientific) knowledge. Picturing answers to the question of how our linguistic practices can be taken to be part of the world of which they non-­relationally produce knowledge. However, our linguistic practices crucially do include entry- and exit-moves, and so they produce knowledge in virtue of both causal-relational and normative-non-relational, mutually sustaining patterns. This entails that we only ever move horizontally in our scorekeeping activity, while transcendentally taking these moves to be embedded in the world, but in a way that never comes to be disclosed to us through the picturing activity of our languagings.

Notes 1. As will emerge, our disagreement with the position developed in Haag (2007) does not concern its interpretive substance but rather the label he chooses. On our reading, to call Sellars a transcendental realist is misleading. I prefer the label ‘phenomenalism’ because it helps understand that for Sellars our epistemological situation never changes in the sense that we will come to perceive things as they are in themselves. Rather, according to Sellars, the manifest objects retain their status as being transcendentally ideal. Yet, there is a further sense in which Sellars’ transformation of the Kantian concept of passivity imports a commitment to a realist presupposition. Thus, the term ‘realism’ in this book is reserved to characterise the transcendentally motivated consequence of Sellars’ transcendental argument about the concept of an object of experience. 2. Here I agree with O’Shea’s interpretation in which he characterises the status of the concept of picturing with regard to its presuppositional role in the following way: ›My own view is that Sellars’ account of picturing and of empirical truth as correspondence represented a searching if sketchy attempt to argue that the normative aspects of meaning, reference, and truth are not reducible to, yet presuppose for their possibility, various specific underlying causal patterns and representational mappings […]. It is the latter mappings which systematically relate cognitive and linguistic systems to the world which they thereby succeed in being about – even though aboutness itself is not a further mysterious relation to the world. Investigation of these modes of cognitive representation has since become the cooperative business not only of philosophers of mind, epistemologists, and philosophers of language, but of linguists, cognitive psychologists, and neuroscientists as well. Central to Sellars’ philosophical quest was the attempt to envision the overall conceptual space in which those sorts of detailed epistemological and scientific investigations might be seen to make sense‹ (O’Shea 2007, 158). I also find inspiration in deVries (2010) for the general thrust of my interpre-

Transcendental Phenomenalism 149











tation of the concept of picturing. However, I take my position to divert from deVries’ (2012b) interpretation of the status of scientific realism. Cf. Chapter 6.3. 3. Cf. ›The correlation between objects and their linguistic pictures must not be confused with the pseudo-relations standing for and denoting‹ (SM V §36). 4. Cf. ›The fundamental job of singular first-level matter-of-factual statements is to picture, and hence the fundamental job of referring expressions is to be correlated as simple linguistic objects by matter-of-factual relations with single non-linguistic objects‹ (SM V §26). 5. Cf. ›What is the basic job of empirical statements? The answer is, in essence, that of the Tractatus, i.e. to compete for places in a picture of how things are, in accordance with a complex manner of projection. Just how such a manner of projection is to be described is a difficult topic in its own right‹ (NI §29). 6. Cf. Stovall (2022) for an elaboration of this idea with regard to shared intentionality, rational agency and deontic, discursive, and practical cognition. 7. One might have to qualify this observation, as Sellars states in the introductory passages to his last book, with regard to metaphysical theses such as: ›‘Time is unreal.’ ‘Sense data are constituents of physical objects.’ ‘Mind is a distinct substance.’ ‘We intuit essences’‹. Sellars holds that they are of real philosophical interest and should not be brushed aside: ›These are issues you can get your teeth into‹. In reflecting on his philosophical upbringing, Sellars writes: ›By contrast, Pragmatism seemed all method and no results. […] As for Naturalism. That, too, had negative over tones at home. It was as wishy-washy and ambiguous as Pragmatism. One could believe almost anything about the world and even some things about God, and yet be a Naturalist. What we needed was a new, nonreductive materialism. […] I, however, do not own the term, and I am so surprised by some of the views of the new, new Materialists, that until the dust settles, I prefer the term ‘Naturalism,’ which, while retaining its methodological connotations, has acquired a substantive content, which, if it does not entail scientific realism, is at least not incompatible with it‹ (NAO I §2–7). According to Brassier (2011, 49), Sellars ›emphasizes the normative valence of knowing at the cost of eliding the metaphysical autonomy of the in-itself‹ — an assessment we take up in Chapter 6 and the Conclusion. 8. That is, I stress the Kantian elements in Sellars’ thinking over his scientific inclinations. This is, so I hope to show, reasonable, as his commitment to Kant’s transcendental methodology of explicating the concept of intentionality runs deeper and in this sense is the ground on which to reconstruct his commitment to scientific realism. In this regard, I prefer the label ‘transcendental phenomenalism with a transcendentally demanded realism’ over the label scientific naturalism, as is used by O’Shea (2007, 2010, 2016b, 2021b). 9. That this is true of ‘acts of speaking out loud’ can be made clear. The corollary, that it also extends to thought and other mental states (that characteristically share their intentionality), depends on a further thesis: The thesis that language expresses thought in Sellars’ hands becomes the thesis that methodologically our understanding of the logical structure of thought depends on our grasp of the logical structure of utterances. Such that we depend on our knowledge of language to interpret silent episodes in which we take our interlocutors to carry on in an inner dialogue. Sellars argues in extenso for this reading in EPM, LTC,

150  Transcendental Phenomenalism











MFC, and SM III. Cf. Related considerations in favour of this linguistic approach to capturing the relation between thought and language are given by Stekeler-Weithofer (2014, 28–31, 85–86), Kambartel & Stekeler-Weithofer (2005, 193ff.), and Barth (2010, Chapters 1–3). 10. In O’Shea’s terminology, the fact that languagings need to be understood as rule-governed to play out their function does not prevent them from also having a causal shadow: ›Linguistic and mental representations are always Janus-faced on Sellars’ view, simultaneously backed up by norms and fronting the world; or, to switch metaphors, the causal-­representational regularities are in a crucial respect the slaves of our rule-governed reasons, and it is essential to the resulting ‘purely descriptive’ pictures of the world that they are […] the causal shadows of norms‹ (O’Shea 2010, 467). 11. For an extended reconstruction of the significance of this point see O’Shea (2007, 138ff.). 12. Cf. Christias (2015, 556–557) mobilising this point against Sachs’ (2014) invocation of somatic intentionality. Since Sachs, as Christias points out, sails past the transcendental setup of Sellars’ overall argument, we have not engaged with his attempt to bring Merleau-Ponty’s phenomenalism in conversation with Sellars’. 13. Cf. ›My argument is that, in the case of matter-of-factual statements (and, in the last analysis, the acts of thought to which they give expression), [their] role is that of constituting a projection in language users of the world in which they live‹ (NAO V §118). 14. Cf. I §9, fn 3 where Sellars puts it as follows: ›For an elaboration of the ontological framework which supports this analysis see my essay [KTE] and [SM II]‹. 15. In a related project, Stekeler-Weithofer (2014, 64–102) articulates the conditions and possibilities of contentive empirical assertions as inherently dependent on generic sentences. Although couched in a different philosophical terminology, his in-depth analysis shows how generic sentences serve as material-conceptual rules for classifying contents. 16. In reconstructing Sellars’ engagement with Kant’s position it is helpful to know that he reads Kant as only committed to articulating the coherence of the concept of an object of knowledge. For Sellars this means: ›It is […] obvious, on reflection, that Kant is not seeking to prove that there is empirical knowledge, but only to show that the concept is a coherent one and that it is such as to rule out the possibility that there could be empirical knowledge not implicitly of the form ‘such and such a state of affairs belongs to a coherent system of states of affairs of which my perceptual experiences are a part’‹ (KTE §11). In the closing paragraphs of the article Sellars picks up this observation again. We can read this as showing that Sellars is well aware of Kant’s commitment to a strong version of immanence: ›Thus Kant is in a position to grant that empirical knowledge involves a uniformity of conceptual response to extra-conceptual items and even that extra-conceptual items conform to general laws, without granting that the character of the items to which we conceptually respond, or the laws to which God knows them to conform, are accessible to finite minds. We could expect Kant to say that, if there is empirical knowledge, there must be such uniformities (once again, he is not attempting to prove that there is empirical knowledge, but to articulate its structure) and that, in the absence of particular reasons for thinking that something has gone wrong, we are entitled to suppose that our conceptual machinery is functioning properly‹ (KTE §45).

Transcendental Phenomenalism 151 17. Limiting concepts are relative to our own case only by way of negation. For a discussion of four different types of limiting concepts Förster (2012) finds in Kant’s resolving of the antinomies of the teleological power of judgment, cf. Haag (2012a, 994ff.). 18. Cf. NAO V §88–96, and §63: ›That languagings are evoked (in contexts) by happenings of certain kinds is a causal fact which is nevertheless essential to their conceptual character‹. 19. Even then we can distinguish between the notion of a world existing absolutely outside of our acts of representing, and the notion of a world existing outside a particular framework or set of representings. The idea that there is an absolute standpoint outside the set of representings is importance as a transcendentally identified vantage point, not as one that is in fact epistemically obtainable. Endorsing the method of abstraction does not entail endorsing attainability of a standpoint ‘absolutely outside’. Note that to Sellars himself his improvement with respect to Kant becomes visible precisely by taking into account the idea of there being various ‘conceptual frameworks’. In PH VII, he explicitly speaks of how talking of frameworks is a great new philosophical tool enabling us to solve many philosophical problems. In particular, Sellars’ idea seems to be that we can replace the notion of a world outside a set of representings with the notion of adopting different stances towards different frameworks (i.e. different, more adequate sets of representing): ›To say that physical objects do not exist is to make a point about the framework of physical objects and not in the framework‹. Cf. Chapter 1.3. 20. That is, although I agree with Stang (2016, 5) that the distinction is messy and unhelpful I think it makes sense to say that Sellars defends a ›one world‹ reading with regard to the phainomena-nouomena distinction. Stang gives two reasons to be dissatisfied: ›[1] because it is not a commitment of “two object” readings that, for each appearance, there is one and only one thing in itself that appears as that object. The “two object” interpreter can hold that each appearance is the appearance of an indefinite plurality of things in themselves. [2] Nor is the other standard term, “one world” versus “two world”, helpful, either, for “world” is a technical term in Kant’s metaphysics and has a very specific meaning. [Thus, one] can coherently hold a “non-identity” interpretation while denying that appearances in space and time constitute a “world” at all‹. The main point of Sellars transcendental phenomenalism together with a transcendentally demanded realism is to amount to a position which takes the objects of common sense to be transcendentally speaking unreal with regard to the present conceptual framework, while maintaining that it is conceivable that in a later conceptual framework there are resources to cash in the requirements such that their reality can be accounted for (cf. Chapters 5.1 and 5.2 below and SM V §§101–102). Cf. also Willaschek (1998), Rosefeldt (2015), see also Christias (2015), Matsui (2020), and Hoeppner (2020, 255–256). 21. This may sound as if we have already taken care of the ‘vertical’ dimension. However, we are not finished just because we have a place for the analogical formation of concepts. The reason is that Sellars’ aim is not that of accounting for the possibility of true statements in one given conceptual structure. Rather, he is interested in explaining the relation between (a) framework relative truth and a non-semantic notion of truth that ranges in grades across different conceptual structures. For that we need to consider the argument from counterparts (Chapter 5.3) and discuss the viability of the resulting conception of truth qua picture (Chapters 6.1 and 6.2)

152  Transcendental Phenomenalism 22. That this horizon includes unobservable entities poses no threat to the Sellarsian version of immanent knowledge as we can see from (1) his discussion of instrumentalism in SM V, SRI, SRT, and EPM and from (2) his rejection of reductive phenomenalism treating the conceptual framework of physical objects as subordinate to sense impression, cf. SM II §36 and PHM. 23. Even in passages such as the following: ›I agreed [in SM II] with Kant that the world of common sense is a ‘phenomenal’ world, but suggested that it is ‘scientific objects,’ rather than metaphysical unknowables, which are the true things-in-themselves‹ (SM V §79). 24. Cf. Chapter 2.2.3 where we have covered Sellars’ discussion of this point in KTI, which underlines that Sellars is well aware of the fact that this cannot be implied. If there is an isomorphism, it needs to be more abstract, mainly because it is a result of the method of projection inherent in the conceptual structure by which we respond to our environment in our perceptual takings, inferrings and actions. 25. This is the reason why see Sachs’ (2019a) attempts at semantically adapting the notion of picturing to recent work in cognitive neuroscience of predictive processing to be executed in a more Millikanian style than Sellars wanted. It certainly doesn’t mean it’s on a false track, just a different one. Sellars’s notion of picturing seems to be ideally suited to the sorts of inquiries involved in cognitive scientific theories of ‘neural representations’. However, I wouldn’t call the latter ‘semantic’, stressing Sellars’ normative-functional usage. This I need to underline, for I agree that Sellars himself does think of ‘picturing’ as a kind of natural-causal theory of mental and linguistic representation. TTP is a late, clear example, but see also BBK, MEV, NAO V in which Sellars often uses ‘representational’ for non-intentional, officially ‘non-semantic’ (non-normative) picturing-­ relations. The terms ‘semantic’ and ‘representational’ are especially slippery in this regard, both within Sellars, and certainly in discussing how his views related to ‘views at large’ (cf. Chapter 6). I side with O’Shea (2011, 2016b) in thinking of his theory of picturing as exploring the presuppositional basis for articulating a naturalistic theory of representation. This entails placing it within the larger scope of a theory according to which all such representation relations have to be understood as constrained by and produced by ‘norms’ within a wider ‘system’ of one kind or another, i.e. a ‘space of reasons’, a ‘proper-functional biological’ space, etc. Its key insight is that one can maintain the normative-functional perspective as primary while embedding within it theories of representation of the cognitive-­scientific kind. 26. One might suspect that Sellars needs to argue for the inverse claim to achieve his argumentative aim: i.e. that because knowledge is part of the world, we can see it to be knowledge. But it is the very point of the transcendental methodology that this cannot simply be assumed to be a fact. The situation here rather resembles a complex presupposition, one that the transcendental perspective of analysis helps to show: our language as a whole can be taken to yield knowledge only on the assumption that it comprises languagings that picture and that it thereby vertically is part of the world. This, however, is a world that the language users can only make horizontal claims about by moving inside the language. This is what Sellarsian immanence amounts to. We need the assumption that something like picturing has to be happening even though our non-relational meaning claims never take us outside the bounds of the language we actually use.

Transcendental Phenomenalism 153 27. Cf. O’Shea (2007, 153): “The basic idea is that we are now to consider human languages themselves from the naturalistic ‘engineering’ standpoint, but as always with one eye on the conceptually irreducible normative standpoint of intentionality. [For …] although we know that ‘linguistic objects are subject to rules and principles — are fraught with ‘ought’ — we abstract from this knowledge in considering them as objects in the natural order’, i.e. as […] natural-linguistic–objects (TC 212)”. 28. Notice how Sellars here speaks of ›correct‹ picture, which I think is misleading. After all, he points out, that picturing is not a semantic but a matter-of-factual, causal relation between objects in the real (TC 215). That we tend to conceive of pictures as correct is matter of our scorekeeping stance, the normative eye with which we need to see things to see their value and function. This, however, is a feature we bring to the theoretical object of concern, it is not a feature it has independently of our perspective on it. In this sense, we can distinguish (a) the correctness involved in the normative perspective of judging a certain claim to be correctly made from (b) the adequacy the utterance of the claim has when considered in abstraction, with an engineering eye, as a natural linguistic object by being isomorphically related to an object. Cf. Sicha’s (2014) criticism of Rosenberg (2007, 104ff.). To illustrate my point: These two will be systematically related to each other. If I say or think now, ‘The dog here is on this mat’, this will be both a correct ‘entry’ move and a ‘correct picturing’ of the world – insofar as my conceptual framework is adequate as far as picturing goes. The fact that we can only evaluate the latter through norm-governed empirical inquiry does not mean we cannot (fallibilistically) say that my sentence or thought, as far as I know, ‘correctly pictures’ the empirical situation, qua physical sentence, and qua whatever it is ‘little thoughts are made of’. The ‘correctness’ claim – the naturalistic ‘picturing-representational relation’ – will of course always be parasitic on the relevant norm-governednesses that have generated and sustained those causal relations, however (in)adequately. But with all of that in mind, the phrasing ‘correctly pictures’ can be used. Considering how ‘the mat is on the dog’ by contrast, would have been both semantically incorrect and in a non-accidental way. That is, thanks to the ‘rules’, we would treat it as a very bad, inadequate, and so derivatively, we can say ‘incorrect’ picturing of the objective state of affairs. We have a rough sense of why and how the latter natural-linguistic object (i.e. ‘the mat is on the dog’) would, as a tokening of that linguistic type, not be causally related in general in the ‘right way’ to those types of empirical objects, given how we generally have used ‘is on’s, ‘dog’s, and ‘mat’s in English. 29. ‘Truth as semantic assertibility’ applies to all truth claims, e.g. empirical (matter-of-factual), moral, and mathematical. Sellars does here (SM V §9) say that the empirical-representational dimension is primary in the sense of being necessary if there are to be any of the other practices and claims (as he does in PSIM). But in SM V §9–10 he is primarily making a distinction within the empirical domain, between ‘first-level matter-of-factual discourse’ and other kinds of matter-of-factual discourse, in particular claims about laws, quantification, etc. Basically, here he wants to get down to the ‘atomic’ picturing statements that underlie the others, in a Tractarian spirit. 30. Cf. SM V §51 where Sellars delineates his understanding of the ‘absolute sense of true’ as always meaning ‘true with regard to our conceptual structure’. This leaves a difference between ‘semantically assertible truth’ relative to older conceptual structures and the ‘absolute sense of true’ that we now embrace. See Christias (2015), as mentioned in Chapter 5.2.

154  Transcendental Phenomenalism 31. The next subchapter discusses the relevance of the claim that this ‘matter-­ of-factual relation’ can be understood to become gradually more adequate. Here it is of utmost importance to keep in mind that the relation is factual and as such always non-semantic. This point – that picturing never becomes semantically accessible – has been misunderstood, as deVries (2012b) shows for Reider (2012) but also by Rorty (1988), Millikan (2012), Rouse (2015), and Price (2013, 2015, 2016). Since it continues to pose difficulties for interpretations of Sellars’ conception of nature and the possibility of our reference to it, we will discuss it in Chapter 6, but see also Chapter 1.3. 32. To do this would imply resting content with the horizontal view in a way Rorty and Huw Price suggest we should do (cf. Chapters 6.1–6.2). Cf. also Rorty (1972, 1997), Price (2011, 2013) and Rorty & Price (2010). The metaphor of horizontal versus vertical ways of looking at representations appears already in Rorty (1961, 219) and is taken up Price (2013, 46, 106, 188). Cf. also the related discussion in Macbeth 2010. 33. According to Sellars, this consideration is in no way restricted to the English language only. To make his point, Sellars invokes the idea that different languages can share one conceptual structure. This claim rests on the idea that in any two given languages a sign design can be interpreted to fulfil similar functional roles: ›Thus, to characterize a statement in a foreign language, for example, French as true is, in effect, to treat this language as a “dialect” of a language game which we play, i.e. to treat speakers of French as speakers of our language, as players of a common game. Since the term ‘language’ as it is ordinarily used refers to the specific linguistic materials (sign designs and surface grammar) which differentiate, e.g. French from German, we need another term for the common game which is played by users of such differing resources. I shall use the expression ‘conceptual structure’ to serve this purpose‹ (SM V §49).   To which Sellars adds in a footnote: ›It should be borne in mind that ‘conceptual structure’ in this sense refers to [overt] language games. It does not refer to conceptual activity in the sense of “inner episodes.” I am assuming, as before, that once the epistemic and ontological categories with which we are concerned have been clarified in their application to Rylean items the extension of this clarification to “inner-episodes” poses no difficulty of principle‹ (SM V §49 fn21). 34. To give this claim further nuance: There is a difference between saying that we always must evaluate correspondence from within some framework or other, and the claim made that no correspondence evaluations can ever be made of a framework from within (and parasitic on) its own resources. For example, we do study our own perceptual and linguistic representational systems (e.g. facial recognition patterns) both relying on our normatively classificatory ‘correctness’ conditions and seeing how these patterns match up isomorphically (causally, extensionally). The patterns thus identified as covarying isomorphically are themselves assessed and individuated in terms and from within of our current conceptual frame. As such they too have a causal underbelly. Can we thus grant the ‘shadows of norms’ point and examine the norm-parasitic causal relations within our own framework. This, however, does not mean that we can epistemically step outside our current frame and evaluate its potential from outside. As of course is also the case with cross-framework corrections, etc. We study and correct the correspondence failures, always relying on our current concepts. I am indebted to James O’Shea for prompting me to clarify this point.

6

Objections and Consequences

The neo-pragmatist sees in Sellars’ work the possibility of giving up on ‘theoretical philosophy’, as bad-old metaphysics. Sellars refuses to do this, and it is worth articulating why. Doing so will, among other things, allow us to shed light on Sellars’ scientific realism, one crucial locus of neo-pragmatist concern. The problem objections from pragmatism revolve around has to do with the place of picturing in Sellars’ philosophy, and I want to distinguish a more radical and more conciliatory version of the complaint. Rorty’s radical complaint is going to be that an interest in picturing as such betrays Sellars’ backsliding into representationalism and bad-old-metaphysics (Chapter 6.1). Price’s more conciliatory objection is that picturing can be accommodated, in a basically Rortyan picture, by interpreting it as e-representation (Chapter 6.2). In effect, I am going to argue that both of these lines of thought overlook the transcendental role of picturing, which, in turn, makes Sellars scientific realism seem far more innocent than the skeptical pragmatist recognises.

6.1  Objection from Pragmatism For a pragmatist such as Richard Rorty, to appeal to something outside of our practices of using concepts in order to explain the truth and meaning of our statements is a dubious strategy. So when Rorty explicates the background from which he comes to discuss Sellars’ concept of picturing he writes: Many of those who owe their philosophical formation to Heidegger […] see his desire to save this traditional problematic […] as pious nostalgia, further evidence of the dominion of Greece over Germany. They view that desire, and that kind of talk, as a slide back into metaphysics. (Rorty 1988, 216) The traditional problematic Rorty has in mind here is the question of just how we are to understand our cognitive relation to the world. DOI: 10.4324/9781003221364-7

156  Objections and Consequences And in Rorty’s eyes, Sellars’ answer is unsatisfactory for the following reason: Picturing is for Sellars […] the extra dimension which relates social practices to something beyond themselves, and thus recaptures the Greek problematic of humanity’s relation to the non-human (of nomos vs. physis). In Sellars’ case this non-human something is “the world”. […] Many of us whose minds were formed by reading Sellars think of Sellars’ doctrine of picturing as an unfortunate slide back into representationalism. (Rorty 1988, 216) Yet Sellars’ backsliding is tacit, so Rorty holds, as the commitment to representationalism is concealed by the complexity of Sellars’ account of truth across conceptual frameworks. That is why Rorty argues that Sellars has never quite managed to shake off that representationalist craving:1 Sellars is still held captive by a representationalist picture. In this picture, Neanderthal or Aristotelian sentences have meaning — that is, are translatable by us — by virtue of their referring, albeit unperspicuously, to what really exists — viz., the objects referred to in the ideal, Peircian, conceptual system. (Rorty 1988, 219) But it is not just the claim that the Peircian ideal conceptual system can be construed as telling us what really exist by being able to represent the objects as they are, that is problematic. According to Rorty, Sellars’ hidden commitment to a bad kind of representationalism also engenders the further problematic that it invites skepticism, this is so since Sellars cannot give meaning to the idea that any given picture can be more or less adequate: For if Sellars were free of this picture [read: commitment], it would not seem of importance to him to set up the baroque Tractarian apparatus with the aid of which [in SM V §67] he tries to explicate “the concept of a domain of objects which are pictured in one way (less adequate) by one linguistic system, and in another way (more adequate) by another.” [SM V §67] As with all other accounts of meaning which insist on a tie with the world as a condition of meaningfulness, Sellars opens the gates to skepticism. For now he has to give an account of the notion of “more adequate picturing” which will serve as what he calls “an Archimedean [sic]

Objections and Consequences 157 point outside the series of actual and possible beliefs.” [SM V §75] But any such account will lead back to skepticism. (Rorty 1988, 219) Rorty’s interpretation here rests on the claim that Sellars’ account of meaning relies on a tie with the world. However, this would only make sense to object against Sellars on the condition that the concept of ‘picturing’ could be read as having a semantic function, as a notion we can actually use (against this cf. NAO V §48), a notion we could intentionally integrate into our epistemic practices. In Sellars, though, it has only the status of a transcendental postulate, a requirement, as Sellars holds, that needs to be in place for our language to be part of the world it purports to be about (NAO V §61–64). Seen this way, its role is a constitutive one and belongs on the level of principles of interest at the level of analysis we identified above (in Chapter 2) as M1 (concerning unschematised categories), not at the level O (spatiotemporally schematised categories), as Rorty suggests. The skepticism Rorty sees on the horizon for Sellars here has its source in the fact that the concept of picturing is epistemologically idle: For Sellars’ very description of the picturing relation raises doubts of the sort associated with what Putnam has called “metaphysical realism”. We begin to wonder how we could ever know whether our increasing success at predicting and controlling our environment as we moved from Neanderthal through Aristotelian to Newtonian was an index of a non-intentional “matter-of-factual,” relation called “adequate picturing”. (Rorty 1988, 219–220) Rorty sums up the upshot of the problem in an image: ›Perhaps the gods see things otherwise. Perhaps they are amused by seeing us predicting better and better while picturing worse and worse‹ (Rorty 1988, 220). In other words, for Rorty, it is a distinct possibility that correctness and adequacy substantially diverge. This would mean that the criteria of correctness of assertion by which we sanction each others’ linguistic moves bear no systematic relation to the adequacy of picturing by which our language as a whole is characterised. What is peculiar about the position of this skeptic is that it does not cast any doubt on the truth of assertions, for assertions would remain intact. And the point goes deeper for Rorty, precisely because there is just no way that we could ever go and ascertain just how adequately any given language or conceptual scheme pictures the world. Rorty seems to insist that to be able to successfully deal with this skeptical worry; we would have to assume a point of view on our picturing activity that could translate it back into semantically or epistemically

