Sellars and the History of Modern Philosophy 9781138065680, 1138065684

Table of contents :
Introduction: The Lingua Franca of Thought: Willfrid Sellars and the History of PhilosophyPart I: Sellars and Modern Philosophy1. Sellars on DescartesChristian Barth2. The Lingua franca of Nominalism: Sellars on LeibnizAntonio M. Nunziante3. Sellars and Hume on the Ontological Status of Theoretical-Explanatory EntitiesDavid Landy4. Sellars' Interpretive Variations on Kant's Transcendental Idealist ThemesJames R. O'Shea5. Hegel After Sellars: Conceptual ConnectionsLuca CortiPart II: Sellars and the Beginning of the Contemporary Age6. Peirce and Sellars on Non-Conceptual ContentCatherine Legg7. Sellars and Frege on Concepts and LawsDanielle Macbeth8. `We pragmatists mourn Sellars as a Lost Leader': Sellars' Pragmatist Distinction Between Signifying and PicturingCarl Sachs9. The Varieties and Origins of Wilfrid Sellars' BehaviorismPeter Olen10. Sellars and Carnap: Science and/or MetaphysicsCarlo Gabbani11. Sellars and Wittgenstein, Early and LateGuido Bonino and Paolo Tripodi12. Wilfrid Sellars and Roy Wood Sellars: Theoretical Continuities and Methodological DivergencesFabio GironiConclusions13. Thinking with Sellars and Beyond Sellars: On the Relations Between Philosophy and the History of Philosophy

Citation preview

Sellars and the History of Modern Philosophy

This edited volume systematically addresses the connection between Wilfrid Sellars and the history of modern philosophy, exploring both the content and method of this relationship. It intends both to analyze Sellars’s position in relation to singular thinkers of the modern tradition, and to inquire into Sellars’s understanding of philosophy as a field in reflective and constructive conversation with its past. The chapters in Part I cover Sellars’s interpretation and use of Descartes, Leibniz, Hume, Kant and Hegel. Part II features essays on his relationship with Peirce, Frege, Carnap, Wittgenstein, American pragmatism, behaviorism and American realism, particularly his father, Roy Wood. Sellars and the History of Modern Philosophy features original contributions by many of the most renowned Sellars scholars throughout the world. It offers an exhaustive survey of Sellars’s views on the historical antecedents and metaphilosophical aspects of his thought. Luca Corti is FCT Post-Doctoral Fellow at Mind, Language, Action Group at the University of Porto, Portugal, and at the International Center for Philosophy at the University of Bonn, Germany. He has published two books and several articles on Kant, Hegel and contemporary Hegelisms, as well as on Sellars and Sellarsian themes, including Senses and Sensations: on Hegel’s Later Picture of Perceptual Experience (2018), Conceptualism, Non-Conceptualism, and the Method of Hegel’s Psychology (2016), Ritratti hegeliani (2014), Crossing The Line: Sellars on Kant on Imagination (2012). Antonio M. Nunziante is Associate Professor at the University of Padua, Italy. His research is in the history of ideas and is mainly focused on issues concerning naturalism and normativity in the Early Modern Philosophy (Leibniz), in the Classical German philosophy (Kant, Hegel) and in the pre-analytic American philosophy (early American naturalism). His works include: Infinite vs. Singularity. Between Leibniz and Hegel (2015), The “Morbid Fear of the Subjective,” Privateness and Objectivity in Midtwentieth Century American Naturalism (2013), Lo spirito naturalizzato. La stagione pre-analitica del naturalismo americano (2012), Representing Subjects, Mind-dependent Objects. Kant, Leibniz and the Amphiboly (with A. Vanzo, 2009), Individuals, Minds and Bodies: Themes from Leibniz (ed., “Studia Leibnitiana”, 2004).

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

Sellars and Contemporary Philosophy Edited by David Pereplyotchik and Deborah R. Barnbaum Pragmatism and Objectivity Essays Sparked by the Work of Nicolas Rescher Edited by Sami Pihlström The Quantum of Explanation Whitehead’s Radical Empiricism Randall E. Auxier and Gary L. Herstein Peirce on Perception and Reasoning From Icons to Logic Edited by Kathleen A. Hull and Richard Kenneth Atkins Pierce’s Speculative Grammar Logic as Semiotics Francesco Bellucci Pragmatism, Pluralism, and the Nature of Philosophy Scott F. Aikin and Robert B. Talisse Pragmatism and the European Traditions Encounters with Analytic Philosophy and Phenomenology Before the Great Divide Edited by Maria Baghramian and Sarin Marchetti Community and Loyalty in American Philosophy Royce, Sellars, and Rorty Steven A. Miller Sellars and the History of Modern Philosophy Edited by Luca Corti and Antonio M. Nunziante For a full list of titles in this series, please visit www.routledge.com

Sellars and the History of Modern Philosophy Edited by Luca Corti and Antonio M. Nunziante

First published 2018 by Routledge 711 Third Avenue, New York, NY 10017 and by Routledge 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN Routledge is an imprint of the Taylor & Francis Group, an informa business © 2018 Taylor & Francis The right of the editors to be identified as the authors of the editorial material, and of the authors for their individual chapters, 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 A catalog record for this book has been requested ISBN: 978-1-138-06568-0 (hbk) ISBN: 978-1-315-15954-6 (ebk) Typeset in Sabon by Apex CoVantage, LLC

Contents

Introduction: The lingua franca of Thought. Wilfrid Sellars and the History of Philosophy

1

LUCA CORTI AND ANTONIO M. NUNZIANTE

PART I

Sellars and Modern Philosophy

13

  1 Sellars on Descartes

15

CHRISTIAN BARTH

 2 The lingua franca of Nominalism: Sellars on Leibniz

36

ANTONIO M. NUNZIANTE

  3 Sellars and Hume on the Ontological Status of TheoreticalExplanatory Entities

59

DAVID LANDY

  4 Sellars’s Interpretive Variations on Kant’s Transcendental Idealist Themes

79

JAMES R. O’SHEA

  5 Hegel After Sellars: Conceptual Connections

97

LUCA CORTI

PART II

Sellars and the Beginning of the Contemporary Age

117

  6 Peirce and Sellars on Nonconceptual Content

119

CATHERINE LEGG

vi Contents   7 Sellars and Frege on Concepts and Laws

138

DANIELLE MACBETH

  8 “We pragmatists mourn Sellars as a Lost Leader”: Sellars’s Pragmatist Distinction Between Signifying and Picturing

157

CARL B. SACHS

  9 The Varieties and Origins of Wilfrid Sellars’s Behaviorism

178

PETER OLEN

10 Sellars and Carnap: Science and/or Metaphysics

197

CARLO GABBANI

11 Sellars and Wittgenstein, Early and Late

216

GUIDO BONINO AND PAOLO TRIPODI

12 Wilfrid Sellars and Roy Wood Sellars: Theoretical Continuities and Methodological Divergences

233

FABIO GIRONI

Conclusions255

13 Thinking With Sellars and Beyond Sellars: On the Relations Between Philosophy and the History of Philosophy

257

DIONYSIS CHRISTIAS

Index277

Introduction The lingua franca of Thought. Wilfrid Sellars and the History of Philosophy Luca Corti and Antonio M. Nunziante Wilfrid Sellars: Philosophy in Relation to Its Past In recent decades, Wilfrid Sellars has been progressively recognized as a central figure for contemporary philosophy: his ideas have been shown to offer great potential, deeply influencing many areas of the current philosophical discussion in debates touching subfields ranging from philosophy of mind and action to philosophy of science and semantics. Sellars’s followers— which are usually divided along a spectrum that moves from ‘left-wing’ to ‘right-wing’—have developed his insights in various directions. They have highlighted the many (and sometimes opposing) faces of his philosophy. In this sense, as James O’Shea has recently claimed, it is no longer necessary to justify ranking Sellars among the most important philosophers in twentieth century (O’Shea 2016, 1).1 What renders Sellars such a unique figure, however, is not only his philosophical insights on topics like meaning, norms or science, or his unique, systematic philosophical attitude, but the way that Sellars came to elaborate his views: namely, through constant dialogue with the history of philosophy. His confrontation with philosophers of the past is constitutive of his philosophical work, and thus essential to understanding Sellars’s philosophy. As Sellars himself stated in the Preface of his Essays in Philosophy and its History, I cannot conceive that my views on such topics as abstract entities, mental acts, induction, and the relation between theoretical and practical reasoning would have taken the form they have, if they had taken form at all, if I had not devoted as much time and energy to teaching and research in the history of philosophy as I did to these topics an sich betrachtet. (EPH, vii) Sellars was one of the first thinkers in the context of the so-called analytic tradition to value and engage in study of the past. When he arrived at the University of Iowa in 1938, he oversaw all of the university’s classes on history of philosophy. During that period, he says,

2  Luca Corti, Antonio M. Nunziante I became increasingly convinced of the importance of the subject and also came to see that much of the current literature on the subject was imperceptive and uninformed. The probing of historical ideas with current conceptual tools, a task which should be undertaken each generation, was long overdue. (AR, 290) From that period on, Sellars pursued pioneering work that served as a continuous and dynamic engagement with the history of philosophy. His work is notably based on the slogan that “philosophy without the history of philosophy is if not empty or blind, at least dumb” (SM, 1). There are a number of reasons that this aspect of Sellars’s thinking is interesting and worth discussion. First because, to best understand Sellars’s thought, it is fundamental to situate its relationship to the history of philosophy, for—according to Sellars himself—the origin of his positions is often rooted in such a dialogue. Second because, thanks to this dialogue, Sellars was able to originally connect some central themes of the analytic tradition (the question of meaning and functional role semantics, for instance, as well as the role of inference, intentionality and conceptual activity) to classical themes from the modern philosophical tradition. In many respects, Sellars was among the first philosophers to inaugurate a dialogue between analytically inspired thinking and the history of philosophy. In so doing, he was able to view the tradition in a new light, drawing insights that are important for modern and contemporary thinkers. The paradigmatic example is his treatment of Kant, which was key to Sellars’s elaboration of his own views on semantics. Within Sellars’s system, as he wrote, 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, 285) This kind of attitude can be seen on other issues too, for instance, Sellars’s variations on the classical notion of ‘analogy,’ or his reflections on the distinction between formal and objective reality as it appears in Descartes (SM, 31). By taking such a stance, Sellars initiated a dialogue that has in various ways been the source for many important contemporary works that bring classic ideas back to the stage of current philosophical thought—including, the rediscovery of Kant advanced by both ‘right-wing’ and ‘left-wing’ Sellarsians, for instance, the Rortyan project, the works of Jay Rosenberg and his pupils, the Pittsburgh Neo-Hegelians and Danielle MacBeth’s work on Frege. All of these philosophers take Sellars’s ideas and attitude as a starting point.

Introduction  3 Third, as some of the essays in this volume aim to show, Sellars’s ideas can still open up interesting new insights on classic and more contemporary readings of Descartes, Leibniz and Hume but can also be fruitfully related to pragmatism and ‘new realism,’ adding new perspectives to the work that has already been done in these areas. Finally, from a metaphilosophical point of view, Sellars tries to develop an interesting approach to the history of philosophy itself, one which is methodologically interesting and—although not unproblematic—deserves further attention in light of contemporary concerns on the relation between philosophy and its history. This volume explores these various strands of this complex relationship between Sellars and the history of philosophy, throwing light on various modes, methods, and contents to illuminate the stakes of this dialogue.

Philosophy and Its History: Sellarsian Views in Context Philosophy has a unique and often troubled relationship with its past, one which ranges between two extremes: what we might call ‘historiodicy’ or ‘extreme historicism,’ on the one hand, and ‘historiophobia’ on the other. A negative metaphilosophical attitude towards the past is a quite widespread phenomenon among philosophers, old and new. Starting at least from Descartes, the idea that study of the past is, at best, not useful—if not downright dangerous—had a certain success, until the time of Husserl and beyond. In this perspective, one is intoxicated by the past and has to free oneself from the burden of past mistakes and learn to think on one’s own. Descartes, most famously, wanted to start anew, guided by the motto “be my own guide.”2 This attitude is considered, among other things, the hallmark of analytic philosophy. Some articulations of this perspective have become classic, including John Searle’s reported statement, echoing the Cartesian attitude, that “I am an analytic philosopher. I think for myself” (Mulligan 2003, 267). Such an anti-historical attitude often involves defense of a strong division of labor between historians and philosophers. Here the famous dictum by Quine has become equally paradigmatic: “There are two sorts of people interested in philosophy, those interested in philosophy and those interested in the history of philosophy.”3 This view was common in Sellars’s time, and it influenced the institutional organization of the field. As Sellars recalls in his Autobiographical Reflections, within American philosophy departments of the 1940s, “it will be remembered . . . that the history of philosophy was not only neglected, there was an active campaign to delete it, or at least downgrade it, as a requisite for the Ph.D.” (AR, 290). It was during such debates over the question of whether more history of philosophy was worth being taught that another key episode occurred: prominent philosopher Gilbert Harman hung the controversial sign on his door, “Just say no to the history of philosophy!”4 Historiophobia was thus part of the beginning of analytic philosophy, or at

4  Luca Corti, Antonio M. Nunziante least part of its founding myth. That this was the context in which Sellars was working is what makes him stand out so sharply. Such an uneasy relationship with the past has become part of analytic philosophy self-image, to the point that, as Hans-Johann Glock notices, even “historians of the analytic movements . . . regularly deplore its lack of historical self-consciousness” (Glock 2008b, 867). From that initial period on, however, many things have changed: the rise of analytic histories of philosophy and analytical approach to classical texts (including works such as Bennet’s and Strawson’s on Kant, Dummet’s Frege, Mate’s work on Leibniz, as well as the works of Sellars himself, some of which are now classics) has opened new perspectives and ways of dealing with the past for analytic philosophers. The same goes for the history of analytic philosophy, which shows that analytic philosophy itself has a particular relationship with its past: it does not develop ‘in a vacuum’ but has its own pantheon of philosophers and canon of texts and its own set of problems. In this sense, it can be defined as a proper tradition (Beaneay 2013a; Cottingham 2005; Crane 2015). Finally, the work of Richard Rorty gave additional impetus to the reappraisal of history within the so-called analytic world, together with the work of Charles Taylor and, of course, Bernard Williams’s plea for more historically self-conscious work in philosophy (Rorty 1979; Williams 2002, 2006). Several of these scholars have brought arguments for the idea that the two enterprises of philosophy and history of philosophy cannot ultimately be separated and that it is impossible to do philosophy without engaging, in some way or another, with its past.5 The growing interest in history in the analytic camp has generated a lively debate with historians of philosophy that involves not only contents but also questions about the methodology for doing history of philosophy and the possibility of intersection between the two domains. These questions necessarily involve metaphilosophical issues concerning the nature, aims and scope of philosophy’s “uses of the past.” To what extent can philosophy make use of its past? How should it be done? What are the different approaches? Are there perennial questions in philosophy? Why should we care about history? Several volumes inquiring into these areas and defending opposite positions have appeared.6 Others include interesting contributions from the so-called continental tradition, where the bounds between philosophy and history of philosophy were never loosened up (LaerkeSmith-Schliesser 2013). Within this context, Sellars is highly interesting: not only because he was a pioneer, but also because he is an engaging case study from a methodological point of view.

Sellars and the History of Philosophy: Questions of Method Before addressing Sellars’s approach, it is helpful to contextualize the various existing ways of understanding the history of philosophy. Generally speaking, the current debate on the history of philosophy—especially in

Introduction  5 the contemporary Anglophone context—can be usefully divided into two approaches, which have been called ‘assimilationist’ and ‘contextualists.’ The terminology is not fixed, and many positions occupy a place somewhere between these two extremes.7 ‘Assimilationists’ think that past philosophers can, in a relevant sense, be treated as interlocutor for contemporaries. Another less negatively connoted name for this general attitude is ‘conversational’ (Rorty-Schneewind-Skinner 1984). The conversational approach is not a new one and can be found as early as in Descartes, who famously wrote that reading good books is like having a conversation with the most distinguished men of past ages—indeed, a rehearsed conversation in which these authors reveal to us only the best of their thoughts. (Descartes 1985, 113) Among those who have developed this metaphor in more recent times are Paul Grice (“I treat those who are great but dead as if they were great and living, as persons who have something to say to us now”—Grice 1986, 66) and scholars like Jonathan Bennett, a prominent representative of this approach (Bennett 2001, I, 1). Such an attitude is often (but not always) accompanied by other traits. First, it goes together with an attempt to translate past idioms into contemporary terms. Bennett, for instance, is confident that we understand past philosophers like Kant, “only in proportion as we can say, clearly and in contemporary terms, what his problems were, which of them are still problems and what contribution Kant makes to their solution” (Bennett 1966). The possibility of translating past philosophical vocabulary and insights into a current idiom is highly debated but constitutes an important part of many histories (RortySchneewind-Skinner 1984; Glock 2008b). Second, many ‘conversational’ philosophers think history of philosophy can be of use for current debates by providing arguments and new insights for them. Third, the idea of translation is therefore often (but not always) connected with the idea that, once a translation is performed, we will be rewarded with interesting ‘arguments’ from the past philosophers, which can be injected into current discussions.8 Note that these three characteristics do not necessarily always appear together, though often they do, and they have been taken as hallmarks of what is called ‘analytic history of philosophy.’ The dangers of such an approach are as many as its advantages and have been variously highlighted. The first is anachronism, or distortion and misunderstanding of past texts. This charge goes together with the allegation that such an approach to looking at the past results in a mere ‘mirroring’ of one’s existing position, retrojecting one’s present interests onto history, without really attending to history and context. Various scholars have called this a “programme of flattening the past into the present in the name of understanding it” (Ayers 1978, 55).

6  Luca Corti, Antonio M. Nunziante This lack of sensitivity for differences, in fact, produces “mythological histories” or “historical mythologies” instead of proper histories (Vermeir 2013, 54; Garber 2005, 131). On the opposite pole of this position are ‘contextualist’ historians of philosophy. Historians of this kind are suspicious of the ‘conversational’ approach and defend a conception of the history of philosophy that is not driven by any sort of current concerns. Garber and many others call this a ‘historical history of philosophy’ (putting emphasis on the first term), and contrast it with ‘philosophical history of philosophy.’ First, such a history does not aim prima facie to be useful for contemporary philosophy nor is it driven by the aspiration to contribute to it. Contextualists instead try to do history “without conducting an anachronistic search for contemporary issues in the past” (Osler 2002, 533).9 Second, from this perspective, past figures should rather be studied for their own sake and translation into our contemporary idiom is secondary: the most important thing is to understand the figures’ “own vernacular” (Watson 2002; Ayers 1978). Such a history is therefore often further qualified as “unapologetic” and “disinterested” (Garber 2005, 129). There is nothing necessarily philosophical in such histories: it is just history. Third, this approach maintains that understanding past philosophers— and their statements—occurs by situating them in their context. How broad that context should be, and what it should include, is matter of debate, but the period’s scientific theories, political and religious beliefs, social structures, and other historical contingencies are often mentioned.10 Practitioners of this kind of history believe they can correct widespread historical mistakes and misrepresentations. For instance, it turns out, to use two slogans, that ‘Descartes was not an epistemologist’ and that the ‘Magna Charta is not the origin of modern constitutionalism.’ This approach often (but not always) implies other corollaries: for instance, the denial that there is a fixed set of ahistorical and atemporal philosophical problems to be addressed by a discipline called ‘philosophy’ with a specific essence. Of course, this approach raises its own difficulties. The alleged ‘purity’ of such an account can be questioned. The same goes for the idea of understanding a text without translating it into any other idiom, which has been accused of incoherence (Rorty-Schneewind-Skinner 1984, 7).11 A history so conceived—critics argue—necessarily commits to some version of the myth of a historical given. Needless to say, generalizing both accounts results in caricatures, but they are useful tools for both mapping the current controversy and understanding the position of Sellars.12 For against the backdrop of this tension between conversationalism and contextualism, Sellars represents a challenging case. His most famous quote declares the history of philosophy as the “lingua franca of thought” that makes communication between philosophers possible: “A dialectical use of historical positions is the most reliable way of anchoring arguments and making them intersubjectively available” (AR, 292). But communication is certainly not Sellars’s principal interest in

Introduction  7 the history of philosophy. An intensive analysis of traditional texts, according to Sellars, can provide new insights into current debates by revealing hidden assumptions and, perhaps, misconceptions that affect the contemporary philosophical discussion in various ways. To do this, Sellars takes up the traditional idea of ‘conversation with the ancients’ that we have seen at work in Descartes. As he states: Philosophy is a continuing dialogue with one’s contemporaries, living and dead, and if one fails to see oneself in one’s respondent and one’s responded in oneself, then is confrontation but no dialogue. (EPH, vii) As this quote interestingly illustrates, however, Sellars grants a ‘right of reply’ to the ancient, and this can only be attained through historical sensitivity and the attempt to penetrate the context of a thinker. One can find several statements in which he praises such an approach: The true measure of success as an historian of philosophy lies in the ability to think oneself into the divergent and often radically incommensurable conceptual frames which have made their appearance since Thales stumbled into his well and saw the stars.13 To penetrate past ideas, one must resist the temptation to simply project oneself into the position of the other. Sellars seems thus to embrace an engaging position with respect to the two extremes of contextualism and conversationalism. On the one hand, he likes to talk about philosophia perennis— intimating the existence of perennial problems—and he is even reported to have encouraged his students to treat historical positions ahistorically, exercising the practice of imagining philosophical hybrids: working out the position of “Leibnoza,” “Spinhobbes” or even “Deskant” (Watson 2002, 526). On the other hand, however, Sellars seems clearly conscious of the importance of historical work. He recognizes that, in attributing some contemporary distinctions to past philosophers, we are—to use his words—“guilty of anachronism” (ML, 162; SM 31). He therefore appears to be aware of the historical difference between what a philosopher argued and what we can say by developing her position. To be sure, Sellars is not, in any sense, what we would call today a historian of philosophy. He does not dwell on historical details or careful historical contextualization, but he recognizes the importance of such work and he seems to defend the possibility of a relationship to the past that is contextually sensitive and philosophically conversational. His position appears to be closer to those who think that the difference between contextualism and conversationalism is not mutually exclusive: a conversational approach does not need to mean ventriloquizing one’s position into the past, just as a contextual approach does not need to be simple repetition of idioms that are not ours and that we cannot fully grasp.

8  Luca Corti, Antonio M. Nunziante Finally, another interesting aspect of Sellars’s approach to the history of philosophy is how he balances between two perspectives he adopts on the past. On the one hand, he looked with favor at a micro-level analysis: he suggested focusing on very specific questions in an author and inquiring into the smallest details. On the other hand, however, he also pled for a broader perspective on history that would tackle general questions from which a philosopher’s whole production of thought could be viewed from a different and original angle. This—to put it in his own terms—would provide a synoptic view on the whole. According to Sellars, both the ‘macro’ level and the ‘micro’ level are necessary: In the history of philosophy, as in philosophy itself, we must continually shift between analysis and synopsis, embracing the extremes of both. To stay at or near the middle is to be safe but uninspired. To give Kant’s dictum one more twist: analysis without synopsis is blind; synopsis without analysis is empty. (PP, 5) Thus, Sellars is not only an interesting philosopher in his own right, but also worth studying for his way of approaching the history of philosophy. Tim Crane has recently argued that the notion of tradition is necessary to understanding philosophical questions and the range of their possible answers (Crane 2015). Only by interrogating this tradition will it be possible to make ourselves more intelligible to ourselves—and Sellars, which is now part of our tradition, has understood it.

Synopsis of the Volume Our volume widely explores Sellars’s relationship to the modern philosophical tradition (Part I) and to the tradition of the early nineteenth-century philosophical world (Part II). Part I, which consists in five chapters, addresses Sellars’s interpretation and use of Descartes, Leibniz, Hume, Kant and Hegel. Part II, consisting in seven chapters, analyzes Sellars’s relationship with Peirce, Frege, Carnap, Wittgenstein, American Pragmatism, Behaviorism and American Realism (as expressed, in particular, by Sellars’s father, Roy Wood Sellars). The contributors explore how Sellarsian ideas can open up new perspectives on the history of philosophy (MacBeth, Landy)—reading classical authors with a new light but also focusing on where Sellars gets things wrong, correcting him with more accurate historical insights (Barth, Nunziante, Leggs). Others engage in dialogue and reconstruct how Sellars’s ideas have influenced recent scholarship in past figures on the history of philosophy (Corti, O’Shea) or how Sellars’s ideas are related to Pragmatism, American Realism, Behaviorism (Sachs, Gironi, Olen) or make use of ideas coming from his contemporaries such as Wittgenstein or Carnap (Bonino-Tripodi, Gabbani). In this way, the

Introduction  9 volume extends until recent years, in various directions. The final contribution is explicitly devoted to the metaphilosophical assumption guiding Sellars’s approach to the history of philosophy (Christias). The contributors come from different backgrounds and employ a wide range of approaches to philosophy and its history: some are ‘historically’ oriented, others are more ‘philosophically’ oriented. We think that this variety of orientations constitutes the richness of the volume and mirrors the variety of possible ways of doing philosophy and its history, as well as the flourishing number of insights inspired by Sellars’s philosophy. The aim of the volume is to attain a stereoscopic vision, which has Sellars at its center—but we are conscious that this is just a first step, one that needs to be expanded and extended.

Notes 1 On the rise of Sellars, see O’Shea (2007, 2016, 2010, 2011), deVries (2009) Olen (2016), MacBeth (2014), Reider (2016), Pereplyotchik and Barenbaum (2016) and Gironi (2017). 2 On this point, see for instance Garber (2005, 131): “Modern philosophy was born with the rejection of the history of philosophy, and the status of the study of past philosophy has been problematic ever since. In his Discourse on the Method, Descartes began by rejecting what he learned in school in favour of what he could discover for himself through reason and experience.” See also Lepenies (1984, 145). 3 Originally quoted in MacIntyre (1984, 39–40). 4 On the story of Harmann’s dictum and its contextualization, see Sorell (2005). For other examples of such an attitude, see Glock (2008a) and Putnam (1997); see also Fodor’s boast (2003, 1–2) of being able to write a book on Hume without having actually read Hume, and Ayers (1978, 76 ff). 5 See Taylor (1984) and MacIntyre (1984), as well as more recent philosophers belonging to the analytic tradition, such as Tim Crane (2015). Such an attitude is more common in the so-called continental tradition. 6 For methodology, see the debates and contributions contained in Reck (2013), Sorell and Rogers (2005), Rorty, Scheewind and Skinner (1984), Piercey (2003, 2009), Rée, Ayers and Westonby (1978), but also Glock (2008a, chap. 4, 2008b) Laerke, Smit and-Schliesser (2013). 7 We follow here the terminology of Laerke, Smith and Schliesser (2013, 1–6). For further distinctions, see what follows and Glock (2008a, 2008b), as well as Garber (2005). 8 The title of one of the relevant book series—The Arguments of the Philosophers, ed. Ted Honderich (London: Routledge, 1970)—exemplifies such an attitude. That does not mean, however, that philosophy should be seen as only a repository or storehouse of arguments (as other scholars suggest). 9 Osler’s position is stated eloquently: “Analytic history of philosophy is bunk.” 10 Some authors tend to construct the methodological opposition between the two ways of doing history as ‘rational reconstruction’ versus ‘historical reconstruction,’ see Rorty (1984). For a history of the concept, see Beaney (2013b); for a critique see Ayers (1978). 11 According to Rorty (1984): “Translation is necessary if ‘understanding’ is to mean something more than engaging in rituals of which we do not see the point” (Rorty 1984, 52, n. 1). An argument in favor of translation can also be found in Glock (2008b).

10  Luca Corti, Antonio M. Nunziante 12 Some scholars take a middle-ground position, arguing that the difference between the two is a matter of grade and not of nature (cf. Vermeir 2013). This debate has grown and has various strands; for a finer-grained approach, see Glock (2008b) and a series of articles in the Journal of Philosophy, and see Watson (2002) and Osler (2002) and Garber (2001, 2003, 2005, 2013), who defend a contextualist approach. 13 Sellars 1934 (unpublished). Review of “A History of Western Philosophy”, in Wilfrid S. Sellars’s Papers, 1899–1990, ASP.1991.01, Archives of Scientific Philosophy, Special Collections Department, University of Pittsburgh, Box 39 Folder 9.

References Ayers, Michael. 1978. ‘Analytical Philosophy and the History of Philosophy’. In Philosophy and Its Past. Edited by Michael Ayers, Jonathan Reé and Adam Westoby, 42–66. Hassocks, UK: Harvester Press. Beaney, Michael. (Ed.). 2013a. The Oxford Handbook of the History of Analytic Philosophy. Oxford: Oxford University Press. Beaney, Michael. 2013b. ‘Analytic Philosophy and Historyof Philosophy: The Development of the Idea of Rational Reconstruction’. In The Historical Turn in Analytic Philosophy. Edited by Erich H. Reck, 231–60. London: Palgrave Macmillan. Bennett, Jonathan. 1966. Kant’s Analytic. Cambridge: Cambridge University Press. Bennett, Jonathan. 2001. Learning from Six Philosophers: Descartes, Spinoza, Leibniz, Locke, Berkeley, Hume, 2 vols. Oxford and New York: Oxford University Press. Cottingham, John. 2005. ‘Why Should Analytic Philosophers Do History of Philosophy?’ In Analytic Philosophy and History of Philosophy. Edited by Tom Sorell and John G.A. Rogers, 25–42. Oxford: Clarendon Press. Crane, Tim. 2015. ‘Understanding the Question: Philosophy and Its History’. In Experimental Philosophy, Rationalism, and Naturalism: Rethinking Philosophical Method. Edited by John Collins and Eugen Fischer, 72–84. London: Routledge. Descartes, René. 1985. ‘The Philosophical Writings of Descartes’. Translated by John Cottingham, Robert Stoothoff and Dugald Murdoch, vol. I. Cambridge: Cambridge University Press. deVries, Willem A. (Ed.). 2009. Empiricism, Perceptual Knowledge, Normativity, and Realism: Essays on Wilfrid Sellars. Oxford: Oxford University Press. Fodor, Jerry. 2003. Hume Variations. Oxford: Oxford University Press. Garber, Daniel. 2001. ‘Au–delà des arguments des philosophes’. In Comment écrire l’histoire de la philosophie? Edited by Yves-Charles Zarka, 231–45. Paris: Presses Universitaires de France. Garber, Daniel. 2003. ‘Towards and Antiquarian History of Philosophy’. Rivista di storia della filosofia, 58/2: 207–17. Garber, Daniel. 2005. ‘What’s Philosophical About the History of Philosophy?’ In Analytic Philosophy and History of Philosophy. Edited by Tom Sorell and John G.A. Rogers, 129–46. Oxford: Clarendon Press. Garber, Daniel. 2013. ‘Does History Have a Future? Some Reflections on Bennett and Doing Philosophy Historically’. In Debates in Modern Philosophy: Essential Readings and Contemporary Responses. Edited by Stewart Duncan and Antonia Lolordo, 347–62. New York and London: Routledge. Gironi, Fabio. (Ed.). 2017. The Legacy of Kant in Sellars and Meillassoux: Analytic and Continental Kantianism. New York and London: Routledge.

Introduction  11 Glock, Hans-Johann. 2008a. What Is Analytic Philosophy? Cambridge: Cambridge University Press. Glock, Hans-Johann. 2008b. ‘Analytic Philosophy and History: A Mismatch?’ Mind, 117: 867–97. Grice, Herbert Paul. 1986. ‘Reply to Richards’. In Philosophical Grounds of Rationality: Intentions, Categories, Ends. Edited by Richard E. Grandy and Richard Warner. Oxford: Oxford University Press. Laerke, Mogens, Smith, Justin, Halldòr, Erik and Schliesser, Eric. (Eds.). 2013. Philosophy and Its History: Aims and Methods in the Study of Early Modern Philosophy. Oxford: Oxford University Press. Lepenies, Wolf. 1984. ‘Interesting Questions’ in the History of Philosophy and Elsewhere’. In Philosophy in History: Essays on the Historiography of Philosophy. Edited by Richard Rorty, Jerome B. Schneewind and Quentin Skinner, 141–72. Cambridge: Cambridge University Press. MacBeth, Danielle. 2014. Realising Reasons. Oxford: Oxford University Press. MacIntyre, Alasdair. 1984. ‘The Relationship of Philosophy to Its Past’. In Philosophy in History: Essays on the Historiography of Philosophy. Edited by Richard Rorty, Jerome B. Schneewind and Quentin Skinner, 31–48. Cambridge: Cambridge University Press. Mulligan, Kevin. 2003. ‘Searle, Derrida and the Ends of Phenomenology’. In John Searle. Edited by Barry Smith. Cambridge: Cambridge University Press. Olen, Peter. 2016. Wilfrid Sellars and The Foundation of Normativity. London: Palgrave Macmillan. O’Shea, James R. 2007. Wilfrid Sellars: Naturalism with a Normative Turn. Cambridge and Malden, MA: Polity Press. O’Shea, James R. (Ed.). 2010. ‘Naturalism, Normativity, and the Space of Reasons’. Special Issue of the International Journal of Philosophical Studies, 18. O’Shea, James R. 2011. ‘How to Be a Kantian and a Naturalist About Human Knowledge: Sellars’ Middle Way’. Journal of Philosophical Research, 36: 327–59. O’Shea, James R. (Ed.). 2016. Wilfrid Sellars and His Legacy. Oxford: Oxford University Press. Osler, Margaret J. 2002. ‘The History of Philosophy and the History of Philosophy: A Plea for Textual History in Context’. Journal of the History of Philosophy, 40/4: 529–33. Pereplyotchik, David and Barenbaum, Deborah R. (Eds.). 2016. Sellars and Contemporary Philosophy. New York and London: Routledge. Piercey, Robert. 2003. ‘Doing Philosophy Historically’. Review of Metaphysics, 56: 779–800. Piercey, Robert. 2009. The Uses of the Past from Heidegger to Rorty. Cambridge: Cambridge University Press. Putnam, Hilary. 1997. ‘A Half Century of Philosophy, Viewed from Within’. Daedalus, 126, 1: 175–208. Reck, Erich H. (Ed.). 2013. The Historical Turn in Analytic Philosophy. London: Palgrave Macmillan. Rée, Jonathan, Ayers, Michael and Westoby, Adam (Eds.). 1978. Philosophy and Its Past. Hassocks: Harvester. Reider Patrick. (Ed.). 2016. Wilfrid Sellars, Idealism, and Realism Understanding Psychological Nominalism. London: Bloomsbury.

12  Luca Corti, Antonio M. Nunziante Rorty, Richard. 1979. Philosophy and the Mirror of Nature. Princeton: Princeton University Press. Rorty, Richard, Schneewind, Jerome B. and Skinner, Quentin. (Eds.). 1984. Philosophy in History: Essays on the Historiography of Philosophy. Cambridge: Cambridge University Press. Sellars, Wilfrid. 1934. (unpublished). Review of “A History of Western Philosophy”, in Wilfrid S. Sellars’s Papers, 1899–1990, ASP.1991.01, Archives of Scientific Philosophy, Special Collections Department, University of Pittsburgh, Box 39 Folder 9. Sellars, Wilfrid (ML). 1959. ‘Meditations Leibnitziennes’. In Philosophical Perspectives. Edited by Wilfrid Sellars, 153–81. Springfield, IL: Charles Thomas Publisher. Sellars, Wilfrid (PP). 1967. Philosophical Perspectives. Springfield, IL: Charles Thomas Publisher. Sellars, Wilfrid (SM). 1968. Science and Metaphysics: Variations on Kantian Themes. Atascadero, CA: Ridgeview Publishing Company. Sellars, Wilfrid (EPH). 1974. Essays in Philosophy and Its History. Dordrecht: D. Reidel Publishing Company. Sellars, Wilfrid (AR). 1975. ‘Autobiographical Reflections: (February 1973)’. In Action, Knowledge and Reality. Edited by Hector Neri Castañeda, 277–93. New York: Bobbs-Merrill. Sorell, Tom. 2005. ‘On Saying No to History of Philosophy’. In Analytic Philosophy and History of Philosophy. Edited by Tom Sorell and John G.A. Rogers, 43–59. Oxford: Clarendon Press. Sorell, Tom and John, Rogers G.A. (Eds.). 2005. Analytic Philosophy and History of Philosophy. Oxford: Clarendon Press. Taylor, Charles. 1984. ‘Philosophy and Its History’. In Philosophy in History: Essays on the Historiography of Philosophy. Edited by Richard Rorty, Jerome B. Scheewind and Quentin Skinner, 17–30. Cambridge: Cambridge University Press. Vermeir, Koen. 2013. ‘Philosophy and Genealogy: Ways of Writing History of Philosophy’. In Philosophy and Its History: Aims and Methods in the Study of Early Modern Philosophy. Edited by Mogens Laerke, Justin Erik Halldòr Smith and Eric Schliesser, 50–70. Oxford: Oxford University Press. Watson, Richard A. 2002. ‘What Is the History of Philosophy and Why Is It Important?’ Journal of the History of Philosophy, 40/4: 525–8. Williams, Bernard A.O. 2002. ‘Why Philosophy Needs History’. London Review of Books, 17 (October): 7–9. Williams, Bernard A.O. 2006. ‘Descartes and the Historiography of Philosophy’. In The Sense of the Past. Essays in the History of Philosophy. Edited by Myles Burnyeat, 258–9. Princeton: Princeton University Press.

Part I

Sellars and Modern Philosophy

1 Sellars on Descartes Christian Barth

1. Introduction For Wilfrid Sellars, conversation with past figures is an essential aspect of doing philosophy.1 René Descartes belongs to his preferred dialogue partners. This is not the case because he considers Descartes’s views to be true, but because he thinks that Descartes is a philosopher whose views, although in many respects wrong, provide the opportunity to learn important insights.2 However, it is clear to Sellars that serious exegetical work has to be done before we can take advantage of this opportunity. It is thus no surprise that Sellars takes issue with Descartes’s views in several writings and lectures.3 This exegetical ambition distinguishes Sellars from twentiethcentury and contemporary authors such as Donald Davidson and Hilary Putnam who present themselves as radical critics of what they call the ‘Cartesian picture of the mind.’ Often, these authors operate rather carelessly in exegetical matters.4 Not so Sellars. Although we will find several of his exegetical claims disputable, there is no doubt about the carefulness and seriousness of his exegetical work. It provides the basis on which Sellars develops his insightful critique of Descartes’s views. In order to get an overview of Descartes’s conception of the mind, we will present his core claims and distinctions in the next section. In section 3, we will then expound and critically discuss Sellars’s exegetical claims about Descartes’s conceptions of intentionality and consciousness. Sellars’s critique of Descartes’s views that revolves around the Myth of the Given and the conflation of sensing and conceiving will be addressed in section 4. Section 5 summarizes the chapter’s main points.

2. Descartes on the Mind: Core Claims and Core Distinctions In his account of the mind, Descartes famously replaces the Aristotelian conception of intentionality as reception of forms by his theory of ideas.5 The backbone of this theory consists in a short list of core claims and distinctions. These claims and distinctions form the backdrop against which we will discuss Sellars’s exegetical views.

16  Christian Barth (a) Soul Monism: Mechanizing the living functions that Aristotelians attributed to the vegetative souls of plants and the sensory souls of animals,6 Descartes’s category of souls reduces to minds.7 (b) Intellectualism: The essential feature of the mind consists in the intellect.8 The will as well as the capacities to sense and to imagine merely are accidental capacities the mind as thinking thing exhibits.9 Hence, while there could exist a mind as pure intellect without will, sense and imagination, there could not exist a mind that wills, has sense perceptions or imagines without an intellect. Furthermore, for Descartes, this implies that all10 mental acts, even acts of willing, sense perception and imagining, essentially are intellective acts. Taking up a traditional term for the acts of the intellect, Descartes calls all mental acts ‘thoughts’ (cogitationes).11 (c) Ideas vs. non-ideational thoughts: In the 3rd Meditation, Descartes divides thoughts into subclasses, namely ideas in the proper sense and thoughts that are not properly called ‘ideas.’12 I will call the former ‘ideas’ and the latter ‘non-ideational thoughts.’13 The first class of thoughts, i.e., the class of ideas, includes all and only perceptions (in Descartes’s technical sense of the term), i.e., acts or operations of the intellect, which Descartes characterizes as essentially representational.14 It is less clear, however, what the second class of thoughts, i.e., the class of non-ideational thoughts, includes. As we will see, Sellars understands them as acts of will such as acts of desiring, fearing or judging, where these acts are considered in abstraction of any object they are directed at. Non-ideational thoughts would thus be dependent acts in the sense that they cannot occur without an idea that represents an object and which the act of the will is directed at. As an alternative to Sellars’s view, we will defend the reading that thoughts are complex thoughts composed of an idea and an act of will. Non-ideational thoughts would then be self-contained acts of desiring object X, affirming or denying that p, fearing object X, etc., that, in principle, can occur independent of any other act. (d) Ideas considered as act and considered as presenting an object: Ideas possess two aspects. The first aspect concerns the fact that ideas are acts (actus) or operations (operationes) of the intellect,15 the second that they represent an object. When considered under the former aspect, Descartes calls them ‘material ideas’ (idea materialiter); when considered under the latter aspect he calls ideas ‘objective ideas’ (idea objective).16 The notions of materiality and objectivity applied here are technical terms from scholastic vocabulary and have nothing to do with their usual connotations in contemporary parlance. They just pick out aspects under which ideas can be considered: as act and as representing an object. As Descartes points out, the notion of representing an object involved here does not imply the commitment that object O exists outside of the idea.17 This naturally suggests that the object represented

Sellars on Descartes  17 by the material idea is an immanent, i.e., inner-mental, object. Under this assumption, we are confronted with two possible readings of the according to the first reading, defended by Sellars, the immanent object is a relatum of two representation relations. They are represented by the material idea and they represent an object external to the idea (if such an object exists). The second reading proposes that the first relation is not one of representation, but of presentation. The material idea presents the immanent object to the mind. This second reading builds its case on two observations: the Latin term ‘repraesentare’ often means ‘presenting’ in the sense of “mak[ing] something immediately available” rather ‘representing,’18 and Descartes often uses the verb ‘exhibere’ (to exhibit, to present) in order to denote the relation between act and immanent object.19 We will come back to these two readings later. (e) Formal vs. objective existence: Descartes distinguishes between two kinds of existence, namely formal existence (existentia formaliter) and objective existence (existentia objective).20 Formal existence is, so to speak, the usual way in which things exist. Sellars calls it ‘existence simpliciter.’21 Objective existence, as Descartes explains, is the kind of existence that applies to things insofar they are in mental acts. Accordingly, what we think of ” being as it were in the [external] object of our ideas, exists [as immanent object] in the ideas themselves objectively.”22 For Sellars, the distinction between objective and formal existence is the core idea of conceptual representation because it gives expression to the fact that the things represented in conceptual representation as so and so need not exist simpliciter.23 (f) ‘All thoughts are conscious’: According to Descartes, minds are conscious (conscius) of all thoughts (cogitationes) that occur in them.24 In other words, there are no thoughts that are not conscious to the mind, not even during sleep or in coma.

3.  Sellars on Descartes on Intentionality and Consciousness 3.1 Intentionality Descartes is revolutionary in his mechanical natural philosophy and his rejection of scholastic hylomorphism, which he replaces by his dualism of substance types.25 One may be inclined to add Descartes’s theory of ideas, which is supposed to replace the Aristotelian account of intentionality as reception of forms, to the list of revolutionary turns in Descartes. For Sellars, however, in his account of intentionality “Descartes adheres to the scholastic tradition”26 such that we still find the scholastic distinctions alive in Descartes’s writings on this subject. But Descartes’s account is not only in substantial continuity with scholastic philosophy, but also with contemporary philosophy. Indeed, for Sellars, “the Cartesian distinctions . . . can be

18  Christian Barth translated into contemporary terms”27 because they are “less sophisticated counterparts of distinctions which are drawn with more or less rigor in those contemporary philosophies of mind which have been influenced by formal semantics.”28 This implies that an intensive analysis of Descartes’s views is not merely of historical interest, but has the prospect of revealing assumptions and, perhaps, misperceptions that effect the contemporary philosophy of mind. As Sellars sees it, Descartes presents us at best merely with a proto-theory rather than a full-blown theory of intentionality.29 But this proto-theory contains the approach that will dominate two centuries and will even leave its traces in Kant.30 With the notion of ideas, Descartes proposes a representational account of intentionality. Ideas are about things in the sense of being intentionally directed at them in virtue of the fact that they represent them. Sellars contends in BD that Descartes models mental representation on the intellective act of conceiving.31 However, this is not a fully adequate account of Descartes’s view since his intellectualism does not merely imply that mental acts of representing are to be modeled on the intellective act of conceiving; instead, he believes that representational mental acts essentially are intellective acts of conceiving.32 Sellars’s presentation in SM, in which he diagnoses a straightforward misunderstanding of sensation as a kind of conceiving, is more faithful to Descartes’s view.33 Sellars recognizes that Descartes applies two different metaphors in his account of mental representation: the image metaphor and the container metaphor.34 He thinks that the image metaphor is not powerful and does not even lead to a proto-theory because it involves strong negative analogies: images can instantiate the same properties as the represented things do. Therefore, in the case of images, representation of material things can be understood in terms of co-instantiation of properties. But mental acts of Cartesian immaterial substances cannot co-instantiate properties of material things. Accordingly, the image metaphor does not lead very far because representation by images does not help us in explaining mental representation. The second, more fruitful metaphor Sellars identifies is the container metaphor.35 By talking of the container metaphor, he refers to Descartes’s assumption that things exist ‘in’ ideas.36 Sellars is right that Descartes’s account is a proto-theory at best, since Descartes nowhere properly spells out the ‘in’-metaphor. By applying this metaphor, Descartes ‘explains’ representation of an external thing by an idea in terms of the objective existence of the thing in the idea; or, as Sellars puts it, in terms of the idea of containing the objectively existing thing.37 The technical notion of objective existence is supposed to make room for the prima facie puzzling thought that even material substances and their properties can exist in mental acts of an immaterial mind. The all-important notion of objective existence remains, however, a primitive and, indeed, mysterious notion in the Cartesian framework. According to Sellars’s reading, Cartesian ideas contain an immanent, i.e., inner-mental, object in virtue of which they represent the external object

Sellars on Descartes  19 they do.38 In addition, for Sellars, Descartes distinguishes between two relations of representation: first, mental acts represent their immanent object since “for a thing or modification to exist ‘in’ a mental act is for the latter to represent it.”39 Second, immanent objects represent, at least potentially, objects external to the idea. They do so if the thing existing objectively in the idea also exists formally external to the idea.40 Sellars’s reading implies that the objects in ideas are intermediaries between the representing mind and the represented object external to the mind. This is a classic reading of Descartes as a so-called ‘indirect realist.’ While the direct realist claims that the mind is in direct contact with external objects, the indirect realist states that the mind is in contact with them only via the direct contact with immanent objects. We find this reading of Descartes already in the works of Thomas Reid; it was affirmed in the twentieth century among others by Anthony Kenny and Richard Rorty.41 However, in the last decades the direct realist position has become popular and has been defended by several authors.42 The dispute between the proponents of both readings cannot be fully addressed here. However, it needs to be pointed out that the dispute between indirect and direct realist readings can be understood in at least two different ways which often are not sufficiently distinguished in the scholarly debate: first, the indirect realist might claim that the mind directly represents immanent objects, and only by virtue of representing them, it indirectly represents external objects. Second, the indirect realist may contend that the immediate objects of our beliefs and knowledge are immanent objects and that only by inference we indirectly acquire beliefs and knowledge about external objects. There is considerable textual evidence that Descartes is an indirect realist of the second sort.43 It is far from clear, however, that he also is committed to an indirect realist position of the first type. Sellars attributes to Descartes an indirect realism of the first kind. This follows from his contention that the material idea represents the objective idea and that by virtue of this it represents external objects.44 The Achilles heel of Sellars’s reading is his understanding of the relation between material and objective ideas as a representation relation. But as already mentioned in section 2, this reading is not without an alternative. Indeed, the alternative reading according to which the relation is one of presentation rather than representation is textually better supported as is witnessed by Descartes’s frequent use of the term ‘exhibere.’45 But if this is correct, it is open to Descartes to claim that the mind directly represents external objects. The mind does so by means of immanent objects that are directly presented (and thus conscious) to it. Accordingly, the container metaphor gives expression to the presentation and not to the representation relation. 3.2.  Consciousness (conscientia) Apart from the notion of representation, the second key notion of Descartes’s conception of the mind is that of consciousness (conscientia). In

20  Christian Barth BD, Sellars discusses Descartes’s conception of consciousness at length and begins his discussion with the important addendum to the 2nd Replies. In this addendum, Descartes states a set of crucial definitions, the second of which refers to the notion of idea. In BD, Sellars works with the following translation: “Idea is a word by which I understand the form of any thought, that form by the immediate awareness of which I am conscious of that said thought” (AT VII 160/HR I 159). This translation is problematic because it operates with a controversial identification of immediate perception with immediate awareness. A more literal translation reads: “Idea is a word by which I understand the form of any thought, that form by the immediate perception of which I am conscious of that said thought.”46 In the following, I will work with the latter translation. Furthermore, I will call the form of thoughts that is immediately perceived ‘formC.’ Sellars thinks that this definition by implication conflicts with Descartes’s account of thoughts in the 3rd Meditation. Since Descartes believes that all and only thoughts are immediately conscious to the mind, the definition of ‘idea’ seems to imply that all thoughts are ideas. But this seems to contradict the 3rd Meditation claim that only a subclass of all thoughts are ideas, while there also are thoughts of a second kind, namely non-ideational thoughts.47 It has to be noted, however, that Sellars’s diagnosis of a conflict is not compelling. This becomes clear if we take into account the distinction between material and objective idea from the Meditation’s Preface to the Reader48 and if we take seriously Descartes’s talk of form. In the definition of ‘idea,’ Descartes does not identify the idea with the thought of which the mind is conscious, but merely with ‘the form’ of the thought “by the immediate perception of which [the mind] is conscious of that said thought.” The way in which Descartes puts the definition leaves room for a view that coheres well with the 3rd Meditation.49 Applying the definition to thoughts that are ideas leads to the claim that the mind is conscious of ideas in virtue of immediately perceiving them. We are not forced to understand this as an appeal to a higher-order perception that has the idea as its object. Instead, we can—and, indeed, we should—read it as saying that the act of perceiving that constitutes the idea considered materially presents to the mind the idea’s immanent object. This means that in the case of ideas, the mind is conscious of only with regard to their immanent object. Applying the definition to non-ideational thoughts results in the statement that the mind is conscious of them in virtue of the act of perceiving that is part of the idea included in that thought. Again, the consciousness only concerns the thoughts’ immanent. Hence, in both cases, the notion of form Descartes is applying in the definition of ‘idea’ turns out to be the notion of objective idea from the Meditation’s Preface to the Reader. In other words, the notion of form in play is the notion of presenting an object. This reading does not only bring into coherence the definition of ‘idea’ of the addendum to the 2nd Replies with the 3rd Meditation, but also leads

Sellars on Descartes  21 to a different and arguably more adequate understanding of Cartesian consciousness (conscientia).50 Most scholars—Sellars is no exception51— understand Descartes as saying that in the case of non-ideational thoughts the mind is conscious of the kind to which the thought belongs: the mind is conscious of them as desires, judgments, fears, etc. The reading of the notion of form suggested here leads to a deflated view of consciousness in the sense that the scope of consciousness merely embraces the immanent object of thoughts, but not the kind of thought to which they belong. We will come back to this contrast between inflationary and deflationary readings of Cartesian consciousness later on. Sellars makes use of the conflict he diagnoses between the addendum of the 2nd Replies and the 3rd Meditation for an expositional purpose. When he later “return[s] . . . to the problem posed by the definition of the term ‘idea’,”52 he introduces a different reading of Descartes’s notion of things existing objectively in ideas. This new reading is supposed to solve the diagnosed conflict.53 As Sellars points out, Descartes understands conceiving of a triangle in terms of a triangle existing objectively in the act of conceiving, i.e., in the idea. Sellars claims, however, that “[t]he terminology is flexible” and adds that we can alternatively say that “the [act of conceiving] can even be said to be triangular.”54 In a nutshell, Sellars claims that Descartes’s talk of a triangle existing objectively in an idea is interchangeable with talk of the idea objectively exhibiting the character of being triangular. However, we do not find the latter way of talking in Descartes. He does not use terms such as ‘triangular idea,’ ‘green idea’ or ‘tree-like idea’ in order to denote ideas of a triangle, of green and of a tree. This is significant because Sellars’s way of reformulating Descartes’s position is not trivial. Sellars’s reformulation rules out that Descartes believes that in the case of ideas of being triangular, being green and being a tree, there must be an object, namely a substance, carrying these characters while objectively existing in the ideas. In view of Descartes’s way of expressing his position, this should be considered to be a life option. Hence, Sellars’s claim of terminological flexibility requires defense that he does not provide. Leaving this worry aside, Sellars eventually introduces his reformulation of Descartes’s position in order to save him from incoherence. He wants to point out that if we follow the character reading rather than the object reading, we can prevent him from incoherence in his account of the mind’s consciousness of its thoughts.55 Sellars’s line of thought goes as follows: for Descartes, all thoughts are conscious to the mind. According to the definition of ‘idea’ in the addendum to the 2nd Replies, minds are conscious of their thoughts in virtue of the immediate perception of their formC. What is this formC? Sellars suggests that it is “a character by virtue of which it is the sort of thought it is.”56 In the case of non-ideational thoughts such as volitions it is the character of being a volition such that consciousness of volitions is consciousness of them as volitions. Now, if we want to apply this reading of the definition of ‘idea’ we also need to identify the formC of ideas, i.e., the

22  Christian Barth character in virtue of which ideas are the ideas they are. And here Sellars’s character reading pays off. Assuming this reading, we can say that the formC of an idea of a triangle is its character of being objectively triangular. Minds are conscious of such ideas in virtue of being immediately perceiving their characters of being triangular. Hence, applying the character reading we can, Sellars claims, provide a coherent reading of the definition of ‘idea’: in the case of non-ideational thoughts as well as in the case of ideas, consciousness is grounded in the immediate perception of the thoughts’ “character by virtue of which it is the sort of thought it is.” If Sellars is right, Descartes defines a notion of idea that differs from the one he presents in the 3rd Meditation as well as in the Meditation’s Preface to the Reader. According to the definition, ideas are representational acts (the ideas from the 3rd Meditation) aspects of them (the material or objective ideas from the Preface), but characters of mental acts. This is, as Sellars acknowledges, a “wider definition of ‘idea’ ”57 compared to the notion of idea from the 3rd Meditation. But it resolves the apparent conflict between the definition of ‘idea’ in the addendum of the 2nd Replies and the notion of idea from the 3rd Meditation because in the addendum Descartes now turns out to define a notion of idea that differs from the one developed in the 3rd Meditation. There is no conflict because Descartes is concerned with different notions of idea. However, Sellars’s interpretation of the addendum and the 3rd Meditation conflicts with key passages in Descartes. As he remarks himself, according to his reading there is an important difference between the formC of ideas and the formC of non-ideational thoughts. The former exist objectively (e.g., being triangular) in the mind, the latter does so formally (e.g., being a volition). This implies that two very different mechanisms must be responsible for the mind’s consciousness of ideas and of non-ideational thoughts respectively. Minds are conscious of non-ideational thoughts in virtue of having a meta-perception of them. For instance, I am conscious of my act of desiring in virtue of having a separate meta-act of perceiving this act as an act of desiring. In the case of ideas, by contrast, the act of perceiving that constitutes the material aspect of the very idea is responsible for the consciousness of the formC, for instance, the idea’s character of being objectively triangular. But then we end up with a hybrid conception of Cartesian consciousness, or rather with two radically different kinds of consciousness applying to ideas and to non-ideational thoughts, respectively.58 However, Descartes’s characterizations point to a unified account that equally applies to all thoughts. Consider again the definition of ‘idea’: Descartes claims that it is the immediate perception of the thought’s form that is responsible for being conscious of them. There is no place in his work at which he gives reason to think that the immediate perception is radically different in the cases of ideas and of non-ideational thoughts. I suspect that things start going in Sellars’s interpretation at an early stage, namely in his account of the division of thoughts in the 3rd

Sellars on Descartes  23 Meditation. As we have seen, Sellars understands Descartes as sub-dividing thoughts into representational ideas such as the idea of the sun and non-representational acts of the will such as the mere act of desiring.59 Now, two points count against this reading of the 3rd Meditation passage: first, the wording of the passage strongly indicates that Descartes understands the second kind of thoughts to be self-contained complex acts such as desiring object X or judging affirmatively that p rather than to be merely the abstracted and dependent acts of desiring and affirming. The passage reads: Other thoughts have various additional forms: thus when I will, or am afraid, or deny, there is always a particular thing which I take as the object of my thought, but my thought includes something more than the likeness of that thing. (AT VII 37/CSM II 25–6; emphasis added) Descartes here uses the scholastic term ‘similitude’ (likeness) in order to denote the idea’s immanent object that is ‘like’ the supposed external thing. As the passage makes clear, the ‘other,’ i.e., the non-ideational, thoughts do not only include ‘various additional forms’ such as the act of willing, of fearing or of denying, but also ‘the likeness of that thing,’ i.e., the immanent object that exists objectively in an idea of that object. Non-ideational thoughts are self-contained complex thoughts involving an idea. Second, in this passage Descartes intends to distinguish between two kinds of thoughts in the context of investigating which thoughts are truthevaluable. His general aim is to find out which mental acts can be false.60 Hence, Descartes is interested in whether the desire to eat ice cream, the fearing of an approaching lion or the judgment that the sun is rising can be true or false. He surely does not raise the absurd question whether the mere act of desiring, fearing or judging can be so. Thus, we find strong evidence that Descartes intends to distinguish in the 3rd Meditation between simple, representational thoughts (ideas) and complex thoughts that involve an idea and a dependent act stemming from will (non-ideational thoughts). With this understanding of the classification of thoughts, a unified account of Cartesian consciousness becomes available. We can now understand him as saying that thoughts in general are conscious to minds because thoughts are or include ideas and because ideas are perceptions that present immanent objects to the mind. It is these perceptions that Descartes refers to in the definition of ‘idea’ and that he claims to be responsible for the mind’s consciousness of its thoughts. However, this also means that the ubiquitous Cartesian consciousness only concerns the immanent objects of thoughts, but not the kind to which these thoughts belong.61 In KPT, Sellars understands the consciousness of mental acts as a nonrepresentational affair.62 This partly conflicts with the reading presented in BD according to which consciousness of non-ideational thoughts rests on a

24  Christian Barth meta-perception and, thus, is representational. Perhaps, Sellars presents Descartes’s view differently for didactic purposes in the student lectures in KPT; or he only has in mind the consciousness of ideas, but not of non-ideational thoughts. Be that as it may, the way in which Sellars there understands the non-representational character of the consciousness of ideas also runs into difficulties. Sellars presents his reading in order to show that Descartes does not fall prey to a vicious regress.63 His account of Cartesian consciousness solves this problem, but at the cost of ignoring the definition of ‘idea’ from the addendum to the 2nd Replies. As we now know, in the definition Descartes states that consciousness is due to an immediate perception. But Sellars understands non-representational consciousness as a non-perceptual affair. He explicitly separates the non-representational consciousness from perception; the latter only grounds representational consciousness.64 But Sellars’s understanding of Cartesian consciousness of ideas as non-representational does not fit the wording of the definition of ‘idea’ that refers consciousness back to an immediate perception. The deflationary reading of Cartesian consciousness, which we developed earlier, is in better position in two respects: first, it avoids the regress because according to the deflationary reading Cartesian consciousness also does not rest on a higher-order perception, but on the act of perceiving that is the material aspect of the idea in question. Second, the deflationary reading does justice to the wording of the definition of ‘idea’ from the addendum to the 2nd Replies since it contends that Cartesian consciousness indeed rests on an immediate perception.

4.  Sellars’s Critique of Descartes 4.1.  The Myth of the Given To a large extent, the critical side of Sellars’s engagement with Descartes’s conception of the mind revolves around the famous Myth of the Given. As Sellars points out at the beginning of EPM, philosophers have been appealing to the Given against the backdrop of construing in specific ways the philosophical situations in which they find themselves. Sellars calls this way of construing the situations “the framework of givenness,”65 i.e., the situations are construed as frameworks within which the idea of the Given becomes attractive or appears even without alternative. Throughout history, philosophers construed very different situations as frameworks of givenness and identified different things as the Given in these situations.66 The Myth of the Given is, then, a general scheme that involves two variables that take on different values with regard to different philosophers: The first variable concerns the question which specific situation is construed as ‘the framework of givenness.’ The second variable concerns the question what is identified as the Given by the philosopher in the situation that is construed in this way.