158  Objections and Consequences efficacious information. Lest we achieve that, it has to remain a fact that we can never meaningfully obtain a viable standpoint outside our current conceptual scheme to check its relation to other conceptual schemes.2 Since that view from nowhere is no option, Rorty argues: This sort of skeptical doubt […] can never be resolved. For Sellars himself has to admit that there is no super-language, neutral between the three conceptual schemes just mentioned, in which we can formulate a criterion of adequacy. His own principles force him to agree with the point which Putnam makes against Kripke: that you cannot specify a non-intentional Archimedean tie with the world, a point outside a series of beliefs. For the non-intentional relations you specify will be as theory-relative, as belief-relative, as everything else. (Rorty 1988, 220) This implies that what we would really be doing in assessing the comparative adequacy of a succession of conceptual schemes (CSi, CSj … CSP) were a semantically loaded form of relating it to our current conceptual system and never to anything else. In other words, there’s just no stepping outside the confines of our current language game. And from that pragmatist perspective, Sellars’ obligation to makes sense of the notion of ‘more adequate’ picturing can never be cashed in, other than from a language-game internal place of making claims. Claims, that is, which are subject to our current standards of S-assertibility. 3 So either “CSj pictures more adequately than CSi” just means “CSj is better suited to our needs than CSi”, or it does not. If it does, then we can dispense with the Tractarian apparatus which is supposed to unite all such conceptual systems. If it does not, and if we cannot say anything more about what it does mean, then surely we can forego talk of picturing altogether. As Rosenberg [1980, 186] puts it, talk of correct picturing is “in a sense idle4” because “the sense of such claims of ontological adequacy or absolute correctness is given only in terms of the notion of conceptual schemes and retrospective collective justifiabilities constitutive of the very diachronic process we have been describing”. (Rorty 1988, 220) But with that established, the very conception of picturing as a position outside the series of actual and possible conceptual schemes or structures would collapse into a mere fiction. Such that it would not just be idle, it would dissolve into mere warranted assertibility. Then the objection from pragmatism would amount to recommending itself: The real basis

Objections and Consequences 159 of all our explanatory pursuits in doing philosophy is the analysis of our linguistic practices – with a specific focus on the function that individual claims and the terms they involve have in human practices. Although very radical in its assumptions, this form of pragmatism that takes the notions of social practices and their use of certain terms as basic has gained some traction. Robert Brandom (1994, 2000, 2008, 2011, 2015 and 2016) elaborates this idea into a theory of intentionality operating within the confines of a methodological pragmatism, i.e. ›the view that meaning should be thought of as a theoretical concept and meanings as postulated to explain proprieties of use, that is, of the activities of those who express them‹ (Brandom in Price 2013, 96). This captures the intuition Rorty appeals to when proposing that we only need to describe our practices.5 6.1.1  Rortyian Immanence In light of this criticism, Rorty recommends Sellars accept a strong version of pragmatism, the view that a ›philosophical account of our practices need not take the form of descriptions of our relation to something not ourselves, but need merely describe our practices‹ (Rorty 1988, 221). It would be only natural for Sellars to accept this offer, Rorty argues, because it has already been on the table in an earlier publication of Sellars: The desired “relation to the world” which representationalists fear may be lacking is, Sellars was implying [in SSMB], built into the fact that these are our practices – the practices of real live human beings engaged in causal interaction with the rest of nature. [This amounts to the] claim that reference to the practices of real live people is all the philosophical justification anybody could want for anything, and the only defense against the skeptic anybody needs. (Rorty 1988, 221–222) Thus, the version of immanence which Rorty advocates takes social practices to be the rock bottom of the explanation of our place in the world and of the objects we find in it – underlining that ›[his] own leanings are obviously toward radicalism‹ (Rorty 1988, 227): The question of whether there is anything for philosophers to appeal to save the way we live now, what we do now, how we talk now — anything beyond our own little moment of world-history — is the decisive issue between representationalist and social-practice philosophers of language. […] The alternative to this [practice first] assumption would seem to be [an appeal to] a gift from heaven called

160  Objections and Consequences “clarity of thought” or “powerful analytic techniques” or “critical distance” — a heaven-sent ability to wrench one’s mind free from one’s community’s practices, to turn away from nomos toward physis. (Rorty 1988, 223–224) Of the charges launched by Rorty, three stand out in particular: (1) Backsliding: The charge that Sellars’ concealed commitment to representationalism implies he is still committed to a realist metaphysics. (2) Archimedeanism: The charge that to make sense of the very idea of picturing Sellars needs to assume a view from sideways on, a view outside of language, to substantiate the idea that picturing functions like a vocabulary that is neutral towards the different conceptual schemes and allows us to assess these in turn. (3) Idleness of picturing: The charge that since Sellars cannot give sense to the qualification ‘more adequate picturing’ other than from within our current conceptual scheme the very idea of picturing is idle and collapses into mere S-assertibility and is nothing over and above language-game internal standards of correctness.6 6.1.2  Meeting Rorty’s Challenges The three challenges are interrelated, and together they amount to a rejection of the thesis that the ›ideal, Peircian, conceptual system‹ would yield something like definite and final knowledge of ›the objects referred to‹ (Rorty 1988, 219). Notice that Rorty here introduces the concept of ‘reference’ into his characterisation of Sellars’ position in a way that Sellars would not underwrite. This is a key point because it reveals how Rorty reads Sellars as advocating a kind of epistemology under the heading of ‘picturing’. Picturing, on that construal, would enable us to know Kant’s things as they are in themselves. Inherent in Rorty’s understanding of the very concept of picturing, accordingly, is the thesis that to picture is nothing other than Sellars’ roundabout way of construing our relation to the objects known (albeit future objects) in representationalist terms: To picture is to refer to objects more perspicuously than we currently do in the conceptual structure we presently have. In his review of SM Rorty is upfront about this detail of his interpretation: [T]o say that the language-game which will succeed ours is more adequate than ours is to say that the future “counterparts” of certain past notions (as, e.g., “swirl of particles with such-and-such features” is the counterpart of “brown desk,” and “such- and such an hallucinatory image” is the counterpart of “the Devil”) are such that the sentences

Objections and Consequences 161 which describe them, considered as marks, form better maps of the world. But this way of speaking seems just to postpone the problem: we now want to know in what parameters we are to describe these objects common to all language-games in order to see that they have been better mapped. This looks like an unanswerable riddle, since it is equivalent to asking how the thing-in-itself can be known — how objects can be described which are no more the objects of some given language-game than of any other. What Sellars needs here is a vocabulary which is common to all possible language-games, and which is suitable for formulating criteria of adequacy of mapping. He thinks he has one. (Rorty 1970, 68–69) To Rorty our relation to the objects pictured cannot be construed in other terms than as a relation between a language and the objects it enable us to speak about. What counts as an object is settled in some language-game, while without a language game, there are no objects. Accordingly, all Sellars seems to be doing is to use the concept of picturing to characterise the epistemic activities of future agents in a future language game. In other words, Rorty’s reading of Sellars’ conception of picturing misses the central point that to Sellars picturing is a non-semantic relationship (SM V §58). The fundamental non-relationality or horizontality of all semantic notions debars Sellars from proposing anything like it. So what inspires Rorty to interpret Sellars in these terms? For Rorty, the thought that conceptual systems could differ in their respective adequacy as a whole is unthinkable without a tertium comparationis. And since such a tertium could only be the point of view of an interpreter looking at an object from within a language-game, if picturing is to fulfil that role, it has to fulfil it by being a (meta-)language ‘neutral’ between the object-languages under consideration. This, however, construes the relationship between a given conceptual structure (language) and its capacity to picture not as one between natural-linguistic objects and objects (NAO V §116) but as one where the language under consideration is looked upon from a meta-standpoint which on pain of paradox, cannot be a non-linguistic viewpoint. By arguing this, Rorty skews the point central to Sellars’ construal of picturing as a standpoint outside the series of actual and possible conceptual structures. Sellars does not argue that we could ever read off of the picturing capacity of any one natural linguistic object that it pictures better than any other natural linguistic object. To claim this would be to treat picturing in an instrumental sense as a first-order descriptive use, or as one could put it, to invoke a semantic use of the concept of picturing.

162  Objections and Consequences Sellars makes it difficult to see this, but it is clear that the Peircian conceptual system in its capacity to picture given objects only serves as a theoretical notion that is introduced to allow looking at conceptual schemes from an abstractive engineering perspective (cf. Chapters 5.1– 5.3). This means, the Peircians do not get to read the S-assertibility off of their picturing activity, and in this sense, there cannot be any semantic usage of the concept of picturing something as something. That Rorty has something akin in mind stands out from his appeals to a language that would have to be ‘neutral and looking on the series of conceptual systems from outside’. Thus the first charge of (1) backsliding into a realist metaphysics by adhering to representationalism can be dismissed, insofar as Sellars does not argue that the Peircian community refers to objects by way picturing them. On the contrary, the Peircian community too will be making meaningful claims only along non-relational horizontal lines. And the thought that these – when considered as natural linguistic objects – can be conceived to stand in a more adequate relation to counterpart objects of our objects is a transcendentally identified idea that serves as a regulative ideal.7 Does Sellars really propose an inconsistent (2) Archimedean point of view? On the contrary, he is evidently aware of the problem that in thinking about pictures as of a certain way, as having certain features, we predicate such features from within a normative perspective. Rorty argues that we cannot ever leave such a perspective. Against this, Sellars holds that we can reflectively do so much as productively probe into the most basic structure of our current epistemological situation – on the assumption that this involves abstracting from our current position within a specific conceptual system. It involves reflecting on the uniformities involved in meaningfully speaking a language. And what is involved in thinking the thought of a conceptual structure thereby picturing an object? Sellars argues that it involves an epistemic context in which we are in one way or another affected by the object that gets pictured.8 [T]he concept of a linguistic or conceptual picture requires that the picture be brought about by the objects pictured […], in thinking of pictures as correct or incorrect we are thinking of the uniformities involved as directly or indirectly subject to rules of criticism. (SM V §56) And furthermore, it involves thinking of the uniformities in responding to such affections as yielding a certain pattern. A pattern which we tend to read with semantic or deontic eyes.9 But aside from that irreducibly normative perspective, we can also consider the uniformities as resulting in natural linguistic objects. Given that we abstract from the fact that our assessments of their proprieties, like the (extra-)

Objections and Consequences 163 linguistic objects they concern, ›are subject to rules and principles — are fraught with ought — we abstract from this knowledge in considering them as objects in the natural order‹ (TC 212; cf. O’Shea 2007, 153). This is a point about transcendental methodology insofar Sellars contends that in abstracting from the specific details of any given language game, we can reflect on and articulate which most basic facts necessarily need to be in place for a language to be vertically part of the world it purports semantically and normatively (i.e. horizontally) to be about.10 If we read Sellars’ insistence on it being a fact that any meaningful language pictures as his way of giving flesh to the skeleton of the project of a transcendental linguistics, this enables us to see it as consistent with his interpretation of Kant. This continuity is exhibited by the fact that Sellars adheres to a phenomenalism about objects in the Manifest Image objects while defending a transcendentally demanded realism (cf. Chapter 6.3). After all, pictures are evoked in us, and so are the world stories that thereby get selected. In endorsing these claims, Sellars pays tribute to Kant’s thesis about the passivity characteristic of our affective capacities of being receptive with regard to the sensible world. And the continuity also shows in the fact that Sellars maintains that the conceptual nature of meaning-statements and our endorsements of their semantic assertibility is also the result of our spontaneity. It is a result in that we bring our spontaneity to bear in our responsiveness to the environment, i.e. on the items which we classify to be of a certain kind. But these two commitments are compatible, Sellars maintains, with the thesis that the semantic principles or linguistic rules shaping our conceptual responses are reflected in uniform patterns.11 And it is these patterns that we can consider as objects in nature, as natural linguistic objects. Thus we can step outside of our use of language, but not very far. For some like Rorty that is already too far because he assumes that looking at ‘causal’ features inevitably can only be done with semantic eyes. Sellars discusses a related objection: [A]n objection can be raised to the whole enterprise. For, it might be argued, even if it were made to work, it could not do what I want it to do. For, surely, I have at best indicated how a structure of natural-linguistic objects might correspond, by virtue of certain ‘rules of projection’ to a structure of nonlinguistic objects. But to say that a manifold of linguistic objects correctly pictures a manifold of nonlinguistic objects is no longer to consider them as mere ‘natural-linguistic objects’ — to use your term — but to consider them as linguistic objects proper, and to say that they are true. Thus, instead of finding a mode of ‘correspondence’ other than truth that accompanies [S-assertible] truth in the case of empirical statements, your ‘correspondence’ is simply [S-assertible] truth all over again.

164  Objections and Consequences So the objection. I reply that to say that a linguistic object correctly pictures a nonlinguistic object in the manner described above is not to say that the linguistic object is true, save in that metaphorical sense of ‘true’ in which one geometrical figure can be said to be a ‘true’ projection of another if it is drawn by correctly following the appropriate method of projection. If it is objected that to speak of a linguistic structure as a correct projection is to use normative language and, therefore, to violate the terms of the problem which was to define ‘picturing’ as a relation in rerum natura, the answer is that, while to say of a projection that it is correct is, indeed, to use normative language, the principle which, it will be remembered, I am taking as axiomatic assures us that corresponding to every espoused principle of correctness there is a matter-of-factual uniformity in performance. And it is such uniformities, which link natural-linguistic objects with one other and with the objects of which they are the linguistic projections, that constitute picturing as a relation of matter of fact between objects in the natural order. (NAO V §115–116) Aware of the charge of illegitimately assuming an Archimedean point, Sellars intends to argue is that we can stipulate from within the normative perspective on our use of concepts (1) that such endorsements or espousals of principles of correctness are reflected in uniformities of performance and (2) that such uniformities constitute natural linguistic objects. Rorty’s point was that what Sellars cannot claim is (3) that he has a position up his sleeve, a ›super-language, neutral between the […] conceptual schemes‹ (Rorty 1988, 220). What is so inconsistent to Rorty is that this entails the thought that we could non-normatively make a normative assessment, i.e. make a judgement stating which one of the natural linguistic objects (or conceptual structures as a whole) pictured ‘more or less adequately’ (Rorty 1970, 69). In Sellars’ response to the fictitious objection, we see that he agrees with Rorty about this latter point, that to assess a performance as correct or incorrect is ›no longer to consider them as mere ‘natural-­linguistic objects’‹ (NAO V §114) means switching back into the normative perspective. And Sellars is prepared to acknowledge that our own language game is our hometown when it comes to making claims about the functioning of expressions in related conceptual structures (different language games): [W]e can define a sense in which expressions in a different but related conceptual structure can be said to refer to or denote that which is denoted by expressions in our conceptual structure. (SM V §62)

Objections and Consequences 165 As in the case of [matter-of-factual] truth [i.e. truth qua picture], the importance of this analysis lies in the fact that it permits the extension of epistemic notions12 to conceptual items in a framework which is other than, but related to, the conceptual structure which is embedded in our language as it now stands. In other words, the connection of these epistemic notions with our current conceptual structure (which is necessarily the point of view from which we view the universe) is loosened in a way which makes meaningful the statement that our current conceptual structure is both more adequate than its predecessors and less adequate than certain of its potential successors. (SM V §63) Thus far from being idle (3), the concept of picturing is introduced to allow Sellars to establish the transcendental thought that the statements of a language as a whole can be seen from outside, from the engineering standpoint that abstracts from its normative structure. But the very idea of this viewpoint – considering that the older conceptual structure CSi stands in a different degree of adequacy to the objects it purports to be about than a related but younger conceptual structure CSj might do – comes with an important qualification: Thus the fact that, using the conceptual framework of common sense, we quite properly say, Jones saw that O was red does not commit us to the idea that there is such a thing as O as conceived in the framework of common sense, nor that O is red as redness is conceived in this framework. Jones sees that O is f involves that Jones has a conceptual episode of the • O is f • kind. This includes a component which refers to O, and, assuming that the conceptual structure in question is of the subject-predicate kind, a component by virtue of which it characterizes O as f. That there is no such thing as O as conceived in the framework of common sense, is compatible with the idea that there is such a thing as O as conceived in another framework, thus that of physical theory. (SM V §64, cf. also SM V §95) That is, Sellars grants that although we might extend our stipulations, by introducing the concept of picturing as a transcendental presupposition such that it pertains to any conceptual structure, the only cash we have for giving meaning to the terms in which we think the material contents of the claims we thus consider, as doing the picturing has to come from within our own conceptual structure (CSO). This, however, does not hinder us from thinking the thought – and that is the aim of construing our relation to the world with the help

166  Objections and Consequences of the concept of picturing – of changes in conceptual structures and between conceptual structures and of an accompanying gradation in the concept of truth. And we can do this to the extent that some objects we took to exist in earlier conceptual structures can be shown to not exist at all. The criterion of ontological existence is being pictured: It is a truism that we do not speak a more adequate language than we do. On the other hand, it makes sense to speak of people who speak a more adequate language than we do. The putative concept of a linguistic structure which permits a more adequate picturing of objects than we are able to do raises the question: In which framework are these objects conceived? If in CSO, then how can they be more adequately pictured than they are in CSO, i.e. by its method of projection? How, it might be asked, can a common-sense object be more adequately pictured than in common-sense terms? Are the individual variables we use tied exclusively to the individual senses of our current conceptual structure? Are the predicate variables we use tied exclusively to our conceptual resources? It is obvious that the only cash we have for these variables is to be found in our current conceptual structure, but it is a mistake to think that the substituends for a variable are limited to the constants which are here-now possessions of an instantaneous cross-section of language users. [These considerations] give us a way of speaking which abstracts from those features which differentiate specific conceptual structures, and enables us to form the concept of a domain of objects which are pictured in one way (less adequate) by one linguistic system, and in another way (more adequately) by another. And we can conceive of the former (or less adequate) linguistic system as our current linguistic system. (SM V §65–67) Sellars argues that if we abstract from the features of a specific conceptual structure, we can isolate a sense in which such a conceptual structure (seen from the engineering perspective) pictures its objects (it normatively makes claims about). If we generalise this by adding the thought that sequence of conceptual structures will come to exist throughout time, this allows us to define a sense in which each of these conceptual structures will (a) form not only more or less adequate pictures about the objects we hold (normatively) dear in the current conceptual structure (CSO). But also a sense in which each of these conceptual structures will (b) form more or less adequate pictures of the objects as they will (or would be, or might) be pictured by an ideal Peirceish conceptual structure (CSP).

Objections and Consequences 167 What matters is that this train of thought concerns the conceivability of a sequence of conceptual structures which, when considered with regard to their capacity to picture their objects, can be construed as increasingly more adequate (cf. Chapter 5.3). Notice that although the concepts of “ideal truth” and “what really exists” are defined in terms of a Peirceian conceptual structure they do not require that there ever be a Peirceish community. Peirce himself fell into difficulty because, by not taking into account the dimension of “picturing,” he had no Archimedeian [sic] point outside the series of actual and possible beliefs in terms of which to define the ideal or limit to which members of this series might approximate. (SM V §75) And from this, it stands out that far from holding the Archimedean point of view to be one we can really adopt, Sellars here makes it clear, pace Rorty that in his analysis it enables the thought about an ideal or limit. Sellars here underlines that the concept of picturing functions as a transcendental postulate enabling us to think about the relation between our language and the world in non-semantic terms.13 Its function as a limiting concept in Kant’s sense implies that it can never be used in the sense in which we use ordinary empirical concepts. Here the level of philosophical reflexion on framework features is to be sharply separated from the matter-of-factual level on which we make judgements about empirical objects inside any given framework. In this regard, Rorty is correct in likening the concept of picturing to the limiting concept of the in itself. It is a concept that plays a constitutive role in our theorising about our relation to the world. But in Sellars, this amounts to a radically non-semantic and thereby also non-intentional conception of being related. For the second-order isomorphism, he has in mind is construed to be a matter-of-factual relation, in rerum natura. On that note, we can conclude that Rorty misconstrues the relation between prediction and picturing for the following reason. On his understanding, we cannot but have a semantic and normative perspective on the contents we concern ourselves with, be they empirical or abstract objects applied in theory-building. But that is explicitly denied by Sellars. As soon as we talk about ‘the correctness of a picture’ we are slipping back into the intentional, semantic view, or scorekeeping stance. Yet, Sellars’ conception of picturing is transcendentally required and pushes a realist intuition. Our language as a whole, when considered from the engineering perspective, is a part of the world, exactly because it involves a complicated method of projection. It is because we make predictions and claims inside our language that the language has the empirical shape it does. But the empirical shape in its isomorphic relationship to objects in the vicinity needs to be in place for our language to be part

168  Objections and Consequences of the world. For that reason, it cannot come to happen that we happily predict certain outcomes correctly (by our language-internal standards of correctness) while underneath, we picture less and less adequately. By contrast with what Rorty insinuates, to produce linguistic pictures is not the same as stating matter-of-factual propositions. Rather: The criterion of the correctness of the performance of asserting a basic matter-of-factual proposition is the correctness [read: adequacy] of the proposition qua picture, i.e. the fact that it coincides with the picture [which] the world-cum-language would generate in accordance with the uniformities controlled by the semantical rules of the language. (SM V §57) That is, in the long run, in so far as our claims happen to picture the world adequately, we will get better in predicting. From a transcendental point of view, this can be expressed by inverting the pragmatist intuition that ‘we make the rules around here’ to be the claim that ›the [adequacy] of the picture is not defined in terms of the correctness of a performance but vice versa‹ (SM V §57). When Sellars writes ‘is defined’, he has in mind that our ability to picture the world is what we are answerable to – an ideal we approximate.. This means it is not a criterion we can actually apply in assessing the normative qualities of a language as a whole. To ask: ›How do we know one conceptual structure is preferable to another?‹ is to misunderstand our situation. We have no independent standpoint for addressing this question. The comparative success of a later conceptual structure we ascertain by successfully living in it (cf. Chapter 6.3). Rosenberg is clear about this in one of his last articles: ›Sellarsian picturing‹ wherein he revises his previous judgements (the one Rorty endorsed) and now concludes that ›the concept of picturing per se is the concept of a mode of [correspondence], not a mode of correctness, and so has no independent epistemic function‹ (Rosenberg 2007, 120). Rorty pushes a semantic characterisation of the function of the concept of picturing that has epistemic consequences – picturing just is correctness but in the future CSP. And since Rorty has no use for a distinction of levels of abstraction, he does not recognise the level (of unschematised categories) M1, nor that we can meaningfully ask what is involved in moving from level M1 to level O (of how to think schematisation in accordance with our forms of sensibility). Sellars would have to give the following argument against the skepticism Rorty formulated: The divergence between prediction and picturing can never become a reality because picturing always is a matter of the method of projection constituting our current conceptual structure (cf. SM V §3, 56, 76, 88).14 In so far as languagings are evoked, their double nature (Janus-face) implies that the way we picture is the empirical reflection (qua linguistic

Objections and Consequences 169 performance) of the semantical assertibility norms we adopt in speaking the language in which we make our predictions and observations. By contrast with Rorty’s assumption, it is not the case that any two isolated objects picture one another. Rather, it is because the very idea of thinking our epistemic uptake of an object involves the thought of our being affected by it that we can think the thought of it causing us to respond by way of a language entry move. And such a move, in turn, is reflected in the uniformity of a certain pattern. The pattern, considered as natural-­linguisticobject, is isomorphic to the causal origin of our conceptual response. Thus, if we predict better, we do so because, comparatively speaking, we have a better grasp on the question which of the objects around us (as they are conceived of in CSO) cause us to respond to them. And this goes hand in hand with our picturing them. That is the very idea of there being an isomorphic relation. The relation only obtains to the object that invokes a conceptually shaped response in us, not to any object whatsoever. For that reason, the Gods would, by definition, see us making better predictions only if we thereby pictured with a higher degree of adequacy. According to Rorty, the entire appeal of the Peircean ideal can only be cashed out in epistemic terms. And this makes it clear that he fails to see the transcendental role the notion of picturing plays in Sellars’ account of immanence. On an epistemic reading of ‘criterion’, it would be one that we can actually make use of. And Sellars’ formulation invites that misreading. A reading on which the whole question comes down to ‘Can we give a non-trivial non-circular reason for saying that we ought to adopt the theory which is ‘ideal’ in this sense?’ If we can, then realistic projects of ‘legitimation’ make sense. (Rorty 1979, 84) Yet it is not the case that Sellars is engaged in a project of legitimising our knowledge claims in this sense. Especially not by way of arguing that the concept of picturing can be interpreted in an epistemological key. However, the challenges Rorty formulates constitute a productive misunderstanding in that they allow us to bring into focus how the Kantian background of Sellars thinking ties in with those of Sellars’ commitments that aim beyond Kant.15 6.1.3 The Transcendental Function of the Concept of Picturing The standpoint from which one can envision Sellars to address Rorty’s challenges would be one that combines two of Kant’s with three of Sellars’ insights. According to Kant’s transcendental psychology, the (K1) categorial frame and (K 2) the non-conceptual, passive component

170  Objections and Consequences in our demonstrative reference never change. The forms of thought, no matter how we interpret their deduction into a table of categories, stay the same. The defence we have seen Sellars giving this aspect of Kant’s transcendental psychology amounted to the argument that no matter what content we think, we think it in virtue of classifying a content with regard to its epistemic powers (Chapter 2.1). And doing that, in turn, is nothing other than understanding the subject matter of our thinking to be of some categorial form or other (Chapter 2.2). Thus when Sellars gives the Kantian categories a transcendental role in accounting for the consistency of the concept our epistemic relation to the world as the content of such acts of taking something to be a certain way, he is interpreting ›Kant’s “epistemological turn”‹ to consist in ›the claim that the distinction between epistemic and ontological categories is an illusion‹ (KTE §9; cf. Chapter 1.3, cf. Cumpa 2020 and O’Shea 2021b). And it is an illusion because to say ‘what exists’ is precisely to already be guided in one’s judgements by the rules and meta-rules for forming conceptually articulated contents. What in Kant’s version of the phenomenological fact of our fundamental passivity with regard to the world are our intuitions, in Sellars’ re-interpretation become basic ‘this-suches’ (SM II §33–34). At its foundation, our receptiveness towards the impingements from outside rests on the demonstrative reference component of intuitions because it is the fundamental job of intuitions to present objects to the mind for their consideration (cf. SM V §27). In Sellars, this Kantian insight implies that what the categories bring from above to the shaping of intuitions has its correspondence in what materials demonstrative reference brings to the shaping from below. In this very basic form, the Aristotelian hylomorphism of matter and form is retained in the Kantian conception of an object of experience as that which we come to grasp when cognising synthesised unities in the perceptual act of taking something to be of a certain categorial form, some S to be p (cf. B150–152). The results we need from Sellars are (S1) the non-relationality of our conceptually articulated intentionality (Chapter 4), (S2) the fact that the Manifest Image is unstable and the conceptions of its objects under pressure from rival conceptualisations (Chapter 3.8) and (S3) the methodological claim that to understand our thinking the very thought of our representations’ being intentionally related to something that corresponds to them, we need a transcendental methodology of differentiating levels of abstractions (Chapter 1). With the help of theses K1-2 and S1-3, we can uncover the nature of the presuppositions binding for all immanent knowledge (Erkenntnis). To think ourselves as part of a world, which we can come to have knowledge of, is to think ourselves as simultaneously to be subject of acts of intuiting and of acts of categorically shaping and classifying their very contents into unities we can

Objections and Consequences 171 come to structure conceptually according to the rules governing our conceptual structure (CSO). In a simplified sense, this entails thinking the thought that an individual item affects us and brings conceptualisations about in us (Chapter 3.1–3.6). It entails, in other words, that the categories are schematised in the specific way that is characteristic of the conceptual structure we feel at home in. But in as far as we can conceive of unschematised categories, we can also think that there will be future schematisations. Those are new schematisations in which the very same, though in this regard empty and only schematic, forms of thought, when conceived of as categories, will be applied in novel ways in conceptualising the very same non-conceptual impingements we are delivered with by the senses. Thus in moving from the level M1, the level of considering our intentional relation to an object überhaupt as necessarily subject to the categories, to the material level O at which we perceive a ‘Gegenstand der Erfahrung’ in taking it to be spatio-temporally schematised, we are thinking a schematisation with regard to our present conceptual structure.16 When Kant (10:130–131) asks in the letter to Herz what could ground the reference of that in us which we call representations (Vorstellungen), he poses a question that concerns the justifiability of our intentional relation, our being related at all. If thinking through the presuppositions of our intentional relatedness requires a method of abstraction, this method of abstraction is accompanied by thinking the categories (on O as schematised, on M1 unschematised and finally, on M2 as pure forms of thought) to govern the formation of the corresponding concept of an object to which our intentional reference points. In Kant, this takes the form of thinking a categorically structured individual intuition to relate us to an appearance. In Sellars, in so far as the Manifest Image conception of perceiving the objects around us is unstable and under stress, the more appropriate level of reflecting on the general features of our intentional relation to the world we are a part of is the level of thinking the dispositional capacity of a language to produce matter-of-factual pictures and to make matter-of-factual knowledge claims about the world it is used in.17 In so far as we can think the thought that a language at a given point in time is used with more or less success in predicting what is going to happening, reporting what is happening, and documenting what has happened, we are thinking the thought that a given conceptualisation and schematisation applies to non-linguistic individuals with more or less success. Thinking this thought then is not a matter of assuming a position outside of language. It is a matter of reflecting on what is presupposed in thinking a sequence of conceptual structures to schematise the categories anew. In accordance with the Kantian insights that every conceptual structure comprises a categorial frame (K1) and demonstrative reference (K1), Sellars can argue (S3) that this combines with thinking

172  Objections and Consequences our intentional relation at bottom to require the fundamental commitment to a Gegenstand überhaupt, about the reality, possibility or impossibility nothing is presupposed, at the level M2. But where we begin to give flesh to this notion of there being an object, we need to move down the levels of abstraction and put contentive material to these bare bones. Doing this amounts to schematising the categories as they govern our thinking of something as something to yield knowledge of individuals, which, in so far as the Manifest Image is unstable, gives way to the idea (S2) that corresponding to the concepts of objects as we understand them to be now, there will be counterpart objects and counterpart concepts in future conceptual structures (Chapter 5.3). Sellars’ claim that an equivalent to Kant’s commitment to (K 2) needs to be preserved finds expression in his endorsing the claim that the non-conceptual component of our language-entry move contains a demonstrative reference to that which impinges on us. And as far as its impinging is concerned, this demonstrative reference is invariant across conceptual schemes. Sellars’ interpretation of Kantian intuitions implies that their non-conceptual import (their affecting our receptivity and correspondingly effecting determinate sensory bodily states in us) amounts to claiming that this passive component is invariant, while its conceptualisation and a fortiori schematisation may change. It changes with new ways of looking at the world emerging from the sciences’ methodologically applying analogical concepts to account for the intuitable in experience: For, as I see it, the use of analogy in theoretical science, unlike that in theology, generates new determinate concepts. It does not merely indirectly specify certain unknown attributes by an “analogy of proportion.” One might put this by saying that the conceptual structures of theoretical science give us new ways of schematizing categories. (SM II §49) But to say that new ways of schematising categories emerge (when the theoretical sciences generate new concepts by analogically extending the concepts of objects currently in use) is to move from the level M1, where unschematised categories go with thinking an object überhaupt, to level O, where spatiotemporal schematisation of the categories goes with thinking an ‘Erfahrungsgegenstand’ in space and time of determinate characteristics. When Sellars writes that pictorial adequacy concerns the method of projection (SM V §56), he has the empirical form or shape of a natural-­ linguistic-object in mind, which, as a specific shape is brought about in accordance with the rules of that particular conceptual structure. But in the same sense in which the empirical form of analog recording

Objections and Consequences 173 devices have changed in the 20th century, the shape of natural linguistic objects may change when new inferences and new schematisations come to govern the way in which we project our theories on the world as we intuit it.18 Conceptual structures, on this reading, contain or manifest themselves in increasingly more complex definitions of object-hood and individuality (cf. SM V §97–98). But according to Sellars, it makes sense to understand the shape of the utterances produced in a conceptual structure to comprise not only their logical form but also their empirical form. The empirical form of an utterance here is not the content of empirical statements. In light of the double nature argument, we can put it like this: the empirical form concerns tokenings of sentences considered as natural-linguistic objects – not empirical descriptions of items in the world considered with a semantic eye. For as we recall, languagings are Janus-faced and causally evoked by the objects they are (a) normatively seen to be about and (b) externally seen as isomorphic to (cf. Chapter 5.1).19 If we understand the concept of picturing in this way, we understand it in its function of uncovering the realist intuition inherent in thinking a language to be an epistemic instrument because through its capacity to produce pictures it is causally isomorphic to the world it claims to yield knowledge of. 20 This shows how Sellars hopes to cash in the requirement that a transcendental linguistics uncovers the most general requirements that come with thinking a language to be part of the world it is about. In so far as individual tokenings of utterances, acts in which we perceptually respond by taking something to be the case, are matterof-­factual; they are part of the very same world by which they are evoked (cf. Chapter 5.2). It also shows that Sellars can grant Rorty the point that we cannot ever leave our current conceptual structure and maintain a full-fledged normative perspective on linguistic activity. But what we can do is, by way of abstracting from the current format of our conceptual structure, reflect on its components and its most basic and crucial conditions of possibility – carry out the project of transcendental linguistics. 21 The argument from counterparts supports the thesis that we can think successor concepts to stand in a more adequate picturing relation to the objects we experience here and now. However, a conceptual structure later than ours will not know the world better by virtue of knowing that they are tokening epistemically more correct concepts, having checked such tokening against the concomitant pictorial adequacy. Rather, in a later conceptual structure too the epistemological status of descriptive concepts is constrained by their non-relationality. This implies that the deliberate use of epistemically significant empirical concepts will likewise be bound to be only non-relationally related to further concepts of objects. That is, although the same extra-linguistic objects may evoke conceptual responses, those responses have their conceptual articulation not on the basis of this evoking (i.e. direct imprinting of categories status),

174  Objections and Consequences but because of the inferential patterns they stand in. Here one may ask, does the non-relationality thesis imply that we can never ‘know’ that later conceptual structures picture the world more adequately? And what is the sense of knowing at issue here? A first reply is to say the following: In a later conceptual structure, we may respond to our sensory states in epistemic contexts and project these states outward more reliably (i.e. as seen from within, in accordance with the semantic norms of that conceptual structure). But that does not amount to a self-transparent direct meta-knowledge of the fact that we are grasping the world more directly. Rather, we may entertain the hope that we are more successful in light of comparative improvements regarding our previous attempts. But the cash value remains our evaluation here and now.22 Correctness entails thinking the thought that a statement of a conceptual structure qua picture is (in degrees) adequately isomorphic to an object. To think the thought of progress from CSO to a later conceptual structure is to think the thought of our making more coherent and successful predictions. This involves thinking that a later conceptual structure pictures ‘more adequately’ than a previous conceptual structure. Correctness contains at bottom adequacy. But this is a matter of thinking the presuppositions inherent in our intentional relatedness, not a matter of suggesting that one can adopt a standpoint of normative assessment outside of a language on its epistemic efficaciousness. The claim Sellars thereby advances can be summed up as follows: Even though we may evaluate the correctness of our epistemic success more positively – the pictorial adequacy of our acts of claiming never becomes their correctness never becomes epistemically accessible. That is why Sellars underlines that he endorses a transcendental phenomenalism with regard to the objectual content of any conceptual structure: The thesis I wish to defend, but not ascribe to Kant, though it is very much a “phenomenalism” in the Kantian (rather than Berkeleyian) sense, is that although the world we conceptually represent in experience exists only in actual and obtainable representings of it, we can say, from a transcendental point of view, not only that existence-­initself accounts for this obtainability by virtue of having a certain analogy with the world we represent but also that in principle we, rather than God alone, can provide the cash. (SM II §49) The epistemological idea that we can come to know the world better and better involves thinking the thought that we can improve our grasp of that, which invokes the kind of experiences in us that we de facto have. The immanent nature of this enterprise implies that there is no external guarantee. The standards of correctness we assess our own

Objections and Consequences 175 epistemic success by are products of our own making, but they are also formulated in response to and constrained by a world we never made. So if the process of improving our understanding of the causal origin of our perceptual states has the status of ‘providing the cash’, it is so for immanent reasons. Only we can provide the material explanations by schematising categories in developing new conceptual structures. 23 However, this, too, is nothing that we do intentionally. We can intend to do better science, but if we do is a fact, we can ascertain only ex post. Rather, to think of an individual object as having particular qualities, I am already involving the mathematical categories (e.g. of quantity and quality) and in judging a stone to be heavy I am already involving the dynamical categories (e.g. causality). Which further distinctions go into the acts of judging depend on the specifics of our conceptual structure (cf. SK II §65). In this regard, we can say that the materials which we describe by the concepts of e.g. gold and silver have existed all along, only recently we began to apply the concepts in new ways and use them to build, e.g. computer chips. It is controversial whether we are any closer to these metals, but we have found ways of according them a role in CSO that was not conceivable in earlier CSj or CSi. And while this case concerns the use of empirical concepts, it is a further and open question what is involved in schematising categorial concepts in novel ways, ways that are not bound by the forms of intuition, i.e. time and space (cf. Gomes & Stephenson & Moore forthcoming). A second reply is to acknowledge this immanent conception of the correctness of our knowledge claims but to hold that we can make a further point about pictorial adequacy. This point is based on the thought that picturing relations may not be directly epistemically accessible to us, but that they may nonetheless be understood as fostering our epistemic ends. This opens up the idea of an internal sense of accounting for an increase in a degree of pictorial adequacy. A conceptual structure may function better than another, earlier one. 24 And it is legitimate to attribute this comparative ‘better functioning’ (greater success at tracking objects, prediction outcomes of experiments) to a higher degree of pictorial adequacy. The argument here is the following: For a conceptual structure to function better, it has to picture more adequately. To think the thought of an increase in the degree of adequacy is to think a direction in which the sequence of conceptual structures advances. If there is any meaning in the idea of a Peirceian Ideal Image (CSP), it is a practical meaning coming with the fact that a regulative ideal implies a vector. In the epistemological scenario of concern, our aim is to ‘get around in the world better’. That is, the abstract idea that a Peirceian act of perceiving pictures an object in an ideally adequate manner can only mean that this translates into ideally successful predictions of that object’s behaviour and a body of practical knowledge about what we can do with it and in

176  Objections and Consequences relation to it (cf. Stovall 2022, Koons & Sachs forthcoming). The idea of an increase of pictorial adequacy amounts to the thought that we can rationally, or structurally improve our practical access to the world we encounter, because we can formulate a comparative sense in which we achieve our epistemic ends more successfully. Such a comparative evaluation has to happen for us finite knowers always retrospectively and in contrast to earlier or competing conceptual structures. And as such, it always is an immanent evaluation.

6.2  Objection from Global Expressivism At first glance, Huw Price seems to share Sellars’ methodology of presuppositional analysis and, in a sense, also his understanding of levels of abstraction. In Price’s work, we find a further response to Sellars’ account of picturing and non-relational aboutness which we have not given room yet. This response comes to the fore in his defence a ‘global expressivism’, which he has developed and refined in recent years (Price 2011, 2013, 2016, 2017, and 2020). Two features make Price’s response relevant. First, by contrast with Rorty and Brandom, Price seeks to interpret Sellars’ conception of picturing constructively by arguing that it should be given an expressivist frame and function (Price 2013, 159–166). But not only that, he unfolds an interpretation according to which the concept of picturing can be aligned with his version of anti-representationalism if it is read to be a theoretical notion in play when we describe the environment-tracking dimensions of certain discourses. What is at issue is whether Price understands picturing as an activity we can intentionally commit to. That is if Price subjects picturing to a semantic reading that overlooks its transcendental function. The question is whether Price does what Rosenberg said cannot be done: seek to give picturing an independent epistemic function. Second, and this warrants an extended presentation of his position, like Sellars, Huw Price defends a version of naturalism that distinguishes itself against flat-footed version of metaphysical naturalism according to which the world is made up only by what makes the statements of physics true (Price 2013, 156, 160–166). I want to distinguish object naturalism from a second view of the relevance of science to philosophy. According to this second view, philosophy needs to begin with what science tells us about ourselves. Science tells us that we humans are natural creatures, and if the claims and ambitions of philosophy conflict with this view, then philosophy needs to give way. This is naturalism in the sense of Hume, then, and arguably Nietzsche. I’ll call it subject naturalism. (Price 2013, 5)

Objections and Consequences 177 Paul Redding has identified a similarity between Price’s project of developing a subject naturalism and the question Kant formulates in his letter to Herz (10:130–131) – a similarity in the level of abstraction characteristic of the kind of reflection Price adopts towards the concept of representation: Kantian features indeed seem very close to the surface of Price’s “subject naturalism”. One might say that Kant had adopted a stance toward traditional metaphysicians analogous to that of Price toward the object naturalist, in that it was central to Kant’s philosophical project to bring into question untheorized assumptions about the conception of representation implicit in traditional metaphysics. The centrality of this semantic issue is clear in a letter by Kant during the period of the gestation of the Critique of Pure Reason. In 1772 Kant writes to Marcus Herz that up to that time his philosophy had “lacked something essential, something that in my long metaphysical studies I, as well as others, had failed to pay attention to and that, in fact, constitutes the key to the whole secret of hitherto still obscure metaphysics”. This neglected topic concerned the nature of representation. “I asked myself” Kant goes on, “What is the ground of the relation of that in us which we call ‘representation’ to the object”. More particularly, we might consider Kant’s critical attitude to Locke in the first Critique as analogous to Price’s deflationary, subject naturalist critique of the “object naturalist’s” implicit semantic theory. (Redding 2010, 270–271) I will leave Redding’s claim that Kant is asking a semantic question to the side and read it to mean that Kant and Price share the level of abstraction in formulating a question regarding the purport of representations we find ourselves concerned with. Before we come to assess in how far Redding is justified in taking Price to ask an M1 question, we need to provide the context in which Price raises his objection to object naturalism. Price’s objection entails the claim (1) that Sellars’ loses his grounds for defending his own version of a transcendentally motivated naturalism or realism (Price 2016, 135). And by consequence, it amounts to the claim (2) that Sellars, therefore, should accept a globalised expressivism instead (Price 2015, 13–14; 2016, 124, 135). The two claims combined would imply that Sellars’ search for further vertical aspects in an account of the world-directedness of our intentionality (cf. Chapter 4.4) came to a finish at this point. The bite of Price’s objection consists in the conjecture that Sellars’ project could be assimilated to an expressivist pragmatist position, a viewpoint from which there’s no need to ever leave the horizontal domain. 25 By implication, that would amount to giving up

178  Objections and Consequences the hope for a more nuanced, vertical account of our being in touch with what is conceived of as the extra-linguistic world, independently of how we represent it to be. 26 While one type of reader might be dissatisfied with Sellars’ analysis of the intentionality of our states and expressions, another might be comfortable with it. Huw Price can count as one who is, in fact, so comfortable with Sellars’ account of non-relationality that the real mistake, from this perspective, lies with Sellars’ further vertical ambitions. For such a project now appears to overstretch its entitlement by making claims about the ‘real world’. In its most direct formulation, global expressivism rejects the idea that we are ever entitled to compare ‘objects in nature’. The objection Price raises against Sellars can be put in the form of the question: Why should we follow Sellars in stipulating that we are in a position to make matter-of-factual claims about matter-of-factual relations between natural-linguistic objects and other objects in nature? For an expressivist, this appears like a version of object naturalism, which sneaks realist representationalism in through the backdoor again, in Sellars’ case, by appealing to picturing activity. And worse, Sellars seems in effect to be giving up the critical project of transcendental philosophy based on a transcendental methodology when he claims that we are able to compare natural-linguistic objects in nature to other objects in nature, such that between them a factual and increasingly more adequate isomorphic relation can be known to exist. But has Sellars, in effect, given up on the insight that in the sciences, we collectively develop conceptual structures to handle our experiences? Has he given up on the idea that there is no such thing as being a single-­ minded realist about any one of these conceptual structures? From the perspective of global expressivism it appears so. Since Sellars’ invocation of the concept of nature as characterised by objects standing in matter-­ of-factual relations is all but innocuous. While on the face of it, Sellars’ account accommodates and appreciates the importance of normative phenomena internal to our conceptual scheme, the version of realism he comes to endorse seems to allow only the underlying picturing relation to be real. In the reply, we developed to answer the objections raised by Rorty, we underlined the transcendental function of the conception of picturing, but we left open what exactly this implies for an elaborated understanding of Sellars’ commitment to a version of realism. Insofar as Price does not acknowledge this dimension, he is bound to read Sellars as over-stepping his entitlement and veering towards a metaphysically dogmatic scientific realism (Price 2016, 137). In other words, Price worries that Sellars might ultimately subscribe to a sophisticated kind of reductive object naturalism. According to this position, all those vocabularies would be inferior, which explicate the complex logical and inferential, presuppositional, and scientifically

Objections and Consequences 179 relevant relations among the objects and contents constituting the manifest image phenomena. Modal, mathematical and moral vocabularies would be mere epiphenomena to their causal base.27 Price’s reading of Sellars suggests that if only the causal features of tokenings of utterances were real, 28 this would seem, in effect, to amount to a realist stance towards ›objects in nature‹. 29 This would be a stance Sellars has not entitled himself to. Furthermore, from the angle of a more traditional understanding of the aims of transcendental philosophy, it is doubtful that he can avail himself of such a metaphysically loaded conception of nature without thereby giving up on the critical project altogether (cf. Chapter 6.3). Huw Price’s philosophical project is pertinent here because against such an outlook, he proposes a radical anti-representationalism with a deflationary conception of truth. 30 Price defends his anti-­descriptivism by way of rejecting what he calls the ‘old bifurcation thesis’, the conviction that some areas of discourse are genuinely descriptive some other areas should be given an expressivist reading. What is distinctive of Price’s expressivist methodology is that he claims we should treat all areas of discourse as non-descriptive. This core commitment sets his version of pragmatism apart from related theoretical standpoints such as Simon Blackburn’s (1993, 1998, 2005) quasi-realism about moral facts and Robert Brandom’s (1994, Chapter 8, 2008) inferentialist semantics about empirical matters of fact. Huw Price argues in favour of globalising Blackburn’s method of analysis, that is, for treating all fact-stating vocabularies along expressivist lines. In so far as that can be achieved, it would amount to a thorough rejection of all forms of object naturalism. In its place, Price suggests, we should endorse a kind of subject naturalism: Object naturalism gives science not just centre stage but the whole stage, taking scientific knowledge to be the only knowledge there is (at least in some sense). Subject naturalism suggests that science might properly take a more modest view of its own importance. It imagines a scientific discovery that science is not all there is-that science is just one thing among many that we do with ‘representational’ discourse. (Price 2011, 247) In this regard, Price agrees with Sellars’ non-relational account of the aboutness of descriptive terms (Price 2011, 321; 2013, 38–39; 2015, 315; 2016, 132). And he endorses Sellars’ analysis of truth in terms of semantic assertibility.31 This makes Sellars an attractive candidate for Price’s own program, so relevant in the fact that Price interprets Sellars’ anti-descriptivism as a proto-expressivim to which we only need to add a global expressivism (Price 2013, 2016).

180  Objections and Consequences This entails a methodological recommendation to Sellars: to give up the inquiry for a thorough articulation of the concept of intentionality as a project that puts us in touch with extra-linguistic reality. According to Price, to pursue such a project is to be committed in one way or another to a version of representationalism or bad metaphysics. Thus, the objection from global pragmatism consists in the claim that Sellars does not have any grounds for being a naturalist because if his conception of non-relationality about meaning and truth can be globalised to encompass all regions of discourse, there is no conception of a ‘world’ left that it would be worth striving for. For Price, ›the many things that we do with ‘representational’ discourse‹ are rock bottom in other words: speakers and their usage of concepts. From this, it is clear in what way Price builds on Richard Rorty (1979, 2003), Robert Kraut (1990), and Robert Brandom’s (1994) scorekeeping game. As Brandom points out, they share the following background: [O]ne of the vocabularies I am a local expressivist about is representational vocabulary itself. By this I do not mean deflationism about traditional technical semantic vocabulary: ‘true’, ‘refers’, ‘denotes’ and like cognates. I do in fact endorse a distinctive kind of deflationism about such locutions, understanding them as anaphoric proform-forming operators. In spite of specific differences, generically, this view belongs in a box with Paul Horwich’s […]. (Brandom 2013, 102) On Brandom’s account, this approach has far-reaching consequences that lead him to label his own view as a global anti-­representationalism.32 The difference to Price’s global expressivism lies in the fact that Brandom aims to retain some pragmatic function for representational vocabulary, a move Price is highly critical of, maintaining that language has no downtown (cf. Price 2011, 309). To this, Brandom retorts: I mean something possibly more fundamental. For I am also a certain kind of deflationist about the representational dimension of intentionality itself. […] I offer an account of intentionality as a ‘pragmatically mediated semantic relation’. I do not discuss representational vocabulary [in BSD]. But the account of its expressive role […] is an expressive, deflationary one. (Brandom 2013, 102–103) Despite these technical differences, the appeal of Brandom’s methodological approach lies in treating discursive intentionality as prior in the order of explanation (and thus as a source of deflationism) to the representational dimension of epistemic intentionality. In the background, as Brandom states, is Rorty’s claim that

Objections and Consequences 181 we have been driven to a philosophical impasse when we find ourselves committed to representations characterised by a sort of intrinsic epistemic privilege that is magical in virtue of its supposed intelligibility independently of the role the representings in question play in our actual reason-giving practices. (Brandom 2013, 92) Thus, to be a pragmatist in this sense is to take the scorekeeping practices as rock bottom because they confer onto the relevant concepts all their value, role, and significance. Since Sellars does not adhere to a version of representationalism that would endorse their epistemic privilege in the form castigated by Rorty or Brandom, the question is, rather, whether Price and Sellars share an understanding of the meaning of ‘representation’. Does Price agree with Sellars’ account of intentionality? To assess the validity of Price’s interpretation in this regard, we need to turn to the point where Price engages with Sellars’ work. Price argues for a close similarity between his bifurcation thesis and Sellars’ distinction of truth into truth as S-assertibility and truth qua picture. 6.2.1  Truth as Picture and as S-Assertibility According to Price, Sellars too proposes a deflationary approach to truth, ’that focuses on the role of the so-called equivalence principle‘ (Price 2015, 316): [W]e see that what we have here is the principle of inference: That that snow is white is true entails and is entailed by that snow is white which governs such inferences as That snow is white is true. So, Snow is white. But if the word ‘true’ gets its sense from this type of inference, we must say that, instead of standing for a relation or relational property of statements (or, for that matter, of thoughts), ‘true’ is a sign that something is to be done—for inferring is a doing. (Price 2015, 316 citing TC 206) Price also cites the opening of SM V, in which Sellars argues that to grasp the concept of truth ›essential distinctions must be drawn‹ which requires introducing: [T]he difference between the primary concept of factual truth (truth as correct picture), […] and the generic concept of truth as

182  Objections and Consequences S-assertibility, which involves the quite different mode of correspondence […] in terms of which the ‘correspondence’ statement (i.e. equivalence statement) That 2 plus 2 = 4 is true ↔ 2 plus 2 = 4 is to be understood. (Price 2015, 316 citing SM V §9) Note that Price here quotes a passage in which Sellars uses the misleading phrase ‘correct picture’. That it is misleading can be gleaned from the fact that in the following paragraph, Price takes picturing and S-assertibility to be two kinds of ‘semantic notions’. This interpretation of the picturing relation as fulfilling a semantic function is at the core of Price’s interpretation of picturing as e-representation. Price is not interested in the dependency claim which Sellars makes about the two conceptions of truth (cf. Chapters 5.2.2 and 5.4). Nor does Price acknowledge the fact that in related passages, Sellars underlines that picturing is a non-semantic relation among objects in the real. For Price, Sellars is articulating how different regions of discourse, say two scientific discourses, or common sense discourse and the discourse of specialised scientific experimenting, stand to one another. Price writes that according to his ›diagnosis‹: What is happening here is that a cluster of notions—what we might loosely call the semantic notions—are being pulled in two directions, one inclusive and one exclusive. In these passages, […] Sellars [is] making this point with respect to the notions of ‘descriptive’, ‘fact’, ‘proposition’, and ‘true’ itself. In all these cases, he ends up saying, there’s a generic notion applicable to declarative statements of all kinds, and a local notion applicable much more narrowly—to the matter-of-factual […]. (Price 2015, 316) Price reads Sellars’ as deploying both notions in a horizontal manner, as notions we can will-fully deploy to have them serving semantically deflated epistemic ends. According to Price, we use the generic terms to capture intra-linguistic or normative affairs, and we use narrow, local descriptive terms where we are concerned with matters of fact. In the following, we will expand on the issue and articulate why the supposed ability to use both notions is of concern, and why this is an issue arises out of a semantic reading of picturing. 6.2.2 The Distinction of i-Representation and e-Representation Central to Price’s interpretation of Sellars’ two notions of truth is his own distinction of two types of representations. According to Price, our representational activities fall into two classes, those which we treat as

Objections and Consequences 183 environment-tracking, or ‘e-representations’ and those by virtue of which we monitor and explicate intra-linguistic affairs, so-called i-representations.33 Invoking his interpretation of picturing in semantic terms, Price now makes the move of reading picturing as equivalent34 to his notion of e-representation: [M]y response to this fundamental terminological tension [is] to see it as reflecting the fact that all these notions are trying to serve two quite different masters. [W]e get a much clearer view of the landscape by making this explicit—by recognizing that we have two quite distinct notions or clusters of notions in play, misleadingly being forced together by our failure to recognize the distinction and to modify our terminology accordingly. My terms e-representation and i-representation were my attempt to mark this distinction. This recommendation seems to accord very closely with Sellars’s own conclusions—in particular, with the remarks I have just quoted about the distinction between two notions of truth. Hence it is very natural, from my point of view, to regard Sellars’s account of picturing as an attempt to spell out a particular kind of e-representation, and to regard his account of truth as S-assertibility as a contribution to an understanding of an important kind of i-representation. I see Sellars’s strenuous efforts to insist that these notions of truth are distinct, and ‘belong in different boxes’, as of a piece with my insistence that e- representation and i-representation are not different attempts to get at the same thing, but different attempts to get at different things […]. (Price 2015, 316–317) But how does Price understand the sense in which both get at different things? He reads Sellars in the following way: Sellars puts the difference in terms of the idea that picturing is a natural relation between objects—linguistic items considered as objects, on one side, and objects in our environment, on the other— whereas truth in the generic sense is a ‘pseudo-relation’, to be understood in terms of its inference-supporting role within the language game. I think we can emphasize this distinction even further (as I do in Price et al. 2013: chapter 2), by noting the different theoretical stance we employ in each case. (Price 2015, 317) In effect, Price here underlines that to him picturing is nothing over and above a slightly different perspective on attempts of understanding natural matter-of-factual relations. Relations, that is, which we theorise about from within the standpoint of specialised empirical sciences,