Sellars on Descartes  25 Depending on which values the variables take on, we receive “various forms”67 of the Myth of the Given. Yet, in the first Carus Lecture, Sellars puts forward a hypothesis about the principle that “is, perhaps, the most basic form of what I have castigated the ‘The Myth of the Given’.”68 The principle reads: “If a person is directly aware of an item which has categorial status C, then the person is aware of it as having categorial status C.”69 If Sellars is right, what philosophers identify as the Given in all forms of the Myth satisfies this principle. In all cases, some item is thought to “impose [its categorial structure] on the mind”70 such that awareness of the item as having this categorial structure results. Importantly, in this process the mind does not actualize its conceptual capacities. The awareness in question is preconceptual, yet a classificatory awareness in which something particular is recognized as exhibiting a general feature. For Sellars, almost all philosophers71 construed in their work situations as frameworks of givenness. Descartes is no exception. Indeed, we will see Sellars claiming that Descartes appealed to the idea of the Given twice. The first situation in which Descartes appeals to it is his critical epistemic project culminating in the Meditations. In this text, the meditator’s aim is to figure out which of her cognitions are indubitable and certain such that they qualify as foundation of all other knowledge. In this situation, the meditator applies the skeptical method and eventually comes to identify her own current mental acts as providing this foundation. According to Sellars, the way in which Descartes develops his position fulfills the criteria of the Myth of the Given: the Cartesian mind is pre-conceptually aware of its inner acts as having the categorial status they have, i.e., as its inner acts. Indeed, for Sellars, Descartes puts forward an even stronger thesis: the mind is aware of its own inner acts as belonging to the kind of act to which they belong (judgments, willings, sense perceptions, etc.) and as being of the object of which they are. For instance, as Sellars understands Descartes, in the case of a perception of a triangle, “there would be a[n] . . . awareness of the perception of a triangle as a perception of a triangle.”72 And this is a classic form of the Myth of the Given. However, Sellars’s charge rests on his inflationary understanding of Cartesian conscientia according to which it implies awareness not only of the inner act’s object, but also of the kind to which the act belongs. Seen in light of the deflationary reading of Cartesian conscientia we developed previously, the charge reduces to a ‘mythical’ awareness of the object as the object it is, but not of the inner act as belonging to the kind to which it belongs. Yet, this reduced form of the Myth is correctly attributed to Descartes. Sellars believes that Descartes applies his Myth of the Given also in a second situation, namely when confronted with the task of explaining our possession of mental ideas. In BBK, he attributes to Descartes this view with regard to the idea of sensation: [B]y virtue of having sensations we experience sensations as sensations . . ., and . . . from this experience, by an act of so-called abstraction,

26  Christian Barth the intellect can acquire a non-analogical understanding of what it is to be a sensation. (BBK §25) If Sellars is right, for Descartes we acquire our mental ideas by abstracting them from our preconceptual awareness of our own current mental states. It is surprising, however, that Sellars attributes this position to Descartes, since Descartes is well known for his thoroughgoing innatism of ideas. As an innatist, Descartes would not claim that the idea of a sensation is abstracted from our awareness of sensations, but that the awareness of them provides the occasion to actualize the innate capacity to conceive of sensations as sensations.73 Even if we take Descartes’s innatism into account, one might still think that Sellars is right that Descartes exploits his Myth of the Given a second time. He does not do so with respect to the question of the possession of mental ideas, but with respect to their actualization. For, one might argue that Descartes can only account for the actualization of innate mental ideas by appealing to our preconceptual awareness of our mental acts as the acts they are. To put it metaphorically: how could the mind ‘know’ that the actualization of the innate idea of sensation (rather than a different innate idea or no innate idea at all) is in order if the mind were not aware of a current sensation as a sensation? However, we do not find textual evidence for this kind of view in Descartes. He nowhere explains the actualization of innate ideas by an appeal to consciousness and, in particular, by claiming that consciousness provides an awareness of mental acts as belonging to the kind to which they belong. Indeed, we do not find any detailed account of the actualization of innate ideas in Descartes at all. Of course, Descartes appeal to God and the way in which he arranged the world including the inner workings of minds. The innate ideas get actualized as they do because God created things the way he did. Yet, without a concrete account of how these inner workings and the actualization of innate ideas proceed, this appeal to God is not merely heavy metaphysical baggage, but also amounts to sheer hand-waving. In the end, we do not find anything that amounts to an explanation of the actualization of innate ideas in Descartes. This is theoretically dissatisfying, but it also implies that we think that Descartes understands consciousness in the way Sellars assumes. The abstractionist conception (as well as the innatist conception) of mental ideas implies that mental ideas are primary in the order of conceiving. More precisely, they are primary in the order of conceiving compared to ideas that apply to forms of thinking that take place in public verbal utterances, i.e., what Sellars calls ‘thinking-out-loud.’ For Descartes, the ubiquitous preconceptual consciousness of one’s own current mental acts immediately provides the mind with mental ideas (or provides occasion to actualize innate mental ideas). Our own mental acts belong to the first items of which

Sellars on Descartes  27 we have conceptions and the conceptions we have of them do not presuppose conceptions of any other things. Sellars emphasizes that there is a radically different conception of thought available which Descartes did not have at his disposal.74 This radically different conception avoids the Myth of the Given and innatism. The key moves are, first, to recognize that thinking-out-loud is an everyday phenomenon that belongs to our normal behavior and, second, that there is a concept of thinking-out-loud available that does not depend on the concept of inner thinking in the order of conceiving. This concept of thinking-out-loud is behaviorist in kind. Combined with an inferential-functional concept of meaning, it constitutes the core of what Sellars calls the “verbal behaviorist model of thinking” (VB model).75 The essential point is that the VB model could have been rationally formed and mastered independent of and prior to the concept of inner thinking. Sellars summarizes his anti-Cartesian strategy nicely in SK: 48. Many have thought that to explicate the concept of non-inferential knowledge of ‘what is going on in one’s mind at the present moment’, one must return to the Cartesian framework. From the point of view sketched in this lecture, the essential feature of the latter framework is that it denies that ‘thinking-out-loud’ makes sense save as analyzable into ‘thinking occurrent non-verbal thoughts and giving them expression in one’s verbal behavior’. In short, the Cartesian argues that the concept of thinking-out-loud includes the concept of thoughts as Cartesian inner episodes. According to the VB position, on the other hand, the concept of thinking-out-loud stands on its own feet as the primary concept pertaining to thought, so that if a concept like the Cartesian concept of thought episodes which are not propensities to think-outloud does turn out to be needed in giving a full account of ‘what thinking is,’ those who are inclined to accept something like the VB position would argue that it is a concept built on a VB foundation and is in some sense a derivative or secondary concept. As a useful parallel, consider the case of microphysics. Macro-objects, we say, if we are scientific realists, are really systems of microphysical particles. Yet our concepts of these particles are built—not on direct observation—but on a foundation of knowledge at the perceptual level. In short, though VB argues that even if there is a sense in which Cartesian thoughts are prior in the order of being to thinking-out-loud, the latter is prior in the order of knowing (SK 327). In the famous ‘Myth of Jones’ in EPM, Sellars rationally reconstructs how a theoretical concept of inner thinking could have been developed from the VB model and how this theoretical concept could then have gained a reporting first-personal use.76 In this way, Sellars gives an anti-Cartesian explanation of the development of our capacity to self-reflect and to gain knowledge

28  Christian Barth our own mental acts. Whereas in Sellars’s Descartes, mental acts are Given, i.e., minds have a direct, preconceptual awareness of them as what they are, Sellars presents a different story77 that is supposed to work without any Given, while displaying the rationality of our mental concepts and legitimizing our claims to have self-knowledge and first-person authority. 4.2.  Sensory Representation: Conflating Conceiving and Sensing Descartes applies the term ‘cogitatio’ in an extremely wide sense. In particular, not only purely intellective acts such as those of metaphysical and mathematical thinking count as thoughts, but also sensations and imaginations. We noted above that Descartes’s wide application of the term ‘cogitatio’ is due to his intellective conception of the mind, i.e., his view that the essence of the mind consists in the intellect and that, for him, all mental acts therefore are intellective acts. Sellars recognizes the intellectualist drive in Descartes’s conception of the mind and its effects on his conception of sensation, i.e., on mental acts such as the sensing of blue or the feeling of pain. He observes that Descartes explicitly refers to sensations as “confused thoughts of the external cause”78 and that he is committed to an intellectualist conception since he takes conceivings “as his paradigm of the modifications of the mind that represent.”79 In many texts, Sellars emphasizes this intellectualist shape of Descartes’s understanding of sensation. For instance, in EPM he explicitly calls it a “conceptualistic interpretation of sensation”80 and in SM, Sellars points out that Descartes rightly saw that sensations have a representational function, but did not understand the representationality of intellective perceptions and sensations “in the spirit of analogy, the positive being counterbalanced by the negative, but literally, the negative analogy construed as specific difference.”81 As Sellars rightly points out, Descartes’s approach results in a conflation of the non-cognitive representationality of sensations with the cognitive representationality of truly intellective acts and that he misconstrues the representationality of sensations as a specific form of the latter.82 If sensations represent as intellective acts do by means of objectively existing immanent objects, it is unclear why there is so much a phenomenal difference between conceiving and sensing; why there is so much a phenomenal difference between conceiving of a triangle and seeing a triangle, between conceiving of pain and feeling pain, and between conceiving of blue and sensing something blue.83 Descartes’s contention that sensory ideas are clear, but confused, while intellective ideas are clear and, at least potentially, distinct merely hints at a difference that he neither properly characterizes, nor explains.84 In other texts, however, Sellars develops a more positive account of Descartes’s position. In KPT, Sellars pushes the view according to which Descartes believes, first, that sensations and conceivings are both representational,

Sellars on Descartes  29 but, second, that only the representationality of conceivings amounts to a classificatory and cognitive awareness, while the representationality of sensations does not do so. Instead, sensations are representational states of which we are or, at least, can be aware, but they do not themselves amount to classificatory awareness.85 According to this reading, Descartes registers deep differences between the non-cognitive representationality of sensations and the cognitive representationality of conceivings, i.e., Cartesian perceptions, such that “we have that important distinction in Descartes, I was suggesting, between sensations and perceptions.”86 Most importantly, while conceivings contain an objectively existing object, sensations do not do so, but exist simpliciter through and through.87 The proposed reading according to which Descartes sees “that important distinction . . . between sensations and perceptions’ ” is not in line with Descartes’s official intellectualist account of the mind. Furthermore, although this more perceptive Descartes recognizes an important distinction, Sellars rightly points out that he still is confronted with two difficult problems that he does not have the resources to resolve: First, the idea that sensations exist simpliciter through and through leads to the absurd view that the mind literally turns blue when seeing a blue object. As Sellars suggests the solution to this problem, which was not available to Descartes, is to think of sensations as exhibiting properties that are analogical to but not the same as the properties exemplified by physical objects.88 Second, Descartes famously distinguishes between the primary qualities size, shape and movement and secondary qualities such as color, sound and smell. While the former are objective categorical properties of physical objects that are clearly and distinctly conceivable, the latter are dispositions in physical objects to cause in minds noncognitive sensations of colors, sounds and smells. Furthermore, the phenomenality of experiences results from the cognitive conceivings of primary qualities and non-cognitive sensations. For instance, the phenomenality of the visual experience of a blue triangle results from a conceiving of a triangle and a sensation of blue. The problem is that if Descartes draws the distinction between conceiving and sensation in the way Sellars, then he is in trouble explaining why the phenomenality of the experience is seamless. How can it be that an objectively existing color and a simpliciter existing shape go together seamlessly to constitute the experience of a blue triangle? To conclude, even the more perceptive position Sellars at times attributes to Descartes lacks the resources to develop a proper account of the distinction between the sensory and intellective elements as well as their cooperation in sense perception.89

30  Christian Barth

5.  Evaluating Sellars’s Critique In his writings, Sellars presents a classic interpretation of Descartes as an indirect realist who accounts for intentionality in terms of representing immanent objects that exist in mental acts. While there is some truth to this interpretation, Sellars misconstrues important details. Sellars believes that the immanent object is represented by the mental act and itself represents an external object, if such exists. However, in fact Descartes does not understand the former relation in terms of representation, but in terms of presentation. The mental act presents its immanent object to the mind such that the mind is conscious (conscius) of it, and by means of the act-cum-object the mind (potentially) represents an external object. In a nutshell, contrary to what Sellars believes, Descartes does not understand the immanent objects as representational intermediaries, although he does understand them as epistemic intermediaries. Sellars’s account of Descartes’s conception of consciousness (conscientia) is inflationary in kind in that the mind’s consciousness concerns the immanent objects of current mental acts as well as the kind to which these acts belong. For Sellars, Cartesian consciousness is consciousness of one’s current mental acts as the acts they are and as being of the objects of which they are. We have argued that this inflationary reading is not well supported and defended a deflationary account according to which Cartesian consciousness merely concerns the acts’ immanent objects, but not the kind to which the mental. Cartesian consciousness is rooted in the fact that mental acts present their immanent objects to the mind. The acts do not in addition present themselves to the mind as what they are. Sellars sees Descartes as subscribing to a version of the Myth of the Given. More precisely, for Sellars, Descartes identifies one’s own current mental acts as what is given to the mind in the problematic sense of the Myth. This objection to Descartes is developed against the backdrop of Sellars’s inflationary account of Cartesian consciousness. If we take seriously the proposed deflationary reading, Sellars’s charge to Descartes still stands, but reduces in its scope. What is Given in the problematic sense then merely are the immanent objects of thoughts, not the mental acts as belonging to this or that kind. Sellars also attributes to Descartes an abstractionist view of the acquisition of mental ideas. However, this attribution is misguided since Descartes is a thoroughgoing innatist. Finally, we have seen that in many texts Sellars criticizes Descartes for conflating sensation and intellection. As a consequence of this conflation, Descartes cannot properly account for the phenomenological difference between both kinds of acts, e.g., between thinking of pain and feeling pain or between thinking of blue and seeing something blue. In other texts, however, Sellars believes that Descartes has some awareness of this important distinction. Yet, even if Descartes draws this distinction, he faces the challenge of explaining how the intellective thought of primary qualities such as

Sellars on Descartes  31 shape and the sensation of secondary qualities such as being blue go seamlessly together as they do in sense perceptions of blue rectangular surfaces. As Sellars rightly points out, Descartes does not have resources to properly address this problem. Indeed, for Sellars, no one will have these resources earlier than Kant.

Notes 1 See EPH, vii. 2 At times, Sellars is even prepared to defend Descartes—at least to some extent— against criticism. See, for instance, MCP, 238, where he takes issue with Lichtenberg’s objection to the cogito reasoning. For a critical discussion, see Sievert (1980). 3 See, in particular, EPM, BBK, PSIM, SM, BD, KTM, and KPT. 4 For a convincing complaint in this direction, see Hatfield (2001). 5 The second major difference between Aristotle and Descartes concerns their ontology of the mind. Descartes replaces Aristotelian hylomorphism with Platonistic dualism. This difference, however, only plays a marginal role in Sellars’s engagement with Descartes (see SM, 31–2). 6 To be more precise, the functions of the vegetative and the sensory souls are fully mechanized in non-rational beings and the vegetative functions of the soul also are mechanized in rational beings. The sensory functions of the soul of rational beings are partly mechanized and partly intellectualized. The intellectualized part will be of interest in this contribution. 7 See AT VII 356. 8 See AT VII 12 and 78. 9 See AT VII 78 and AT VIIIA 25. 10 Descartes assumes acts stemming from will that contribute to complex mental acts (we will call them ‘non-ideational thoughts’), but cannot occur independent of a representational intellective act (idea) at which they are directed. If considered in abstraction, these acts stemming from will might appear to one as independent of the intellect. However, since these acts are modes of the mind whose essential feature is the intellect and since modes ‘inherit’ the essence of what they modify, the acts stemming from will eventually are also essentially intellective acts. They essentially are modifications of intellective acts, namely ideas (see AT VIIIA 25). 11 Descartes’s wide notion of thought that includes acts of willing, sense perception and imagination has been a constant issue in the scholarship. The rationale for this wide notion is not, as has often been claimed, that Descartes redefines the essence of acts of thinking in terms of consciousness. Instead, the wide notion follows from Descartes’s intellective understanding of the mind and of these acts. 12 See AT VII 37. 13 In order to avoid confusion, it needs to be kept in mind that although nonideational thoughts are not ideas, they nevertheless essentially involve ideas. We will elaborate on this shortly. 14 Descartes illustrates the representational character of ideas by drawing an analogy to material images (see AT VII 37). 15 Descartes uses ‘act’ and ‘operation’ interchangeably in this context (see AT VII 246). As Sellars notices, the notion of act is to be understood in the sense of an actuality rather than an action (see NAO, 66). 16 See AT VII 8. 17 See AT VII 8.

32  Christian Barth 8 See Simmons 2016, 645. 1 19 See AT VII 39, 40, 42, 43, 44, 46, 82, and 83. 20 For Sellars’s account of this distinction, see SM, 33. 21 See SM, 33 and 36. 22 AT VII 161. 23 See SM, 31. 24 See AT VII 246 and KPT, 207. 25 Sellars mentions Descartes’s radical break with Aristotelian hylomorphism at SM, 31–2. 26 KPT, 190. 27 SM, 31. 28 SM, 32. 29 See BD, 261–2. 30 See SM, 31. 31 BD, 260 32 See section 2. 33 See SM, 32. 34 Sellars also mentions that Descartes describes the representational character of ideas by saying that ideas represent things in virtue of containing likenesses of these things (see KPT, 208–9). Here, Sellars refers to Descartes’s frequent use of the Latin term ‘similitude’ in a technical meaning inherited from scholastic philosophy (see, for instance, AT 37). 35 See BD, 261 and SM, 34–6. 36 See BD, 261 and SM, 33. 37 See BD, 261. 38 See KPT, 188–90. 39 SM, 33. Accordingly, Sellars calls the mental acts ‘representings’ (SM, 34) and the objective ideas ‘representables’ (SM, 33). 40 See SM 33. 41 See Kenny (1968) and Rorty (1979). 42 See Nadler (1989), Perler (1996) and Pessin (2009). 43 See AT III 474. 44 See KPT, 197–8. 45 See section 2. 46 Ideae nomine intelligo cujuslibet cogitationis formam illam, per cujus immediatem perceptionem ipsius ejusdem cogitationis conscius sum. 47 BD, 260. See also section 2. 48 See for this distinction section 2. 49 This reading presupposes that we do not understand the non-ideational thoughts from the 3rd Meditation as dependent acts of will, but as self-contained complex acts that involve ideas. 50 See Barth (2017) for a comprehensive defense of this deflationary reading of Cartesian consciousness. 51 See BD, 262. 52 BD 260. 53 See BD, 263. 54 BD, 261. 55 See BD, 262–3. 56 BD, 262. 57 BD, 263. 58 The view that the consciousness of ideas as well as of non-ideational thoughts is grounded in meta-perceptions is no option because it inevitably leads into a vicious regress. Sellars is well aware of this (see KPT 210).

Sellars on Descartes  33 9 See also BD, 265. 5 60 See AT VII 37. 61 In BD, Sellars presents another line of thought that is supposed to support his meta-perception account of the consciousness of non-ideational thoughts in Descartes (see BD 263–5). The passage from the Passions of the Soul Sellars discusses is important, but I cannot discuss it here in any detail. The main problem with Sellars’s textual evidence in this case is that the term ‘conscientia’ just does not occur in this passage. It is far from clear that Descartes is concerned with consciousness (‘conscientia’) in this passage at all. 62 Sellars ascribes to Descartes the view that there is representational awareness of extramental things and non-representational awareness of intramental things (see KPT 210–12). 63 See KPT 210–12. 64 See KPT, 212. 65 See EPM, 253. 66 See EPM, 253–4. 67 EPM, 286 and FMPP, 3. 68 FMPP, 12. 69 FMPP, 11. 70 FMPP, 12. 71 See EPM, 253. 72 KPT, 211. See also BBK §25 and §57. 73 Perhaps, Sellars thinks that Descartes’s innatism merely concerns mathematical and metaphysical concepts because in SM he attributes an innatist position concerning these kinds of concepts to Descartes. If this is correct, it suggests that Sellars only takes into account the narrow notion of innateness from the 3rd Meditation (see AT VII 37), but not the broad notion of innateness from the document A Certain Broadsheet (see AT VIIIB 357–8, 360–1, and 366). Both notions are not in conflict with each other since Descartes is concerned with different issues in these two passages. While the former concerns ideas as acts and asks in which circumstances these acts occur, the latter is related to dispositions to have ideas and claims that all ideas are innate to the mind in the sense that dispositions to have them are so. 74 For a thoughtful comparison of Descartes’s and Sellars’s conceptions of thought, see Alanen (1992). 75 For the VB model, see LTC 514–20, SK 318–27, and BLM 8–11. The VB model includes additional concepts of thought such as the concepts of proximate and of remote propensities to think-out-loud. Yet, the concept of thinking-out-loud is “[t]he central concept” of this network of concepts of thoughts within the VB model (see MEV, 327). It is “the primary concept” in the order of conceiving and “[a]ll other concepts of thinking are to be understood in terms of their connection with it” (BLM, 9). 76 EPM §§56–9. 77 Another myth? See EPM §63. 78 EPM §25. See also BD, 267. 79 BD, 267. 80 EPM §25. 81 SM, 32. 82 SM, 32. 83 See Haag (2007) for a more extensive treatment of this issue. 84 The idea that conceiving is like inner speech could have safeguarded Descartes from the conflation because understanding conceiving as inner speech makes obvious the deep difference between conceiving and sensing. But the inner speech

34  Christian Barth model of conceiving is missing in Descartes (SM, 35). The container model, in contrast, supports the conflation of conceiving and sensing (SM 35–6). 85 KPT, 201–3. 86 KPT, 203. 87 KPT, 202–3. 88 See BD, 270 and KPT, 201. 89 See BD, 273–4.

References Descartes’s Writings Descartes, René (AT). 1897–1910. Oeuvres. Edited by C.E. Adam and P. Tannery. Paris: Leopold Cerf. Descartes, René (HR). 1934. The Philosophical Works of Descartes. Translated by E.S. Haldane and G.T.R. Ross. 2 vols. New York: Dover. Descartes, René (CSM). 1984. The Philosophical Writings of Descartes. Translated by J. Cottingham, R. Stoothoff, and D. Murdoch. 2 vols. Cambridge: Cambridge University Press.

Sellars’s Writings Sellars, Wilfrid (EPM). 1956. ‘Empiricism and the Philosophy of Mind’. In Minnesota Studies in the Philosophy of Science. Edited by H. Feigl and M. Scriven, Vol. I, 253–329. Minneapolis, MN: University of Minnesota Press. Sellars, Wilfrid (BBK). 1960. ‘Being and Being Known’. Proceedings of the American Catholic Philosophical Association, 34: 28–49. Sellars, Wilfrid (PSIM). 1962. ‘Philosophy and the Scientific Image of Man’. In Frontiers of Science and Philosophy. Edited by Robert Colodny, 35–78. Pittsburgh: University of Pittsburgh Press. Sellars, Wilfrid (SM). 1968. Science and Metaphysics: Variations on Kantian Themes (The John Locke Lectures for 1965–66). London: Routledge & Kegan Paul. Sellars, Wilfrid (MCP). 1969. ‘Metaphysics and the Concept of a Person’. EPH, 214–41. Sellars, Wilfrid (EPH). 1974. Essays in Philosophy and its History. Dordrecht: D. Reidel Publishing Company. Sellars, Wilfrid (SK). 1975. ‘The Structure of Knowledge: (1) Perception; (2) Minds; (3) Epistemic Principles’ (The Matchette Foundation Lectures for 1971 at the University of Texas). In Action, Knowledge and Reality: Studies in Honor of Wilfrid Sellars. Edited by Hector-Neri Castañeda, 295–347. Indianapolis, IN: Bobbs-Merrill. Sellars, Wilfrid (BD). 1977. ‘Berkeley and Descartes: Reflections on the “New Way of Ideas” ’. In Studies in Perception: Interpretation in the History of Philosophy and Science. Edited by P.K. Machamer and R.G. Turnbull, 259–311. Columbus, OH: Ohio State University Press. Sellars, Wilfrid (BLM). 1980. ‘Behaviorism, Language, and Meaning’. Pacific Philosophical Quarterly, 17: 81–101. Sellars, Wilfrid (NAO). 1980. Naturalism and Ontology (The John Dewey Lectures for 1973–74). Atascadero, CA: Ridgeview Publishing Company.

Sellars on Descartes  35 Sellars, Wilfrid (FMPP). 1981. ‘Foundations for a Metaphysics of Pure Process’ (The Carus Lectures). The Monist, 64: 3–90. Sellars, Wilfrid (MEV). 1981. ‘Mental Events’. Philosophical Studies, 39: 325–45. Sellars, Wilfrid (KPT). 2002. Kant and Pre—Kantian Themes: Lectures by Wilfrid Sellars. Edited by P.V. Amaral. Atascadero, CA: Ridgeview Publishing Company. Sellars, Wilfrid (KTM). 2002. Kant’s Transcendental Metaphysics: Sellars’ Cassirer Lectures Notes and Other Essays. Edited by Jeffrey F. Sicha. Atascadero, CA: Ridgeview Publishing Company.

(C) Other Writings Alanen, Lilli. 1992. ‘Thought Talk: Descartes and Sellars on Intentionality’. American Philosophicaly Quarterly, 29/1: 19–34. Barth, Christian. 2017. Intentionalität und Bewusstsein in der frühen Neuzeit: Die Philosophie des Geistes von René Descartes und Gottfried Wilhelm Leibniz. Frankfurt am Main: Klostermann. Haag, Johannes. 2007. ‘Sinnliche Ideen: Descartes über sinnliche und begriffliche Aspekte der Wahrnehmung’. In Sehen und Begreifen: Wahrnehmungstheorien in der frühen Neuzeit. Edited by Dominik Perler and Markus Wild, 95–121. Berlin: de Gruyter. Hatfield, Gary. 2001. ‘Transparency of the Mind: The Contributions of Descartes, Leibniz, and Berkeley to the Genesis of the Modern Subject’. In In Departure for Modern Europe: A Handbook of Early Modern Philosophy (1400–1700). Edited by Hubertus Busche, 361–75. Hamburg: Meiner Verlag. Kenny, Anthony. 1968. Descartes: A Study of his Philosophy. New York: Random House. Nadler, Steven. 1989. Arnauld and the Cartesian Philosophy of Ideas. Manchester: Manchester University Press. Perler, Dominik. 1996. Repräsentation bei Descartes. Frankfurt am Main: Klostermann. Pessin, Andrew. 2009. ‘Mental Transparency, Direct Sensation, and the Unity of the Cartesian Mind’. In Topics in Early Modern Philosophy of Mind. Edited by J. Miller, 1–37. Dordrecht: Springer. Rorty, Richard. 1979. Philosophy and the Mirror of Nature. Oxford: Oxford University Press. Sievert, Donald. 1980. ‘Sellars and Descartes on the Fundamental Form of the Mental’. Philosophical Studies, 37/3: 251–7. Simmons, Alison. 2016. ‘Representation’. In The Cambridge Descartes Lexicon. Edited by Lawrence Nolan, 645–55. New York: Cambridge University Press.

2 The lingua franca of Nominalism Sellars on Leibniz Antonio M. Nunziante

Introduction In Sellars’s interpretation of Leibniz we can identify two quite distinct, albeit internally linked, fields of inquiry: the first focuses on some technical questions of logic and philosophy of language; the second on some broader questions of epistemology, ontology and history of philosophy. In the first case, reference to Leibniz is helpful for disambiguating some internal topics of analytical metaphysics. By investigating notions such as “particulars,” “proper names” and “abstract entities,” Sellars, in fact, takes a stand against a contemporary debate initially triggered by Bertrand Russell, which counted Nelson Goodman, Peter Strawson and Gustav Bergmann as its closest referents (Rauzy 2009, 87). Such line of inquiry was developed in works like On the Logic of Complex Particulars (LCP 1949), Particulars (P 1952), Meditations Leibnitziennes (ML 1959) and Abstract Entities (AE 1963). In the second case, the framework of reference consists of some peculiar themes in Leibniz’s philosophy, such as his theory of relations, his representational theory of mind and the definition of truth within his overall nominalist strategy. Sellars focuses on some specific epistemological and ontological issues that influenced much of the modern epistemological tradition, recalling authors such as Descartes, Berkeley, Hume and, above all, Kant. He refers directly to these leibnitian topics in ML, but other, more indirect references can be found in works like Some Remarks on Kant’s Theory of Experience (KTE 2002), Science and Metaphysics (SM 1968), Berkeley and Descartes (BD 1977) and in the seminars on pre-Kantian and Kantian topics (KPT). The first line of inquiry has been analyzed by Jean-Baptiste Rauzy, who has convincingly shown how Sellars connected the controversy on “particulars” to Leibniz, as well as to Bergmann’s re-reading of Russell (Rauzy 2009, 94).1 This very solid work has the merit of demonstrating how the broad topic of universals was linked in Sellars’s early writings to the more technical, and seemingly remote, issue of the nature of particulars. The second line of inquiry still remains to be thoroughly investigated, since Sellars’s interpretation of Early Modern epistemology was conducted by scholars

The lingua franca of Nominalism  37 almost exclusively in relation to Descartes (Alanen 1992), Hume (Landy 2008) and Kant in particular (Macbeth 2000; McDowell 2006; O’Shea 2011, 2016, 2017; Haag 2017; Brandom 2015, 2017). The present paper aims to fill this gap to some degree. I will therefore not refer to the issues already analyzed by Rauzy (unless it is strictly necessary), and will focus instead on the previously mentioned epistemological and ontological topics. Significantly, Sellars claimed in his Introductory Remarks to the Class (KPT 2) that “One of the most interesting topics a person wants to work on is the relation of certain features of Kant’s thought to the corresponding features of Leibniz’s thought.” Paraphrasing his words, we could perhaps say that one of the most interesting topics a person could work on is the relation of certain features of Sellars’s thought to the corresponding features of Kant’s and Leibniz’s thought.

1 Taking some first steps along the path of Sellars’s interpretation of Leibniz, we can start with how he interprets the doctrine of the complete concept.2 Sellars begins by referring to the classic quotation from section 8of the Discourse on Metaphysics: It is indeed true that when several predicates are attributed to a single subject and this subject is attributed to no other, it is called an individual substance; but this is not sufficient, and such explanation is merely nominal. We must therefore consider what it is to be attributed truly to a certain subject. Now it is evident that all true predication has some basis in the nature of things and that, when a proposition is not an identity, that is, when the predicate is not explicitly contained in the subject, it must be contained in it virtually. That is what the philosophers call in-esse, when they say that the predicate is in the subject. (Philosophical Essays, 40–1) In ML Sellars begins to disambiguate the notion of individual concept by using two kinds of conceptual tools, which will gradually become more and more entwined. The first concerns the question of proper names: the individual concept nominates the individual—not in the sense that we, finite subjects, nominate a possible individual, but rather in the sense in which God gives a proper name to each individual in the very act of creation (ML, 153–4). The second conceptual tool focuses instead on the distinction between concept and nature, specifying that the individual concept has a double level of existence: it exists in the divine’s understanding as a logical concept, and in re as the nature of a substance (ML, 154). Individuals are conceived in modal terms (as mere logical possibilities of the divine’s mind) on the one hand, while on the other they are factually considered as natural substances existing in the world.

38  Antonio M. Nunziante Sellars says that the “venerable” notion of individual nature thus took a new twist in Leibniz’s hands, since he was the first to see clearly that the individuality of a substance can only be understood in terms of episodes in its history, and to conclude that if the nature of a substance is to account for its individuality, it must account for episodes, and not merely the capacities, powers, dispositions . . . which were traditionally connected with the natures of things. (ML, 154) Accordingly, the nature of a substance indicates more than just a dispositional property (as traditionally claimed by the post-Aristotelian tradition (APM, 545; Rauzy 2009, 93), but Leibniz reworks it better in terms of conditional episodes. He says that the nature of an individual substance, if we but knew it, would explain why an individual “behaves as it does in the circumstances in which it is placed” (ML, 154). The nature of a substance somehow captures all the episodic premises that lie behind the behavior of a given individual. In this context, Sellars applies to Leibniz a logical model drawn from C. D. Broad, which enables episodes to be interpreted as conditional premises of a series, such that “if at any time S were to be involved in an episode of kind E1, it would be involved in an episode of kind E2” (ML, 154). Sellars thus describes Leibniz’s concept of nature by logically appealing to a whole series of conditionals that, chained together, form the premise “of a syllogism in re” (ML, 155; Rauzy 2009, 93). The result produces a sort of modal stress, however, because—as Sellars points out—Leibniz conflates two different models here: he uses the so-called “thing-nature” framework of the Aristotelian tradition on the one hand, while on the other he refers to the “event-law” framework that dominated the lexicon of science from the seventeenth century onwards (APM, 546, 565–6). The first model, more closely reflecting common sense, assumes that some natural kinds of things exist, grow and develop according to an internal framework. It consequently admits the (ontological) presence of substantial forms representing the metaphysical core of each substance. By contrast, instead of referring immediately to ‘things,’ the second model focuses specifically on the lawful dimension of nature: scientists do not know what kinds of things are there until they arrive at laws “which can be translated into the thing language” (APM, 566). In this case, it is the nomological element that determines what properties are needed in order to be ‘a thing’ (res), so we are no longer speaking of a substance, but rather of ‘events.’ The latter indicates a sort of relational structures that does not spring from a central core, but consists in a process of reciprocal functional correlations. Returning to Leibniz, the nature of a substance would therefore occupy an amphibious position, since it would contain both episodic facts (in accordance with the thing-nature framework of the Aristotelian tradition) and

The lingua franca of Nominalism  39 hypothetical facts conforming to the modern event-law lexicon. Indeed, given the operational presence of the Principle of Sufficient Reason, episodes prove to be reducible into hypotheticals, so the first model can be collapsed into the second (ML, 156).3 Individual substances are therefore the outcome of a strange twist taken by Leibniz, and Sellars argues that the source of this conceptual mélange should be sought in his remote doctrine of relations. The Aristotelian realistic conception of relations met with some well-known difficulties in explaining the connection between particulars and universals, as Sellars recalls. Leibniz, for his part, worked within this late-Aristotelian context. Renewing the lexicon of inherence (accidents that inhere to matter), he established a new epistemic status for relations (KPT, 257), in what Sellars judges a “brilliant metaphysical move.” This, in turn, produced a kind of snowball effect as it soon changed the traditional approach to the very concept of ‘truth’ and the modern concept of ‘representation.’ More generally, it reshaped some of the most solid epistemic principles of Early Modern times (like the general distinction between ‘subject’ and ‘object’), giving new power to a pure nominalist position.

2 At the beginning of the previously quoted passage, Leibniz said that “when several predicates are attributed to a single subject and this subject is attributed to no other, it is called an individual substance.” (Philosophical Essays, 40–1). He promptly added, however, that this is not enough, because such an explanation is ‘merely nominal.’ So we need to investigate what it is to be ‘attributed truly’ to a certain subject. What seems to interest Sellars is the term ‘attribute,’ and in particular the fact that attributes cannot be conceived as abstract entities (KPT, 249). Sellars says we are not dealing with linguistic facts, since the context has a rather theological background, referring to God’s knowledge of individuals. The idea is, more or less, that statements like ‘Socrates is wise’ are only true if the attribute of being wise is a constitutive element of the corresponding judgment pronounced by God (KPT, 250). Hypothetically, wisdom might even not figure among our criteria for identifying Socrates (as it was for many of his Athenian peers), because what really matters is the ‘basis in the nature of things.’ It is only in divine judgments that predicates are necessarily involved in the subject and therefore metaphysically true. The same topic can be considered from a different point of view, focusing on the problem of relations: if the nature of Socrates involves everything that is true of him, then it will also be true of him that, at some time in his life, he stood on the Agora. In this case, we have to consider how a spatial relation (being on the Agora) can be true of Socrates, and how it can be included in his nature (KPT, 252). We can tell in advance where the whole argument is leading: relations (of any kind) have a mere phenomenic nature, insofar as they can be considered as predicates that inhere to a subject, and

40  Antonio M. Nunziante they are eventually transformed by Leibniz into representative states of individual substances (monads). But what we need to understand is precisely what this means. One important element to consider is the historical distinction drawn by Descartes between ‘formal’ and ‘objective’ reality. In an act-content model of representation, Sellars argues, we have to distinguish between two ways of being. There is a “second class of existence,” in which the content depends entirely on the act of representing (“objective” existence); and there is a “first class of existence,” in which being is not dependent on a mental act (“formal” or “actual” existence)(KPT, 9). There are consequently also two kinds of truth. There are things that are metaphysically true, in the sense that they actually have a kind of existence; but there are also secondclass truths concerning represented objects, which are considered true only by virtue of the fact that they correspond to some entities in the first class. Sellars attributes the utmost importance to this distinction in many of his writings. In his view, it represents “the whole key of the epistemology of this period,” for here we have the primary source of “a correspondence theory of experience” (KPT, 8–9).4 Leibniz’s theory of relations does indeed seem to develop along the lines of a correspondence theory of truth, though this is actually undermined from within because he completely rewrites notions like ‘correspondence’ or ‘counterpart’ (as does Kant later on). While for Descartes and Newton, actual things are those existing in the absolute dimension of space and time, for Leibniz every sort of relational structure can have no such existence, but must always inhere to a substance. Space and time have no actual existence, only an objective one. The fact that the individual S1 represents a triangle does not mean that a real triangle ‘absolutely’ exists. Descartes might have endorsed it, but for Leibniz (and later for Kant) there can be no actual space because the whole point is to determine precisely what it means to be “actual” (KPT, 253). According to Leibniz, the idea of an extended universe being formally there makes no sense. As a matter of fact, a radical schism gave rise to the Early Modern age: for the Cartesians and Newtonians, as well as for most of the empiricist tradition, the physical universe has an eminently mind-independent space-time existence. Leibniz had another kind of intuition, which was quite the opposite: the notion of a stand-alone universe would make no sense without the actual presence of individuals endowed with perception, so the objective existence of the former can be considered as part of the formal representative character of the latter. The physical universe is not self-contained, but compatible and integrated with a system of actual representing subjects.5 This brings us, of course, to the very core of Leibniz’s nominalism. In a metaphysical sense, only substances (and their properties) exist: “Indeed, considering the matter carefully, we must say that there is nothing in things, but simple substances, and in them, perception and appetite” (Philosophical Essays, 181).6 The consequences are clearly enormous because the very

The lingua franca of Nominalism  41 idea of ‘objective’ existence (like the space-time primary dimension) is called into question. Only representing substances are actual, while the physical universe (with its space-time relations) becomes an objectively represented world. Leibniz stretches the old Cartesian distinction between ‘formal’ and ‘objective’ existence to its limits, and Sellars is keen to remind us that, if we do not take this idea seriously, then “Leibniz doesn’t exist at all” (KPT, 253). He adds that Kant will suffer the same fate because he packs “a new meaning into the word actual” in his Critique, and he is indeed “very close to Leibniz” (KPT, 253–4). Briefly put, the very idea of ‘actuality’ was reshaped (albeit in different ways) by Leibniz and Kant, and that is why they were both driven to explore a new meaning for the expression ‘being true.’ The sense of the old dictum adaequatio mentis et rei now lacked the material counterpart that consists, according to the Cartesian and Newtonian traditions, in the absolute spatiality of bodies. To clarify this main idea, which marked a pivotal turning point in the epistemology of the Modern age, Sellars proposes to approach the issue from a different angle. If we imagine a system of three monadic individuals, S1, S2 and S3, we can try to understand how the phenomenal domain of space is engendered by their representations. In outer space, Sellars says, “an object is between object and object,” meaning that spatial relations essentially involve ongoing continuity (‘something being beyond something’). The object in question occupies a certain region in space, which is further determined by its relation to another region in space, and so on (KPT, 254–5). In our tiny monadic system, we thus have three monads that are not spatially related, but they are capable of representative states. As Sellars suggests, let us suppose that the monad S2 represents S1, with the result that S1 will objectively exist in S2 (being represented therein). If we also suppose that S1 represents S3 (considering such representing acts in the broad terms of what Leibniz would call petites perceptions), then S2 not only represents S1 but also (albeit unwittingly) represents the state of S1 that represents S3. To comment on this situation, we can say that S2 has a firstorder representation of S1 and a second-order representation of S3; or that S2 represents S1 directly and S3 indirectly (KPT, 255). Needless to say, the game can go on and on, in a process based on the idea that every monad always represents every other monad, as well as every state of every other monad. In other words, Leibniz thinks that our petites perceptions are infinitely complex, not just in the sense that there are infinitely many of them, but that they are complex in an interesting dimension that is usually overlooked: they are infinitely complex in this ‘nestedness’ as I put it. (KPT, 255)7 What emerges from this continuous process of internal representative mediation is that the phenomenic experience of ‘something that is beyond something’ is grounded in the indirect nature of a substance’s representation.

42  Antonio M. Nunziante “That is all I’m driving at,” Sellars concludes, emphasizing the structural similarity between such a “nesting of representations” and the corresponding spatial “beyondness.”8 The metaphysical key to the story is thus as follows: Leibniz begins by describing what we ordinarily regard as a real relation between objects, then turns it into an ideal factor that must be explained in terms of the nature of the perceiving substance. Notions like ‘truth’ and ‘existence’ hover there too, insofar as they are implicit in his commitment to a strong form of nominalism that we will try to develop later on. Now the main question becomes: why did Leibniz question the idea that objects have spatial relationships instead of leaving things as they stood? Sellars suggests that the origin of his very counterintuitive solution was rooted in classical puzzles concerning the (mainly Aristotelian) concept of substance, particularly those treating the relationships between substances and accidents. To say ‘this leaf is green,’ for instance, was classically regarded as an example of the inherent relation between substance and accident, which posed the tough problem of how to interpret the relationship between the particular green of a single leaf and ‘greenness’ in general. The problem can be better defined, says Sellars, if we think of two leaves sharing the same shade of green (like a ‘Forest Green, Pittsburgh Paint #59’): do each of the two substances possess their own green, or should we assume that there is only one ‘Forest-Green-59’ instantiated by two different substances? Here we come up against the classical contrast between the idea of an individual qualitative identity and the notion of a numerically identical universal (KPT, 258). Leibniz’s insight, at least according to Sellars’s reconstruction of it, was more or less as follows: relational predicates can be expressed in the general form of: S1 is R to S2 Hence the puzzle of exactly where R should be placed, because it can go on the side of S1 (considering R as part of it), or on the side of S2, or we can even imagine that R inheres to both terms or to neither of them. Leibniz instead took the view that such a proposition should be treated as a special case of: S is P and thus S1 is R-to-S (P) Leibniz supported the idea that an R-to-S2 predicate is inherent in S1. So he accepted the consequence that S2 must be in S1 (inesse) and, as Sellars points out, he reinterpreted such commitments in the light of the Cartesian

The lingua franca of Nominalism  43 distinction between ‘objective’ and ‘actual’ being (ML, 159). He therefore interpreted facts in the form of S1 is R to S2 as facts in the form of S1 represents S2. Once again, relations are transformed into representative states of a formally existing substance. The relational fact is but a well-founded phenomenon, the objective existence of which has nothing to do with the idea of the naive realism mainly accepted by the empiricists. From a realistic stance, there is a kind of extra-epistemic fact (‘being green’) that exists in itself, irrespective of the current representative status of S1.9 Leibniz’s extreme nominalist position was intended precisely as a way to escape from both naive realism and abstract Platonism. Once and for all, “there is nothing in things except simple substances, and in them perception and appetite.” (Philosophical Essays, 181)The key is understanding the structure of the inner nature of substances.

3 Needless to say, Sellars was strongly in favor of a theoretical shift that enabled relations to be transformed into categorical facts (remember The logic of ‘looks’ or The logic of ‘means’ in EPM)10 In fact, one of the most interesting possible applications of this approach regards causality. In Leibniz’s world, facts in the form of S2 is acted on by S1 (i.e., by being in a state f, S1 causes S2 to become p) (ML, 160) easily become facts taking the form S2 represents the fact that S1 is in the state f and thus S2 represents S1 The notion of causality is therefore revised on the basis of the Cartesian distinction between objective and formal being. Like relations, causes are merely represented, not representing. Leibniz’s universe is composed of individual substances, and the worldly things they run into during their life span must be seen as part of their personal story, as the representative

44  Antonio M. Nunziante stuff making up the fabric of their nature. This view, however, brings us back to a problem considered earlier. Assuming that Leibniz has effectively turned causes and relationships into ideal representations, how do we envision the bond that collectively holds together the representative states of a substance? How should we interpret the nature of this bond? Why, Sellars asks, should we not continue alongside Broad in interpreting episodic facts in terms of hypothetical facts, taking for granted “that both episodes and hypotheticals are grounded in Necessary Being?” (ML, 162). The starting point of the problem, we remember, was: “if at any time S were to be involved in an episode of kind E1, it would be involved in an episode of kind E2.” According to Sellars, Leibniz would not have been wholly comfortable with such a conclusion because of the metaphysical premises of his system. Nowadays, we tend to think along the lines of a distinction between causal properties (as general hypotheticals) and occurrent states (as categoricals), but Leibniz—like most of his predecessors—interpreted causal properties in terms of ‘desires, plans, personal commitments.’ Thus, whereas we might be inclined to interpret the statement ‘Jones has a strong desire to go to New York’ in terms of conditional facts about Jones, Leibnitz thinks of a strong desire as a continuing series of episodes which tends to develop into going to New York and will continue to develop if not impeded. (ML, 162; emphasis original) Leibniz sees S2 becoming p as a matter of S2 having a plan to become p. In other words, he interprets the ‘becoming’ as a sort of actually ‘doing’ something, so though the ‘plan’ to become p may be hypothetical, it is nonetheless interpreted as a categorical fact regarding S2. We can hardly avoid referring here to the pivotal part played by the distinction that Sellars himself drew between the ‘ought to do’ and the ‘ought to be’ rules in the construction of his own nominalist position, but there is nonetheless a sense, according to Leibniz’s own position, that “all the fundamental facts about a substance are episodic facts” (ML, 162).11 The consequences are huge, since ‘nature’ no longer indicates a law-like hypothetical function (as would seem obvious to us), but rather a “life-plan” and “as such it has esse intentionale as the content of an abiding aspiration” (ML, 162). Broad’s hypothetical law-like notion is ultimately replaced by a concept of nature as “something that is always there” (KPT, 252) and wholly involved in every single episode in the life of an individual substance. Sellars uses the sentence ‘Socrates is wise’ as an example of a statement that helps us to see that ‘wisdom’ is not a momentary state of Socrates, but an enduring trait that is somehow “a characteristic of this life-plan” (KPT, 252). To sum up this line of reasoning we can say that the individuality of a substance is identified by the complete concept; the latter is explained as the sum of the representative states of the former; every single representative episode is logically described by the nature of the substance (i.e., by the

The lingua franca of Nominalism  45 presence of a life-plan that acts as a self-governed source of justification); the epistemic status of relations is founded on metaphysical, and therefore not relational, grounds (the divine’s decision to create a world of multiple, mutually compatible individuals); finally, causes and relations are reduced to representative states of the subject and, as such, they are treated like predicates that inhere to a substance (praedicatum inest subjecto) and, once again, to its nature. The circle is almost complete. What is still missing is a key term that we now need to analyze because, in some ways, it provides the ultimate sense of Sellars’s interpretation of Leibniz. Of course, the term in question is ‘truth.’

4 Leibniz’s thesis of the inesse is supported by several considerations regarding the nature of truth. The issue can first be summarized as follows: the representative, temporally tensed, episodes of a substance conform to a corresponding set of timeless facts in the mind of God. Adopting a correspondence theory of truth, a given representative episode R1 is true if and only if it corresponds to a timeless fact F1 that represents the real counterpart of R1 in the mind of God. A statement like this calls into question both the notion of truth and the concept of time. As Sellars points out, we can say of an episode it took place, is taking place, or will take place, but a fact is a fact: it is a fact that 2 + 2 = 4, and it makes no sense to say that 2 + 2 was 4 or that it will be 4 (ML, 163). In some contexts at least, ‘being a fact’ is a timeless mode of being. In Sellars’s view, Leibniz assumes that there is a timeless set of entities (i.e., facts) which are about what happens to a substance at different times, and such that it is by virtue of corresponding to these entities that our statements and judgments about the substance are true. (ML, 163) The plot for this ‘ontology of truth’ is quite difficult to unfold, however, since Leibniz, in Sellars’s view, conflated the notion of timeless fact and the notion of life-plan. He, in fact, assumed the very first notion as the actual guidance of his thought and as a consequence a sort of internal pressure was generated in his whole system (ML, 164). The core question thus becomes: what does it mean to be a ‘fact’? Leaving aside all the surrounding statements, it is Sellars himself who champions an ‘over-simplified thesis,’ according to which claims of the form: It is a fact that-p are simply another way of saying ‘P’ is a true statement in our language (ML, 167).

46  Antonio M. Nunziante The purported idea here is that the element of truthfulness lies not in a kind of supporting extralinguistic factor (as in the most classic correspondence theory of truth) but in the intrinsic coherence that ties some statements in our language together. In Sellars’s opinion, this whole issue is pivotal and also calls Kant into question, because in both cases it seems possible to detect a peculiar form of coherentism proposed as the ground floor of truth. So, before proceeding with Leibniz, let us take a brief look at Kant’s own insights.

5 According to Sellars, correspondence theories of truth, in the Early Modern age at least, have been mainly grounded on a representational model of knowledge (KPT, 8). Employing a lexicon that refers to the classical quasiCartesian epistemological framework, Sellars claims that, in considering such theories, we have to be very careful to distinguish between: (i) the ‘act’ ” of representing performed by the subject; (ii) the objective ‘content’ of the intended representation; and (iii) the real object in the outer spacetime world.12 We thus have the individual’s actual thoughts (‘representings’), some represented content (‘represented’), and transcendent real objects (‘unrepresented representings’) that together constitute the ur-type of a relational model of knowledge, which Sellars calls “the Cheshire cat form” of relation (KPT, 12).13 On several occasions Sellars emphasizes that here we are only dealing with an illusory form of relation because, on the basis of the content represented, we would be led to imagine a one-to-one correspondence between the immanent represented content and the transcendent object of the real world. Yet, this is the weak point of the model, since the only real relationship entertained is the one occurring between the act of representing and the immanent represented content, not the one between represented contents and transcendent objects. The claim that we can immediately relate our knowledge to things in themselves raises the prospect of a “transcendental realism” that Kant himself helps us to unveil (KPT, 25; SM 47).14 Transcendental realists are those who mistake the epistemic conditions of our referring to objects (like space and time) for properties of things in themselves. What has only epistemic value is assumed instead to be a transcendent object. In other words, the ‘an sich’ world would be the ultimate ground of givenness and would provide the possibility for a truthful theory of correspondence. Sellars’s insight is that at the bottom of such relational structure lies the attempt to offer a plausible explanation for the case of intersubjectively shared representations. If we consider two people sharing the same representative content, we are almost compelled to imagine that they both refer to a common item of experience, given independently from any form of subjective representation. Furthermore, according to the Cartesian tradition, the existence of such independent content would reflect God’s

The lingua franca of Nominalism  47 representation of it in his divine mind (KPT, 20). So we have the represented item, the real-worldly objects, and the archetypal items in the mind of God. This complex theory of transcendence later on evolves into what Sellars calls a “theological conceptualism,” the most famous interpreter of which was Kant himself (KPT, 36). Indeed leaving aside the theological premises of the Cartesian discourse, the idea that the very possibility of representing an object has to do with the existence of a certain domain of ‘representables’ can still be found in Kant: “in a certain sense, there must be a domain of representables, qua representables, and among them is, for example, a triangle” (KPT, 19).15 Kant needs this sort of assumption to break the private dimension of the subject, since representables are conceived of as a public domain, as a public source of possible knowledge. We need to bear in mind that, in Kantian terms, the argument does not claim that there is a class of representables and, consequently, that the foundation of empirical knowledge is also provided; instead, it shows that the concept [of empirical knowledge, A.N.] is a coherent one and that it is such and 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, 271)16 At the heart of the argument there is a purely epistemological issue and, in Sellars’s opinion, the nearest ancestor of this view was Leibniz himself, with his theory of God’s continuous representation of all possible worlds (KPT, 22). The main point is that, since Kant says that one cannot refer to ‘things’ outside the realm of representations (the ‘in itself’ of a thing is not representable, not even by analogy), the result is that the notion of truth is construed from inside the notion of representability. The next step, in fact, is that such a domain of representables becomes “the domain of items which are candidates for being transcendent objects” (KPT, 23–4).17 This is a truly fundamental remark because it implies that the transcendent domain of things is construed on the basis of an epistemic shift: a particular class of well-cohering representeds almost unwittingly becomes a class of transcendent items that, from that point on, will be regarded as part of a mind-independent world. As Sellars sees it, the point is that what we usually call an ‘actual state of affairs’ is effectively ‘a conceptual response’ endowed with an internal ‘judgmental form.’ So much so that: “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, 275)18

48  Antonio M. Nunziante The whole business implicit in the lexicon of ‘correspondence’ is therefore highly metaphorical: this is the final destination of Sellars’s conceptualhistorical analysis. Kant was an idealist with respect to the world of appearances, in the sense that he simply denied that anything in space and time could have any form of being other than a second-class one (KPT, 25). Yet, truth statements involve something more to be defined: they demand the presence of a small group of available representable contents that ‘cohere in a certain way.’ In a curious sense, they are the actual, or true, ones, but, for Kant, they are not true in any “correspondence” sense. They are just privileged ones which must be there if the notion of truth is to make sense. So, Kant is not trying to prove that there is truth, that there is knowledge or that there are objects: he is explicating the very concept of an objectively true experience. (KTE, 275) In Kant we find a notion of truth that partly corresponds to the realistic insights of the Early Modern age. The very notion of ‘privileged content’ somehow replaces the old metaphysical facts of the space-time world of Descartes or Newton and, as a result, a certain form of realism (scientific realism, one might cautiously say) has indeed been effectively achieved, by an internal development of the notion of representability itself. Of course, Kant was not a nominalist—and this is another, remarkable point (and probably the most crucial) where his and Leibniz’s paths cross. In Kant’s theory of truth, Sellars observes, there is “a great deal of weight on coherence,” (KPT, 140) since it is the very concept of coherence that shoulders the whole burden of truthfulness. More concretely, in the Analogy of the Experience Kant considers a special subclass of coherent representables as though they were “contents pertaining to the physical world obeying the laws of physics” (KPT, 141). This is the decisive epistemic step: such representables have a ‘physical lawfulness’ in the technical sense that they owe this kind of coherence to the properties studied by mechanics in the physics of the time. But this also shows how Kant was constructing a sort of scientific realism, though its meaning had been dramatically reshaped by comparison with the naive realism of Early Modern times. In other words, Kant gave a new shape to the traditional issue of the primary qualities of bodies, partly by excluding the world of secondary qualities (e.g., colors) from the range of knowable things (KPT, 45). This move marked a great departure from Leibniz, who saw little difference between primary and secondary qualities, given that the latter too played an essential role in determining the nature of an individual substance. In Leibniz’s nominalist world, as we have seen, there are only representing monadic individuals, to which relations are reduced. Therefore, unlike Kant, space, time and other sorts of relations possess an ontological counterpart

The lingua franca of Nominalism  49 for which we are able to account (KPT, 51–3). Leibniz’s final theory, in fact, is that our representative knowledge is but the counterpart of the actual existence of a large set of compossible substances, whose representation nevertheless is always confused, because of the infinite complexity of their nestedness. For Kant, on the other hand, representations can hardly be confused, because the very concept of ‘actual’ was deprived of any metaphysical sense, and the ‘in-itself’ of things was by no means intended to have an individuating nature. Kant’s stance is rather transcendental, since it considers the problem of the objectivity of representations in the epistemic terms of their construability (KPT, 75). Yet, even when Kant bids farewell to Leibniz, the theoretical structure of his thought brings him back to the latter. According to Sellars, the system of representables draws a sharp distinction between what one actually represents and what one would represent ‘if’ she found herself in a potentially different situation. In other words, it sets up a distinction between ‘actual’ and ‘possible’ experiences. Indeed, the whole matter of coherence generates a modal issue, since it does not simply refer to some aseptic content but always presupposes the possession of some scientific perspective. If we consider the history of physics, for example, the objective existence of any coherent content is unavoidably consistent with a certain blend of variable perspectives that imply paradigm shifts, new discoveries, scientific progress and so on. The privileged representables are hence by Sellars called ‘iffy’ representables, in the sense that they are available from a certain perspective, like the current scientific image of the world, but are not eternal like a Platonic idea. Acknowledging this therefore involves the possibility of other different perspectives, so much so that the whole system of coherence must be “temporalized” (KPT, 141). In short, the law-like problem of Kant’s physical world comes face to face with Leibniz’s modal question concerning the nomological legitimacy of other possible worlds (KPT, 141). According to Sellars, this topic is only hinted at in Kant’s works, and that is why “the small clues that Kant throws out when he is discussing possible experiences” are so baffling (KPT, 141). This final comment thus gives us a chance to return to Leibniz for the last time.

6 In Leibniz we investigated the idea of an individual substance conceived as a set of representative episodes; we explored the idea that this would correspond to a timeless class of events; we then faced the question ‘what does it mean to be a fact?’, reporting Sellars’s claim that a fact is simply a true statement in our language. The reference to Kant allowed us to clarify the profound nature of facts: they are subclasses of coherent statements in our language, namely a ‘conceptual response.’ This is true for Leibniz too, even if in his case the situation is possibly more complicated, given the difference between our language and God’s. In Leibniz we also find a seemingly verificationist approach, which is however internally undermined, since the

50  Antonio M. Nunziante world of timeless facts is but a world of true statements. A first consequence of this is that the concept of nature can be formulated in such a way that it requires no use of facts. Sellars notes that when we assert something as: The statement ‘S1 will be f3 in 1959’ is true because it is a fact that S1 will be f3 in 1959 (ML, 167). the proper ‘because statement’ can be formulated differently as: The statement ‘S1 will be f3 in 1959’ is true because S1 will be f3 in 1959. This is the ‘promissory note character’ of Leibniz’s notion of nature, which “requires no ontology of facts” because it carries in itself “a pervasive feature of the statements we are in a position to make about the world” (ML, 167). Truth has to do with an internal development of the metaphysical nature of substances and—insofar as they are all distinguishable, and therefore nameable—the next step is to disambiguate the role played by proper names in the structure of God’s language. We have already discussed the distinction between our language and the divine’s language, but now we need to put it under pressure, since it is through the latter that proper names were originally chosen.

7 Opening this new line of inquiry, Sellars makes the point that although proper names are essentially related to “definite descriptions” or “demonstratives,” they are actually reducible to neither (ML, 168). The use of demonstratives assumes that the speaker locates herself in a shared world of space-time objects, but this can only be accomplished by referring to names and definite descriptions of enduring things and it therefore demands the ability “to recognize a named or described object as this object” (ML, 168). The first provisional conclusion is, thus, that a mutual relationship exists between demonstratives and descriptions (ML, 168). The issue is subtler and intriguing, however: “Granted that names are an irreducible mode of reference, what are the implications of the idea that every individual thing is nameable?” For if anything is a central fact in Leibnitz’s metaphysics, it is that he clearly assumes that every substance is nameable, and I believe that the recognition of this fact throws a flood of light on his system. (ML, 169) Names are allegedly an irreducible mode of reference and the ‘Principle of Nameability’ says that every individual substance is different, reminding

The lingua franca of Nominalism  51 us of the principle of indiscernibles. Yet, the question is: what kind of substances are we referring to? Are we dealing with real or possible individuals? Strictly speaking, the proper name is a criterion that distinguishes its nominatum from all other individuals. But while we tend to think that the individual concept specifies only a few facts about the nominatum, Leibniz believes that the individual concept specifies everything the nominatum does or experiences throughout its career (ML, 170). Once again, Leibniz is concerned with God’s sense of names. If we assume that God can nominate every logically possible substance (not only actual ones), it follows that we might wonder whether the possible coincides with the nameable, or whether the former are broader in extent than the latter (ML, 171). The question is striking, given that Leibniz distinguishes the concept of possibility from the more restrictive one of compossibility. In Leibniz’s scheme of things, being possible is not enough to ensure actual being, since every possible individual must also prove to be compatible with a number of other individuals. Hence the problem is to understand to which of the following two cases the nameability will belong: (1) if nameability coincides with logical possibility, then the individual concept must provide a complete description of its nominatum, so that it can be distinguished from any other individual (whether it exists or not); (2) if nameability is conceived in the more restrictive sense of compossibility, the individual concept will provide a sufficient, though incomplete description of its nominatum. In this case, the distinguishability of substances would not be completely internal, but would follow from the “distinguishability of the worlds” (ML, 171–2). Sellars suggests that Leibniz opted for the latter solution and, in so doing, he undercut his requirement that “the individual concept selects a substance in terms of a complete description.” But what is at stake behind the nameability of a possible substance? To find out, we have to change our approach. As Sellars says, we have to consider that the primary sense of ‘possible’ in Leibniz resembles a ‘state of affairs,’ like when we say: It is possible that Tom will get well. (ML, 172) One could say that such a claim naturally presupposes the actual existence of an individual, namely a person called Tom. Yet, this objection is not as sound as it seems, however, since we may well ask: “How can one properly argue that there are no possible things on the ground that possible states of affairs concern actual things?” (ML, 172). We might actually accept the idea that a ‘derivative use’ of possible can be introduced, as in:

52  Antonio M. Nunziante There is a possible man in the corner = It is possible that there is a man in the corner. (ML, 173) Basically, the idea is that we can refer to possible things because of the possibility of producing true statements in our language. So, when we say “it is possible that Tom will get well” or “it is possible that there is a man in the corner,” the common ground is that they are both used to produce true statements. The way in which Leibniz approached this modal topic thus has to do with the possibility of constructing true linguistic statements: the idea is that God needs possible worlds in order to build true statements, and thereby produces an actual world. In Sellars’s view, the final page of the story consists in intending the possibles as part of a process that imply the creation of an actual world, according to the model of a fictional speech. Here is the passage: What I am suggesting is that at the back of Leibnitz’ mind is the picture of God as making use, within the fictional rubric, of alternative languages, and by so doing conceiving of alternative sets of individual substances. (ML, 180) and the conclusion: According to this picture, the model for creation is obviously the removing of the fictional rubric from one of these languages; the move, on God’s part, from “Suppose that there were such and such things” to “There are such and such things,” via “Let there be such and such things.” (ML, 180) The actual world takes shape as the result of a functional process in which fictional degrees (possible-but-not-actual statements) gradually decrease until a complete individual has been produced. To create means to declassify the fictionality of a language. Needless to say, the very possibility of a fictional discourse presupposes some sense of actuality, just as a Dickens has to exist prior to an Oliver Twist. But Leibniz seems to extend this dimension of possibility even to God, given that on his account God is an entity that necessarily exists but only if it is possible (ML, 181). This is the last metaphysical step that Sellars develops in his Meditations. The Divine Understanding is the locus of possibles, namely the place where a creative process occurs and different alternatives are taking shape. This also explains why the final output (the actual world) remains contingent, since the very fact of being created brings with it a whole bundle of not-actuated possibilities that contribute to defining how the process is accomplished.19 The matter could even be approached in reverse: we might say that possible substances are a prerequisite in the process, since God is but the thought

The lingua franca of Nominalism  53 of possibilities. In that case, the conclusion is that, while defining the nature of His possible characteristics, God also defines Himself. In defining the nature of the individuals that will populate the world, God also defines the actuality of His own language. Sellars, thus, paradoxically concludes his essay on Leibniz by saying that it is through us and our present existence that God discovers Himself: “If my positive argument is correct, the actuality of God, as of anything else, would presuppose our existence as discoverers of God” (ML, 181). It is as if God were like the mind of a novelist lighting up when facing the possibility of giving life to some possible characters in a coherent novel. Such activity gradually takes more definite shape and finally gives names to its creatures. By naming them, God also names Himself as their author—adding His signature, as it were, to His own creation.20

Conclusion The analysis conducted so far has attempted to retrace the main lines along which Sellars interpreted Leibniz. It is now time to look at how and, above all, from what standpoint his interpretation should be assessed. This is no easy task, and the reader will have realized that we have largely avoided introducing interpretative issues so as not to complicate matters even further. However, if we were to adopt the criteria of a historiographic assessment, then we should also acknowledge that Sellars:

• refers very little to direct sources (he almost exclusively cites the Dis• •

• •

course on Metaphysics, with a few indirect references to the New Essays and Monadology, at most); refers very little to any critical literature (only mentioning Russell and Broad, while the works of Couturat, for example, which also marked a decisive turning point in Leibniz studies, are never named); sometimes comes up with philologically inaccurate claims (the doctrine of relationships is not as Sellars presents it; the notion of complete concept is not so unequivocal; and the question of representational solipsism is less straightforward than it appears); employs decontextualized analytical tools (the whole issue of proper names and the reference to particulars are things that do not belong to Leibniz’s apparatus); sometimes advances arguments that are not reflected in the texts (such as the way in which God conceives creation, and ML’s final arguments).

Even so, there are just as many important elements to consider. Adopting broader filters in our assessment, we realize that Sellars:

• clearly grasps some of the most insightful and crucial points of Leibniz’s thought: the internal pressure between a ‘thing-nature framework’ and a ‘law-event framework’ is a core feature of his historically shaped philosophy on which contemporary interpreters largely agree (Di Bella

54  Antonio M. Nunziante

•

2005); the doctrine of relationships, though internally more complex, is certainly a fundamental part of his thought (Mugnai 1992); and the theme of heritage of topics from Leibniz in Kant is equally crucial and surprisingly topical (Look, in press); also construes part of his own philosophical strategy in terms of a comparison with Leibniz: the idea of ​​‘fact’ as “true statement in our language” (SM, 116; ML, 167), and the idea that relationships can be transformed into categorical facts (ML, 162) are just two of the most striking examples of this.

We can thus arrive at some general considerations about how Sellars saw the relationship between philosophy and the history of philosophy. It could be said that he espouses a holistic conception of the history of philosophy. To paraphrase the myth of the genius Jones, we might say that, as private speeches are forged from public language, so too can ‘private’ philosophical insights be framed solely on the grounds of an internalization of public (historically produced) philosophy. From this point of view, the idea of the history of philosophy as the lingua franca of thought would be a matter not only of mere communication, but also of the production of ideas. It is true that Sellars’s reconstructions are very often ideal-typical rather than historical, but it is also true that he was working at a time when specialization had yet to produce its disruptive effects. The subsequent fortunes of the historiography on Leibniz help us to understand this point: up until the Sixties, works published on Leibniz were relatively few and often unsatisfactory. But the decades that followed saw an exponential increase in intensity of critical debates, due to the parallel activities and publications produced by the Leibniz-Archiv (1962), the international Leibniz-Kongressen (from 1966 onwards), and the Studia Leibnitiana review (1969). After their arrival on the scene, most previous studies suddenly became obsolete. Sellars was a true precursor from this point of view because, despite all his philological limitations, his approach was new in the analytical world. Through his ideas on the relationship between philosophy and the history of philosophy, he generated something that simply had not been done before, starting a trend that is still of great interest today: the idea that philosophy is both a theoretical and, at the same time, an intrinsically historical discipline. This is a notion that, in some respects, would be interesting to apply to the contemporary metaphilosophical debate.

Notes 1 Sellars attributes to Russell a crucial mediating role, both in terms of his analytical interpretation of Leibniz (Russell 1900) and his reflection on particulars and proper names (Russell 1910, 1911, 1940). He can thus be seen as the traitd’union between Leibniz’s two lines of inquiry discussed here. 2 The first references to Leibniz can be found in his M.A. thesis in Buffalo (Substance, Change and Event, 1934), in which he read Leibniz through the lens of Russell’s interpretation. There are some lecture notes too, presumably written

The lingua franca of Nominalism  55 by Sellars at Oxford (1934–36), while attending a course on Leibniz held by J. L. Austin (“Austin-Leibniz”, undated, in Wilfrid S. Sellars Papers, 1899–1990, ASP.1991.01, Archives of Scientific Philosophy, Special Collections Department, University of Pittsburgh, Box 7, Folder 23). 3 Sure enough, the principle of reason refers to the idea that the whole series of episodes is grounded in something that is not episodic but self-subsistent: God. 4 In his essay on Berkeley and Descartes (BD, 363), Sellars traces such a distinction by referring to the Appendix of the second set of Objections and Replies to Descartes’s Meditations on First Philosophy. The distinction also has an important role in SM 31. A careful analysis focused on the subtle, yet decisive, distinctions among the Cartesian’s notions of ‘formal reality,’ ‘objective reality’ and ‘material falsity’ of ideas is developed in Moran (2014, pp. 67–92). 5 “There is no motion when there is no change that can be observed. And where there is no change that can be observed, there is no change at all” (Leibniz 1989, 340). This quotation was not produced by Sellars himself: I report it here for the sake of clarity. 6 “So it is obvious that unless there were change in simple things, there would be no change in things at all. Indeed, not even change can come from without, since, on the contrary, an internal tendency to change is essential to finite substance, and change could not arise naturally in monads in any other way” (Leibniz 1989, 177). Here again, I refer to quotations that might help to clarify the metaphysical meaning of Leibniz’s nominalism. For more historical details, and to see how Leibniz managed to combine idealism, nominalism and phenomenalism, see Adams (1994). 7 On the different models of Leibniz’s conception of ‘nestedness,’ see Nachtomy (2006, 225–6). 8 “There is a continuum of beyondness which corresponds to the continuum of mediacy of representing.” Sellars plays here with the ‘direction’ of spatial continuity and the continuity of the ‘indirectness’ of the representation. It is perfectly clear, he adds, “that you can map a spatial continuum (supposing that there is such a thing) into a continuum of indirectness of representation”: Leibniz gives us “a metaphysical model for geometry” (KPT, 256). 9 The idea that there might be something actual corresponding to the representation of S1 pertains to the nature of divine decrees, since it refers to God’s decision to create both terms. Some scholars have rightly pointed out that “analytically deriving the existence of other individuals or external objects from the concept of perception is totally misleading”. What we can say is that “the perceptual nature of all substantial states brings it about that a certain series of internal (we would say ‘intentional’) objects is part of the individual’s essence”. Paris’s love for Helen “does not properly end with Helen, to be sure, hence it does not necessarily imply Helen’s existence in flesh and blood ‘out there in the world’; it is directed, however, towards a certain representational object” (Di Bella 2005, 345–6). 10 An interesting question is whether it was Leibniz who influenced Sellars on this topic or, vice versa, whether Sellars’s theoretical tools allowed him to interpret Leibniz in this way. 11 For the distinction between rules of criticism (ought-to-be) and rules of action (ought-to-do), see LTC, 507–8; DeVries (2005, 43–6); O’Shea (2007, 79–83). 12 According to Sellars, in Early Modern epistemology (Kant included) we find no clear-cut distinction between representation as a “representing act” and representation as a “represented object” (KTE 269; KPT 27; SM 36). 13 The idea that the act-content model is a false relational model, and that this kind of model provides the conceptual key to understanding the epistemology of modern thought, was developed by Sellars right at the start of his lectures on Kant (KPT, 8–9).

56  Antonio M. Nunziante 14 According to Henry Allison, “the defining characteristic of transcendental realism is its confusion of appearances, or ‘mere representations’, with things in themselves”. It was only Kant’s critical philosophy that succeeded in clarifying this distinction (Allison 2004, 22–3). 15 The argument is also developed in KTE, 269. 16 “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, 270). 17 According to Kant, cognitive content can exist only as represented, or as representable, not in an absolute sense—as a pure realist would have it. 18 Here the quotation resembles Wittgenstein’s Philosophical Investigations § 265. As Sellars points out in KTE, “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, in 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” (KTE, 282). For an analysis of what is involved here for Sellars’s idea of Kant, see O’Shea (2007, 134–6). 19 Piro (2009) remarks that every event is the bearer of a not-actuated bundle of possibilities. Each event becomes “actual” precisely by virtue of this bundle of not-actuated possibilities. See Piro (2009, 540–2). 20 Interestingly enough, the conclusion was rather different and, to my mind, less radical in an earlier version of the paper. Here, in fact, Sellars concluded his analysis by saying: “If, then, our promissory note—which can be abbreviated, with proper precautions into the claim that facts are true statements, and categories of fact, categories of true statement—if our promissory note can be cashed, our discussion will have achieved two purposes: (1) It will have exhibited the basic role which an unformulated in re conception of truths or facts played in rationalistic thinking; (2) It will have served to indicate the radical character of the measures which must be taken if the many important insights contained in rationalistic systems are to be translated into contemporary terms” (Leibnitz Rationalism: Scaffolding for a Reconstruction (Part I), undated, in Wilfrid S. Sellars Papers, 1899–1990, ASP.1991.01, Archives of Scientific Philosophy, Special Collections Department, University of Pittsburgh, Box 33, Folder 9, 17).

References Adams, Robert Merrihew. 1994. Leibniz. Determinist, Theist, Idealist. New York and Oxford: Oxford University Press. Alanen, Lilli. 1992. ‘Thought—Talk: Descartes and Sellars on Intentionality’. American Philosophical Quarterly, 29: 19–34. Allison, Henry. 2004. Kant’s Transcendental Idealism. New Haven and London: Yale University Press. Brandom, Robert. 2015. From Empiricism to Expressivism: Brandom Reads Sellars. Cambridge, MA: Harvard University Press. Brandom, Robert. 2017. ‘Two Kantian Axes of Sellars’ Thought’. In Sellars and Contemporary Philosophy. Edited by David Pereplyotchik and Deborah R. Barnbaum, 197–218. New York and London: Routledge.

The lingua franca of Nominalism  57 de Vries, Willem A. 2005. Wilfrid Sellars. Montreal and Kingston: McGill-Queen’s University Press. Di Bella, Stefano. 2005. The Science of the Individual: Leibniz’s Ontology of Individual Substance. Dordrecht: Springer. Haag, Johannes. 2017. ‘A Kantian Critique of Sellars’ Transcendental Realism’. In Wilfrid Sellars, Idealism and Realism: Understanding Psychological Nominalism. Edited by Patrick J. Reider, 149–71. London: Bloomsbury Academic. Landy, David. 2008. ‘Sellars on Hume and Kant on Representing Complexes’. European Journal of Philosophy, 17: 224–46. Leibniz, Gottfried Wilhelm. 1989. Philosophical Essays. Edited and translated by Roger Ariew and Daniel Garber. Indianapolis, IN: Hackett Publishing Company. Look, Brandon. (Ed.). Forthcoming. Leibniz and Kant. Oxford: Oxford University Press. Macbeth, Danielle. 2000. ‘Empirical Knowledge: Kantian Themes and Sellarsian Variations’. Philosophical Studies: An International Journal for Philosophy in the Analytic Tradition, 101, 2/3: 113–142. McDowell, John. 2006. ‘Sensory Consciousness in Kant and Sellars’. Philosophical Topics, 34: 311–26. Moran, Dermot. 2014. ‘Descartes on the Formal Reality, Objective Reality, and Material Falsity of Ideas: Realism through Constructivism?’ In Realism, Science and Pragmatism. Edited by Kenneth R. Westphal, 67–-92. New York and London: Routledge. Mugnai, Massimo. 1992. ‘Leibniz’ Theory of Relations’. Studia Leibnitiana, Supplementa XXVIII. Nachtomy, Ohad. 2006. Possibility, Agency and Individuality in Leibniz’s Metaphysics. Dordrecht: Springer. O’Shea, James R. 2007. Wilfrid Sellars: Naturalism with a Normative Turn. Cambridge: Polity Press. O’Shea, James R. 2011. ‘How to Be a Kantian and a Naturalist About Human Knowledge: Sellars’s Middle Way’. Journal of Philosophical Research, 36: 327–59. O’Shea, James R. 2016. ‘What to Take Away from Sellars’s Kantian Naturalism’. In Sellars and his Legacy. Edited by James R. O’Shea, 130–48. Oxford: Oxford University Press. O’Shea, James R. 2017. ‘Thought, Freedom, and Embodiment in Kant and Sellars’. In Sellars and Contemporary Philosophy. Edited by David Pereplyotchik and Deborah R. Barnbaum, 15–35. New York and London: Routledge. Piro, Francesco. 2009. ‘Lo scolastico che faceva un partito a sé (faisat band à part): Leibniz su Durando di San Porziano e la disputa sui futuri contingenti’. Rivista di Medioevo, 34: 507–43. Rauzy, Jean-Baptiste. 2009. ‘Sellars et Bergmann Lecteurs de Leibniz: La querelle des particuliers’. In Gustav Bergmann: Phenomenological Realism and Dialectical Ontology. Edited by J.M. Monnoyer and B. Langlet, 87–102. Frankfurt am Main: Ontos Verlag. Russell, Bertrand. 1900. A Critical Exposition of the Philosophy of Leibniz. Cambridge: Cambridge University Press. Russell, Bertrand. 1910. ‘Knowledge by Acquaintance and Knowledge by Description’. Proceedings of the Aristotelian Society, 11: 108–28. Russell, Bertrand. 1911. ‘On the Relations of Universals and Particulars’. Proceedings of the Aristotelian Society, 12: 1–24.

58  Antonio M. Nunziante Russell, Bertrand. 1940. An Inquiry into Meaning and Truth. London: George Allen and Unwin Ltd. Sellars, Wilfrid. 1934. Substance, Change and Event. MA thesis, State University of New York at Buffalo. http://www.ditext.com/sellars/sce.html. Sellars, Wilfrid (APM). 1949. ‘Aristotelian Philosophies of Mind’. In Philosophy for the Future: The Quest of Modern Materialism. Edited by Roy Wood Sellars, Vivian Jerauld McGinn and Marvin Farber, 544–70. New York: The MacMillan Company. Sellars, Wilfrid (LCP). 1949. ‘On the Logic of Complex Particulars’. Mind, 58: 306–38. Sellars, Wilfrid (P). 1952. ‘Particulars’. Philosophy and Phenomenological Research, 13: 184–99. Sellars, Wilfrid (ML). 1959. ‘Meditations Leibnitziennes’. In Philosophical Perspectives. Edited by Wilfrid Sellars, 153–81. Springfield, IL: Charles Thomas Publisher. Sellars, Wilfrid (EPM). 1963. ‘Empiricism and the Philosophy of Mind’. In Science, Perception and Reality. Edited by Wilfrid Sellars, 127–96. London: Routledge & Kegan Paul. Sellars, Wilfrid (AE). 1963. ‘Abstract Entities’. Review of Metaphysics, 16: 627–71. Sellars, Wilfrid (SM). 1968. Science and Metaphysics: Variations on Kantian Themes. Atascadero, CA: Ridgeview Publishing Company. Sellars, Wilfrid (LTC). 1969. ‘Language as Thought and Communication’. Philosophy and Phenomenological Research, 29: 506–27. Sellars, Wilfrid (BD). 1977. ‘Berkeley and Descartes: Reflections on the “New Way of Ideas” ’. In Studies in Perception: Interpretation in the History of Philosophy and Science. Edited by P.K. Machamer and R.G. Turnbull, 259–311. Columbus, OH: Ohio State University Press. Sellars, Wilfrid (KPT). 2002. Kant and Pre-Kantian Themes: Lectures by Wilfrid Sellars. Edited by Pedro Amaral. Atascadero, CA: Ridgeview Publishing Company. Sellars, Wilfrid (KTE). 2002. ‘Some Remarks on Kant’s Theory of Experience’. In Kant’s Transcendental Metaphysics. Edited by Jeffrey Sicha, 269–82. Atascadero, CA: Ridgeview Publishing Company. Sellars, Wilfrid. Leibnitz Rationalism: Scaffolding for a Reconstruction (Part I), undated, in Wilfrid S. Sellars’s Papers, 1899–1990, ASP.1991.01, Archives of Scientific Philosophy, Special Collections Department, University of Pittsburgh, Box 33, Folder 9.

3 Sellars and Hume on the Ontological Status of Theoretical-Explanatory Entities David Landy

Throughout his long and prolific career, Sellars makes very clear that he is, in more senses than not, a Kantian, through and through. From his explicit endorsements of various Kantian theses to his subtle critique and amendments to pieces of Kant’s philosophical system,1 there is perhaps no philosopher that Sellars associates with more closely.2 It is not surprising, then, that like Kant,3 at times Sellars presents his philosophical system by way of a contrast with Hume’s. That contrast most commonly takes the form of objections to one of the most prominent aspects of Hume’s (often implicit) account of mental representation: his imagism.4 As Sellars understands him, Hume holds that our thoughts represent what they do by virtue of the fact that they are images: sights, sounds, smells, touches and tastes conceived of as directly introspectable ‘perceptions,’ or as Sellars would call them, sensedata. There is certainly something to this interpretation of Hume. Consider, for instance, Hume’s account of the origin of our representations of spatial complexes. The table in front of me is alone sufficient by its view to give me the idea of extension. This idea, then, is borrow’d from, and represents some impression, which this moment appears to the senses. But my senses convey to me only the impressions of colour’d points, dispos’d in a certain manner. If the eye is sensible of any thing farther, I desire it may be pointed out to me. But if it be impossible to show any thing farther, we may conclude with certainty, that the idea of extension is nothing but a copy of these colour’d points, and of the manner of their appearance. (T 1.2.3.4; SBN 34) Our complex idea of a spatial complex comes to represent the spatial complex that it does by being a collection of simple ideas of colored points arranged in a way that exactly resembles the arrangements of the spatial complex being represented. We represent the relation that some simple impressions stand in to one another by arranging simple representations of each of these impressions into the same relation. We represent a as being next to b by placing an idea of a next to an idea of b. The idea of a spatial

60  David Landy complex is nothing more than a spatial complex of ideas. Hume is clear that our representation of temporal complexes works in the same manner. The idea of time being deriv’d from the succession of our perceptions of every kind, ideas as well as impressions, and impressions of reflection as well as of sensation, will afford us an instance of an abstract idea, which comprehends a still greater variety than that of space, and yet is represented in the fancy by some particular individual idea of a determinate quantity and quality. (T 1.2.3.6; SBN 34) Our idea of time is “deriv’d from the succession of our perceptions.” Hume’s thought is that we represent two items as being related, now temporally, by placing them in a temporal relation to one another. That is, for example, we represent one thing as happening before another by having a representation of the former followed by a representation of the latter. So, whereas we represent a spatially complex state of affairs by forming a kind of picture before our mind’s eye, we represent a temporally complex state of affairs by forming a kind of movie there. While Sellars certainly does not oppose the suggestion that there is something to the idea of mental images, he consistently and vehemently resists the thesis that these can play the specific role in determining the content of our representations that he takes Hume to assign to them. For example, in Science and Metaphysics, Sellars takes Hume to task precisely for the account of representation above. Hume, in any case, strode over all these complexities with seven league boots, for, as is notorious, he confused between: (1) an impression of a green square; (2) a conviction that a green square exists; (3) a green square. And, correspondingly, between: (1) an impression of a green square adjoining a red square; (2) a conviction that a green square adjoining a red square exists; (3) a green square adjoining a red square. (SM, §71) Sellars takes Hume to hold that one represents a green square as adjoining a red square by doing nothing more than having a mental image that consists of a green square adjoining a red square (and that one believes in the existence of such a state of affairs by properly enlivening that mental image). As he argues earlier in that section, however, Sellars follows Kant in holding that representations of complex states of affairs as complex cannot consist of mere complexes of representations, and specifically these require distinctly normative-conceptual resources that Hume’s view disallows.5

Sellars and Hume  61 Similarly, in “Empiricism and the Philosophy of Mind,” Sellars famously affords himself a digression—the section entitled, “Impressions and Ideas: and Historical Point”—to critique Hume (along with Locke and Berkeley) for holding the thesis that, “the idea that the awareness of certain sorts—and by ‘sorts’ I have in mind, in the first instance, determinate sense repeatables— is a primordial, non-problematic feature of ‘immediate experience’ ”(EPM §79). In other words, Sellars takes all three of these philosophers to task for holding a version of the Myth of the Given, one manifestation of which is the thesis that merely by being in a certain naturalistically describable state—in this case the state of having a certain sense impression—we thereby come to represent the world as being some way—in this case we represent ourselves as having that very sense impression. Furthermore, Sellars sees Hume as adopting the specific version of the Myth also propagated by the proponent of sense-data theories, namely that our talk of sensible qualities, e.g., red, is constructed from talk about mental states, e.g., looks-red. As Sellars sees it, Hume et al. reverse the actual order of logical priority. It is our talk of, e.g., an object’s being red that is logically prior to our talk of objects as merely looking red, and our talk of the mental states that are in some sense necessary for a representation of red to be an ostensible seeing of something as red are of a still different logical order (one belonging to the scientific image of ourselves as objects in the world first heralded by mythical Jones).6 Thus, it is no surprise at all that Sellars sees Hume as one of his, and Kant’s main philosophical antagonists. Sellars sees Hume as holding that being in a certain mental state is sufficient for representing the world as being in a certain way, that this provides us with direct and unproblematic access to these mental states, and that these mental states are the logical (rather than merely causal) foundation from which our representations of the world are constructed. Sellars makes other assorted criticisms of Hume along these lines elsewhere, but the gist of the matter is that Hume’s imagism is a version of the Myth of the Given, and is simply inadequate to its task at hand: accounting for the representational content of our mental representations. With all of those very serious points of disagreement, what would be a surprise is if there was anything left over for Sellars and Hume to agree on. Of course, that there are such points is precisely what I will argue for in what follows, and to do so, I will turn from Sellars’s views of philosophy in the Modern period to his views of philosophy in his own time. At the turn of the previous century the so-called Logical Positivists appropriated Hume as their historical forefather, casting him as endorsing something very much like their own account of scientific explanation, the Deductive-Nomological account (DN).7 According to DN, the explicandum of a scientific explanation is an observed particular, and its explanans is an inductively established universal generalization. For example, we observe the position of Mercury in the night sky and explain this position by deducing it from Kepler’s laws of planetary motion and the previously observed

62  David Landy position of Mercury. Scientific explanation and progress takes the form of discovering propositions of increasing generality that subsume an increasing variety of particular phenomena under them. That picture of scientific explanation has long been recognized as inadequate, and yet the belief that Hume holds a view very similar to it has proven remarkably and unfortunately durable. I believe that something like it is an implicit presupposition of many current approaches to Hume, and versions of this interpretive line have recently been explicitly defended by several scholars.8 I have argued that despite the prevalence of this interpretation of Hume, a close examination of his actual procedures in pursuing what he calls the science of human nature reveals that he is working with a more sophisticated, if largely inchoate, explanatory methodology. In fact, I have argued that Hume’s understanding of scientific explanation can be helpfully interpreted by using Sellars’s as a model. Consider a few of Sellars’s main critiques of DN. First, Sellars argues that DN misidentifies both the explicandum and the explanans of scientific explanations. As Sellars sees it, the explicandum of a scientific theory is not observed particulars, but rather the fact that the particulars are subject to the empirical generalizations that they are. So, in the previous example, it is not the position of Mercury alone that demands explanation, but rather the regular motions of the planets, or Kepler’s law itself. Furthermore, as Sellars sees it, the explanans of such phenomena makes essential appeal to the nature of the particulars at hand. theories about observable things do not “explain” empirical laws in the manner described, they explain empirical laws by explaining why observable things obey to the extent that they do, these empirical laws; that is, they explain why individual objects of various kinds and in various circumstances in the observation framework behave in those ways in which it has been inductively established that they do behave. Roughly, it is because a gas is—in some sense of ‘is’—a cloud of molecules which are behaving in certain theoretically defined ways, that it obeys the empirical Boyle-Charles law.9 Along these lines, despite his famous dictum, “Hypothesis non fingo,” Newton nonetheless attempts to explain the law of universal gravitation, the inductively established universal generalization that supposedly itself explains the position of Mercury, by appeal to the differing densities of the aether surrounding massive bodies.10 That is, he attempts to explain it by casting that empirical regularity as itself being a feature of an underlying unobservable substance. Misunderstanding the proper explicandum and explanans of a scientific theory in turn leads to another important mistake that Sellars finds in DN: it overlooks the methodological importance of explanation’s dependence on perceptible models.

Sellars and Hume  63 the fundamental assumptions of a theory are usually developed . . . by attempting to find a model, i.e. to describe a domain of familiar objects behaving in familiar ways such that we can see how the phenomena to be explained would arise if they consisted of this sort of thing. The essential thing about a model is that it is accompanied, so to speak, by a commentary which qualifies or limits . . . the analogy between the familiar objects and the entities which are being introduced by the theory.11 On Sellars’s understanding of theoretical explanation, one explains the fact that certain particulars are subject to certain empirical generalizations via an appeal to the fact that these particulars have—as Hume puts it—certain natures, qualities or powers.12 These qualities of the particulars, in turn, are represented via perceptible models, which are specified to resemble and differ from the posited entities in determinate ways. To switch hackneyed examples for a moment, consider a familiar case of such modeling from the sciences: the Bohr model of the atom. Taking the solar system as its perceptible model, Bohr proposes that we understand the structure internal to an atom as analogously consisting of a large nucleus being orbited by smaller electrons, and also specifies determinate differences between the theoretical entity posited and its perceptible model: The former is orders of magnitude smaller, it is held together by electrostatic forces instead of gravity, the electrons’ orbits are circular rather than elliptical, etc. Finally, Sellars argues that in miscasting scientific explanation as aiming only at subsuming particulars under empirical generalities of increasing scope, DN also accords undue ontological priority to the objects represented by the common-sense framework that is the (merely) methodological starting place of such explanations. For to suppose that particular observable matters of fact are the proper explananda of inductive generalizations in the observation framework and of these only, is to suppose that, even though theoretical considerations may lead us to formulate new hypotheses in the observational framework for inductive testing and may lead us to modify, subject to inductive confirmation, such generalizations as have already received inductive support, the conceptual framework of the observation level is autonomous and immune from theoretical criticism.13 Contra DN, Sellars famously argues that, “In the dimension of describing and explaining the world, science is the measure of all things, of what is that it is, and of what is not that it is not,”14 and so while the commonsense picture of the world does enjoy a methodological priority, because of its ultimate explanatory success, it is the scientific picture that will represent the world’s real ontology.

64  David Landy So, what we have in Sellars is a rejection of DN that includes the following alternative theses: (a) The explicanda of a scientific explanation are not observed particulars, but the fact that these particulars obey the inductively established universal generalizations that they do. (b) The explanans of a scientific explanation is not an empirical generalization that subsumes particular phenomena, but a representation of the substance underlying that phenomena features of which account for its observed behavior. (c) This representation of the substance underlying the explicanda is constructed by specifying determinate similarities and differences between some antecedently familiar model object and the posited substance. Surprising as it may be, I believe that, contra to the long history of casting Hume as accepting something like DN, Hume would actually accept all three of these theses.15 Now, defending an interpretation of Hume as employing something like this Sellarsian understanding of scientific explanation is an ambitious project that I certainly cannot complete here. What I can do, however, is present at least a few parts of that defense, and specifically those parts that make the most contact with the Sellarsian model as I have introduced it. To that end, in what follows I will address an aspect of Hume’s understanding of his pursuit of the science of human nature that I believe has been largely overlooked and misinterpreted. It is in his explication of the proper understanding of the idea of substance that I will argue we can see Hume committing to theses (a)-(c) above. The idea of substance, properly conceived, is the idea of positing an unobserved cause underlying observed regularities using some antecedently familiar phenomena—for Hume this will always be some perception—as a model. Hume devotes a small but important early section of Book I, 1.1.6, “Of modes and substances,” to accounting for the difference between substances and modes, and my plan here is to work through that section in some detail.16 Hume begins there by challenging those philosophers who have relied on the specious distinction between substances and accidents to produce the impression from which the idea of substance is derived. After demonstrating that they will not be able to do this because ex hypothesi this notion of substance is one according to which a substance is that in which perceptible qualities inhere, but which itself is entirely different from such qualities, Hume concludes: We have, therefore, no idea of substance, distinct from that of a collection of particular qualities, nor have we any other meaning when we either talk or reason concerning it. (T 1.1.6.1; SBN 16)17

Sellars and Hume  65 Having demonstrated that our idea of substance is nothing over and above the idea of a collection of particular qualities, Hume recognizes that this leaves him with an explanatory burden insofar as this appears to undermine the intuitively plausible distinction between substances and modes. The ideas of substance as well as that of a mode, is nothing but a collection of simple ideas, that are united by the imagination, and have a particular name assign’d them. (T 1.1.6.2 SBN 16) The reference to having a particular name assign’d to them is an anticipation of the theory of general ideas that Hume will introduce in the following section. In a moment, we will consider exactly how Hume puts that theory to service in this context. What is important now is just that at this point in the section, Hume appears to be claiming that the idea of a substance and the idea of a mode are both just instances of the theory of general representation, and so is in danger of losing the distinction between them. Thus, Hume’s next move is to recover that distinction in a series of dense and obscure claims that require unpacking. He begins with the idea of substance: the particular qualities, which form a substance . . . are at least suppos’d to be closely and inseparably connected by the relations of contiguity and causation. (T 1.1.6.2; SBN 16) The idea of substance, properly conceived, is the idea of some particular qualities, assigned the same name, that are “suppos’d to be closely and inseparably connected by the relations of contiguity and causation.” The relation of contiguity that Hume refers to here is the contiguous appearance of each of the qualities at hand, but that turns out to be a non-starter because as Hume notes a moment later, it is easy enough to imagine a scenario in which a mode displays this same relation. The simple ideas of which modes are form’d, either represent qualities, which are not united by contiguity and causation, but are dispersd’d in different subjects; or if they be all united together, the uniting principle is not regarded as the foundation of the complex idea. (T 1.1.6.3; SBN 17) As Hume notes here, while modes are typically “dispers’d in different subjects,” they need not be, and when they are not, what distinguishes a mode from a substance is that with respect to the former, “the uniting principle is not regarded as the foundation of the complex idea.” ‘The complex idea’ here refers to the general representation that is used to represent the substance, and it is striking that Hume takes such ideas to operate via a uniting principle that serves as their foundation. That is surprising because in

66  David Landy the section immediately following this one Hume presents his theory of general representation, and there portrays general representation as depending on only resemblance relations, making no mention of the uniting principles serving as the foundation of such representations. In 1.1.7 Hume adopts Berkeley’s arguments against Locke for the conclusion that all representation proceeds via perceptions that are themselves fully determinate.18 He then goes on to explain how such perceptions can themselves be used to represent, for example, not just this or that person, but persons in general. When we have found a resemblance among several objects, that often occur to us, we apply the same name to all of them, whatever differences we may observe in the degrees of their quantity and quality, and whatever other differences may appear among them. After we have acquir’d a custom of this kind, the hearing of that name revives the idea of one of these objects, and makes the imagination conceive it with all its particular circumstances and proportions. But as the same word is suppos’d to have been frequently apply’d to other individuals, that are different in many respects from that idea, which is immediately present to the mind; the word not being able to revive the idea of all these individuals, only touches the soul, if I may be allow’d so to speak, and revives that custom, which we have acquir’d by surveying them. They are not really and in fact present to the mind, but only in power, nor do we draw them all out distinctly in the imagination, but keep ourselves in a readiness to survey any of them, as we may be prompted by a present design or necessity. (T 1.1.7.7; SBN 20–1) When we encounter items that resemble one another in certain ways, we call them all by the same name. Because this name is frequently used in the presence of these items our minds form an association between the two such that whenever that name is used we call to mind some one, or a few, ideas of these items. Furthermore, we stand disposed to call to mind more such ideas (ideas of the items between which this resemblance was found) upon further prompting. So, we represent persons in general by forming certain associations between our ideas of particular persons. Namely, those ideas all resemble one another, and are called to mind in the appropriate situations (primarily when the word ‘person’ is used, etc.) because of this resemblance. While that is Hume’s account of how general terms represent, it leaves open the question of what we represent using them. I have argued elsewhere that what is represented by a general term is that the objects that we are disposed to recall upon hearing that term (and further prompting) all resemble one another.19 That is, according to Hume’s account, what prompts us to form general ideas in the first place is that, “we have found a resemblance among several objects,” and what is represented by an idea prompted by such encounters is that these objects as resembling one another. So, it is not

Sellars and Hume  67 just that our general idea ‘personhood’ represents persons, but more specifically, it represents persons as members of a set of objects among which we have found a resemblance. It represents persons as persons.20 Notice that as expected this account of general terms applies to both modes and substances. We can equally well represent persons (qua substances) as resembling one another by using the general term ‘person’ as we can represent smiles (qua modes) as representing one another by using the general term ‘smile’. Thus, Hume must account for this difference via something other than his account of general representation itself, and so appeals to the uniting principle that serves as the foundation of the representation of substance. The representation of a substance, unlike that of a mode, has a uniting principle and a foundation. That is, it is more than a mere collection of qualities: it is a collection of qualities that together compose the powers or nature of some real thing.21 That is getting a bit ahead of ourselves. Where things stand in our exegesis to this point is as follows. What we have just discovered is that even in cases in which a certain quality can be found among all of the ideas that compose the revival set of some general term, that quality will not be counted as belonging to the substance represented by that general term unless it meets a further criterion: falling under the uniting principle that is the foundation of the complex idea. Having eliminated contiguity as a reliable criterion by which to distinguish modes from substance, we are left with the causal relations that Hume mentions as the only remaining contender for playing the role of this uniting principle. What I now want to show is that the most plausible such causal principle is a theoretical-explanatory one. At this point, we have seen Hume cast substances and modes as both being collections of certain qualities. Hume then distinguishes these by noting that the former, but not the latter, stand in certain relations of contiguity and causation. He then quickly recants the importance of the former relation, leaving causation as the foundation of our proper representation of substance. Without further specification, however, it is unclear precisely which causal relation Hume has in mind here. Here are three possibilities. The ideas that compose the idea of a substance might be supposed to: (a) stand in causal relations to each other (Q1 causes Q2), (b) collectively play a certain causal role (Q1 and Q2 together cause E). (c) have a common causal explanation (the existence of both Q1 and Q2 are explained by the nature of substance S). By way of explaining each of these three possibilities, consider Hume’s own example of the qualities of the substance gold. After noticing that what qualities we take to be substantial ones can change as we discover additional qualities that stand in the appropriate causal relation (whatever it may be), Hume writes: Thus, our idea of gold may at first be a yellow colour, weight, malleableness, fusibility; but upon the discovery of its dissolubility in aqua

68  David Landy regia, we join that to the other qualities, and suppose it to belong to the substance as much as if its idea had from the beginning made a part of the compound one. (T 1.1.6.2; SBN 16) Our idea of gold is the idea of a certain kind of substance, and that idea is composed of the ideas of certain qualities: color, weight, malleability, fusibility and eventually solubility in aqua regia. Every sample of gold, though, will also have certain other qualities—e.g., a shape—and so we can ask: what is it that makes the former qualities substantial qualities of gold, rather than modes? As we saw earlier, Hume’s answer is that the substantial qualities, but not the modes, stand in a certain causal relation, and so the next question to be considered is precisely which casual relation this is. For example, to take (a) to be the relevant causal relation would be to suppose that the it is in virtue of gold’s color causing its weight or its malleability causing its fusibility, etc. that these qualities are substantial qualities. That does not seem very plausible at all. For one, Hume is clear that causal relations are never simultaneous, and any and all of these qualities might well be perceived simultaneously. For another, we simply do not think of these qualities as causing one another, despite our thinking of them together constituting the substance gold. Taking (b) to be the relevant causal relation is to suppose that in thinking of certain qualities as substantial qualities, we think of those qualities as playing a unified causal role. For example, we might think that it is the color and malleability of gold that are together the cause of a particular brick of gold’s falling to the ground when dropped. Here again we encounter a conflict between the suggestion at hand and our common conception of the relation of these properties to one another. For example, we do not, in fact, think of the color of gold, much less its fusibility or its solubility in aqua regia, as playing any causal role in a brick of gold’s falling to ground. As Hume explicitly notes in 1.3.13 “Of unphilosophical probability,” just as we need to distinguish between substantial qualities and modal qualities, we must likewise distinguish those causes that are essential to particular event’s occurrence and those that are merely accidental to it. In almost all kinds of causes there is a complication of circumstances, of which some are essential, and other superfluous; some are absolutely requisite to the production of the effect, and others are only conjoin’d by accident. (T 1.3.13.9; SBN 148) The color of gold is “only conjoin’d by accident” to the brick of gold’s falling when dropped, and so is a superfluous rather than an essential cause. Since its color is a substantial quality of gold, but is not always causally

Sellars and Hume  69 efficacious, then it cannot be that causal efficacy is a necessary condition of a quality’s being a substantial quality. Furthermore, because each of these qualities will be causally efficacious in different circumstances, rather than being united by their causal roles, they seem rather to be divided by these. Malleability is a quality that is more closely related to other metals than it is to the yellow color of gold. So, it does not appear to be their collective causal role that unites these properties into a representation of substance, since they do not appear to have such a collective causal role at all. That leaves (c), and I believe that we have good reason to think that Hume has just such a causal relation in mind as that which serves as the ‘uniting principle’ and ‘foundation’ of the general representation of substances. Taking (c) to be the relevant causal relation is to suppose that in thinking of certain qualities as substantial qualities, we suppose that the causal explanation of each quality will make reference to some single thing. It is easy enough, for example, (for us) to consider all of the qualities that Hume lists as substantial qualities of gold—its color, weight, malleability, fusibility and solubility in aqua regia—as explained by its atomic structure. Here again is Hume’s proposal for what makes a quality a substantial quality. the particular qualities, which form a substance . . . are at least suppos’d to be closely and inseparably connected by the relations of contiguity and causation. (T 1.1.6.2; SBN 16) The current suggestion is that to suppose that gold’s color, weight, malleability, et al. are “closely and inseparably connected by the relations of contiguity and causation” is to suppose that there is a single causal explanation tying these qualities together. In the specific case of gold, we have such an explanation ready at hand—it is gold’s atomic structure that accounts for all of these properties—but similar explanations can be given, at least provisionally, for anything that we suppose to be a substance.22 Of course, even to consider this option as a live possibility for Hume, this single thing cannot be a mere “unknown something,” but will have to be (at least potentially) representable by a perceptible model. It cannot be, that is, that we take the substance gold to be a via negativa in which the relevant qualities inhere.23 Thus, in 1.4.3 “Of the antient philosophy,” Hume castigates the Peripatetics for holding that substance is something “unknown and invisible.” In order to reconcile which contradictions the imagination is apt to feign something unknown and invisible, which it supposes to continue the same under all these variations; and this unintelligible something it calls a substance, or original and first matter. (T 1.4.3.4; SBN 220)

70  David Landy What is objectionable here is not seeking an explanation for the constancy and coherence of our perceptions, nor in making an appeal to substance as the means of that explanation, but rather in the appeal to an entirely unintelligible understanding of substance. Later in that same section, discussing how this empty notion of substance leads to an equally empty notion of qualities as accidents that inhere in such a substance, Hume contrasts the specious reasoning of the Peripatetics with the causal reasoning of the scientist of human nature. For having never discover’d any of these sensible qualities, where, for the reasons above-mention’d, we did not likewise fancy a substance to exist; the same habit, which makes us infer a connection betwixt cause and effect, makes us here infer a dependance of every quality on the unknown substance. (T 1.4.3.7; SBN 222) Where the Peripatetic illegitimately infers that accidents inhere in an unknown substance, the scientist of human nature legitimately infers that the qualities of a substance (represented via a perceptible model) are related via a common cause. Again, it is not positing that a substance exists and can explain the behaviors of certain qualities that is specious, but only misunderstanding the nature of that substance and its relation to these qualities that is. Hume makes all of this explicit in the opening paragraph of the following section, 1.4.4 “Of the modern philosophy.” I must distinguish in the imagination betwixt the principles which are permanent, irresistible, and universal; such as the customary transition from causes to effects, and from effects to causes: And the principles, which are changeable, weak, and irregular; such as those I have just now taken notice of. The former are the foundation of all of our thoughts and actions, so that upon their removal human nature must immediately perish and go to ruin. The latter are neither unavoidable to mankind, nor necessary, or so much as useful in the conduct of life . . . For this reason the former are receiv’d by philosophy and the latter rejected. (T 1.4.4.1; SBN 225) Philosophically sound practice relies on the kinds of causal explanations that we have seen Hume take to provide the foundation for the proper notion of substance. Where the Peripatetics go wrong is in attempting to base their explanations on notions that are entirely empty. Hume makes this contrast even clearer in the continuation of this passage in which he draws an analogy between the reasoning of scientist of human nature and that of the Peripatetics to two people reasoning about what is in the darkness around them.

Sellars and Hume  71 One who concludes somebody to be near him, when he hears an articulate voice in the dark, reasons justly and naturally . . . But one, who is tormented he knows not why, with the apprehension of spectres in the dark, may, perhaps, be said to reason, and to reason naturally too: But then it must be in the same sense, that a malady is said to be natural; as arising from natural causes, tho’ it be contrary to health, the most agreeable and most natural situation of men. (T 1.4.4.1; SBN 225–6) The proper remedy to the Peripatetics’ appeal to substance as an unknown something—a specter in the dark—as a specious explanation is not to abandon the natural and legitimate practice of substantial explanation per se, but rather is to ensure that this is done properly. It is perfectly legitimate to explain an articulate voice in the dark by supposing that voice to belong to a person. What is illegitimate is to seek explanations where none are needed and offer as an explanation a mere unknown something. Hume summarizes his critique of the Peripatetics as follows. But these philosophers carry their fictions still farther in their sentiments concerning occult qualities, and both suppose a substance supporting, which they do not understand, and an accident supported, of which they as imperfect an idea. The whole system, therefore, is entirely incomprehensible, and yet is deriv’d from principles as natural as any of those above-explain’d. (T 1.4.3.8; SBN 222; emphasis added) The problem with the peripatetic philosophy is its reliance on spurious explanations, but as the final clause here again makes clear, this is not a condemnation of explanation in general—the impulse here is a natural and perfectly legitimate one—but only of explanatory posits that are ultimately occult, not understood, imperfect and incomprehensible. The proscription on such specious explanations, however, does not prevent us from thinking of the substance gold as itself being the common causal explanation of each of these properties, so long as we are able properly to give content to this supposition. It is no accident, for example that in taking gold to be a substance with a certain atomic structure, we employ the Rutherford-Bohr model of the atom, which we saw takes the solar system as its perceptible model, and specifies determinate ways in which an atom both resembles and differs from this.24 Thus, my suggestion is that in thinking of certain qualities, but not others, as being substantial qualities, we think of those qualities as being the result of some underlying nature of the substance at hand.25 In fact, Hume proposes just such a criterion in presenting what it is that makes us take a particular object to be identical across time (which, as Hume notes, will

72  David Landy itself depend on thinking of such an object as being of a certain kind of substance). Whenever we discover such a perfect resemblance [of the former and present qualities of the object], we consider, whether it be common in that species of object; whether possibly or probably any cause cou’d operate in producing the change and resemblance; and according as we determine concerning these causes and effects, we form our judgment concerning the identity of the object. (T 1.3.2.2; SBN 74) We think of an object as being identical across time just in case we think of it is as being of a certain “species of object” in which “any cause cou’d operate in producing the change and resemblance.” That is, we think of an object as being identical across time just in case we think that there is some univocal causal explanation for how objects of that kind come to have the qualities that they do. Supposing that representing an object as being of a certain species is equivalent to representing it as being a certain kind of substance, we can conclude that to represent something as a substance is to represent there being some univocal causal explanation for its having the qualities that it does. To return to the contrast between substances and modes, a modal property is one that is not supposed to relate to the underlying nature of the substance at hand. Hume’s own examples of modes are of a dance and beauty. The simple ideas of which modes are formed, either represent qualities, which are not united by contiguity and causation, but are dispers’d in different subjects; or if they be all united together, the uniting principle is not regarded as the foundation of the complex idea. The idea of a dance is an instance of the first kind of modes; that of beauty of the second. The reason is obvious, why such complex ideas cannot receive any new idea, without changing the name, which distinguishes the mode. (T 1.1.6.3; SBN 17) A dance is not a substance because it is “dispers’d in different subjects.” Of course, this presupposes that we have a working notion of the identity of different subjects, and that is precisely the work that an account of substances provides. Plausibly, though, the causal explanation of a dance would be given in terms of the conglomerate causal explanations of the individual dancers. That is, “a dance” is most plausibly construed as a nominalization of the verb “dancing,” which applies most readily to persons, which in this context would be the subjects, or substances, at hand, the substantial qualities of which we take to be explained by some uniting principle. By contrast, beauty is not a substance because while it is “united together, the uniting principle is not regarded as a foundation of the complex idea.”

Sellars and Hume  73 That is, while it is a single object that is beautiful—again presupposing that we have a notion of the identity of the substance at hand—that object is not the object that it is in virtue of being beautiful. For example, a sculpture is considered a substance in virtue of the causal structure of that object, whereas its being beautiful is not attributable to this same structure (at least not per se). In accounting for the causal structure of the sculpture, we would make reference to the properties of the material of which it is composed: its color, hardness, malleability, rigidity, etc. We would expect these qualities to have a common causal explanation that would make reference to the nature of this substance—e.g., the properties of clay or stone—but we would not expect the beauty of the statue to require a similar explanation. Finally, unlike a substance, a mode “cannot receive any new idea” because a quality that is a mode is not “closely and inseparably connected” to any other qualities. By contrast to a substance, the general term that is annexed to the qualities that comprise a mode do not have a uniting principle that provide their foundation. They are simply those qualities that are called by a certain name. Should we change what qualities these are, we change the meaning of term, and thus refer to a different mode. What this implies, though, is that the opposite is true of substances. The general term that represents a substance does not merely represent a list of qualities, but represents these qualities as being connected via some uniting principle or foundation. That is, these qualities belong to a certain substance, which is why we can discover that further qualities also belong to that substance, or that some qualities that we took to be substantial are not so. We are now in a position to confirm some of our earlier suggestions about how Hume conceives the science of human nature. We observe that certain mental particulars obey certain empirical generalities. This fact demands an explanation. That explanation takes the form of a theoretical posit, the representation of which requires a perceptible model, which resembles and differs from that which is posited in determinate ways.26 Such a theoretical posit performs its explanatory role by casting the substance as having as qualities, power or nature such that it obeys the empirical regularities that it does. All of which is to say that the explanations employed by the science of human nature are essentially ontological explanations. It is because a gas is a collection of atoms obeying the known laws of motion that it obeys the Ideal Gas Law. It is because gold is an atom with a certain atomic structure that is yellow, malleable and dissolvable in aqua regia. It is because our complex perceptions are composed of simple ones that we can explain the variety of human thought given the paucity of experience, etc. What all of this amounts to, in turn, is a picture of Hume’s understanding of the aims and methods of scientific explanation that stands in direct contrast to his appropriation by the Logical Positivists. Contra DN, the explicanda of a scientific explanation are the observed regularities of experience, and their explanans is the nature of the substance underlying such phenomena. Such substance is not represented as a mere via negativa, but

74  David Landy rather specifying determinate similarities and differences between such a theoretical posit and some perceptible model. That is all to say that at least insofar as these admittedly brief considerations have shown, Hume is more of a Sellarsian scientific realist than he is a Logical Positivist.

Notes 1 Sellars’s most explicit and prominent critique of Kant is throughout his engagement with him in Science and Metaphysics. In Chapter I, Sellars offers what he takes to be a more perspicuous version of Kant’s account of intuition and inner and outer sense; in Chapter V he endorses a form of transcendental realism that he suggests was unavailable to Kant. In both of these cases, I believe that Kant has more resources to accommodate Sellars’s objections than Sellars anticipates. See Landy, Kant’s Inferentialism. As for Sellars’s endorsements of various Kantian theses, these are too numerous to list. 2 Sellars does say that he considers Aristotle the philosopher par excellence of the manifest image, but of course Sellars is a scientific realist who holds that the descriptive content of the manifest image eventually must be replaced with that of the scientific image. That replacement itself, in turn, calls for a new philosophical understanding of our place in the world, and as Sellars sees it, that understanding will look mostly Kantian. 3 Kant, Prolegomena, 4, 260; 10. 4 Recent scholarship on Hume’s account of representation has downplayed, to various degrees, his imagism. See, for example, Garrett, “Hume’s Naturalistic Theory of Representation,” Schafer, “Hume’s Unified Theory of Mental Representation,” Landy, “Recent Scholarship on Hume’s Theory of Mental Representation,” and Cottrell, “Copying, Representation, Sympathy: A Puzzle in Hume.” 5 For example, Sellars follows Kant in holding that in order to represent a manifold as a manifold, “this manifold first [must] be gone through, taken up, and combined in a certain way in order for a cognition to be made out of it” (A77/ B102–3). 6 Sellars, EPM, XVI. 7 Cf. Weinberg, An Examination of Logical Positivism, 3. 8 E.g., De Pierris, Ideas, Evidence, and Method, and Qu, “Prescription, Description, and Hume’s Experimental Method,”. 9 Sellars, “Language of Theories,” 113–14. 10 I discuss Hume’s comments on this attempted explanation in Landy, Hume’s Science of Human Nature, Chapter 2. 11 Sellars, EPM, §142. 12 E.g., in Into.8; SBN xvii Hume urges that the only way to form any notion of the “powers and qualities” of the human mind is via “careful and exact experiments, and the observation of those particular effects, which result from its different circumstances.” For citations from Hume’s A Treatise of Human Nature I employ the standard convention of citing the book, chapter, section, and paragraph number from the Clarendon edition, followed by the page number from the Selby-Bigge/Nidditch edition. 13 Sellars, “Language of Theories,” 113. 14 Sellars, EPM, §41 15 To be clear, despite sharing a commitment to these four theses, Hume and Sellars differ significantly in many other respects, including but not limited to how to understand the implications of their shared commitments. Cf. Landy, “Sellars on Hume and Kant on Representing Complexes,” and Landy, “A (Sellarsian) Kantian Critique of Hume’s Theory of Concepts.”

Sellars and Hume  75 16 At present there is not a particularly large body of secondary literature on this brief but important section. I believe that the best, most thorough and most current extant reading of it is Baxter, “Hume on Substance,” and so, for interested readers, I will present an ongoing commentary highlighting my agreements and disagreements with Baxter here in the endnotes. While Baxter does advance his own particular line on Hume’s conception of ontology in that piece, much of what he says about this section in particular is, I think, representative of what might be called the standard interpretive line, such as there is one. 17 Thus, Baxter is mostly right to point out that, “When Hume denies that we have an idea of substance, he is not denying that we have an idea of a collection of qualities. . . . Thus it must be the idea of pure substance he is questioning”(Baxter, “Hume on Substance,” 46). Hume does not deny that we have any idea of substance, but only that we do not have an idea of substance like the one described by those who hold that a substance is a that in which observable qualities inhere. Still I hesitate to endorse Baxter’s conclusion in its entirety because of what Baxter means by “pure substance”: “the principle of unity and identity for an individual in which its accidents such as ‘Colour or Weight’ inhere. It is a ‘Substratum’, a ‘pure Substance in general” (Baxter, “Hume on Substance,” 45). Notice that Baxter’s definition of pure substance encompasses a number of theses that it is at least possible to separate. One is that pure substance is a substratum in which accidents inhere. Another distinct thesis is that pure substance is the principle of unity and identity for an individual. These two aspects can be combined, but need not be. Farther along, I will defend the thesis that Hume, in fact, rejects the former conception of substance, but accepts the latter. That is, substance is not that in which accidents inhere, but it does serve as a principle of unity and identity for an individual. Specifically, the idea of substance is the idea of that which explains the regular co-occurrence of certain observable qualities. For example, we (but not Hume) might hold that the atomic structure of gold explains its manifest qualities such as its yellow color, malleability, etc. Notice that while in that case gold atoms would provide a principle of unity and identity for an individual, the relation between these manifest qualities and gold atoms is not one of inherence. (Also worth noting is that because we represent gold atoms via modeling them on some antecedently represented object, they are more than a mere via negativa, represented solely by contrast with their observable qualities.) 18 As Hume puts it, “every thing in nature is individual” (T 1.1.7.6; SBN 19). Recall that it is this theory of general representation that Sellars critiques in EPM on the grounds that it seeks to explain merely determinable representation by presupposing that determinate representation is itself unproblematic. 19 Cf. Landy, “Recent Scholarship on Hume’s Theory of Mental Representation.” 20 Lest the reader think that Hume has dodged Sellars’s criticism of his account of general representation by equating representing as with general representation, it should be noted that (as previously) for Hume particular spatial arrangements of mental ideas represent particular spatial arrangements of impressions as such without the deployment of general terms, as do temporal complexes temporal complexes, etc. So, Sellars’s criticism that Hume’s theory of general representation explains merely determinable representation by presupposing that determinate representation is unproblematic is prima facie on target. 21 Thus, Baxter cannot be right that, “For Hume the idea of a purported individual substance is really just a collection of ideas of various qualities. For instance the idea of a peach is just the ideas of its colour, taste, figure, solidity, etc. And for Hume there is no real unity or existence to such collections other than the individual existences of their members” (Baxter, “Hume on Substance, 50). If the

76  David Landy idea of a purported individual substance were nothing more than just a collection of ideas of various qualities, then Hume would lose the distinction between substances and modes that he spends almost the entirety of this section attempting to recover! As we have seen, Hume in fact explicitly writes that the idea of substance is a collection of ideas of various qualities with a uniting principle as their foundation. Whether or not one accepts the account of that foundation uniting principle that I present above, it should be clear that some account of it needs to be given, and that that account must respect Hume’s claim that the idea of a substance differs from that of a mode precisely insofar as the latter, but not the former is a mere collection of various qualities. The contrast between Baxter’s claim that “there is no real unity” amongst the collection of qualities that compose the idea of a substance and Hume’s claim that such collections have a “uniting principle as their foundation” could hardly be more striking. 22 The availability of such an explanation is what leads to one of Sellars’s repeated touchstones: Eddington’s two tables. “It would seem, then, that if kinetic theory is a good theory, we are entitled to say that molecules exist. This confronts us with a classical puzzle. For, it would seem, we can also say that if our observation framework is a good one, we are entitled to say that horses, chairs, tables, etc., exist. Shall we then say that both tables and molecules exist? if we do, we are immediately faced with the problem as to how theoretical objects and observational objects ‘fit together in one universe.’ To use Eddington’s well-worn example, instead of the one table with which pretheoretical discourse was content, we seem forced to recognize two tables of radically different kinds. Do they both really exist? Are they, perhaps, really the same table? If only one of them really exists, which?” (LT §33). As I have attempted to show elsewhere (Landy, Hume’s Science of Human Nature, Chapter 3), Hume’s solution to this puzzle is structurally similar to Sellars’s. When one successfully explains some manifest phenomena by appeal to the nature of the substance underlying it, the explanatory force of that posit warrants a change in one’s ontology. Just as Sellars holds that we ought (in some sense) to replace the language of tables and chairs with the language of a molecules and atoms, so too does Hume hold that the language of the scientist of human nature ought to replace that of the vulgar. (For the sense in which Sellars denies that we ought to affect this change immediately, see Scientific Realism and Irenic Instrumentalism.) 23 As Baxter points out, Hume, “finds that we can have no impression from which such an idea could be copied. . . . Hume later renews this objection, making clear the Berkeleyan presupposition that ideas represent by resembling. Since an impression is nothing like a pure substance, it cannot represent a pure substance” (Baxter, “Hume on Substance,” 46). Impressions and ideas cannot represent that which they in no way resemble. Of course, that leaves open the possibility that impressions and ideas can represent that which they resemble in only some way. That is, it leaves open the possibility that unobservable substances can be represented via the use of a model by specifying determinate ways that the posited substance both differs from and resembles the model. Of course, this procedure is precisely what I want to suggest Hume endorses. 24 The notion of models plays an important role throughout Sellars’s philosophical writings, although he does not, as far as I know, give a detailed or thorough account of their use. (Nor does Hume.) The important points of Sellars’s understanding of models appears to be encapsulated in the passage quoted earlier: “the fundamental assumptions of a theory are usually developed not by constructing uninterpreted calculi which might correlate in the desired manner with observational discourse, but rather by attempting to find a model, i.e., to describe a domain of familiar objects behaving in familiar ways such that we can

Sellars and Hume  77 see how the phenomena to be explained would arise if they consisted of this sort of thing. The essential thing about a model is that it is accompanied, so to speak, by a commentary which qualifies or limits . . . the analogy between the familiar objects and the entities which are being introduced by the theory” (EPM §142). As Sellars sees it, it is the instrumentalist, the philosopher who holds that the manifest image is sacrosanct and that all theory is in principle otiose, that understands theoretical posits as being nothing more than an uninterpreted calculus that is correlated with observable phenomenon. On that line, since our ultimate ontology will always be limited to what is “directly observable,” theoreticalexplanatory activity cannot amount to anything more than idle speculation. By contrast, as the scientific realist sees it, the explanatory force of a theoretical posit can warrant a change in ontological commitment, but can earn that force only if it has some representational content that is distinct from that of our representation of observable phenomena. The former comes by that content by being modeled on the latter. For example, atoms are like the solar system in such-and-such ways, but unlike it in these other-and-other ways. Thus, atoms are represented via, as Sellars puts it, an analogical extension of our ordinary concepts. The content of the representation ‘atom’ is different from the content of the concept ‘solar system’, but also not constructed entirely ex nihilo. 25 Again I find myself in agreement with Baxter, with certain caveats. Hume continues the line of thought about perceptions and qualities to conclude that they do not inhere in anything: “Inhesion in something is suppos’d to be requisite to support the existence of our perceptions. Nothing appears requisite to support the existence of a perception. We have, therefore, no idea of inhesion” (THN 1.4.5.6). It follows as well that there is no such thing as inhesion (Baxter, “Hume on Substance,” 57). I agree entirely with Baxter that Hume holds that there is no such thing as inhesion. All perceptions are themselves distinct existences, and therefore substances, and therefore do not inhere in anything else. (I argue in Landy, “A Puzzle about Hume’s Theory of General Representation,” and Landy, Hume’s Science of Human Nature, that simple ideas themselves are theoretical posits.) Since we cannot have an idea of anything other than a perception, it also follows that we can have no idea of inhesion. That said, even this strong thesis leaves room for a conception of substance as that which explains the regular co-occurrence of manifest qualities. The crucial move to make will be to hold that this substance is not “requisite to support the existence of our perceptions,” and I argue in Hume’s Science of Human Nature, Chapter 5 that this is a move that Hume can and does make. What an explanatory appeal to a posited substance does, as Hume sees it, is render the regular co-occurrence of certain qualities expected, whereas that would have previously been surprising and puzzling. It is not that we take these qualities to inhere in that substance, but rather that we take them to be the observable manifestation of the nature of that substance. Malleability does not inhere in the ‘pure substance’ gold atoms. Rather, malleability just is the ability of gold atoms to be repositioned without breaking their metallic bonds. 26 Again, Sellars: “Theories about observable things . . . explain empirical laws by explaining why observable things obey to the extent that they do, these empirical laws; that is, they explain why individual objects of various kinds and in various circumstances in the observation framework behave in those ways in which it has been inductively established that they do behave. Roughly, it is because a gas is—in some sense of ‘is’—a cloud of molecules which are behaving in certain theoretically defined ways, that it obeys the empirical Boyle-Charles law” (LT §41). Putting aside the differences between Sellars and Hume on the status of the representation of determinate mental particulars, at this point is appears

78  David Landy that they would agree to at least the following. We observe particulars behaving in certain regular ways. (For Hume, this will primarily be the co-occurrence of certain phenomenal qualities.) We inductively establish certain empirical generalizations to describe these. That these particulars obey these empirical generalizations demands explanation, which explanation takes the form of a theoretical posit that casts the regularities as the manifestation of the nature of a posited underlying substance.

References Baxter, Donald L.M. 2015. ‘Hume on Substance: A Critique of Locke’. In Locke and Leibniz on Substance. Edited by Paul Lodge and Tom Stoneham, 45–62. New York: Routledge. Cottrell, Jonathan. Unpublished Manuscript. ‘Copying, Representation, Sympathy: A Puzzle in Hume’. De Pierris, Graciela. 2015. Ideas, Evidence, and Method: Hume’s Skepticism and Naturalism Concerning Knowledge and Causation. Oxford: Oxford University Press. Garrett, Don. 2006. ‘Hume’s Naturalistic Theory of Representation’. Synthese, 152, 3: 301–19. Hume, David (T). 1974. A Treatise of Human Nature. Edited by L.A. Selby-Bigge. New York: Oxford University Press. Hume, David (SBN). 2000. A Treatise of Human Nature. Edited by David Fate Norton and Mary J. Norton. New York: Oxford University Press. Landy, David. 2007. ‘A (Sellarsian) Kantian Critique of Hume’s Theory of Concepts’. Pacific Philosophical Quarterly, 88: 445–57. Landy, David. 2009. ‘Sellars on Hume and Kant on Representing Complexes’. European Journal of Philosophy, 17: 224–46. Landy, David. 2015. Kant’s Inferentialism: The Case Against Hume. New York: Routledge. Landy, David. 2016. ‘A Puzzle About Hume’s Theory of General Representation’. Journal of the History of Philosophy, 54: 257–82. Landy, David. 2017. Hume’s Science of Human Nature. New York: Routledge. Landy, David. Forthcoming. ‘Recent Scholarship on Hume’s Theory of Mental Representation’. European Journal of Philosophy. Qu, Hsueh. 2016. ‘Prescription, Description, and Hume’s Experimental Method’. British Journal of the History of Philosophy, 24: 279–301. Schafer, Karl. 2015. ‘Hume’s Unified Theory of Mental Representation’. European Journal of Philosophy, 23, 4: 978–1005. Sellars, Wilfrid (EPM). 1956. ‘Empiricism and the Philosophy of Mind’. In Science, Perception and Reality. London: Routledge & Kegan Paul; Reissued by Atascadero, CA: Ridgeview Publishing Company. Sellars, Wilfrid. 1961. ‘The Language of Theories’. In Science, Perception and Reality. London: Routledge & Kegan Paul; Reissued by Atascadero, CA: Ridgeview Publishing Company. Sellars, Wilfrid (SM). 1968. Science and Metaphysics: Variations on Kantian Themes, The John Locke Lectures for 1965–66. London: Routledge & Kegan Paul; reissued by Atascadero, CA: Ridgeview Publishing Company 1992. Weinberg, Julius. 1936. An Examination of Logical Positivism. London: Routledge & Kegan Paul and Trench, Trubner & Co., Ltd.

4 Sellars’s Interpretive Variations on Kant’s Transcendental Idealist Themes James R. O’Shea

Wilfrid Sellars’s 1966 John Locke Lectures, Science and Metaphysics: Variations on Kantian Themes (1968 ‘SM’) were delivered the same year as the publication of P. F. Strawson’s groundbreaking book, The Bounds of Sense: An Essay on Kant’s Critique of Pure Reason (1966). Both books are philosophically sophisticated and ambitious, both in the interpretation of Kant and in relation to expressing the authors’ own philosophical views. As we shall see, however, Sellars’s book presents some exceptional interpretive challenges for the reader, and I hope to provide some clarifying guidance to the terrain. In this chapter I will examine some of the main contours of Sellars’s interpretation of Kant’s theoretical philosophy, focusing on this occasion on its most controversial aspect: his complex engagements with Kant’s transcendental idealist distinction between objects as ‘appearances’ and as ‘things in themselves’. Whereas elsewhere I have primarily been concerned to isolate those aspects of Sellars’s analytic Kantian or (as I call it) ‘Kantian naturalist’ theory of experience that I find most defensible, in what follows I want to explore Sellars’s more full-blooded treatments of Kant’s transcendental idealism in particular, both for its own sake and as an interpretation of Kant. In The Bounds of Sense Strawson distinguished centrally between ‘two faces of the Critique’ (1966, part one). One face for Strawson is represented by Kant’s deeply insightful ‘Metaphysics of Experience,’ which turned out, not surprisingly, to bear strong affinities to the ‘descriptive metaphysics’ of our experience defended in Strawson’s own classic work, Individuals (1959) (cf. 1996, part two). The other face, as Strawson sees it, is the ‘incoherent’ and ‘disastrous model’ of Kant’s ‘transcendental psychology’ or ‘transcendental subjectivism’: the ‘theory of the mind making Nature’ as a realm of mere ‘appearances’ in what Strawson derides as Kant’s ‘Metaphysics of Transcendental Idealism’ (1966, 16–23, and part four). The former, defensible a priori or ‘transcendental’ theory of experience “is concerned with the conceptual structure which is presupposed in all empirical inquiries” (1966, 18), involving for both Kant and Strawson—and for Sellars, too— the conception of one unified consciousness of one directly perceivable and objective spatiotemporal world of lawfully interacting and persisting

80  James R. O’Shea material substances. Sellars outlined his own interpretations and updated defenses of Kant’s theory of experience not only in Science and Metaphysics, which tended to focus more heavily on Kant’s transcendental idealism, but in particular in such articles as ‘Some Remarks on Kant’s Theory of Experience’ (1967 KTE), “this I or he or it (the thing) which thinks“ (1970 ‘I’), ‘Kant’s Transcendental Idealism’ (1976 KTI), and ‘The Role of Imagination in Kant’s Theory of Experience’ (1978 IKTE), all of which were collected posthumously in Kant’s Transcendental Metaphysics (KTM) in 2002.1 I and many of the other authors referred to in this chapter (for example, Brandom, Landy, McDowell, Rosenberg, Sicha and Westphal) have laid out and defended key aspects of Sellars’s views on Kant’s analysis or theory of experience, in relation to the thinking self, conceptual content, causal laws, perceptual knowledge, the refutation of (empirical) idealism, normativity and so on. My present purpose, however, is to take a more careful look at Sellars’s interpretation of Kant’s transcendental idealism as found primarily in that deep but enigmatic book, Science and Metaphysics.

1. Sellars’s Manifest Image ‘Appearances’ and Scientific Image ‘Things in Themselves’ Whereas Strawson rejected as entirely incoherent Kant’s distinction between objects (and persons) as appearances to us in nature as opposed to as nonspatiotemporal things in themselves, Sellars takes Kant’s transcendental idealism to be a near-miss that can be successfully reformulated in terms of Sellars’s own famous distinction between the manifest (MI) and scientific (SI) images of the human-being-in-the-world: 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) Having explored Sellars’s scientific realism and his distinction between the ‘two images’ elsewhere (2007, 2016), I will present the key points succinctly here without full textual justification. The “world of everyday experience” or manifest image (MI) for Sellars is defined in terms of (i) the nature of persons as concept-using thinkers and intentional agents,2 and (ii) the nature of the ‘strictly’ perceptible3 physical objects that are basic in our common-sense conceptual framework (Sellars PSIM, passim). Regarding (ii), let us follow Sellars and use visually perceived expanses of color—for example, the redness of an apple’s surface—as our main

Sellars’s Interpretive Variations  81 example of the sorts of ordinary sensible qualities that characterize the everyday objects of our experience. I will not present or evaluate his arguments on this particular matter, but it is well known that Sellars argued throughout his works (e.g., PSIM) that, phenomenologically considered, MI physical objects possess ‘homogeneous’ color-contents as their constituent content-characters in a way that he contends is ostensibly incompatible with the ‘postulational’ scientific image (SI) conception of physical objects as exhaustively composed of colorless particles or fields as conceived by ongoing modern scientific theoretical explanations. Through a long journey of highly sophisticated categorical-ontological and phenomenological analyses (cf. Rosenberg 2007, Ch. 9, and deVries 2005, Ch. 8), Sellars argued that the expanses of color that populate the MI-world of common-sense physical objects must eventually be reconceived (ontologically ‘relocated’, as it were) as the contents solely of the sensory states of the perceivers’ central nervous systems when they are sensing-redly. That is, such a perceiver is having a sensation of red in the manner that is typically caused either by the visible presence of (what in MI is responded to as) a ‘red’ physical object, or, in the case of hallucinations, non-standard lighting, etc., by some other stimulation of the visual cortex that is responsible for the experience. Furthermore, Sellars argued that since our MI-experiences of colored physical objects are of bounded regions or expanses of color—put crudely, since Berkeley was right that perceived color and perceived shape (extension) go seamlessly together in our experience—the sensory states of perceivers must somehow preserve (that is, coherently be conceived as the ‘true home’ of) both color and shape, as for example when we have a sensation in the ‘of-a-red-rectangle’ manner. So finally, since it is not coherent to suppose that such sensory states of perceivers are literally colored and shaped in the way that MI-physical objects are conceived to be, the sensory states or ‘sense impressions’ of the perceiver must be reconceived (and will be, in a future, more adequate SI-neurophysiological theory) to have corresponding intrinsic qualities and geometrical relationships that are analogous to and systematically represent the qualities and relations that we conceive (within MI) as their standard physical causes. “Succinctly put, impressions have attributes and stand in relations which are counterparts of the attributes and relations of physical objects and events” (SM I, §65). In the end, crudely put, Sellars’s view is that our common-sense MIconception of colored physical objects reflects the fact that our sensory systems evolved to get us around our environment and to avoid poisonous berries, not to reveal to us directly the ultimately correct categorial ontology of the natural world. This is one main source of Sellars’s famous (or infamous) scientia mensura view, that is, his Feyerabendian thesis (SRII V, §48) that, strictly speaking, the ordinary perceptible objects of the manifest image do not really exist, i.e., as so conceived within MI—they are ‘mere appearances’ or objective ‘representables’ in the Kantian sense of being ‘actual’ (cf. SM II, KTE, IKTE, KTI). They really exist per se, i.e., as the things they

82  James R. O’Shea are ‘in themselves’, only as radically ontologically reconceived in the ongoing theoretical or postulational scientific image of the human being and its world. Within the ‘ought-to-be’ rules that implicitly govern the language of our ordinary MI conceptual framework, it is of course true to assert that bananas are yellow and apples are red (i.e., true with respect to MI, as opposed to what is ‘really’ or ‘ultimately’ true: cf. SM V). But, speaking as a philosopher, I am quite prepared to say that the common-sense world of physical objects in Space and Time is unreal— that there are no such things. Or, to put it less paradoxically, that in the dimension of describing and explaining the world, science is the measure of all things, of what is that it is and of what is not that it is not. (EPM IX, §41) If we now look back to the passage from Science and Metaphysics quoted earlier, we have seen one main reason why Sellars believed “that a consistent scientific realist must hold that the world of everyday experience is a phenomenal world in the Kantian sense” (SM VI, §61). It is important to recognize, however, that the previous argument from the nature of MI-color and other sensible qualities is not the only sort of consideration that Sellars mobilizes in order to support his contention that our ordinary MI-conception of objects is merely ‘phenomenal’ in the Kantian sense. As Sellars argues in SM lectures V-VI as well as in PSIM and all of his other writings pertaining to scientific realism, there are many other dimensions in which the ongoing scientific image of the world, with its postulation of processes, fields and particles that are strictly speaking not perceivable in principle in terms of the MI-ontology of colored physical objects, provides demonstrably more adequate explanations of the nature of MI-phenomena than the MI-ontology can provide of itself. This applies not only to the nature of perceived color and other sensible qualities, no matter how carefully they are phenomenologically analyzed within the ‘life-world’ of MI; it applies also to the explanation of how it is that we ourselves became concept-users and rational agents in the first place. Furthermore, Sellars argues that even our ongoing scientific conceptions of the world, too, are continually being shown to be ‘mere appearances’ that are better explained by being reconceived in more adequate scientific theories of the world. For example, Einstein’s relativity theory reconceived the very nature of space, time, mass, velocity and so on, so as to explain precisely why and how Newton’s theory and its accompanying object-ontology both succeeded and failed to the extent that it did, thereby revealing Newton’s world to be one of ‘mere appearances in the Kantian sense’ relative to its grounding in the real Einsteinian nature of ‘things in themselves’, which, as Sellars puts in the passage above on Kant (SM VI, §61), grounds the ‘obtainability’ of those regularities that are captured (and those that are not captured) by Newton’s laws. And so on.