184  Objections and Consequences such as technical engineering or applied physics. 35 This entails that Price interprets Sellars to argue that we use the generic notion of truth, which equates truth with semantic assertibility, when dealing with the correctness of, say, meta-linguistic or intra-linguistic statements, while we use the local notion of truth qua picture to speak about matter-of-factual relations: In the case of the generic notion [of truth as S-assertibility], we are interested in a notion we find in use in ordinary language. To the extent—the very great extent, in my view—that the explanatory, pragmatist approach recommends itself in such a case, our theoretical focus will be on the use of the notion. We will be asking, in effect, ‘What are creatures like us doing when they use this notion? Why do they have it in their language in the first place?’ In the case of picturing, however, our focus, as Sellars himself always stresses, is first-order, matter-of-factual, and highly theoretical. There is no reason whatsoever to imagine that the notion we find ourselves investigating will be in play in folk usage. And our theoretical interest is in the relation in the world [sic!], not in the use of certain terms in ordinary language. (Price 2015, 317) It is not entirely clear what Price means when he claims that in the case of picturing our focus is ‘first-order’. For in the terminology we have developed in reply to Rorty’s concerns, the status of picturing is second order in being a feature that matter-of-factual reporting discourse is postulated to have (cf. NAO V §105, i.e. aside from its semantic features). If Price, instead, construes the function of picturing in semantic, epistemically efficacious terms, it might be because he does not acknowledge the double nature argument. Sellars puts this as follows: The fundamental job of singular first-level matter-of-factual statements is to picture, and hence the fundamental job of referring expressions is to be correlated as simple linguistic objects by matter-­ of-factual relations with single non-linguistic objects. (SM V §26) To the extent that Price does not acknowledge this distinction, he would lose the difference between empirical matter-of-factual statements (e.g. ›This pyramid is red‹) and the fact that as natural-linguistic-object such an utterance bears a second-order isomorphic relation to the object mentioned (cf. Chapter 5.4). Then it would be an open question if Price is justified in assimilating Sellars’ distinction of truth as an adequate picture with e-representations and truth as semantic assertibility with his i-representation. By claiming that the two are nothing over and above his

Objections and Consequences 185 i- and e-representations, Price would omit the crucial qualification that picturing is a non-semantic relation, which enables, transcendentally speaking, the horizontal dimension and its matter-of-factual discourse (cf. SM V §9, 27, 36 and Chapter 5.3). For our purposes, we can leave it open whether Price simply fails to see this or has the resources to sign Sellars up for his version of bifurcation. For Sellars, it is clear that we cannot ever check on the adequacy of our picturing activity since we have no theory-independent access to the fact that our horizontal (s-assertible) claims, in fact, picture. One elegant solution open to Price would be to grant Sellars the in principle inaccessibility of the picturing relation but argue that his e-­representations are just what Sellars needs to account for their immanent significance (Chapter 6.1.3). The knowledge claims we make and by which we purport to track the world should by Price’s lights be called e-representations, on the conditions that they, in fact, enable us to better track objects, make better predictions about their behaviour, etc. From Price’s point of view, it makes sense to conceive of e-representational terms as semantically analysable representations. They are what we subject to our expressivist analysis. An analysis, that is, the methodology of which sounds Rortyian enough: [T]he view that for some interesting topics, the path to philosophical illumination lies not, as others have thought, in an inquiry into the nature of the (apparent) subject matter, but in asking about the distinctive role of the concepts in question — how we come to have such concepts, what roles they play in our lives, and so on. […] The view in question seems appropriately called a kind of pragmatism. (Price 2017, 149) And according to this expressivist methodology, both deflated notions of representation can be relegated to playing their role solely by allowing us to specify aspects of our language game moves. With this, Price, in effect, proposes to translate Sellars’ bifurcation of the notion of truth into a pragmatist key. One that covers both the internal i-representational and the external e-representational sense, such that the term ‘true’ is but an immanent product of the language game: In my view, the most illuminating route to a pragmatic theory of truth is to see it as associated with this kind of in-game externality — as a normative constraint, external to any individual speaker, to which speakers necessarily take themselves to be subject, in playing the game of giving and asking for reasons. It is an ‘in-game’ notion in precisely the sense that winning itself is an in-game notion, in a game such as chess: you don’t understand the notion of winning unless you understand what it is to play the game.

186  Objections and Consequences The second notion of external constraint goes with that of covariance — and hence ‘normal’, ‘intended’ or ‘proper function’ covariance — between a tokening of a representation and an element of an external environment. In this sense, a token fails the constraint if it is a counter-example to the general or intended pattern — a ‘false positive’ or a ‘false negative’. As these terms themselves indicate, it is easy to run this notion together with the in-game norms of truth and falsity (thus confusing in-game answerability for environmental answerability), but this is a mistake. (Price 2013, 38) To which Price, in a footnote, adds an important detail: [O]ne source of the confusion seems to be that the in-game notion has the character of faithfulness to an external realm of facts, which can make it seem like the second notion, if we fail to notice that the realm of facts in question is itself a product of the game. (Price 2013, 38 fn28) And with this, Price makes explicit that from a pragmatist position as he holds it, there cannot be such a realm of facts that is not in one way or another dependent on our uptake of it. Just here, it is an open question if Price captures all the aspects of the concept of picturing or whether he omits its transcendental function in favour of highlighting the kind of vocabularies we need to understand our grasp of the relations we postulate to be picturing the world. And with this, we can briefly return to Redding’s stipulation that there is a similarity between Kant’s and Price’s reflections on the concept of representation. We can affirm that both call into question untheorised assumptions about representational purport. But in as much as Kant also pursues the project of defending a transcendental idealism, one is surely justified in seeing Price’s a posteriori philosophical anthropology as distinct from it. For our argument, the more relevant question in the vicinity is if the distinctions Price establishes can be of help in understanding the role of picturing for Sellars’ account of intentionality. 6.2.3  Discussing Global Expressivism When Price asks: ›Is Sellars ready to set metaphysics aside?‹ (Price 2016, 137), he suggests Sellars should give up on metaphysical naturalism (as the ultimate answer to the ontological question of what there is), and holds that his version of a globalised expressivism reveals all there is to know about nature (Price 2016, 138).36 Price’s suggestion is to treat the m-worlds as ›different but not inferior‹ (Price 2016, 138), urging that ›Sellars needs the metaphysical quietism that Cambridge [has]‹

Objections and Consequences 187 (Price 2016, 138). Price, prepared to take the game metaphor at face value, concludes: What has happened here is that for fact, as for other semantic notions, we have had to recognize that the notion has an inclusive sense and an exclusive sense. In the exclusive or narrow sense, it is a matter of definition that all the facts there are are natural facts (that’s what the narrow notion is). In the inclusive or broad sense, it is immediate—not quite a matter of stipulation, perhaps, but an observation easily made about our language, once the question is in front of us—that this is not the case. Either way, then, there is no interesting metaphysical thesis in the offing. So Sellars’s account of matter-of-factual truth, far from supporting an argument for the kind of bare naturalism […], actually provides us with grounds for denying that there could be such an argument. And this should be no surprise, if we bear in mind how Sellars comes to this point, having given up on the idea that there is any interesting account of descriptive language to be had—any account, that is, that might do the positivists’ work of vindicating the thought that only scientific language is genuinely factual. […] Recall Sellars’s own insistence [in CDCM §79] that ‘once the tautology ‘The world is described by descriptive concepts’ is freed from the idea that the business of all non-logical concepts is to describe, the way is clear to an ungrudging recognition that many expressions which empiricists have relegated to second-class citizenship in discourse are not inferior, just different’ […]. (Price 2015, 318) Price here, on the one hand, underlines to what extent he agrees with Sellars’ non-relationality thesis, according to which the analysis of meaning-statements precisely does not take us down the semantic ladder from words to the world or to their truth-makers. Price invokes a version of Rorty’s anti-Archimedeanism, arguing that the non-relationality thesis should be interpreted to the exclusion of any appeal to game-external facts. If Sellars were, in fact, suggesting such an appeal, Price’s criticism would bite. The question, therefore, turns on whether Sellars’ concept of picturing is a non-semantic notion with a transcendental function, or whether it is identical to Price’s e-representings we produce and with a semantic eye keep score on. One solution to this seeming dilemma is to see Price as articulating the general framework of which we need to give an immanent reading as we do of our grasp of picturing relations. Since we do not have direct epistemic access to picturing relations, we need an account of the kind of representational activities we can, in fact, deploy to explicate the world-tracking dimension of our epistemic practices.37

188  Objections and Consequences In assessing a language game, we can, from the standpoint of philosophical anthropologists, say that its players use certain terms to track objects in their environment.38 On his construal of ‘meta-linguistic’ talk, such a standpoint is one according to which we deploy the notion of something picturing something else in order to account for an objectlevel explanandum: Both notions [i- and e-representation] may be useful, for various theoretical purposes, but we shouldn’t confuse them, and we should keep in mind that they present themselves as objects of enquiry in quite different ways: the in-game notion presents itself (like de re ascriptions) as a feature of our linguistic practice that we want to explain – the term ‘true’ is part of a linguistic explanandum. The other notion is a relation we postulate in the world, as part of a theoretical explanans. (Price 2013, 191) The pertinent question, therefore, is just how Price understands the status of such a theoretical explanans. To be a global expressivist is to treat all vocabulary as serving us in expressing our deontic attitudes. In this reading, Price’s metasemantics explicates what it takes to develop a semantics for the functioning of various different languages games within our current conceptual structure. To stipulate that there will be future conceptual structures with counterpart concepts and counterpart objects is just another one of the language games we may play – in this case, in the philosophy of science. But such a game will never be able to situate itself on a different level than the various other games we play, the terms we choose to use in this or that way. And with this result in view, we have a final reason to reject Redding’s interpretation of the methodological similarity in the approaches of Kant and Price (cf. Chapter 6.2 above): None of the passages we considered supports a reading according to which Price investigates the concept of intentionality in the way Sellars derives from Kant. What they do share, however, is a continued interest in explicating untheorised assumptions in the concept of representational purport and of what it is to take oneself to be in possession of a ‘correct representation’. This stands out from Price’s insistence that an appropriate account of our conceptual practices requires distinguishing between the two types of representation he conceives of as on a par (Chapter 6.2.2). It is interesting to note that Price arrives at this distinction by way of an a posteriori reflection on the most generic characteristics of the language games of the natural sciences and of environment tracking discourses. With regard to its methodological approach, he characterises his philosophical project as a ‘side-ways on philosophical anthropology’. This approach comes with a constraint. The constraint holds that we cannot

Objections and Consequences 189 say much beyond our current understanding of what our relation to the world consist in, simply because all our estimations about our relation to the i-world or e-world are nothing but results downstream from our conceptual practices. This means Price places importance on the notion of fallibility. Our conception of what objects there are is a result of the conceptual articulations we can currently give (cf. Price 2011, 183). The higher-order reflection on the fact that and the ways in which we develop such conceptions of objects within special sciences, according to Price, amounts to a form of philosophical anthropology in the sense that his preferred focus is on what the natural, moral and social sciences reveal to us about ourselves. This is one way of accounting for the expressivist claim that i- and e-representations are not subordinate to, or subsets of one another but that the two amount to complementary forms of voicing claims (cf. Price 2011, 167; 2013, 38–39). And with this, we have all the material in place to show more acutely the status of Price’s terminology in contrast to Sellars’ notion of picturing. According to Sellars, we need the distinction between the order of being and the order of understanding to make sense of the basics of our epistemological situation. E-representations come first in the order of being. They are what allowed us to emerge as perceivers and agents and to evolve into language using creatures, engaging in i-­­representational activity. That is, with e-­­representations, we track items and affairs in the external world, and with i-­­representational vocabulary, we interpret and correct (on a meta-level) our own tracking. But with regard to the order of explanation, natural-­linguisticobjects are brought about only in contexts that are governed by i-representational rules (cf. Chapter 4.2). This was also the point of the principle (in TC 216) that the endorsement of linguistic norms shows in the specifics of how we make claims and is necessarily reflected in uniform (behavioural) patterns. It is reflected in a specific structure (empirical form) that the natural linguistic utterance tokens thereby have (Chapter 5.1). In other words, Sellars is concerned with keeping two directions of dependence separate here. In Sellars’ terminology, this can be expressed as the claim: Unless our perceptual responses to the environment were evoked by the objects they manage to picture (considered abstractly), our predictive control and explanatory success would not advance in the way we judge it to do. While at the same time, we could not produce the natural-linguistic-object tokens we do, in accordance with the method of projection inherent in our current conceptual structure, unless our perceptual responses occurred within the ambience of rules and norms that shape the language that we currently speak. Thus while Price and Sellars both endorse what could be called a broadly pragmatist understanding of the category of object-hood as downstream from our social-inferential practices, Sellars accords the

190  Objections and Consequences matter-of-factual picturing relation a transcendental function as enabling the thought of our world-directedness as a directedness that is in constant progress. For Sellars e-representational activity has a special status which Price is not prepared to underwrite. A further difference comes with a divergence in their philosophical outlook. For methodological reasons, Sellars holds that the science of a transcendental linguistics can advance our knowledge of our fundamental epistemological situation in light of a methodology of abstraction – as reliably bringing out necessary or invariant features of the framework with which we encounter the world. Thus, leaving aside the question whether Price’s interpretation of Sellars’ concept of picturing goes through as an exegetical interpretation, in Sellars’ systematic thinking, the postulate of picturing serves as an independent higher-order explanation of general features that those practices have, which we take to be in the business of world-tracking, of producing e-representations. A final theme in the discussion of Price’s reading of Sellars concerns the status of scientific knowledge.39 However, instead of addressing this question from an expressivist point of view, it is more fruitful to delineate Sellars’ position with regard to the Kantian frame in which he develops his understanding of the world-disclosing dimension of scientific knowledge.

6.3  Consequences for Scientific Realism A question that results from our discussion of the objections from pragmatism and from expressivism concerns the relation between picturing and scientific realism. It stands out that both opposing camps agree with that side of Sellars’ writings which stresses the irreducibility of the normativity of meaning, the functional approach to thoughts and linguistic aboutness, and the meta-linguistic treatments of modalities and abstract entities. Both Rorty and Price, alongside Brandom and McDowell and other students of Sellars seek ways to either navigate around or reject Sellars’ commitment to the claim that the sciences are to be accorded explanatory priority. This is evident from Rorty’s rejection of the idea that there could ever be a privileged language or neutral perspective for saying ‘how things really are’ (Chapter 6.1), and it is likewise evident from Price’s question if Sellars is ready to set metaphysics aside (Chapter 6.2.3). However, if the concept of picturing constitutively plays a transcendental role in Sellars’ account of intentionality, a role which neither of them acknowledges, this opens up a novel space for assessing the relationship between picturing and realism. What specific shape of realism is Sellars ultimately endorsing? One result of our preceding discussion was that picturing does not play an epistemological role but underlies and constitutes the uniformities which all first-level matter-of-factual acts of experiencing are reflected in. According to this understanding, the role of picturing in scientific

Objections and Consequences 191 activity cannot be construed as contributing to a direct knowledge of how things are in themselves. A better way to conceive of the relationship between picturing and realism in Sellars then is to highlight that in his commitment to scientific realism, he remains a transcendental idealist for the reason that he remains a transcendental phenomenalist. This can be explicated as a constellation of commitments that takes the following form. (1) In his epistemology, Sellars advocates that the entities we perceive are transcendentally speaking, ideal, but they are nonetheless entities about which, in the first instance, we have to be a realist. We do have to grant them some reality to the extent that they are capable of affecting us. (2) But if we replace our conceptions of what these entities are and what functions they play as scientific postulates and concepts of objects, we do this in light of a transcendentally demanded realism. We do not do this because the instruments of science carve nature at the joints. (3) That is, we do not adhere to scientific postulates on the basis of the dogma that with the aid of the sciences, we will come to experience things as they are in themselves, independently of a conceptual structure. This can also be put in terms of the negative thesis: The notion of adequacy of picturing never collapses into correctness of assertion. Sellars at no point argues that the degree of adequacy of picturing becomes readable to the scientists of the future, such that they could assess the correctness of their claims against it.40 It makes more sense to interpret Sellars as arguing that a robust notion of relations among items in the space of nature is required to account for the specific shape of our basic passivity with regard to impingements from the world. Such a realism, however, can well be characterised as a transcendentally demanded realism, without holding that Sellars ever leaves the framework of transcendental philosophy. Instead of envisaging a future in which the Scientific Image will, in a whole-sale manner, simply replace the Manifest Image, Sellars argues that an appropriate specification reflected in a commentary in accordance with the limits and powers of scientific knowledge has to be given under the constraint of and within the bounds of transcendental philosophy.41 A look into writings later than EPM or SM confirms a more nuanced account of the understanding of realism he underwrites: If my interpretation of his [Kant’s] views on the nature of reason are essentially correct, then they challenge first those who take the cartesian line — dare I call them ‘rational psychologists’? — to take the scientific image of man seriously. But second, they also challenge “scientific realists”, those who are already committed to the scientific image — perhaps they should […] be called ‘rational physicists’ — to come up with a concept of nature which not only finds a place for reason and the causality of reason (tasks which any naturalist

192  Objections and Consequences will undertake), but also for the autonomy of reason and the reality of the moral point of view. (I §87) To be a realist about the results of scientific theorising, for Sellars then, necessarily goes with an acknowledgement of our freedom as thinkers and as moral agents (cf. ORAV, RAL, OMP and recently Dach 2019; Klemick 2020; Koons 2016, 2019; Wolf & Koons 2016). It means being a realist within the bounds of his project of a transcendental linguistics, of articulating the conditions of possibility of human, finite knowledge. The degree to which Sellars reflects this also stands out from the following self-characterisation: But though I am indeed a Scientific Realist, and think that the domain of basic individuals consists of those which scientific theory will ‘in the long run’ find it necessary to postulate, I also regard the [Manifest Image, i.e.] conceptual framework in terms of which man experienced himself and the world long before the revolution in physics was even a twinkle in the eye of Democritus, as a coherent, delicately articulated whole which it is necessary to understand in its own terms before one can be in a position to determine the precise sense in which it, or part of it, is ‘replaceable’ by the world-picture presented by theoretical science. (SK I 22) Notice that this leaves the precise scopes of ‘theoretical sciences’ open, such that those could, in principle, also encompass the moral and social sciences. Why? Because according to Sellars’ mature understanding of picturing, every framework, and in it every first level matter-of-factual statement to some extent pictures. One consequence of this is that readings of Sellars that make him adhere to a strong reductionism sail past his understanding of empirical knowledge. A further consequence to be drawn from this interpretation is that this self-labelling invites an adjustment of Sellars’ earlier self-characterisation around the end of the 1960s. When in SM II he argued that the ›gulf between appearances and things-in-themselves, though a genuine one, can in principle be bridged‹ (SM II §51), it appeared as if he was advocating direct knowability of things in themselves. In light of our response to objections from Rorty and Price, it is fair to question this outlook. Also, taking on board passages from later writings, it seems that the concepts of knowledge and of intentional directedness for Sellars always remain to be transcendentally understood. One would think that Sellars agrees with Kant in holding that ‘without an account of synthesis, no combining of the manifold in the concept of an object’ is to be had. After all, the concept of an object of experience is

Objections and Consequences 193 always subject categorial schematisations we bring to bear in sensible intuiting. If Sellars has good grounds to argue in favour of a transcendentally motivated realism within the bounds of a transcendental phenomenalism, it is because he aimed beyond Kant’s conception of empirical reality finding the Manifest Image as a framework to be unstable (Chapter 2.4). According to Sellars, Kant unduly restricted the schematisation of the categories to time and space only, and in this, perhaps, was too much a child of his time.42 Against this, Sellars holds that we are in no position to ascertain that the schematisation of the categories could not also take new forms, which changed our concept of empirical reality at its foundation (cf. CC and FMPP). One example is Sellars’ speculative anticipation that the category of substance is in principle replaceable by the category of a pure process.43 However, here Sellars is not doing first-order ontology, although he has been read in this way.44 What he is engaged in is best described as a speculative anticipation of the new schematisation of the categories, understood as pure forms of thought. In this, his project is akin to that of medieval logicians anticipating future developments of scientific knowledge. But, and this is what matters since categories are meta-classificatory rules for generating determinate conceptual contents, his understanding of the very concept of an object of empirical and a fortiori scientific knowledge and of the requirements on human knowability remains committed to a critical, transcendental frame. In light of this, it is also clear that picturing never becomes ‘describing’. This is also evident from the fact that Sellars at no point gives up his criticism of the Myth of the Given. Were he to claim that our picturing activity displays the world as it is in itself, this would amount to a relapse into giveness. It would be an awkward consequence to find Sellars in his later work to fall back behind his earlier insights. For these reasons, it appears sound to conclude that Sellars’ commitment to a form of realism does not entail him dogmatically believing that we can know, in the sense of ‘experience’, the real structures underlying our perceptual experience. Rather, it must mean that we have to think such structures to be in place – but that the reality and nature of these structures is a matter of the distinctions and evaluations which we presuppose as valid and binding for all of us. Nonetheless, there’s always more to be discovered by us finite, fallible beings. This includes anticipating future stages in the progress of the sciences: Thus, to say that theoretical statements are capable of factual truth in the full sense is to say that a stage in the development of scientific theory (including the theory of sentient organisms) is conceivable in which it would be reasonable to abandon mediation by substantive correspondence rules in favor of a direct commerce of the conceptual framework of theory with the world. Such direct commerce exists

194  Objections and Consequences already in limited contexts, and, to the extent that it does exist, theoretical frameworks enjoy in anticipation the first class status which would be theirs in that ‘long run’ in terms of which, according to Peirce, we conceive the scientific enterprise and the ‘truth’ about ‘what really exists,’ which is its formal, final, and efficient cause. (SRII 189) When above, we developed an interpretation of the notion of gradually more adequate degrees of picturing (Chapter 6.1.3), we laid focus on the pragmatic feature of the regulative ideal of a Peircian community.45 A regulative ideal gives us a practical sense of ‘better’, better relative to our need to get around in the world. This is echoed in the conjecture that after all, we were given our perceptual abilities not for the purpose of ontological insight, but to enable us to find our way around in a hostile environment - just as we were given pain to get our hands quickly off hot stoves. (SSOP §89) And yet, we also strive to know the world metaphysically speaking better and better and take ourselves have produced more adequate pictures to the extent that we have greater epistemic, predictive, and explanatory results. It is clear that there remains a tension between Sellars’ commitment to a scientific realism and the fact that we can evaluate our scientific success only immanently. But it is a tension that has been very productive ever since it was set on the scene. On the account that we have developed over the course of this book, Sellars does not defend an ontological project. This is a methodological one of uncovering the presuppositions and requirements requisite for understanding invariant framework features necessary for an immanent conception of knowledge. And by this, he means evolving finite knowledge of an evolving world of which we can meaningfully consider ourselves to be a finite part.

Notes 1. To be precise: Rorty frames this accusation as one he supposes Davidson would come up with. However, since Rorty has made essentially the same point in his prior review of SM (Rorty 1970, 69), nothing depends on that detail. 2. McDowell (1994, 42), suggested to call such a standpoint a point outside of the conceptual purportedly looking on the conceptual as a whole – a view from side-ways-on. See Shapiro (2013) for a defence of Sellars against McDowell’s accusations. 3. Rorty agrees with the Rosenberg of 1980 who, according to Rorty ›has made “pictures more correctly than” mean something like “accepted (for good reasons, in a relatively domination-flee communication situation) later than”‹ (Rorty 1988, 226). To Rorty’s satisfaction, this has been a

Objections and Consequences 195









desirable development in Rosenberg’s thinking who in 1974 ›took [SM V] at face value and developed an account of proto-correlational isomorphisms, Jumblese et al. [And] wanted, at that time, to preserve Sellars’ notion of “one truth about the world” — the one told according to the semantic rules of CSP. But in [1980] [Rosenberg] drops the idea of “one truth” and settles for that of “one world”. Now he says that the fact that a “successor conceptual scheme is more nearly (absolutely) correct than its predecessor consists in its adoption or espousal as a successor being warranted or justified’’ (Rosenberg 1980, 117)‹. According to Rorty, the importance of that could not be overstated, because ›That sentence closes off the skeptical “metaphysical realist” possibility which was left open in both [Sellars’ SM] and [Rosenberg’s 1974 book]. These two books had picturing be a matter-of-factual relation causally independent of social practices and so they left open the possibility that successive schemes might predict better and better while picturing worse and worse. The later Rosenberg precludes this possibility‹ (Rorty 1988, 225–226). On the point Rorty makes here: The threat that picturing might come to be causally independent gives away Rorty’s misunderstanding of the relation between social practices and picturing. Since picturing is to be conceived as the causal shadow of our language use, the objection of causal independence falters (see Chapter 5.3 and O’Shea 2010, 467). 4. To this reading, Rosenberg returns in (2007, 104–127) offering a novel interpretation in which he rejects this previous judgement. Now he reads the concept of picturing as fulfilling a transcendental and in our sense ‘constitutive’ role, not a semantic or epistemological one. 5. In Chapter 6.2 we encounter a form of anti-metaphysical philosophical anthropology Huw Price ends up defending. He presents it as an extension of Rorty’s pragmatism, affirming that ›I remain a card-carrying Rortyian anti-representationalist. For in my view the bad kind of representationalism – the kind I join Rorty in opposing – depends precisely on failing to notice the distinction between two harmless, in fact useful, kinds of [i- and e-]representation (loosely so-called, at least) (Price 2013, 192). 6. This is evident in the conclusion Rorty draws: ›[…] Until Sellars fills out this passage, I think, we must say that his project of giving a sense to “picturing” remains up in the air. Perhaps he should just say that the mere notion of “better picturing” is sufficient to give us the notion of “more adequate” (and thus of truth by correspondence, rather than intra- structural truth by coherence) even if we do not have a vocabulary in terms of which we can isolate the objects pictured in a way that is neutral between language-games. Or perhaps he should just construe “more adequate” in terms of the familiar though complex criteria by reference to which we now say that our science is more adequate than Greek science. But I do not think that he can rest in the position he has adopted‹ (Rorty 1970, 69). 7. Crucially, this makes intelligible that we could be saying true things, in our language, that could be recognised as true by Peirceans, just as we can recognise the Ancient Egyptians as saying truly that the sun rises in the east when uttering something in their language. Because we classify their utterance as falling in the same propositional family as our ‘the sun rises in the east’, in virtue of the way we see the two sentences picturing the world, we can recognise them as having said something true by our lights even though they do not have our language. And so we make it conceivable that we, too, could be saying true things recognised as such by those who come after us, without having to rely on a hypothesised End of Inquiry. Without picturing, Sellars claims, Peirce could only make sense

196  Objections and Consequences













of truth in terms of a Final Language. But with picturing, we can make sense of truth in our language as a possibility even if we don’t know what a Final Language will look like. I am indebted to Preston Stovall for this way of putting things. 8. Of course, Rorty could object here that this kind of reflection is itself nothing but a language game we may decide to play but which we could just as well drop. By which he would want to call into question that there is anything special to be gained from deploying a method of abstraction. From this point of view Sellars would not be in a position to meet Rorty’s challenges but the discussion would end in a stalemate between Rorty’s ‘practice first’ and Sellars ‘presuppositional analysis about intentionality first’. However, looked at closely, the two positions are only prima facie on a par, because Rorty already presupposes a substantial commitment to our being able to refer to worldly practices of language usage, whereas Sellars’ analysis goes deeper in bringing out what is presupposed for that to be possible in the first place. 9. This is what O’Shea has in mind against Huw Price, namely the argument that pattern individuation is always already guided by normative seeing (private conversation). deVries (2012a) makes the same point against ­M illikan contrasting Brandom and Millikan against Sellars. 10. That the transcendental argument concerns what it takes to move from M1 to O also implies that objections of the sort that Fodor and Lepore (2001) raise against Brandom are not applicable here because they take place on a lower level of abstraction, where presuppositions about (object level) representational purport are already assumed to be accounted for. Insofar as Sellars is interested in delineating the credentials of a framework that can explain such purport and contrast it with different forms of purport in rival conceptual structures, his philosophical project can only be adequately criticised on the same level of abstraction, i.e. M1 (cf. Chapter 1.3). 11. Again, this is not a claim about the reducibility of the normative to this natural underside, rather, it is a claim to the effect that to be realised in a world of which we can be a part, linguistic and epistemic activities need to be reflected in uniform patterns. Cf. O’Shea (2007, 158 ff.). 12. Among these Sellars counts: Individual senses, families of individual senses and propositions – and their respective counterparts in related conceptual structures. Cf. SM V §53, 62. 13. A further way of putting this is to distinguish between the matter-of-factual level about which we can make ordinary empirical claims, and a level requisite prior to even having the concept of such a domain of intersubjectively available individual contents. The level of reflexion the concept of picturing belongs on is this latter one. 14. Sellars explains the limiting notion of an ideal method of projection in the following way: ’To deny the physical reality of Cantorian entities one does not need to construe a Cantorian conceptual framework as a useful tool for dealing with a quantized world (cf. Whitehead). One can suppose that the world is continuous in a more Aristotelian sense, and, hence, that though any mesh in terms of which we conceptually cut up the world into objects to be pictured will have a finite grain, it can, however, be replaced, in principle, by a still finer mesh. In this case the concept of an ideally adequate method of projection would be an “idealization” in the sense in which mathematical geometry is an idealization‘. Sellars’ commitment to a form of nominalism prevents him from endorsing sets. (SM V §94 fn25).

Objections and Consequences 197 15. One can read Rorty as saying ‘the Peircian ideal is of our own design; it does not exist, in the same sense in which neither god nor the gods exist’. And, one can imagine Sellars to have had something like this objection in mind when responding: ›Notice that although the concepts of “ideal truth” and “what really exists” are defined in terms of a Peirceian conceptual structure they do not require that there ever be a Peirceish community‹ (SM V §75). 16. Here I side with Haag (2007, Chapters 9 and 23–24), who finds evidence for this distinction of levels of abstraction in Kant’s concept of categories and correspondingly in the concept of an object by interpreting IKTE §44–45 (cf. Chapter 2.1). 17. Since Sellars distinguishes between different languages and the conceptual structures they can be taken to all share, he does not run into the difficulties Davidson raises for philosophical positions adhering to the scheme/ content dogma. Sellars’ criticism of the Myth of the Given entails that he rejects any position that tries to accord an epistemic function to the ‘content’ to which a scheme applies. Rather, Sellars non-relational conception of the aboutness of statements expressing the ‘meaning’ or ‘truth’ of terms shows to what extent he agrees with Davidson. We will refrain from discussing the plausibility of Rorty’s contention that languages and conceptual schemes are to be treated as identical. My reason is methodological in that Rorty ultimately defends a quietism about intentionality, arguing that the project of epistemology should be given up entirely (Rorty 1979, 85). Here it suffices to see that Rorty’s reasons for rejecting the scheme-content distinction do neither apply to Sellars semantics nor his conception of truth as S-assertibility. 18. Thus, when Sellars speaks of the form relevant to a concept, I read him as having ‘categorial form’ in mind, a form that in the here relevant sense is invariant across conceptual structures. Cf. ›The objects which are pictured by a linguistic picture can thus be genuinely extra-linguistic (though, of course, linguistic episodes as items in rerum natura can also be pictured). The concepts of these objects are, of course, relative to a conceptual scheme, but the form of these concepts is not‹ (SM V §61). 19. Sellars here is careful to avoid speaking of ‘objects’ at this point and instead resorts to the more general and indefinite term ‘happenings’. Cf. ’That languagings evoked (in contexts) by happenings of certain kinds is a causal fact which is nevertheless essential to their conceptual character‘ (NAO V §69). 20. In other words, the commitment to realism in the background here is one to the effect that Sellars’ transcendental methodology demands a conception of ‘world’ as the entirety of objects invoking pictures on the basis of a method of projection. This, however, is precisely not the conception of world that would realise a T-schema sentence. A ‘world’ is always what a conceptual structure is about, but also that, in which the conceptual structure as a whole is realised (qua uniformities of patterns) and of which it therefore is a part. From these distinctions it stands out that Sellars commitment to a transcendentally demanded realism is not a metaphysical realism. For neither does picturing have an epistemological function, nor does adequacy ever collapse into correctness of assertion. 21. Insofar as not all inner episodes are linguistic episodes, but only some of them, the project of a transcendental linguistics is but one if central area within the larger project of transcendental psychology. cf. MEV, and Haag (2007, Chapters 1–2, 2012b).

198  Objections and Consequences 22. In SM V §53, Sellars underlines this in the following way: ›[H]owever many sophisticated senses of ‘true’ may be introduced, and however important they may be, the connection of truth with our current conceptual structure remains essential, for the cash value of S-assertibility is assertion by us hic et nunc‹.   On this, see Christias (2018) and Matsui (2020, 15–17) who terms this Sellarsian commitment ‘ideal successor externalism’ and contrasts Sellars’ position with Putnam’s. 23. My interpretation here aligns with what O’Shea (2011) in debate with Rottschaefer (2011a, 2011b) has called Sellars’ middle way. A way according to which the formal criteria for our intentional relations to worldly contents are already intact and remain relatively stable. Insofar as these are the forms of thought governing intentional relatedness to any subject whatsoever, they also govern the postulational methodologies of moral and natural sciences. However, according to O’Shea (2011, 334), Sellars has strong leanings towards a view that takes the material to be under constant revision. This is a matter of counterpart concepts in terms of which future conceptual structures carve up the material content of concern to it. Yet here, O’Shea adds – setting his interpretation aside from Rottschaeffer’s right-wing scientific realism – one needs to be clear about the fact that Sellars holds our concepts of persons, thought and intention to be exempt from this process of revision (ibid.). On this I agree with O’Shea while reading the significance of Sellars’ self-labelling as a scientific realist slightly differently (but see Haag 2007, 414 for yet a different take). As this concerns my interpretation of the bearing the role of picturing has here, I pick this up in Chapter 6.3. 24. More generally speaking, this can be extended to cases in which Einstein’s physics can explain in the sense of predicting outcomes that Newtonian physics is unable to incorporate or account for. 25. Price’s argument for why we never have to leave the horizontal domain, combines the pragmatist credo ‘practice first’ with the thesis that functional analysis of concept use suffices to account for the positive role of the concept of matter-of-factual truth. Truth, in Price’s hands, thereby becomes ‘convenient friction’. A friction which results from a ›special purpose disposition [we have …] to disapprove of speakers with whom we disagree. This disposition is the mark of the [truth] norm‹ (Price 2011, 175). That this normative friction is a horizontal matter only comes out in the following passage: ›To use a Rylean metaphor, my view is thus that truth supplies factual dialogue with its essential esprit de corps. As the metaphor is meant to suggest, what matters is that speakers think that there is such a norm — that they take themselves to be governed by it — not that their view be somehow confirmed by science or metaphysics‹ (Price 2011, 165). 26. That it also amounts to a version of quietism comes out in Price (2015). Price there confronts McDowell’s conceptions of quietism, second nature, and therapy urging him to accept his version of philosophical anthropology. 27. Notice how problematic this position would be. To the object naturalist they would be inferior, even though modal and mathematical notions would be needed to explicate and discover causal relations. Cf. Price (2013, 18 fn8). 28. According to Price this would mean ‘real’, in the only sense that the object naturalist allows. 29. Price (2016, 138) puts this by saying: ›In other words, I think that Sellars should accept that mathematical facts, moral facts, modal facts, and the like, are “not inferior, just different”‹. A recommendation which in light of CDCM §79 is odd to make (cf. Chapter 6.2.3).

Objections and Consequences 199 30. For my argument, Price’s divergence from related deflationary accounts of truth does not play as essential a role as does his use-based account of its pragmatic function. The reason is that on the one hand Price takes his approach to agree with ›familiar disquotationalist minimalists such as Quine […] and Paul Horwich [… who hold that] truth is not a substantial property, about the nature of which there is an interesting philosophical issue‹ and although, like them, he thinks ›that the right approach to truth is to investigate its function in human discourse — to ask what difference it makes to us to have such a concept‹. What matters is that on the other hand, Price stresses that on his view, a pragmatist can do more: ›Unlike such minimalists, however, I don’t think the right answer to this question is that truth is merely a grammatical device for disquotation. I think that it has a far more important function, which requires that it be the expression of a norm. But like other minimalists, again, I think that there is no further question of interest to philosophy, once the question about function has been answered‹ (Price 2011, 166–167). For further discussion of the question in how far Sellars’ account of truth as S-assertibility poses a challenge to deflationary accounts of the function of truth talk, see Shapiro (2014b). 31. Cf. Price (2016, 134–135). See also Price (2003) and Rorty & Price (2010) for a defence along non-relational lines which Price develops building on the work of Richard Rorty. 32. Cf. also Brandom (2011, 190–219) and for an extended reconstruction of the differences and similarities between the positions of Price and Brandom see Bacon (2012, Chapter 7) and Loeffler (2017, Chapter 7). For a first encounter see Price (2008). 33. When pressed to the hilt, Price admits that on his view, e-representations are representations we treat as reflecting our understanding of our environment, while at bottom, their status is really the same as that of i-representations. Metaphysically speaking: neither of them is privileged in disclosing ‘the world’ to us, both are merely ways of presenting our understanding to ourselves, ways of voicing our de-re-commitments. But they possess their status due to our evaluating them in a specific way, not in virtue of a reference relation to an extralinguistic nature in any metaphysical sense. Price is anxious to stress that we fare well in thinking of them as distinct, such that neither of them is reducible to the other (Price 2013, 38–39). But that alone does not prevent him from arguing that both share the meta-property of being radically non-representational notions. He clarifies this in his reply to Brandom: ›In one sense, I find it easy to agree with Brandom at this point. Indeed, […] far from rejecting representation, as Rorty does, I now recognised not one but two varieties of it! However, I also noted that that somewhat tongue-in-cheek concession did not imply that I did not remain a Rortyian anti-representationalist in a deeper sense. My faithfulness to Rorty turns on the fact that neither of my two notions of representation (i-representation and e-representation) is what traditional representationalist had in mind – on the contrary, I think, the traditional view rests on confusing them‹ (Price 2013, 186). 34. In fact, in the following passage makes an even stronger claim according to which picturing is ›a kind of e-representation‹. The issue one can have with this turns on the question, whether picturing itself is a (semantic) representational relation, or whether it is a matter-of-factual relation enabling representational relations. See Chapter 6.3 for a discussion of this point. See also the related discussion in Chapter 5.3 and Chapter 6.1.3 about the status of the very idea of an ‘increase in degrees of pictorial adequacy’.

200  Objections and Consequences 35. From this it stands out that Price construes it as if this was all part of a natural science question – that is, he has no use for the transcendental line of thought Sellars pursues in introduction of the concept of picturing. 36. On this cf. Kraut (2016), who argues that Price’s objection is misgiven in that Sellars’ aims were never metaphysical in the sense Price retains. Rather, Sellars’ account of abstract entities, for instance, has a much more pragmatist shape than Price acknowledges, which ›constitutes a theory of what such entities are: a theory of what is going on in traditional metaphysical discourse‹ (Kraut 2016, 70). That is, with regard to abstract entities, Kraut finds Sellars too endorses a practice first approach. 37. A number of recent interpretations of Sellars have failed to make this distinction and read ‘to picture’ in the sense of ‘to describe’, cf. Reider (2012), Rouse (2015), Koch (2011, 2016 Chapters 4 and 7). Cf. also the critical discussion of Millikan’s theoretical extension within a teleosemantic framework in deVries (2012a). 38. Against McDowell’s contentions Price (2015) defends a version of the claim that one can envision such an innocuous anthropological side-stepping. 39. That Price’s reading is largely confined to Sellars’ writings between 1956 and 1967 is perhaps reason enough to leave the expressivist perspective to the side. The full flavour of Sellars’ commitment to a transcendental methodology comes to out in his writings from the mid-1960s onwards. In what follows we turn to addressing the consequences for an assessment of Sellars’ commitment to realism by widening our focus to include later writings. 40. But see Koch (2011, 2016, 2018), for a defence of such a reading. 41. A similar point is made by O’Shea in Pereplyotchik & Barnbaum (2016, 237–238ff.) against Brandom and deVries. 42. Cf. Haag (2007, Chapter 9) for a reading that takes issue with Sellars’ criticism of Kant here. 43. Against this, Brassier (2013, 112–113) has articulated a reading, that places the developments within the theoretical sciences in narrower, Manifest Image, confines. According to him ’Pure processes are postulated at the metacategorial level in order to explain the covariation between patterns of representings and patterns of represented objects. Yet this postulation is perfectly in keeping with Sellars’s [methodology]. It serves as a model that will be necessarily transformed in the course of its deployment by future empirical science. In this regard, Sellars[…] is critical rather than dogmatic precisely insofar as it retains a role for a priori philosophical theorising. However, the ontological categories first catalogued and then postulated by philosophers are constrained by their explanatory role relative to empirical investigation and hence necessarily subject to future empirical revision‘. Cf. also Cumpa (2020) for a discussion of developments in category. 44. Cf. for instance Seibt (1990, 2016) and Rouse (2015). Cf. Landy (2019) for a recent shot at reconstructing the argument of the Carus lectures. 45. The idealisation inherent in this foreshadowing stands out from passages in which Sellars’ firmly locates such success in the future, e.g. TE 453: ›As I see it, then, substantive correspondence rules are anticipations of definitions which it would be inappropriate to implement in developing science, but the implementation of which in an ideal state of scientific knowledge would be the achieving of a unified vision of the world in which the methodologically important dualism of observation and theoretical frameworks would be transcended, and the world of theory and the world of observation would be one‹.

Conclusion

Retrospect This book set out by introducing the problem of intentionality as a problem central to a range of intersecting philosophical domains. Over the course of our discussion, we saw that Sellars’ strategy for addressing this problem required him to defend his position on issues in the philosophy of mind, in category theory, in the philosophy of language, and in what some call the metaphysics of epistemology. With regard to the particular claims he defends and especially with regard to the question of how he defends his claims, it has become clear that Sellars does not follow one single method but to subsume various methods within the overarching frame of a transcendental analysis. However, we also saw him to be consistent in defending his views within a transcendental frame. Accordingly, our project of reconstructing Sellars’ positive account of intentionality was intended to take the shape of a transcendental argument. It was designed to cash out Sellars’ overall position as a transcendental phenomenalism which, at its presuppositional basis, combined with a transcendentally motivated realism. In looking back, we offered a positive answer to the question: Is Sellars a Kantian? But our discussion required us to further nuance our response in important details. In this, the guiding criterion for the assessment of his position has been whether and to what extent Sellars develops answers that Kant could not have given, answers pointing beyond Kant’s immanent conception of objective knowledge. This required broaching a novel path in Sellars scholarship. Instead of working within received distinctions of left-wing versus right-wing Sellarsian questions (as coined by Rorty 1979), or instead of solely working in the terminology of opposing Manifest Image with Scientific Image concerns, the solutions we developed took their cues from details uncovered in a systematic reconstruction of Sellars’ Kant exegesis and tried to show where Sellars saw reason to reject core Kantian conclusions. The tactical decision of basing our approach on this particular dimension in Sellars’ work came with a benefit. It required detailing the relationship DOI: 10.4324/9781003221364-8

202  Conclusion between Sellars’ conception of philosophical methodology in relation to the role of framework analysis. Chapter 1 introduced the method of analysis as a method Sellars finds in Kant’s writing about epistemology and critical philosophy. This provided the occasion to address the second desideratum identified in the introduction, the need for an explicit positioning of Sellars’ transcendental methodology on the map of received philosophical methods. Sellars’ methodology of distinguishing levels of abstraction – as a decidedly normative project – was delineated against the more traditional conception of descriptive metaphysics. One result of this reconstruction was that Sellars’ method amounts to a transcendental method precisely in substantiating the analytic truth that objects of empirical knowledge conform to logically synthetic universal principles. To that end, the presuppositional analysis was introduced as a procedure for identifying necessary pre-conditions and presuppositions of our intentional relation to the objects we encounter in acts of representing. Chapter 2 articulated further aspects of Sellars’ commitment to Kant’s critical project by defending Kant’s transcendental psychology. Sellars’ affirmative reconstruction of this central part of Kant’s project was shown to take the shape of a transcendental linguistics. This allowed us to make explicit the requirements which, according to this transcendental methodology of analysis, need to be in place to say what it is for a consciousness to be characterised by intentionality. In a first step this involved systematically reconstructing Sellars’ account of the Kantian categories in terms of meta-classificatory rules for generating contents. This also involved applying the method of abstraction introduced earlier to the analysis of categorial aspects of matter-of-factual judgments. The understanding of what function the Kantian categories have was put to the test by considering two objections. In a second step, we systematically reconstructed the role categories for mental acts play in Sellars’ understanding of the concept of an object. If an object, in the most general sense, is what we purport to be intentionally related to, it matters in what categorial terms we think the contents which we take to be perceptually available to us, and which we take to exist independently of our representing them such. One result was that Sellars by and large agrees with Kant’s treatment of the difference between subjective and objective successiveness. Another result was the finding that the concept of an object and the concept of an act of perceptually taking it to be a certain way are interdependent. Articulating what this interdependence implies for the nature of our intentional access involved discussing the role of synthesis in perceptual processes. This required clarifying the respective roles of conceptual and sensory components and how they interact in empirical synthesis. One conclusion we drew there was that Sellars’ understanding of the ontological status of objects is deeply Kantian, in that the deeper thrust

Conclusion 203 of Kant’s idealism contains an interdependence thesis about the self and the world. It meant situating the understanding of objects of experience in terms of the category of actuality, between merely imagined contents and things in themselves. Chapter 3 presented Sellars’ unified theory of perceptual experience with the aim of assembling all the requisite elements involved in perceiving objects as actual. This meant accounting for the role of the in itself, the status of sense impressions, and the influence of image models on the constitution of the contents of intuitions. Most importantly, this required explicating the demonstrative component of intuitions as an invariant feature of our intentional relation to a world causally affecting us and to which we also fruitfully respond. A subsequent discussion of the interaction of image-models and empirical schemata in the production of three-dimensional object perceptions allowed us to detail the interplay of elements and dynamics in Sellars’ account of object perception. We rounded up this reconstruction by clarifying the distinction between two types of direct reference that come with Sellars’ separation of unconscious from conscious aspects of intuitions. In this way, we foreshadowed a third kind of matter-of-factual directedness between objects and discursively articulated perceptual responses. One central result was a brief presentation of Sellars’ dissatisfaction with the Kantian account he builds on. Kant’s theory of experience leads to a problematic transcendental idealism, in that the empirical realism it contains is a conception with an internal tension. From a transcendental point of view, Kant’s immanent account of the kind of knowledge available in the Manifest Image conception of objects and persons means knowing only appearances. While Sellars agrees with this result, from his position this result is untenable because the Manifest Image as a framework of experiencing the world is itself evolving and therefore inherently unstable. This was summed up in the formulation of the instability thesis. If Kant’s conception of empirically real objects is unstable for reasons internal to the framework of common sense, this implies a motivation for Sellars to formulate an alternative account of the notion of object-hood and of intentionality. Chapter 4 presented Sellars’ own non-relational conception of intentionality and semantic aboutness. This was introduced as a theory designed to be able to do justice to the aforementioned instability and to be able to account for conceptual change. To this end, the meta-linguistic theory of functional classification and linguistic roles was introduced to give a general account of what it is to be intentionally related to the content of one’s acts of representing something to be the case. However, the nature of this theoretical approach to intentionality left a certain type of world-directed expectation unanswered. Chapter 5 took up the discussion right at this point and, in response, reconstructed three arguments on the basis of which Sellars’ overall

204  Conclusion phenomenalism was interpreted as a consequence of his transcendental perspective on the project of a transcendental psychology: Namely, giving an immanent account that shows how the idea of a mind that gains knowledge of a world it is part of can be made coherent. The argument from double nature was introduced and discussed as an argument accounting for the empirical form of intentional phenomena. The argument from immanence was introduced as an argument for understanding our relation to the world we live in as resting on a conception of acts of judgment. This conception has us being confined to a transcendental understanding of the concept of nature as that which occasions our perceptual responses but is also the target of our knowledge claims about it. The aim of giving an explicit reconstruction of the argument from counterparts was to show in what sense Sellars thinks he has a conception that allows distinguishing between grades of matter-of-factual truths. The relevance of reconstructing these three arguments consists in the fact that only on their basis important nuances in the concept of intentionality came to light. For, according to Sellars, our understanding of this concept requires not only semantic but also non-semantic empirical aspects of contents. This involves relating not just acts to objects of the same frame, but a sense in which frames as wholes can be superseded by later frames. This enabled an understanding of the arc of Sellars’ transcendental argument for a particular understanding of the concept of intentionality according to which the transcendental character of our knowledge can be made explicit. Knowledge is what results between the two invariant points, the formal categorial and the material demonstrative. It is what is generated as material content by the theoretical structures we develop collectively. In cases where we anticipate future forms of scientific research, this development, in a different characterisation, amounts to schematising the categories in new ways. The categories are conceived of as meta-conceptual rules for the classification of contents, according to their respective logical and epistemic powers. This amounts to a process that would force us to give up the Kantian conception of empirical objects. It is in this sense that Sellars’ conception of counterparts and degrees of adequacy in matter-of-factual truths across conceptual structures makes possible envisaging development in the validity of knowledge claims, a form of reflection that is unavailable to Kant. Chapter 6 can be read as an attempt to meet the third desideratum specified in the introduction, namely, to contribute to a re-evaluation of the relationship between Sellars’ nominalistic, expressivist leanings, and his commitments to a scientific form of realism. This was carried out by presenting two objections to the overall account of transcendental phenomenalism developed in the previous chapter. Our discussion of the objection from pragmatism and from expressivism revealed a serious tension in Sellars’ conception of grades of truths. Our diagnosis found

Conclusion 205 that this tension can only be resolved by a reading stressing the immanent and thoroughly pragmatic character of assessments of scientific progress. That an Archimedean point of view from which to evaluate the correctness of our claims of knowledge is unavailable can only mean that we assess a given region of theory-production from the viewpoint of a further region. Thus, our articulation of the material contents we purport to be intentionally related to always depends on the categorial-inferential features of the framework we live and theorise in. And it is clear that we can challenge singled-out areas and claims of related frameworks, and even of our own framework, but not all at once.

Outlook In rounding up our investigation, we will now briefly touch on desiderata for further research which have come into view based on the above results. One theme we largely did not discuss is Sellars’ interpretation of the distinction between primary and secondary qualities. Such a discussion would surely belong in an extended reconstruction of the instability thesis, because Sellars argues in favour of this thesis on the basis of his criticism of Kant’s conception of extension and colour (cf. Rosenberg 2007, Chapter 9, Rosenthal 2016, and Landy 2019). A related and perhaps slightly more demanding desideratum would be to address the question of what one would have to do to carry out a reduction of normative and qualitative phenomena. This is a strand in Sellars’ overall philosophical ambition that we also left to the side. It is the topic of considerations about a complex double commitment on Sellars’ part (PSIM §16), holding that the normative domain of persons and intentions is in one regard irreducible and in another causally reducible (cf. MMB and SMBB, and O’Shea’s (2007, Chapter 1). Of concern here are not just questions pertaining to the ultimate ontological home of intentions, as explainable in light of the technical notion of we-intentions (cf. Koons 2019, Dach 2021, Klemick 2020), but also questions concerning the relocating of qualia, of qualitative phenomena more generally, in terms of Sellars’ later developments in his theory of sensation (cf. Haag 2014). A project to this effect would have to live up to the challenges of cashing in the notion of the in-principle-reducibility of all phenomena to the laws of micro-physical theory. The idea that the laws of the natural sciences can be stated in the language of physics has long been under pressure (cf. Oppenheim & Putnam 1958; Nagel 1961; Dupré 1995; Fodor 1974; Wimsatt 1976, 1994, 2007; Brooks et al. 2021). The question points beyond the confines of the philosophy of science, especially when aiming to assess the case of Sellars’ later philosophical writings (such as FMPP, TTP, and TTC) as an endeavour that is inseparably intertwined with his category theory in terms of pure processes (cf. O’Shea 2021b.). Therefore,

206  Conclusion a further desideratum would be to investigate the relationship between Sellars’ category-theoretic project and the more recent developments in metaphysical category-theory (cf. Thomasson 2018; Cumpa 2020 for a survey, and Cumpa 2021 for recent contributions to this field). This is also of interest as Sellars, in his later works, begins to criticise Kant for having held a much too Cartesian, dualistic conception of the relationship between thought and embodiment. This combines with him urging that ‘Aristotle had been along sounder lines than Kant’, and should, therefore, be considered to be the philosopher of the Manifest image (MP §60). The resurgent interest in Hegel’s reception and transformation of Aristotle’s categories and logic has begun to be of interest to the younger generation of Sellars scholars. Such an investigation of the relationship between Hegel’s and Sellars’ category theories would throw new light on Sellars as a transcendental philosopher in the German idealist tradition (cf. deVries 2020b; Corti 2019; Giladi & Sachs 2019; Redding 2007, 2012, 2020). A further desideratum flowing out of the two latter chapters concerns the question of how Sellars construes the relationship between practical and theoretical reason and agency. If regulative ideals are projective, orientating rational constraints, but also exert a pragmatic guidance function in that they direct us to lead our theoretical lives towards a certain end point (cf. KNDW and OAFP), this commends investigating the relationship of Sellars’ practical philosophy to his transcendental linguistics (cf. Koons 2016, 2019; Brassier 2018; Hicks 2020). A final desideratum invites responding to the overall methodological program of this book. In recent years, Sellars’ earlier publications, prior to the 1956 essay EPM, have begun to receive more attention. It is in this ambience of research that the thesis was advanced that Sellars underwent a meta-philosophical shift, notably away from the formal project of a pure pragmatics to a position that is more in line with the post-linguistic-turn mentality. Does this imply (as e.g. Olen 2016; Corti 2021b and Stefanie Dach, in personal correspondence, contend) that Sellars’ commitment to a transcendental methodology in accounting for intentionality is a relatively novel side to his philosophical profile? This would mandate researching the exact nature of the influence Carnap and Wittgenstein exerted on Sellars’ philosophical development (cf. Brandt 2011, 2019; Gomes 2017; Miller 2018), to make clear at what point exactly Sellars began thinking and working in a Kantian methodological frame.