Sellars’s Interpretive Variations  83 With this initial background understanding of Sellars’s wider philosophical views in place, we can roughly put it that he offered two main reasons for regarding the MI-world of common-sense perceptible objects to be ‘transcendentally ideal’ phenomenal appearances in Kant’s sense: First, (i) there is the (allegedly) required ontological ‘relocation’ or explanatory reconception of perceived color-shape-expanses and other MI-constitutive proper and common sensible qualities ‘from’ MI-objects ‘to’ states of MIperceivers (this is putting it very crudely), and then further SI-reconceiving such sensory states in a future, radically revised SI-neurophysiology that will preserve, analogically, the initial consciously experienced color-shape qualitative characters and structure; and second, (ii) Sellars appeals to the demonstrably more adequate explanations of the nature of reality that have already occurred, and will continue to occur, not only in relation to (i) but in the ongoing scientific theory-replacements themselves. Crucially, an explanatory burden placed on SI-explanations in both cases is that their resulting reconceived ontologies, which reveal how things are ‘in themselves’, explain why things approximately lawfully appeared as they did, both in our ordinary MI perceptual world, and in the progress of superseded SI-ontologies such as Newton’s that are successfully corrected and incorporated within more adequate SI-theories such as Einstein’s. With the rough outlines of Sellars’s own big picture in place,4 we can now take up some central and perplexing themes that arise particularly in Chapters 1 and 2 of Sellars’s Science and Metaphysics. Certain features of Sellars’s interpretation of Kant’s transcendental idealism should be clear from what we have seen already, however. Whereas Graham Bird had already published his book on Kant’s Theory of Knowledge in 1962 (and cf. Bird 2006, Ch. 23), arguing that Kant’s first Critique was not in fact committed to the real existence of ‘things in themselves’ (though we are of course, for Kant, committed to the indispensable but problematic idea or empty thought of such a grounding); and Henry Allison would later develop his highly influential ‘two standpoint’ interpretation of Kant’s transcendental idealism, which would similarly render problematic any straightforward ontological interpretations of the real significance for Kant of our admittedly indispensable idea of ‘things in themselves’; both Strawson and Sellars, by contrast, interpreted Kant’s non-spatiotemporal things in themselves in a straightforwardly ontological manner, as Kant’s view on what really exists. Strawson reached the more usual conclusion at the time of finding Kant’s transcendental idealism, so interpreted, to be hopelessly problematic. On his reading, furthermore, Kant’s own conception of the ‘appearances’ was infected as a result with the sort of implausibly subjectivist phenomenalism (unlike the better ‘face’ or analysis of objectivity and the unity of experience that Strawson also found in Kant) that Bird had already diagnosed as the typical absurd consequence of traditional strongly ontological conceptions of Kant’s idea of things in themselves. Sellars’s interpretation was unusual

84  James R. O’Shea among English-language interpreters of Kant at the time in defending a more plausible non-phenomenalist, intentionalist interpretation of Kant’s own empirical realism (that is, in contrast to the more usual senses of ‘phenomenalism’ typified in some respects by Bennett (1966) and revitalized in more recent times by Van Cleve 1999),5 but nonetheless conjoined (pace Bird and Allison) with a strongly ontological reading of Kant’s own things in themselves.6 An important example of Sellars’s attempt to merge something like both of these ‘faces’ of Kant together is provided by the topic of the next section, which focuses on Chapter 1 of Science and Metaphysics, ‘Sensibility and Understanding’.7

2. Nonconceptual Sensibility, Conceptual Understanding and Singular Intuitions For Sellars, it was a crucial insight of Kant’s to recognize “the need for a sharp distinction between sensibility and understanding” (SM I, §3), and thus to break the classical empiricist and rationalist sensory-cognitive continuum that Kant had correctly diagnosed as characterizing both Locke’s sensualizing of the understanding and Leibniz’s intellectualizing of the appearances (cf. A271/B327).8 Central to Sellars’s own philosophy, correspondingly, was a complex but firm distinction between the (intensional, but not intentional) nonconceptual sensory representation ‘of’ objects and the (intentional, and so intensional) conceptual representation ‘of’ objects, with both dimensions of ‘aboutness’ being integrated in adult human sense perceptual knowledge (cf. SM I, §§56–9; EPM V, §§24–5). On Sellars’s view, which he finds ‘implicit in’ and regards as philosophically required by Kant’s account, the two different sorts of states, conceptual and nonconceptual, possess different types of object-representational content, both of which are integrated in human perception. The first chapter of Science and Metaphysics partly defends and partly criticizes Kant’s version of this distinction by seeking to expose and clarify its ambiguities when presented in terms of Sellars’s own distinctions. The result presents challenges for the reader along multiple dimensions that I hope to clarify. I will begin first with the topic of nonconceptual sensory representations, which picks up on a central theme from section 1 above. Having a sense impression or nonconceptual representation of a red rectangle is to sense in the manner normally and reliably caused by red rectangles (for example, by facing sides of bricks), which explains the logical non-extensionality that is involved in our (and other animals’) being able to have such an impression ‘of’ or sensorily represent a red rectangle when there is in fact no red rectangular physical object in the environment (SM I, §§56–9). For Sellars, here knowingly going well beyond anything in Kant, it was important that such ‘mediating’ nonconceptual sensory states of the perceiver are not merely characterized in topic-neutral functional terms (‘a state of the kind normally caused by ’ etc., as for example in J.J.C. Smart’s

Sellars’s Interpretive Variations  85 physicalism). Rather, the theorist must introduce ‘new predicates’ that provide positive intrinsic characterizations of such conscious, qualitative sensory states (cf. SM I, §55; EPM IV and XVI; PSIM VI). Such contents are conceived by analogy with the colors, shapes and other attributes and relations of their (normally) corresponding physical causes, which such sense impressions thereby nonconceptually track and represent (or ‘picture’: cf. SM V). Sellars also defends the further controversial thesis, which I will not explore here, that such qualitative sensings, though they are indeed broadly ‘physical’ in the sense of being fully located and causally operative within the natural world (partly constituting central nervous systems), they cannot according to Sellars be ‘physical’ as that term has most often been (he thinks, over-restrictively) understood by physicalists.9 In key respects outlined earlier I believe that Sellars ought to be classified in contemporary philosophy of perception and in relation to the recent Kant literature (cf. Schulting 2016) as a defender of nonconceptual (objectrepresentational) content, though one who sought, perhaps in some ways similar to Gareth Evans (1982) subsequently, to embed such nonconceptual sensory-informational ‘tracking’ content within, and as having ‘a strong voice in’ (SM I, §39) the integrated outcome of an otherwise strongly conceptualist understanding of adult human perceptual knowledge. For understandable reasons, however, Sellars has to the contrary almost universally been classified in these debates through the lens of John McDowell’s subsequent conceptualism: the position, crudely put, that the representational content of adult human perception is exhaustively conceptual content. Robert Hanna, for example, writes of Sellars’s (alleged) “rejection of non-conceptual content” due to his (and later, McDowell’s) rejection of the Myth of the Given. Hanna thus takes it as obvious that “Davidson and Sellars are both clearly thoroughgoing conceptualists avant la lettre” (Hanna 2006, 82, 89). While we cannot enter here into the complex details of the contemporary debates about nonconceptual content, both in general and in Kant (Schulting 2016 is a good place to start),10 Brady Bowman in the following passage provides a useful snapshot of the typical roles of some of the main players in the current debates concerning Kant and nonconceptualism (my interpolations added): Like the more recent figurehead Gareth Evans [1982], Kant figures ambiguously in the debate between conceptualists and non-conceptualists, being claimed as an illustrious predecessor by both sides. For example, some of McDowell’s more recent [2009] discussions of the content of sense experience are framed as interpretations of Kant that elicit from him a more consistent and nuanced conceptualism even than that propounded by Sellars, thus vindicating [Kant] against Sellars’s criticisms [in SM I]. Apparently with equal plausibility, Hanna [e.g., 2006] interprets Kant as a powerful theoretician of non-conceptual content, drawing on Kant’s doctrine of the specific difference between intuitions and

86  James R. O’Shea concepts. While Hanna acknowledges the famous passages that seem to place Kant unambiguously in the conceptualist camp [e.g., ‘intuitions without concepts are blind’ (A51/B75)], he also draws attention to remarks in the Critique of Pure Reason that seem equally unambiguously to state that objects can appear to us without having any relation to the functions of the understanding, i.e. to concepts [e.g., ‘appearances would nonetheless offer objects to our intuition, for intuition by no means requires the functions of thinking’ (A89–90/B122–3)]. (Bowman 2011, 419) Let us examine the aspects of Sellars’s view as expressed in SM Chapter 1 that make it plausible to classify him as a conceptualist, which I contend he sought to integrate with his substantive conception of nonconceptual, distinctively sensory-representational content. Sellars’s view was that Kant nearly succeeded in expressing a view having this shape (i.e., the shape of Sellars’s own position), but that he crucially failed to clarify the relevant distinctions required. Sellars throughout his works stresses Kant’s view that, for beings like us, the possibility of our representing any empirically mind-independent object of experience at all—that is, in such a way that we are capable of representing the object as located and persisting independently of our representing it—requires the sorts of conceptual syntheses in judgment that Kant articulates throughout the transcendental analytic. In particular, both Sellars and subsequently McDowell interpret Kant’s claim in such a way that it extends the requirement of conceptualization to Kant’s account of the singular and immediate sensible intuition of any objects that may appear to us. Sellars focuses, for instance, on this well-known passage from Kant’s ‘metaphysical deduction’ or ‘clue’ to the transcendental deduction (cf. Sellars SM I, §10, and McDowell 2009, Ch. 14: ‘Avoiding the Myth of the Given’): The same function that gives unity to the different representations in a judgment also gives unity to the mere synthesis of different representations in an intuition, which expressed generally, is called the pure concept of understanding. (A79/B104–5) Sellars explains (in particular in KTI) that it is the ‘same function’ of conceptual synthesis that governs what Kant later in the transcendental deduction calls the ‘productive imagination,’ which is involved constitutively in all our perceptual cognitions. Hence our ‘immediate’ sensible intuitions of objects in perception are, in this primary sense of ‘intuition’ on Sellars’s reading of Kant, a distinctive kind of singular, indexical (this-here-now), conceptualized thought of the given empirical object that directly evokes this noninferential response in the perceiver (e.g., this red cube). Such a singular intuitive perceptual ‘taking-to-be’ involves the productive imagination’s

Sellars’s Interpretive Variations  87 “intimate blend of sensing and imaging and conceptualization” (IKTE II, §23). This aspect of Sellars’s reading of Kant on sensible intuition is plausibly regarded as ‘conceptualist’ in at least the sense that it takes our most basic and immediate sensory awarenesses of objects (as the intuited subjectterms available for judgment) to be constituted by a certain kind of conceptual representational content. In this way such sensible intuitions or immediate perceptual takings are thus shown already to involve the categories in our cognition of “whatever objects may come before our senses” (B159), and this is the linchpin of Kant’s transcendental deduction. What Sellars further argues in SM I, however (and similarly in the ‘Appendix’ to SM as far as the form of time rather than space is concerned), as we have already seen, is that “our conceptual representations of the spatial structure of physical states of affairs are guided by [analogous] ‘counterpart’ features of our [nonconceptual] sense impressions” (SM Appendix, §3, my interpolations added). And as we have seen, the same account holds for such intrinsic qualitative sensory contents as the perceived red and rectangular surface of a brick: our conceptualized perceptual responses in such cases are ‘guided by’ the counterpart quasi-red, quasi-rectangular nonconceptual contents of the corresponding sense impressions. That the ‘guiding’ nonconceptual sensory states that have a ‘strong voice in the outcome’ are genuinely intensional representational contents, for Sellars, has to do not only with the postulated theoretical (for Sellars’s Kant, ‘transcendental’) hypothesis that “impressions have attributes and stand in relations which are counterparts of the attributes and relations of physical objects and events” (SM I, §65). It also has to do with the further fact that these substantive sense impression/ outer object isomorphisms normally ‘track’ objects and thus ‘guide us’ reliably in our actions and in our judgments. On Sellars’s own wider views (cf. Sellars MEV 1981), these animal and logical representational or isomorphic ‘picturing’ achievements are thanks both to our bodily nature (for Sellars, our proper biological functioning due ultimately to natural selection) and our conceptual capacities (reflecting a logical space of reasons derived from culturally inherited norms or ‘ought-to-bes’ of assertional practice). He sees Kant as essentially having put forward a nativist faculty version, involving Kant’s pure a priori forms of sensibility (‘Space and Time’) and understanding (the categories), of an outlook on cognition that otherwise is very close to Sellars’s own integrated conceptualist and nonconceptualist picture. Sellars characterizes the nonconceptual aspect of Kant’s views on sensible intuition as a “sheer receptivity” (e.g., SM I §§9–10, 17–19, 73–8), in contrast to Kant’s fully developed conception of intuitions as singular conceptualized thought-responses to objects that are guided by and incorporate such nonconceptual contents. However, it is important to bear in mind that such sensible contents are ‘sheer’ only in the sense of “being in no sense conceptual” (SM I §16). They are not ‘sheer’ in the sense of lacking those reliable patterns of specific counterpart relational structures and intrinsic characters that make them

88  James R. O’Shea the nonconceptual representations of objects of corresponding kinds and structures that they are. The question is: How much of Sellars’s own robust conception of such nonconceptual sensory representations, as I have sketched it in this chapter, can one actually find in Kant? How much goes beyond anything in Kant? And how much does Sellars himself recognize goes beyond Kant’s own views? In relation to ‘sheer receptivity’ Sellars refers us to Kant’s characterizations (e.g., A99ff.) of sensory receptivity as providing us with a “manifold of representations, but not with a representation of a manifold, which latter [Kant] proceeds to equate with representation of a manifold as a manifold” (SM I, §19), i.e., with intuition in the conceptualized sense. But here Sellars argues that Kant, unlike Sellars’s own view of the structured counterpart sensory relations that are isomorphic to the outer spatial relations of the physical objects they (nonconceptually) represent, was unfortunately committed to the following dead-end claim: [Kant is committed to the claim] that what the representations of sheer receptivity are of is in no sense complex, and hence that the representations of outer sense as such are not representations of spatial complexes. If I am right, the idea that Space is the form of outer sense is incoherent. Space can scarcely be the form of the representings of outer sense [JOS: hence the need for Sellars’ own theory of analogous counterpart relations for sense-manifolds]; and if it is not the form of its representeds, i.e. if nothing represented by outer sense as such is a spatial complex, the idea that Space is the form of outer sense threatens to disappear. (SM I, §19) Contemporary interpreters who have defended the view that Kant is a nonconceptualist (again, cf. Schulting 2016) might well step in with other arguments to bolster Sellars’s contention that nonetheless “Kant ‘implicitly’ gives some such account” as Sellars has outlined of the nonconceptual intuitional representation of space (and of time, and of the sensible ‘particulars’ that occupy them), or that Kant “must have done so” (SM I, §78). But the actual ‘cash’ that Sellars provides for finding such a story in Kant is minimal by his own admission, and his main conclusion is that Kant’s failure to make the distinction clearly (in particular by “overlooking the importance of analogical concepts—save in theological contexts”) had the unfortunate effect of sending subsequent philosophers on a nineteenth century reprise of the sensory-cognitive continuum of Locke and Leibniz: Kant’s failure to distinguish clearly between the ‘forms’ of receptivity proper and the ‘forms’ of that which is represented by the intuitive conceptual representations which are ‘guided’ by receptivity—a distinction

Sellars’s Interpretive Variations  89 which is demanded both by the thrust of his argument, and by sound philosophy—had as its consequence that no sooner had he left the scene than these particular waters were muddied by Hegel and the Mills, and philosophy had to begin the slow climb ‘back to Kant’ which is still underway. (SM I, §75)

3. Conclusion: Sellars’s Variations on Kant’s Appearances and Things in Themselves Sellars’s strongly ontological reading of Kant’s things in themselves has stressed Kant’s use of analogical thinking ‘in theological contexts’, no doubt thinking of Kant’s various ways of spelling out reason’s experiencetranscending ideas of the ‘unconditioned’ (God, freedom, a community of souls) as ‘ground of the appearances’ in nature, in both morally practical and regulative-theoretical terms. Occasionally Sellars, as we have seen, and his student Jay Rosenberg after him, suggest that Kant also held something like Sellars’s own view of the analogical structure of ‘things in themselves’ and possibly even its Sellarsian reflection in the theoretically postulated analogical structure of our sensings ‘in themselves’ (i.e., Sellars’s analogically spatial and analogically temporal sensory σ-manifolds and τ-manifolds: e.g., SM Appendix §§17–23). Thus, both of them refer to Kant’s reference to a ‘duration’ that is “not a time” (B149, Sellars SM II.17, Rosenberg 2005, 78–9; Sellars there also cites Kant’s reference at B798 to the concept of “a presence that is not spatial” as at least “non-contradictory”). Sellars, however, is very cautious in this regard, since he also holds that “Kant’s treatment of sensation is notoriously inadequate and inept” (KTE I, §3n; cf. SM II, §39n). And in fact, in his own lectures on Kant at Pittsburgh (reproduced in Sellars 2002 KPT 131), Sellars in fact makes clear that “Kant himself never says” that the manifold of sense is “quasi-spatial, somehow spatial,” that is, analogously spatial in Sellars’ postulated sense, in contrast to (but on Sellars’ view undergirding) the form of space that we conceptually represent via the transcendental imagination. Or again, at SM II, §72 Sellars remarks that it would still have been open to Kant to say that things-in-themselves, in so far as they affect our sensibility, have, like sense impressions, attributes and relations which are in their own way analogous to those of perceptible things, and by virtue of which they elicit sense impressions which are in their different way endowed with Space-like characteristics (confused by Kant with the form of outer intuition), and perform the guiding role described above. That Kant implicitly accepted some such view of things-in-themselves is, I think, clear. Yet if the fact had been brought to his attention he would most certainly have claimed that this

90  James R. O’Shea transcendental use of analogy is empty. The abstract concept of such Space-like characteristics could have ‘cash value’ only for God. (SM II, §72) I think, pace Sellars, that it was not ‘clear’ that Kant ‘implicitly accepted some such view of things-in-themselves’, indeed partly for the reason of emptiness that Sellars here cites. Or again, Sellars remarks that Kant nowhere denies, and is not committed to denying, that the manifold of external sense as such is a relational structure. Indeed, the more general point can be made that Kant nowhere denies or need deny that the initself has a relational structure. (SM I, §76) In these passages and elsewhere I find the cited evidence to be extremely thin that Kant actually held or implicitely believed anything like the (nontheologically) analogical conception of either things in themselves or of our sense impressions that Sellars and Rosenberg suggest Kant either held, or implicitly held, or at least did not deny and had it ‘open’ to him to hold. Sellars himself clearly wants to find this more committal ‘analogical’ view to be ‘implicit’ in Kant’s views on sensibility, but it is in fact, I conclude, nowhere to be found in Kant. Rosenberg, who along with Sellars has taught me more truths worth preserving from Kant’s theory of experience than any other Kant commentator, asserts these doubtful claims concerning Kant’s things in themselves more strongly than Sellars did, asserting straightforwardly that “Kant’s story of spatial experience” (2005, 80) in effect includes all of the analogical elements conjectured by Sellars to have implicitly informed Kant’s views on sensibility and things in themselves. Sellars concludes his remark above that Kant ‘nowhere denies’ that things in themselves or our sense impressions have structures analogous to those we perceive in physical objects, by stating that what Kant “does deny, whether for good reasons or for bad, a topic for subsequent discussion, is that the relations we conceptually represent are the relations which the in-itself exemplifies” (SM I, §76). Sellars in fact goes on to argue in SM Chapter II, ‘Appearances and Things in Themselves: 1. Material Things’, that Kant’s own reasons for denying that “the relations we conceptually represent are the relations which the in-itself exemplifies” were indeed unsound arguments. However, Sellars also indicates that there exists a sound argument “for the transcendental ideality of the [MI] perceptual world,” one which “lies in the distinction between perceptible physical objects and the objects of theoretical science, a distinction which was blurred by Kant” (SM II, §71n; cf. Landy 2011, ‘Postscript’; O’Shea 2016, Ch. 7). The subsequent discussion (SM II, §§72–4) shows that Sellars here has in mind, of course, his own arguments pertaining to the ontological relocation and theoretical reconception of color-expanses that we rehearsed briefly in section 1 earlier.

Sellars’s Interpretive Variations  91 This, Sellars concludes, is an “argument for the transcendental ideality of perceptible things which really works, and is the one I shall espouse” (SM II, §69). My own view is that Sellars’s general suggestion in Science and Metaphysics that Kant may have sought, as did Sellars himself, to integrate the conception of a certain kind of nonconceptual sensory representation of objects within an otherwise strongly conceptualist outlook on our perceptual cognition, represented an important anticipation of later ‘nonconceptualist’ views both in general and in the interpretation of Kant. Furthermore, I have argued elsewhere that one can retain Sellars’s ‘Kantian naturalist’ hypothesis of a middle way on that and other topics without accepting Sellars’s (which is not Kant’s) radical ‘transcendental idealist’ thesis that the manifest image or common-sense world of colored objects “is radically false, i.e., there really are no such things as the physical objects and processes of the common-sense framework” (though of course Sellars emphasizes that there are scientifically reconceived counterparts to those objects) (SRII V, §23). Sellars’s arguments specifically based on color and other sensible qualities rest on premises and qualia intuitions that have increasingly been shown to be highly contentious. Furthermore, while I am also sympathetic to Sellars’s outlook on conceptual change in science and to his anti-instrumentalist scientific realism in general, the idea that the paradigms of scientific progress provided by the revolutionary shifts from Newton to Einstein, or from the Charles-Boyle phenomenal gas law to the kinetic theory of gases, should lead us to conclude that all scientific progress (including in all the ‘special sciences’) will or must be characterized by this ontological replacement model, strikes me as just the sort of matter that Sellars’s Peircean fallibilism ought to have left open. So with respect to both the problem of color and the fact of scientific revolutions, I resist Sellars’s philosophical contention that either of them entails the ‘strictly speaking’ falsity of higher-level ontologies, including our common-sense conception of persisting colored objects. I conclude that Sellars’s and Rosenberg’s attempts to preserve the core truths in Kant’s theory of experience, and to integrate them with an overall scientific naturalist outlook on the nature of things—including the nonconceptual representational dimension of perceptual experience examined in this chapter—can and should survive the rejection of some of their more doubtful interpretations and adaptations of Kant’s ‘transcendental idealism’ in particular. Perhaps the perplexities arising from Sellars’s interpretive replacement of Kant’s transcendental idealism by his own manifest image/ scientific image distinction should also lead us to reassess wholesale, once again, the strongly ontological reading of Kant’s own views on our largely empty thought of ‘things in themselves’ as the ‘ground of appearances.’ It may well be that Kant’s own ‘empirical realism,’ which included within it, for Kant whatever lawfully successful scientific postulations might happen to fall outside the particular limits or thresholds of our sense organs

92  James R. O’Shea (A225–6/B272–3),11 is realism enough for any plausible scientific realism. It is certainly realism enough, I believe, to support the sort of Kantian naturalist outlook on human beings and the world that Sellars sought to envision.12

Notes 1 Thanks to the work of Jeffrey Sicha, KTM and Sellars’s other publications continue to be available at very affordable prices at www.ridgeviewpublishing.com, and Sicha has also been at work (together with Pedro Amaral among others) making Sellars’s publications available in electronic editions. Also important for understanding Sellars’s views on Kant are his 1975–76 lectures to students at the University of Pittsburgh, Sellars’s KPT. For book-length interpretations of Kant’s Critique of Pure Reason strongly influenced by Sellars, see especially Rosenberg (2005); and influenced by both Sellars and Rosenberg on Kant are Powell (1990), O’Shea (2012) and Landy (2015) (all students of Rosenberg). See also Sicha’s substantial introduction to Sellars’s KTM. Also deeply influenced by Sellars and Kant, and central to current debates (including on Sellars’s reading of Kant) are Brandom (2015) and McDowell (2009). For more on Sellars as a pivotal figure in the history of Kant-influenced analytic philosophy, see Westphal (2010). 2 There is an important sense in which for Sellars, as for Kant, persons as thinking selves and intending agents are in crucial, normative-functional respects not ‘objects’ in such a way as to render them candidates for ontological elimination and replacement by more adequate scientific successor conceptions of their nature. This is compatible with there also being many other aspects of ourselves as embodied persons that are subject to such explanatory reconceptions in the ongoing development of the scientific image. The distinction between persons and objects is in fact the key to Sellars’s synoptic fusion or integration of the MI and SI ontologies: Crudely, the MI-objects get replaced by SI-processes, Kantianthinking agents remain through the integration, albeit with enormously significant reconceptions of their nature. In particular our understanding of qualitative sensory consciousness (e.g., color consciousness) will have been radically transformed, but along many other dimensions as well (cf. Christias 2016, all of whose recent writings on Sellars I highly recommend). See O’Shea (2007 passim) and (2016, Chapter 7) for the textual support for these and other characterizations of Sellars’s overall views appealed to in this chapter. 3 It is seldom appreciated that Sellars’s distinction between MI and SI rests definitively, as far as the ‘objects’ of those images or conceptual constructs are concerned, on a twofold understanding of the nature of ‘perception’ (e.g., SRII V, §55–8; O’Shea 2007, 33–8): crudely put, (1) a ‘strict’ or ‘manifest’ sense in which what we directly perceive are the Aristotelian ‘proper and common sensible qualities’ (e.g., color and shape) possessed by ordinary physical objects in the constitutive way discussed briefly in the main text here (call these ‘manifest perceptible’ physical objects); as opposed to (2) an all-inclusive ‘pragmatic’ conception of perceptual observation or detection developed by Feyerabend and Sellars, and inherited by Rorty and Brandom, as any reliable, noninferential response to objects (any ‘language entry transition’, in Sellars’s terms). Call the latter, i.e., (2), any ‘reliably observed’ objects or processes. This pragmatic conception covers all noninferential perception for Sellars, whether it be the ‘strict’ perception of MI-objects, or the observational detection of SI-processes (for example, ‘there goes an electron’ uttered by a trained physicist as a reliable noninferential response to a streak in a cloud chamber). The objects of MI are defined or stipulated by Sellars to be those that are strictly manifest perceptible, whereas the SI is defined in terms of those evolving theories of modern

Sellars’s Interpretive Variations  93 science that incorporate the ‘postulation’ of new basic objects, processes, fields, etc., that are manifestly imperceptible. The latter (manifestly) ‘unobservable’ processes not only may, but Sellars argues must become directly ‘reliably observable’ in sense (2), e.g., in the manner of the electrons example given above (but successfully ‘gone global’, as it were), if scientific realism is to be fully vindicated (cf. SM V, §90–1; SRII V, §§55–8, VII §§72–4, VIII §§89–91. Sellars’s conception of scientific realism was in this last respect unusual by standard post-1970 Putnam-Kripke lights (for further discussion, cf. O’Shea 2007, Chs. 2, 6, 7). 4 In O’Shea (2016, Ch. 7, ‘What to Take Away from Sellars’s Kantian Naturalism’), I have indicated how I would reject certain aspects of the picture above, retaining both Sellars’s outlook on Kant’s theory of experience and central components of his scientific naturalism/scientific realism, but rejecting his contention that problems pertaining to sensible qualities and the nature of scientific progress must entail the falsity ‘strictly speaking’ of our common-sense ontology of colored, persisting substantial kinds. The latter contention I regard in fallibilist terms as a hypothesis that, in light of debates both in the philosophy of perception and concerning the status of the ‘special sciences’ over the last four decades, is neither inconceivable nor mandatory, but unlikely to provide the best explanation of how all things ‘hang together’ (PSIM I, §1). See Landy (2015, including his concluding ‘Postscript to Transcendental Idealism’), for a recent fruitful investigation of many of the same issues in Kant and Sellars that I am exploring in this chapter. I will not attempt to sort out the similarities and differences between our two accounts on this occasion, but I am happy to note that there are substantial aspects of similarity. 5 In O’Shea (2015) I support the claim that Kant and Sellars both defended nonphenomenalist conceptions of empirical realism, causal lawfulness and objectivity, despite the perennial temptation to read Kant in such phenomenalist terms. When Sellars writes of ‘Kant’s phenomenalism’ he is not interpreting Kant’s ‘phenomena’ or appearances in a way that is inconsistent with Sellars’s own well-known critiques of phenomenalism, which he regarded as based on an incorrect, non-Kantian analysis of the conceptual structure of the (Kantian) manifest image. The issue is a delicate one, since Sellars’s characterizations of Kant’s appearances “as the contents of actual and obtainable conceptual representings” (SM VI, §61, quoted earlier) obviously bear a surface resemblance to typical phenomenalist ‘if . . . then’ analyses of our experience. In the article cited I clarify the distinction between Kant’s view and the ‘phenomenalist temptation’ with reference to Kant, Strawson, C. I. Lewis, Sellars and Brandom. 6 The last two decades have witnessed a resurgence of strongly ontological interpretations of Kant’s things in themselves (including non-phenomenalist readings), particularly viewed as developing out of, but preserving in key metaphysical respects, Kant’s pre-critical Leibnizian heritage. Rae Langton (1998) was a particularly influential book along those lines. 7 For a probing discussion of Sellars’s SM Chapter 1, see McDowell (2009, Chapters 1 and 2). I have examined McDowell’s conceptualist outlook in that book in detail elsewhere. 8 Page references to Kant 1787 are to the standard marginal A/B page references, to the original first and second editions respectively. 9 Sellars notoriously held the view that such qualitatively contentful sensings or sense impressions (of a red cube, for example), while they are ‘physical’ in the all-inclusive sense of being natural spatio-temporal-causal items (Sellars’s ‘physical1’), they must in another, narrower sense (‘physical2’) be non-physical2 states of consciousness, in that the physical2 primitives defined as those adequate to explain inorganic, insensate life, are in principle (Sellars contended) inadequate to explain the intrinsic contents of qualitative sensory consciousness. Sellars’s philosophical hypothesis is thus that a future physics of neurophysiological

94  James R. O’Shea processes will need to introduce new (non-physical2) primitives in order to explain sensory consciousness. This further analogical hypothesis is operative in SM I, but Sellars is well aware that it goes beyond anything explicitly in Kant. 10 I am not convinced, for instance, that the recently much discussed distinction between ‘state’ vs. ‘content’ nonconceptualism makes it more plausible to deny that Sellars is a nonconceptualist in any interesting sense (I take it that Hanna holds this with regard to Sellars), but I cannot engage that particular issue here. 11 See also Kant’s 1790 reply to Eberhard (On a discovery . . .), Akademie edition, 8:210, Kant’s Theoretical Philosophy after 1781, p. 302. 12 I would especially like to thank Professor Antonio Nunziante and Luca Corti of the Università degli Studi di Padova, who organized the wonderful conference in Padua that gave birth to this paper and to this volume, and for whose patience and guidance throughout I am very grateful.

References Allison, H. 2004. Kant’s Transcendental Idealism: An Interpretation and Defense, revised 2nd edition (1st edition, 1983). New Haven, CT: Yale University Press. Bennett, Jonathan. 1966. Kant’s Analytic. Cambridge: Cambridge University Press. Bird, Graham. 1962. Kant’s Theory of Knowledge. London: Routledge & Kegan Paul. Bird, Graham. 2006. The Revolutionary Kant: A Commentary on the Critique of Pure Reason. Chicago and LaSalle, IL: Open Court. Bowman, Brady. 2011. ‘A Conceptualist Reply to Hanna’s Kantian Non— Conceptualism’. International Journal of Philosophical Studies, 19: 417–46. Brandom, Robert B. 2015. From Empiricism to Expressivism: Brandom Reads Sellars. Cambridge, MA and London: Harvard University Press. Christias, Dionysis. 2016. ‘Can “Ready-to-Hand” Normativity be Reconciled with the Scientific Image?’ Philosophia, 44: 447–67. deVries, Willem A. 2005. Wilfrid Sellars. Chesham, Bucks: Acumen Publishing and Montreal & Kingston: McGill-Queen’s University Press. Evans, Gareth. 1982. The Varieties of Reference. Edited by John McDowell. Oxford: Oxford University Press. Hanna, Robert. 2006. Kant, Science, and Human Nature. Oxford: Clarendon Press. Kant, Immanuel. 1787. Critique of Pure Reason. Translated by Paul Guyer and Allen Wood. Cambridge: Cambridge University Press 1997. Kant, Immanuel. 1790. ‘On a Discovery Whereby Any New Critique of Pure Reason Is to Be Made Superfluous by an Older One’. In Theoretical Philosophy After 1781. Edited by H. Allison and P. Heath, 271–336. Cambridge: Cambridge University Press. Landy, David. 2015. Kant’s Inferentialism: The Case Against Hume. London and New York: Routledge. Langton, Rae. 1998. Kantian Humility: Our Ignorance of Things in Themselves. Oxford: Clarendon Press. McDowell, John. 2009. Having the World in View: Essays on Kant, Hegel, and Sellars. Cambridge, MA and London: Harvard University Press. O’Shea, James R. 2007. Wilfrid Sellars: Naturalism with a Normative Turn. Cambridge: Wiley-Blackwell and Polity Press. O’Shea, James R. 2012. Kant’s Critique of Pure Reason: An Introduction and Interpretation. London: Routledge.

Sellars’s Interpretive Variations  95 O’Shea, James R. 2015. ‘Concepts of Objects as Prescribing Laws: A Kantian and Pragmatist Line of Thought’. In Pragmatism, Kant, and Transcendental Philosophy. Edited by Robert Stern and Gabriele Gava, 196–216. London: Routledge. O’Shea, James R. (Ed.). 2016. Sellars and his Legacy. Oxford and New York: Oxford University Press. Powell, C. Thomas. 1990. Kant’s Theory of Self-Consciousness. Oxford and New York: Oxford University Press. Rosenberg, Jay F. 2005. Accessing Kant: A Relaxed Introduction to the Critique of Pure Reason. Oxford: Clarendon Press. Rosenberg, Jay F. 2007. Wilfrid Sellars: Fusing the Images. Oxford: Oxford University Press. Schulting, Denis, ed. 2016. Kantian Nonconceptualism. London: Palgrave Macmillan. Sellars, Wilfrid (EPM). 1956. ‘Empiricism and the Philosophy of Mind’. In Minnesota Studies in the Philosophy of Science, Vol. I. Edited by H. Feigl and M. Scriven, 253–329. Minneapolis, MN: University of Minnesota Press). Reprinted in SPR. Sellars, Wilfrid (PSIM). 1962. ‘Philosophy and the Scientific Image of Man’. In Frontiers of Science and Philosophy. Edited by Robert Colodny, 35–78. Pittsburgh, PA: University of Pittsburgh Press. Reprinted in SPR and ISR. Sellars, Wilfrid (SPR). 1963. Science, Perception and Reality. London: Routledge & Kegan Paul. Reissued by Atascadero, CA: Ridgeview Publishing Company 1991. Sellars, Wilfrid (SRII). 1965. ‘Scientific Realism or Irenic Instrumentalism: A Critique of Nagel and Feyerabend on Theoretical Explanation’. In Boston Studies in the Philosophy of Science, Vol. II. Edited by Robert Cohen and Marx Wartofsky, 171–204. New York: Humanities Press. Reprinted in Sellars (PPME), Atascadero, CA: Ridgeview Publishing Company. Sellars, Wilfrid (‘I’). 1970. ‘. . . 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 ISR and KTM. Sellars, Wilfrid (KTE). 1967. ‘Some Remarks on Kant’s Theory of Experience’. Journal of Philosophy, 64: 633–47. Presented in a symposium on Kant at the 1967 meeting of the American Philosophical Association (Eastern Division). Reprinted in ISR and KTM. Sellars, Wilfrid (SM). 1968. SM Science and Metaphysics: Variations on Kantian Themes, The John Locke Lectures for 1965–66. London: Routledge & Kegan Paul. Reissued by Atascadero, CA: Ridgeview Publishing Company 1992. Sellars, Wilfrid (KTI). 1976. ‘Kant’s Transcendental Idealism’ (presented at an International Kant Congress at the University of Ottawa). Published in volume 6. In Collections of Philosophy, 165–81. Reprinted in KTM. Sellars, Wilfrid (IKTE). 1978. ‘The Role of Imagination in Kant’s Theory of Experience’. The Dotterer Lecture 1978. In Categories: A Colloquium. Edited by Henry W. Johnstone Jr., 231–45. University Park, PA: Pennsylvania State University Press. Reprinted in ISR and KTM. Sellars, Wilfrid (MEV). 1981. ‘Mental Events’. Philosophical Studies, 39: 325–45. Reprinted in ISR.

96  James R. O’Shea Sellars, Wilfrid (KPT). 2002. Kant and Pre-Kantian Themes: Lectures by Wilfrid Sellars. Edited by Pedro Amaral. Atascadero, CA: Ridgeview Publishing Company. Sellars, Wilfrid (KTM). 2002. Kant’s Transcendental Metaphysics: Sellars’ Cassirer Lectures Notes and Other Essays. Edited by Jeffrey F. Sicha. Atascadero, CA: Ridgeview Publishing Company. Sellars, Wilfrid (ISR). 2007. In the Space of Reasons: Selected Essays of Wilfrid Sellars. Edited by Kevin Scharp and Robert Brandom. Cambridge, MA: Harvard University Press. Strawson, P. F. 1959. Individuals: An Essay in Descriptive Metaphysics. London: Methuen. Strawson, 1966. The Bounds of Sense: An Essay on Kant’s Critique of Pure Reason. London: Methuen. Van Cleve, James. 1999. Problems from Kant. Oxford: Oxford University Press. Westphal, Kenneth R. 2010. ‘Kant’s Critique of Pure Reason and Analytic Philosophy’. In The Cambridge Companion to Kant’s Critique of Pure Reason. Edited by Paul Guyer, 401–30. Cambridge: Cambridge University Press.

5 Hegel After Sellars Conceptual Connections Luca Corti

For now that philosophy has gone ‘back to Kant’ for the second time, can a Hegelian ‘trip’ be far behind? W. Sellars, Notre Dame Lectures, 268

1. Introduction My aim in this article is to explore the role Sellars has played in shaping current discussions of Hegel. Although the claim that Sellars was a key contributor to today’s revival of Hegel is widely accepted, the question of precisely how this occurred has yet to be fully determined. I will present a brief, non-technical overview of the main topics in Sellars that have migrated into the current conversation on Hegel, mainly thanks to their mobilization in several influential interpretations of the German philosopher that rely on Sellarsian ideas. To approach the origins of Sellars’s influence, I will make use of his idea that the history of philosophy is a lingua franca. I will track how some Hegelian distinctions have been, to paraphrase him, ‘translated into Sellarsian terms’ and vice versa (SM, 31). I will point to the main historical stages in this process of convergence between Sellarsian and Hegelian vocabularies. This reconstruction will hopefully help to contextualize and elaborate some aspects of the current debate on Hegel that have now become part of the standard reading of Hegel’s work. As I will show, there are two core areas that have guided the Sellarsian revival of Hegel, namely (i) Sellars’s idea of conceptual role semantics, which defines meaning and conceptual activity as inherently normative, such that they cannot be fully grasped by any other kind of discourse, and (ii) his idea of a conceptually informed element in receptivity. Before addressing these views, I will offer some brief contextualizing remarks on the American philosophical reception of Hegel in the years before Sellarsian Hegelianism emerged.

2.  Hegel in America Before Sellars The relation between Hegel and American philosophy in the last century was notably an uneasy one that saw Hegel appearing and disappearing from the

98  Luca Corti American main stage several times. As Bruce Kuklick noted in his study on the historical formation of the American philosophical canon, “Hegel was very much part of the canon at the end of the nineteenth century” (Kuklick 1984, 133). During those years, in “American philosophical circles there were more Hegelians of various sorts than one could shake a stick at” (ibid., 132).1 Within a few decades however, Hegel’s reputation in the US declined drastically, eventually leading to his complete disappearance from the canon. According to Kuklick, “although he [Hegel] may have been knocked about previously, he was killed in World War I” by a mixture of philosophical arguments and widespread aversions against all sorts of German influence (ibid.). In fact, the years after World War I, and then World War II notably marked the nadir of Hegel’s American reputation. The year 1945 in particular is the annus horribilis. This is the year in which Karl Popper published his Open Society and its Enemies, where he notably accused Hegel of being the father of what he called the “poisonous intellectual disease of our own time which I call oracular philosophy” (Popper 1945, II, 8). In the same year Bertrand Russell published his History of Western Philosophy, in which he attacked Hegel with arguments close to Popper’s and closed with the famous sarcastic remark that Hegel “illustrates an important truth, namely, that the worse your logic, the more interesting the consequences to which it gives rise” (Russell 1945, 99). The anti-Hegelian tendencies already present in both traditions considered ‘distinctive’ of American thought—the pragmatist tradition, on the one hand, and analytic philosophy, still in its early phases, on the other— crystallized, and, to state it in the words of Henry S. Harris, “sealed in amber, so to speak, the image of Hegel which would become canonical for the next years” (Harris 1983, 77).2 This image was not a positive one and heralded the disappearance of Hegel from the American philosophical mainstream.3 After a long absence, in the last decades Hegel seems to have come back on stage—so much so that in 2009 Frederick Beiser observed in his introduction to the Cambridge Companion to Hegel and Nineteenth-Century Philosophy a “remarkable growth of interest in Hegel” within the Anglophone world large enough to be called a “Hegel renaissance” (Beiser 2008, 1). This seemingly unexpected rehabilitation of Hegel was in fact the result of a longer process that involved several different phenomena, which cannot all be tracked here.4 One of the main paths leading to this renaissance, however, was certainly forged by Sellars. The strong effect of Sellars’s ideas on Hegel scholarship has nonetheless manifested itself only quite recently. The 1983 article entitled ‘The Hegel-Renaissance in the Anglo-Saxon World’ by Harris, for instance, was one of the first to announce a resurgence of Hegelism, but pointed to a different, somewhat more confined, phenomenon (Harris 1983). Harris not only did not mention Sellars but also predicted that the imminent rise of an Anglo-Saxon Hegelian school would revive Hegel’s Philosophy of Nature in light of “the dominant aspect

Hegel After Sellars  99 of English speaking thought, the tradition of scientific empiricism” (ibid., 94). Seen from today’s perspective, Harris’s ‘prophecy’ appears quite misguided. Indeed, very different aspects of Hegel’s thought—especially related to Hegel’s notion of Geist—than those Harris identified seem to have become important and Sellars’s influence has posthumously become visible as vital to this process.

3.  Sellarsian Ideas and their Hegelian Upshots Initially beyond the radar of mainstream of American Philosophy, Sellars in the 1940s began developing a system of thought that would eventually also affect the study of Hegel. At first glance, Sellars’s appreciation of Hegel seems anything but positive. Sellars mentions Hegel rarely in his work and mostly in a negative way. In EPM, for instance, Sellars criticizes the image of a “Hegelian serpent” of knowledge (EPM, 57); in LRB he attributes to Hegel an attempt to attain a priori knowledge of laws of nature (LRB, 146) and in SM accuses him of having “mudded the philosophical waters” of the relationship between understanding and sensibility (SM, 16, 29). Only in his Notre Dame Lectures does Sellars say anything positive, acknowledging that Hegel saw something important about indexicals—but Sellars then immediately adds that Hegel “hadn’t quite known how to cope with it” (WSNDL, 265). So Sellars appears to have seen Hegel primarily as a philosophical option to be avoided.5 The path that connects Sellars to contemporary Hegel scholarship is thus not a direct one; the link rather emerges via different circuitous branches. Two views are crucial to Sellars’s philosophical program and have been particularly influential, in the long run, for the Hegel renaissance: (i) one is related to Sellars’s renowned normative conception of meaning and conceptual thought (cf., §§4–7), (ii) the other to Sellars’s equally famous views on the contents of perceptual experience (cf., §8). In both cases, Sellars appears to have developed his view in constant dialogue with the tradition, in particular with Kant. The encounter between his ideas and Kant produced a convergence of idioms and problems that paved the way for further connection to Hegel. While Sellars did not provide the step to Hegel, he established a conceptual space that allowed others to do so.

4.  Conceptual Role Semantics and the Notion of ‘Rule’ The first relevant issue is also one of the core views in Sellars’s system, namely the so-called ‘normative functionalism’.6 Sellars put forth a non-relational account of meaning that has been highly influential and, to some extent, pioneering for subsequent forms of Conceptual Role Semantics (Dennett 1987, 341). His approach conceives of semantic significance in terms of a logically autonomous, “self-sustaining, holistic system of rule-governed,

100  Luca Corti contextually dependent, normative types embodied in natural token” (deVries 2009, 235, cf. LRB, LTC, SRLG, MFC). Put crudely, according to Sellars, linguistic episodes—as episodes in the conceptual order—are semantically significant in so far as they are assertable according to certain rules of language. In his model, there are three basic kinds of moves that define the structure of a language game: Language Entry Transition, Intra-Linguistic Moves, and Language Departure Transitions (MFC, 424, NAO, 69, SRLG, 329, SM, 114).7 Intra-Linguistic Moves, in particular, are those that confer conceptual significance on a certain linguistic item and the rules governing such moves are preeminently inferential: “the meaning of a term lies in the materially and formally valid inferences it makes possible” (PPPW, 265). Leaving aside more finely grained distinctions within Sellars’s sophisticated picture of linguistic practice, what is relevant is that Sellars uses this conception of language as a model for analogically understanding conceptual episodes (cf. PSIM, 32, SK, 306 ff., SRTT, 105–7). Two things are important here. First, such a model sees conceptual content as essentially normative, or rather, as dependent on a set of rules whose normative significance is, according to Sellars, logically irreducible to natural scientific explanations.8 As has been shown, Sellars progressively builds normative elements into his definition of what a conceptual-linguistic ‘rule’ is (Olen 2016, Ch. 6). The logical autonomy of the normative space will become a core thesis in his philosophy, one which he will defend throughout his career and will become a central tenet of Sellarsian Hegelianism. Second, Sellars’s attempt to define conceptual elements in normative ‘inferential’ terms played a key role enabling the first translation of his conceptual role semantics into Hegelese. As has been shown elsewhere, at least from the late 1950s onward Sellars expresses his core views on language and concepts using a Wittgensteinian vocabulary with slogans such as “The linguistic meaning of a word is entirely determined by the rules of its use” (PPPW, 142, cf. also NAO, 78). Understood along these lines, the notion of ‘rule’ or ‘norm’—laden with a normative force—acquires a particular role, one which is central not only to Sellars’s own philosophical vocabulary, but which also brings it to converge with Wittgensteinian views and, finally, mediates his approach to Kant. While developing his program—which he eventually calls “transcendental linguistics” (KTM, 281), Sellars is inclined to approach Kant through such a view of conceptual activity.9 Sellars’s confrontation with Kant notably dates back to his years at Oxford and continued throughout most of his career, shaping various aspects of his thought (cf. Haag 2012; O’Shea 2006, 2018; Westphal 2010). As he made clear in his Autobiographical Reflections, Sellars recognized that especially his functional theory of concepts had developed out of confrontation with Kant.10 Kant, according to Sellars, was among the first thinkers to conceive concepts in a ‘normative functional’ way. To quote Sellars again, “Kant’s revolutionary move was to see the categories as concepts of functional roles in mental activity” (KTM, 346). This made Sellars particularly attentive to

Hegel After Sellars  101 the notion of ‘rule’ as it appeared in Kant’s theory of judgment and to Kant’s definition of the understanding as the faculty of rules (KTM, 261). The possibility of convergence between Kant’s theory of conceptual activity and the Sellarsian linguistic-normative picture of thinking sketched above—or, to put it in Sellars’s words, the possibility of developing a “linguistic version of Kant’s position” (KTM, 281)—began playing out in the very first phases of the Hegel renaissance.

5.  Hegel’s Logic as a Conceptual Role Semantics These Sellarsian views on language affected interpretations of Hegel as early as the late 1970s. Sellars himself did not apply his views to a reading of Hegel, but such an attempt was made by Terry Pinkard—who as far as I know was among the first to synthetize the two perspectives of Sellars and Hegel, thanks to his education in a context where Sellarsians and Hegelians were working together.11 Pinkard used the Sellarsian idea of conceptual role semantics to interpret Hegel’s Science of Logic. In 1979 Pinkard published two papers arguing that, in order to explain Hegel’s theory of conceptual thinking in the Logic, we can proceed “following Wilfrid Sellars” (LHL, 419). Pinkard uses some famous Sellarsian examples (like Texas Chess) and makes the first effort to ‘translate’—so to speak—conceptual role semantics into the lingua franca of Hegelese (or maybe, more accurately, to translate Hegel’s terminology into the lingua franca of Sellarsianism). Given a particular conceptual item in the Logic, Pinkard says, “to put it in Hegelian language, we can say that determinateness . . . [or] conceptual meaning . . . is constituted by its moves, its logic” (ibid.). Therefore, Pinkard concludes: “The Science of Logic would be then an elaborate construction in which the determinateness of concepts is constituted by their . . . inferential relations . . . vis à vis one another” (HIHL, 214). Such application of Sellarsian ideas to Hegel can be found also in other influential works of the period. Robert Pippin, for instance—although promoting a different take on Hegel—also relied on a Sellarsian account of conceptual activity to approach Hegel. In his famous book Hegel’s Idealism, Pippin addresses Hegel’s Science of Logic by mobilizing the basic insight of conceptual role semantics. Hegel, Pippin argues, treats “all concepts functionally.” This means that “according to Hegel . . . the concept’s determinacy (its own particularity or content) is primarily a function of the role it can and cannot play in judgments.” Concepts are contentful only in so far as “there are specific judgmental roles for that concept to play within a range of judgments . . . and a system of inferences grounding such possible judgments” (HI, 238). The connection between a Sellarsian normative conception of meaning and conceptual activity, on the one hand, and Kant’s and Hegel’s theory of concepts, on the other, was also facilitated and fostered by other tendencies in American philosophy, such as the debates on Wittgenstein, and in

102  Luca Corti particular, on the alleged presence of a ‘transcendental idealism’ both in his early and later thought.12 One of the issues at stake in this debate was the idea that a norm-governed, conceptual structure could be seen as some sort of transcendental condition for meaningful language and contentful thought. In that respect, the connection between Kantian and Hegelian vocabulary, on the one hand, and the Wittgensteinian-Sellarsian vocabulary of ‘norms’ and ‘like-mindedness’, on the other, started to seem less eccentric.13 In this sense the Wittgensteinian debates not only produced new attention for the issue of ‘normativity’ but also fostered interpretative connections and an interplay of idioms between Wittgenstein scholarship and Hegel scholarship that continues today. The notion of conceptual ‘norm’ or ‘rule’ offered a common ground for this convergence.

6.  Geist as a Normative Realm This tendency to systematically connect Sellarsian and Hegelian insights took force a few years later with the publication of various contributions, including three important books in 1994: Brandom’s Making It Explicit, Pinkard’s Hegel’s Phenomenology and McDowell’s Mind and World. Albeit in different ways, these three works made use of previous debates and interpretative moves not only to develop some Sellarsian insights on normativity and Kant, but also to address and revive Hegel’s theory of Geist. McDowell, Brandom and Pinkard not only adopted a typical left-Sellarsian stance, according to which the normative realm is logically autonomous and cannot be reduced to any descriptive or causal account. They also made use of specific Sellarsian tools to approach Hegel. Pinkard, for instance, employed the notion of ‘social space’ to read Hegel. The notion was coined by the Sellarsian philosopher Jay Rosenberg to describe a Sellarsian normative space made of formal and material inferences. In Pinkard’s words: A social space in Rosenberg’s sense consists of a set of material and formal inferences. Hegel shares something like this view, since he thinks of these social spaces as having a logic—that is, as having an inferential structure, which can be outlined by the dialectician. (HP, 346 fn. 12) For Hegel, then, being self-conscious within a particular shape of spirit would thus mean, according to Pinkard Assuming a position in a ‘social space’, that is, assuming a whole set of inferences that licenses the agent to move from one position in the space to the another. A ‘move’ in a ‘social space’ is an inference licensed by that space. (HP, 47)

Hegel After Sellars  103 Robert Brandom, on the other hand, famously puts forth a normative account of conceptual content that is inspired by Sellars. Beginning with Making It Explicit and then more consistently in his later writing, Brandom links his philosophical project to Kant and especially to Hegel. By following a Sellarsian insight, Brandom presents Kant as the father of the “normative turn” (POP, 1) that reconstructs mind, meaning and rationality as essentially normative.14 From this perspective, Hegel further develops Kant’s approach, bringing it to completion. According to Brandom, Hegel developed a semantic theory similar to Brandom’s own, which uses inference to delimit the conceptual.15 Conceptual content, according to Brandom, has to be understood in terms of material inference. The notion of ‘material inference,’ which is primitive in Brandom’s semantics and derived from Sellars, can be translated into Hegelese as ‘mediation.’ According to Hegel, to be conceptually contentful would mean to stand in relations of material inference (or ‘mediation’), which in turn arise from relations of material incompatibility, translated into Hegelese as “determinate negation” (cf. SPCR, 140). In short, for Brandom: “What Hegel means by conceptual (begrifflich) . . . is thus to stand in material relations of incompatibility and (so) consequence (inference)” (TMD, 181).16 In all of these cases, Hegel’s notion of Geist was approached through the Sellarsian notion of a normative realm articulated through conceptual norms—often conceived as inferential, which are not reducible to or graspable by some sort of causal stories. As Pippin, another proponent of such an approach, has stated, “Hegel treats the problem of conceptuality as in general the problem of normativity” (HPP, 236) and his notion of Geist “is his theory of normativity” (TPS, 97).17 To exemplify it in a recent statement: spirit is nothing but ways of actively holding each other to account by the demanding and giving of reasons for beliefs and actions in a social community . . . the most generic name for such a realm could simply be ‘the normative’. (HPP, 122, 128) The same is true for Brandom, who claims: I understand the geistig as the realm of conceptually articulated norms, of authority and responsibility, commitment and entitlement. Spirit as a whole is the recognitive community of all those who have such normative statuses, and all their normatively significant activities. (TMD, 227) Such a normative understanding of Geist proved to be quite successful. Indeed it was the notion of Geist—and not Hegel’s Philosophy of

104  Luca Corti Nature as Harris had predicted—that “has catapulted Hegel back into the world, especially Anglophone” (HPP, 235), driving the new attention to his philosophy. Furthermore, thanks to the conceptual mediation of Sellars’s ideas on normativity and conceptual content, crucial questions regarding normativity, such as, “Where do norms come from and why?,” can now be asked in Hegelian terms, as “Where does Geist come from and why?” This question has found a variety of answers in Hegelian thought.

7.  The Origins of Norms The contemporary interest in ‘normativity’ goes together with an emphasis on sociality and the social nature of norms. Sellars himself had stressed that, given that meaning and conceptuality are inherently normative, it is the “linguistic community as a self-perpetuating whole which is the minimum unit in terms of which conceptual activity can be understood” (LTC, 512). In PSIM Sellars also underlined “the essentially social character of conceptual thinking,” by saying that “there is no thinking apart from common standards of correctness and relevance, which relate what I do think to what anyone ought to think” (PSIM, 16–17). There Sellars also claimed that Hegel was among the first to recognize this. Thanks not only to Sellars, sociality became a crucial character in our understanding of Hegel’s notion of Geist. However, the notion of sociality also came to intersect questions about the origins of norms and has given rise to different form of Sellarsian Hegelism. Debate over the origins and status of norms has created one of the major divides among Sellarsian Hegelians. One group of interpreters makes use of the paradigm of recognition to account for the genesis of norms. Recognition, however, means different things to different interpreters: Brandom for instance uses the notion of deontic attributive statuses to account for the genesis of normativity. He identifies a process of mutual attribution of social statuses (in Brandom, commitments and entitlements) that paradigmatically unfolds between two individuals. As Vincent Descombes says, “Robinson Crusoe meets Friday”: this is the paradigm of an emerging sociality (Descombes 2014, 120). To put it in Brandom’s own words: Taking someone to be responsible or authoritative, attributing a normative deontic status to someone, is the attitude-kind that Hegel (picking up a term of Fichte’s) calls ‘recognition’ (Anerkennung). (AI, 70, cf. also MIE, 663 fn. 1, TMD, 216) Once two participants in a practice attain a mutuality of ‘deontic attribution’ (or, in Hegel’s term, mutual recognition) then normativity comes on stage. In other words, both participants come to treat and assess each other’s performances normatively. Interpreters like Brandom also attribute

Hegel After Sellars  105 such view to Hegel (cf. SDR; AI, Ch. 2). Critics of such a view, however, argue that mutuality between two individuals is not enough to ensure the objectivity of normative standards. ‘Correctness,’ within this model, would simply amount to being ‘taken to be correct’ by the other participant in the practice. This is, at best, a form of “normative positivism,” as Pippin labels it (BH, 393). Brandom appears to acknowledge the problem when he admits: I do want things both ways. I want to say that they [norms] are socially instituted and attitude-transcendent. (RAP, 176) By pointing to this weakness, other scholars reject this dyadic, I-You model of recognition, and instead turn to other Hegelian solutions to the question. According to Pippin, for instance, normativity does not arise out of a normless context but should rather be conceived as pertaining to an institutional level, or an ongoing historical process of institutional formation that is not reducible to single attribution by participants. There is no original normless situation, only an on-going, continuous historical process of initiation or socialization into a community’s normative practices, demanding allegiance in all sorts of practical, engaged and largely implicit ways and receiving it in an equally various number of practices of consent, affirmation, sustenance, in a variety of modalities of self-legislation and self-obligation. (BH, 391) The notion of ‘institution’ provides an attitude-transcendent but still social dimension for norms, enabling one to says that “there are reasons to do things that exist independently of any individual or group acknowledging that reason” (HPP, 74). Another group of interpreters rejects the entire idea of a social origin of normativity, and therefore opposes any variation on the slogan “every transcendental institution is a social constitution.” McDowell is a paradigmatic case: for him, the notion of social institution or communal self-legislation does not make the origins of normativity intelligible (HWV, 107). If self-legislation of rational norms is not to be a random leap in the dark, it must be seen as an acknowledgment of an authority that the norms have anyway. . . . But their authority is not a creature of one’s recognition. (HWV, 105) McDowell relies on the notions of custom, initiation and second nature to understand how we came to acquire responsiveness to reasons. To dispel the idea that the normative realm is something nonhuman, McDowell notably

106  Luca Corti invites us to conceive our responsiveness to reason as natural, although “in a different sense” (EI, 248).18 If we turn to Sellars, we see that the notion of ‘recognition’ was not part of his conceptual apparatus. He rather casts the problem of the origins of norms in terms of how normative responsiveness to the ‘rules’ of a language is learned in the first place. Sellars used a Wittgensteinian notion of ‘training’ to explain the transition from preconceptual patterns of behavior to full-blown responsiveness to normative claims. In his words: as children we learn to understand the noise ‘blue’ in much the same way as the dog learns to understand the noise ‘bone,’ but we leave the dog behind in that the noise ‘blue’ also comes to function for us in a system of rule-regulated symbol activity, and it is a word, a linguistic fact, a rule-regulated symbol only in so far as it functions in this linguistic system. (PPPW, 142) Thanks to this training, human individuals would become responsive to the normative connection characteristic of the space of reasons—the connection that makes up the functional role of a linguistic item. “Only when the child has got the hang of how the sounds function in the language can he be properly characterized as saying ‘this is a book’ ” (SK, 320). Thus, to put it more eloquently, “The upshot of the training is that the child becomes a thinker, exhibiting rational discourse in sequences which are not fixed responses to intra-organic episodes. . . . The child, so to speak, is trained to be free” (BLM, 12–3). McDowell explicitly mentions Sellars when he argues for the idea that “acquiring command of language . . . is coming to inhabit the logical space of reason” (EI, 247). Despite the specific emphases of the various models, these readings are nonetheless intelligible on the basis of a common background. Sellarsian conceptual role semantics, mediated by the renaissance of Wittgensteinian themes, shaped a notion of ‘norm’ that enabled these authors to newly interpret Hegel’s theory of concepts and his notion of Geist.

8.  Perceptual Experience Another topic that Sellars treats in close relation to Kant, and which has paved the way for much current Hegelianism, is notably the issue of perception.19 For Sellars, Kant offers a powerful conception of perceptual content (“Once it is appreciated that Kant’s account of the conceptual structures involved in perception can be given a linguistic turn” [KTM, 268]). In reading Kant’s notion of receptivity, Sellars notably draws a sharp distinction between intuitions and sensations. While, on the one hand, Sellars considers Kant’s ‘sensations’ as blind nonconceptual episodes analogous to his own ‘sense impressions’, on the other, he puts forth a reading that considers

Hegel After Sellars  107 ‘intuitions’ in Kant as conceptually shaped perceptual items that take on a linguistic structure (the famous ‘this-such’). Such a distinction within receptivity is part of Sellars’s overall strategy to answer the problem raised by the Myth of the Given. At the same time, Sellars’s thesis of the conceptuality of intuitions in Kant—together with the textual evidence that Sellars offers to support it—constitutes the starting point of much current Hegelism. Sellars’s ‘conceptualist’ reading of the Metaphysical and Transcendental Deduction of the Critique of Pure Reason—together with his insistence on passages like A79/B 105, in particular—has shaped contemporary readings of Hegel in various ways. Among the first to approach Hegel through this particular reading of Kant was Pippin, who had the opportunity of taking Sellars’s classes on Science and Metaphysics.20 According to Pippin, Hegel has understood the conceptualist upshot of Kant’s Transcendental Deduction better than Kant himself. Such an interpretation of the Kant-Hegel relationship, mediated by a reading of Kant that had been defended by Sellars, would become the basis from which other influential readings were put forth in the following years. McDowell follows Sellarsian thought on the conceptuality of intuition to put forth his philosophical proposal, ‘Hegelian, at least in spirit’ (RS, 269), that conceptual capacities entirely inform conceptual content.21 But Sellarsian insights on perception were already applied to Hegel by Sellars’s students: Willem deVries, for instance, has done pioneering work on Hegel’s theory of mental activity (DeVries 1988) and further connected Sellarsian and Hegelian perspectives (DeVries 2008). Although Sellars notably maintained a role for nonconceptual episodes in his theory of perception and knowledge—a role that left-leaning Sellarsian Hegelians aim to reject—the conceptualist elements of his views of perception and his interpretation of Kant have become constitutive of the Hegel renaissance.

9.  New Directions After Sellars Beyond the disagreements over single issues, these views on normativity and perception constitute a common background linking a distinct family of current interpretations of Hegel whose resemblance sets them apart from other approaches to Hegel. These interpretations have undoubtedly helped bring Hegel back into the debate. From a first phase, in which the translation from Sellars’s idiom into Hegel’s was quite straightforward, more nuanced interpretations have been put forth in recent years. The early equation between Geist and a set of Sellarsian rules—for instance—has been progressively put into question and given up in favor of a broader understanding of Hegel’s notion of Geist. Pinkard, one of the first to apply inferential role semantics to Hegel, revised his own view in order to resist the assimilation of a form of life with a determinate set of rules (including higher order rules for criticism). Hegel himself tried to show

108  Luca Corti that we should understand the series of the shapes of spirit not as a series of systems of rules, but rather as something like a series of forms of life, which function as the conditions under which we can formulate beliefs and rules about beliefs. (IAL, 292, my translation) Understood along these lines, forms of life, or, in Hegelian terms, ‘shapes of spirit,’ can be seen as involving natural-biological aspects in addition to social-normative and linguistic ones. What keeps together a form of life are not only its linguistic practices, but rather all other things constituting it: human mortality, birth, social systems, mundane facts such as the fact that we get tired and need rest and so on. (Ibid., 268, my translation) This broader conception of Geist shed light on other aspects of Hegel’s system, including Hegel’s understanding of more ‘natural’ aspects of normativity. In this way, it has made room for additional lines of inquiry into the question of how to articulate the relation between norm and nature.22 At the same time, Hegel’s Science of Logic has become the object of new readings as the notion of ‘inferential move’ grew to be understood as too narrow to understand the scope and structure of the text. Crudely: Hegel has come to be recognized as putting forth a broader and much more ambitious program identified with his ‘theory of intelligibility,’ i.e., about the conditions for making sense of something at all (Zambrana 2015; Pippin 2016, 2017), or with his theory of explanation or ‘metaphysics of reasons’ (Kreines 2015). In this respect, distinctive normative forms of accountgiving appear to be only one kind of explanation among others that are also addressed by Hegel’s text. This broader approach to Hegel’s work could potentially open up new ways of reading Hegel that may even lead him away from or to oppose Sellars. Indeed, Hegel appears to be critical of some aspects of the very Sellarsian notions, such as the ‘reasons’ that appear in the game of giving and asking for reasons, that have been crucial to Hegel’s recent revival. Hegel notably discusses the topic of ‘reasons’ and justification in the Grund chapter of his Science of Logic. There, he appears to criticize various understandings of the notion, in particular those that see Gründe as the basic form of intelligibility for understanding actions. An action can have several reasons (Gründe); as something concrete, it contains a manifold of essential determinations, each of which can therefore be offered as the reason. (WdL, 11.311–12)

Hegel After Sellars  109 And Hegel goes on to add The search and the assigning of reasons . . . [is] an endless meandering without final destination; for each and every thing good reasons can be adduced, but so they can for its opposite just as well, and there can be a great many reasons with nothing following from them. (Ibid.) To some extent Hegel thus seems to be suspicious of the game of giving and asking for reasons, at least as a key for understanding actions, and sees the exchange of reasons as a partial form of account-giving. As has been noted, “part of Hegel’s argument . . . involves a thoroughgoing critique of the concept of a Grund” and Sellarsian ‘reasons’ appear at least in some ways to be “still instances of what German calls Gründe” (Rand 2016, 508).23 This opens a new space for dialogue between Hegel and Sellars with the potential for new elements to come to the fore.

10. Conclusion In this paper, I have tried to point to some moments in the history of American philosophical thought that have brought Sellars’s ideas to bear on Hegel. The core idea leading the Hegel revival was Sellars’s Conceptual Role Semantics—with his inherent account of conceptual activity and its inferential nature. Key ideas related to Conceptual Role Semantics were translated, via Sellars, into the lingua franca of Kantianism. Then, starting in the late 1970s, this translation into Kantianese continued thanks to additional influences and it became the basis from which to interpret Hegelian texts. A key term for understanding the new interpretation of Hegel is ‘rule’ or ‘norm’—especially in the inferential account of norm Sellars conceived as the paradigmatic ‘intra-liguistic move.’ On the other hand, issues on perception and Kant have opened the second big Sellarsian gateway to Hegel. Sellars’s interpretation of Kant’s notion of intuition had a profound influence on Hegel scholarship, shaping the first interpretation of Hegel’s Phenomenology of Spirit as a radicalization of Kant and building Hegel’s current reputation as a conceptualist on perceptual experience. Of course, there are areas of Sellars’s thought that have not played any role in the Hegelian renaissance—his scientific naturalism, for instance, and all the aspects that today are central to what is called ‘right-wing-Sellarsianism,’ which follows Sellars’s scientia mensura dictum and makes normativity secondary. Furthermore, compared to the entirety of the systematic contents of Hegel’s philosophy, Sellarsian insights bear on relatively few parts of Hegel’s thought. However, at least in the first phases of the revival, they played a constitutive role, helping to shed new light on some aspects of Hegel’s

110  Luca Corti text and paving the way for critics to work out other aspects of Hegel’s philosophy, such as the Hegelian theories of action, or aesthetic intelligibility. At the same time, they have catalyzed a discussion on Hegel that has enabled more accurate study of Hegel’s texts in the Anglophone world and the formulation of antagonist Hegelian views and interpretation, including those that draw thoughtful attention to the differences between Hegel and Sellars and possible Hegelian responses to Sellarsian topics. To use a phrase by Dennett: “Sellars’ influence has been ubiquitous but almost subliminal” (Dennett 1987, 341) in all these respects.

Notes 1 This included the influence of the Saint Louis Hegelians, the interest in Hegel of some pragmatists such as Dewey, as well as the philosophy of Josiah Royce and other circles. On this cf. Muirhead (1928), Riedl (1973), Pochmann (1948), Kuklick (2001, Ch. 5), Good (2002, 2006). 2 For the anti-Hegelianism in pragmatism, see Bernstein (2010); for anti-Hegelianism as a ‘foundational myth’ in analytic philosophy, see Hylton (1993), Welsch (2005), Redding (2007, 2011). 3 Kuklick (1984, 133): “After the war Hegel became, for Americans, a silly, pompous, and defeated figure, unworthy of the great tradition. Indeed, the wonder is not that Hegel vanished, but that Kant remained.” 4 See Corti (2015), Harris (1983) and Beiser (2007). 5 I agree on this point with Paul Redding (2007), 16. 6 The expression ‘normative functionalism’ has been used by Brandom in BAR, 95, O’Shea (2007), Lance (2008), and then by Maher (2012a, 2012b). 7 For an overview cf. deVries (2005, 30–1), O’Shea (2007, 60–1) and Lance (2008, 408–9). 8 Sellars thinks that there is indeed a sort of reducibility, which is causal but does not affect the autonomous logical status of the space of reasons. To quote Sellars: “Let us agree to put this by saying that although Ought is not logically reducible to Is, Ought is causally reducible to Is” (MMB, 85). He also states: “The task of the philosopher cannot be to show how, in principle, what is said by normative discourse could be said without normative discourse, for the simple reason that this cannot be done” (SSMB, 214). 9 For a discussion of Sellars’s use of the word ‘transcendental,’ see Olen (2016, 64 ff). 10 “I would have to work out a whole new way of looking at the conceptual order. . . . 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, 285). 11 At the University of Austin, Terry Pinkard was exposed to Sellars’s ideas thanks to a young Sellarsian teacher, Laurence Bonjour, while at the same time he was also exposed to a non-metaphysical reading of Hegel put forth by John Findlay and Klaus Hartmann, who also taught at Austin while Pinkard was a student there. 12 The debate featured various articles, see Lear (1982), Lear and Stroud (1984), Moore (1985), Williams (1974). 13 The term ‘like-mindedness’, which has become a standard gloss for Hegel’s Geist, was used for the first time in a Wittgensteinian context. It was however soon put to work in interpretations of Hegel. Pippin for instance writes (HI, 278, fn 5) that “The problem Hegel faces can be usefully compared with similar issues in Wittgenstein,” and in particular with the “understanding of what the

Hegel After Sellars  111 Wittgensteinean would call a fundamental human ‘like-mindedness’.”. See also Pippin WQ, 123 and HPP, 104. Pippin, however, reminds us that Hegel never uses what can be considered the German equivalents (gesinnt, gleichgesinnt), (HPP, 195, fn 6). A Wittgensteinian reading of Hegel was notably put forth by Findlay (1958) and Lamb (1979). 14 In particular, Kant would grasp the ‘normative character of concept use’. “Kant understands concepts as having the form of rules, which is to say that they specify how something ought (according to the rule) to be done” (BAR, 163. Cf. also KL, 70, TMD, 21, MIE, Ch. 1 and AI). 15 Based on the slogan “Conceptual contents are inferential roles” (MIE, xiv, 85 ff.; TMD, 6). 16 The correspondence appears to be straightforward. According to Brandom “’Mediation,’ is his [Hegel’s] term for inferential articulation . . . ‘Determinate negation’ is his term for material incompatibility” (MIE, 92). Things are different for ‘logical’ concepts, from which Brandom adopts an ‘expressivist’ conception of logic. We will not dig into the distinction here (cf. SPCR, 134 and Corti 2014, 114ff.). Brandom’s semantic theory is in turn grounded pragmatically, in the specific sense that it is dependent on a pragmatic theory described in terms of deontic normative statuses. These pragmatic grounds, according to Brandom, have a Hegelian counterpart as well (see below). 17 For reference to Sellars account of normativity in Pippin’s reading of Hegel, see TPS, 45ff: “it seems that Kant and Hegel are making a point similar to one made by Sellars in a linguistic context, that the objective grasp of a concept is nothing other than mastery of the correct use of a word.” 18 Alongside that, McDowell interprets Hegel’s famous passages on Master-Servant dialectics but rather as an ‘allegory’ to understand the relation between an empirical I and an apperceptive I. Therefore, in the whole passage, “only one biological individual is really in play” (HWV, 153). In McDowell’s words: “The description of the struggle to the death works as an allegorical depiction of an attempt, on the part of a single self-consciousness, to affirm its independence, by disavowing any dependence on ‘its objective mode,’ which is the life that has come to stand in for the otherness of the world whose scene that life is” (ibid., 161). 19 On this point, see Redding (2007, cap. 1; 2011). 20 See Corti (2014) and Pippin (2007). 21 There is disagreement between Kant’s position and Hegel’s critique between Pippin and McDowell, cf. TPS, 166 ff, MG, and HWV, 185 ff. Brandom focuses on the notion of judgment and does not pay attention to Kant’s notion of intuition. In his book he states “Among the topics I did not find it necessary so much as to mention are intuition, sensibility, receptivity” cf. AI, 50. 22 Pinkard (2012), Papazoglou (2012), Giladi (2014). 23 For a further discussion of the notion of Grund, see Yeomans (2012, 75ff).

References Beiser, Frederick C. 2007. ‘Dark Days: Anglophone Scholarship Since the 1960s’. In German Idealism: Contemporary Perspectives. Edited by E. Hammer, 70–90. London and New York: Routledge. Beiser, Frederick C. 2008. ‘Introduction: The Puzzling Hegel Renaissance’. In The Cambridge Companion to Hegel and Nineteenth-Century Philosophy. Edited by Frederick Beiser, 1–14. Cambridge: Cambridge University Press. Bernstein, Richard. 2010. ‘Hegel and Pragmatism’. In The Pragmatic Turn. Edited by Richard Bernstein, 89–105. Cambridge: Polity Press.

112  Luca Corti Brandom, Robert (MIE). 1994. Making It Explicit: Reasoning, Representing, and Discursive Commitment. Cambridge, MA: Harvard University Press. Brandom, Robert B. (BAR). 2000. Articulating Reasons. Cambridge, MA: Harvard University Press. Brandom, Robert (TMD). 2002. Tales of the Mighty Dead: Historical Essays in the Metaphysics of Intentionality. Cambridge, MA: Harvard University Press. Brandom, Robert (RAP). 2008. ‘Replies’. In Robert Brandom: Analytic Pragmatist. Edited by Bernd Prien and David P. Schweikard, 163–94. Frankfurt am Main: Ontos Verlag. Brandom, Robert (SDR). 2007. ‘The Structure of Desire and Recognition: SelfConsciousness and Self-Constitution’. Philosophy Social Criticism, 33: 127–50. Brandom, Robert (SPCR). 2008. ‘Sketch of a Program for a Critical Reading of Hegel: Comparing Empirical and Logical Concepts’. Internationales Jahrbuch des Deutschen Idealismus, 3: 131–61. Brandom, Robert (AI). 2009. Reason in Philosophy: Animating Ideas, Cambridge, MA: Harvard University Press. Brandom, Robert (POP). 2011. Perspectives on Pragmatism: Classical, Recent and Contemporary. Cambridge, MA: Harvard University Press. Brandom, Robert (KL). 2006. ‘Kantian Lessons About Mind, Meaning, and Rationality’. Southern Journal of Philosophy, 44: 49–71. Corti, Luca. 2014. Ritratti hegeliani: Un capitolo della filosofia americana contemporanea. Rome: Carocci. Corti, Luca. 2015. ‘Hegel in America: momenti di una rinascita’. Il pensiero, 55, 1–2: 70–95. Dennett, Daniel. 1987. The Intentional Stance. Cambridge, MA: MIT Press. Descombes, Vincent. 2014. The Institutions of Meaning: A Defense of Anthropological Holism. Cambridge, MA: Harvard University Press. deVries, Willem A. 2005. Wilfrid Sellars. Montreal: McGill-Queen’s University Press. deVries, Willem A. 2009. ‘Getting Beyond Idealisms’. In Empiricism, Perceptual Knowledge, Normativity, and Realism: Essays on Wilfrid Sellars. Edited by Willem DeVries. Oxford: Oxford University Press. deVries, Willem A. 1988. Hegel’s Theory of Mental Activity: An Introduction to Theoretical Spirit. New York: Cornell University Press. deVries, Willem A. 2008. ‘Sense Certainty and the This-Such’. In Hegel’s Phenomenology of Spirit: A Critical Guide. Edited by Dean Moyar and Michael Quante. Cambridge: Cambridge University Press. Findlay, John N. 1958. Hegel: A Re-Examination. London: Allen & Unwin. Giladi, Paul. 2014. ‘Liberal Naturalism: The Curious Case of Hegel’. International Journal of Philosophical Studies, 22, 2: 248–70. Good, James A. (Ed.). 2002. The Early American Reception of German Idealism. 5 vol. Bristol: Thoemmes. Good, James A. 2006. ‘John Dewey’s “Permanent Hegelian Deposit” and the Exigencies of War’. Journal of the History of Philosophy, 44, 2: 293–313. Haag, Johannes. 2012. ‘Some Kantian Themes in Wilfrid Sellars’ Philosophy’. Paradigmi, 1: 111–26. Harris, Henry S. 1983. ‘The Hegel-Renaissance in the Anglo-Saxon World Since 1945’. The Owl of Minerva, 15, 1: 77–106. Hylton, Peter. 1993. ‘Hegel and Analytic Philosophy’. In The Cambridge Companion to Hegel: Cambridge University Press. Edited by F. Beiser, 445–85. Cambridge: Cambridge University Press.

Hegel After Sellars  113 Kreines, James. 2015. Reason in the World: Hegel’s Metaphysics and Its Philosophical Appeal. Oxford: Oxford University Press. Kuklick, Bruce. 1984. ‘Seven Thinkers and How They Grew: Descartes, Spinoza Leibniz: Locke, Berkeley, Hume, Kant’. In Philosophy in History. Edited by R. Rorty, J.B. Schneewind and Q. Skinner, 125–39. New York: Cambridge University Press. Kuklick, Bruce. 2001. A History of Philosophy in America (1720–2000). Oxford: Clarendon Press. Lamb, David. 1979. Language and Perception in Hegel and Wittgenstein. New York: St. Martin’s Press. Lance, Mark. 2008. ‘Placing in a Space of Reasons: Sellarsian Philosophy in the 21st Century’. In Oxford Companion to American Philosophy. Edited by Cheryl Misak, 403–30. Oxford: Oxford University Press. Lear, Jonathan. 1982. ‘Leaving the World Alone’. The Journal of Philosoph, 79, 7: 382–403. Lear, Jonathan and Stroud, Barry. 1984. ‘The Disappearing “We” ’. Proceedings of the Aristotelian Society, Supplementary Volumes, 58: 219–58. Maher, Chauncey. 2012a. The Pittsburgh School of Philosophy: Sellars, McDowell, Brandom. London: Routledge. Maher Chauncey. 2012b. ‘Normative Functionalism’. Social Epistemology Review and Reply Collective, 2, 1: 13–5. McDowell, John (MW). 1994. Mind and World. Cambridge, MA: Harvard University Press. McDowell, John (EI). 2009. The Engaged Intellect: Philosophical Essays. Cambridge, MA: Harvard University Press. McDowell, John (HWV). 2009. Having the World in View: Essays on Kant, Hegel, and Sellars. Cambridge, MA and London: Harvard University Press. Moore, Adrian W. 1985. ‘Idealism in Wittgenstein, and Theories of Meaning’. The Philosophical Quarterly, 35, 139: 134–55. Muirhead, John H. 1928. ‘How Hegel Came to America’. The Philosophical Review, 37, 3: 226–40. Olen, Peter. 2016. Wilfrid Sellars and the Foundations of Normativity. London: Palgrave Macmillan. O’Shea, James. 2006. ‘Conceptual Connections: Kant and the Twentieth Century Analytic Tradition’. In A Companion to Kant. Edited by G. Bird, 513–27. Oxford: Wiley-Blackwell. O’Shea, James. 2007. Wilfrid Sellars: Naturalism with a Normative Turn. Cambridge: Polity Press.Papazoglou, Alexis. 2012. ‘Hegel and Naturalism’. Hegel Bulletin, 33: 74–90. Pinkard, Terry (LHL). 1979. ‘The Logic of Hegel’s “Logic” ’. Journal of the History of Philosophy, 17, 4: 417–35. Pinkard, Terry (HIHL). 1979. ‘Hegel’s Idealism and Hegel’s Logic’. Zeitschrift für Philosophische Forschung, 33, 2: 210–26. Pinkard, Terry (HP). 1994. Hegel’s Phenomenology: The Sociality of Reason. Cambridge: Cambridge University Press. Pinkard, Terry (IAL). 2004. ‘Innen, Außen, und Lebensformen: Hegel und Wittgenstein’. In Hegels Erbe. Edited by C. Halbig, M. Quante and L. Siep, 254–94. Frankfurt am Main: Suhrkamp. Pinkard, Terry (HN). 2012. Hegel’s Naturalism: Mind, Nature, and the Final Ends of Life. Oxford: Oxford University Press.

114  Luca Corti Pippin, Robert (HI). 1989. Hegel’s Idealism: The Satisfaction of Self-Consciousness. Cambridge: Cambridge University Press. Pippin, Robert (WQ). 2000. ‘What Is the Question for Which Hegel’s “Theory of Recognition” Is the Answer?’ European Journal of Philosophy, 8, 2: 155–72. Pippin, Robert (BH). 2005. ‘Brandom’s Hegel’. European Journal of Philosophy, 13, 3: 381–408. Pippin, Robert (TPS). 2005. The Persistence of Subjectivity: On the Kantian Aftermath. Cambridge: Cambridge University Press. Pippin, Robert (I). 2007. ‘Interview with Robert Pippin’. The Yale Philosophy Review, 3: 79–97. Pippin, Robert (MG). 2007. ‘McDowell’s Germans: Response to “On Pippin’s Postscript” ’. The European Journal of Philosophy, 15, 3: 411–34. Pippin, Robert (HPP). 2008. Hegel’s Practical Philosophy: Rational Agency as Ethical Life. Cambridge: Cambridge University Press. Pippin, Robert. 2016. ‘Logik und Metaphysik—Hegels‚ Reich der Schatten’. HegelJahrbuch, 371–86. Pippin, Robert. 2017. ‘Hegel on Logic as Metaphysics’. In The Oxford Handbook to Hegel. Edited by D. Moyar. Oxford: Oxford University Press. Pochmann, Henry A. 1948. New England Transcendentalism and St. Louis Hegelianism. Philadelphia: Carl Schurz Memorial Foundation. Popper, Karl. 1945. The Open Society and Its Enemies. London: George Routledge & Sons. Rand, Sebastian. 2016. ‘Review of James Kreines’ “Reason in The World” ’. Journal of the History of Philosophy, 54, 3: 508–509. Redding, Paul. 2007. Analytic Philosophy and the Return of Hegelian Thought. Cambridge: Cambridge University Press. Redding, Paul. 2011. ‘The Analytic Neo-Hegelianism of John McDowell and Robert Brandom’. In The Blackwell Companion to Hegel. Edited by Stephen Houlgate and Michael Baur, 576–93. Oxford: Wiley-Blackwell. Riedl, John O. 1973. ‘The Hegelians of Saint Louis, Missouri and Their Influence in the United States’. In The Legacy of Hegel. Edited by J.J. O’Malley, K.W. Algozin, H.P. Kainz and L.C. Rice, 268–87. Dordrecht: Springer. Russell, Bertrand. 1945. A History of Western Philosophy: And Its Connection with Political and Social Circumstances from the Earliest Times to the Present Day. New York: Simon & Schuster. Sellars, Wilfrid (LRB). 1949/2005. ‘Language, Rules and Behavior’. In Pure Pragmatics and Possible Worlds: The Early Essays of Wilfrid Sellars. Edited by Jeffrey Sicha, 117–34. Atascadero, CA: Ridgeview Publishing Company. Sellars, Wilfrid (MMB). 1952. ‘Mind, Meaning, and Behavior’. Philosophical Studies, 3: 83–95. Sellars, Wilfrid (SSMB). 1953. ‘A Semantical Solution of the Mind-Body Problem’. In Pure Pragmatics and Possible Worlds: The Early Essays of Wilfrid Sellars. Edited by J. Sicha, 186–214. Atascadero, CA: Ridgeview Publishing Company. Sellars, Wilfrid (SM). 1968. Science and Metaphysics: Variations on Kantian Themes. Atascadero, CA: Ridgeview Publishing Company. Sellars, Wilfrid (SRLG). 1954/1963. ‘Some Reflections on Language Games’. In Science, Perception and Reality, 321–58. Atascadero, CA: Ridgeview Publishing Company.

Hegel After Sellars  115 Sellars, Wilfrid (PSIM). 1962/1963. ‘Philosophy and the Scientific Image of Man’. In Science, Perception and Reality, 1–40. Atascadero, CA: Ridgeview Publishing Company. Sellars, Wilfrid (EPM). 1963. ‘Empiricism and the Philosophy of Mind’. In Science, Perception and Reality. Edited by Wilfrid Sellars, 127–96. London: Routledge & Kegan Paul. Sellars, Wilfrid (SRTT). 1967. ‘Some Reflections on Thoughts and Things’, Noûs, 1, 2: 97–121. Sellars, Wilfrid (LTC). 1969. ‘Language as Thought and Communication’. Philosophy and Phenomenological Research, 29: 506–27. Sellars, Wilfrid (MFC). 1974. ‘Meaning as Functional Classification’. Synthese, 27: 417–37. Sellars, Wilfrid (AR). 1975. ‘Autobiographical Reflections: (February 1973)’. In Action, Knowledge and Reality. Edited by Hector Neri Castañeda, 277–93. New York: Bobbs-Merrill. Sellars, Wilfrid (SK). 1975. ‘The Structure of Knowledge: (1) Perception; (2) Minds; (3) Epistemic Principles’ (The Matchette Foundation Lectures for 1971 at the University of Texas). In Action, Knowledge and Reality: Studies in Honor of Wilfrid Sellars. Edited by Hector-Neri Castañeda, 295–347. Indianapolis, IN: Bobbs-Merrill. Sellars, Wilfrid (BLM). 1980. ‘Behaviorism, Language, and Meaning’. Pacific Philosophical Quarterly, 17: 81–101. Sellars, Wilfrid (NAO). 1980. Naturalism and Ontology. Atascadero, CA: Ridgeview Publishing Company. Sellars, Wilfrid (PPPW). 1980. Pure Pragmatics and Possible Worlds: The Early Essays of Wilfrid Sellars. Edited by J. Sicha. Atascadero, CA: Ridgeview Publishing Company. Sellars, Wilfrid (KTM). 2002. Kant’s Transcendental Metaphysics: Sellars’ Cassirer Lectures Notes and Other Essays. Edited by Jeffrey F. Sicha. Atascadero, CA: Ridgeview Publishing Company. Sellars, Wilfrid (WSNDL). 2009. Wilfrid Sellars Notre Dame Lectures (1969–1986). Edited with an Introduction by Pedro Amaral, published online by Andrew Chrucky at www.ditext.com/amaral/wsndl.pdf. Welsch, Wolfgang. 2005. ‘Hegel und die analytische Philosophie’. In Wissen und Begründung. Edited by K. Vieweg and B. Bowmann, 11–74. Würzburg: K& N. Westphal, Kenneth. 2010. ‘Kant’s Critique of Pure Reason and Analytic Philosophy’. In Cambridge Companion to the Critique of Pure Reason. Edited by Paul Guyer, 401–30. Cambridge: Cambridge University Press. Williams, Bernard. 1974. ‘Wittgenstein and Idealism’. In Understanding Wittgenstein. Edited by Royal Institute of Philosophy Lectures, 76–95, Vol. 7. London: Palgrave Macmillan. Yeomans, Christopher. 2012. Freedom and Reflection: Hegel and the Logic of Agency. Oxford: Oxford University Press. Zambrana, Rocìo. 2015. Hegel’s Theory of Intelligibility. Chicago: University of Chicago Press.

Part II

Sellars and the Beginning of the Contemporary Age

6 Peirce and Sellars on Nonconceptual Content Catherine Legg

How do sensory experiences (or qualia) manage to combine with concepts to produce the rich human thinking we all enjoy? Obviously it happens, as for example when I form a perception of a yellow chair in which an experience of yellowness and a concept of ‘chairhood’ both participate. But the question of how exactly it happens seems worthy of philosophical reflection in that these two things seem crucially different from one another. They differ first in that sensory experiences enjoy a certain felt immediacy or presence (that particular yellow, which I may or may not be able to express in words), while concepts participate in wide-ranging logical relationships across an entire space of reasons (for instance that the yellow chair, qua chair, is not identical with any single electron in the Universe: actual or even potential). But the puzzle extends even further—sensory experiences and concepts seem possessed of incompatible properties in that the former seem to enjoy a kind of ‘private’ infallibility, while judgments constructed from concepts are clearly truth-apt and fallible. Previously I have explicated this ‘Experience-Truth Gap’ as follows: On the one hand, my perceptions are suffused with immediately felt experience (for instance, the juicy, sweet ‘cherryness’ of a cherry I am biting into), which it seems that in some important sense ‘no-one can take away from me.’ Thus the nature of our sensory feels appears to enjoy some degree of infallibility. (‘Even if that cherry was a total hallucination, I can’t be wrong about how it tasted to me.’) On the other hand, much of the point of perception seems to be to enable us to endorse new propositions about the world that are truth-apt. (‘This cherry is delicious! But is it really a cherry, or rather a small plum?’) In this regard our perceptions seem perfectly fallible. This is all rather confusing. (Legg 2017, 42–3) Insofar as this disparity does represent a tension in philosophy of perception, it obviously bleeds profusely into epistemology more generally.

120  Catherine Legg Sellars was far-sighted in his day in noting and highlighting this tension. He thought (rightly, I believe) that it was being quite overlooked by his empiricist contemporaries, including noted logical positivists such as Carnap and Ayer, leading to “perplexities . . . when one tries to think [empiricism] through” (Sellars 1997, §7, 25). He diagnosed the problem as lying with the peculiar infallibility attributed to sensory experience. He famously dubbed this the Myth of the Given. But the Myth’s exact legacy, and in what way philosophers should best avoid it, if they should, is still being determined. Important work here has been done by Sellars’s students—notably John McDowell. In McDowell (2004a, 2004b), he explored how sensibility and the understanding (or more poetically, following Kant, receptivity and spontaneity) combine to create human cognition that is subject to logical appraisal. In seeking to reconcile these two faculties, as he understood them, McDowell sought a middle way between the bald naturalism that he argued results from over-reliance on receptivity and the ‘frictionless coherentism’ that he argued results from over-reliance on spontaneity. He argued that Sellars’s identification of the Myth led him to excessively deprecate the contribution of sensory experience to human cognition, developing instead a purely conceptual epistemology (i.e., an inferentialism) which in McDowell’s view falls into frictionless coherentism. Thus, McDowell charges Sellars with failing to account for the primal confrontation between mind and world which constitutes genuine perception, and thus knowledge. But since then, McDowell has not been entirely satisfied with his appraisal of Sellars. In Having the World in View he has returned to the same ground and offered a somewhat different take, which I shall discuss below. Robert Brandom has also done much to bring these issues to light (Brandom 1994, 2000, 2007). His dogged defense of an inferentialist as opposed to a representationalist account of content—according to which the valid inferences into which a given proposition may enter constitute its entire meaning—has been (among other things) an attempt to rehabilitate the conceptual, and thus rationalism, against the unreflective empiricism of most analytic philosophy. The systematicity he has managed to achieve in working out his new inferentialist doctrine has brought fascinating new currents to mainstream philosophy. But it has also dismayed a number of Brandom’s pragmatist peers, who feel that he has swung so far towards rationalism that he has landed in intellectualism.1 Brandom’s analytic pragmatism has been charged with an ‘experience problem.’ For instance, Paul Redding has argued that Brandom fails to recognize the way in which by means of perception we form beliefs de re as well as de dicto. Giving the example of himself wearing a blue tie, Redding claims that perceptual experience will most likely show the tie to be some particular shade of blue which, although it can be more and more precisely described, can never be fully captured in concepts, and thus in propositional form: Perceptual experience, it might be said, is more fine-grained than what is actually captured by any general concept . . . this feature of perceptual

Peirce and Sellars on Nonconceptual Content  121 experience does not seem to be captured in the de dicto expression: the semantics of a de dicto expression depends simply on whether the proposition is true or false. The way in which my tie makes the proposition true or false drops from consideration altogether. (Redding 2014, 668) In other words, perception seems to be a belief-forming mechanism characterized by a kind of direct confrontation between mind and world which, in its felt immediacy and the bottomless determinacy of its detail, differs crucially from merely understanding that the proposition “Paul Redding is wearing a blue tie” is true. Admittedly, Brandom has tried to clear some space in his epistemology for sense-experience. He acknowledges that for apparently simple pure qualia such as the color red, it seems counterintuitive to account for their meaning solely in terms of inferences licensed both ‘upstream’ (in propositions from which it follows that something is red) and ‘downstream’ (in propositions which follow from something being red). What would such inferences be? Didn’t Frank Jackson show us with his famous ‘Black and White Mary’ thought experiment (Jackson 1982) that it is conceivable that a person might be able to license all valid red-related inferences, yet still not understand the true meaning of red—the unique and distinctive experience of seeing it?2 For such concepts, then, Brandom makes an exception to his inferentialism, allowing their meaning to be at least partly specified by simple (i.e., mere stimulus-response) ‘reports’ on the part of relevant sensory mechanisms: in Articulating Reasons, he calls this the color red’s “noninferential circumstances of appropriate application” (Brandom 2000, 21). This concession to qualia is what makes Brandom a strong rather than a hyperinferentialist. He describes himself as thereby offering a ‘two-ply’ theory of perception, which weaves together an experiential and an inferential strand, thereby showing itself: the product of two distinguishable sorts of abilities: the capacity reliably to discriminate behaviorally between different sorts of stimuli, and the capacity to take up a position in the game of giving and asking for reasons. (Brandom 2002, 349) But in order that his first ‘strand’ avoid the Myth of the Given, Brandom qualifies that “even such noninferential reports must be inferentially articulated” (Brandom 2000, 47). This qualification is what makes Brandom a strong rather than a weak inferentialist. How can a noninferential report be inferentially articulated? Brandom gives the example of a culture where “white is the colour of death and things associated with death are to be shunned or avoided.” Here the concept white “can be understood to be inferential in a broad sense, even when the items connected are not themselves sentential” (Brandom 1994, 658).

122  Catherine Legg To sum up this section, sensory experience and conceptual understanding seem prima facie quite different things. It would be difficult to deny that sensory experience plays a vital role in perception—yet where the border exists between it and the conceptual, and how they manage to combine to produce thought, is still at this point rather mysterious. One way of posing these issues is as a discussion of so-called nonconceptual content. If we admit that experience does play a vital role in perception, should we not then understand perception as delivering nonconceptual as well as conceptual content? Lively debates on this matter currently abound in mainstream philosophy (e.g., Peacocke 1992, 1998; Bermúdez 1995; Speaks 2005). Here I wish to bring Charles Peirce to the table as I believe he has ideas to offer these characteristically ‘Pittsburgh School’ debates that are highly congenial, yet to date largely unexplored. In prior work I have situated Peirce with respect to Brandom’s inferentialism, arguing that Peirce is a hyperinferentialist due to his early (1860s) unequivocal rejection of intuition in epistemology, claiming instead that all thought is in signs. I argued that Peirce works out an inferentialist account even of qualia such as red through naturalistic arguments which challenge the self-certainty of introspection, showing that we are not able to tell by intuition whether a thought is an intuition or an inference, and (what comes to the same thing) sketching a model of our sensory capacities in which many inferences occur below conscious awareness (Legg 2008). In later papers (Legg 2014, 2017) I have begun to explore Peirce’s more mature commentary on these issues, in a full-fledged theory of perception that he presented around 1902. Here Peirce accords a more explicit and direct role to experience in an entirely preconceptual percept, but this is overlaid with a level of perceptual judgment which (I believe) corresponds to Sellars’s space of reasons. What is most interesting here is the relationship Peirce charts between percept and perceptual judgment: the latter does not describe, nor is it justified by the former, rather, it indexes it—how will be explained below. As I have not yet related Peirce’s theory of perception specifically to Sellars—who might justly be described as ‘inferentialism’s original fount’— this is my task here. I will begin by explicating Sellars’s classic account of the Myth of the Given, and how to avoid it, in Empiricism and Philosophy of Mind (henceforth: EPM). Then I will consider McDowell’s most recent (2009) critique of Sellars’s account of perception, in order to sharpen our understanding. Then I will present Peirce’s mature theory of perception, and finally relate it to Sellars’s view, thereby showing how Sellars is able to avoid McDowell’s critique of him. We will discover a remarkable amount in common between Sellars and Peirce on perception. This includes a direct realism that is rare in modern philosophy, paired somewhat paradoxically with a crucial role played by diachronic habit as against Cartesian ideals of immediate apprehension. (It will, moreover, be explained why the paradox is only

Peirce and Sellars on Nonconceptual Content  123 apparent.). These two features lead to an overall critical common-sensist account of rationality, which Sellars counterposes to the mainstream foundationalism of his day, and which I shall argue is also very Peircean. But I shall begin by reviewing the famous Myth of the Given, and its dire warnings.

The Myth of the Given The Myth of the Given has arguably itself attained somewhat mythical status in certain philosophical circles as simultaneously deeply important not to commit, yet difficult to understand. But the underlying idea is relatively straightforward. Myth-makers purport to analyze “epistemic facts” into “non-epistemic facts” (Sellars 1997, §5, 19), which doesn’t work. As noted, Sellars diagnoses the Myth as most pressingly a problem for classical empiricism.3 Although the claim that all knowledge derives from experience is defining for empiricism, Sellars asserts that a crucial ambiguity infects precisely that concept. By Sellars’s day, experience was frequently being referred to as sense-data, and he adopts this terminology for purposes of critical scrutiny. Famously, he claims that the concept of sense-data confuses two importantly different notions. The first is: “1) The idea that there are inner episodes—e.g., sensations of red . . . which can occur to human beings . . . without any prior process of learning or concept formation.” This is a purely causal notion of bodily impingement which accords, for instance, with Hume’s official account of an impression. The second notion is: “2) The idea that there are certain inner episodes which are . . . noninferential knowings that certain items are, for example red . . . and that these episodes are the necessary conditions of empirical knowledge as providing the evidence for all other empirical propositions” (Sellars 1997, §7, 21–2). This is what Hume’s notion of an impression actually needs to be in order for it to play the role that Hume accords it in his epistemology. Why is this confusion pernicious? First and foremost, it is intellectually dishonest. Sense-data notion 2) may be a basis for an epistemology— although Sellars’s sympathies ultimately do not lie with this variety of foundationalism (Sellars 1997, §38, 78), but it is not non-epistemic, as ‘knowings’ occur in it unreduced. Only sense-data notion 1) actually presents a non-epistemic analysans, but it is no proper basis for an epistemology. Sellars’s argument for this is lengthy and subtle, but essentially he argues that sense-data so understood cannot play an appropriate justificatory role with regard to our knowledge. It is simply not helpful to state that because X ‘looks’ red to P, P is justified in believing that X is red, and therefore (if X is indeed red) ‘P knows that X is red.’ Sellars’s argument against this philosophical move carefully disentangles the way that ‘looks talk’ is actually used. Aided by a legendary thoughtexperiment about a tie shop with confusing lighting, he notes that the statement, ‘This tie looks green’ only arises when it is known that surrounding

124  Catherine Legg conditions are liable to make a tie appear green when it is not really green (e.g., if one takes the tie outside into sunlight). Otherwise, all parties involved in the situation will simply state, ‘This tie is green.’ So, rather than the statement ‘X looks green’ routinely offering epistemic support for the knowledge ‘X is green,’ the statement is used precisely to signal an appearance that one has reason to believe may well differ from reality, due to: (i) extensive background knowledge of standard conditions for something to look the color that it is, (ii) knowledge that those conditions are not met in this case. And so, when it comes to epistemic priority, ‘X looks green’ is parasitic on ‘X is green,’ rather than the reverse (Sellars 1997, §18, 43). Such logical dependencies of color-concepts on extensive background knowledge of suitable conditions in which to view them are anathema to traditional empiricism’s logical atomism, according to which something’s ‘looking green’ might be the only thought one ever has (Sellars 1997, §19, 44). Lest it be argued that the particular contingent tie shop scenario is a very slender reed to support the strong conclusion that claims of the form ‘X looks P’ are parasitic on claims of the form ‘X is P’ in all circumstances, Sellars develops further strands to his argument. These include trying to imagine a community that uses ‘is talk’ but not ‘looks talk’—which he argues is conceivable—and trying to imagine a community that uses the latter but not the former—which he argues is not.4 How does Sellars propose to avoid the Myth? He counterposes to his contemporaries’ unreflective empiricism a rationalism that holds that there is no direct or intrinsic characterization of immediate experience (Sellars 1997, §25, 57). Nor is it right, he argues, to build a systematic account whereby immediate experience plays a ‘self-authenticating’ role for our knowledge. Rather than building our epistemology around the search for ideal prior conditions for knowledge-formation in the individual, we should understand knowledge-formation as a group endeavor in which critical scrutiny can be applied by anyone, anywhere, at any time: the correctness of a report does not have to be construed as the rightness of an action. A report can be correct as being an instance of a general mode of behavior which, in a given linguistic community, it is reasonable to sanction and support. (Sellars 1997, §35, 74) Relatedly, in characterizing an episode or state as that of knowing, we are not giving an empirical description of that episode or state; we are placing it in the logical space of reasons, of justifying and being able to justify what one says. (Sellars 1997, §36, 76)

Peirce and Sellars on Nonconceptual Content  125 Yet isn’t it counterintuitive (so to speak!) to hold that there is no direct or intrinsic characterization of immediate experience? Sellars does not deny that sensations are present, and that they play a causal role in our perception: One can certainly admit that the tie between ‘red’ and red physical objects . . . is causally mediated by sensations of red, without being committed to the mistaken idea that it is ‘really’ sensations of red, rather than red physical objects, which are the primary denotation of the word ‘red.’ (Sellars 1997, §29, 64) So is Sellars committed to nonconceptual content or not? We turn now to McDowell for further enlightenment.

McDowell on Sellars: Rational Capacities Permeate Our Perception (But Not Rational Judgments) We noted that in Mind and World McDowell charges Sellars with an excessively rationalist coherentism. In his 1997 Woodbridge lectures, published in 2009 as Having the World in View, McDowell adopts a more nuanced evaluation of his great teacher. Much of this discussion departs from a certain statement that Sellars makes when rejecting foundationalism in epistemology as standardly construed. Sellars remarks that although he does not wish to say that empirical knowledge has no foundation, the metaphor of ‘foundation’ is misleading, because: it keeps us from seeing that if 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. (Sellars 1997, §38, 78) This second logical dimension, which runs to observation statements from other empirical propositions, was evident in the tie shop example, where the salesperson takes the tie outside in the sun and observes ‘This tie is blue’ not only because he is seeing blueness, but also because he knows that sunlight constitutes reliable conditions for detecting blueness. In order to further explicate the two-way nature of this logical relationship, McDowell develops the metaphor of a line demarcating Sellars’s allimportant “logical space of reasons”: Sellars’ thesis is that the conceptual apparatus we employ when we place things in the logical space of reasons is irreducible to any conceptual apparatus that does not serve to place things in the logical space of reasons. So the master thought [of EPM] as it were draws a line; above

126  Catherine Legg the line are placings in the logical space of reasons, and below it are characterizations that do not do that. (McDowell 2009, 5) Now our question concerning whether there is any nonconceptual content can be put by asking: Does anything lie ‘below the line’ in Sellars’s account of perception? And if so, what role does it play? McDowell begins by noting that Sellars seems to straightforwardly acknowledge some ‘below the line’ contribution to perception towards the end of EPM (X onwards), where he vindicates a role for ‘inner episodes’ in order to distinguish himself from behaviorists such as Skinner. But what contribution, exactly, is made by these inner episodes? One might suppose that it is some kind of phenomenology that renders (for instance) my perception that Redding’s tie is blue characteristically visual. But McDowell argues that the visual blueness of Redding’s tie occurs ‘above the line’ (McDowell 2009, 14). What lies above the line is not “phenomenologically colourless.” So why posit anything below the line? We noted earlier that Sellars admits that “the tie between ‘red’ and red physical objects . . . is causally mediated by sensations of red.” McDowell considers the possibility that such posited sensations are needed to scientifically explain similarities between blue things and blue hallucinations (McDowell 2009, 15). But why should these ‘below the line’ entities be understood as sensations? Why not go straight to a physiological story to explain blue hallucinations? McDowell remarks, “The sensations look like idle wheels” (McDowell 2009, 16). McDowell goes on to suggest that in Sellars’s later work, sensations are posited as less of a scientific explanation and more of a transcendental argument establishing that our knowledge is objective through “having the world in view.” Sensations are “directed towards showing our entitlement to conceive subjective occurrences as possessing objective purport” (McDowell 2009, 17). In this view, sensations are not, to use Kantian language, apperceived (i.e., ‘themselves visible’ in perception). Nevertheless it is helpful to say that they perform a “guiding function” for “the flow of one’s conceptual representations” (McDowell 2009, 22). What does this mean? Using a probably too-crude illustration, one might say that when one directly perceives a cat, one experiences a ‘flow of cat-related concepts’ (whiskers, tail, fur and so forth) which together fall into a certain overall ‘shape’, which constitutes a non-concept-involving, and therefore below the line, impression of a cat. Let us pause and note that what is described here might be understood as a form of indexicality. This is an interesting thought which will be addressed again shortly. For McDowell, though, this Sellarsian ‘guiding’ and ‘shaping’ which cannot be apperceived and is in principle unconceptualizable is just too mysterian and unhelpful. (He even claims that, ironically, it itself is redolent of the Myth of the Given.) He closes his volume with a chapter designed to finally sort out the Myth. Here he claims that “intuitional

Peirce and Sellars on Nonconceptual Content  127 content” is still conceptual (and thus above the line) although it is “not discursive content at all” (McDowell 2009, 270). The model for this is “having something in view”—for instance a red cube. McDowell claims that having a red cube in view “is complete in itself” (McDowell 2009, 270). It is not, as it stands, a judgment in propositional form (such as ‘This cube is red’). Nevertheless, it draws on ‘conceptual capacities’—our understanding of redness and cubehood, and to that extent it can be understood as having “judgment-shaped contents” when viewed in the appropriate light—that is, by a language user from within their language (McDowell 2009, 35). Yet at the same time as McDowell claims that the perception draws on conceptual capacities, he claims that the concepts themselves do not ‘figure’ in the experience itself. He claims that if he and another person view a particular bird which McDowell knows to be a cardinal and the other does not, it is not the case that the concept ‘cardinal’ features in his perception and not the other person’s. As both persons have the same bird in view, their experience has the same intuitive content (McDowell 2009, 259). This complex, multimodal view is how McDowell explicates Kant’s difficult notion of an ‘intuition.’ He claims that this undermines Sellars’s transcendental argument which posits ‘manifolds of sheer receptivity’ (below the line) to secure objectivity in our knowledge by providing an ‘external constraint’ on our thinking. McDowell claims that, by contrast, in his understanding of perception the only external constraint required is “objects themselves . . . becoming immediately present” (McDowell 2009, 39), along with whatever judgment-shaped properties users of relevant languages are able to recognize. To sum up, then, McDowell argues against Sellars that sense-experience and understanding are importantly different, and yet in perception nothing occurs below the line and there is no nonconceptual content. This determination has seemed to a number of philosophers to be disappointingly ad hoc in its separation (albeit somewhat equivocal and confusing) of perception from judgment, where one had hoped for an account which would explain the seamless continuity between them. What does it mean to claim that a perception is conceptual if it is not at all discursive?5 If having a red cube in view does not in itself involve judging that the cube is red, how is it that we do come to judge that the cube is red?6 There is also arguably an opportunity missed here to develop a coordinated causal and rational account of the contribution of sensation to perception—which is precisely what we shall now explore.

Peirce’s Mature Theory of Perception Interestingly, Peirce may also be understood as offering a ‘two-ply’ theory of perception. He also separates immediate experience, which he calls the percept, from truth-apt propositions derived from that experience, which

128  Catherine Legg he calls the perceptual judgment. Peirce’s realm of perceptual judgment thus corresponds nicely to Sellars’s space of reasons. Moreover, Peirce takes a position on the autonomy of the space of reasons which is very Sellarsian, while arguing against an antecedent of the empiricism of Carnap and Ayer which he finds in the philosophy of John Venn: Mr. Venn belongs to a school which considers the logical process as starting at the percepts, if not at impressions of sense. . . . But I maintain that logical criticism cannot go behind . . . the first judgments which we make concerning percepts. (Peirce CP, 7.198) Peirce’s stated reason for this claim, though, focuses less on the possibility of justification than criticism. He writes that criticism never pertains to a perception taken on its own, but only in juxtaposition with further facts, which requires that the whole be propositionally formulated in order to generate a contradiction: I look at an object and think that it seems white. That is my judgment of the object perceived . . . [it] does not, in itself, call for any explanation. On the contrary, it can only do that when it has been connected with other facts which taken by themselves would justify an expectation of the contrary of this fact. For example, if we should find that this object which seemed white, in the first place was white, and then that it was a crow, and finally that all the crows known were black. (Peirce CP, 7.198) Peirce notes that by contrast the percept, as it is outside of the space of reasons, strictly cannot be described, in a manner interestingly reminiscent of Travis (2004): I recognize that there is a percept or flow of percepts very different from anything I can describe or think. What precisely that is I cannot even tell myself. . . . I am forced to content myself not with the fleeting percepts, but with the crude and possibly erroneous thoughts, or selfinformations, of what the percepts were. (Peirce CP, 2.141) The percept cannot be described as it is singular. It “is a single event happening hic et nunc,” which “cannot be generalized without losing its essential character” (Peirce CP, 2.146). To put the same point another way, the percept cannot be described as it is fully determinate. Peirce at one point likens the perceptual judgment to “the printed letters in a book, where a Madonna of Murillo is described,” and the percept to the exquisite picture itself (Peirce CP, 5.54). Here we see Peirce articulating Redding’s insight

Peirce and Sellars on Nonconceptual Content  129 that perception differs from mere de dicto judgment in the ‘indescribable’ quantity of information it presents. (This similarity is not surprising since both thinkers share deep Kantian roots.) At this point, though, one might wonder: why does Peirce’s percept not grievously recreate the Myth of the Given,7 if it is some kind of nonepistemic faculty that is posited at the root of all perception? But we shall see that the percept does not stand in a justificatory relation to the perceptual judgment. So how do the percept and the perceptual judgment relate to one another, as they must if Peirce is to present a coherent theory of perception? Not in the way that might seem intuitively obvious—and the way that empiricism has traditionally supposed—according to which the perceptual judgment describes or copies the percept (as for instance British Empiricist Hume claimed that all our ideas form as copies of impressions).8 Peirce also explains why the percept cannot justify the perceptual judgment: as it is not in propositional form (not even close), it cannot serve as any kind of premise from which the perceptual judgment is concluded. Instead, Peirce claims, “there is no relation between the predicate of the perceptual judgment and the sensational element of the percept, except forceful connections” (Peirce CP, 7.634). In other words: the role of the perceptual judgement is solely to index the percept.9 This simply means that the percept causes the perceptual judgment whilst not providing any of its content. We earlier saw McDowell consider and reject the idea of positing something that performs a ‘guiding function’ for a flow of conceptual representations in perception, while not being itself apperceived. Peirce’s percept performs precisely this role. So although the percept was aligned above with ‘immediate experience,’ this should not be understood as immediate experience as sense-data, but immediate experience as direct contact with an object. Thus, this is a form of Reidian direct realism (which in contemporary philosophy of perception is now frequently dubbed relationalism).10 Perception is a process that takes place across time, though, and over time our constant causal triggers from percept to perceptual judgment gradually become enmeshed in an ever more smooth and predictable network of habits of association between certain kinds of percepts and certain kinds of perceptual judgments. How this occurs is best explicated in the broader framework of Peirce’s pragmatist theory of meaning, whereby belief consists in nothing but habits of expecting certain experiences in certain circumstances, and acting to bring about desired future experiences in that light. In short, our evolutionary legacy ensures that from birth our percepts produce “direct and uncontrollable interpretations” (Peirce CP, 7.648). These are trained, guided and corrected by our language-community until we produce perceptual judgments that are appropriate to our circumstances. For this educative process, the nature of the percepts lying behind these judgments matters not, as only the latter are publically accessible to our teachers.11 Crucially, this repetition of (noninferential) causal triggers until they become (inferential) habits of association is how concepts are built in

130  Catherine Legg Peirce’s philosophy. There is no other (dualist) ‘mind-stuff’ from which concepts are made.12 This is the great elegance of Peirce’s view, and its prospects for underpinning a thoroughly naturalistic philosophy of mind. One further feature of Peirce’s account of how percept and perceptual judgment combine requires remarking on before I turn to Sellars. Despite their great difference from one another, the presence of stable habits of expectation and action enables the two to interlock in a process of mutual modification which is entirely two-way. Peirce puts it like this: I have no means whatever of criticizing, correcting or recomparing [percepts], except that I can collect new perceptual facts relating to new percepts, and on that basis may infer that there must have been some error in the former reports, or on the other hand I may in this way persuade myself that the former reports were true. (Peirce CP, 2.141) This means that strictly speaking there is no way that the percept is, in and of itself (or ‘in view’). The percept only exists as interpreted—and thus potentially reinterpreted—by the perceptual judgment:13 Now let us take up the perceptual judgment ‘This wafer looks red.’ It takes some time to write this sentence, to utter it, or even to think it. It must refer to the state of the percept at the time that it, the judgment, began to be made. But the judgment does not exist until it is completely made. It thus only refers to a memory of the past; and all memory is possibly fallible and subject to criticism and control. The judgment, then, can only mean that so far as the character of the percept can ever be ascertained, it will be ascertained that the wafer looked red. (Peirce CP, 5.544) Here Peirce differs crucially from McDowell. We saw that McDowell claimed that where two people view a cardinal, one of whom has the concept ‘cardinal’ and the other not, they can have the very same experience. In this sense, McDowell retains a preconceptual ‘space of experience,’ but for Peirce there is no such space. In that sense McDowell’s ‘line’ metaphor is potentially misleading: more on this as follows.

Sellars Viewed in the Light of Peirce Peirce’s Reidean direct realism, which captures the de re dimension of perception highlighted by Redding, is arguably the best way of making sense of what Sellars means when he speaks in EPM of sensations performing a ‘guiding function’ for a flow of conceptual representations while not themselves being apperceived. In a later extended discussion of phenomenalism,

Peirce and Sellars on Nonconceptual Content  131 Sellars is even more explicit about this. In a section entitled “Direct Realism: Causal vs Epistemic Mediation,” he writes: the direct perception of physical objects is mediated by the occurrence of sense impressions which latter are, in themselves, thoroughly noncognitive. Furthermore, this mediation is causal rather than epistemic. Sense impressions do not mediate by virtue of being known. (Sellars 1963, 91) Ironically this kind of direct realism sounds a great deal like how McDowell says he wants to understand “having the world in view”—solely as “objects themselves . . . becoming immediately present” (McDowell 2009, 39). How McDowell managed to reject Sellars’s proffered indexicality by confusing it with a kind of transcendental phenomenology is difficult to understand.14 I would guess that a significant role is here played by a certain ambiguity in the technical philosophical term impression: between an ‘experience’ (phenomenological) of a sensation and an ‘experience’ (transactional or indexical) of an object.15 Sellars is arguably thinking of the latter, yet McDowell cannot help reading the former. This ambiguity—and the resulting confusion generated in philosophy—goes back to the early modern philosophy’s ‘veil of ideas’16—indeed is arguably created by it. Sellars also echoes Peirce’s focus on habit. He shows keen awareness of a diachronic dimension to perception, and makes it crucial to his attack on the Myth of the Given, when he criticizes classical empiricists as follows: they have taken givennness to be a fact that presupposes no learning, no forming of associations, no setting up of stimulus-response connections. (Sellars 1997, §6, 20) A third point of confluence between Sellars and Peirce arguably consists in some form of critical-common-sensism. In the discussion of phenomenalism cited above, Sellars not only claimed that the mediation of perception by sensations is causal not epistemic, he also claimed that realizing this constitutes a crucial step “towards an adequately critical direct realism” (Sellars 1963, 90).17 Why? I would suggest, precisely to capture the two-way correction between percept and perceptual judgment we saw Peirce elaborating earlier. Ironically, by removing sensory experiences from the role of justifying our judgments, and rendering them instead preconceptual indices to a world independent of our minds, they become a much more powerful scaffold for self-correction as we go about our business of trying to know reality. This, I think, is the essence of the account of rationality that Sellars puts forward against the foundationalism that is still widely pervasive in mainstream philosophy:

132  Catherine Legg 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. (Sellars 1997, §38, 79) With so much commonality now established between Sellars and Peirce, it is worth considering whether there are any important differences. One feature of Sellars’s thought that comes instantly to mind is his psychological nominalism, whereas Peirce famously vows to oppose nominalism in all its forms. Yet Sellars defines psychological nominalism as the view that, “all awareness of sorts, resemblances, facts etc. . . . is a linguistic affair” (Sellars 1997, §29, 63)—in other words, general properties consist in ‘mere words’ (and their uses). In an interesting sense of this claim, I don’t think that Peirce would disagree.18 I hope to investigate this matter more fully in a future study. A second potential major disagreement concerns the ‘naturalizing of normativity.’ We have noted that Peirce presents an elegant platform for this in holding that concepts, with their rich network of normative forces, are formed through the repetition of (noninferential) causal triggers until they become (inferential) habits of association. Sellars, on the other hand, famously separates the world’s normative manifest image from a scientific image that he envisages to be purely descriptive. Yet whether such a separation is what Sellars most wants, or rather all that he feels he is able to have given the state of development of his own philosophy, is an interesting question.19 Although the separation is reinforced in McDowell’s strong dualism between ‘first’ and ‘second’ nature, and Brandom’s Hegelian counterposing of Natur and Geist, the fact that there is an ongoing dispute between this ‘left-wing’ camp of Sellars adherents, and ‘right-wingers’ such as Ruth Milikan and the Churchlands who emphasize Sellars’s unflinching scientific realism, suggests that possibilities lie germinal in Sellars’s texts for a more thoroughgoing and integrated naturalism.

Conclusion The fact that we are beings who perceive means that our thinking is constrained by a world which, to the degree that it constrains us, it makes sense to describe as ‘external.’ As philosophers, it is tempting to wonder at what point in our encounter with this world are concepts ‘put onto’ what we experience? But we have now learned that just because preconceptual experience is part of the explanatory story of perception, doesn’t mean it should be reified into a discrete component of perception (that is itself apperceived). Just this insight arguably constitutes a significant step forward from mainstream empiricism, and much current debate around the existence or otherwise of ‘nonconceptual content.’ Yet arguably all the Pittsburgh School players largely agree on this point—so what is the difference between them?

Peirce and Sellars on Nonconceptual Content  133 We’ve seen that Brandom holds a two-ply theory of perception according to which we apperceive not only concepts whose meaning is fully inferential, but also noninferential (mere stimulus-response) sensory ‘reports.’ In this way, unlike Peirce, he attempts to put both strands ‘above the line.’ But by assigning such sensory reports their own, sui generis, explicitly noninferential category within the semantic space he arguably compromises his inferentialism. By contrast, McDowell claims that perception is conceptual through and through, and to that degree arguably manages to achieve a hyperinferentialism. Yet he feels obliged to draw a distinction between ‘intuitional’ and ‘discursive’ content, in order to postulate (in the form of intuitional content) a brute synchronic encounter between mind and world—a ‘having the world in view’—as the basis for all our knowledge, just as Kant taught. In one of the many ironic twists and turns which characterize these debates, I would claim that this posited brute synchronic encounter resembles sufficiently what Sellars called a “direct or intrinsic characterization of immediate experience” (Sellars 1997, §25, 57) to constitute yet another face of the Myth of the Given. I would like to venture the same point in another way, as follows. Inferentialism should mean more than just that concepts are somehow present (‘sitting there’!) in perception. (Is the concept cardinal ‘in’ one’s perception of a cardinal, or is it not?) Please forgive me for remarking that that seems a very representationalist view of the inferential. Let us recall that concepts are the basic ingredients not of physical space but of a space of reasons. As such, their role—the whole point of their presence, if you will—is to reason: to be always poised to correct and reinterpret any previous thoughts, according to the logical relationships between them. So to the extent that McDowell is able to think thoughts about the cardinal qua cardinal that his companion is not, his perception of it is not the same. This essentially dynamic quality of the conceptual is, I would argue, the vital insight that classical pragmatism can bring to these debates.20 To sum up McDowell, then, his fundamental claim that he sees as distinguishing himself from Sellars—that in perception ‘intuitional content’ lies ‘above the line’—is a mixture of incorrect and misconceived. First, if we understand intuition not as qualitative and (in that sense) phenomenological, but as indexical, and indexicality as preconceptual (as I have argued that we should), then we must see ‘intuitional content’ as a contradiction in terms. Second—and even more profoundly—we must cast a critical eye on ‘the line.’ This metaphor was used by Sellars himself, and is in some ways extremely helpful in enabling us to discuss what can be part of a space of reasons and what can not. Yet at the same time it is arguably not helpful insofar as it suggests that there might be a space ‘containing’ reasons, and a space ‘containing’ only something else. The connection between the conceptual and the indexical in our thinking is much more intimate than that.21 In

134  Catherine Legg fact, in Peirce’s thought we have seen how it is possible to explain concepts as themselves created from congeries of indices that are sufficiently repetitious to fall into intelligible patterns. We have seen that both Sellars and Peirce managed to combine both a rare direct realism and a diachronic account of perception such that there is no direct or intrinsic characterization of experience—all perception is interpretative. The combination of these two claims has frequently been viewed as contradictory. But it is not. As Haack clearly explains: Perception is interpretative. But it is neither necessary nor desirable to take this to mean that perception is not of things and events around us, but of images. . . . It is both possible and plausible to take it as meaning, rather, that though perception is of the things around us, our perceptual beliefs involve explanatory filling-in of the often limited information afforded us by our perceptual interactions with those things. (Haack 1994, 32)22 Seeing astutely through this apparent paradox enabled Sellars to replace foundationalism with a critical common-sensism still largely untried in mainstream epistemology. But although Sellars was a devastating critic of the dominant empiricism of his day, his positive account is arguably not as clearly or thoroughly worked out as Peirce’s account of meaning in terms of habit, embedded in a broader pragmatist theory of meaning (which was in turn embedded in an ambitious general theory of signs, but that is another story)—to show how hyper-inferentialism can be operationalized.

Notes 1 See, for instance, Koopman (2007), Levine (2012). A related critique, though not from a self-avowed pragmatist, is McDowell (2005). 2 For this point I am grateful to Dave Beisecker. 3 Although intriguingly he suggests that it is also a problem for a putatively later Wittgensteinian understanding of language in terms of rule-following that is triggered by circumstances of rule-application that are themselves understood to be ‘given’ (Sellars 1997, §30, 64–5). 4 I am grateful to Yuri Cath for pressing me on this point. 5 It should be noted that Paul Redding makes an interesting attempt to explicate this in terms of Hegel’s distinctive theory of judgment in Redding (2017). 6 For me, the matter is not fully clarified by McDowell claiming that when he judges that there is a red cube in front of him, “a conceptual capacity corresponding to ‘red’ and a conceptual capacity corresponding to ‘cube’ have to be exercised with a togetherness corresponding to the togetherness of ‘red’ and ‘cube’; in ‘There is a red cube in front of me’,” and that this provides, “a partial specification of the function that gives unity to the various representations in an ostensible seeing that there is a red cube in front of me” (McDowell 2009, 31). What exactly is this ‘function’? Which part of it is here being specified? And what is the other part?

Peirce and Sellars on Nonconceptual Content  135 7 For first raising this point to me in 2014, and for other comments which improved this paper, I am grateful to Willem DeVries. 8 Could the relationship be rightly viewed as a ‘noninferential report’, in Brandom’s sense? In my view this is an interesting question, worthy of a separate study. 9 One of the first Peirce scholars to clearly highlight this key point was Susan Haack—see, for instance, Haack (1994). 10 For an interesting recent paper tracing links between these debates, see Sant’Anna (forthcoming). 11 In this regard Forman (2007) gives a nice account of how although McDowell dismisses sensations as ‘idle wheels’ for explaining concurrent perception of an object, he misses the way in which they form a necessary condition for acquiring empirical concepts in the first place. 12 For a brilliant unpicking of how this works, and some of its profound consequences for philosophy of mind and epistemology, see Massecar (2012). 13 This point is extremely well explicated in (Rosenthal 2001). Similar revisionary potential has been posited in human agency, whereby narrativity confers upon our actions a ‘trajectory-dependent property.’ For instance if X happens, then it is true that I was in love with M at a given time, whereas if Y happens, it is false (Jones 2008, 271–5). Given that Peirce’s is a pragmatist theory of perception, this convergence is arguably no coincidence. 14 As I understand it, essentially the same point is made in DeVries (2006). See also DeVries and Coates (2009), Macbeth (2009). 15 Levine puts the same point well by stating that in Sellars’s view, “the cues that sensation provides are structural instead of qualitative” (Levine 2007, 65). Interestingly, this structural dimension to sensory causes arguably bestows on them an iconic, over and above the indexical function that has been the focus of this paper. I thank Vera Saller for pointing this out to me. It is difficult to say much explicitly about this iconicity as the percept is so inarticulable, but once again resonances with Sellars appear in that he famously (somewhat controversially) talked of ‘picturing’—albeit between tokens of linguistic expressions and the world (and thus ‘above the line’). 16 This ambiguity was noted in the last section where it was remarked that ‘immediate experience’ may be understood both as ‘sense-data’ and as ‘direct contact with an object.’ 17 See also Sellars (1982). 18 For instance, Peirce writes, “the external world . . . does not consist of existent objects merely, nor merely of these and their reactions; but on the contrary, its most important reals have the mode of being of what the nominalist calls ‘mere’ words, that is, general types and would—bes. The nominalist is right in saying that they are substantially of the nature of words; but his ‘mere’ reveals a complete misunderstanding of what our everyday world consists of” (Peirce CP, 8.191). 19 Explored to some degree in (Olen 2016). 20 I endeavored to argue this in more depth at the end of Legg (2017). 21 At some points in Having the World in View McDowell does gesture towards understanding this, e.g., in Chapter 2, “The Logical Form of an Intuition,” but I would argue that he doesn’t thoroughly integrate it. 22 Levine (2007) also ably defends the compatibility of these two positions, with specific reference to Sellars’s philosophy of perception, although I take issue with his interpretation of Sellars’s sensations as able to be ‘led into’ the space of reasons.

136  Catherine Legg

References Bermúdez, José Luis. 1995. ‘Nonconceptual Content: From Perceptual Experience to Subpersonal Computational States’. Mind and Language, 10, 4: 333–69. Brandom, Robert B. 1994. Making It Explicit. Cambridge, MA: Harvard University Press. Brandom, Robert B. 2000. Articulating Reasons: An Introduction to Inferentialism. Cambridge, MA: Harvard University Press. Brandom, Robert B. 2002. ‘The Centrality of Sellars’ Two—Ply Account of Observation to the Arguments of “Empiricism and the Philosophy of Mind” ’. In Tales of the Mighty Dead, 523–53. Cambridge, MA: Harvard University Press. Brandom, Robert B. 2007. ‘Inferentialism and Some of Its Challenges’. Philosophy and Phenomenological Research, 74, 3: 651–76. DeVries, Willem A. 2006. ‘McDowell, Sellars, and Sense Impressions’. European Journal of Philosophy, 14, 2: 182–201. DeVries, Willem A. and Coates, Paul. 2009. ‘Brandom’s Two—Ply Error’. In Empiricism, Perceptual Knowledge, Normativity and Realism: Essays on the Anniversary of “Empiricism and the Philosophy of Mind”. Edited by W.A. deVries, 131–45. Oxford: Oxford University Press. Forman, David. 2007. ‘Learning and the Necessity of Non—Conceptual Content in Sellars’s “Empiricism and the Philosophy of Mind” ’. In The Self—Correcting Enterprise: Essays on Wilfred Sellars. Edited by Michael P. Wolf and Mark Norris Lance, 115–45. Poznan Studies in the Philosophy of the Sciences and the Humanities. Vol. 92. New York: Rodopi. Haack, Susan. 1994. ‘How the Critical Common—Sensist Sees Things’. Histoire Épistémologie Langage, 16, 1: 9–34. Jackson, Frank. 1982. ‘Epiphenomenal Qualia’. Philosophical Quarterly, 32, 127: 127–36. Jones, Karen. 2008. ‘How to Change the Past’. In Practical Identity and Narrative Agency. Edited by C. Mackenzie and K. Atkins, 269–88. London: Routledge. Koopman, Colin. 2007. ‘Language is a Form of Experience: Reconciling Classical Pragmatism and Neopragmatism’. Transactions of the Charles S. Peirce Society, 43, 4: 694–727. Legg, Catherine. 2008. ‘Making It Explicit and Clear: From “Strong” to “Hyper—” Inferentialism in Brandom and Peirce’. Metaphilosophy, 39, 1: 105–23. Legg, Catherine. 2014. ‘ “Things Unreasonably Compulsory”: A Peircean Challenge to a Human Theory of Perception: Particularly with Respect to Perceiving Necessary Truths’. Cognitio, 15, 1: 89–112. Legg, Catherine. 2017. ‘Idealism Operationalized: How Peirce’s Pragmatism Can Help Explicate and Motivate the Possibly Surprising Idea of Reality as Representational’. In From Icons to Logic: Peirce on Perception and Reasoning. Edited by K. Hull and R.K. Atkins, 40–53. London: Routledge. Levine, Steven. 2007. ‘Sellars’ Critical Direct Realism 1’. International Journal of Philosophical Studies, 15, 1: 53–76. Levine, Steven. 2012. ‘Brandom’s Pragmatism’. Transactions of the Charles S. Peirce Society, 48, 2: 125–40. Macbeth, Danielle. 2009. ‘Un’antinomia del giudizio empirico: Brandom e McDowell’. In Lo Spazio Sociale della Ragione: Da Hegel in Avanti. Edited by L. Ruggiu and I. Testa, 343–60. Milan: Mimesis.