References

Works by Wilfrid Sellars BBK  (1960). “Being and Being Known”, Proceedings of the American Catholic Philosophical Association, 34, 28–49. Reprinted in ISR and Science, Perception and Reality (pp. 41–59). London: Routledge & Kegan Paul. Reissued by Atascadero, CA: Ridgeview, 1991. CC  (1973). “Conceptual Change”, in P. Maynard and G. Pearce (eds.), Conceptual Change (pp. 77–93). Dordrecht: D. Reidel. Reprinted in Essays in Philosophy and its History (pp. 172–188). Dordrecht: D. Reidel. CDCM  (1957). “Counterfactuals, Dispositions, and the Causal Modalities”, in H. Feigl, M. Scriven, & G. Maxwell (eds.), Minnesota Studies in the Philosophy of Science (Vol. II, pp. 225–308). Minneapolis, MN: University of Minnesota Press. CIL (1948). “Concepts as Involving Laws and Inconceivable without Them”, Philosophy of Science, 15, 287–315. EPH (1974). Essays in Philosophy and Its History, Dordrecht: D. Reidel. EPM (1956). “Empiricism and the Philosophy of Mind”, in H. Feigl, & M. Scriven (eds.), Minnesota Studies in the Philosophy of Science (Vol. I, pp. 253–329). Minneapolis, MN: University of Minnesota Press. (Originally presented at the University of London Special Lectures in Philosophy for 1956 under the title “The Myth of the Given: Three Lectures on Empiricism and the Philosophy of Mind”.) Reprinted in SPR with additional footnotes. Published separately as Empiricism and the Philosophy of Mind: With an Introduction by Richard Rorty and a Study Guide by Robert Brandom, edited by Robert Brandom (Cambridge, MA: Harvard University Press, 1997). Also reprinted in Knowledge, Mind, and the Given: Reading Sellars’s “Empiricism and the Philosophy of Mind”, in Willem deVries and Timm Triplett (eds.). Indianapolis, IN: Hackett, 2000. FMPP (1981). “Foundations for a Metaphysics of Pure Process” (The Carus Lectures), The Monist, 64, 3–90. Lecture I, ‘The Lever of Archimedes’. Reprinted in ISR. I (1972). “‘…this I or he or it (the thing) which thinks’”, the presidential address, American Philosophical Association (Eastern Division) for 1970, Proceedings of the American Philosophical Association, 44, 5–31. Reprinted in Essays in Philosophy and its History, (pp. 62–90). Dordrecht: D. Reidel, ISR, and KTM.

208  References IKTE  (1978). “The Role of Imagination in Kant’s Theory of Experience”, The Dotterer Lecture 1978, in H. W. Johnstone, Jr. (ed.), Categories: A Colloquium (pp. 231–245). University Park, PA: Pennsylvania State University Press. Reprinted in ISR and KTM. ISR (2007). K. Scharp & R. Brandom (eds.), In the Space of Reasons: Selected Essays of Wilfrid Sellars. Cambridge, MA: Harvard University Press. ITM (1957). “Intentionality and the Mental”, a symposium by correspondence with Roderick Chisholm, in H. Feigl, M. Scriven, & G. Maxwell (eds.), Minnesota Studies in the Philosophy of Science (Vol. II, pp. 507–539). Minneapolis, MN: University of Minnesota Press. Reprinted in A. Marras (ed.). (1972). Intentionality Mind and Language. Chicago, IL: University of Illinois Press. IV (1964). “Induction as Vindication”, Philosophy of Science, 31: 197– 231. Reprinted in Essays in Philosophy and its History (pp. 367–418). Dordrecht: D. Reidel. KPT (2002). Kant and Pre-Kantian Themes: Lectures by Wilfrid Sellars, in P. V. Amaral, Atascadero, CA: Ridgeview. KTE (1967). “Some Remarks on Kant’s Theory of Experience”, Journal of Philosophy, 64, 633–647. Presented in a symposium on Kant at the 1967 meeting of the American Philosophical Association (Eastern Division). Reprinted in Essays in Philosophy and its History, (pp. 44–61). Dordrecht: D. Reidel., ISR, and KTM. KTI (1976). “Kant’s Transcendental Idealism”, Collections of Philosophy, 6, 165–181. KTM (2002). Kant’s Transcendental Metaphysics: Sellars’ Cassirer Lectures Notes and Other Essays, in Jeffrey F. Sicha, Atascadero, CA: Ridgeview. LTC (1969). “Language as Thought and as Communication”, Philosophy and Phenomenological Research, 29, 506–527. Reprinted in Paul K. (ed.). (1971). Language and Human Nature, St. Louis, MO: Warren H. Green. Also reprinted in Essays in Philosophy and its History, (pp. 93–117). Dordrecht: D. Reidel. and ISR. MMB (1952). “Mind, Meaning, and Behavior”, Philosophical Studies, 3, 83–95. ME (1989). The Metaphysics of Epistemology: Lectures by Wilfrid Sellars, in P.V. Amaral, Atascadero, CA: Ridgeview Publishing Co. Lectures given during the Fall of 1975. MEV (1981). “Mental Events”, Philosophical Studies, 39: 325–345. Reprinted in ISR. MFC (1974). “Meaning as Functional Classification (A perspective on the relation of syntax to semantics)”, (with replies to Daniel Dennett and Hilary Putnam) Synthese, 27: 417–437. Reprinted in J. G. Troyer and S. C. Wheeler, III (eds.). (1974). Intentionality, Language and Translation, Dordrecht: D. Reidel. An expanded version of BEB, “Belief and the Expression of Belief”, in H. E. Kiefer and M. K. Munitz (eds.), Language, Belief, and Metaphysics (pp. 146–158). Albany, NY: State University of New York Press, 1970. Also reprinted in ISR.

References 209 MGEC (1979). “More on Givenness and Explanatory Coherence,” in G. Pappas (ed.), Justification and Knowledge (pp. 169–182). Dordrecht: Reidel. Reprinted in Jonathan D. (ed.). (1988). Perceptual Knowledge (pp. 177– 191). Oxford: Oxford University Press. MP (1969). “Metaphysics and the Concept of a Person”, in K. Lambert (ed.), The Logical Way of Doing Things (pp. 219–252). New Haven, CT: Yale University Press. Reprinted in Essays in Philosophy and its History (pp. 214–241). Dordrecht: D. Reidel. NAO (1980). Naturalism and Ontology (The John Dewey Lectures for 1973– 1974). Atascadero, CA: Ridgeview. NI (1964). “Notes on Intentionality”, Journal of Philosophy, 61, 655–665. Reprinted in PPME. NS  (1962). “Naming and Saying”, Philosophy of Science, 29, 7–26. Reprinted in Science, Perception and Reality (pp. 225–246). London: Routledge & Kegan Paul. Reissued by Atascadero, CA: Ridgeview, 1991. OAPK (1970). “Ontology, the A Priori and Kant”, in J. F. Sicha (ed.), Kant’s Transcendental Metaphysics: Sellars’ Cassirer Lectures Notes and Other Essays (pp. 261–268). Atascadero, CA: Ridgeview. OAFP  (1988). “On Accepting First Principles”, in J. E. Tomberlin (ed.), Philosophical Perspectives 2, Metaphysics and Epistemology (pp. 301– 314). Atascadero, CA: Ridgeview Publishing Co. The paper was written in the sixties. OMP  (1966). “‘Ought’ and Moral Principles”, (unpublished). http://www. ditext.com/sellars/omp.html (last access 14.09.2021). ORAV (1980). “On Reasoning about Values”, American Philosophical Quarterly, 17, 81–101. One of three Tsanoff Lectures presented at Rice University, October 1978. PHM (1963). “Phenomenalism”, Science, Perception and Reality (pp. 60– 105). London: Routledge & Kegan Paul. Reissued by Atascadero, CA: Ridgeview, 1991. Reprinted in ISR. PP (1967). Philosophical Perspectives. Springfield, IL: Charles C. Thomas, Publisher. Reprinted in two volumes by Ridgeview Publishing Co.: Part 1: History of Philosophy, Part 2: Metaphysics and Epistemology. PR (1955). “Physical Realism”, Philosophy and Phenomenological Research, 15, 13–32. Reprinted in pp. 7–24. PSIM (1962). “Philosophy and the Scientific Image of Man”, in R. Colodny (ed.), Frontiers of Science and Philosophy (pp. 35–78). Pittsburgh, PA: University of Pittsburgh Press. Reprinted in Science, Perception and Reality (pp. 1–40). London: Routledge & Kegan Paul. Reissued by Atascadero, CA: Ridgeview, 1991 and ISR. RAL  (1973). “Reason and the Art of Living in Plato”, in D. Riepe (ed.), Phenomeno­logy and Natural Existence: Essays in Honor of Marvin Farber (pp. 353–377). New York, NY: The University of New York Press. Reprinted in Essays in Philosophy and its History, (pp. 3–26). Dordrecht: D. Reidel.

210  References RM (1973). “Reply to Marras”, Canadian Journal of Philosophy, 2, 485–493. SK (1975). “The Structure of Knowledge: (1) Perception; (2) Minds; (3) Epistemic Principles”, in H-N. Castañeda (ed.), Action, Knowledge and Reality: Studies in Honor of Wilfrid Sellars (pp. 295–347). Indianapolis, IN: Bobbs-Merrill. The Matchette Foundation Lectures for 1971 at the University of Texas. SM (1968). Science and Metaphysics: Variations on Kantian Themes. London: Routledge & Kegan Paul. The John Locke Lectures for 1965– 1966, reissued by Atascadero, CA: Ridgeview, 1992. SPR (1963). Science, Perception and Reality. London: Routledge & Kegan Paul. Reissued by Atascadero, CA: Ridgeview, 1991. SRII  (1965). “Scientific Realism or Irenic Instrumentalism: A Critique of Nagel and Feyerabend on Theoretical Explanation,” in R. Cohen & M. Wartofsky (eds.), Boston Studies in the Philosophy of Science (Vol. II, pp. 171–204). New York, NY: Humanities Press. Reprinted in PPME. SRLG (1954). “Some Reflections on Language Games”, Philosophy of Science, 21, 204–228. Reprinted in ISR. Reprinted with extensive additions in SPR. Roughly: §1–31 are the same in both, but SPR §32–46 are added; so ISR §32–37 = SPR §47–52 (§53 added); ISR §39–49 = SPR §54–65 (§62 is ISR fn. 3); ISR §50–56 was revised in SPR §66–74; and ISR §57–72 = SPR §75–89. SRT (1976). “Is Scientific Realism Tenable?”, (Presented at a symposium at the 1976 Philosophy of Science Association Meeting in Chicago). Published in Vol. II, Proceedings of PSA: 307–334. SSIS (1971). “Science, Sense Impressions, and Sensa: A Reply to Cornman”, Review of Metaphysics, 25, 391–447. SSMB (1953). “A Semantical Solution of the Mind-Body Problem”, Methodos, 5, 45–82. Reprinted in PPPW. SSOP (1982). “Sensa or Sensings: Reflections on the Ontology of Perception”, Philosophical Studies, 41 (Essays in Honor of James Cornman), 83–111. TC  (1962). “Truth and ‘Correspondence’”, Journal of Philosophy, 59, 29–56. TE (1963). “Theoretical Explanation”, in B. Baumrin (ed.), Philosophy of Science: The Delaware Seminar, (Vol. II, pp. 61–78). New York, NY: John Wiley. TTC (1974). “Toward a Theory of the Categories”, Essays in Philosophy and its History (pp. 318–339). Dordrecht: D. Reidel. TTP (1983). “Towards a Theory of Predication”, in J. Bogen & J. McGuire (eds.), How Things Are (pp. 281–318). Dordrecht: D. Reidel. Presented at a conference on predication at Pitzer College in April, 1981. WSNDL (2009). Wilfrid Sellars Notre Dame Lectures (1969–1986), edited with an introduction by Pedro Amaral, published online by Andrew Chrucky. http://www.ditext.com/amaral/wsndl.pdf (last access 14.09.2021). A revised edition was published in 2018, in David Landy and Pedro V. Amaral (eds.), Atascadero, CA: Ridgeview Publishing Company.

References 211

Works by Other Authors Allais, L. (2004). “Kant’s one world: Interpreting ‘transcendental idealism’”, British Journal for the History of Philosophy, 12(4), 655–684. Allais, L. (2015). Manifest Reality: Kant’s Idealism and His Realism. Oxford: Oxford University Press. Azari, B., Westlin, C., Satpute, A., Hutchinson, J., Kragel, Ph., Hoemann, K., & Barrett, L.F. (2020). “Comparing supervised and unsupervised approaches to emotion categorization in the human brain, body, and subjective experience”, Scientific Reports, 10(1), 20284. Bacon, M. (2012). Pragmatism: An Introduction, Cambridge: Polity. Barrett, L. F. (2006). “Solving the emotion paradox: Categorization and the experience of emotion”, Personality and Social Psychology Review, 10(1), 20–46. Barth, C. (2010). Objectivity and the Language-Dependence of Thought: A Transcendental Defence of Universal Lingualism. New York, NY: Routledge. Becker, K. & Thomson, I. (eds.) (2019). The Cambridge History of Philosophy, 1946-2010. Cambridge: Cambridge University Press. Blackburn, S. (1993). Essays in Quasi-Realism. Oxford & New York: Oxford University Press. Blackburn, S. (1998). Ruling Passions: A Theory of Practical Reasoning. Oxford: Clarendon Press. Blackburn, S. (2005). “Quasi-Realism No Fictionalism”, in M. E. Kalderon (ed.), Fictionalism in Metaphysics (pp. 322–338). Oxford: Oxford University Press. Brandom, R. B. (1994). Making It Explicit: Reasoning, Representing and Discursive Commitment. Cambridge, MA: Harvard University Press. Brandom, R. B. (2000). Articulating Reasons: An Introduction to Inferentialism. Cambridge, MA: Harvard University Press. Brandom, R. (2001). “Modality, Normativity, and Intentionality”, Philosophy and Phenomenological Research, 63(3), 587–609. Brandom, R. B. (2002). Tales of the Mighty Dead: Historical Essays in the Metaphysics of Intentionality. Cambridge, MA: Harvard University Press. Brandom, R. B. (2008). Between Saying and Doing: Towards an Analytic Pragmatism. Oxford & New York: Oxford University Press. Brandom, R. B. (2011). Perspectives on Pragmatism: Classical, Recent, and Contemporary. Cambridge, MA: Harvard University Press. Brandom, R. (2013). “Global Anti-Representationalism?”, in H. Price (ed.), Expressivism, Pragmatism and Representationalism (pp. 85–111). Oxford: Oxford University Press. Brandom, R. B. (2015). From Empiricism to Expressivism: Brandom Reads Sellars, Cambridge, MA: Harvard University Press. Brandom, R. B. (2016). “Sellars’s Metalinguistic Expressivist Nominalism”, in J. R. O’Shea (ed.), Sellars and His Legacy (pp. 19–46). Oxford: Oxford University Press. Brandt, S. (2011). Wittgenstein and Sellars on Intentionality. Oxford: Oxford  University. https://ora.ox.ac.uk/objects/uuid:0d9c1102-17bf-493ba1a0-aa983d277717 (last access 14.09.2021).

212  References Brandt, S. (2019). “Sellars and Wittgenstein on Following a Rule”, in S. Brandt & A. Breunig (eds.), Wilfrid Sellars and Twentieth-Century Philosophy (pp. 75–92). New York: Routledge. Brandt, S. & Breunig, A. (2019). Wilfrid Sellars and Twentieth-Century Philosophy. New York, NY: Routledge. Brassier, R. (2011). “Concepts and Objects”, in The Speculative Turn: Continental Materialism and Realism (pp. 47–65). Melbourne: Re.press. Brassier, R. (2013). “Nominalism, Naturalism, and Materialism: Sellars’ Critical Ontology”, in B. Bashour & H. D. Muller (eds.), Contemporary Philosophical Naturalism and its Implications (pp. 101–114). New York, NY: Routledge. Brassier, R. (2018). Review of “D. Pereplyotchik & D. R. Barnbaum (eds.), Sellars and Contemporary Philosophy. New York, NY: Routledge, 2017”. https://ndpr. nd.edu/news/sellars-and-contemporary-philosophy/ (last access 14.09.2021). Bremer, J. (1997). Rekategorisierung statt Reduktion. Göttingen: Vandenhoeck & Ruprecht. Brentano, F. C. (1874). Psychologie vom empirischen Standpunkt. Leipzig: Duncker & Humblot. Brooks, D. S., DiFrisco, J. & Wimsatt, W. C. (eds.) (2021). Levels of Organization in the Biological Sciences. Cambridge, MA: MIT Press. Burge, T. (1992). “Philosophy of language and mind: 1950–1990”, The Philosophical Review 101, 3–51. Burgess, A. & Sherman, B. (eds.). (2014). Metasemantics: New Essays on the Foundations of Meaning (1 edition). Oxford & New York: Oxford University Press. Callanan, J. (2006). “Kant’s transcendental strategy”, The Philosophical Quarterly, 56: 360–381. Cappelen, H., Gendler, T. S. & Hawthorne, J. (eds.). (2016). The Oxford Handbook of Philosophical Methodology. New York, NY: Oxford University Press. Cassam, Q. (1987). “Transcendental arguments, transcendental synthesis and transcendental idealism”, The Philosophical Quarterly, 37 (149), 355–378. Castañeda, H.-N. (ed.). (1975). Action, Knowledge, and Reality: Critical Studies in Honor of Wilfrid Sellars. Indianapolis, IN: Bobbs-Merrill. Christias, D. (2014). “An interpretation and extension of Sellars’s views on the epistemic status of philosophical propositions”, Metaphilosophy, 45(3), 348–371. Christias, D. (2015). “Somatic intentionality bifurcated: A Sellarsian response to Sachs’s Merleau-Pontyan account of intentionality”, International Journal of Philosophical Studies, 23(4), 539–561. Christias, D. (2018). “On the proper construal of the manifest-scientific image distinction”, Synthese, 195(3), 1295–1320. Cobley, P. (2019). “Afterword. Semiotics and languaging”, Chinese Semiotic Studies, 15(4), 697–709. Conant, J. (2017). “Kant’s Critique of the Layer-Cake Conception of Human Mindedness”, in J. R. O’Shea (ed.), Kant’s Critique of Pure Reason (pp. 120– 139). Cambridge: Cambridge University Press. Corti, L. (2019). “Between causes and reasons: Sellars, Hegel (and Lewis) on “sensation””, International Journal of Philosophical Studies, 27(3), 422–447.

References 213 Corti, L. (2021a). “The notion of sensation in Sellars’ theory of perception”, European Journal of Philosophy, 1–21. https://doi.org/10.1111/ejop.12626 (last access 14.09.2021) Corti, L. (2021b). “Review of Stefan Brand & Anke Breunig (eds.), Wilfrid Sellars and Twentieth-Century Philosophy. New York: Routledge, 2019, ISBN9780815384991”. https://ndpr.nd.edu/reviews/wilfrid-sellars-and-­twentiethcentury-philosophy/ (last access 14.09.2021). Corti, L. & Nunziante, A. M. (eds.) (2018). Sellars and the History of Modern Philosophy. London: Routledge. Crane, T. (1998). “Intentionality as the Mark of the Mental”, Royal Institute of Philosophy Supplement, 43, 229–251. Crane, T. (2008). “Fraught with Ought: In the Space of Reasons: Selected Essays of Wilfrid Sellars, in Kevin Scharp and Robert Brandom (eds.); Wilfrid Sellars: Fusing the Images by Jay Rosenberg”, London Review of Books, 30(12), 33–35. Crane, T. (2013). The Objects of Thought. Oxford & New York: Oxford University Press. Cumpa, J. (2020). “Categories”, Philosophy Compass, 15(1), e12646. Cumpa, J. (ed.) (2021). Categorial Ontologies: From Realism to Eliminativism. New York, NY: Routledge. Dennett, D. C. (1987). The Intentional Stance. Cambridge, MA: MIT Press. Dach, S. (2018). “Sellars’s Two Images as a Philosopher’s Tool”, Metaphilosophy, 49, 568–588. Dach, S. (2021). “Sellars, we-intentions and ought-statements”, Synthese, 198(5), 4415–4439. Daly, C. (2010). An Introduction to Philosophical Methods. Peterborough: Breakwater Books. Delaney, C. F., Loux, M. J., Gutting, G., & Solomon, W. D. (eds.). (1977). The Synoptic Vision: Essays on the Philosophy of Wilfrid Sellars. Notre Dame: University of Notre Dame Press. deVries, W. A. (2005). Wilfrid Sellars. Montreal: McGill-Queen’s University Press. deVries, W. A. (2010). “Naturalism, the autonomy of reason, and pictures”, International Journal of Philosophical Studies, 18(3), 395–413. deVries, W. A. (2012a). “All in the Family”, in D. Ryder (ed.), Millikan and Her Critics (pp. 259–280). Hoboken, NJ: Wiley-Blackwell. deVries, W. A. (2012b). “Sellars, realism, and Kantian thinking”, Social Epistemology Review and Reply Collective, 2(1), 57–67. deVries, W. A. (2016a). “Images, descriptions, and pictures. Personhood and the Clash”, in J. R. O’Shea (ed.), Sellars and His Legacy (pp. 47–59). Oxford: Oxford University Press. deVries, W. A. (2016b). “Just what is the relation between the manifest and the scientific images?”, International Journal of Philosophical Studies, 24(1), 112–128. deVries, W. A. (2016c). “The Causal Articulation of Practical Reality”, in D. Pereplyotchik & D. R. Barnbaum (eds.), Sellars and Contemporary Philosophy (pp. 155–171). New York, NY: Routledge. deVries, W. A. (2019). “Some Remarks on Sellars’s Theory of Experience”, in S. Brandt & A. Breunig (eds.), Wilfrid Sellars and Twentieth-Century Philosophy (pp. 207–220). New York, NY: Routledge.

214  References deVries, W. A. (2020a). “Wilfrid Sellars”, in Edward N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Fall 2020 Edition). https://plato.stanford.edu/ archives/fall2020/entries/sellars/ (last access 14.09.2021). deVries, W. A. (2020b). “Hegel’s Pragmatism”, in M. Bykova & K. Westphal (eds.), The Palgrave Hegel Handbook: Palgrave Handbooks in German Idealism (pp. 541–557). London: Palgrave Macmillan. D’Oro, G. & Overgaard, S. (eds.). (2017). The Cambridge Companion to Philosophical Methodology. Cambridge: Cambridge University Press. Dupré, J. (1995). The Disorder of Things: Metaphysical Foundations of the Disunity of Science. Cambridge, MA: Harvard University Press. Erhard, C., Meißner, D. & Noller, J. (eds.). (2017). Wozu Metaphysik?: HistorischSystematische Perspektiven. Freiburg: Verlag Karl Alber. Fodor, J. A. (1974). “Special sciences (or: The disunity of science as a working hypothesis)”, Synthese, 28(2), 97–115. Jerry, F. & Ernie, L. (2001). “Brandom’s burdens: Compositionality and inferentialism”, Philosophy and Phenomenological Research, 63(2), 465–481. Förster, E. (2012). The Twenty-Five Years of Philosophy: A Systematic Reconstruction (B. Bowman, Trans.). Cambridge, MA: Harvard University Press. Franz, J. H. (2010). Geist und Handlung: Wilfrid Sellars’ Theorie des Handelns im Manifesten und Wissenschaftlichen Weltbild. Würzburg: Königshausen & Neumann. Gabbani, C. (ed.). (2012). Humana.Mente, 21. Special issue: Between Two Images. The Manifest and Scientific Conception of the Human Being, 50 Years On. https:// www.humanamente.eu/index.php/HM/issue/view/33 (last access 14.09.2021). Garfield, J. L. (2012). “Sellarsian synopsis: Integrating the images”, HUMANA. MENTE Journal of Philosophical Studies, 5(21), 103–121. Giladi, P. & Sachs, C. (2019) “Editors’ introduction to ‘Hegel and Sellars’” (Special Issue) International Journal of Philosophical Studies, 27(3), 359–362. Gironi, F. (ed.). (2017). The Legacy of Kant in Sellars and Meillassoux: Analytic and Continental Kantianism. New York, NY: Routledge. Gomes, A. (2017). “Kant, the Philosophy of Mind, and Twentieth-Century Analytic Philosophy”, in A. Stephenson & A. Gomes (eds.) Kant and the Philosophy of Mind: Perception, Reason, and the Self (pp. 5–24). Oxford: Oxford University Press. Gomes, A., Stephenson, A. & Moore, A. W. (forthcoming). “On the Necessity of the Categories”, The Philosophical Review. Guyer, P. (2005). “Review of Kant and pre-Kantian themes: Lectures by Wilfrid Sellars and Kant’s transcendental metaphysics: Sellars’ Cassirer lecture notes and other essays”, Philosophical Review, 114(4), 535–539. Haag, J. (2007). “Weltbezug und Objektivität in nicht-relationalen Konzeptionen der Intentionalität”, in A. Rami & H. Wansing. (eds.) Referenz Und Realität (pp. 39–59). Padaborn: Mentis. Haag, J. (2009). “Sinneseindrücke und die ‘Führung von Außen”, in R. Schantz (ed.), Wahrnehmung und Wirklichkeit (pp. 19–57). Frankfurt: Ontos. Haag, J. (2012a). “Die Funktion von Grenzbegriffen. Symposium zu: Eckart Förster: Die 25 Jahre der Philosophie”, Deutsche Zeitschrift für Philosophie, 60(6), 993–1001.