Peirce and Sellars on Nonconceptual Content  137 Massecar, Aaron. 2012. ‘Peirce’s Interesting Associations’. Transactions of the Charles S. Peirce Society: A Quarterly Journal in American Philosophy, 48, 2: 191–208. McDowell, John. 1994. ‘The Content of Perceptual Experience’. The Philosophical Quarterly, 44, 175: 190–205. McDowell, John. 1994. Mind and World. Cambridge, MA: Harvard University Press. McDowell, John. 1997. ‘Brandom on Representation and Inference’. Philosophy and Phenomenological Research, 57, 1: 157–62. McDowell, John. 2005. ‘Motivating Inferentialism: Comments on Making It Explicit (Ch. 2)’. Pragmatics and Cognition, 13, 1: 121–40. McDowell, John. 2009. Having the World in View: Essays on Kant, Hegel, and Sellars. Cambridge, MA: Harvard University Press.Olen, Peter. 2016. Wilfrid Sellars and the Foundations of Normativity. London: Palgrave Macmillan. Peacocke, Christopher. 1992. A Study of Concepts. Cambridge, MA: MIT Press. Peacocke, Christopher. 1998. ‘Nonconceptual Content Defended’. Philosophy and Phenomenological Research, 58, 2: 381–8. Redding, Paul. 2014. ‘An Hegelian Solution to a Tangle of Problems Facing Brandom’s Analytic Pragmatism’. British Journal for the History of Philosophy, 23, 4: 657–80. Redding, Paul. 2017. ‘Hegel and Sellars “Myth of Jones”: Can Sellars Have More in Common with Hegel Than Rorty and Brandom Suggest?’ In Wilfrid Sellars, Idealism and Realism. Edited by P.J. Reider, 41–58. London: Bloomsbury. Rosenthal, Sandra. 2001. ‘The Percipuum and the Issue of Foundations’. www. digitalpeirce.fee.unicamp.br/perros.htm. Sant’Anna, André. Forthcoming. ‘Perception Pragmatized: A Pragmatic Reconciliation of Representationalism and Relationalism’. Philosophia. Sellars, Wilfrid (EPM). 1956/1997. Empiricism and the Philosophy of Mind. With an Introduction by Richard Rorty and a Study Guide by Robert Brandom. Cambridge, MA: Harvard University Press. Sellars, Wilfrid (SPR). 1963. Science, Perception and Reality. London: Routledge. Sellars, Wilfrid. 1982. ‘Sensa or Sensings: Reflections on the Ontology of Perception’. Philosophical Studies, 41, 1: 83–114. Speaks, Jeff. 2005. ‘Is There a Problem About Nonconceptual Content?’ The Philosophical Review, 114, 3: 359–98. Travis, Charles S. 2004. ‘The Silence of the Senses’. Mind, 113, 449: 57–94.

7 Sellars and Frege on Concepts and Laws Danielle Macbeth

Although Sellars often had occasion to refer to Frege and Fregean themes in his writings, only in ‘Grammar and Existence’ (1960) does Sellars engage at any length with a Fregean thesis. The thesis is that a (Fregean) concept is, though not an object, nevertheless an objective entity, the Bedeutung of a concept word. This thesis, Sellars aims to show, is quite wrong. What matters for meaning in the case of a concept word is not reference, Bedeutung, but instead the functional role of the word. There is, according to Sellars, no such thing as a concept as conceived by Frege. Interestingly, this divergence in the views of Sellars and Frege is coupled with what I will argue is a deep convergence on the nature and role of inference in cognition. Both Sellars and Frege hold that, as Sellars puts it in a famous title of one of his early essays (CIL), concepts involve laws and are inconceivable without them. And such laws are, for both, fully objective, discovered in the course of scientific inquiry. Why, then, do Sellars and Frege diverge on the question of the Bedeutung, the reference or signification, of concept words?

1.  The Rationality of Inquiry Sellars and Frege are centrally concerned with scientific inquiry, the striving for truth; and according to both we can and do achieve fully objective knowledge in the course of such inquiry. There is an important difference: whereas Frege focuses exclusively on logic and the practice of mathematics, Sellars is primarily interested in the practice of the empirical sciences. Nevertheless, they are agreed that scientific inquiry, properly speaking, is answerable to reality, to how things actually are however we take them to be. What they need to explain is how it is that we are answerable to things as they are, how it is that the reality on which thought aims to bear can rationally constrain the judgments we make. Knowledge is answerable to things as they are. It can seem natural, then, to hold that we are, in our empirical inquiries, answerable to the tribunal of experience, that such inquiry is rational just if it has a foundation in the testimony of our senses. Sellars rejects the view. Although he does think that empirical knowledge has a foundation in what he calls observation reports,

Sellars and Frege on Concepts and Laws  139 it is not in virtue of having such a foundation that empirical inquiry is rational. Empirical inquiry is rational, Sellars holds, “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). Nothing that we think we know is unrevisable in principle. Indeed, it is just this, we will see, that is the lesson of Sellars’s critique of the Given. To a first approximation, the Myth of the Given is the idea that something nonconceptual can serve as a reason, or more generally, have normative significance for thought. Sense-datum theories are an obvious case in point insofar as, according to such theories, our empirical knowledge is grounded in sensings caused in us, and in other animals and prelinguistic infants, by the impacts of things in the environment on our sense organs. Such sensings, or their contents, are nonconceptual and yet also, in us, a form of knowing or at least cognitively relevant to knowing. Nothing, Sellars argues, can play such a role, can be at once merely causal, not conceptually articulated, and cognitively significant. What is cognitively significant is and must be conceptually articulated. But sense-datum theories are only one episode in the myth, and Sellars’s critique of such theories only “a first step in a general critique of the entire framework of givenness” (EPM §1). And this matters insofar as one might otherwise think that what is at issue here is only conceptual articulation, not also revisability in principle. In fact, we will see, what it means to be conceptually articulated, according to Sellars, just is to be revisable in principle. This is indicated already by the fact that Sellars offers as other episodes in the Myth “intuited first principles” and “synthetic necessary connections” (EPM §1). Such items are clearly to be taken to be at once conceptually articulated and unrevisable in principle. But again, nothing could possibly be both. What is unrevisable in principle is what one is merely caused to think, and what one is caused to think is not and cannot be an item of knowledge. The rationality of inquiry lies in self-correction, our capacity to call anything we think into question as reason sees fit. Perhaps it will be objected that if absolutely anything we think is subject to critically reflective inquiry, if there is no (absolute) ground on which the edifice of empirical knowledge is built, no (absolute) foundation for empirical knowledge, then we are left, at best, with a merely coherent set of belief, not empirical knowledge at all. In fact, it is just this either/or, either foundationalism or coherentism, that Sellars rejects. And he rejects it because it assumes a mistakenly static conception of knowing: “Above all, the picture [‘of human knowledge as resting on a level of propositions . . . which do not rest on other propositions in the same way as other propositions rest on them’] is misleading because of its static character” (EPM §38). It is the dynamic character of inquiry, the fact that it is constitutively a process of discovery and self-correction that grounds our capacity for knowledge. That absolutely anything we think we know can turn out to have been mistaken is not a barrier to knowledge of things as they are but instead its sine qua

140  Danielle Macbeth non. However things actually are, we can through the course of our ongoing critically reflective inquiries achieve knowledge of them as they are. And Frege agrees. “The goal of scientific endeavor is truth. . . . What is true is true independent of our recognizing it as such. We can make mistakes” (PW, 3). “Where truth is concerned there is the possibility of error” (PW, 132). Our ability to get things right is and must be an ability also to get things wrong. Indeed, Frege thinks, even logic is a science, a domain of knowledge, and that means that even in logic we can make mistakes. Even the law of identity, that a = a, which would seem to be unquestionably true, might one day be shown to be false, in need of revision or outright rejection. As Frege puts the point, this impossibility of our rejecting the law in question [the law of identity, a = a] hinders us not at all in supposing beings who do reject it; where it hinders us is in supposing that these beings are right in so doing, it hinders us in having doubts whether we or they are right. (GG, 15) We have, at this moment in history, no reason at all to doubt the law of identity. But as Frege says, we can imagine beings who have such doubts, which is to say that we can imagine ourselves coming to have such doubts. We cannot now imagine what such doubts would be; if we could so imagine, we would have those doubts. We do not doubt the law of identity, but we cannot know that we will not one day come to doubt it. After all, Aristotle held that subalternation—the rule licensing the move from ‘all S is P’ to ‘Some S is P’ and from ‘no S is P’ to ‘some S is not P’—is a valid rule of logic, though we do not. And intuitionists deny that the rule of double negation elimination is valid. One can have doubts, even in logic; the fact that no one now has doubts about the law of identity is not enough to show that such doubts will not arise. Even in logic there is no in principle unrevisability. The rationality of inquiry, the striving for truth, is intelligible only with the recognition that anything we think we know may turn out to have been mistaken. Our capacity to get things right just is the capacity to find out that we had been wrong. Frege and Sellars are agreed that, as Sellars puts it, the rationality of inquiry lies in self-correction. It follows, as Sellars notes, that “human discourse is discourse for finding things out as well as for expressing, textbook style, what we already know” (CDCM, 250). But what sort of discourse is discourse for finding things out? According to both Sellars and Frege it is discourse that enables the expression of laws governing what is a reason for what, what may be inferred in light of this or that claim. If inquiry is to be possible, it is not enough that we have various habits of response, (Humean) dispositions to make claims in light of other claims. Such dispositions must be made explicit in judgments that can be defended and appealed to in defense, that is, explicitly endorsed or rejected just as claims about what is the case can be.1

Sellars and Frege on Concepts and Laws  141 To find out that one had been wrong is not merely to have a change of mind, to have thought that p and now think instead that not-p. Suppose, for example, that I habitually conclude regarding things that are F that they are also G, perhaps because always in the past things I have found to be F have also been experienced to be G. Now I experience an F-thing that is not, my experience tells me, G. Given that my inferring that F-things are G was merely habitual, my propensity so to infer will be correspondingly weakened. I had thought that Fs are invariably G, but no longer. My mind, more exactly, my mind’s habit, has been changed by the ongoing course of my experience; it has been caused so to change. Such changes are, Sellars holds, to be distinguished from reasoned changes of mind, where a reasoned change of mind requires the capacity to express what are otherwise merely habits of inference as rules licensing inference. If I have and know myself to have a rule of inference to the effect that being F entails being G, then on experiencing something to be (apparently) F but not G I am faced with a contradiction. If the thing is F then I can infer that it is G, appearances notwithstanding. If it is not G, as my experience suggests, then I can infer that it is (appearances notwithstanding) not-F. It is also possible, of course, that my rule that being F entails being G is not after all a good rule. I am faced with an inconsistent triad; reason demands that something must go. The capacity to have second thoughts, reasoned changes of mind as such changes contrast with changes of mind that are merely caused in one, is grounded in this way in one’s language having the resources needed in the expression of laws governing inferences as well as resources for expressing claims about what is the case. In the empirical case of concern to Sellars, such laws take the form of material rules of inference, rules such as that being a dog entails being a mammal. Standard logic does not recognize such (counterfactual-supporting) laws, material rules of inference, but only rules of logic (which are understood, pace Frege, as strictly formal, without content or truth)—as well as, of course, empirical claims about what is in fact the case. Sellars argues that language is unintelligible without them: “Material rules are as essential to meaning (and hence to language and thought) as formal rules, contributing the architectural detail of its structure within the flying buttresses of logical form” (IM, 7). It is only because the expressions in terms of which we describe objects, even such basic expressions as words for the perceptible characteristics of molar objects, locate those objects in a space of implications, that they describe at all, rather than merely label. The descriptive and the explanatory resources of language advance hand in hand; and to abandon the search for explanation is to abandon the attempt to improve language period. (CDCM, 306–7) To describe rather than merely label requires just those capacities that enable explanation, that is, material rules of inference that track, or purport

142  Danielle Macbeth to track, causal and other necessary relations. And it does so because in describing one is answerable to how things are however one takes them to be. It is intelligible that one is describing, then, only if one is responsive to how things are, answerable to the norm of truth. It must be in one’s power to get things right, and hence in one’s power to correct one’s errors; and this requires in turn that there be material as well as formal rules of inference because only so can there be conflict and second thoughts. “The presence in the object language of the causal modalities . . . serves . . . to provide the framework for the thinking by which we reason our way . . . into the making of new commitments and the abandonment of the old” (CDCM, 302–3). As Sellars also argues, laws expressing inferential relations cannot be understood to be different from empirical generalizations (quantified conditionals) only in being necessary, true not just in the actual world but in all possible worlds. Rather, they have the form of inference licenses that permit or enjoin commitment to particular descriptive claims given that one is committed to others. Although they can be the contents of acts of judgment, laws have the form of rules rather than of claims. Frege, despite being concerned with logic and mathematics rather than the empirical sciences, agrees that language, as it matters to inquiry, needs to be able to express laws governing inferences. And it does so because such inquiry is constitutively systematic. “The first requirement—knowledge of [one’s] own ignorance” entails “the ideal of a system of mathematics” (PW, 221). “Science comes to fruition only in a system. . . . Only through a system can we achieve complete clarity and order” (PW, 242). In the introduction to Grundgesetze, Frege describes this “ideal of a strictly scientific method in mathematics”: It cannot be demanded that everything be proved, because that is impossible; but we can require that all propositions used without proof be expressly declared as such, so that we can see distinctly what the whole structure rests upon. After that we must try to diminish the number of these primitive laws as far as possible, by proving everything that can be proved. Furthermore, I demand—and in this I go beyond Euclid—that all methods of inference employed be specified in advance; otherwise we cannot be certain of satisfying the first requirement. (GG, 2) Science requires that methods of inference, including strictly logical methods, be specified in advance, made articulate in judgments. It follows, as I argue in Frege’s Logic (Macbeth 2005), that generality in Frege’s logic is not to be understood quantificationally. Generalized conditionals of Frege’s concept-script have the form of (counterfactual-supporting) statements of entailment. Because logic, for Frege, is a science, in Frege’s system even formal rules of inference are not merely formal, devoid of all content. They are grounded in the meanings of the logical particles and hence are to be

Sellars and Frege on Concepts and Laws  143 understood on the model of what Sellars thinks of as material rules of inference. An axiomatization of the laws of logic makes explicit the meanings, as far as one understands them, of the primitive signs of the system. If a contradiction is derived on the basis of those axioms then one knows that there is something one has not yet understood regarding one’s primitive signs, that second thoughts are in order. Frege claims at the end of the introduction to Grundgesetze that no one will be able either to build “a better more durable edifice . . . upon other fundamental convictions” or to show that his “principles lead to manifestly false conclusions” (GG, 25). He is clearly very confident that he has gotten things right. Nevertheless, Frege thinks, the real test of one’s convictions lies in what follows from one’s fundamental convictions and principles. Again, if a contradiction is derived then we know that something is amiss in the starting points. This is just what happened to Frege. Frege included Basic Law V as an axiom of his Grundgesetze system, and on that basis Russell derived a contradiction. Thus, although the discovery was a crushing personal blow to Frege, intellectually he saw it as promising “a great advance in logic” (PMC, 132). Russell’s paradox showed Frege that something he had thought he knew he did not know, where this is, again, “the first prerequisite for learning anything . . . the knowledge that we do not know” (GL, iii). The capacity to make explicit in an axiomatization one’s conception of the primitive concepts of logic such as the conditional and negation, their significance for inference, is by the same token the capacity to have second thoughts about just what those concepts actually mean, the inferences they properly license. In mathematics, it is explicit definitions that play this role. Definitions of mathematical concepts are stipulations introducing some new simple sign, the definiendum, and assigning it a meaning (Bedeutung) that is given in the definiens by way of an inferentially articulated sense (Sinn)—as we will soon see in relation to some examples. Because the definiens is formulated using the same signs as are employed in the axioms of logic, theorems can be proved on the basis of definitions by employing the primitive and derived rules of logic. And here again, one can discover that something one thought one knew needs to be rethought. If one’s definition leads to a contradiction, it is mistaken. According to Frege, “almost all errors made in inference . . . have their roots in the imperfections of the concepts” (PW, 34)—that is, in our conceptions of those concepts. (Fregean concepts cannot have imperfections.) And we discover the imperfections in our conceptions of concepts by making explicit, in axioms and definitions, our understanding of their contents, and reasoning on the basis of those axioms and definitions. In the case of defined concepts in particular, “if the content is not just indicated but is constructed out of its constituents by means of the same logical signs as are used in the computation [that is, in inference] . . . the computation must quickly bring to light any flaw in the concept formation” (PW, 35). And this content is itself inferentially articulated according to Frege: concepts, at least those

144  Danielle Macbeth that are mathematical, contain as their contents inference licenses, rules of material inference. Consider, for example, the concept prime number. Although we tend to think of the content of this concept in terms of what is the case if it applies, the necessary and sufficient conditions of its application, Frege thinks of it instead as involving an inference license. To say that a number is prime is to issue an inference license to the effect that if one is given any other number (save the number one) then it can be inferred that that number will not divide the original number without remainder. Just the same is true of Frege’s own definitions, of being hereditary, of following in a sequence, and of being a many-one procedure, as provided in Part III of Begriffsschrift. In each case, as Frege explicitly notes in his natural language paraphrases, the definiens takes the form of an inference license setting out what can be inferred given that some condition is satisfied. To judge, for example, that the property F is hereditary in the f-sequence is, by definition, to judge that “from the proposition that d has the property F, whatever d may be, it can always be inferred that each result of an application of the procedure f to d has the property F” (BGS §24; emphasis added). We have seen that for both Frege and Sellars the rationality of inquiry lies in self-correction, the recognition that anything one takes oneself to know may turn out to have been mistaken, and that this requires in turn that we recognize material as well as formal (that is, strictly logical) rules of inference. Concepts, according to both, involve just such rules. To be conceptually articulated just is to be inferentially articulated, which is to be revisable in principle. Nevertheless, we have seen, there is a notable difference between Frege and Sellars on the inferential articulation of non-logical concepts. Because his concern is with the science of mathematics, Frege focuses on the internal inferential articulation of concepts, the inferential contents of concepts as they can be displayed in explicit definitions. For Sellars, by contrast, inferential articulation is primarily between concepts; it involves rules of inference such as that being a dog entails being a mammal.

2.  Sellarsian Conceptual Meaning and Fregean Sense We have seen that for both Frege and Sellars, each in his own way, concepts involve laws and are inconceivable without them. Without laws there is no self-correction and without self-correction there is no scientific inquiry. It is furthermore the case for both Sellars and Frege that such inferential articulation contrasts in some fundamental way with what is to be found in objective reality. Although the inferential articulation of a concept is something objective in that one can get it right and can get it wrong, it is not itself an objective entity. Again we begin with Sellars and the case of empirical inquiry, with what Sellars calls conceptual meaning as it contrasts with applications of concepts to things, particulars in the empirical world. What Sellars calls conceptual meaning is exhausted by material and formal rules of inference: “The conceptual meaning of a descriptive term is

Sellars and Frege on Concepts and Laws  145 constituted by what can be inferred from it in accordance with the logical and extra-logical rules of inference of the language (conceptual frame) to which it belongs” (ITSA, 317). Although language use, and hence meaning in a broad sense, involves also language entry and language exit transitions, it is only the inferential moves that contribute to what Sellars thinks of as conceptual meaning. Sellars’s grounds for this thesis are clear: although there is an obvious sense in which a rule of inference, whether material or formal, is a rule, something that sets out what is to be done in certain specified circumstances, language entry transitions and language exit transitions cannot in the same sense be governed by rules. [To characterize] the learning to use a language or system of concepts as learning to use symbols in accordance with two types of rules: (a) rules of syntax, relating symbols to other symbols; (b) semantical rules, whereby basic factual terms acquire extra-linguistic meaning . . . involves a radical mistake. A rule is always a rule for doing something in some circumstance. Obeying a rule entails recognizing that a circumstance is one to which the rule applies. If there were such a thing as a semantical rule by the adoption of which a descriptive term acquires meaning, it would presumably be of the form ‘Red objects are to be designated by the word “red”.’ But to recognize the circumstances in which this rule has application, one must already have the concept of red! (ITSA, 312) We have already seen that for Sellars it is not enough that one have habits of response to claims; one needs to be able to express rules of inference in judgments so as to appeal to them as rules and also if need be subject them to critically reflective scrutiny. The rules governing applications of concepts cannot in the same way be expressed as rules. Although it is true that, other things being equal, an English speaker ought, on Sellars’s view, respond to, say, red things with an utterance of the word ‘red’, this ought is what he calls an ought-to-be as contrasted with a rule properly speaking, which is an ought-to-do.2 English speakers ought to be such as to be able to tell red things when they see them, and they are so trained in the course of their entry into the language. But this, as Sellars sees, is very different from the fact that such speakers ought to infer, given that a thing is, say, a dog, that it is also a mammal, and indeed ought to be able to express this rule as a claim, as the claim that being a dog entails being a mammal. In the latter case, although we begin with a mere habit of response, an ought-to-be, we must also learn to affirm the inference rule explicitly. Such a rule so affirmed has the form of an ought-to-do. Despite the fact that an English speaker ought to be disposed to utter, say, the word ‘red’ in the presence of red things, Sellars denies that a word such as ‘red’ has empirical meaning in virtue of “its role as a conditioned response to red things” (ITSA, 314). The fact that a word such as ‘red’

146  Danielle Macbeth ought to be applied to red things has no role at all to play in the content of a concept as Sellars understands it. The familiar notion (Kantian in its origins . . .) that the form of a concept is determined by ‘logical rules’, while the content is ‘derived from experience’ embodies a radical misinterpretation of the manner in which the ‘manifold of sense’ contributes to the shaping of the conceptual apparatus applied to the manifold in the process of cognition. . . . There is nothing to a conceptual apparatus that isn’t determined by its rules, and there is no such thing as choosing those rules to conform to antecedently apprehended universals and connexions, for the ‘apprehension of universals and connexions’ is already the use of a conceptual frame, and as such presupposes the rules in question. (IM, 26) Again, there is a kind of rule associated with one’s disposition to utter, say, ‘red’ in the presence of red things, but it is not, according to Sellars, a rule in the same sense as that in which an inference rule, whether material or formal, is a rule. As the point is put in ‘Some Reflections on Language Games’ (SRLG, 334), whereas intralinguistic moves are inferences according to syntactical rules, whether material or formal, language entries are instead observations. Conceptual meaning, as Sellars understands it, is given by the material and formal inferences that a descriptive term is involved in; what is correctly described using the term is no part of meaning in this sense. It is illuminating to compare this distinction that Sellars draws, between the conceptual meaning of a term and that to which it is correctly applied, with the distinction Frege draws between the sense (Sinn) of an expression and its designation or meaning (Bedeutung). The Fregean sense of an expression, which is what is grasped in one’s having the use of a word, is its inferential content, what follows if it applies; the meaning or designation is that which is referred to, if anything, by the word. To take a very familiar example, the two expressions ‘the morning star’ and ‘the evening star’ designate the same entity, Venus, but they express different senses. One can, for example, infer from the fact that something is the morning star, though not from the fact that it is the evening star, that it is seen in the morning. Thus, while it is true that the morning star is (identical to) the evening star, one and the same thing, nevertheless, it may take extensive astronomical investigations to realize this. The two definite descriptions, despite having the same designation, Bedeutung, express different Fregean senses. Similarly, it is in virtue of the inferential content of the expression ‘the largest prime number’ on the basis of which to draw inferences, that we can discover, that is, prove, that this definite description has no designation or meaning, that there is nothing that is the largest prime number. Thus, for both Sellars and Frege, in addition to the thing referred to (Frege) or correctly described (Sellars)

Sellars and Frege on Concepts and Laws  147 by a word, there is also its conceptual meaning (Sellars), its sense (Frege), determining what follows, the inferential consequences of the use of the word. For both, the meaning of a sentence is given not only by what is the case if it is true but also by what follows if it is true. I have suggested that what Sellars calls conceptual meaning, that is, the formal and material rules of inference that for him constitute the contents of concepts, can be usefully compared to what Frege thinks of as the sense of a concept word. But as already noted, whereas for Frege, those contents are internally inferentially articulated, for Sellars, inferential articulation is primarily between concepts. Thus, whereas for Frege, it is meaningful to say of one concept in particular that we have or have not adequately grasped its sense, on Sellars’s understanding it is the whole inferentially articulated language that is the smallest unit of account. According to Sellars, “there are an indefinite number of possible conceptual structures (languages) or systems of formal and material rules, each one of which can be regarded as a candidate for adoption by the animal which recognizes rules, no one of which has an intuitable hallmark of royalty” (IM, 26). It is furthermore clear that Sellars thinks that it is a future language of science that is the language we need in order to describe and explain things as they actually are.3 Again, because, according to Frege, concepts are internally inferentially articulated, it is intelligible that we might achieve clarity regarding one concept in particular. In his Habilitationsschrift (1874, included in Frege CP), Frege provides this example of quantity: According to the old conception, length appears as something material which fills the straight line between its endpoints and at the same time prevents another thing from penetrating into its space by its rigidity. . . . The introduction of negative quantities made a dent in this conception, and imaginary quantities made it completely impossible. Now all that matters is the point of origin and the endpoint; whether there is a continuous line between them, and if so which, appears to make no difference whatever; the idea of filling space has been completely lost. All that has remained is certain general properties of addition, which now emerge as the essential characteristic marks of quantity. The concept has gradually freed itself from intuition and made itself independent. This is quite unobjectionable, especially since its earlier intuitive character was at bottom mere appearance. (CP, 56) We need, then, to distinguish between a concept and our knowledge of that concept, between “the logical and objective order” and “the psychological and historical order” (CP, 136). A logical concept does not develop and it does not have a history. . . . If we said instead ‘history of attempts to grasp a concept’ or ‘history of the

148  Danielle Macbeth grasp of the concept’, this would seem to me much more to the point: for a concept is something objective: we do not form it, nor does it form itself in us, but we seek to grasp it, and in the end we hope to have grasped it, though we may have mistakenly been looking for something when there was nothing. (CP, 133) Because all our awareness of what is objective is mediated by a sense (Sinn) according to Frege, we can think that we have grasped a concept (in its pure form) when we have not. But equally, we can, in time, come to grasp a concept (in its pure form) through achieving an adequate conception of the sense through which it is given. Both Frege and Sellars take inferentially articulated senses to be objective, something about which we can be right or wrong, without being objective entities. And according to both, it can take real intellectual work to achieve knowledge of the material and formal rules of inference that are needed if we are to achieve knowledge of things as they are in themselves. Just as Frege says, [O]ften it is only after immense intellectual effort, which may have continued over centuries, that humanity at last succeeds in achieving knowledge of a concept in its pure form, in stripping off the irrelevant accretions which veil it from the eyes of the mind. (GL, vii) And clearly Sellars would agree, at least with the general idea: only through our ongoing scientific inquiries can we achieve the conceptual structure, the language, that is needed adequately to address things as they are. Nevertheless, as we know, Frege and Sellars fundamentally disagree insofar as Frege holds that concept words not only express senses but also designate objective entities of some sort, whereas Sellars does not. According to Sellars, concepts are nothing more than functional roles; they just are Fregean senses, albeit holistically conceived. Sellars’s oft-repeated account of sentences of the form “ ‘s’ (in language L) means . . .” provides a useful illustration of the central idea. The surface grammar of a sentence of the form “ ‘s’ (in language L) means . . .” suggests that logically the sentence has the form “xRy” on the basis of which one might naturally conclude that “meaning is a relation between a word and an nonverbal entity” and “that the relation in question is that of association” (EPM §31). In fact, Sellars argues, appealing now to the particular case of the sentence “[In German] ‘rot’ means red,” the rubric “ ‘. . .’ means . . .” is a linguistic device for conveying the information that a mentioned word, in this case “rot”, plays the same role in a certain linguistic economy, in the case the linguistic economy

Sellars and Frege on Concepts and Laws  149 of German-speaking peoples, as does the word “red”, which is not mentioned but used—used in a unique way; exhibited, so to speak—and which occurs “on the right-hand side” of the semantic statement. (EPM §31) One who already has the use of the word ‘red’ in English—who has grasped its conceptual meaning, the formal and material inferences in which it figures, and can tell red things by looking—can learn in this way how the German word ‘rot’ is used, its role in German. No nonverbal entity need be invoked to account for sentences of the form “ ‘s’ (in language L) means . . .”. There is no thing that is the meaning, Bedeutung, of a concept word according to Sellars. There is only its role, its place as a node in the whole inferentially articulated structure that is a language. Although, according to Sellars, concept words, that is, predicates, have inferentially articulated meanings and can be applied in judgments, they do not designate anything objective. It follows, as Sellars argues in “Naming and Saying,” that a truly perspicuous language such as Sellars’s Jumblese would have no property or relation words but only names. By writing names in various ways—for instance, in certain spatial relations, or particular fonts, sizes, or colors—one thereby shows, in Jumblese, how things stand with those objects: “The fact that two objects stand in a dyadic relation would be represented by making their names stand in a dyadic relation” (NS, 228). For Sellars, concepts are nodes in inferential webs of material and formal inferential relations, nothing over and above their conceptual meanings. They are, in essence, what Frege teaches us to call senses.

3.  A Diagnosis Despite many differences of detail, Sellars and Frege are in fundamental agreement about the role of inference, and the material and formal laws that govern it, in our pursuit of knowledge of things as they are. And yet they disagree on the status of concepts, and thereby on the precise nature of the fruits of scientific inquiry. One might think that this divergence can be explained by the fact that Sellars focuses on the empirical sciences while Frege’s interest lies with mathematics. Certainly, there is something to this idea. But we will see, what is most significant is the fact that the mathematical practice Frege has in mind is essentially post-Kantian (in a sense to be explained), while the scientific practice Sellars is concerned with is instead Kantian. Frege’s concern is with mathematics as it came to be practiced over the course of the nineteenth century, that is, with the practice of deductive reasoning on the basis of the contents of concepts as that practice contrasts both with ancient Greek diagrammatic reasoning and with the sort of constructive algebraic problem solving that was the norm in mathematics through the seventeenth and eighteenth centuries. Sellars’s concern, by contrast, is with postulational science as that practice contrasts not only

150  Danielle Macbeth with ancient Greek and medieval correlational science but also (I will suggest) with twentieth-century fundamental physics. It is these features of the respective views of Sellars and Frege, we will see, that promises to account for their divergence on the status of concepts as the meaning or designation of concept words. We begin with a very brief overview of the contours of the history of the exact sciences. As everyone knows, European thought underwent a (scientific) revolution in the seventeenth century. What is less well known is that this revolution was followed by another. Indeed, as I show in Realizing Reason (Macbeth 2014), the intellectual history of the West has been marked by two waves of revolution, each involving first mathematics, then physics and finally philosophy.4 It is Descartes’ Geometry (1637) transforming the practice of mathematics that first sets things in motion: ancient Greek diagrammatic practice was henceforth to be replaced by the sort of constructive algebraic problem solving that Descartes pioneered and we all learn in school. Fifty years on, enabled by the new mathematics of Descartes and others following him, Newton’s Principia transformed the practice of physics. Ancient and medieval observational and correlational science gave way to theories postulating unobservable entities together with mathematically expressed laws governing their behaviors on the basis of which to explain observable phenomena.5 And a century after that, Kant’s Critique of Pure Reason (1781/87) transformed the practice of philosophy, replacing traditional logic and metaphysical speculation with, on the one hand, a purely formal general logic, and on the other, a transcendental logic that would uncover the conditions of possibility of our knowledge of things.6 The second wave was begun in the nineteenth century with the new mathematical practice of Riemann (among others) in which, constructive algebraic problem solving having been eschewed, it was to be by deductive, logical reasoning directly from the contents of concepts that mathematical discoveries were to be made.7 The fruits of this new practice enabled, in the twentieth century, the development both of Einstein’s theories of special and general relativity and of quantum mechanics, neither of which is postulational.8 In contemporary fundamental physics there is no physical correlate that the theory aims correctly to describe or model; there are only the mathematical structures.9 The case of philosophy is more complex. Although the seeds of the revolution in logic and philosophy that is needed in the wake of developments in mathematics and fundamental physics in the nineteenth and twentieth centuries were sown already by Frege at the end of the nineteenth century, Frege’s work was not understood. As a result, philosophy has not yet had its second revolution. Frege’s work was not understood. His logic in particular, although essentially post-Kantian in respects to be outlined, was taken to be merely an extension of Kant’s monadic predicate calculus to a full logic of relations. As a result, the fundamental character of the nineteenth-century revolution in mathematics was likewise misunderstood. Sellars, despite his profound,

Sellars and Frege on Concepts and Laws  151 and in the context very radical, insights into the role of material inferences in cognition and the need for a means of expressing laws in addition to quantified generalities, is heir to these misunderstandings. Sellars did not see what Frege, the first truly post-Kantian thinker, saw. And because he did not, Sellars remains a Kantian. It is just this, we will see, that can account for their deep disagreement over the status of concepts. We begin again with Descartes, in particular with the fact that, in light of his transformative work in mathematics, Descartes was able selfconsciously to invert the traditional order of existence and essence that we find in Aristotle.10 On Descartes’s new understanding, the ancient and scholastic idea that knowledge of existence is and must be prior to knowledge, if any, of essence is simply wrong. Instead, “according to the laws of true logic,” Descartes insists, “we must never ask about the existence of anything [never ask if it is, an est], until we first understand its essence [what it is, quid est]” (CSM II, 78). From being what is essential to some actually existing object, that in virtue of which it is what it is, its nature, essence is now to be understood as a purely mental entity, a meaning, or concept, or as Descartes thinks of it, an idea. Kant inherits this conception of concepts as prior to and independent of any relation to any object, of concepts as they contrast with what Kant calls intuitions, singular representations immediately of objects. From Kant’s perspective—which is possible at all only in the wake of Descartes’ groundbreaking work—a term logic such as Aristotle’s, in which all categorematic expressions have both a referring and a predicative function, conflates the logical function of referring that is played by intuitions with the logical function of predicating played by concepts. On a Kantian conception of logic and language, it is referring expressions, or their mental correlate intuitions, that give objects, and predicative expressions, Kantian concepts, that enable the characterization of given objects as thus and so. The problem, for Kant, is that intuitions without concepts are of no cognitive significance, and thus have no (direct) role in judgment. In Kant’s logic, it is quantifiers, and only quantifiers, that are the means of referring to objects.11 Sellars rejects this aspect of Kant’s view; for Sellars, although sensings are not cognitively significant, intuitions, singular representations of objects, can be.12 Sellars’s language Jumblese perspicuously displays the workings of just such a Kantian language, a language in which there are referring and predicative expressions conceived on the model of Kantian intuitions through which objects are given and Kantian concepts through which objects are thought. In Jumblese, it is manifest that predicate expressions do not designate, that they serve only to characterize objects otherwise given as thus and so. Sellars’s most fundamental conception of cognition is in this way essentially Kantian. Building on Descartes’s radically new conception of concepts, Kant teaches us to distinguish logically between intuitions through which alone objects are given and concepts through which (alone) given objects are thought. Now we need to draw another distinction, this time following

152  Danielle Macbeth Frege. We need to see that much as, from the Kantian perspective, ancient Aristotelian logic conflates referring and predicating in the notion of a term, so from Frege’s perspective, Kant’s division of representations into concepts, through which alone objects are thought, and intuitions, through which alone objects are given, conflates two different distinctions, that between the Sinn and the Bedeutung of an expression with that between Fregean concepts and objects. Whereas for Kant all cognitive significance, all being for a thinker, is through concepts, and all objectivity lies in relation to an object or objects—either as given in sense experience or merely with respect to their form in mathematics,Frege teaches us to distinguish, on the one hand, between cognitive significance (Fregean sense, Sinn) and concepts, which are the Bedeutung or signification of concept words, and on the other, between objective significance, Bedeutung, and objects.13 If Frege is right, Kant was after all wrong to think that all objective significance lies in relation to an object, wrong to think that all cognitive significance, all being for a thinker, is predicative through (Kantian) concepts. As Frege understands them, the two distinctions, that of (Fregean) concept and object and that of sense (Sinn) and meaning or signification (Bedeutung), seem to be orthogonal: all object names and all concept words both express Fregean senses and designate, or at least ought to designate, something objective, that is, an object or a Fregean concept. This seems, however, not to be quite right. Certainly it is true, as I have argued both Sellars and Frege clearly recognize, that the words of any language that can serve as a medium of our cognitive access to reality must express senses as well as designating something objective. The Sinn/Bedeutung distinction applies to language as such. But as the work of Sellars and Frege already indicates, the concept/object distinction seems to serve instead to distinguish two very different sorts of languages, everyday natural language, including its “sophisticated extension” (EPM §38) in the postulational sciences, on the one hand, and mathematical language as first fully actualized in the nineteenth century, on the other. The designation or signification of words in natural language and the postulational sciences are objects, particulars with their empirically discoverable properties and relations. The designations or significations of words of mathematical language are instead concepts, laws of correlation with their mathematically discoverable properties and relations. If so, Sellars’s Kantian perspective is in fact that of natural language (and the postulational sciences), in which case Sellars is right to deny that predicate expressions designate Fregean concepts. Such words do express Fregean senses, but what they refer to, if anything, are objects. But Frege is also right insofar as the concept words of a properly mathematical language designate (Fregean) concepts. Sellars was obviously deeply influenced by Kant; although he was not a Transcendental Idealist, his most fundamental conceptions were Kantian.14 Frege, we have seen, is not Kantian but instead essentially post-Kantian in distinguishing between, on the one hand, (Fregean) concepts and Fregean sense (Sinn), and on the other, objects and Fregean signification (Bedeutung).

Sellars and Frege on Concepts and Laws  153 And as Frege saw, it was developments in mathematical practice over the course of the nineteenth century that showed that Kant was wrong, that he was wrong not only about the practice of mathematics—wrong to think that mathematics involves an intuitive use of reason, the construction of concepts in pure intuition—but wrong also about logic. Pace Kant, even the deduction of a theorem, that is, purely logical reasoning directly from the contents of mathematical concepts as laid out in explicit definitions, can extend our knowledge.15 But if so, then Kant must be wrong to think that all objectivity lies in relation to an object. Hence, Frege comes to think, even concept words, paradigmatically, mathematical concept words, not only express senses but also signify something objective, concepts conceived as laws of correlation, arguments to truth-values.

4. Conclusion I have suggested that Frege and Sellars are both right both in regard to the thesis on which they are in agreement, that the inferential articulation of concepts is constitutive of our capacity for knowledge, and in regard to the thesis regarding the objectivity of concepts, on which they are in complete disagreement. They are both right about the first thesis because the rationality of inquiry lies in the capacity for self-correction that is made possible by, and only by, the inferential articulation of concepts. And they are both right regarding the second thesis, despite their disagreement, because the concepts Sellars is concerned with are utterly different from those Frege is concerned with. Sellars is right: natural language knows no Fregean concepts. Concepts, in our ordinary understanding, are rather what Frege teaches us are the senses expressed by words; they are what Sellars thinks of as conceptual meanings. Such (Kantian) concepts do not have sharp boundaries, and cannot be defined as mathematical concepts can. But Frege is also right: in mathematics, concepts are objective entities. They have sharp boundaries and except for a handful of primitives can be defined in terms of primitives of the language. We were able to understand how both could be right in this way because, as we saw, Frege and Sellars are concerned not only with different disciplines, mathematics and empirical science, respectively, but also with different historical moments in the ongoing unfolding of these disciplines. In the opening paragraph of his Locke Lectures, Sellars remarks that “philosophy without the history of philosophy is, if not empty or blind, at least dumb” (SM, 1). My reflections here have suggested that we need more than the history of philosophy. We need also the history of science, both of the empirical sciences and of mathematics. Only so can we understand not only how radically our conception of reality has changed over the last three millennia but also how mathematical language has become autonomous, a medium of cognitive access to things as they are. And it is only if we understand that that we can understand how Sellars and Frege might both be right not only about the laws that articulate our concepts but also about those concepts themselves.

154  Danielle Macbeth

Notes 1 Sellars writes: “Once the development of human language left the stage when linguistic changes had causes but not reasons, and man acquired the ability to reason about his reasons, then, and this is a logical point about having the ability to reason about reasons, his language came to permit the formulation of certain propositions which, incapable of proof or disproof by empirical methods, draw in the heart of language militant, a picture of language triumphant” (CDCM, 307). 2 Sellars introduces the distinction between ought-to-do’s and ought-to-be’s in ‘Language as Thought and Communication’ (LTC). 3 As Sellars famously puts it, “in the dimension of describing and explaining the world, science is the measure of all things, of what is that it is, and of what is not that it is not” (EPM §41). 4 Obviously it would be impossible to establish such a claim here. Interested readers are urged to consult Macbeth (2014). 5 This contrast between ancient and modern scientific practice is the basis for Sellars’s contrast between what he calls the manifest image (of man-in-theworld) and the scientific image. See ‘Philosophy and the Scientific Image of Man’ (PSIM). 6 See, for example, the discussion of Kant’s logic, both general and transcendental, in Heis (2012). 7 See Stein (1988) and Gray (2004) for preliminary discussion of this development, also Chapter 5 of Macbeth (2014). 8 See, for example, Friedman (2001) for the case of Einstein’s theories. 9 See Chapter 9 of Macbeth (2014). 10 See Secada (2000). 11 See Thompson (1972). 12 We need, then, to distinguish “between a special sub-class of conceptual representations of individuals which, though in some sense a function of receptivity, belong to a framework which is in no sense prior to but essentially includes general concepts, and a radically different kind of representation of an individual which belongs to sheer receptivity and is in no sense conceptual” (SM, 7). 13 Frege writes in some comments on his distinction between sense and meaning: “It is easy to become unclear about this by confounding the division into concepts and objects with the distinction between sense and meaning, so that we run together sense and concept on the one hand and meaning and object on the other. To every concept-word or proper name, there corresponds as a rule a sense and a meaning, as I use these words” (PW, 118). 14 In Macbeth (2000), I explore respects in which Sellars is, and is not, a Kantian. 15 Frege makes this point explicitly in relation to Kant in Grundlagen (GL), section 88. In Macbeth (2014), Chapters 7 and 8, I explain in detail how this works on Frege’s view.

References Descartes, René. 1637. The Geometry of René Descartes. Translated by David Eugene Smith and Marcia L. Latham. New York: Dover 1954. Descartes, René (CSM II). 1984. The Philosophical Writings of Descartes, vol. II. Translated by J. Cottingham, R. Stoothoff and D. Murdoch. Cambridge: Cambridge University Press. Frege, Gottlob (BGS). 1879. [Begriffsschrift] ‘Conceptual Notation: A Formula Language of Pure Thought Modelled upon the Formula Language of Arithmetic’. In

Sellars and Frege on Concepts and Laws  155 Conceptual Notation and Related Articles. Translated by T.W. Bynum. Oxford: Clarendon Press 1972. Frege, Gottlob (GL). 1884. [Grundlagen] Foundations of Arithmetic. Translated by J.L. Austin. Evanston, IL: Northwestern University Press, 1980. Frege, Gottlob (GG). 1893. [Grundgesetze] The Basic Laws of Arithmetic: Exposition of the System. Translated and Edited by Montgomery Firth. Berkeley and Los Angeles: University of California Press 1964. Frege, Gottlob (PW). 1979. Posthumous Writings. Edited by H. Hermes, F. Kambarel and F. Kaulbach. Translated by P. Long and R. White. Chicago: University of Chicago Press. Frege, Gottlob (PMC). 1980. Philosophical and Mathematical Correspondence. Edited by G. Gabriel, H. Hermes, F. Kambartel, C. Thiel and A. Veraart. Translated by Hans Kaal. Chicago: University of Chicago Press. Frege, Gottlob (CP). 1984. Collected Papers on Mathematics, Logic, and Philosophy. Edited by Brian McGuinness. Translated by M. Black, V. Dudman, P. Geach, H. Kaal, E-H. Kluge, B. McGuinness and R. Stoothoff. Oxford: Basil Blackwell. Friedman, Michael. 2001. Dynamics of Reason. Stanford: CSLI Publications. Gray, Jeremy. 2004. ‘Anxiety and Abstraction in Nineteenth-Century Mathematics’. Science in Context, 17: 23–47. Heis, Jeremy. 2012. ‘Attempts to Rethink Logic’. In The Cambridge History of Philosophy in the Nineteenth Century (1790−1870). Edited by Allen Wood and Songsuk Hahn. Cambridge: Cambridge University Press. Kant, Immanuel (CPR). 1781/87. Critique of Pure Reason. Translated by Paul Guyer and Allen Wood. Cambridge: Cambridge University Press 1998. Macbeth, Danielle. 2000. ‘Empirical Knowledge: Kantian Themes and Sellarsian Variations’. Philosophical Studies, 101: 113–42. Macbeth, Danielle. 2005. Frege’s Logic. Cambridge, MA: Harvard University Press. Macbeth, Danielle. 2014. Realizing Reason: A Narrative of Truth and Knowing. Oxford: Oxford University Press. Secada, Jorge. 2000. Cartesian Metaphysics: The Late Scholastic Origins of Modern Philosophy. Cambridge: Cambridge University Press. Sellars, Wilfrid (CIL). 1948. ‘Concepts as Involving Laws and Inconceivable Without Them’. Philosophy of Science, 15: 287–315. Sellars, Wilfrid (IM). 1953. ‘Inference and Meaning’. Mind, 62: 313–38. Reprinted in ISR. Sellars, Wilfrid (SRLG). 1954. “Some Reflections on Language Games”. Philosophy of Science, 21: 204–28. Reprinted, with additions, in SPR. Sellars, Wilfrid (EPM). 1956. ‘Empiricism and the Philosophy of Mind’. In Minnesota Studies in the Philosophy of Science, Vol. I. Edited by Herbert Feigl and Michael Scriven, 253–329. Minneapolis, MN: University of Minnesota Press. Reprinted, with additional footnotes, in SPR. Sellars, Wilfrid (CDCM). 1957. ‘Counterfactuals, Dispositions, and the Causal Modalities’. In Minnesota Studies in the Philosophy of Science, Vol. II. Edited by Herbert Feigl, Michael Scriven and Grover Maxwell, 225–308. Minneapolis, MN: University of Minnesota Press. Sellars, Wilfrid (ITSA). 1958. ‘Is There a Synthetic A Priori?’ In American Philosophers at Work. Edited by Sidney Hook, 135–59. New York: Criterion Press. Reprinted in SPR.

156  Danielle Macbeth Sellars, Wilfrid (GE). 1960. ‘Grammar and Existence: A Preface to Ontology’. Mind, 69: 499–533. Reprinted in SPR. Sellars, Wilfrid (PSIM). 1962. ‘Philosophy and the Scientific Image of Man’. In Frontiers of Science and Philosophy. Edited by Robert Colodny, 35–78. Pittsburgh, PA: University of Pittsburgh Press. Reprinted in SPR. Sellars, Wilfrid (NS). 1962. ‘Naming and Saying’. Philosophy of Science, 29: 7–26. Reprinted in SPR. Sellars, Wilfrid (SPR). 1963. Science, Perception, and Reality. London: Routledge & Kegan Paul. Sellars, Wilfrid (SM). 1968. Science and Metaphysics: Variations on Kantian Themes. The John Locke Lectures for 1965–66. London: Routledge & Kegan Paul. Sellars, Wilfrid (LTC). 1969. ‘Language as Thought and Communication’. Philosophy and Phenomenological Research, 29: 506–27. Reprinted in ISR. Sellars, Wilfrid (ISR). 2007. In the Space of Reasons: Selected Essays of Wilfrid Sellars. Edited by Kevin Scharp and Robert Brandom. Cambridge, MA: Harvard University Press. Stein, Howard. 1988. ‘Logos, Logic, and Logistiké: Some Philosophical Remarks on Nineteenth Century Transformations of Mathematics’. In History and Philosophy of Modern Mathematics. Edited by William Aspray and Philip Kitcher. Minnesota Studies in the Philosophy of Science, vol. XI, 238–59. Minneapolis, MN: University of Minnesota Press. Thompson, Manley. 1972. ‘Singular Terms and Intuitions in Kant’s Epistemology’. Review of Metaphysics, 26: 314–43.

8 “We pragmatists mourn Sellars as a Lost Leader” Sellars’s Pragmatist Distinction Between Signifying and Picturing Carl B. Sachs Introduction The title of this chapter is taken from Richard Rorty (1979a) in which he began to distance himself, via criticism of Jay Rosenberg (1974), from Wilfrid Sellars.1 Rorty was a committed Sellarsian for the first twenty years of his professional life (Rorty 2010, 8), and his criticisms of Sellars led him to the remarkable Aufhebung of Sellars and W.V.O. Quine of ‘Privileged Representations’, in Philosophy and the Mirror of Nature (1979b). Yet it was only a few years previous that Rorty ruefully remarked that he thought that Sellars would have, and should have been, a pragmatist, and yet failed to do so. Here I shall situate Sellars relative to the history of pragmatism, without simply assimilating Sellars to the pragmatist tradition, to demonstrate that Sellars’s relationship to pragmatism is far closer than is widely assumed and that Rorty’s rueful remark is misplaced. Sellars’s proximity to the pragmatist tradition has been underappreciated partly because pragmatists still pay too little attention to Clarence Irving Lewis. In his time Lewis was one of America’s most distinguished pragmatist epistemologists. Yet Lewis was sharply critical of his fellow pragmatist John Dewey. Accordingly, I shall begin with a short reconstruction of an underlying tension between Dewey and Lewis as to what pragmatism involves (§1) before turning to an explication of Sellars’s distinction between ‘signifying’ and ‘picturing’ (§2). There I shall argue that Sellars’s distinction can be understood as, among other things, a reconciliation between Dewey and Lewis. I will then conclude with a brief examination of the signifying/picturing distinction based on recent work by Jay Rosenberg, Johanna Seibt and Huw Price (§3). I shall argue that this distinction, best understood as a distinction between discourse (signifying) and cognition (picturing), is crucial for contemporary pragmatism (§4). One note before I begin: I will not address the relationship between pragmatism and what Sellars calls ‘the Myth of the Given.’ I agree with Richard Bernstein’s (2010) assessment that the rejection of the Given is already to be found in Charles S. Peirce (as well as William James and Dewey). Even Lewis, for all his talk of ‘the given,’ does not commit himself to the Myth of the Given.2 Any

158  Carl B. Sachs further discussion of the Myth of the Given in relation to pragmatism on my part would be redundant.

Lewis’s Anti-Deweyan Pragmatism To set the stage for Sellars’s contribution to pragmatism, I shall begin with an examination of Clarence Irving Lewis. In his intellectual autobiography Sellars remarks that “the highlight of that year (at least I think it was that year) was a seminar in C. I. Lewis’s Mind and the World Order led by John Austin and Isaiah Berlin” (Sellars AR, 287).3 Here I shall briefly sketch some prominent themes from Dewey’s version of pragmatism to indicate how Lewis offers a deeply non-Deweyan or anti-Deweyan pragmatism, and why this matters for the tensions that Sellars undertook to resolve. In contrast to the Platonic/Cartesian/Kantian axis, Dewey—building on the work of previous pragmatists—argued that human mindedness, in all its various permutations and elaborations, developed from forms of intelligent behavior in the animal world. What we can call culture—what Hegel called Geist or ‘spirit’—is not something set apart from ‘nature’ but rather is itself natural and developed from other kinds of natural intelligence.4 Dewey’s naturalized philosophical anthropology is richly informed by the psychology and biology of his day, including German psychophysics and Darwin’s revolution in evolutionary theory. For Dewey, philosophy is not a self-enclosed conversation amongst Great Minds but a critical reflection on the problems and issues of concern to society and to human flourishing. The Deweyan philosopher is a public intellectual engaged with the problems as she finds them in the society of which she is a member and citizen. Thus she cannot, as a philosopher, disengage herself either from politics or from science. In his ‘The Influence of Darwinism on Philosophy’ (MW 4), Dewey argues that Darwin’s revolution is directly relevant to how we think of what we are doing when we do philosophy. Since to be a living thing is to be an organism situated in an environment, and continually confronted with the problems immanent to the situations in which it finds itself, so that it must engage in some rudimentary problem-solving—whether by learning or by evolving—then this must be true of human beings as well. Hence we should think of human beings less as ‘rational animals’ and more as ‘clever beasts.’ The marks of human uniqueness—language, culture and technology—not only have antecedents elsewhere in nature but were shaped by the same general forces that shaped the webs of spiders, the dams of beavers and the songs of whales. Dewey’s comprehensive philosophical vision stems from reconstructing traditional concerns of epistemology, metaphysics, ethics, aesthetics and political philosophy from a vantage-point afforded by the growth of psychology and biology in the sciences and the experiments in democracy in the decades between the Civil War and the Cold War. Dewey does not, however, have the last word as to what counts as pragmatism. C. I. Lewis developed a version of pragmatism in his Mind and the

“We pragmatists mourn Sellars”  159 World Order (1929) (hereafter MWO), also indebted to Peirce and James, but strikingly different. Unlike Dewey, Lewis had little interest in the empirical sciences and had no political theory to speak of; also unlike Dewey, Lewis was primarily interested in problems arising in the development of mathematical logic; he wrote a textbook of symbolic logic and developed what he called ‘conceptual pragmatism’ by reflecting on the fact of plurality of logical systems. Conceptual pragmatism consists of (1) recognition of open-ended plurality of incompatible conceptual systems, including both natural and formal languages; (2) the absence of any definitive criterion for asserting that some conceptual system carves nature closer to the joints than its rivals; and hence (3) the role of human interests in deciding choice of conceptual system, which includes the trade-off between simplicity of notation and approximating patterns of human inference, and also, in the case of empirically meaningful language, the choice of categories that constrain how the given is interpreted. All this clearly puts Lewis in the pragmatist heritage, and he saw himself as the heir of Peirce, Royce and (to a lesser extent) James. Yet while he admired Dewey and agreed with him about much (Lewis 1930/1970), the differences are no less crucial. Though Lewis also thinks that epistemology must be reconstructed in response to new developments in human knowledge, he differs from Dewey as to what the most salient developments are: “Whoever has followed the developments in logistic and mathematical theory in the last quarter century [1900–25, CS] can hardly failed to be convinced that the consequences of these must be revolutionary” (Lewis 1929, vii). In other words, modern symbolic logic is to Lewis as evolutionary theory is to Dewey. Lewis shares Dewey’s commitment that epistemology should be engaged with and responsive to the most epistemically authoritative developments of one’s own time and place. Epistemological reflection cannot be independent from actual instances of our best examples of what counts as knowledge, whether in modern symbolic logic or general relativity. But although Lewis’s epistemology is informed by current developments in mathematical logic and fundamental physics, there is one important qualification: The reflective attitude is pragmatic in the same sense that it is empirical and analytic. It supposes that the categories and principles which it seeks must already be implicit in human experience and human attitude. . . . But the reflective method is not, or need not be, pragmatic in the sense of supposing, as current pragmatism sometimes seems to do, that the categories of biology and psychophysics have some peculiar advantage for the interpretation of the practical attitudes of thought. (Ibid., 34–5) It is difficult to avoid the supposition that the ‘current pragmatism’ to which Lewis alludes here is that of Dewey, whose Experience and Nature

160  Carl B. Sachs was first published four years before Mind and the World Order. Though Lewis agrees with Dewey that epistemology should be responsive to the sciences of the day, he strongly disagrees that biology and psychophysics have any specific priority, relative to the other sciences, for pragmatism. Why does Lewis take this position, and why is it relevant to understanding Sellars’s own relation to the history of pragmatism? To answer this I shall now turn to Lewis’s pragmatic theory of concepts. In keeping with previous pragmatists, but with perhaps greater clarity and attentiveness to the challenge posed by alternative conceptual frameworks in logic and in physics, Lewis emphasized the social nature of concepts: The coincidence of our fundamental criteria and principles is the combined result of the similarity of human animals, and of their primal interests, and the singularities of the experiences with which they have to deal. More explicitly, it represents one result of the interplay between these two; the coincidence of human modes of behavior, particularly when the interests which such behaviors serves involve coöperation. (Ibid., 20) More bluntly: “Our categories are guides to action” (ibid., 21), and “concepts and principles reveal themselves as instruments of interpretation; their meaning lies in the empirical consequences of the active attitude” (ibid., 31). This in turn licenses a version of verificationism about meaning: The concept is a definitive structure of meanings, which is what would verify completely the coincidence of two minds when they understood each other by the use of language. Such ideal community requires coincidence of a pattern of interrelated connotations, projected by and necessary to cooperative, purposeful behavior. . . . It is concepts . . . which must be implicitly present in our practice, which constitute the element of interpretation which underlies our common understanding of our common world. (Ibid., 89) In short, the function of conceptual meaning is to enable the coincidence of two (or more) minds that allows for cooperative behavior.5 Notice that although Lewis upholds verification about conceptual meaning, the verification lies in diachronic patterns of behavior and not (as for some logical positivists) in synchronic presentations of sense- data. Concepts must be public, Lewis thinks, because what is available to anyone through introspective awareness is just sensation and imagery. Since I cannot guarantee that my sensations and images are identical to those of another, it would be a triumph for skepticism if concepts were reducible to private mental phenomena. To account for conceptually structured discourse, Lewis explains concepts in terms of patterns of cooperative behavior rather than in terms of introspectively accessible mental phenomena.

“We pragmatists mourn Sellars”  161 Though he admits that no individual mind would find concepts meaningful if they could not be associated with sensations and images (ibid., 77), the meaning of the concepts themselves is not constituted by those introspectively available associations but by patterns of cooperative, purposive behavior.6 Throughout MWO Lewis rightly stresses the anticipatory character of knowledge: the function of concepts and categories is to anticipate future experiences and to prepare us to meet them. In this regard Lewis builds on the pragmatist theory of concepts developed by Peirce and above all by Royce.7 Yet he also does not think that the natural sciences have any specific relevance to epistemology; on the contrary, the reflective attitude undertaken by the pragmatist philosopher is wholly independent of all scientific theories. As Lewis sees it, while it certainly falls under the purview of science to determine the ratio essendi of things, it would be a mistake to conflate this project with the explication of the ratio cognoscendi: Epistemological investigation is, naturally, by way of the ratio cognoscendi: that is its peculiar task. Those ‘theories of knowledge’ which reverse the direction of explanation and give a causal, naturalscientific account, merely substitute a more or less uncritical and psychological methodology, based upon dubious assumptions, for their proper business. (Ibid., 426) Since the ratio essendi and ratio cognoscendi must be distinguished, then although science may have some privileged status with regard to the latter, it falls to philosophy alone to investigate the former.8 Though they will hopefully converge at the Peircean limit, for the here-and-now the intellectual vocations of philosophy and science are distinct. Thus, Lewis’s conceptualistic pragmatism is restricted to the explication of the ratio cognoscendi. The ratio essendi or order of being has nothing to do with it; nothing in biology or any other science is relevant to conceptualistic pragmatism. By contrast, Dewey’s own pragmatism is grounded in the organism-environment transaction; it is not much exaggeration to say that for Dewey, pragmatism is what philosophy ought to become in response to Darwinism (Popp 2007; Rogers 2008). Such a thought could not be further from Lewis’s Kantian pragmatism.

Sellars’s Distinction Between Signifying and Picturing Despite his ongoing criticisms of Lewis (and, following Lewis’s death, Firth), Sellars expressed an infrequent but consistent admiration for classical pragmatism; consider the following selected texts: 1. Now I would argue that Pragmatism, with its stress on language (or the conceptual) as an instrument, has had hold of a most important

162  Carl B. Sachs insight—an insight which, however, the pragmatist has tended to misconceive as an analysis of ‘means’ and ‘is true.’ On the other hand, if the pragmatist’s claim is reformulated as the thesis that the language we use has a much more intimate connection with conduct than we have yet suggested, and that this connection is intrinsic to its structure as language, rather than a ‘use’ to which it ‘happens’ to be put, then Pragmatism assumes its proper stature as a revolutionary step in Western philosophy (SRLG §33–4; emphasis added).9 2. It wasn’t until my thought began to crystallize that I really encountered Dewey and began to study him. . . . He caught me at a time when I was moving away from ‘the Myth of the Given’ (antecedent reality?) and rediscovering the coherence theory of meaning. Thus it was Dewey’s Idealistic background which intrigued me the most. I found similar themes in Royce and later in Peirce. I was astonished at what I had missed (NAO, 1; emphasis added). 3. Unless I am very much mistaken, the argument of this chapter [Chapter V–CS] also provides that missing ingredient, the absence of which from Peirce’s account of truth leaves the ‘would-be’ acceptance ‘in the long run’ of propositions by the scientific community without an intelligible foundation; a fact which has obscured the extent to which this gifted composer of variations on Kantian themes succeeded in giving metaphysics a truly scientific turn (SM, vii; emphasis added). These texts show that Sellars’s interest in pragmatism is not incidental to the larger strokes of his thought. At present I am most interested in Sellars’s suggestion that Peirce really had succeeded in the Kantian aim of giving metaphysics a truly scientific turn, but that his accomplishment was marred by the absence of a concept that Sellars undertook to provide. The concept is what Sellars, borrowing the term from the early Wittgenstein, calls ‘picturing,’ which must be sharply distinguished from what Sellars calls ‘signifying’ or ‘semantic assertability.’ Unfortunately, Sellarsian picturing is a difficult notion, and it is unsurprising that Rorty objected to it. Nevertheless, I shall attempt to show that picturing, correctly understood, is not (pace Rorty) a detraction from Sellars’s pragmatist credentials but a vital component of them. Put most generally, picturing is the ability to reliably track and respond to causal regularities in the environment. As such it is distinguished from what Sellars calls ‘signifying,’ or the ability to engage in intentional discourse and discursively structured thought. Sellars’s relation to the history of pragmatism can be most clearly seen if we see him as aligned with Dewey with regards to picturing and with Lewis with regards to signifying. Sellars introduces the signifying/picturing distinction in his 1960 ‘Being and Being Known’ (BBK) by distinguishing between two ways of talking about the mind-world relation:

“We pragmatists mourn Sellars”  163 There is an isomorphism in the real order between the developed intellect and the world, an isomorphism which is a necessary condition of the intellect’s intentionality as signifying the real order, but is to be sharply distinguished from the latter . . . a confusion between signifying and picturing is the root of the idea that the intellect as signifying the world is the intellect as informed in a unique (or immaterial) way by the natures of things in the real order. (BBK, 50; emphasis original) While there is something right about traditional philosophical conceptions of mindedness, nevertheless crucial distinctions must be made to understand the place of mind both in nature and as a part of nature. We need to understand what mindedness is not just ‘in the order of understanding’—the ratio cognoscendi of C. I. Lewis—but also what it is ‘in the order of being’—the ratio essendi that Lewis neglects. To understand the problem of mindedness as a part of nature, Sellars asks us to imagine a highly sophisticated robot: Suppose such an anthropoid robot to be ‘wired’ in such a way that it emits high frequency radiation which is reflected back in ways which project the structure of its environment (and its ‘body’). . . . Suppose such a robot to wander around the world, scanning its environment, recording its ‘observations’, enriching its tape with deductive and inductive ‘inferences’ from its ‘observations’ and guiding its ‘conduct’ by ‘practical syllogisms’ which apply its wired-in ‘resolutions’ to the circumstances in which it ‘finds itself’. It achieves an ever more adequate adjustment to its environment, and if we permitted ourselves to talk about it in human terms (ad we have been) we would say that it finds out more and more about the world, that it knows more and more facts about what took place and where it took place, some of which it observed, while inferred others from what did observe by the use of inductive generalization and deductive reasoning. (Ibid., 52–3; emphasis original) The analogy is not perfect; the robot is not literally carrying out observations and inferences, partly because it lacks sensations or sensory consciousness, but also because the robot is entirely ‘mechanistic’; it cannot hold itself accountable to the kinds of norms that govern our perception, reasoning and action. Despite this, the analogy suggests that just as there are patterns of interaction through which the robot can navigate its environment with increasing skill, an ability that can be explained in terms of how the robot represents its environment, so too our own intentionality has a ‘mechanistic’ explanation. How the intellect tracks the world can explained in terms of the causal

164  Carl B. Sachs relationship between computational functions and environmental features. The robot need only be equipped with a rudimentary system for detecting features of its environment and responding to that detection by modifying its own states, where the modification of the inner states produces changes in behavior. If it can do this, then it has everything it needs in order to count as representing the regularities it ‘observes’; hence the robot can implement activity recognizable to us as functionally analogous to our own cognitive activity. Thus, even though the intellect initially encounters itself as signifying or as discursively structured, that fact has little relevance for explaining how the intellect relates to the world in rerum natura. Minded animals, including ourselves, are not fundamentally different from a robot that “comes to contain an increasingly adequate and detailed picture of its environment in a sense of picture which is to be explained in terms of the logic of relations” (ibid., 53; emphasis original). To understand how this matters to human activity, we need only consider what the nascent cognitive sciences tell us about how brains process information: I submit that as belonging to the real order it [the intellect—CS] is the central nervous system, and that recent cybernetic theory throws light on the way in which cerebral patterns and dispositions picture the world . . . what we know directly as thoughts in terms of analogical concepts may in propria persona be neurophysiological states. (Ibid., 59; emphasis original) Picturing thus turns out to be a crucial part of the Sellarsian account of mind. If (as Sellars argues at length elsewhere) semantic notions such as ‘means’ or ‘refers to’ are strictly metalinguistic concepts that serve to indicate how a piece of language functions, then our best explanation of the relation between language and the world will be the work of cognitive neuroscience, not semantics or epistemology.10 Yet Sellars, like Lewis, is a social pragmatist about intentionality and conceptual meaning.11 For that very reason he realizes that the question of how the intellect relates to the world cannot be understood in terms of intentionality: If intentionality is a socio-linguistic affair, then how the intellect is related to the world is distinct from intentionality. We need a different kind of relation—an empirical relation between two relational systems in the natural world—to account for how the intellect relates to the world. We need picturing in the comprehensive account of the place of mind in the natural world precisely because intentionality is not a relation—that is, not a relation between mind and world.12 Put otherwise: given that the intellect is related to the world at all (which only a skeptic or solipsist would genuinely doubt), can the true nature of that relation be ascertained solely by reflecting on the manifest image? The Great Minds of ‘Western’ philosophy, from Plato to C. I. Lewis, would not hesitate to say ‘yes.’