References 215 Haag, J. (2012b). “Philosophische Abstraktionsebenen”, in C. Barth & H. Sturm (eds.), Brandoms Expressive Vernunft. Historische und systematische Untersuchungen (pp. 261–285). Paderborn: Mentis. Haag, J. (2012c). “Some Kantian Themes in Wilfrid Sellars’s Philosophy”, in B. Centi (ed.), Kant in the 20th Century (Vol. 30, pp. 111–126). Sonderband von: Paradigmi. Rivista di critica filosofica. Haag, J. (2014). “McDowells Kant und McDowells Sellars”, in C. Barth & D. Lauer (eds.), Die Philosophie John McDowells (pp. 145–168). Paderborn: Mentis. Haag, J. (2015). “Die Unergründlichkeit der Einbildungskraft”, Internationales Jahrbuch für Philosophische Anthropologie, 5(1), 159–178. Haag, J. (2016). “A Kantian Critique of Sellars’s Transcendental Realism”, in P. Reider (ed.), Wilfrid Sellars, Idealism and Realism: Understanding Psychological Nominalism (pp. 149–171). London: Bloomsbury Publishing. Haag, J. (2017). “Analytic Kantianism: Sellars and McDowell on sensory consciousness”, Con-Textos Kantianos. International Journal of Philosophy, 1(6), 18–41. Haag, J. (2019). “Transcendental Principles and Perceptual Warrant: A Case Study in Analytic Kantianism”, in S. Brandt & A. Breunig (eds.), Wilfrid Sellars and Twentieth-Century Philosophy (pp. 130–150). Now York, NY: Routledge. Hall, E. (1961). Our Knowledge of Fact and Value. Chapel Hill, CA: University of North Carolina Press. Haug, M. C. (ed.). (2013). Philosophical Methodology: The Armchair or the Laboratory? (1 edition). London; New York: Routledge. Haugeland, J. (1990). “The intentionality all-stars”, Philosophical Perspectives, 4, 383–427. Hicks, M. R. (2020). “Wilfrid Sellars and the Task of Philosophy”, Synthese. https://doi.org/10.1007/s11229-020-02646-8 (last access 14.09.2021). Hoeppner, T. (2020). “Analytische Philosophie (III): Wilfrid Sellars, Peter F. Strawson und Barry Stroud. Analytischer Kantianismus”, in J. Urbich & J. Zimmer (eds.), Handbuch Ontologie (pp. 248–257). Stuttgart: J.B. Metzler. Horstmann, R.-P. (2018). Kant’s Power of Imagination. Cambridge: Cambridge University Press. Hübner, J. (2015). Einführung in die Theoretische Philosophie. Stuttgart: Metzler & Springer-Verlag. Jacob, P. (2014). “Intentionality”, in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Winter 2014). Metaphysics Research Lab, Stanford University. https://plato.stanford.edu/archives/win2014/entries/intentionality/ (last access 14.09.2021). Kambartel, F. & Stekeler-Weithofer, P. (2005). Sprachphilosophie: Probleme Und Methoden. Stuttgart: Reclam. Kant, I. (1900). Gesammelte Schriften Bd.1-24 = AA. (Bd. 1-22 Preussische Akademie der Wissenschaften, Bd. 23 Deutsche Akademie der Wissenschaften zu Berlin, & ab Bd. 24 Akademie der Wissenschaften zu Göttingen, Eds.). Berlin. Kant, I. (1999). Critique of Pure Reason, in P. Guyer & A. W. Wood (eds.). Cambridge: Cambridge University Press. Klemick, G. (2020). “Sellars’ metaethical quasi-realism”, Synthese, 197(5), 2215–2243.

216  References Koch, A. F. (2011). “ Die Bildtheorie des Elementarsatzes und die Lesbarkeit der Dinge (Wittgenstein, Sellars, Kant)”, in S. G. Carson, J. Knowles, & B. K. Myskja (eds.), Kant: Here, Now, and How Essays in Honour of Truls Wyller (pp. 179–192). Paderborn: Mentis. Koch, A. F. (2016). Hermeneutischer Realismus. Tübingen: Mohr Siebeck. Koch, A. F. (2018). “Den Raum der Gründe gründen. Bekanntschaft ohne Mythos und der Grund des Raums der Gründe”. (Manuscript). Koons, J. R. (2016). “Toward a Sellarsian Ethics for the 21st Century”, in D. Pereplyotchik & D. R. Barnbaum (eds.), Sellars and Contemporary Philosophy (pp. 36–52). New York, NY: Routledge. Koons, J. R. (2019). The Ethics of Wilfrid Sellars. New York, NY: Routledge. Koons, J. & Sachs, C. (forthcoming). “The Role of Picturing in Sellars’s Practical Philosophy”, Journal of Philosophical Research. Kraut, R. (1990). “Varieties of pragmatism”, Mind, 99(394), 157–183. Kraut, R. (2016). “Norm and Object: How Sellars Saves Metaphysics from the Pragmatist Onslaught”, in J. O’Shea (ed.), Sellars and His Legacy (pp. 60–79). Oxford: Oxford University Press. Kreines, J. (2007). “Between the bounds of experience and divine intuition: Kant’s epistemic limits and Hegel’s ambitions”, Inquiry, 50(3), 306–334. Landy, D. (2009). “Sellars on Hume and Kant on representing complexes”, European Journal of Philosophy, 17(2), 224–246. Landy, D. (2015). Kant’s Inferentialism: The Case against Hume. London/­ New York, NY: Routledge. Landy, D. (2018). “Sellars and Hume on the Ontological Status of TheoreticalExplanatory Entities”, in L. Corti & A. M. Nunziante (eds.), Sellars and the History of Modern Philosophy (pp. 59–78). New York & Oxford: Routledge. Landy, D. (2019). “Sellars’ argument for an ontology of absolute processes”, Journal for the History of Analytical Philosophy, 7(1), 1–25. Levine, S. (2016). “Sellars and nonconceptual content”, European Journal of Philosophy, 24, 855–878. Levine, S. & Wanderer, J. (2013). “Review of Chauncey Maher, The Pittsburgh School of Philosophy: Sellars, McDowell, Brandom”, https://ndpr.nd.edu/ reviews/the-pittsburgh-school-of-philosophy-sellars-mcdowell-brandom/ (last access 14.09.2021). Loeffler, R. (2017). Brandom. Malden, MA: Polity. Macbeth, D. (2010). “Inference, Meaning, and Truth in Brandom, Sellars, and Frege”, in B. Weiss & J. Wanderer (eds.), Reading Brandom: On Making It Explicit (pp. 197–212). New York, NY: Routledge. Maher, C. (2012). The Pittsburgh School of Philosophy: Sellars, McDowell, Brandom (1st edition). New York, NY: Routledge. Matsui, T. (2020). “Inferentialism and semantic externalism: A neglected debate between Sellars and Putnam”, British Journal for the History of Philosophy, 10.1080/09608788.2020.1801380 McDowell, J. (1994). Mind and World. Cambridge, MA: Harvard University Press. McDowell, J. (1998a). “Lecture II: The logical form of an intuition”, The Journal of Philosophy, 95(9), 451–470. McDowell, J. (1998b). “Lecture III: Intentionality as a relation”, Journal of Philosophy, 95(9), 471–491.

References 217 McDowell, J. (2000). “Sellars’s Transcendental Empiricism”, in Rationalität, Realismus, Revision/Rationality, Realism, Revision. Vorträge des 3. internationalen Kongresses der Gesellschaft für Analytische Philosophie vom 15. bis zum 18. September 1997 in München/Proceedings of the 3rd international Congress of the Society for Analytical Philosophy September 15–18, 1997 in Munich (Reprint 2012). Berlin, Boston: De Gruyter. McDowell, J. (2013). Having the World in View: Essays on Kant, Hegel, and Sellars (Reprint). Cambridge, MA: Harvard University Press. Miguens, S. (ed.). (2020). The Logical Alien: Conant and His Critics. Cambridge, MA: Harvard University Press. Miller, S. A. (2018). Community and Loyalty in American Philosophy: Royce, Sellars, and Rorty. London & New York: Routledge. Millikan, R. (2012). “Reply to deVries. A Daughter’s Respectful Disagreement with Sellars”, in D. Ryder, J. Kingsbury, & K. Williford (eds.), Millikan and Her Critics (pp. 276–280). Hoboken, NJ: Wiley-Blackwell. Misak, C. & Price, H. (eds.). (2017). The Practical Turn: Pragmatism in Britain in the Long Twentieth Century. Oxford, New York: Oxford University Press. Nagel, E. (1961). The Structure of Science: Problems in the Logic of Scientific Explanation. New York: Harcourt, Brace & World. Noûs. (1973) 7(2). “Studies in Wilfrid Sellars’ Philosophy”. http://www.jstor.org/ stable/2214484 (last access 14.09.2021). Olen, P. (2016). Wilfrid Sellars and the Foundations of Normativity. London: Palgrave Macmillan. Oppenheim, P. & Putnam, H. (1958). “The Unity of Science as a Working Hypothesis”, in H. Feigl, M. Scriven, & G. Maxwell (eds.), Concepts, Theories, and the Mind-Body Problem (Vols. 1–2, pp. 3–36). Minneapolis, MN: University of Minnesota Press. O’Shea, J. R. (2007). Wilfrid Sellars: Naturalism with a Normative Turn. Cambridge & Malden, MA: Polity Press. O’Shea, J. R. (2010). “Normativity and scientific naturalism in Sellars’ ‘JanusFaced’ space of reasons”, International Journal of Philosophical Studies, 18(3), 459–471. O’Shea, J. R. (2011, July 1). “How to be a Kantian and a naturalist about human knowledge: Sellars’s middle way”, Journal of Philosophical Research, 36, 327–359. O’Shea, J. R. (2014). Kant’s Critique of Pure Reason: An Introduction and Interpretation. Durham: Routledge. O’Shea, J. R. (2016a). Sellars and His Legacy, in J. R. O’Shea (ed.). Oxford: Oxford University Press. O’Shea, J. R. (2016b). “What to take away from Sellars’s Kantian naturalism”, in ibid. Sellars and His Legacy (pp. 130–149). Oxford: Oxford University Press. O’Shea, J. R. (2017). Kant’s Critique of Pure Reason: A Critical Guide. New York City, NY: Cambridge University Press. O’Shea, J. R. (2021a). “What is the myth of the given?”, Synthese. https://doi. org/­10.1007/s11229-021-03258-6. O’Shea, J. R. (2021b). “On Sellars’s Analytic-Kantian Conception of Categories as Classifying Conceptual Roles”, in J. Cumpa (ed.), Categorial Ontologies: From Realism to Eliminativism. New York, NY: Routledge.

218  References Parent, T. (2016). Self-Reflection for the Opaque Mind: An Essay in NeoSellarsian Philosophy (1st edition). New York, NY: Routledge. Peregrin, J. (2001). Meaning and Structure: Structuralism of (Post)Analytic Philosophers. Aldershot, Hants, England; Burlington: Routledge. Peregrin, J. (2018). “Davidson and Sellars on ‘Two Images’”, Philosophia, 46(1), 183–192. Pereplyotchik, D. & Barnbaum, D. R. (eds.). (2016). Sellars and Contemporary Philosophy. New York, NY: Routledge. Perler, D. (2004). Theorien der Intentionalität im Mittelalter. Frankfurt am Main: Klostermann Vittorio. Pitt, J. C. (1978). The Philosophy of Wilfrid Sellars: Queries and Extensions. Papers Deriving from and Related to a Workshop on the Philosophy of Wilfrid Sellars held at Virginia Polytechnic Institute and State University 1976, D. Reidel Publishing Company, Dordrecht, Holland: Springer Science & Business Media. Price, H. (1997). “Naturalism and the Fate of the M-worlds”, Aristotelian Society Supplementary Volume, 71(1), 247–268. Price, H. (2008). “Will there be blood? Brandom, Hume, and the genealogy of modals”, Philosophical Topics, 36(2), 87–97. Price, H. (2011). Naturalism without Mirrors. Oxford; New York: Oxford University Press. Price, H. (2013). Expressivism, Pragmatism and Representationalism. Cambridge: Cambridge University Press. Price, H. (2015). “Idling and Sidling Toward Philosophical Peace”, in S. Gross, N. Tebben & M. Williams (eds.), Meaning without Representation (pp. 307–330). Oxford: Oxford University Press. Price, H. (2016). “Wilfrid Sellars Meets Cambridge Pragmatism”, in D. Pereplyotchik & D. R. Barnbaum (eds.), Sellars and Contemporary Philosophy (pp. 123–140). New York, NY: Routledge. Price, H. (2017). “Ramsey’s Ubiquitous Pragmatism”, in C. Misak & P. Huw (eds.), The Practical Turn: Pragmatism in Britain in the Long Twentieth Century (pp. 149–162). Oxford: Oxford University Press. Price, Huw (2020). “Global Expressivism by the Method of Differences”, Royal Institute of Philosophy Supplement 86, 133–154. Price, H. & McDowell, J. (1997). “Review of ‘mind and world”, Philosophical Books, 38(3), 169–181. Quine, W. V. O. (1960). Word and Object. Cambridge, MA: MIT Press. Redding, P. (2007). Analytic Philosophy and the Return of Hegelian Thought. Cambridge: Cambridge University Press. Redding, P. (2012). “McDowell’s radicalization of Kant’s account of concepts and intuitions: A Sellarsian (and Hegelian) critique”, Verifiche: Rivista Trimestrale Di Scienze Umane, 41(1–3), 9–37. Redding, P. (2020). “The “Pittsburgh” Neo-Hegelianism of Robert Brandom and John McDowell”, in M. Bykova & K. Westphal (eds.), The Palgrave Hegel Handbook. Palgrave Handbooks in German Idealism. Cham: Palgrave Macmillan. https://doi.org/10.1007/978-3-030-26597-7_28 (last access 14.09.2021). Reider, P. (2012). “Sellars and knowing the thing-in-itself”, Social Epistemology Review and Reply Collective, 2(1), 68–82.

References 219 Reider, P. (2016). Wilfrid Sellars, Idealism, and Realism: Understanding Psychological Nominalism. London: Bloomsbury Publishing. Rorty, R. (1961). “Pragmatism, categories, and language”, The Philosophical Review, 70(2), 197–223. Rorty, R. (1970). “Science and Metaphysics: Variations on Kantian Themes”, Philosophy, 45(171), 66–70. Rorty, R. (1971). “Verificationism and transcendental arguments”, Noûs, 5(1), 3–14. Rorty, R. (1972). “The world well lost”, Journal of Philosophy, 69(19), 649–665. Rorty, R. (1979). Philosophy and the Mirror of Nature. Princeton, NJ: Princeton University Press. Rorty, R. (1988). “Representation, social practise, and truth”, Philosophical Studies, 54(2), 215–228. Rorty, R. (1997). “What do you do when they call you a ‘Relativist’?”, Philosophy and Phenomenological Research, 57(1), 173–177. Rorty, R. (1998). Achieving Our Country: Leftist Thought in Twentieth Century America. Cambridge, MA: Harvard University Press. Rorty, R. (1999a). Philosophy and Social Hope. New York: Penguin Books. Rorty, R. (1999b) “Trotsky and the Wild Orchids” in ibid. Philosophy and Social Hope (pp. 140–153). Harmondsworth, UK: Penguin Books. Rorty, R. & Price, H. (2010). “Exchange on ‘Truth as Convenient Friction’”, in M. de Caro & D. Macarthur (eds.), Naturalism and Normativity. New York: Columbia University Press. Rosefeldt, T. (2015). “Noumenon/Phaenomenon”, in M. Willaschek, J. Stolzenberg, G. Mohr, & S. Bacin (eds.), Kant-Lexikon (pp. 1686–1690). Berlin, Boston: De Gruyter. Rosen, G. (2001). “Brandom on modality, normativity and intentionality”, Philosophy and Phenomenological Research, 63(3), 611–623. Rosenberg, J. F. (1974). Linguistic Representation. The Netherlands: Springer. Rosenberg, J. F. (1980). One World and Our Knowledge of It: The Problematic of Realism in Post-Kantian Perspective. Springer Science & Business Media. Rosenberg, J. F. (2005). Accessing Kant: A Relaxed Introduction to the Critique of Pure Reason. Oxford, New York: Oxford University Press. Rosenberg, J. F. (2007). Wilfrid Sellars: Fusing the Images. Oxford, New York: Oxford University Press. Rosenthal, D. (2016). “Quality Spaces, Relocation, and Grain”, in ed. J. O’Shea (ed.), Sellars and His Legacy (pp. 149–185). Oxford: Oxford University Press. Rottschaefer, W. A. (2011a).”The middle does not hold: Why It’s always better to be right with the right-wing-Sellarsians”, Journal of Philosophical Research 36, 361–369. Rottschaefer, W. A. (2011b). “Why Wilfrid Sellars is right (and right-wing): Thinking with O’Shea on Sellars, Norms, and Nature”, Journal of Philosophical Research 36, 291–325. Rouse, J. (2015). Articulating the World: Conceptual Understanding and the Scientific Image. Chicago, IL: University of Chicago Press. Ryder, D., Kingsbury, J. & Williford, K. (eds.). (2012). Millikan and Her Critics. Hoboken, NJ: John Wiley & Sons. Sachs, C. B. (2011). “The shape of a good question: McDowell, evolution, and transcendental philosophy”, The Philosophical Forum, 4(1), 61–78.

220  References Sachs, C. B. (2013). “Rorty’s Debt to Sellarsian Metaphysics: Naturalism, Secularism, and the Enlightenment”, Metaphilosophy 44(5), 682–707. Sachs, C. B. (2014). “Discursive and somatic intentionality: Merleau-Ponty contra ‘McDowell or Sellars’”, International Journal of Philosophical Studies, 22(2), 199–227. Sachs, C. B. (2015). Intentionality and the Myths of the Given: Between Pragmatism and Phenomenology (1st edition). London, Brookfield: Routledge. Sachs, C. B. (2019a). “In defense of picturing; Sellars’s philosophy of mind and cognitive neuroscience”, Phenomenology and the Cognitive Sciences, 18, 669–689. Sachs, C. B. (2019b). “A Conceptual Genealogy of the Pittsburgh School: Between Kant and Hegel”, in K. Becker & I. Thomson (eds.), The Cambridge History of Philosophy, 1946–2010 (pp. 664–676). Cambridge: Cambridge University Press. Schlösser, U. (2013). “Concept Formation, Synthesis and Judgment: Kant’s Theory of the Logical and Cognitive Activities of the Mind”, in Dina Emundts (ed.), Self, World, and Aesthetics: Metaphysical Topics in Kant and Hegel (pp. 177–206). Berlin, New York: Walter De Gruyter. Schönwälder-Kuntze, T. (2015). Philosophische Methoden zur Einführung. Hamburg: Junius Hamburg. Schulting, D. (ed.). (2016). Kantian Nonconceptualism. London: Palgrave Macmillan. Seibt, J. (1990). Properties as Processes: A Synoptic Study of Wilfrid Sellars Nominalism. Atascadero, CA: Ridgeview Publishing Co. Seibt, J. (2007). Wilfrid Sellars. Paderborn: Mentis. Seibt, J. (2016). “How to Naturalize Sensory Consciousness and Intentionality Within a Process Monism With Normativity Gradient”, in J. O’Shea (ed.), Sellars and His Legacy (pp. 186–220). Oxford: Oxford University Press. Shapiro, L. (2013). “Intentional relations and the sideways-on view: On McDowell’s critique of Sellars”, European Journal of Philosophy, 21(2), 300–319. Shapiro, L. (2014a). “Linguistic function and content: Reflections on Price’s pragmatism”, The Philosophical Quarterly, 64(256), 497–506. Shapiro, L. (2014b). “Sellars on the function of semantic vocabulary”, British Journal for the History of Philosophy, 22(4), 792–811. Shapiro, L. (2019). “Sellars, Truth Pluralism, and Truth Relativism”, in S. Brandt & A. Breunig (eds.), Wilfrid Sellars and Twentieth-Century Philosophy (pp. 174–206). New York, NY: Routledge. Sicha, J. F. (2002). Kant’s Transcendental Metaphysics: Sellars’ Cassirer Lectures Notes and Other Essays. Atascadero, CA: Ridgeview Publishing Co. Sicha, J. F. (2014). “Defending the unpopular Sellars: Picturing and ‘the descriptive’”, Journal of Philosophical Research, 39, 127–163. Siewert, C. (2017). “Consciousness and Intentionality”, in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Spring 2017). Metaphysics Research Lab, Stanford University. https://plato.stanford.edu/archives/spr2017/entries/ consciousness-intentionality/ (last access 14.09.2021). Stang, N. F. (2016). “Kant’s transcendental idealism”, in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Spring 2016). Metaphysics Research Lab, Stanford University. https://plato.stanford.edu/archives/spr2016/entries/ kant-transcendental-idealism/ (last access 14.09.2021).

References 221 Stekeler-Weithofer, P. (2009). “Generisches Wissen in kategorialen Inferenz­ strukturen: Zur Metaphysik des Begrifflichen”, in V. Munz, K. Puhl & sJ. Wang (eds.), Language and World Part Two: Signs, Minds and Actions (pp. 191–215). Heusenstamm: Ontos. Stekeler-Weithofer, P. (2014). Sprachphilosophie: Eine Einführung (1st edition). München: C.H. Beck. Stern, R. (2003). Transcendental Arguments: Problems and Prospects. Oxford: Oxford University Press. Stern, R. (2020). “Transcendental Arguments”, in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Fall 2020 Edition). https://plato.stanford.edu/ archives/fall2020/entries/transcendental-arguments/ (last access 14.09.2021). Stovall, P. (2022). The Single-Minded Animal. Shared Intentionality, Normativity, and the Foundations of Discursive Cognition. New York: Routledge Strawson, P. (1966). The Bounds of Sense. London: Methuen. Stroud, B. (2002). Understanding Human Knowledge: Philosophical Essays. Oxford: Clarendon Press. Stroud, B. (2013). Engagement and Metaphysical Dissatisfaction: Modality and Value. New York, NY: Oxford University Press. Thomasson, A. (2018). “Categories”, in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Spring 2018). Metaphysics Research Lab, Stanford University. https://plato.stanford.edu/archives/spr2018/entries/­categories/ (last access 14.09.2021). Tomasello, M. (2014). A Natural History of Human Thinking. Cambridge, London: Harvard University Press. van Fraassen, B. C. (1999). “The Manifest Image and the Scientific Image”, in D. Aerts, J. Broekaert, & E. Mathijs (eds.), Einstein Meets Magritte: The White Book: An Interdisciplinary Reflection (pp. 29–52). Dordrecht: Kluwer. Wanderer,J.& Levine,S.(2013).Review of“Chauncey Maher,The Pittsburgh School of Philosophy: Sellars, McDowell, Brandom, London/New York: Routledge, 2012”, https://ndpr.nd.edu/reviews/the-pittsburgh-school-of-­p hilosophy-sellarsmcdowell-brandom/ (last access 14.09.2021). Westphal, K. R. (2004). Kant’s Transcendental Proof of Realism. Cambridge: Cambridge University Press. Westphal, K. R. (2017). Grounds of Pragmatic Realism: Hegel’s Internal Critique and Reconstruction of Kant’s Critical Philosophy. Leiden & Boston: Brill. Westphal, K. R. (2020). Kant’s Critical Epistemology. Why Epistemology Must Consider Judgment First. New York & Oxford: Routledge. Willaschek, M. (1998). “Phaenomena/Noumena Und Die Amphibolie Der Reflexionsbegriffe”, in M. Willaschek & G. Mohr (eds.), Immanuel Kant, Kritik Der Reinen Vernunft (pp. 325–351). Berlin: Akademie-Verlag. Willaschek, M., Stolzenberg, J., Mohr, G. & Bacin, S. (eds.). (2015). Kant-Lexikon. Berlin, Boston: De Gruyter. Williams, M. (1977). Groundless Belief: Essay on the Possibility of Epistemology. New Haven: Yale University Press. Williamson, T. (2014). “How Did We Get Here from There? The Transformation of Analytic Philosophy”, Belgrade Philosophical Annual, 27, 7–37. Wilson, M. (2017). Physics Avoidance: And Other Essays in Conceptual Strategy. Oxford: Oxford University Press.

222  References Wimsatt, W. C. (1976). “Reductionism, levels of organization, and the mindbody problem”, in Consciousness and the Brain (pp. 205–267). Boston, MA: Springer. Wimsatt, W. C. (1994). “The ontology of complex systems: Levels of organization, perspectives, and causal thickets”, Canadian Journal of Philosophy, 24, 207–274. Wimsatt, W. C. (2007). Re-Engineering Philosophy for Limited Beings: Piecewise Approximations to Reality. Cambridge, MA: Harvard University Press. Wittgenstein, L. (1921). Tractatus Logico-Philosophicus: Logisch-Philosophische Abhandlung. Frankfurt am Main: Suhrkamp Verlag. Wolf, M. P. & Koons, J. R. (2016). The Normative and the Natural. London: Palgrave Macmillan. Wolf, M. P. & Lance, M. N. (eds.). (2006). The Self-Correcting Enterprise: Essays on Wilfrid Sellars. Amsterdam: Editions Rodopi BV. Wunsch, M. (2007). Einbildungskraft und Erfahrung bei Kant. Berlin/New York: de Gruyter. Wunsch, M. (2014). “Zur argumentativen Rekonstruktion der Theorie der Einbildungskraft in Kants ‘Kritik der reinen Vernunft’”, in M. Egger (ed.), Philosophie nach Kant. Neue Wege zum Verständnis von Kants Transzendentalund Moralphilosophie (pp. 127–140). Berlin & Boston: De Gruyter.