“We pragmatists mourn Sellars”  165 Sellars objects to this venerable verity because the manifest image cannot explain itself without succumbing to the Myth of the Given. Though Sellars is sufficiently faithful to Aristotle, Kant, and C. I. Lewis (not to mention Strawson and Austin) to think that the manifest image explicates the world as we experience it, the manifest image cannot explain itself. We cannot look to the manifest image to understand why we happen to have the kind of manifest image that we do. And yet this problem must arise for any philosopher who acknowledges—as Lewis did but Aristotle and Kant did not—that it is a contingent fact of history that we do have the kind of manifest image that we do.13 The actuality of our contingent manifest image is an empirical fact to be explained. In this regard Sellars continues the Deweyan tradition of reflecting on the influence of Darwinism for philosophy. Since we must turn to scientific explanation to understand why we have the manifest image that we do, epistemology must incorporate the insights of evolutionary theory and cognitive science. To better situate Sellars’s philosophy of mind relative to the pragmatist tradition, we need to examine more closely the function of picturing: what is picturing for? Sellars suggests that we understand picturing in terms of the organism-environment relationship. Varying metaphors from picturing to mapping, Sellars observes that “the essential feature of the functioning of a map as, in a primary sense, a map is its location in the conceptual space of practical reasoning concerning getting around in an environment” (NAO, 109).14 This is, as I see it, crucial for understanding the affinity between Sellars and Dewey: the essential function of cognition in rerum natura is to guide purposive activity in response to detectable regularities. It is deeply unfortunate that Sellars’s choice of terminology, borrowed from the Tractatus, obscured the pragmatist dimension of Sellars’s thought, and produced the impression that we would have a more thoroughgoing pragmatism if picturing were abandoned. Pace Rorty, a pragmatist theory of cognition as picturing is precisely what we need in a comprehensive account of mindedness once we accept a social pragmatist account of intentionality per se.

Making Picturing Even More Pragmatic I now undertake a brief examination of how picturing has been interpreted by three contemporary philosophers: Jay Rosenberg, Johanna Seibt and Huw Price. Bringing them together shows how powerful picturing can be as an explanatory concept in philosophy of mind and why it is concept that pragmatists should re-examine. I will first turn to Rosenberg’s explanation of the epistemological status of the concept of picturing, followed by Seibt’s interpretation of picturing in terms of nonlinear dynamics, and concluding with Price’s use of picturing in relation to pragmatist philosophy of language and metaphysics. Recall that Rorty’s rejection of picturing, noted at the beginning, was aimed at Rosenberg’s (1974) defense of picturing, which Rosenberg subsequently rejected (1980). Much later, and shortly before his passing,

166  Carl B. Sachs Rosenberg (2007) revised his assessment by drawing on Sellars’s late work.15 Rosenberg notes that, for Sellars, “picturing is evidently the fundamental mode of correctness for any matter-of-factual representational system . . . a useful strategy might be to begin by considering his account of representational systems (RSs) too basic for the atomic-vs-molecular distinction to get a foothold in the first place, i.e., animal RSs” (108). In putting the point this way, Rosenberg stresses what Sellars means by an ‘animal representational system’: To be a representational state, a state of an organism must be the manifestation of a system of disposition and propensities by virtue of which the organism constructs maps of itself in its environment, and locates itself and its behavior on the map. (MEV, 292) For animal RSs do (and indeed must) picture; to be an ARS is to be, put formally, an embodied and embedded picturing relation, since it must picture both itself—that is, its body—as well as the environment in which it is embedded. Rosenberg thus realizes that picturing must as fully embodied and embedded as any philosophy of mind beholden to Heidegger or Merleau-Ponty. To clarify how pictures can represent, Rosenberg stresses how they are analogous to language: “A state will count as representational just in case—and because—it is suitably implicated in analogous to our language-entries, language-exits, and intra-linguistic moves” (Rosenberg 2007, 108). To be a representation is to be a state with a functional role in a system that must systematically coordinate how the environment impinges on the system with how the system should respond to the environment. An ARS does so by enabling primitive inferences by which of which it is able to track both its environment and also itself in relation to its environment. Thus, while we must not attribute full-blown first-person perspectives to non-discursive animals, nevertheless “Sellars explicitly makes some sort of self-awareness an indispensable condition of an animals RS” (ibid., 109). This requires only that some sub-animal cognitive module have the function of tracking the temporal and spatial location of the animal’s body relative to the features of the environment that are relevant to satisfying or thwarting the animal’s goals. (For example, if some module or modules were dedicated to comparing information about perceptual changes correlated with proprioceptive information with information about perceptual changes that was not correlated with proprioceptive information.) How, then, should an animal’s cognitive experience of its world be described in terms of picturing? On Rosenberg’s suggestion, we should understand

“We pragmatists mourn Sellars”  167 an animal’s total visual field at a given time as a single composite state functioning as a ‘pictorially’ complex representation of its then and there visual environment. Insofar as they are appropriately caught up in dispositions to (primitive) inferences and behavior, such sensory states function as highly complex Jumblese-style sentences. To put the point metaphorically, the world ‘speaks’ to organisms through their senses. (Ibid., 113; emphasis original)16 It should be emphasized that an animal’s pictorial representation of its then-and-there environment is rarely restricted to a single sensory modality (e.g., vision), since it integrates multimodal sensory information and proprioceptive information to form a highly complex map-like representation of the environment and the animal’s situation relative to that situation. Put otherwise, the picturing relation is the cognitive dimension of the organismenvironment transaction so important to Dewey. Importantly, Rosenberg now sees that picturing has (contra Rorty) no epistemological role: The strategy which I have here pursued divests picturing of immediate epistemological significance by interpreting it as a functional mode of representation, rather than as a mode of correctness. Being in the ‘picturing line of work’ is the determinative function of matter-of-factual representational systems per se, and so, trivially, the correctness of a basic matter-of-factual representation will be its correctness as a picture— but, crucially, the priority expressed here is conceptual and not epistemological. (126; emphasis original) Rather than think of picturing as an epistemological concept that plays a directly epistemological role, e.g., in terms of justification, we should think of picturing as playing an indirect role that respects the autonomy of epistemology while also doing something that epistemology alone cannot do: explain why we have the kinds of cognitive capacities and incapacities that we presume ourselves to have when we are doing epistemology. In short, the conceptual status of picturing does not belong to epistemic justification (hence respecting the autonomy of epistemology) but to empirical explanation. For this reason, Rorty is mistaken to remark that “Perhaps the gods see things otherwise. Perhaps they are amused by seeing us predicting better and better while picturing worse and worse” (Rorty 1991, 155). Picturing is not a substitute for a God’s-eye view, but a speculative anticipation of an empirical theory of human and nonhuman cognition. Whereas Rosenberg clarifies the conceptual status of picturing, Seibt (2009) clarifies its ontological role by embedding the idea of picturing within a scientific process metaphysics. Although Sellars has mostly been influential

168  Carl B. Sachs on “the ‘social pragmatist movement’ focused on the normative domain in the light of reason, leaving the causal issues to the purview of neuroscientists or behaviorists” (Seibt 2009, 249), this is at best half of the Sellarsian story. The other half, which concerns what social pragmatists happily leave to neuroscientists, is of profound importance for understanding mindedness. Seibt points out that if we think about picturing in strictly Tractarian terms, we will go awry: “Picturing is not an abstract relation but a certain type of nonlinear causal processing” (ibid., 249). This may seem like a bold claim, since there is nothing about ‘nonlinear causal processing’ in anything Sellars wrote. Yet Seibt argues that we can, armed with important discoveries in complexity theory, understand Sellars far better than he understood himself. The result is that we can come to see that [p]icturing is a relationship of causally founded coordination between two concrete collections of natural items. Items in one of these collections, so-called ‘natural linguistic objects,’ fulfill two additional constraints: first, these items must lend themselves to use as material embodiments of the ‘elementary’ (empirical) statements of a language game L; second, they must exhibit the kind of uniformities that are produced once that game is played. (Ibid., 252) Though this is helpful, there is a problem with assimilating all picturing representations to the non-semantic (picturing) functions of linguistic representations. This obscures both (i) how picturing functions in cognitive processes generally and (ii) how the advent of language transforms how the cognitive systems can picture their environments. Although Seibt allows that natural-linguistic objects may be “rhythmic patterns of an acoustic code, machine states of a Turing machine, or neurophysiological states” (ibid., 253), nevertheless machine codes and neurophysiological states cannot be “functional analogues of the observation statements of some natural language” (ibid.). Rather, we need to understand how such systems represent their environments, as required by our theories of computation and neurophysiology. There may be good empirical reasons to think that cognitive systems do represent features of their environments, but also there is no a priori reason to think that languages themselves are best understood in terms of such representations. We may yet conclude that social pragmatists are correct about the uselessness of the concept of representation for semantics of natural languages, and yet also conclude that the concept of representation is useful for cognitive neuroscience. Sellars’s distinction between signifying and picturing opens up the conceptual space for precisely this possibility. However, the social pragmatist theory of intentionality has also obscured a correct understanding of picturing. If the role of socio-linguistic norms is to institute inferential content, as argued at length by Brandom (1994,

“We pragmatists mourn Sellars”  169 2000), then how can there be picturing or mapping representations? How can there be representations that are non-linguistic and yet function as representations, which include the possibility of misrepresentation? Picturing requires norms of correct and incorrect representation which cannot be socio-linguistic, since they characterize cognition in non-linguistic animals. Seibt attempts to solve this problem by distinguishing picturing and signifying in terms of ‘low-grade’ and ‘high-grade’ normativity: the low-grade normativity of biological functions and the high-grade normativity of sociolinguistic deontic scorekeeping.17 Though I agree with Seibt that biological phenomena (including cognition) display a kind of normativity, ‘grades’ of normativity problematically suggests that linguistic thought has much more of what biological functions have generally. Seibt’s view can be usefully contrasted here with Rouse’s (2015) distinction between dimensions of normativity. On Rouse’s view, the emergence and acquisition of conceptually articulated understanding introduces a novel dimension of normativity distinct from, but not ‘better’ or ‘worse’ than, the teleological normativity of living animals in their environments.18 Yet Seibt correctly sees the need for a theoretical account of what is happening at the level of cognition and not just behavior. Though Rouse does not ignore cognitive science, the anti-representationalists he cites cohere with his broader theoretical orientation informed by Heidegger and Gibson. It is time to stop worrying that any talk of ‘representation’ implicates one in ‘Cartesian cognitive science’ (Wheeler 2005), as if merely saying the word requires thinking about representations as static symbols manipulated according to strict rules or as intentional icons directly accessible to consciousness. But if the meaning of a word consists of its use, then what the word ‘representation’ means depends on how it is used. Sellars’s driving intuition is that the basic function of representations is to orient, to allow an animal to navigate its environment. A theory of how animals navigate their environments requires positing something internal to the brain, and there is no reason why the word ‘representation’ cannot do that job.19 Moreover, representations need not be linguistic; a secure grasp on Sellars’s distinction between signifying and picturing allows us to inquire into what non-linguistic representations might be.20 That aside, Seibt helpfully points out that “orientation systems might be embodied in material objects that differ in structure and complexity radically from sentences in natural or programming languages” (ibid., 258). The complexity of an orientation system depends on the different degrees and kinds of causal regularities that the system can reliably track. This kind of normativity is cognitive, physiological and ecological: World-coordinatedness is the evolutionary condition for the possibility of a certain class of natural items that would not exist if they were not involved in normative functioning: that world-coordinatedness and the

170  Carl B. Sachs specific normative functioning of “natural-(pre)-linguistic” objects are selected for in combination. (ibid., 271) In other words, a genuine explanation of cognition in rerum natura requires that we understand cognition just as we understand all biological functions: in terms of the ecological functions of cognitive systems and the evolutionary history of eco-cognitive coupled dynamics. By placing more emphasis on ecology and evolution than Sellars did, and making use of theoretical tools that were unavailable to Sellars (such as complexity theory and nonlinear dynamics), Seibt advances our understanding of what picturing should mean. In contrast with Rosenberg and Seibt, Huw Price rehabilitates the picturing/signifying distinction through a critique of philosophy of language and analytic metaphysics.21 The central target of that critique is ‘Representationalism,’ understood as the idea that “the function of statements is to ‘represent’ worldly states of affairs and that true statements succeed in doing so” (Price 2013, 24). To be a Representationalist is to think that the proper function of assertoric discourse is to track how the world is. An assertion is true if it succeeds in doing so, false if it does not. Thus, Representationalists must distinguish between genuinely world-tracking discourse and other forms of discourse which are not genuinely world-tracking (even if they appear to be). Price’s strategy is to replace this bifurcation with what he calls ‘the new bifurcation thesis’: that there are two different concepts of representation which do not even belong to the same logical category. The distinction he urges is between ‘e-representations’ (‘e’ for environmental and external) and ‘i-representations’ (‘i’ for inferential and internal) (ibid., 36). E-representations are states of a dedicated system or subsystem that systematically co-vary with states of the environment in which that system is embedded. By contrast, i-representations are nodes in an inferential nexus—that is, propositional content. Distinguishing between these two different concepts of representation entails that our talk about truth and reference—our semantic metavocabulary—is not going to establish one way or the other how our discourse is related to the world (ibid., 37). The i-representation/e-representation distinction means we can prise apart ‘the content assumption’ and ‘the correspondence assumption.’ According to the content assumption, “language is a medium for encoding and passing around sentence-sized packets of factual information—the content of beliefs and desires” (ibid., 40). By contrast, the correspondence assumption holds that “these packets of information are all ‘about’ some aspect of the external world, in much the same way” (ibid.). Once we separate these two assumptions, we realize that “there is no requirement whatsoever that each node [in an inferential nexus—CS] have an e-representational role, where the correspondence assumption would have some traction” (ibid.). On Price’s view, there are many kinds of discourse (e.g., empirical,

“We pragmatists mourn Sellars”  171 mathematical, modal, semantic, ethical), all of which contain assertions and other speech acts. Those assertions form the content of beliefs and desires relative to that kind or dimension of discourse. Qua assertions, an assertion made in ethical discourse is neither more nor less of an assertion than an assertion in mathematical discourse. However, “while all assertoric vocabularies are i-representational, some may be much more e-representational than others” (ibid., 153); not all assertoric discourses reliably track features of the environment, or do so equally well. Price happily recognizes that the new bifurcation thesis is close to Sellars’s distinction between picturing and signifying (ibid., 166–7), with signifying (what Sellars eventually comes to call semantic assertability or S-assertability) being an important kind of i-representation and picturing being an attempt to specify a kind of e-representation. However, Sellars’s view is also broader because, as stressed earlier, he does not restrict picturing items to languages. Although picturing may suggest familiar worries about propositional content, mapping does not; maps are not sentences. Once we understand that Sellars’s basic point is that cognitive activity in rerum natura essentially involves mapping or map-like relations between states of some informationprocessing subsystem (the brain of an animal or the CPU of a robot) and detectable regularities in its environment, picturing items need not be linguistic. Sellars emphasizes linguistic objects only because language is essential to the uniquely human mode of picturing, not because only languages picture. This means that Sellars’s view, unlike Price’s, cannot be restricted to a naturalized anthropology but must be grounded, as Dewey’s was, in an ecologically embedded theory of cognitive activity. Regardless of how far we go in using cognitive science to flesh out picturing, the new bifurcation thesis nevertheless allows us “abandon the presupposition at the core of orthodox naturalistic Representationalism, that propositional content and word-natural-world correspondence live in the same box” (ibid., 170). Propositional content or i-representation is a legitimate use of the concept of representation, perfectly useful (or not) for semantical analysis. By contrast, when we talk about representations as reliably tracking features of the environment, we not even talking about the same thing. In responding to Brandom, Price admits that his distinctions are quite nuanced, and yet defends them: “But these distinctions are a necessary part of the nuance, in my view. In so far as neither Brandom nor Rorty seems sufficiently sensitive to them, I may be closer to Sellars than either of them” (ibid., 194). The nuanced distinction is simply this: we can reject Representationalism (as pragmatists like Dewey, Wittgenstein and Rorty have long urged) without giving up on all talk of representations. The signifying/ picturing distinction allows us to distinguish the concept of representation that satisfies the content assumption (signifying or i-representations) and the concept of representation that satisfied the correspondence assumption (picturing or e-representations). Put in those terms, Price shows how Sellars’s

172  Carl B. Sachs signifying/picturing distinction can resolve live issues in philosophy of language and analytic metaphysics.22

Pragmatism Today and the Cognition/Discourse Distinction We can now integrate Price, Seibt and Rosenberg in the following thought: that Sellars’s signifying/picturing distinction is best understood as a distinction between cognition and discourse—a distinction unfortunately elided by all traditional talk about ‘the mind’ or ‘the intellect.’ The former is about how biological cognitive systems, including but not limited to neural computations, process information in order to reliably track (or map) salient patterns in their environments (both physical and, when relevant, social). The latter is about how persons exchange inferentially articulated propositional contents through assertions and other speech acts in a game of giving and asking for reasons in order to facilitate successful cooperative behaviors.23 Though we should be pragmatists about both cognition and discourse, it would be a disaster to conflate them. If we were to conflate cognition and discourse, every concept with an established use in some dimension of discourse would have to refer to some actual or possible entity, just as cognitive representations (such as those of the robot in BBK) that guide sensorimotor ability reliably track relatively stable features of the environment. The consequence is thus the inflationary ontologies of Western metaphysics prior to, and thence in reaction to, the rise of modern science. By contrast, if our metaphysics is to be constrained by the posits licensed by our best scientific theories, then we must distinguish between the assertions and other speech acts whereby we navigate the polydimensional space of reasons and the map-like neurocomputational representations whereby we navigate our physical and social environments. In short, there is what our discourse is about, and there is how that discourse is related to the world. Conversely, if all assertoric discourse were, as such, a way of reliably tracking salient patterns and processes, such that all semantic content stood in a correspondence relation to some kind of object or relation, then our explication of mathematical, modal, and moral assertoric discourse would not be able to avoid the intractable mysteries of Plato or Meinong. We can finally return to our initial project of understanding Sellars in relation to the pragmatist tradition. His distinction between signifying and picturing, which I suggest is best understood as a distinction between discourse and cognition, is an attempt to synthesize Lewis’s pragmatist but antipsychologistic and anti-naturalistic theory of conceptual meaning with Dewey’s pragmatic naturalistic approach to intelligent behavior. With Lewis, Sellars affirms the fundamentally social character of the inferentially articulated concepts as manifest in discursive practices. With Dewey, Sellars affirms the fundamentally ecological character of the anticipatory representings that allow organisms to reliably track and respond to the causal regularities in their physical and social environments.24 Sellars shares both Lewis’s keen

“We pragmatists mourn Sellars”  173 interest in pluralism of conceptual frameworks and their fundamentally normative dimension and Dewey’s interest in the relevance of biology and psychology for epistemology and philosophy of mind. We pragmatists should find in Sellars’s work, and especially his distinction between cognition and discourse, the tools necessary for pragmatism to be understood, as Sellars himself understood it, as a truly revolutionary step in Western philosophy.25

Notes 1 “Those of us who learned from Sellars to think of the Myth of the Given as a confusion of causal conditions with justification are inclined to think that the project of finding connections between inquiry and the world needs elimination rather than naturalization. If one draws this moral, one will not seek ‘an Archimedean point outside the series of actual and possible beliefs’. So we pragmatists mourn Sellars as a lost leader” (1979a, 91). (Rorty is quoting Sellars 1967, 142.) 2 For a detailed examination of this claim, see Hookway (2008) and Sachs (2014). 3 As this would have been spring of 1937, Mind and the World Order would have been published eight years previously. 4 See Experience and Nature (1925/1982); see also Logic: The Theory of Inquiry (1948/1986), esp. Chapter 2, ‘The Existential Matrix of Inquiry: Biological’. 5 Dewey also underscored the social cooperative function of language; see Experience and Nature (LW 1). The difference is that Lewis sees no need to ground the social cooperative function of linguistic exchange in a naturalized anthropology informed by ecology and evolutionary theory. 6 For the importance of cooperation for conceptual meaning, compare: “We have a common reality because—or in so far as—we are able to identity, each in his own experience, those systems of orderly relation indicated by behavior, and particularly by that part of behavior which serves the ends of cooperation. What this primarily requires is that, in general, we be able to discriminate and relate as others do, when confronted by the same situation” (ibid., 110–11) and “both our common concepts and our common reality are in part a social achievement, directed by the community of needs and interests and fostered in the interest of coöperation” (ibid., 116). 7 As Zack points out, Lewis is best understood as giving Roycean idealism an empiricist twist (Zack 2006, 33). 8 Lewis was aware that this distinction depended on the analytic/synthetic distinction, which in turn depended on taking intensional semantics for natural and formal languages. If one were to insist on extensionalism, there would be no justification for the analytic/synthetic distinction and ultimately no justification for distinguishing epistemology from psychology—exactly as we see in Lewis’s renegade student W.V.O. Quine. For Lewis’s relation to Quine, see Sinclair (2012); see also Misak (2013). 9 I will not address whether the misunderstanding that Sellars attributes to pragmatism here is a correct understanding of pragmatism. 10 Cf. “Truth, we have seen, is not a relation. Picturing, on the other hand, is a relation, indeed, a relation between two relational systems. And pictures, like maps, can be more or less adequate . . . the concept of a linguistic or conceptual picture requires that the picture be brought about by the objects pictured; and while bringing about of linguistic pictures could be ‘mechanical’ (thus in the case of sophisticated robots), in the thinking of pictures as correct or incorrect we are thinking of the uniformities involved as directly or indirectly subject to rules of criticism” (SM, 135–6) and “Picturing is a complex matter-of-factual relation

174  Carl B. Sachs and, as such, belongs in quite a different box from the concepts of denotation and truth” (ibid., 136). 11 In Haugeland’s colorful metaphor, both Lewis and Sellars are on ‘third base’ about intentionality; see Haugeland (1998). 12 Or, if McDowell is correct and intentionality is a relation, then we do not need picturing; see McDowell (2009). 13 By contrast, Hegel acknowledges that it is a fact of history that we have the kind of manifest image that we do, but it is not a contingent fact. 14 Cf. “since agency, to be effective, involves having reliable cognitive maps of ourselves and our environment, the concept of effective agency involves that of our IPM judgments [Introspection, Perception, and Memory—CS] being likely to be true, that is, to be correct mappings of ourselves and our circumstances” (MGEC, 190). In a footnote, Sellars adds, “May I call them pictures?” 15 In particular MEV; but see also Sellars’s ‘Behaviorism, Language, and Meaning’ (1980) for a closely related argument as to why methodological behaviorists should not eschew talk of representations. 16 ‘Jumblese’ is a term Sellars coins for a language without terms for the predicate relation, in order to highlight the degree to which predication is dispensable. 17 See also Seibt (2016) on a ‘normativity gradient.’ 18 See also Okrent (2007) for the vindication of teleological normativity that Rouse uses. 19 Interestingly, Rorty’s objections to representationalism in epistemology did not extend to representationalism in cognitive science (Rorty 1979b, 244–56). 20 For cognitive representations as basically map-like, see Camp (2007); see also Huebner (2011) for why complex mental representations in nonhuman animals do not involve attributing propositional contents to them. 21 For an earlier version of that critique, see Price (2011) in which he argues that analytic metaphysics never adequately came to terms with Carnap’s critique of metaphysics. 22 It is a further question whether Sellars’s scientific process metaphysics is compatible with Price’s neopragmatism as naturalism without metaphysics. See Knowles (2014) for an extension of Price’s naturalism similar to Seibt’s Sellarsian process metaphysics. 23 I emphasize ‘assertions and other speech acts’ because there cannot be an exclusively assertoric discourse—a discourse of third-person speech acts with no firstperson or second-person speech act; see Kukla and Lance (2009). 24 See Rockwell (2014) for how Churchland’s computational neuroscience updates and complements Dewey, though Rockwell does not comment on how Churchland’s ‘neurosemantics’ is indebted to Sellarsian picturing. However, there are crucial differences between Dewey and Sellars. Sellars thinks about cognition qua picturing largely in computational or cybernetic terms. By contrast, Dewey emphasizes organisms, habits and feelings. Whether Sellars and Dewey can be completely reconciled remains to be seen. 25 I would like to thank Bryce Huebner, Steve Levine and Peter Olen for helpful criticisms on previous drafts.

References Bernstein, Richard. 2010. The Pragmatic Turn. Malden, MA: Polity Press. Brandom, Robert. 1994. Making It Explicit: Reasoning, Representing, and Discursive Commitment. Cambridge, MA: Harvard University Press.

“We pragmatists mourn Sellars”  175 Brandom, Robert. 2000. Articulating Reasons: An Introduction to Inferentialism. Cambridge, MA: Harvard University Press. Camp, Elizabeth. 2007. ‘Thinking with Maps’. Philosophical Perspectives, 21: 145–82. Dewey, John. 1977. ‘The Influence of Darwinism on Philosophy’. In The Middle Works of John Dewey, Volume 4: 1907–1909. Edited by Jo Ann Boydston, 1–14. Carbondale: Southern Illinois University Press. Dewey, John. 1981. Experience and Nature: The Later Works of John Dewey, 1925–1953, Volume 1: 1925. Edited by Jo Ann Boydston. Carbondale: Southern Illinois University Press. Dewey, John. 1986. Logic: The Theory of Inquiry: The Later Works of John Dewey, 1925–1953, Volume 12: 1938. Edited by Jo Ann Boydston. Carbondale: Southern Illinois University Press. Haugeland, John. 1998. ‘The Intentionality All—Stars’. In Having Thought: Essays in the Metaphysics of Mind. Edited by Haugeland John, 127–70. Cambridge, MA: Harvard University Press. Hookway, Christopher. 2008. ‘Pragmatism and the Given: C.I. Lewis, Quine, and Peirce’. In The Oxford Handbook of American Pragmatism. Edited by Cheryl Misak, 269–89. New York: Oxford University Press. Huebner, Bryce. 2011. ‘Minimal Minds’. In Oxford Handbook of Animal Ethics. Edited by Tom L. Beauchamp and R.G. Frey, 441–68. New York: Oxford University Press. Knowles, Jonathan. 2014. ‘Naturalism Without Metaphysics’. In Realism, Science, and Pragmatism. Edited by Kenneth Westphal, 200–18. New York: Routledge. Kukla, Rebecca and Lance, Mark. 2009. ‘Yo!’ and ‘Lo!”: The Pragmatic Topography of the Space of Reasons. Cambridge, MA: Harvard University Press. Lewis, Clarence Irving. 1929. Mind and World Order: Outline of a Theory of Knowledge. New York: Dover Publications. Lewis, Clarence Irving. 1970. ‘Review of The Quest for Certainty’. In The Collected Papers of C. I. Lewis. Edited by John D. Goheen and John L. Mothershead Jr., 66–77. Stanford: Stanford University Press. McDowell, John. 2009. ‘Intentionality as a Relation’. In Having the World In View. Edited by John McDowell, 44–65. Cambridge, MA: Harvard University Press. Misak, Cheryl. 2013. The American Pragmatists. New York: Oxford University Press. Okrent, Mark. 2007. Rational Animals: The Teleological Roots of Intentionality. Athens: Ohio University Press. Popp, Jerome. 2007. Evolution’s First Philosopher: John Dewey and the Continuity of Nature. Albany, NY: SUNY Press. Price, Huw. 2011. ‘Metaphysics After Carnap: The Ghost Who Walks?’ In Naturalism Without Mirrors. Edited by Huw Price, 280–303. New York: Oxford University Press. Price, Huw. 2013. Expressivism, Pragmatism, and Representationalism. New York: Cambridge University Press. Rockwell, Teed. 2014. ‘How Computational Neuroscience Revealed That the Pragmatists Were Right’. In Neuroscience, Neurophilosophy, and Pragmatism: Brains at Work with the World. Edited by Tibor Solymosi and John R. Shook, 57–70. New York: Palgrave Macmillan.

176  Carl B. Sachs Rogers, Melvin. 2008. The Undiscovered Dewey: Religion, Morality, and the Ethos of Democracy. New York: Columbia University. Rorty, Richard. 1979a. ‘Transcendental Arguments, Self—Reference, and Pragmatism’. In Transcendental Arguments and Science. Edited by Peter Bieri, Rolf Horstmann and Lorenz Krüger, 77–103. Boston: D. Reidel Publishing Company. Rorty, Richard. 1979b. Philosophy and the Mirror of Nature. Princeton: Princeton University Press. Rorty, Richard. 1991. ‘Representation, Social Practices, and Truth’. In Philosophical Papers 1: Objectivity, Relativism, and Truth. Edited by Rorty Richard, 151–61. New York: Cambridge University Press. Rorty, Richard. 2010. ‘Intellectual Autobiography’. In The Philosophy of Richard Rorty. Edited by Randall Auxier and Lewis Hahn, 3–24. Chicago: Open Court. Rosenberg, Jay. 1974. Linguistic Representation. Dordrecht: D. Reidel Publishing Company. Rosenberg, Jay. 1980. One World and Our Knowledge of It: The Problem of Realism in Post—Kantian Perspective. Boston: D. Reidel Publishing Company. Rosenberg, Jay. 2007. ‘Sellarsian Picturing’. In Wilfrid Sellars: Fusing the Images. Edited by Jay Rosenberg, 104–26. New York: Oxford University Press. Rouse, Joseph. 2015. Articulating the World: Conceptual Understanding and the Scientific Image. Chicago: University of Chicago Press. Sachs, Carl. 2014. Intentionality and the Myths of the Given. New York: Routledge. Seibt, Johanna. 2009. ‘Functions Between Reasons and Causes: On Picturing’. In Empiricism, Perceptual Knowledge, Normativity, and Realism. Edited by Willem deVries, 247–81. Oxford: Oxford University Press. Seibt, Johanna. 2016. ‘How to Naturalize Sensory Consciousness and Intentionality Within a Process Monism with a Normativity Gradient: A Reading of Sellars’. In Sellars and His Legacy. Edited by James R.O’Shea, 186–222. Oxford: Oxford University Press. Sellars, Wilfrid (BBK). 1963. ‘Being and Being Known’. In Science, Perception, and Reality, 41–59. Atascadero, CA: Ridgeview Publishing Company. Sellars, Wilfrid (SRLG). 1963. ‘Some Reflections on Language Games’. In Science, Perception, and Reality, 321–58. Atascadero, CA: Ridgeview Publishing Company. Sellars, Wilfrid (SM). 1967. Science and Metaphysics. Atascadero, CA: Ridgeview Publishing Company. Sellars, Wilfrid (AR). 1975. ‘Autobiographical Reflections’. In Action, Knowledge, and Reality. Edited by Hector-Neri Castañeda, 277–93. Indianapolis, IN: Bobbs-Merrill. Sellars, Wilfrid (NAO). 1979. Naturalism and Ontology. Atascadero, CA: Ridgeview Publishing Company. Sellars, Wilfrid. 1980. ‘Behaviorism, Language, and Meaning’. Pacific Philosophical Quarterly, 60: 3–30. Sellars, Wilfrid (MGEC). 1988. ‘More on Givenness and Explanatory Coherence’. In Perceptual Knowledge. Edited by Jonathan Dancy, 177–91. New York: Oxford University Press. Sellars, Wilfrid (MEV). 2007. ‘Mental Events’. In In the Space of Reasons. Edited by Kevin Scharp and Robert Brandom, 282–300. Cambridge, MA: Harvard University Press.

“We pragmatists mourn Sellars”  177 Sinclair, Robert. 2012. ‘Quine and Conceptual Pragmatism’. Transactions of the Charles S. Peirce Society, 48, 3: 335–55. Wheeler, Michael. 2005. Reconstructing the Cognitive World. Cambridge, MA: MIT Press. Zack, Naomi. 2006. ‘Murray Murphey’s Work and C. I. Lewis’s Epistemology: Problems with Realism and the Context of Logical Positivism’. Transactions of the Charles S. Peirce Society, 42, 1: 32–44.

9 The Varieties and Origins of Wilfrid Sellars’s Behaviorism Peter Olen

1. Introduction Wilfrid Sellars’s conception of behaviorism can best be described as ‘niche.’ On the surface, Sellars’s most well-known use of behaviorism turns on his Rylean thought experiment where he advocates for a conception of inner episodes as modeled on external verbal behavior.1 What separates this account from other attempts to include mental concepts within a behavioristic framework is Sellars’s modeling of inner episodes on external behavior, while simultaneously claiming that such episodes are non-reducible marks of the mental and, thus, cannot be analyzed (without remainder) in strict behavioristic terms. And while there have been numerous discussions of Sellars’s behaviorism, his philosophical commitments have not been placed within the larger narrative of twentieth-century American behavioral science. This, at least from my view, is a shame. Far from a uniform movement, the history of American behaviorism contains numerous lines of influence between philosophers and psychologists—many of which can be seen in Sellars’s own theories. Yet why think exploring these lines of influence will yield anything philosophically interesting? Given that behaviorism can safely be classified as a ‘dead’ psychological theory, why think there is anything of interest here beyond appeasing historical curiosity? Initially recognized in Marras (1977) (and more recently explored in deVries 2005; O’Shea 2007 and Olen 2016a), Sellars’s endorsement of behaviorism (with important qualifications explored later) should not be construed as only an endorsement of the science of his day. While Sellars kept a close eye on then-contemporary developments in the behavioral sciences, he also incorporated aspects of behaviorism into his philosophy of language. Sellars’s reliance on a behavioristically grounded conception of meaning (in terms of the realization of functional roles in patterns of behavior) puts joint pressure on the success of both behaviorism and functional role semantics. The theoretical gamble, then, is that the success or failure of Sellars’s semantics is tied to the success or failure of behaviorism. We should see Sellars’s adoption of behaviorism as not only ineliminable from his semantics, but as a dilemma for current appropriations of Sellars’s views on meaning.

Origins of Wilfrid Sellars’s Behaviorism  179

2.  Historical Varieties of Behaviorism There are numerous lenses through which to see the history of behaviorism. What I am interested in doing here is not re-counting every variant of behaviorism, but arguing that one way of telling the history can lead directly to Sellars’s behaviorism. I will construct this history around three different influential figures: John B. Watson, Edward Tolman and B. F. Skinner. The move from Watson to Skinner is anchored by the idea that classic forms of behaviorism were received—by the 1930s—as too restrictive on the use of mentalistic concepts. Staying within the realm of behavioral analysis (i.e., assuming that the proper vocabulary for conceptual analysis and experimentation is the language of observable behavior), the move is to see behaviorism as becoming increasingly liberalized to include cognitive concepts (in Tolman’s case) and sophisticated theories of verbal behavior (in Skinner’s case).2 This amounts to a rejection of restrictive forms of behaviorism, which interpreted all mentalistic concepts as reducible to physicalist terms. These two liberalizations of an initially austere program in psychology made conceptual room for Sellars’s version of behaviorism. Watson, as one of the founders of behaviorism, inspired a particularly austere form of the theory, one that paints a reductionist picture of mentalistic concepts (insofar as they moved beyond a bare stimulus-response analysis). Explanations of language and linguistic capacities, for example, were reduced to physiological explanations of reflexes or physiological responses to external stimuli (the view of language found in Watson’s or Albert Weiss’s work), seen as complex habits, or simply banished from scientific discussion.3 Whether the specific response to assertions of inner mental life to argue for the elimination of mental concepts in favor of behavioral or physiological explanations, or arguing they would eventually be reduced to descriptions of physical events, is somewhat irrelevant. What matters is that the initial approach to studying language and inner episodes is either to see them as ‘habit’ or ‘a form of behavior’, or as simply reducible to physiological characterization (Watson 1925, 181; Weiss 1925, 52). The explanatory resources needed to account for language learning—while acknowledged as a habit that differentiates us from nonhuman animals—were no more than what was found in the analysis of any other kind of behaviorial or physiological explanations. Tolman’s behaviorism—one informed with then-contemporary philosophical debates of the new and critical realists4—functions as a reintroduction of mentalistic concepts by liberalizing the requirements for what counts as a concept grounded on behavioral terms. Any cognitive notion, such as having a goal or purpose, that goes beyond the immediate arrangements and attainment of stimuli-specific goals would be either banished from the vocabulary of psychology or reduced to physiological explanations in the classical model of behaviorism. Insofar as early behaviorist accounts permitted mentalistic concepts, such as beliefs or purposes, they were seen as “manifestations of overt or covert motor activity” (Staddon

180  Peter Olen 1993, 18). Tolman framed his understanding of behaviorism against this austere view. In drawing a distinction between molecular and molar definitions of behavior (a distinction between defining behavior based on underlying causal mechanisms and taking behavioral terms as sui generis), Tolman clearly set himself apart from earlier forms of behaviorism by rejecting the kinds of reductive explanations found in Watson’s work (Tolman 1932/1949, 5–8). Given the rich amount of ‘mental’ representation Tolman built into even animal learning, it is not surprising that he was comfortable leaning on folk psychological notions to explain human behavior (Tolman 1948; Staddon 1993, 25). This also buttressed Tolman’s willingness to see behavioral analysis as sui generis—that the concepts and theories provided by behaviorism should be seen as distinct, though correlated, with their underlying physical causes: An act qua “behavior” has distinctive properties all its own. These are to be identified and described irrespective of whatever muscular, glandular, or neural processes underlie them. These new properties, thus distinctive of molar behavior, are presumably strictly correlated with, and if you will, dependent upon, physiological motions. But descriptively and per se they are other than these motions. (Tolman 1932/1949, 8) This re-working of behaviorism solves numerous problems found in a more restrictive form of behaviorism. While both forms of behaviorism deny any truth to the idea that language or cognitive concepts necessarily entail an inner life (Tolman 1932/1949, 235), cognitive concepts can be reintroduced into the vocabulary of psychological explanation, but on behavioral and instrumental grounds.5 Skinner, although differing substantially from Tolman, offers a thorough behavioral analysis of verbal behavior, and one that plays a large role in Sellars’s thought. One historical consideration to keep in mind is Skinner’s departure from classic behaviorism, especially as it concerns language. Skinner’s (1945) influential paper on operational definitions—one that emphasized the social nature of behavioral control and language acquisition—was published in almost overlapping periods with Sellars’s tenure at Minnesota (Sellars joined the department the year Skinner left for Indiana University). The depiction of how we might understand language as being divorced from a search for the referent of semantic or mentalistic terms is not, in and of itself, a novel move, but Skinner’s explicit endorsement is clearly influential Sellars. Thus, one finds Skinner claiming that: Terms falling within this class are apparently descriptive only of behavior, rather than of other internal states or events, since the possibility

Origins of Wilfrid Sellars’s Behaviorism  181 that the same stimulus may be both public and private (or, better, may have for lack public accompaniments) seems to arise from the unique fact that behavior may be both overt and covert. (Skinner 1945, 274) Cognitive concepts (such as thoughts) are handled on the same stimulusresponse model as verbal behavior. Although granting a difference in the ‘energy level’ between overt and covert behavior, Skinner is committed to the idea that verbal behavior and thought are simply different levels of overt speech and action (Skinner 1957, 435–6). Their analysis, thus, does not concern the terms and concepts of introspection, but the notions of enforcement, punishment and reinforcement gathered from a social setting (i.e., what was already available to one’s community). However, one of the major dividing lines between Skinner and Sellars is that Skinner takes the privacy of inner episodes as a reason to ban reporting role from our conceptual repertoire. By ‘privacy,’ behaviorists were concerned with the fact that our first-person thoughts, sensations and imagery are only immediately accessible to the individual. Thus, the reasoning goes, such inner episodes—insofar as they exist—cannot be part of a scientific study of persons. It is not that inner episodes or private stimuli do not exist for Skinner, but that (i) their characterization is entirely possible in behavioral terms of reinforcement and communal conditions, and (ii) that such private events do not refer to the occurrence of inner episodes, but just the conditioned responses habituated through one’s community (Skinner 1945, 276). ‘Seeing,’ for example, is not a response to the private stimuli of seeing a red patch, but “a term acquired with respect to one’s own behavior in the case of overt responses available to the community” (Skinner 1945, 276; emphasis added). One way to construe the development of American behaviorism is to see it as expanding and liberalizing behaviorism as a psychological theory. Sellars’s philosophy is conceptually primed to solve numerous problems ignored or rejected by earlier behaviorists. This amounts to another liberalization of behaviorism by offering behavioral concepts that account for our inner mental life, linguistic capacities and human freedom6 without succumbing to earlier tendencies towards instrumentalism (e.g., Tolman) or reductionism (e.g., Watson). The mentalistic and semantic resources available for characterizing such domains of discourse will be taken as irreducible aspects of our theoretical and practical accounts of human agency. This re-introduces previously marginalized mentalistic concepts and moves beyond seeing them as merely useful fictions or tools. Starting from a behavioral science basis, the conceptual possibility of re-introducing mentalistic concepts along broadly behavioral lines seems possible only after Tolman and others pushed away from a Watsonian approach to psychology.

182  Peter Olen

3.  Sellars’s Behaviorism While he articulates a variety of different conceptual possibilities for behaviorism, Sellars's use of behaviorism starts with its status as the ‘base’ for an expansive view of human agency (Sellars LRB, 122).7 This behavioral basis is initially given a bland, stimulus-response characterization and is marked by two classic commitments: (i) that the laws and principles applicable to animal learning are equally applicable to humans, and (ii) that a ‘complete’ account of human agency could, in principle, be created out of behavioristic theories. Since the rise of evolutionary theory in the late nineteenth century, the explanation of human agency as a complex case of the laws and theories applicable to nonhuman animals was a theoretical goal, one which found a home with behaviorists (Tolman 1932/1949, 3). Sellars clearly endorses the first claim (Sellars LRB, 122–3), but he clashes with most, if not all, behaviorists over the second point. Although Sellars depicts behaviorism as—at least in principle—capable of explaining human behavior, he will present mentalistic concepts as logically irreducible to behavioral concepts, thus stopping short of endorsing the second claim mentioned above.8 If the classic, Watsonian incarnation of behaviorism were true, there would be little room for a normative depiction of human agency.9 Despite insisting on what initially appears as incompatible commitments, Sellars consistently argues for an understanding of mentalistic concepts as ‘causally reducible, yet logically irreducible’.10 While austere forms of behaviorism analyze mental concepts solely through their connection to causal or behavioral explanations (represented in what Sellars calls ‘logical behaviorism’ or ‘philosophical behaviorism’), Sellars argues these accounts mistakenly claim that mentalistic concepts are logically reducible to behavioral concepts (Sellars SSMB, 194–5). Instead, we should recognize that as a sound methodological thesis (represented in what Sellars calls ‘scientific’ or ‘methodological’ behaviorism),11 behaviorists need not embrace a reducible connection between mental concepts and physical concepts. What is required is a methodological thesis about the applicable concepts used in the study of human behavior—limiting the creation of new concepts to “terms of a basic vocabulary pertaining to overt behavior”—as opposed to a substantive thesis that analyzes human agency solely through behavioral concepts (Sellars EPM, 264–5). Why do we need a conception of mental concepts that is logically distinct from their causal, behavioral basis? Part of the argument—insofar as Sellars argues for this position12—turns on the idea that we simply need the resources only found in semantic metalanguages. This is what Sellars is getting at when he discusses the need to avoid seeing the scientific depiction of persons as “an alien appendage to the world in which we do our living” (Sellars PSIM, 40). This is also why classic accounts of behaviorism are seen by Sellars as unnecessarily stifling for concept formation. If we were wed to a traditional behavioristic understanding of language (either reductionist or

Origins of Wilfrid Sellars’s Behaviorism  183 dispositional), then we would fail to capture the language through which we identify as individuals and members of a community. That is, we would fail to see the world through the concepts and language through which we come to know ourselves as, well, ourselves. Arguably, this is a dimension of human experience and life that is an ineliminable part of human agency, yet one difficult to cash out in strictly behavioral or causal terms.13 Another argument for this position is found in the claim that mentalistic (as well as semantic and normative) concepts are simply saying something different than their material equivalencies. While it may be true there is a strong, law-like between my having a belief that I ought to bet on nine and certain behavioral states (the moving of betting chips, the throwing of dice), this does not mean such material connections exhaust said belief. Even if mental and behavioral concepts convey the same information (the related behavioral dispositions and actions), they do not say the same things. This argument is clearest in Sellars’s work when he discusses normative discourse: The situation is even clearer with respect to normative discourse. Whatever users of normative discourse may be conveying about themselves and their community when they use normative discourse, what they are saying cannot be said without using normative discourse. The task of the philosopher cannot be to show how, in principle, what is said by normative discourse could be said without normative discourse, for the simple reason that this cannot be done. (Sellars SSMB, 214) In terms of behaviorism, the point is that while distinct domains of discourse (e.g., mental, semantic, normative) may be causally reducible to their material correlates, they cannot be seen as logically reducible because of their clear difference in meaning. Even if behaviorism adequately explains the causal story tied to human behavior, this does not entail that such causal explanations are the only explanations when it comes to human behavior. Situating Sellars’s view within the historical narrative of behavioristic theory demands a fairly complex picture.14 As noted previously, Sellars draws various distinctions between interpretations of behaviorism (e.g., scientific, analytic, logical, philosophical, verbal) without identifying who he has in mind in most cases. Sellars’s behaviorism is a radical departure from classical, Watsonian forms of behaviorism.15 Instead of adopting a behavioral basis for understanding human agency and concepts that is radically opposed to mentalistic concepts, Sellars offers a behavioral framework for classifying these concepts without relegating inner episodes to a reductionist or eliminativist picture. Although theoretical terms (e.g., thought, sensation) are modeled on overt behavior, they are not collapsible into such concepts (Sellars EPM, 268). Far from banishing the notion of an inner mental life from the conceptual landscape, behaviorism serves as the basis

184  Peter Olen for understanding all other agential aspects of human thought and action. Placing this variant of behaviorism amongst others allows us to see how Sellars’s modifications of behaviorism fall out of a particular move away from traditional behavioristic theories and towards behaviorists who were willing to embrace cognitive and semantic concepts. This, as some authors have noted, strains against the idea that some thinkers (such as Tolman) should even be classified as behaviorists.16 Tolman’s earlier discussion of purposive behavior is a key example that clearly influenced Sellars, as well as Tolman’s notion of cognitive maps, a focus on explanation (instead of prediction and control), and use of what we now think of as folk psychological concepts.17 We know that Sellars was aware of Tolman’s work, if only through the occasional reference or phrase. Writing towards the end of his career, Sellars frames the historical evaluation of behaviorism as such: Today it is obvious at a glance that in a sense the Behaviorist methodology was unnecessarily restrictive. Once it became clear that the concepts of physical theory could not be explicitly defined in terms of observables, the Behaviorist was open to the challenge—why is it not legitimate to introduce concepts in Behavioristics which are not explicitly definable in terms of observable behavior described in carefully aseptic (non-anthropomorphic) terms? As it was sometimes put: Why not introduce hypothetical constructs into Behavioristics (compare the postulational concepts of physical theory) instead of restricting yourself to intervening variables (the mark of which—though the terminology was misleading—is definability in terms of observables)? To be sure, in the resulting controversy the phrase ‘intervening variables’ was often used with sufficient looseness by its proponents that it could cover ‘variables’ which were not, strictly speaking, definable in observation terms. The latter, however, were treated in an instrumentalist spirit, as merely ‘mathematical,’ and denied psychological reality. Thus, in a sense, the issue was not clearly joined. Nevertheless the methodological point of the stress on the concept of an ‘intervening variable’ was to block a return to Mentalistic concepts. (Sellars BLM, 6) This, in a nutshell, directly targets the instrumentalism inherent in Tolman’s behaviorism. It is not that Tolman’s position does not provide a theoretical basis for discussing mentalistic or semantic concepts in their own metalanguage (he does just so—as argued in section II), but that such concepts are taken to be—in the end—dispensable or reducible to behavioral terms. They are, as Tolman mentions when describing the overall road of the sciences, mere maps or tools for our use, but fall short of approximating reality. Tolman may have opened the door for folk psychological concepts

Origins of Wilfrid Sellars’s Behaviorism  185 (amongst others) to play a role in behavioristic explanations of agency, but this does not mean such concepts carve [mental] reality at the joints. Sellars differs from Skinner on the sui generis status of semantic and mentalistic discourse. While agreeing with Skinner on the linguistic community’s role in instilling and reinforcing linguistic habits and concepts, Sellars quickly departs when it comes to the referential nature of terms. This is not to say that Sellars held something like a reference theory of meaning, but that he thought talk of inner episodes was not shorthand for dispositions to act or verbalize in specific ways. Instead, verbal behavior can reference inner episodes (if only as theoretical concepts). When I say ‘I am having a red experience,’ I’m not only making moves in a public language game, I am also—in a legitimate and seemingly irreducible sense— having a red experience. There is, qua the experience of redness, an episode of redness that counts as an ‘inner episode.’ Skinner’s move to ‘externalize’ verbal behavior as responses may agree with Sellars’s views on the role of the linguistic community, but his rejection of the need for inner episodes to account for our verbal behavior in ways that take inner episodes as legitimate concepts (in the sense that such experiences do, in fact, generate and constrain behavior and, therefore, must play a role in a complete understanding of human agency) constitutes the breaking point between himself and Sellars. Sellars rightly identifies an oddness the claim that even though I am saying ‘I am having a red experience,’ I really mean something else (i.e., the causal or functional analysis: what causes me to say ‘I am having a red experience,’ what I am responding to, the reinforced dispositions to respond, dispositions habituated in me through my linguistic community, etc.). Sellars’s insight is to see the mistake of making it impossible for me to report such experiences as essentially meaning what I explicitly say. This position is perhaps understandable when we start from the assumptions that our inner mental lives are incorrigible and that the semantic resources required for reporting such experiences are, at best, shorthand for behavioral sciences terms or, at worst, chimerical. How does provide a better solution to the problem of inner episodes than Skinner? Here is where the differences between Skinner’s and Sellars’s behaviorisms become clear: My myth has shown that the fact that language is essentially an intersubjective achievement, and is learned in inter-subjective contexts—a fact rightly stressed in modern psychologies of language, thus by B. F. Skinner and by certain philosophers, e.g. Carnap, Wittgenstein—is compatible with the ‘privacy of ‘inner episodes’. It also makes clear that this privacy is not an ‘absolute privacy’. For if it recognizes that these concepts have a reporting use in which one is not drawing inferences from behavioral evidence, it nevertheless insists that the fact that overt

186  Peter Olen behavior is evidence for these episodes is built into the very logic of these concepts. (Sellars EPM, 269–70) While Skinner rejects inner episodes as playing any reporting role above and beyond the objective use of behavioral evidence (i.e., applying the same responses and labels to behavioral evidence recognized in, or inferred from, the behavior of others), Sellars wants to preserve the firstperson reporting role of such concepts. Thus, one key move in the myth of Jones: ‘internalizing’ what were once ‘external’ concepts and categories. Understanding as modeled on external speech allows us to introspect and report through publicly accessible concepts. Insofar as we introspect through concepts (in place of ineffable, private thoughts or ideas) that are built from accessible resources, the intersubjective (and therefore accessible) nature of ‘private’ episodes (Sellars EPM, 264). This maintains the intersubjective verifiability of concepts pertaining to our inner mental life without doing violence to the fact that the privacy of such lives is not a barrier to concepts playing the dual role of reporting and fact-stating (Sellars SRLG, 336). Sellars’s behaviorism allows for the intersubjectivity of what were traditionally private and ineffable episodes. If we start from Sellars’s behavioristic standpoint, where our mental vocabulary and concepts are built from external behavior and then internalized as a part of language acquisition, the mystery behind the ineffability of the mental disappears. Behavioristic categories and concepts are subsequently used (with modification) as the conceptual framework for our inner episodes (Sellars EPM, 267). Although privacy (qua accessibility) may still remain, there is nothing mysterious about this fact (i.e., that I cannot immediately know what thought or sensation you are entertaining this instant). What becomes publicly accessible are the concepts and categories used to guide and control our behavior. Understanding how two individuals have the ‘same’ experience or how two experiences could possibly ‘mean the same thing’ ceases to be a problem. Our different experiences of seeing a red plane and appropriately responding ‘Lo, a red plane!’ are thus explainable through publicly accessible, intersubjective categories. The instrumentalist nature of Tolman’s cognitive concepts and Skinner’s narrow understanding of verbal behavior serve as a set of interlocking problems for the creation of Sellars’s specific kind of behaviorism. That being said, I want to resist ‘final’ classifications of Sellars’s behaviorism for historiographical reasons.18 Sellars’s behaviorism differed from the dominant strands found in Watson, Tolman and Skinner by insisting on the need for additional explanatory resources (e.g., semantic, normative) to account for human agency, thought and action. The complex departures from earlier incarnations of behaviorism marked an ever-increasing picture of how behavioral analysis could sustain, as oppose to eliminate, additional

Origins of Wilfrid Sellars’s Behaviorism  187 semantic concepts. Sellars, like Tolman, adopted a form of behaviorism that embraced cognitive and non-mechanistic concepts of agency, while retaining the behavioristic basis crucial for intersubjective verifiability.

4.  Semantics and Functional Roles Once placed alongside Tolman and Skinner, Sellars’s niche behaviorism sustains a strong connection between behaviorism and semantics. Traditionally, behaviorism left a paltry role for meaning. Even though the study of language acquisition was a key component of behaviorist research platforms, most early twentieth-century theories either offered a physiological characterization of language (as in Watson’s case) or on the role of community membership, reinforcement, and habituation (e.g., Clark Hull, Grace de Laguna, Skinner). One development we can see in Sellars’s views is his employment of functional role semantics to account for classically intensional concepts without starting from a traditionally mentalistic standpoint. Sellars wants to avoid both a reductive form of behaviorism and a classical picture of thoughts, which starts from mental images and moves outward to speech that expresses said images. What is needed, then, is a way to account for meaning in a broad sense (to include not just the meaning of words, but the intentionality of thought and action) without contradicting variety of commitments that come with behaviorism (i.e., external behavior as a starting point for conceptual analysis, the intersubjective nature of concepts, and the seeming ineffability of inner episodes). Functional role semantics indicates that the meaning of a term or concept can be determined by giving “a functional classification of the utterance,” which “involves a special [illustrating] use of expression” (Sellars MFC, 421). When interpreting the semantic claim “ ‘Tortuga’ (in Spanish) means Turtle,” we are not referencing extralinguistic or abstract entities, but pointing out the functional roles both terms fulfill. The semantic equivalence of ‘Turtle’ and ‘Tortuga’ are found in that fact that both terms play a similar role in their respective languages (‘similar’ in the sense that their perceptual cues, syntactical rules and behavioral output—Sellars’s entry, intra and exit transitions—are ‘close enough’). Such roles are exemplified in our verbal and nonverbal behavior through negative and positive uniformities that are gained during the process of learning a language.19 These uniformities constitute a pattern of behavior that is the close connection between terms or concepts meaning something and behaviorism itself: The key to the concept of a linguistic rule is its complex relation to pattern-governed linguistic behavior. The general concept of pattern governed behavior is a familiar one. Roughly it is the concept of behavior which exhibits a pattern, not because it is brought about by the intention that it exhibit this pattern, but because the propensity to emit behavior of the pattern has been selectively reinforced, and the

188  Peter Olen propensity to emit behavior which does not conform to this pattern selectively extinguished. (Sellars MFC, 423) Verbal behaviorism—even as an incomplete and oversimplified framework for explaining inner episodes—places the realization of meaning in the patterns of behavior evinced language users. It is not that terms or concepts are wholly identified by their role in behavioral input or output (Sellars’s language entry and exit transitions), but that these roles are integral for exemplifying what it means for a human language to be particularly human (as opposed to an empty syntactical husk). Functional role semantics allows Sellars to have his cake and eat it too: we get to keep non-extensional discourses (e.g., semantic, modal, mental) in a robust role that goes beyond the instrumentalism granted by Tolman and other behaviorists. More so, it provides a way for behaviorists to salvage the ‘classic’ concepts of ideas or meaning without thereby giving into ‘unscientific’ (i.e., unverifiable) ways of thinking. For example, Skinner initially rejected the idea of meaning outright because of these exact considerations: Perhaps no one today is deceived by an ‘idea’ as an explanatory fiction. Idioms and expressions which seem to explain verbal behavior in term of ideas are so common in our language that it is impossible to avoid them, but they may be little more than moribund figures of speech. The basic formulation, however, has been preserved. The immediate successor to ‘idea’ was “meaning,” and the place of the latter is in danger of being usurped by a newcomer, ‘information.’ These terms all have the same effect of discouraging a functional analysis and of supporting, instead, some of the practices first associated with the doctrine of ideas. (Skinner 1957, 6–7) If Sellars is correct, we can salvage a robust explanatory and constitutive role for talking of meaning, ideas, thoughts and even obligations without additionally running contrary to a causal or function analysis of these terms. Although logically irreducible discourses ‘hover above’ their underlying causal explanations, they are intertwined with such bases. What makes this issue pressing in our current epoch is that while various aspects of Sellars’s philosophy have been endorsed to varying degrees, functional role semantics tends to be a universally acclaimed part of his project. Many contemporary scholars are perhaps willing to abandon Sellars’s nominalism, his conception of abstract entities or his conception of linguistic rules, but functional role semantics is largely accepted as a (somehow) atomistic piece of Sellars’s systematic philosophy.20 The pressing dilemma is whether one can adopt functional role semantics without thereby buying into an outdated form of behaviorism.

Origins of Wilfrid Sellars’s Behaviorism  189

5.  The Dilemma The key issue concerns whether Sellars can hold both of the following commitments: (a) that behaviorism functions as but a moment in the scientific image, one that will eventually give way to more sophisticated, neurophysiological pictures of agency and (b) that behaviorally grounded concepts play an integral role in our semantics. That both of these positions are held by Sellars is largely uncontroversial. Reflecting on the role of behaviorism and conceptual systems towards the end of his career, Sellars claimed that: The substantive framework of verbal behaviorism is now viewed as methodological prelude to a full-blooded theory of representational systems, much as macro-physics is a methodological prelude to a unified theory of the behavior of physical things. It is useful to view this extended theory of conceptual episodes as a placeholder for the future achievements of a neurophysiological theory of representation. (BLM, 14–15; emphasis added) This passage embodies one aspect of the tension I outlined earlier. There is a sense in which this identification of the conceptual order with the progress of neuroscience oddly brings us back to an almost Watsonian picture of how to understand language and behavior. The concern is that if neuroscience will inevitably surpass what behavioral evidence we have, in terms of explaining and accounting for human behavior, it stands to reason that behaviorism and behavioral evidence will disappear as the explainers of human psychology. This, in and of itself, is not a problem, but the tension within Sellars’s philosophy turns on that fact that functional role semantics is tied to a behaviorally grounded conception of ‘roles.’ It is not that functional roles are solely cashed out in behavioral terms (as comparing the different roles played by ‘Turtle’ and ‘Tortuga’ can—at least in principle—be achieved through an analysis of the applicable syntactical rules),21 but that how such concepts ‘fit’ within our patterns of behavior is crucial for understanding their linguistic roles now. Part of what it is for a concept to fulfill a given role just is for it to be realized in certain patterns of behavior, prohibit other patterns of behavior from emerging and demand specific reactions to stimuli. In moving from a behavioristic to neurophysiological framework, the concern is that ‘uniformities’ or ‘patterns-governed behavior’ will cease to be the relevant explainers. If we think what’s really triggering me to respond correctly when I see someone gesture hello is not, in fact, a habituated response to raise my hand, but the firing of mirror neurons (or some other neurophysiological explainer), then it is not clear how—in its present incarnation—functional role semantics is connected to such a causal story. The foundational claim of causal reducibility might change without a problem; what information is being conveyed in the neurophysiological

190  Peter Olen explanation is different from the behavioristic explanation, but it does not seem in-principle impossible for the underlying causal story to change without impacting the logical irreducibility of mentalistic discourse. In other words, the change in underlying causal explanation does nothing to change the fact that causal and semantic vocabularies (regardless of which causal vocabulary we choose) simply do not say the same thing. This kind of reasoning—one that would shrug off the objection by locating functional role semantics on the ‘logical’ (as opposed to causal) side of the story—depends on Sellars’s overall placement of functional role semantics in the manifest and scientific image. This is a vast topic, but the importance of Sellars’s reasoning here should be clear: how he understood the eventual fate of the behavioral sciences will impact the relationship between semantic discourse and behavior. If Sellars was an eliminativist about various ‘levels’ of scientific explanation, as a hypothetical example, then we might think the necessary connection between behavioral evidence and semantic statements would (eventually) be problematic. Following Sellars’s methodological point about behavioristic restrictions on concept formation, we could see behavioral terms as the macro-physical correlates of a [final] unified physical theory—a necessary step to the final scientific image. Their place in our practice of explaining and using language is secure, though simply not the final explanatory word. But treating this methodological restriction on concept formation as a permanent feature of theorizing fails to account for the conceptual change that would necessarily occur through the development of the sciences. Given the systematic nature of his philosophy, the problem for Sellars (and Sellarsians) is the seeming inseparability between his commitments to a specific scientific theory, one which underwrites his semantic theory, and the developing concepts and theories of the sciences. This is a problem for any philosopher—our empirical and theoretical claims are fixed points in time and, on a long enough timeline, we don’t get to revise them. Sellars’s issue, rather, is whether this recognition allows his ideas to survive such a conceptual change. On first pass, one might think so. Sellars was well known his desire for a synoptic vision of the integration of various conceptual frameworks, but this ‘blending’ of ideas does not provide us with a clear answer as to how certain ideas survive or perish under conceptual changes. Where, exactly, functional role semantics fits in an integrated picture of the images is far from clear. Maybe a good Sellarsian has no qualms here. Maybe the fundamentally insightful aspect of Sellars’s argument is the methodological restrictions on concept formation combined with what are now our best informed semantic and scientific theories. The pragmatic and fallibilistic aspects of Sellars’s thought makes room for the eventual conceptual sea change on all fronts. But this leaves current endorsements of functional role semantics—insofar as they are tied to an outdated behavioral science—in a precarious position. Although passed over in contemporary discussions of Sellars’s work, putting

Origins of Wilfrid Sellars’s Behaviorism  191 pressure on the connection between behaviorism and semantics is not a new approach.22 Marras recognized that the close connection between behaviorism and might be too close to survive the transition between images or changes within a given image. For example, Marras argues: Now to the extent that Sellars requires that these functions or roles be specifiable exclusively in terms of overt behavioral regularities, or, to put it differently, that they be completely expressible in terms of a purely Rylean (nontheoretical) language (recall: “Dot-quote common nouns have a purely Rylean sense”), then it is difficult to see how Sellars could avoid logical behaviorism about meaning. (Marras 1977, 672) Marras offers this prognosis in the context of worrying about Sellars’s classification of a pre-Jones, purely behavioristic (i.e., Rylean) verbal community as being explanatorily adequate when accounting for verbal behavior prior to the introduction of theoretical entities. Although granting that Sellars’s methodological behaviorism allows for the introduction of nonverbal theoretical entities, Marras’s point is that that Rylean language itself is perhaps too rich to require the introduction of theoretical terms. This point, as Marras admits, somewhat misses the mark when critiquing Sellars’s project. Looking at the problem from a different direction, one might object that it is difficult, if not impossible, to move from a behavioral to neurophysiological framework when characterizing our verbal and conceptual capacities. The objection can be thought of along the lines of Sellars’s discussion of the trainer/trainee relationship during language acquisition: regardless of the ‘ultimate’ explanation found in future sciences, we will always need behavioral concepts and cues to welcome language users into our community and confirm psychological theories. This is either due to the fact that a broad sense of ‘behavioral criteria’ is ‘always’ used to test and confirm psychological theories (PSIM, 22) or that we need to account for rational or logical connection of semantic discourses. Sellars calls the former requirement ‘good sense’ and it allows him to endorse a seemingly innocuous sense of behaviorism. Predicting the final development of the sciences is equivalent in accuracy to reading tea leaves in order to predict one’s future, but it is not clear that even Sellars was committed to this line of reasoning. Surely it is true that we now require behavioral cues in order to teach language users, but it’s not clear this will always be the case. More so, it is also not clear how that current practice of language learning would help those defending the relationship between behaviorism and functional role semantics. Insofar as the underlying causal explanation changes, one assumes in many cases the ways in which we acquire a given concept changes as well. On the other hand, we could defend the need for behavioristic concepts on largely practical grounds: the move from ‘making

192  Peter Olen sounds’ to ‘expressing claims’ requires the response to, and constraint of, stimuli and patterns of behavior-concepts that are ineliminable to language acquisition regardless of advances in neuroscience. While theoretical explanation may advance beyond observable or behavioral terms, language users need such concepts. But deciding the most practically efficient way for learning a language is a far cry from justifying the continued incorporation of behavioristic considerations on explanatory grounds. This line of reasoning, I believe, is a dead end. What I am worried about is simply one instance of many different clashes between Sellars’s manifest and scientific images. What arguably makes this particular clash significant is the primacy of functional role semantics in Sellars’s philosophy. As mentioned in section II, even ignoring possible conceptual tensions, it seems clear that the [eventual] abandonment of behaviorism will cause problems for Sellars. Unseating functional role semantics would basically undercut Sellars’s ability to structure a complex relationship between a logical framework that hovers above the causal story, and its ‘joining’ with the scientific image.