Index

aboutness 45, 103–108, 110–117n2, 125, 132, 145, 148n2, 176, 179, 190, 197n17, 203 absolute; correctness 158; sense of truth 153; standpoint 151 abstraction 56, 95, 100, 123, 126, 136, 139, 153n28, 168, 196–197; highest point of xiii, 17, 28, 37, 49–50, 72, 79, 161; levels of 5, 14, 19–24, 25n11, 26–28, 100n21, 106, 120n28, 168, 170–172, 176– 177, 196–197n16, 202; meta-level 20–21, 25, 27, 189; method of 30, 33, 58, 126, 151n19, 171–172, 176, 190, 196n8, 196n10, 202 objectlevel 20, 23, 126, 130, 161, 188, 196; standpoint of 27, 153, 162 accusative, empty 13, 20, 28, 45 acquaintance 111 action xiii, 76, 87, 114, 126, 152n24; intentional 6; of rules for 52, 67n23; non-spatial, non-temporal 76 activity 24n4, 66n17, 119n20, 145, 147–148, 154n33, 176, 178, 185, 189–191, 193; (meta)classificatory 39, see synthesis acts, sensible 78 actuality 10, 46, 53, 59, 63–65n12, 79, 80, 87, 130, 135, see (formal) reality, existence (per se); category of 15, 203 adequacy (pictorial), degrees of 7, 137, 143–145, 151, 153n83–158, 160, 162, 165, 168–169, 172–176, 178, 185, 191, 194–195, 197n20, 199n34, 204; criteria of 161; method of projection 196; truth as 184

agent 15–16, 31–32, 49, 49, 64; as perceiver, speaker 122, 141, 189; as thinker 103, 192; epistemic 49, 103, 138; future 161 agnosticism 94, 134, 136 Allais, Lucy 98n9 anthropology 84; anti-metaphysical 195n5; philosophical 186, 188– 189, 195, 198n26, 20 Archimedeanism 157–158, 160, 162, 164, 167, 187, 205 anti-descriptivism 179 antinomies 151 anti-representationalism 113, 176, 179–180, 195, 199 aporia 93 appearance xvii 5, 7, 36–38, 41, 44, 47, 49, 54–57, 59, 62, 66n19– 67n24,67n25, 71–72, 74–79, 85–86, 94, 96, 102, 104, 113, 129, 136–137, 151n20, 171, 192, 203 apperception 55, 66n17; synthetic/ analytic unity of 16; transcendental 47, 49–52, 66, 93; apperceptive act 51–52; unity of 48, 52 assertibility, warranted 7, 158, 176, 195 Aristotle 2, 65, 75, 156–157, 170, 196, 206 autonomy; metaphysical 149n7; of reason 192 axiom 164; axiomatics, geometrical 78 Azari, Bahar 117n7 Bacon, Michael 199n32 Bandini, Aude xixn4 Barnbaum, Deborah 4, 100, 200 Barrett, Lisa Feldman 117n7

224  Index Barth, Christian 150n9 belief 81–82, 120n26, 157–158, 167 Berkeley, George 66n15, 91, 139, 174 bifurcation thesis 7, 179, 181, 185 Blackburn, Simon 179 blind-spot 8n1 bluebird of happiness 111 Brandhoff, Boris xixn4 Brandom, Robert xi–xii, xviii, 4, 8n1, 21, 26n18, 100n25, 117n6, 120n29–30, 159, 176, 179–181, 190, 196n9–10, 199n32, 200n41 Brandt, Stefan 4, 206 Brassier, Raymond xi, 4, 25n10, 66n13, 95, 149n7, 200n43 Bremer, Józef 4, 101 Brentano, Franz 2 Breunig, Anke xixn4 Brooks, Daniel Stephen 205 Burge, Tyler xi Burgess, Alexis 120 Callanan, John 24 Cambridge Pragmatism 186 Cantor, Georg 196 capacity 1, 28, 89, 98n10; conceptual 89; dispositional 171; epistemic 12, 18, 32, 39, 47, 135, 163; finite 25, 28; of consciousness 1, 47, 56, 84; of languagings 144, 147, 161–162, 167; to form images 83; to produce pictures 173, 147; to represent 2 Cappelen, Herman 5 Carnap, Rudolf 206 Cassam, Quassim 24 Castañeda, Héctor-Neri 4 categories passim; Aristotle’s 65n5, 206; as meta-classificatory rules 28, 30, 32, 34, 39, 67n21, 204; Cartesian 36, 40, 73, 107; dynamical 89–90; epistemological 10, 154; for mental acts 37, 93; Kantian xx, 25n, 26–27, 29, 31, 34, 51, 65n7, 65n8, 86, 98n14, 129–130, 136, 170, 197n16, 202; mathematical 87, 89, 99; ontological 10–11, 154n33, 200n43; schematised 157, 171–175, 193; unschematised 28, 157, 168 categorial; classifications 31, 34, 43, 200n43; concepts 93, 175; features 80, 87, 89–90, 101n30, 202; form

30, 170, 197n18; frame 15, 93, 169, 170; knowledge 23, 26n23; ontology 101; schematisations 193; status 105, 107, 146; structure 78, 91–92, 132, 146, 204 causation xiii–xvi, causal passim; causality of reason 191 causes 118n17, 136, 143, 169, 175; forms of 76; noumenal 75; postulated 97; realm of causes 95, 96 Chisholm, Roderick 3 Christias, Dionysis xixn4, 4, 5, 26n19, 118n19, 150n12, 151n20, 153n30, 198n22 Churchland, Patricia 4 Churchland, Paul 4 classification, passim; acts of 35, 43, 52, 70, 106–108, 110, 116, 119; categorial 29, 31, 34, 43, 113, 163, 170; functional 6, 25n16, 65n4, 105, 117, 123, 133, 147, 154, 195, 203; rules for 11, 15, 150, 204, see categories Cobley, Paul 147 cognition passim coherentism 102, 112; coherence 12, 43, 53, 62, 64, 150, 195 colours 1, 65, 100 Conant, James 65n8 concept passim; concept-acquisition 107–109; concept-formation 107, 137, 170 conceptual, (non-) 77, 119, 142–143, 169, 171–172, see guidance; analysis 13, 14, 16, 21–23, 26; framework 22–24, 26, 32–33, 65, 83–84, 94, 98, 119, 135, 143, 151, 153–154, 156, 165, 171, 192–193, 196; scheme passim; system 7, 15, 19, 34, 100, 115, 119, 145, 156, 158, 160, 161, 162 conceptualisation 38, 71, 80, 84, 90, 95–96, 99n15, 100n21, 119n21, 139, 142, 144, 170–171 consciousness passim Cornman, James 26n21 Corti, Luca xixn4, 4, 22, 26n20, 99n17, 100n24, 206 counterparts, argument from 7, 151n21, 173, 204; concepts 100, 142, 160, 172, 188, 196n12, 198n23; objects 7, 23, 25n16, 26n22, 76, 117n11, 119n19, 119n21, 135–137, 142, 160, 162,

Index 225 172, 196; qualities 92, 98n10, 138, 143 co-variances 129, 143 Crane, Tim xi, 3 Cumpa, Javier 65n5, 99n16, 170, 200n43, 206 Dach, Stefanie xixn4, 4, 5, 22, 23, 26n19, 192, 205, 206 Daly, Christopher 5 Davidson, Donald 25n14, 194n1, 197n17 deduction 41, 48, 64n2, 65n8, 66n15, 100n22, 170 deflationism 177, 179, 180, 182, 199n30 Delaney, Cornelius Francis 4 Dennett, Daniel 3, 144 deontological; attitudes 188; cognition 149, 162; modalities 25n16 Descartes, René 35–37, 65n12– 66n15, 41; distinctions 36, 46, 55, 72, 76, 107; tradition 40, 65n11– 12, 191, 206 deVries, Willem 4, 92, 93, 101, 117, 119, 120, 148–149, 154, 196, 200, 206 dimension 16, 35, 40, 66, 79, 82, 203; basic 26n23, 153n29; causal 119n24, 151n21, 156, 167, 176, 178 see picturing; functional 21, 31, 34; horizontal vs vertical xiii–xiv, 6, 94, 110, 115–166, 125, 128, 131, 143–146, 185, 187, 201; reflexive 10, 190 D’Oro, Giuseppina 5 Dupré, John 205 Eerdekens, Fred 99n18 Einstein, Albert 198 embodiment 61, 63–64, 69, 73, 120n26, 125, 141, 206 empiricism xi, xii, 25n16–26n18, 104, 187 epistemology xii, xvii, xx, 3, 11, 15, 24n8, 31, 37, 79, 94, 160, 191, 197n17, 201–202; epistemic power xiv, xx 31, 34–35, 39, 132, 170, 204; epistemological turn 10, 170 Erkenntnis 7, 170, 135, 136 Escher, Maurits Cornelis 99n18 essentialism xii Euclid 119

existence xx, 2, 20, 30, 36–38, 41– 42, 53, 55, 61, 65n12, 70–76, 80, 91–93, 97n2, 97n3, 107, 120n28, 135, 138–139, 151–156, 166–167, 170–174, 193–197, see actuality, see postulation; in-existence 2, 65n12, 80 explication 19, 21, 23, 32, 45, 64, 66n19, 85, 103, 111, 123, 126, 129, 131, 148–149, 155, 188, 191, 203 expressivism 5, 8, 24n6, 122, 176, 177–180, 186–190, 200n39, 204 faculties 17, 42, 71, 76, 78, 81, 88 fallibilism 136, 153n28, 189, 193 Fink, Kevin xixn4 finitude 14, 17, 19–20, 25n11, 27–28, 33, 72, 135, 150n16, 176, 192–194 Fodor, Jerry 196n10 Förster, Eckart 24n9, 65n7, 151n17 foundationalism xii, 91, 111 frame 15, 91, 93, 98, 132, 144, 148, 154, 169, 171, 176, 190, 193, 201, 201, 204, 206 framework passim freedom 192 friction 38, 43, 102, 198 functionalism, see classification Gabbani, Carlo 4 games, see language Garfield, Jay 84 Giladi, Paul 206 Gironi, Fabio xixn4, 4 giveness 193, see Myth of the Given Gomes, Anil 175, 206 guidance 118, 206 Haag, Johannes 4, 8n1, 20, 24n7, 65n7, 93, 98n13, 100n26, 118n16, 119n25, 148n1, 151n17, 197n16, 198n23, 200n42, 205 Hall, Everett 26n23 hallucination 97, 160 Harman, Graham 3 Haug, Matthew 5 Haugeland, John 3 Hegel, Georg Wilhelm Friedrich 16, 94, 120n30, 147, 206 Heidegger, Martin xv, 155 Herz, Marcus 24n1, 171, 177 Hicks, Michael xixn4, 4, 5, 26n19, 101n32, 206

226  Index higher-order; explanation 190; genera of objects 27, 30; knowledge 113; reflection 189; question 2 Hoeppner, Till 151n20 horizontality 6, 105, 108–110, 114– 116, 119n22, 123–128, 132, 137, 146, 148, 152n26, 154n23, 161, 162–163, 177, 182, 185, 198n25, see non-relationality Horstmann, Rolf-Peter 66n14 Horwich, Paul 180, 199n30 Huebner, Bryce xixn4 Hübner, Johannes 67n23 Hume ii, viii, 26n23, 68n32, 176 Husserl, Edmund 23, 24, 26n23 hylomorphism 39, 170 idealisation 196, 200n45 idealism 8, 14, 19, 37–41, 62, 64, 66n15, 68n29–69, 82, 87, 94, 116, 191, 203, 206; dogmatic 7, 57, 66; transcendental 5, 36, 38–40, 63, 66, 68, 103, 186, 203 ideality 50, 81–82, 89, 114, 121, 146 ideals 206 illusion 10, 170 image-models 66n14, 78, 81–89n14, 99, 114, 146, 203 imagination 55, 66n14, 78–79–89, 98n13–99n20, 133 immanence xv, 7, 23, 40, 112, 128– 147, 150, 152n26, 159, 169–176, 185, 187, 194, 201–205; criterion 23, 128, 134, 142 inference xiii, xiv, xix, 25n12, 41, 65, 72, 100, 106, 108, 110, 112, 114, 119n20, 120n26, 126, 147, 174, 178–170, 181–183 innatism 18, 83 instrumentalism, 120n26, 152n22 intention 2, 6, 198 intentionality passim internalism 114, 116 introspection 3, 119 intuition 24n5, 48–49, 51–52, 54, 66n14, 67n22, 70–78, 87–90, 98n14, 99n16n20, 109, 114, 117n11, 119n21, 124, 135–136, 138, 149n7, 170–173, 193, see perceptual taking; forms of 17, 88, 175; intellectual 65n7, 71, 134–137 invariance 13, 61, 64, 90–91, 142, 172, 190, 194, 197, 203–204

isomorphism 7, 115, 118n11, 91, 123, 129, 137, 141, 143–144, 146–147, 152n24, 153n28, 154n34, 167, 169, 173–174, 178, 184, 195n3, see picturing Janus-face xiii, 123, 150n10, 168, 173 Jacob, Pierre 2 Kambartel, Friedrich 150n9 Kant, Immanuel passim; Kantianism, analytic and continental 4 Klemick 4, 192, 205 Koch, Anton Friedrich 200n37 Koons, Jeremy xixn4, 4, 176, 192, 205, 206 Kraut, Robert 4, 180, 200n36 Kreines, James 65n7 Kripke, Saul xixn1, 158 Kukla, Quill Rebecca 4 labyrinth 11, 130 Lance, Marc 4 Landy, David xixn4, 4, 68n32, 100n22, 200n44, 2005 language passim, see transcendental linguistics; entry- and exit-moves 92, 114–115, 118n12, 120, 141, 146, 153n28, 172, 185; extra-linguistic 108, 111, 118n12, 124, 127, 134, 150n16, 173, 178, 180, 197n18, 199n33; intra-linguistic 182–184; language-game 122, 154, 158, 160–161, 164, 187–188, 195, languaging xiii, xx, 6, 7, 119n20, 122–128, 137, 141–148, 150n10, 151n18, 152n26, 168, 173, 197n19 lawfulness 35, 98n10 Legg, Catherine xixn4 Leibniz, Gottfried Wilhelm 99 Lepore, Ernest 196n10 Levine, Steven 4 Lewis, David xixn1 levels of abstractions, see abstraction Loeffler, Ronald 199n32 logic 22, (non-) 26n18, 105, 119, 206; general 17, 28–29, 33; transcendental 12, 34–35 Lyubenova, Vera xixn4 Macbeth, Danielle 98, 154 Maher, Chauncey 4, 8

Index 227 manifold 46–58, 66n18–67n22, 67n24, 75–77, 85–86, 97n2, 97n3, 97n6, 98n8, 110, 118n19, 135, 143, 163, 192, see sense-impressions mapping 161; representational 148n2 Marras, Ausonion 3 Marx, Karl xx materialism, nonreductive 149n7 Matsui, Takaaki 151, 198 McDowell, John 4, 8n1, 38, 116, 118, 120, 190, 194, 198, 200 meaning, theory of 9, 18–20, 25n16, 65n4, 95, 102–126, 148n2, 155– 165, 175, 180–181, 190, 196–197; meaning-statements 109–115, 117n2, 118n12, 119n20–120n28, 152n26, 163, 187, 197n17 Meillassoux, Quentin 4 meta-classification 12, 23, 25n16, 27, 28, 30, 34, 39, 65n4, 109, 129– 130, 193, 202, see classification, see rules meta-knowledge 14, 20, 27, 174 metalinguistic analysis 19, 21, 24n8, 94, 107–109, 115–116, 125, 184, 188, 190, 203, see theory of meaning metaphilosophy 3, 5, 23, 206 meta-principle 126 metasemantics 15, 114, 120n28, 188 meta-theory 10, 13–14, 36 method of projection 123, 129, 144, 147, 149n5, 150n13, 152n24, 163–164, 166–168, 172, 189, 196n14, 197n20 Miguens, Sofia 65n7 Miller, Steven 206 Millikan, Ruth Garrett 4, 117n5, 152n25, 154n31, 196n9, 200n37 Misak, Cheryl 206 modality 13, 25, 95, 100, 190 Moore, Adrian 175 Moore, George Edward 140 Myth of the Given 25n16, 78, 88, 91, 105, 107–109, 127, 129, 143, 146, 193, 197n17, see foundationalism Nagel, Ernest 205 naturalism, (object/subject) xii, xiv, xviii, xix, 95, 116, 149n7, 152–153, 176–179, 186–187 natural-linguistic objects xiii, 6, 121, 129, 137, 145, 147, 153n27,

161–164, 169, 172–173, 178, 184, 189 neo-pragmatism 25n10, 155 Newton, Isaac 157, 198 Nietzsche, Friedrich xviii, 176 nominalism 4–5, 94, 196n14, 204 non-relationality 6–7, 26n22, 102–118n12, 122–134, 145–148, 152, 161–162, 170, 173–180, 187, 197, 199, 203 normativity xiii–xiv, xvi–xvii, 2, 94, 111, 120n27, 123, 148–149n7, 152n25, 153n27, 153n28, 162, 164–165, 167–168, 173–174, 178, 182, 185, 196n9, 198n25, 202, 205; autonomy of the normative xii; irreducibility of 190, 196n11; of correctness conditions 154n34; of justification 2, 5; of thought 65n3 noumena 74–76 Nunziante, Antonio 4 observation xx, 19, 72, 149, 150, 169, 187, 200n45 Olen, Peter xixn4, 4, 5, 206 ontology xviii, 9–11, 35, 37, 40, 46, 64, 97n3, 101n30, 121, 130, 134, 139, 150n14, 154, 158, 166, 170, 186, 194, 200, 202, 205 Oppenheim, Paul 205 ought-to-be rules 19, 87, 119, 127, 134, 153n27, 163, 169 O’Shea, James xii, xiii, xixn3, 4, 64n3, 65n5, 92, 99n16, 101n30, 105, 119n25, 120n31, 126, 148n2, 149n8, 150n10, 152n25, 153n27, 154n24, 163, 170, 195n3, 196n9, 198n23, 200n41, 205 Overgaard, Søren 5 Parent, Ted 4 passivity 38, 47, 71, 74, 88, 95, 97n3, 128, 136, 139, 148, 163, 170, 191 pattern xiii, xviii, 62, 82, 95–96, 112–113, 120n26, 127, 129, 147–148, 154n34, 163, 174, 189, 196n9, 196n11, 197n20, 200n43 Peirce, Charles Sanders 141, 167, 194, 195 Peirceans, 133, 137, 156, 160, 162, 166–167, 169, 175, 194, 195n7, 197n15 Peregrin, Jaroslav 25n14

228  Index perceptual taking 6, 51–56, 60–64, 78, 84, (mis-) 91–93, 100–113, 119n20, 129, 152n24, 170–171, 173, 192, 202, see intuition Pereplyotchik, David xixn4, 4, 100n29, 200n41 Perler, Dominik 2 phenomena 94–95, 103, 117, 151, 178–179, 204–205 phenomenalism 6, 8, 53, 66n16, 91, 93–94, 98n12, 121–123, 129, 133–135, 138–139, 142, 148–152, 163, 174, 193, 201, 204 phenomenology 23–24, 26n23, 47, 79–82, 87, 95, 98, 99, 170 picturing 7–8, 91, 121–122, 131, 141–169, 173–200 Pitt, Joseph 4 Plato 111 positivism 187 postulation xiii–xvi, 32, 70, 75–76, 96–98n11, 104, 144, 157, 167, 186, 188, 190, 192, 198n23, 200n43 presupposition xvi, 2–3, 5, 11–12, 14–15, 18–20, 33, 39, 43–47, 52, 57, 59, 61–63, 70–71, 83, 113, 119n20, 148n1, 152, 165, 170, 171, 174, 193–194, 202; presuppositional analysis 2, 5, 10, 17–19, 27–28, 30–31, 33, 40, 42, 63, 95, 98n11, 114, 148n2, 152n25, 176, 178, 196n8, 201–202 Price, Huw xviii, 5, 8, 24, 25n10, 26n19, 120n33, 154n31, 154n32, 155, 159n5, 176, 196n9, 198n25, 199n30–39, 200 process 7, 11, 14, 22, 32, 39, 41, 62, 64, 78, 86, 99n15, 100n21, 198n23, 202; pure processes 22, 193, 200n43, 205 proper sensibles 89 properties xiii–xiv, 1, 20, 23, 29, 33, 39, 80, 88, 90, 92, 94, 99n19, 106, 111, 118n17, 199n30, 199n33 proposition xx, 111, 121, 168, 182, 195n7, 196n12 proprieties 159, 162 psychology, transcendental 2, 18, 25n15–27, 31–32, 35–37, 64– 65n6, 70, 94, 96, 100n24, 107, 148, 169–170, 191, 197n21, 202, 204 Putnam, Hilary 198n22, 157, 158, 205

qualia 205 quietism, metaphysical 186, 197n17, 198n26 Quine, Willard Van Orman xi, xii, xixn1, 3, 104, 111, 117n4, 199n30 rationalism xii, 35 realism 5, 7, 24n8, 66n19, 67n24, 122–129, 148, 157, 160–167 173, 178–179, 190–192, 195, 200; dogmatic 24; empirical 69; naïve 88, 91, 115; scientific 5, 8, 23, 149, 155, 178, 190–191, 194n2, 194n7, 198n23, 204–205; transcendental 26, 38, 53, 57, 62, 67, 87, 94, 96n1, 98n12, 102, 122, 127–129, 138–139, 142, 148n1, 149, 151n20, 163, 177, 191, 197n20, 200n39, 201–204 reality 65, 75, 77, 91, 94, 135, see actuality; empirical 92, 135, 193; formal 65n12, 72, 75; independent xx, 77, 97n2, 135, 172, 180; manifest 98n12; material 53, 196n14; objective 41, 151n20 receptivity 39–40, 47, 71, 73–76, 89, 116, 128–129, 135–136, 143, 172 recipes 27, 83–86, 99n20 Redding 8n1, 120n30, 177, 186, 188, 206 reducibility xii, 25n15, 54, 94, 123, 148n2, (ir-) 153n27, 196n11, 199n33 reductionism 25n15, 94, 148n2, 149n7, 152n22, (ir-) 153n27, 192, 196n11, 199n33, 205 reference 18, 23, 25n10, 48, 71, 91, 100, 115, 120n28, 148n2, 159, 160, 165, 199n33; causal 91, 108, 131; demonstrative 15, 21, 26n22, 89, 91, 93, 142, 170–172; forms of direct 90–100n23, 118n11, 119n20, 128, 203; intentional 7, 9, 104, 115, 127–130, 137, 145, passim; meta-linguistic reading 148, 180; objective 9, 103; possible 17, 71, 107, 114, 154n31, 171; rulebased 39; überhaupt 13, 25, 44, 171; referentiality 20, 98, 110, 108, see truth-makers; extra-linguistic 108; power of judgements 132 reflexion 167, 196n13, see levels of abstraction; terms of 94, 114; transcendental 88

Index 229 refutation 66 regularities xvi, see rules Reider, Patrick xixn4, 4, 154n31, 200n37 representation passim; e-representation 155, 182–185, 188–190, 199n33–34; i-representation 182–185, 189, 199n33; transcendental object of 24, 52, 57 representationalism xii, xvi, 155–156, 159–160, 162, 178, 180–181, 195n5, (anti-) 195n5, 199; representationality 23 representable; contents 39; objects 27; system of representables 128, 147 representing / represented 16, 29, 37, 54, 72, 76, 86, 90–91, 106, 117n8, 118n14, 138, 151n19, 174, 181, 200n43; acts of 6, 16, 37–41, 44, 46–47, 49, 52, 55–58, 67n24, 73–74, 76–77, 102–104, 107, 138, 151, 202–203; conceptual, 91, 133, see image-models; contents 5, 39, 47, 56–59, 68n29, 89, 117, 135, see appearances; intuitive 9, 54, 135– 136, 138; object of 54, 60, 200n43; sensible 17–18, 37, 41, 50, 71, 73, 193; unconscious 77, 89, 135 Riemann, Bernhard 119 Rorty, Richard xii, xvi, 3, 8, 24n10, 25, 154n31–32, 169, 173, 176, 178, 180–181, 185, 187, 190, 194n1, 194n3, 195n3, 195n5–6, 196n8, 197n15, 197n17, 199n31, 199n33, 201 Rosefeldt, Tobias 98n9, 151n20 Rosen, Gideon 19 Rosenberg, Jay 3, 4, 8n1, 24n7, 53, 65n10, 66n14, 92, 117n4, 119n25, 153n28, 158, 168, 176, 194n3, 195n4, 205 Rosenthal, David 205 Rottschaefer, William 4, 198n23 Rouse, Joseph 5, 154n31, 200n37 rules 11, 39, 51–55, 60, 84, 96, 112, 126, 131, 147, 153, 168, 189, see recipes; classificatory 15, 28, 67n22, 193, 202, 204; correspondence 193, 200n54; default rules of inference 25n12, 112, 134, 153; empirical concept-formation 107, 170; generic 25, 112–113, 150n15, 171; linguistic 90, 126, 153n27–28; necessary 85; of apprehending 44,

see image-models; of projection 163; of synthesis 45, 51 64, 84–85, 131; ought-to-be / ought-to-do 67n23, 87, 134, 162–63; semantic 111, 115, 163, 168, 195n3; rule-governed/guided 46, 48, 53, 59, 61–62, 65n3, 85, 112, 119n20, 120n27, 150n10, 153n28, 196n9, 198n25 Ryleans 154, 198 Sachs, Carl xvi, xixn4, 4, 5, 100n24, 118n19, 150n12, 152n25, 176, 206 Sacilotto, Daniel xixn4 schemata 83–88, 93, 99n15–16, 118, 203; schematisation 28, 31–34, 37, 86, 90, 93, 95, 136, 138, 157, 168, 171–173, 175, 193, 204 scheme-content distinction 197 Schlösser, Ulrich 68n28 Schönwälder-Kuntze, Tatjana 5 Schulting, Dennis 118n15 Seibt, Johanna 4, 20 semantics xv, xviii, 31, 36, 104, 179, 188, 197n17 sensa, 22, 97, 100 sensation 22, 70, 78, 99, 205 sense-impressions xiii, xx, 32, 47, 66n18, 66n19, 72, 75, 77–78, 90, 97n3, 97n6, 98n7, 98n8, 98n10, 100n24, 118n17, 119n22, 133, 143, 203 sensibility 17, 18, 37–41, 50, 74, 78, 81, 88–89, 104, 168, see receptivity sensing 88, 98n10, 99n15, 99n19, 124 Shapiro, Lionel xixn4, 4, 8n1, 25n14, 117n8, 118n16, 120n27–28n, 194n2, 199n30 sheer manifold of sense 75, 110, 118n19, see guidance Sherman, Brett 120n28 Sicha, Jeffrey 4, 153n28 Siewert, Markus 3 skepticism 14, 155–159n3, 168, 195 Socrates 15 sortal 30, 114, 117n9 spontaneity 39, 71, 73–74, 78, 137, 148, 163 Stang, Nick 98n9, 151n20 Stekeler-Weithofer, Pirmin 25n12, 150n9 Stephenson, Andrew 175 Stern, Robert 24n3

230  Index Stovall, Preston xixn4, 4, 149n6, 176, 196n7 Strawson, Peter 24, 65n6 Stroud, Barry 68n30 synthesis 28, 45–49, 51–53, 55, 59, 62–64, 66n19, 67n24, 85–86, 107, 192, 202; empirical 46, 64, 202 teleology 24, 151n17 teleosemantics 200n37 theology 92, 140, 172 theory-construction 12, 36, 70, 82–83, 167, 205 this-suches 93, 109, 118, 119n22, 135, 170, see intuition, perceptual taking Thomasson, Amie 65n5, 206 Tomasello, Michael 3, 84 transcendental linguistics 11, 19, 25n15, 32, 97n1, 103, 123, 129, 147, 163, 173, 190, 192, 197n21, 202, 206 transcendental methodology 5, 9, 13, 36, 39, 70, 97n2–3, 100n24, 102, 105, 121, 146, 148, 149n8, 152n26, 163, 170, 178, 197n20, 200n39, 202, 206 truth passim; analytic 9, 13–15, 24n4; as s-assertibility xv, 7, 68, 145, 153n29–30, 157, 163, 179, 181, 183–185, 197n17, 198n22, 199n30; concept of 7, 9–10, 14, 25n12, 26n23, 65n9, 68n29, 130, 136–137, 139, 145, 166–167, 181– 182, 195n3, 195n6, 196n7, 197n15, 198, 199n30; divine 135–138; ideal 137, 197; matter of factual 68n31,

143–146, 151n21, 165, 181, 184n2, 187, 193, 198n25; T-schema 197; truth-makers 110, 187; truth-talk 120n28 unknowability, Kantian 66n19, 67n24, 73, 92, 94, 96–97n5, 137, 140 utopia xxn13 van Fraassen, Bas 92 verticality 105, 108, 111, 113, 116, 128, 136, 152, 163, see horizontality vinyl xvi, 124–125 vocabulary 160, 161, 178–179, 180, 186, 188, 195n6; fact-stating 179; intentional 106; i-representational 189; mathematical 179; Modal 179; moral 179; perceptual 22; representational 180; semantic 112, 120n27–28, 180 Wanderer, Jeremy 4 Westphal 24n2, 25n13, 53, 97n1 Willaschek, Marcus 98n9, 151n20 Williams, Michael 4 Williamson, Timothy xixn1 Wilson, Mark 92 Wimsatt, William 205 Wittgenstein, Ludwig 67n23, 121, 144, 153, 156, 158, 206 Wolf, Michael Padraic 4 Wolfendale, Peter xixn4 world-directedness 48, 177, 190 world-story 119n20, see method of projection Wunsch, Thomas 66