6. Conclusion Historically grounded depictions of Sellars’s behaviorism are difficult to construct. Sellars’s archival holdings give us little information as to the exact sources for his conception of behaviorism. Shades of Ryle, Skinner, Tolman and others are not difficult to discern in Sellars’s published writings, but this connection is far from exact. Without better records of Sellars’s engagement with specific behaviorists or texts, the issue of how Sellars saw his work in relation to then-practicing behaviorists will remain underdetermined. Yet, this may not be a requirement for a good (i.e., justified and evidentially grounded) account of Sellars’s influences and their role in our understanding of his philosophy. Pluralism can happily rule the day, insofar as we keep in mind exactly what kind of explanation we are offering in terms of historical events, influences and individuals. What I am justifying through the historical record is fairly limited in scope. What needs to be true in order to show the connection between my behavioristic narrative developed in section II and Sellars’s own niche behaviorism is a paltry sum of influences, all of which I believe are fairly evident from the textual evidence cited in this chapter. I’ve suggested Tolman and Skinner as likely sources for a number of Sellars’s ideas and argumentative moves, and I’ve also argued that reading Sellars’s niche behaviorism through a narrative that moves from Watson to Sellars is a useful way to understand a fundamental tension in Sellars’s semantics. I have stayed away from claiming a direct line of influence from specific behaviorists to Sellars (though I think it is likely, given my previous observations), and I am inclined to pull away from claiming Sellars himself understood behaviorism through exact narrative. What I am arguing is that

Origins of Wilfrid Sellars’s Behaviorism  193 placing Sellars’s behaviorism in-line with this kind of reasoning exposes a problem surrounding the interlocking nature of his philosophy, his semantics and conceptual changes resulting from the onward march of the sciences. Locating where Sellars’s behaviorism lands amongst other historical variants is useful insofar as we can use such a location to pinpoint philosophically relevant problems that arise out of its historical antecedents. In terms of the tension between even a liberalized behaviorism and Sellars’s semantics, I have offered no solution. This does not mean the tight connection between behaviorism and semantics is the fatal flaw in Sellars’s views. I would argue a solution to this tension requires more than just a reminder of the synoptic nature of the relationship between the manifest and scientific image. While it is true Sellars desired a ‘final’ picture of humanity that ‘joined,’ rather than ‘reconciled,’ the competing images of persons, such a desire is not equivalent to a solution (PSIM, 40). And even if we proceed on, and are trained through, concepts that start from behavioral or observational vocabularies, this is no guarantee that eventually the sciences will not move beyond the need to be joined with our folk psychological notions. If what is ‘really’ doing the explaining just is the microphysical theory (in Sellars’s terminology), then it might be that functional role semantics cannot be untethered from the behaviorism that serves as its ground.

Notes 1 See section XII of Sellars EPM. 2 As Kurt Salzinger has argued, Skinner was one of the most influential behaviorists to take the idea of language as a phenomenon ripe for behavioral analysis seriously. According to Salzinger, “For years, many behavior analysts viewed verbal behavior as essentially an epiphenomenon that would come along as long as we worked on nonverbal behavior. Yet even a little bit of thought makes clear, that particularly in complex civilizations, it is talk that produces the most important reinforcers and it is talk that allows one to avoid the most egregious consequences. Skinner was challenged to show how behavior analysis could explain this most important of all classes of behavior, thus making the act of convincing his fellow behaviorists to include verbal behavior in their research most important” (Salzinger 2008, 288). 3 For a comparison of their views on this issue, see Mills (1998, 42–4) and Watson (1925, Chapter10). 4 This might be another source for what seems to be a direct influence between Sellars and Tolman. See Tolman (1932/1949, Chapter 25) for his engagement with philosophers, and Smith (1986) for an in-depth account of Tolman’s relationship to new realism, critical realism and logical positivism. 5 Tolman on the instrumental nature of psychological concepts: “All science necessarily presents, it seems to us, but a map and picture of reality. If it were to present reality in its whole concreteness, science would be not a map but a complete replica of reality. And then it would lose its usefulness.”. . . . “One of the first requisites of a science is, in short, that it be a map, i.e., a short-hand for finding one’s way about from one moment of reality to the next—that it be a symbolic compendium and means of which to predict and control” (Tolman 1932/1949, 424–5).

194  Peter Olen 6 Another substantial difference between Sellars and Skinner is their treatment of human freedom and agency. While Skinner roundly rejects the need to account for human agency, Sellars finds the distinction between habituated and free behavior integral when account for aspects of our behavior. See Vargas et al., (2007) and Sellars LRB. 7 DeVries also notes some additional changes in Sellars’s attitude towards behaviorism—see deVries (2005, 174). 8 For an account of the difference between analysis and explanation in Sellars’s philosophy, see deVries and Triplett (2000, 34–5). 9 I’ve argued for the importance of this tension in numerous places. See Olen (2016a and 2016b) for further elaboration on these points. 10 Reflecting on Sellars’s philosophy itself, it should be clear that although consistently adhering to a broad distinction between a behaviorism that rejects all theoretical notions (i.e., those not tied to observable, behavioral concepts) and one that allows for such terms (albeit in a variety of forms), there is a worthwhile tension to be explored that is not, perhaps, always perfectly clear. Especially in his earliest essays (most notably Sellars LRB), the kind of behaviorism initially endorsed seems substantially stronger than Sellars’s later distinction between methodological and philosophical behaviorism. 11 While the distinction between methodological and substantive behaviorism is originally credited to Watson (in at least one form), the most likely sources for Sellars’s adoption of the distinction would be de Laguna (1927) or Bergmann (1956) (a piece Sellars would have had access to substantially earlier than its publication date). 12 Although Sellars does offer arguments in support of his “causally reducible, logically irreducible” claim (see especially Sellars SSMB), much of the power of his insistence on this point is based on simply showing the reader how difficult it would be to account for human agency without such terms, and why such terms (e.g., normative, mentalistic) are not saying the same thing as purely descriptive terms. 13 For an argument that follows this line of thinking about human experience, see Johnson (2014). 14 Although I will briefly discuss some of the philosophers who motivated Sellars’s own behaviorism (most notably Gilbert Ryle), this article primarily focuses on behaviorism’s status as a science. Sellars clearly indicates his initial satisfaction with Ryle’s behaviorism, despite disagreeing with his treatment of inner episodes. See Sellars EPM, 256. 15 For a historically grounded account of changes in notions of behaviorism and radical behaviorism, see Schneider and Morris (1987). 16 See Smith (1986, 69–71); Staddon (1993, 22). 17 I must thank Boris Brandhoff and Jay Garfield for suggesting textual evidence for the connection between Tolman and Sellars. 18 See the conclusion of Olen (2016a) for my arguments about how we ought to treat the history of philosophy. The upshot for my arguments here is that too little evidence exists (in terms of archival resources, correspondence, or citation) to strongly grounded claims of influence between specific behaviorists and Sellars. While it’s clear that Sellars read many behaviorists discussed in this paper, I want to stop short of claiming any direct line of influence. The argument here is simply that Sellars’s positions falls out of a narrative that, in part, structures transitions in the history of American psychology. I take this issue up again in my conclusion. 19 One might suggest that Sellars’s emphasis on the difference between simply making sounds and expressing a piece of language falls directly out of Tolman’s

Origins of Wilfrid Sellars’s Behaviorism  195 distinction between performance and learning (see Tolman 1955 and 1954), although it just as plausibly comes from Kant’s ethical writings. In fact, Sellars runs the two sources together in at least one instance—see Sellars LRB, 124. 20 Sellars’s philosophy should be described as ‘interlocking,’ so much so that it’s difficult to see how one can remove one aspect without thereby bringing numerous other commitments along for the ride. This kind of interconnectedness can be seen at the beginning of Chapter 7 in Sellars SM. 21 This claim is found in Sellars MFC, 431. 22 Although exhibiting difference concerns, the initial exploration of the connection between Sellars’s behaviorism and his conception of meaning is found in Ausonio Marras’s work. See Marras (1976 and 1977).

References Bergmann, Gustav. 1956. ‘The Contribution of John B. Watson’. Psychological Review, 63: 265–76. De Laguna, Grace. 1927. Speech: Its Function and Development. Bloomington: Indiana University Press. DeVries, Willem. 2005. Wilfrid Sellars. Montreal: McGill-Queen’s University Press. DeVries, Willem and Triplett, Timm. 2000. Knowledge, Mind, and the Given: Reading Wilfrid Sellars’s “Empiricism and the Philosophy of Mind”. Indianapolis, IN: Hackett Publishing Company. Johnson, Mark. 2014. ‘Experiencing Language: What’s Missing in Linguistic Pragmatism?’ European Journal of Pragmatism and American Philosophy, 6: 14–27. Marras, Ausonio. 1976. ‘Sellars’ Behaviorism: A Reply to Fred Wilson’. Philosophical Studies, 30: 413–18. Marras, Ausonio. 1977. ‘The Behaviorist Foundation of Sellars’ Semantics’. Dialogue, 16: 664–75. Mills, John. 1998. Control: A History of Behavioral Psychology. New York: New York University Press. Olen, Peter. 2016a. Wilfrid Sellars and the Foundations of Normativity. London: Palgrave Macmillan. Olen, Peter. 2016b. ‘Consequences of Behaviorism: Sellars and De Laguna on Explanation’. Philosophy of the Social Sciences, 47: 111–31. O’Shea, Jim. 2007. Wilfrid Sellars: Naturalism with a Normative Turn. Cambridge: Polity Press.Salzinger, Kurt. 2008. ‘Skinner’s Verbal Behavior’. International Journal of Psychology and Psychological Therapy, 8: 287–94. Schneider, Susan and Morris, Edward. 1987. ‘A History of the Term Radical Behaviorism: From Watson to Skinner’. The Behavior Analyst, 10: 27–39. Sellars, Wilfrid (LRB). 1949/2005. ‘Language, Rules and Behavior’. In Pure Pragmatics and Possible Worlds: The Early Essays of Wilfrid Sellars. Edited by Jeffrey Sicha, 117–34. Atascadero, CA: Ridgeview Publishing Company. Sellars, Wilfrid (SSMB). 1953/2005. ‘A Semantical Solution of the Mind—Body Problem’. In Pure Pragmatics and Possible Worlds: The Early Essays of Wilfrid Sellars. Edited by Jeffrey Sicha, 186–214. Atascadero, CA: Ridgeview Publishing Company. Sellars, Wilfrid (SRLG). 1954/1963. ‘Some Reflections on Language Games’. In Science, Perception and Reality, 321–58. Atascadero, CA: Ridgeview Publishing Company.

196  Peter Olen Sellars, Wilfrid (EPM). 1956/2000. ‘Empiricism and the Philosophy of Mind’. In Knowledge, Mind, and the Given: Reading Wilfrid Sellars’s “Empiricism and the Philosophy of Mind”. Edited by Willem deVries and Timm Triplett, 205–76. Indianapolis, IN: Hackett Publishing Company. Sellars, Wilfrid (PSIM). 1962/1963. ‘Philosophy and the Scientific Image of Man’. In Science, Perception and Reality, 1–40. Atascadero, CA: Ridgeview Publishing Company. Sellars, Wilfrid (MFC). 1974. ‘Meaning as Functional Classification’. Synthese, 27: 417–37. Sellars, Wilfrid (BLM). 1980. ‘Behaviorism, Language, and Meaning’. Pacific Philosophical Quarterly, 61: 3–30. Skinner, Burrhus. 1945. ‘The Operational Analysis of Psychological Terms’. Psychological Review, 52: 270–8. Skinner, Burrhus. 1957. Verbal Behavior. New York: Appleton-Century-Crofts Inc. Smith, Laurence. 1986. Behaviorism and Logical Positivism: A Reassessment of the Alliance. Stanford: Stanford University Press. Staddon, John. 1993. Behaviorism: Mind, Mechanism and Society. London: Gerald Duckworth & Co. Ltd. Tolman, Edward. 1932/1949. Purposive Behavior in Animals and Men. Los Angeles: University of California Press. Tolman, Edward. 1948. ‘Cognitive Maps in Rats and Men’. The Psychological Review, 55: 189–208. Tolman, Edward. 1954. ‘Learning’. Annual Review of Psychology, 5: 27–56. Tolman, Edward. 1955. ‘Principles of Performance’. The Psychological Review, 62: 315–236. Vargas, Ernst, Vargas, Julie and Knapp, Terry. 2007. ‘B. F. Skinner’s Analysis of Verbal Behavior: A Chronicle’. Brazilian Journal of Behavioral and Cognitive Therapy, 9: 21–38. Watson, John. 1925. Behaviorism. London: Routledge & Kegan Paul and Trench, Trubner & Co Ltd. Weiss, Albert. 1925. ‘Linguistics and Psychology’. Language, 1: 52–7.

10 Sellars and Carnap Science and/or Metaphysics Carlo Gabbani

1.  Sellars and Carnap: Some Factual Premises Sellars begins his Autobiographical Reflections by recalling an encounter with Rudolf Carnap in the late 1940s: At the time, I was discussing his Logische Aufbau in my seminar in Philosophical Analysis, and we had scarcely settled in the car for the return journey when I began to bombard my captive audience with questions. I have long since forgotten the detail of what I was after, but I vividly remember that his first reaction was to expostulate, “But that book was written by my grandfather!” The aptness of this remark strikes me anew as I attempt to reconstruct, in outline, the philosopher stages which preceded me—here—now. (Sellars AR, 277) This seems to me the best starting point for a reflection on the relationship between Sellars and Carnap. It also shows the importance and the difficulty of the task—the importance, given that the Auseinandersetzung with Carnap and the tradition of logical empiricism is a central aspect of Sellars’s philosophy and of American philosophy in general. This is due, among other things, to the fact that many leading figures of European culture, including Carnap himself, as well as Hans Reichenbach, Carl Gustav Hempel, Philipp Frank and Herbert Feigl were forced into exile. But this is also a difficult task, even for a reason already evident from this quotation: the main theses, even if not the spirit, of logical empiricists, and of Carnap in particular, changed significantly over the course of a long philosophical career. The first textual evidence that points to a systematic analysis of Carnap’s philosophy by Sellars is an unpublished and undated typescript, titled Some Comments on Carnap’s ‘Scheinproblem’ which focuses especially on Carnap’s theory of knowledge. The terminus post quem of this text is the year 1936, because Sellars made reference (p. 6) to Ayer’s essay Language, Truth and Logic (published at the beginning of that year; cf.: Ayer 1936). It is a graduate student paper and I am inclined to believe that it was written

198  Carlo Gabbani by Sellars as a Ph.D. student at Harvard University (autumn 1937–spring 1938), when he attended the course on logical empiricism and Carnap’s philosophy taught by Quine in the fall of 1937 (cf.: Quine 1985, 130), as Sellars himself recalls in his autobiography.1 The text can now be found in the Sellars’s Collection in Pittsburgh’s Archives of Scientific Philosophy.2 Even if Sellars’s studies were strongly influenced by his long stays in Europe, there is no evidence indicating that he became familiar with logical empiricism during those European sojourns. Thus, we might conclude that Sellars’s acquaintance with logical empiricism probably depends more on Quine’s experience in Europe than on his own. Along with the correspondence between the two (from 1947 to 1962),3 some of the first evidence of Carnap’s extensive discussion of Sellars’s thought is a text, dating back to 1954–55, titled Remarks on Physicalism and Related Topics: Discussions with Wilfrid Sellars. This twelve-page paper is a very interesting document, but largely unknown, because it appeared in mimeographed form only.4 In the paper, Carnap focuses on, and takes issue with Sellars’s peculiar form of emergentism, as well as with Sellars’s distinction between two concepts of ‘physical’ (physical1/physical2),5 defending a traditional physicalistic perspective.6 Coming to more substantive points, Sellars’s engagement with Carnap’s ideas has been long, broad and very influential; several studies have already addressed this very point (see, for instance, Carus 2004; Westphal 2015). Recently, Robert Brandom has underlined how much Sellars’s focus on “the new way of words” in his first published essays is indebted to Carnap.7 And he has also suggested that “Sellars modeled his own Kantian ‘metalinguistic’ treatments of modality and the ontological status of universal explicitly on ideas of Carnap” (Brandom 2015, 33). Obviously, there are also aspects of their philosophies that are traditionally considered to be in stark contrast. For instance, Sellars’s attack on the ‘Myth of the Given’ is widely regarded as the locus classicus of his conflict with empiricism, since Sellars criticizes the very idea of “selfauthenticating episodes” of experience, interpreted as “the unmoved movers of empirical knowledge” and denounces this conception of cognition as a distinctive feature of empiricism, or more precisely as “the framework in which traditional empiricism makes its characteristic claim that the perceptually given is the foundation of empirical knowledge” (Sellars EPM, 169–70). However, I will not deal with this subject.8 For reasons of space, I will concentrate instead on their dialogue on the status of abstract entities and on their respective visions of the status of scientific knowledge, comparing Sellars’s analyses with the doctrines defended by Carnap in the same period. I will suggest that beneath some real and relevant convergences (or dependences) between them, there are deep and very significant differences rooted in different approaches to science, metaontology and metaphysics in general.

Sellars and Carnap  199

2.  The Status of Abstract Entities Sellars’s contribution to the volume of Library of Living Philosophers on Carnap (Sellars EAE) deals with an issue that touches on the most important subject of dialogue between them, namely, the status of abstract entities. This was also a crucial issue for Sellars’s research at that time and in the same year (1963) an important essay, titled Abstract Entities, also appeared (see: Sellars AE). Sellars came back to these issues many times in these years (see: Sellars GE, TTG, CAERP), because he was working on a nominalistic reinterpretation of the linguistic framework of abstract entities, which was interconnected with his general approach to semantics and also related to his naturalistic perspective. Here Sellars tries to offer an alternative to Platonism, which he meant to be “a thesis in the psychology of the higher processes” and, more specifically, as the idea that “the phenomena of meaning (aboutness or reference) involves some sort of commerce (usually spoken of in terms of ‘intuition’, ‘apprehension’ or ‘awareness’) between persons and abstract entities” (Sellars EAE, 442). The kernel of Sellars’s nominalism concerns more than the existence of abstract entities, the idea that they play some role in mental activity. On this point, Kevin Scharp has underlined that already at the end of the 1940s Sellars had written: the difference between the Platonist and the nominalistic empiricist with respect to universals (and propositions) does not consist in the platonist’s saying ‘There are universals’ and the nominalist’s saying ‘No, there are no universals,’ but rather in the platonist’s speaking of psychological relationships between minds and universals, whereas the nominalist finds this to be nonsense. (Sellars LRB, 305; cf.: Scharp 2012, 371) This interpretation is closely associated with Sellars’s non-relational approach to meaning and intentionality. He intends to oppose this Platonism with what he calls “psychological nominalism”: “the denial of the claim, characteristic of the realistic tradition, that a ‘perception’ or ‘awareness’ of abstract entities is the root mental ingredient of mental acts and dispositions” (Sellars EAE, 445). A key point in Sellars’s perspective, however, is the idea that nominalism and the rejection of Platonism do not imply the abandonment of the “linguistic framework of abstract entities,” but rather a reinterpretation of it according to nominalistic means (Sellars EAE, 442).9 In other words, the relevant question for Sellars is: “is it possible for a philosopher consistently to assert ‘There are qualities’ but add, in a different philosophical tone of voice, there really are no such things as qualities and relations?” (Sellars TTC, 322–23). From this point of view, Sellars is more nominalist than

200  Carlo Gabbani many antirealists of his time, as he rejects, for instance, the ontological commitment towards sets, and his ontological landscapes are even more ‘desert’ than Quine’s; but, at the same time, Sellars defends the legitimacy of the presence of abstract terms in our everyday discourse. The core intuition of Sellars’s “syntactical therapy” is that “abstract entities . . . are linguistic entities” (Sellars AE, 49) and that “statements about the abstract entit[ies] . . . are dispensible in favor of statements about conceptual items” (Sellars TTC, 328; cf.: Scharp and Brandom 2007, xvi). The mastery of abstract terms does not involve any “scrutiny of abstract entities” (Sellars GE, 277), but simply a kind of metalinguistic competence with the role of certain singular terms and expressions.10 Moreover, terms like ‘the lion’ are construed by Sellars, in their “explicit form,” as distributive singular terms, or rather, as terms that are “grammatically singular but distribute their reference across an entire class” (deVries 2005, 28). ‘The lion is tawny’ is an equivalent paraphrase of ‘lions are tawny,’ functioning like a singular pro plural. Thus, if one is able to correctly use a common noun and a sentence like the latter, one will also be able to understand and employ the distributive singular term ‘the lion’ without constructing it “as a universal of which lions are instances” (Sellars AE, 53). Sellars, then, introduces expressions like ‘the ˖red˖’ or ˖red˖s (dot— quotation) as a device which applies for all expressions that in whatever language have a conceptual role analogous to that of ‘red’ in English for the English speaker (thus, ‘the ˖red˖’ is not the name of a color, but the name of set of expressions). Accordingly, a sentence like “(in German) ‘rot’ means red” represents for Sellars a material mode of metalinguistic discourse, where ‘rot’ is to be construed as a distributive singular term, and ‘means red’ is to be paraphrased as copula + ‘red’ dot quoted, that is, as the equivalent of: ‘is ˖red˖’. Thus “(in German) ‘rot’ means red” is translated as: “ ‘rot’s is ˖red˖s”, which is to say that (in German) a linguistic expression like ‘rot’ is a form of ˖red˖, or that in German (occurrences of) the word ‘rot’ perform an analogous job as ‘red’ in English for an English speaker. Thus, even to state sentences like: “(in German) ‘rot’ means red” does not imply grasping an abstract property to which both ‘rot’ and ‘red’ refer, but instead recognizing that a given expression of another language has, in that language, a role analogous to that of an expression from our own linguistic practice (cf.: Scharp 2012, 373). As Sellars himself recognizes: Of course, there is always the temptation to say that to be a mental term of the sort to which the general term ‘˖man˖’ applies is to be a mental term which stands for the abstract entity man, or to the character being human. But the above approach can be generalized as the idea that every use of abstract singular terms is essentially classificatory, a matter of classifying conceptual items. (Sellars TTC, 328)

Sellars and Carnap  201 Sellars’s general strategy is inspired by Carnap’s insight on the matter and even his relevant and distinctive innovations recalled here can be seen as an attempt “to refine Carnap’s deflationary expressivist idea that ontological category vocabulary is fundamentally metalinguistic” (Brandom 2015, 240; see also: 262). In fact, in Grammar and Existence (1960), Sellars explicitly appeals to Carnap’s work in order to eliminate “Platonistic anxieties,” asserting that his own analysis paves the way “for a successful use of a therapy essentially the same as the one proposed by Carnap (but which, of course, has a much longer—and indeed, venerable—history).” Sellars also describes the statements of ordinary discourse where abstract terms occur with carnapian labels, as “quasi-syntactical” statements in “the ‘material mode of speech’ ” (Sellars GE, 270).11 Already in The Logical Syntax of Language (Carnap 1934), Carnap argued for the elimination of sets as superfluous (§§ 37–8), while, at the same time, providing an important reinterpretation of “universal words” (§§ 76–8) and the sentences where they occur. Notoriously, Carnap distinguishes between a “material mode of speech” and a “formal mode of speech” (§ 64): in the material mode a (quasi-syntactical) sentence attributes a property to an object, while in the formal mode a (syntactical) sentence attributes a property to the designator of this object. From Carnap’s perspective, the material mode of speech “is a part of ordinary language usage,” but it is also a disguising way of speech and the universal words present in this language “very easily lead to pseudo-problems” (Carnap 1934, 310; emphasis original). According to Carnap, sentences with universal words should be considered quasi-syntactical sentences in the material mode of speech, and they “are formulated as though they refer . . . to objects, while in reality they refer to syntactical forms and, specifically, to the designation of those objects with which they appear to deal” (Carnap 1934, 285). It is then possible to translate the quasi-syntactical sentences of the material mode into equipollent syntactical sentences of the formal mode. Thus, for instance, the sentence “A property is not a thing” (material mode of speech), is translated into: “An adjective (property-word) is not a thing word,” and a philosophical sentence like “Numbers are classes of classes of things,” into: “Numerical expressions are class-expressions of the second level” (Carnap 1934, 297 and 300). Not knowing the difference between these two modes of speech is, in Carnap’s view, the cause of many pseudo-problems in philosophy, while a correct analysis and translation may help us in avoiding them, so as to use universal words without being compromised by the existence of abstract entities. This basic affinity between Sellars’s nominalism and Carnap’s empiricism is not surprising: “Empiricists are in general rather suspicious with respect to any kind of abstract entities like properties, classes, relations, numbers, propositions, etc.” says the incipit of Empiricism, Semantics, and Ontology, and the essay follows with words very similar to those we read, some

202  Carlo Gabbani years after, in Sellars: Carnap aims to demonstrate that “the acceptance of a language referring to abstract entities . . . does not imply embracing a Platonic ontology but is perfectly compatible with empiricism” (Carnap 1950, 205–6). Thus, as you would expect, in his reply to Sellars’s essay for the Library of Living Philosophers, Carnap claims to agree (even if with a caveat) with Sellars’s rebuttal of the Platonistic conception.12 Nevertheless, Carnap also notes that he intends to refute Platonism as far as it is meant as the idea that causal relations hold between persons and abstract entities (cf. Carnap 1963, 924–5), but that he does not intend to exclude the existence of logical (and even psychological) relations between persons and abstract entities. He writes: “I would not reject, as Sellars seems to do, all factual or descriptive relations between material objects and abstract entities, at least not if “relation” is understood in the wide sense which is customary in modern logic.” And then, more significantly, he adds: It seems to me that some psychological concepts may be regarded or reconstructed as relations (in the wide sense of the logical terminology, not in the causal sense) between a person and an abstract entity; e.g., believing may be taken as a relation between a person and a proposition, and thinking-of as a relation between a person and a concept (intension or sense) and the like. In particular, there seems to be no objection to the use of relations of this kind in a theoretical language. (Carnap 1963, 925; emphasis original) This is a very important point of divergence. From a Sellarsian point of view, Carnap’s treatment of universals proves to be too ‘contaminated’ by Platonism and far from “psychological nominalism” no less than many forms of realism, as Carnap is apt to explain concepts and references in terms of a relation between persons and abstract entities. According to André Carus (2004, 334): “It is just this apparently profligate permissiveness that annoys Sellars . . . he regards such uses of abstract entities as obvious hypostatizations, and as beholden to the ‘Myth of the Given’.” More generally, Carnap’s reconstruction of concepts and intentional states in the passage quoted above clearly clashes with Sellars’s non-relational account of meaning and aboutness. From this point of view, Sellars’s flight from Platonism is more radical than Carnap’s, as it also implies eschewing any interpretation of the meaning of a concept in terms of “psychological relationships between minds and universals” (Sellars LRB, 305), while it is not sufficient to simply refute the existence of relations of a causal kind between persons and abstract entities. It thus comes as no surprise that Sellars deemed disastrous “Carnap’s formalization of semantical theory in terms of a primitive relation of designation which holds between words and extralinguistic entities.”13 In short, Sellars finds here what seem to him an inadequate theory of meaning and an unacceptable theory of abstract entities merged together in a semantic which he cannot accept. This is the reason why he ultimately judges Carnap’s contribution to be both pivotal and insufficient or, at least,

Sellars and Carnap  203 in need of a better clarification of the nature and role of the “designation relation” so central to it.14 It is worth mentioning that in the late Carnap the possibility of attributing a (non-causal) role to abstract entities in the reconstruction of some psychological concepts is combined with the idea that quantified variables of a framework express the framework’s ontological commitments. Accordingly, if quantified variables referring to abstract entities are introduced and accepted (within a “theoretical” framework), it seems that abstract entities are necessarily among the entities to which one is committed. For this reason, Quine himself includes Carnap among the realists on abstract entities, by saying that logicism is the new name of realism (cf.: Quine 1953, 14). But this difference between Carnap and Sellars concerning abstract entities is probably symptomatic of a wider divergence regarding the nature and significance of ontological commitment, the approach to traditional metaphysical problems, and the aim of philosophy. This seems to be clear on closer inspection of their different understanding of the role and status of scientific knowledge.

3.  The Status of Scientific Knowledge While Sellars and logical empiricism are perhaps too easily opposed in regard to the problem of the ‘given,’ there is the danger of aligning them too closely with respect to the so-called “scientific image.”15 There is indeed a true convergence between Sellars and the tradition of logical empiricism, for instance, in terms of conceiving science as a unified (or unifiable) enterprise: “the Unified Science” of logical empiricists, on the one hand, and “the Scientific Image” (the singular) of Sellars, on the other.16 However, the differences between the two perspectives are perhaps more relevant and interesting. Here, I would like to consider especially the two following issues: (1) In Empiricism and the Philosophy of Mind, Sellars famously asserts: speaking as a philosopher, I am quite prepared to say that the common sense world of physical objects in Space and Time is unreal—that is that there are no such things. Or, to put it less paradoxically, that in the dimension of describing and explaining the world, science is the measure of all things, of what is that it is, and of what is not that it is not (Sellars EPM, 173). Would Carnap have agreed with such a statement and the protagorean paraphrase? (2) Sellars asserts: “it is ‘scientific objects’, rather than metaphysical unknowables, which are the true things-in-themselves” (Sellars SM, 143), and: As I see it . . . a consistent scientific realist must hold that the world of every day experience is a phenomenal world in the Kantian sense,

204  Carlo Gabbani existing only as the contents of actual and obtainable conceptual representations, 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 (Sellars SM, 173).17 Again, would Carnap (and the logical empiricists) have shared such a stance?18 Protagora’s dictum was quoted (in its original version) also in the manifesto of the Vienna Circle, the Wissenschaftliche Weltauffassung: “Everything is accessible to man; and man is the measure of all things,” that is to say “the scientific world-conception knows no unsolvable riddle” (emphasis original) and what can be meaningfully and scientifically asked can also find, in principle, a meaningful, scientific answer.19 In more general terms, the idea of science as the measure of human knowledge (intended as Erkennen) is certainly consistent with the perspective of logical empiricism and also from a logical empiricist’s point of view science may be regarded as the measure of all description and explanation. But—and there is a big but—in the tradition of logical empiricism, science (present science as well as future, ideal science) is generally not regarded as the measure of ‘reality,’ or, at least, not in a metaphysical sense. This point is an essential one and its origin can be traced back to the years of Vienna Circle. But let me concentrate here on the doctrine developed by Carnap in the post-Vienna stage of his career, especially in Empiricism, Semantics, and Ontology (1950).20 From this perspective, it seems certainly possible for someone to decide that science is “the measure of all things,” that is, the measure of all one’s ontological commitments. But what is the nature, the meaning and the rationale of this decision? First of all, I am speaking here of a ‘decision,’ because, for Carnap the adoption of a certain conceptual framework in describing and explaining is “a matter of practical decision”: A decision external to the framework in question, which can (and should) be justified, but which is not in itself true or false, right or wrong, and is not of a cognitive or metaphysical nature. Once this external choice has been made, issues concerning ontological commitments are fruitful and can be internally settled. Because for Carnap “to be real . . . is to be an element of the system,”21 that is, of the conceptual system one accepts, one will simply be committed to the existence of all (and only) the items included in the framework(s) one adopts. Thus, ontological commitments depend on the conceptual frameworks one adopts in describing and explaining the world and, in conformity with Quiness lesson, the ontological commitments of a theory are indicated by the bound variables present within it. Therefore, one can certainly determine that “in the dimension of describing and explaining the world, science is the measure of all things,” if one

Sellars and Carnap  205 decides that science is for oneself the only framework to be used to describe and explain the world. Once this choice has been made, science will then prove to be the measure of all our ontological commitments. But to say, by the same token, that science is “the measure of all things” (that is, that one accepts all and only the entities that belong to its framework) for Carnap does not imply recognizing the overall reality of the scientific image in relation to the world in itself. As he writes: “An alleged statement of the reality of the system of entities is a pseudo-statement without cognitive content” (Carnap 1950, 214). This means that for Carnap it is not possible to establish that science provides us (or will provide us) with knowledge of things in themselves, or of reality. Here we are touching on a fundamental aspect of Carnap’s empiricism: science is not a way to access things in themselves, and the very distinction between phenomena and things in themselves is a metaphysical relic which cannot be defended any longer. This is also connected to the traditional distinction in Carnap (and in logical empiricism) between an empirical concept of reality (already present in the Aufbau: “empirische Wirklichkeitsbegriff” or “Begriff der (empirischen) wirklichkeit”) and a metaphysical concept of reality (“metaphysische Wirklichkeitsbegriff”).22 The empirical concept of reality is the only legitimate and existing one within science, and it can be used simply to distinguish, for instance, between a mountain which really exists (i.e., whose existence has been empirically ascertained), from a mythical, or imaginary mountain. The metaphysical concept of reality, by contrast, is the one used by metaphysicians when they want to assert the reality of something in an absolute way, that is, to assert its existence as independent from every knowing subject (“Unabhängigkeit vom erkennenden Bewusstsein”)23 and from any conceptual framework: its existence as a thing in itself.24 Now, this second, metaphysical concept of reality is for Carnap a residue of the old, traditional metaphysics to be abandoned. But this means that, in Carnap’s perspective, presumably, two central moves of Sellars’s philosophy amount to (related) metaphysical mistakes:

• One cannot make an ‘external,’ theoretical assertion about the truth-

•

fulness of an overall framework, as Sellars does (“as a philosopher”) when he speaks of the scientific image/the manifest image as a whole. For instance: “To say that there are no such things as the physical objects of the perceptible world is, of course, to make a point about the framework of physical objects, not in it” (Sellars PHM, 97; italics in the original); One cannot use a metaphysical concept of reality in order to state how a framework as a whole deals with the ‘true’ world, as Sellars does, for instance, when he says that the objects of a final, ideal, reformed scientific image are metaphysically real, or “the true things-in-themselves,” no more unknowable.

206  Carlo Gabbani This is probably one of the more important aspects of divergence between Sellars and Carnap. And, obviously, “Carnap’s view on ontology” is at stake here. Significantly, Sellars rejects both (i) the idea of quantification as a general guide to the ontological commitments of a language,25 and (ii) Carnap’s other (independent) idea that ontological matters are issues internal to a given framework whose adoption is based on a pragmatic decision. Let us elaborate on this second point. Sellars has explicitly stated his perplexities about Carnap’s thesis concerning the adoption of a framework: The external question ‘Shall I accept such and such a form of language?’ is, as Carnap points out a practical question in that it calls for ‘decision rather than an assertion’. . . . But although a question of the form ‘Shall I . . .?’ calls indeed for decision, it is generally sensible to ask of a decision ‘Is it reasonable?’ or ‘Can it be justified?’ and these questions call for assertion rather than a decision. Thus the question inevitably arises, Is it proper to ask of a decision to accept a framework of entities, ‘It is reasonable?’ ‘Can this decision be justified, and if so, how?’ This is the crux of the matter, and on this point, it must be remarked, Carnap’s discussion is less incisive. (Sellars EAE, 433) And Sellars then adds: At times . . . he seems to tell us that the demand for a justification is improper. On the other hand, only a few sentences further on he writes that ‘the acceptance can only be judged as being more or less expedient, fruitful, conductive to the aims for which the language is intended.’ Here, as on several other occasions, he implies that such a decision can be justified, that is, shown to be reasonable. As to the nature of such justification, however, he gives no more than a few obscure hints. (Sellars EAE, 433)26 Carnap’s divide between internal and external questions certainly has many problematic features, presuppositions or implications, and many objections have always been raised against it (starting from Quine 1951). But Sellars’s charges here seem not completely convincing. In Carnap’s view, it is possible to justify the rationality of the (external) decision to choose and accept a given framework (or conceptual structure) as more adequate than its alternatives. And Carnap also recognizes that such a decision “although itself not of a cognitive nature, will nevertheless usually be influenced by theoretical knowledge” (Carnap 1950, 208). Nevertheless, this rational justification cannot appeal to the character of reality or to the fact that reality ‘really’ (in a metaphysical sense) includes precisely the entities introduced by the adopted framework. The reasons for accepting a certain framework are instead practical evaluations concerning the possibilities, limits, and virtues

Sellars and Carnap  207 (efficiency, fruitfulness, simplicity, etc.) it offers us in pursuing our epistemic purposes in a given cognitive situation. This choice can also be influenced by already existing knowledge of a certain number of facts (knowledge upon which converges the community which will decide about the adoption of a new framework). But this already existent intersubjective knowledge (i) is in no way absolute knowledge of reality, and (ii) cannot in any way establish the truthfulness of the new conceptual framework that is then adopted also under the influence of this knowledge. From this point of view, it seems that Carnap’s approach does not prevent us from rationally explaining and justifying the acceptance of a given framework. It simply prevents us from saying that we adopt such a framework because it is the adequate or correct one, on the basis of how things really are, and thus it prevents us from a metaphysical kind of commitment towards a conceptual framework, or from granting the framework we adopt the role and value giving us access to reality in itself.27 Thus, Carnap may certainly tolerate, and also share, the idea that science, in describing and explaining, is “the measure of all things,” but not the view according to which science is also the royal road to “things in themselves”: because there is no such road.28 The crucial point is that for Carnap, as he clearly states: “The acceptance of a linguistic framework must not be regarded as implying a metaphysical doctrine concerning the reality of entities in question” (Carnap 1950, 214).29 For this reason, Empiricism, Semantics, and Ontology can be considered as a reaction to Quine’s thesis according to which “the use of bound variables to refer to abstract entities” implies metaphysical realism concerning the latter (cf.: Blatti-Lapointe 2016). In Carnap’s opinion, bound variables express ontological commitment, but this ontological commitment does not have a metaphysical character. And even Carnap’s conviction to remain neutral in the dispute between scientific realists and instrumentalists is based on the idea that to assert the existence of unobservable entities is simply to accept a framework within which they are present as quantified variables. As a consequence, it appears to him equally wrong and senseless to say: ‘They are present in the conceptual framework I adopt and they also really exist,’ as well as: ‘they are present in the conceptual framework I adopt but they do not really exist.’ For this reason, it is not correct to regard the late Carnap (the Carnap of the Philosophical Foundations of Physics, for instance)30 as an instrumentalist in the traditional (metaphysical) sense: Both traditional scientific realism and antirealism are flawed and improper in Carnap’s view, as far as both are based on a metaphysical understanding of the issue.31 It is worth noting that Carnap’s entire ontological perspective here is deeply rooted in his lifelong rejection of metaphysics, but it is not directly based on the idea of a primacy of the observation framework, or, at least, not on a realist (in a metaphysical sense) understanding of it. Furthermore, Carnap’s rejection of a metaphysical concept of reality (or of ‘things in themselves’), his conviction that it is impossible to establish the ontology

208  Carlo Gabbani of the world in se, or to attain knowledge of an alleged ‘noumenal’ world is an insight that is certainly couched within the framework of the distinction between internal and external questions, when the latter is formulated; however, in Carnap’s philosophy this insight seems to be older, more basic and quite autonomous from such a distinction (as well as from the analyticsynthetic distinction). Sellars, by contrast, has been rightly assessed to be “one of the greatest systematic metaphysicians of the mid-twentieth century” (Brandom 2002, 264), and in his writings we can read not only a clear commitment to scientific realism, but also the idea that “realism in any of its forms is a paradigm of metaphysics,”32 or that (in “the long run”) “the ‘truth’ about ‘what really exists’ ” is the “formal, final and efficient cause” of “scientific enterprise” (Sellars SRII, 204). This is a perspective that for Carnap seems to remain unacceptable, even after every possible “liberalization of empiricism.” This different approach may help explain why, even when both Carnap and Sellars seem to agree that a deeper semantic analysis may contribute to better understanding an old philosophical conundrum, they often diverge in their conclusions. For Carnap such philosophical analysis should help us eliminate and dissolve old, senseless metaphysical disputes, which is to say, it should lead us beyond both the horns of a (metaphysical) dilemma. As he sympathetically writes in Empiricism, Semantics, and Ontology, speaking of the heritage of the Vienna Circle: The Circle rejected both the thesis of the reality of the external world and the thesis of its unreality as pseudo-statements; the same was the case for both the thesis of the reality of universals (abstract entities in our present terminology) and the nominalistic thesis that they are not real and that their alleged names are not names of anything but merely flatus vocis (It is obvious that the apparent negation of a pseudostatement must also be a pseudo-statement). It is therefore not correct to classify the members of the Vienna Circle as nominalists, as is sometimes done. (Carnap 1950, 215) Sellars’s philosophical work, on the contrary, aims at giving a better answer and reaching a deeper understanding of the same, old but not meaningless, metaphysical issues. In other words, while Sellars’s philosophy tries to make a contribution to the traditional philosophical disputes, Carnap’s philosophy tries to supersede those same traditional philosophical disputes. There is another historical aspect that may help explain Sellars’s approach as opposed to Carnap’s. The continuity between common sense and science is of great relevance to the thinking of logical empiricists, while in Sellars’s philosophy what is of a paramount importance is the possibility or the risk of a ‘clash’ between common sense, or, more correctly, the manifest image and science. But if the ‘scientific image’ is regarded not simply as broadening, deepening and enriching the manifest one, but also as contrasting,

Sellars and Carnap  209 denying and systematically undermining it, the question may arise: which truly describes the real world? And one may be tempted to say that scientific image reveals to us the ‘real’ world or that, in comparison to the world of scientific image, the world of common sense is only a ‘phenomenal’ world. From this point of view, one may also interpret Sellars’s ideology of a ‘scientia mensura’ as if he were ‘simply’ asserting that scientific objects are the only entities upon which causal relevance may be bestowed and that they are also responsible for the disguising appearances of the phenomenal world of the manifest image.33 Now, it is certainly possible that two different conceptual frameworks come into conflict (not only locally, but systematically), and certainly “one conceptual framework can be more ‘adequate’ than another” (Sellars SM, 134; see also: 138–9). But I suppose that, from an empiricist’s point of view, a correct comparison and a correct justification of this judgment could never make reference to the fact that one of these frameworks allows us to know or to approach things as they are in themselves. As sympathetic to science as a logical empiricist may be, Carnap never considered science as a means to ‘producing’ metaphysics and he would never (at any stage of his career) have entitled a book “Science and Metaphysics,” as Sellars did, but rather have opted for “Science aut Metaphysics.”34

Notes 1 “In the fall of 1937, I went to Harvard to work for my ‘trade-union card’. I remember taking courses with D. W. Prall (Spinoza), C. I. Lewis (Theory of Knowledge), R. B. Perry (Contemporary Philosophy), C. L. Stevenson (Hume), and W. V. Quine (Logical Positivism). It was the last of these which I found most challenging. I had already convinced myself that Ayer’s Language, Truth, and Logic was a brilliant tour de force which led nowhere. . . . In Quine’s course, however, the central figure was Carnap, from the Aufbau to the Logical Syntax of Language” (Sellars AR, 287). 2 Cf. Wilfrid S. Sellars Papers, 1899–1990, ASP.1991.01, Archives of Scientific Philosophy, Special Collections Department, University of Pittsburgh, Box 7, Folder 17, 13-page text, with handwritten notes on the back (available also on-line). 3 Cf. Wilfrid S. Sellars Papers, 1899–1990, ASP.1991.01, Archives of Scientific Philosophy, Special Collections Department, University of Pittsburgh, Box 159, Folder 1 (available also on-line). Besides, Sellars’s Acquaintance and Description Again (Sellars AD), for instance, is quoted in: Carnap (1950, 220). 4 Cf. Carnap 1955 (see Rudolf Carnap Papers, 1905–70, ASP.1974.01, Special Collections Department, University of Pittsburgh, Box 86, Folder 6, available also on-line; further references are to this version). According to the “Bibliography of the Writings of Rudolf Carnap” compiled by Arthur J. Benson, the text is “Based on informal conversations held in Los Angeles, 29 Dec. 1954 to 1 Jan. 1955” (Schilpp 1963, 1045). 5 “Emergence is one form taken by a negative answer to the question: ‘Could a world which includes minds be described with the same primitive predicates (and laws) as a mindless universe?’ ” (Sellars RNWW, 453–4). For Sellars’s two different characterizations of the ‘physical’, see Sellars SSIS, 401–2. 6 Carnap writes: “Once dualism is abandoned, there seems to me good reason left for singling out the occurrence of certain new micro-structures from the

210  Carlo Gabbani occurrence of other new micro-structures, and declaring that the first ones are connected with new qualities or sense-data while the latter are just new physical structure” (Carnap 1955, 6). And he adds: “It seems that emergentism, even in a non-metaphysical version, is in danger of making absolute what is in fact gradual, and of making fundamental what is only a secondary factor” (1955, 7–8). 7 See Brandom (2015, 4). 8 It may be useful to recall at least that, in Empiricism and Abstract Entities, Sellars expressed his view on Carnap’s commitment to the framework of givenness: “if Carnap had once been asked, How is the acceptance of the framework of sense-data to be justified? He would, in effect, have replied ‘because colors, sounds, etc. are given,’ where further questioning would have made it clear that this givenness did not involve a convert [sic] use of a (however rudimentary) symbolic framework of sense-data. I think it is obvious that many empiricists have taken this line. Whether or not Carnap still does (or has ever done) it is not easy to say. Certainly he has, on many occasions, availed himself of the philosophical jargon of givenness, and nowhere he has explicitly discussed and rejected the epistemological views it embodies” (Sellars EAE, 435). 9 Accordingly, what has to be shown is that: “The linguistic framework of abstract entities, which is such an indispensable part of human discourse . . . does not involve a commitment to Platonism” (EAE, 467). 10 It is not possible to adequately reconstruct Sellars’s strategy here, cf. Loux (1977); deVries (2005, 67–93); O’Shea (2007, 63–76); Brassier (2014); Brandom (2015, Chap. 7). 11 And in Empiricism and Abstract Entities he recognizes: “To the creation of this truly revolutionary situation which is just beginning to make itself felt, Carnap’s Logical Syntax of Language and Introduction to Semantics have contributed at least as much as any other single source” (Sellars EAE, 468). 12 “I agree with Sellars in rejecting this Platonistic conception, as it is represented, e.g., in Sellars’s quotation from Alonzo Church” (Schilpp 1963, 924). 13 “This reconstruction commits one to the idea that if a language is meaningful, there exists a domain of entities (the designata of its names and predicates) which exists independently of any human concept formation” (Sellars LOT, 109). In Sellars’s opinion, many of the flaws in Carnap’s account depend on a “much too perfunctory” analysis of “the relation between pure and descriptive semantics” (Sellars EAE, 462); thus Carnap’s studies “provide the essential materials for a non-metaphysical account of abstract entities, but that, by falling to examine more in detail the relation between pure and descriptive semantics, they leave dark corners where metaphysical views can find sanctuary” (EAE, 442). Cf. also Westphal (2015, 45). 14 “Carnap is quite willing to say that the descriptive semantical statement “ ‘fünf’ means a number” asserts that ‘fünf’ stands in the designation relation to a number. He emphasizes that the fact that a number stand in the designation relation no more implies that the number is a datum than facts about electrons imply that electrons are data. . . . But he tells us little if anything positive about the status of this designation relation” (Sellars EAE, 462). It is worth noticing that some years after, Sellars will re-examine Carnap’s doctrine and especially Empiricism, Semantics and Ontology (ESO). In a 1972 talk, he will recognize that in his previous examination he “failed to grasp certain key features of his argument, and that a re-assessment of the bearing of ESO on central issues in ontology is in order” (Sellars IAE, 288). 15 As a matter of fact, distinctions are needed even at the terminological level. According to Carus, “Carnap would never have used terms like ‘scientific image’ and ‘manifest image’ ” (Carus 2004, 327).

Sellars and Carnap  211 16 As Carus notes: “Carnap envisaged the convergence of the increasingly precise and integrated systems of scientific language into a single system. So he could have spoken of ‘the’ scientific image just as Sellars did” (2004, 330–1). 17 And he also observes: “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” (Sellars SM, 50; concerning the limits of Peirce’s conception; see SM, 142). 18 Behind the use of Kantian categories, here there is a radical departure from Kant. Analyzing (and criticizing) this aspect of Sellars’s philosophy, Robert Brandom has written: “Where for Kant, the deliverances of natural (Newtonian) science, no less than the description of the manifest image, describe an empirical nature that belongs to the realm of phenomena, Sellars’s detranscendentalized, naturalized version of the distinction has science limning the realm of the noumena” (Brandom 2015, 59; cf. also: 15 and 25). 19 Chap. 2. The same ideas can be found in Carnap’s Aufbau, see Carnap (1928, § 180). 20 On these aspects of Carnap’s philosophy, see, for instance, Creath (1985); Parrini (1994); Price (1997); Psillos (1999, Chap. 3, and 2008); Bird (2003); Friedman (2008, and 2011); Westphal (2015); Blatti-Lapointe (2016). 21 “To be real in the scientific sense means to be an element of the system; hence this concept cannot be meaningfully applied to the system itself” (Carnap 1950, 207). 22 Carnap (1928, §§ 170 and 176). Obviously, The logical Construction of the World has many features not at all congenial to Sellars. And as he himself, evoking his first approach to logical empiricism, frankly recognized: “I am afraid that I got little out of the Aufbau. I would, I believe, have gained a better appreciation of the power of its technical devices if I had been able to put to one side my violent anti-phenomenalism. Carnap was doing what can’t be done, therefore there must be something wrong about how he is doing it” (Sellars AR, 287). 23 Carnap (1928, 245). 24 The concept of “thing in itself” is criticized and refused as metaphysical by Carnap in Der logische Aufbau; see Carnap (1928, 247). On the contrary, Sellars, in the second chapter of Science and Metaphysics, writes: “The root notion of ‘existing in itself’ is that of existing simpliciter as contrasted with existing as represented, i.e., existing ‘in’ a representing or as ‘idea’ ” (Sellars SM, 36). And he adds (SM, 38): “if the core of the notion of the in-itself is the concept of that which exists simpliciter as contrasted with that which exists as idea (content) or representable, it is clear that Kant adds to this core a theme of ‘unknowability’.” As Sellars himself recognizes: “for Kant, nothing identified in spatio-temporal terms can exist as such in itself” (SM, 44). 25 In Grammar and Existence, Sellars rejects “the idea that the ‘existentially quantified’ formulae of the logician are the counterparts of the statements in everyday discourse in which, to use Quine’s phrase, we make ontological commitments . . . there is no general correspondence between existentially quantified formulae and existence statements” (Sellars GE, 255); and he adds: “to quantify over adjective-common noun-and statement-variables is not to assert the existence of qualities, kinds, or propositions” (267; see also 281). And in The Language of Theories he writes: “I am committing myself to the view that only those existentially quantified statements in which the quantified variable takes names of objects as its substituends have the force of existence statements” (Sellars LOT, 116). Cf.: deVries (2005, 88–9); Westphal (2015, § 6.6). 26 And in the following pages he remarks: “we have found Carnap to be tantalizingly vague as to the circumstances in which it would be reasonable to decide

212  Carlo Gabbani such questions in the affirmative” (EAE, 444). Sellars’s critique is shared and elaborated by Westphal (2015, § 6.7). 27 In The Language of Theories (LOT, 109, footnote 3), Sellars underlines the (unwelcomed) consequences of Carnap’s semantical doctrines on the interpretation of the manifest image. He asserts that, if one accepts Carnap’s theories: “one is committed to the idea that the framework of observable physical things and their properties has an absolute reality”. Probably “absolute reality” here simply means the impossibility of a revisionary approach to the framework of the manifest image. But this could also reveal a misunderstanding: in Carnap’s view the very idea of an “absolute reality” of something (intended as a metaphysical guarantee) results in being meaningless. 28 “It is one thing to endorse, with Carnap, the doctrine that “science is the measure of all things: of those that are, that they are, of those that are not, that they are not”; it is a completely different thing to determine to what that attribution of existence amounts. Science solves the problem of reality by telling us what there is; but science does not tell us what it tells us when it tells us what there is. It is on this topic that Carnap disagreed with the scientific realist” (Coffa 1991, 233). 29 It can be noted that, while Sellars introduces his assertion about the role of science with the clause “as a philosopher,” Carnap underlines that only philosophers have the temptation to wrongly treat external questions (like the one concerning the reality of a total system of entities) as if they were questions of cognitive and metaphysical character. 30 Cf. Carnap (1966). In the second edition (1974, 256), the chapter devoted to the Ramsey-sentence ends with these words: “My own view, which I shall not elaborate here, is essentially this. I believe that the question should not be discussed in the form: ‘Are theoretical entities real?’ but rather in the form: ‘Shall we prefer a language of physics (and of science in general) that contains theoretical terms, or a language without such terms?’ From this point of view the question becomes one of preference and practical decision.” For Sellars’s evaluation of the Ramseysentence (and of Carnap’s employ of it), see Sellars SRII, 177–8. 31 Accordingly, Carnap’s perspective here is not exactly that of the “sophisticated instrumentalist” sketched by Sellars, and according to which: “ ‘Real’ (as contrasted with ‘instrumental’) existence, meaning and truth are limited to objects as conceived at the perceptual level of our current conceptual structure” (Sellars SM, 145). 32 Sellars ISRT, 307; and later on, Sellars speaks of the “very abstract and metaphysical issue of scientific realism” (316). In the same essay, he writes: “Philosophers many years ago amused themselves with the question: Are scientific objects invented or discovered? To this the correct answer is that we invent them and discover that they do the work we require of something that is to count as real” (ISRT, 312). 33 “The thesis I wish to defend, but not ascribe to Kant, . . . 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” (Sellars SM, 49; on this aspect, cf. Brandom 2015, Chap. 1, part II). 34 I have tried to present (and contrast) some aspects of Carnap’s and Sellars’s views on scientific knowledge, but I am well aware that an adequate account should be accompanied by a careful examination of many other aspects, including Sellars’s conception of picturing (see Sellars SM, Chap. 5). However, this task goes far beyond the scope of this essay; cf. O’Shea (2007, Chap. 6).

Sellars and Carnap  213

References Ayer, Alfred Jules. 1936. Language, Truth and Logic. London: Victor Gollancz. Bird, Graham. 2003. ‘Carnap’s Internal and External Questions’. In Language, Truth, and Knowledge. Contributions to the Philosophy of Rudolf Carnap. Edited by Thomas Bonk, 97–131. Dordrecht, Boston and London: Kluwer. Blatti, Stephan and Lapointe, Sandra. (Eds.). 2016. Ontology After Carnap. Oxford: Oxford University Press. Brandom, Robert. 2002. ‘Overcoming a Dualism of Concepts and Causes: The Basic Argument of “Empiricism and the Philosophy of Mind” ’. In The Blackwell Guide to Metaphysics. Edited by Richard M. Gale, 263–81. Oxford and Malden, MA: Wiley-Blackwell. Brandom, Robert. 2015. From Empiricism to Expressivism: Brandom Reads Sellars. Cambridge, MA and London: Harvard University Press. Brassier, Ray. 2014. ‘Nominalism, Naturalism, and Materialism. Sellars’s Critical Ontology’. In Contemporary Philosophical Naturalism and Its Implications. Edited by Bana Bashour and Hans D. Muller, 101–14. New York: Routledge. Carnap, Rudolf. 1928. Der logische Aufbau der Welt: Scheinprobleme in der Philosophie. Hamburg: Meiner 1961. Carnap, Rudolf. 1934. Logische Syntax der Sprache. Wien: Springer. Revised English Translation: The Logical Syntax of Language. London: Routledge & Kegan Paul 1937 (reprint. Chicago and La Salle, IL: Open Court 2002). Carnap, Rudolf. 1950. ‘Empiricism, Semantics, and Ontology’. Revue Internationale de Philosophie, 4: 20–40; In Id., Meaning and Necessity: A Study in Semantics and Modal Logic, 205–21. Chicago: University of Chicago Press 1956. Carnap, Rudolf. 1955. Remarks on Physicalism and Related Topics: Discussions with Wilfrid Sellars. Minneapolis, MN: Minnesota Center for Philosophy of Science (mimeographed version). Typescript available on-line, see: Rudolf Carnap Papers, 1905–1970, ASP.1974.01, Special Collections Department, University of Pittsburgh, Box 86, Folder 6 (references are to this version). Carnap, Rudolf. 1963. ‘Replies and Systematic Expositions’. Schilpp, 859–1013. Carnap, Rudolf. 1966. Philosophical Foundations of Physics: An Introduction to the Philosophy of Science. New York: Basic Books. Carnap, Rudolf. 1974. An Introduction to the Philosophy of Science. New York: Basic Books. Carus, André W. 2004. ‘Sellars, Carnap, and the Logical Space of Reasons’. In Carnap Brought Home: The View from Jena. Edited by Steve Awodey and Carsten Klein, 317–55. Chicago: Open Court. Coffa, Alberto J. 1991. The Semantic Tradition from Kant to Carnap: To the Vienna Station. Cambridge: Cambridge University Press. Creath, Richard. 1985. ‘Carnap’s Scientific Realism: Irenic or Ironic?’ In The Heritage of Logical Positivism. Edited by Nicholas Rescher, 117–31. Lanham: University Press of America. DeVries, Willem A. 2005. Wilfrid Sellars. Chesham: Acumen. Friedman, Michael. 2008. ‘Wissenschaftslogik: The Role of Logic in the Philosophy of Science’. Synthese, 164: 385–400. Friedman, Michael. 2011. ‘Carnap on Theoretical Terms: Structuralism Without Metaphysics’. Synthese, 180: 249–63.

214  Carlo Gabbani Loux, Michael J. 1977. ‘Ontology’. In The Synoptic Vision: Essays on the Philosophy of Wilfrid Sellars. Edited by Cornelius F. Delaney, Michael J. Loux, Gary Gutting and W. David Solomon, 43–72. Notre Dame: University of Notre Dame Press. O’Shea, James R. 2007. Wilfrid Sellars: Naturalism With a Normative Turn. Cambridge and Malden, MA: Polity Press. Parrini, Paolo. 1994. ‘With Carnap Beyond Carnap: Metaphysics, Science and the Realism/Instrumentalism Controversy’. In Logic, Language, and the Structure of Scientific Theories. Edited by Wesley Salmon and Gereon Wolters, 255–77. Pittsburgh and Konstanz: University of Pittsburgh Press-Universitätverlag Konstanz. Price, Huw. 1997. ‘Carnap, Quine and the Fate of Metaphysics’. The Electronic Journal of Analytic Philosophy, 5. http://ejap.louisiana.edu/EJAP/1997.spring/ price976.html. Accessed July 3, 2017. Psillos, Stathis. 1999. Scientific Realism: How Science Tracks Truth. London and New York: Routledge. Psillos, Stathis. 2008. ‘Carnap and Incommensurability’. Philosophical Inquiry, 30: 135–56. Quine, Willard Van Orman. 1951. ‘On Carnap’s Views on Ontology’. Philosophical Studies, 2: 65–72; In Id., The Ways of Paradox and Other Essays, 203–11. Cambridge, MA: Harvard University Press 1976. Quine, Willard Van Orman. 1953. ‘On What There Is’. In From a Logical Point of View, 1–19. Cambridge, MA: Harvard University Press. Quine, Willard Van Orman. 1985. The Time of My Life: An Autobiography. Cambridge, MA and London: MIT University Press. Scharp, Kevin. 2012. ‘Wilfrid Sellars’s Anti—Descriptivism’. In Categories of Being: Essays on Metaphysics and Logic. Edited by Leila Haaparanta and Heikki J. Koskinen, 358–90. Oxford and New York: Oxford University Press. Scharp, Kevin and Brandom, Robert. 2007. ‘Editors’ Introduction’ to In the Space of Reasons: Selected Essays of Wilfrid Sellars. Edited by Kevin Scharp and Robert Brandom, vii–xxv. Cambridge, MA and London: Harvard University Press. Schilpp, Paul A. (Ed.). 1963. The Philosophy of Rudolf Carnap (The Library of Living Philosophers). La Salle: Open Court. Sellars, Wilfrid (RNWW). 1948. ‘Realism and the New Way of Words’. Philosophy and Phenomenological Research, 8: 601–34; In Readings in Philosophical Analysis. Edited by Herbert Feigl and Wilfrid Sellars, 424–56. New York: AppletonCentury-Crofts 1949. Sellars, Wilfrid (AD). 1949. ‘Acquaintance and Description Again’. Journal of Philosophy, 46: 496–504. Sellars, Wilfrid (LRB). 1949. ‘Language, Rules and Behavior’. In John Dewey: Philosopher of Science and Freedom. Edited by Sidney Hook, 289–315. New York: The Dial Press; In Id., Pure Pragmatics and Possible Worlds: The Early Essays of Wilfrid Sellars, 117–34. Edited by Jeffrey F. Sicha. Atascadero, CA: Ridgeview Publishing Company 1980. Sellars, Wilfrid (EPM). 1956. ‘Empiricism and the Philosophy of Mind’. Minnesota Studies in the Philosophy of Science. Edited by Herbert Feigl, Michael Scriven and Grover Maxwell, 253–329. Minneapolis, MN: University of Minnesota Press, 1: 253–329; In Id., Science, Perception and Reality, 127–96. London: Routledge 1963.

Sellars and Carnap  215 Sellars, Wilfrid (GE). 1960. ‘Grammar and Existence: A Preface to Ontology’. Mind, 69: 499–533; In Id., Science, Perception and Reality, 247–81. London: Routledge 1963. Sellars, Wilfrid (LOT). 1961. ‘The Language of Theories’. In Science, Perception, and Reality, 106–26. London: Routledge 1963. Sellars, Wilfrid (CAERP). 1963. ‘Classes as Abstract Entities and the Russell Paradox’. The Review of Metaphysics, 17: 67–90; In Id., Essays in Philosophy and its History, 340–63. Dordrecht: Reidel 1974. Sellars, Wilfrid (EAE). 1963. ‘Empiricism and Abstract Entities’. Schilpp, 431–68. Sellars, Wilfrid (PHM). 1963. ‘Phenomenalism’. In Science, Perception and Reality, 60–105. London: Routledge. Sellars, Wilfrid (AE). 1963. ‘Abstract Entities’. The Review of Metaphysics, 16: 627–71; In Id., Philosophical Perspectives: Metaphysics and Epistemology, 49– 89. Atascadero, CA: Ridgeview Publishing Company 1977. Sellars, Wilfrid (SRII). 1965. ‘Scientific Realism or Irenic Instrumentalism: Comments on J. J. C. Smart’. In Boston Studies in the Philosophy of Science. Vol. 2. Edited by Robert Cohen and Marx Wartofsky, 171–204. New York: Springer Verlag. Sellars, Wilfrid (SM). 1968. Science and Metaphysics: Variations on Kantian Themes. London: Routledge & Kegan Paul. Sellars, Wilfrid (TTC). 1970. ‘Toward a Theory of the Categories’. In Experience and Theory. Edited by Lawrence Foster and Joe W. Swanson, 55–78. Amherst: University of Massachusetts Press; In Id., Essays in Philosophy and its History, 318–39. Dordrecht: Reidel 1974. Sellars, Wilfrid (SSIS). 1971. ‘Science, Sense Impressions, and Sensa: A Reply to Cornman’. Review of Metaphysics, 25: 391–447. Sellars, Wilfrid (IAE). 1974. ‘On the Introduction of Abstract Entities’. In Essays in Philosophy and Its History, 287–317. Dordrecht: Reidel. Sellars, Wilfrid (AR). 1975. ‘Autobiographical Reflections’. In Action, Knowledge and Reality: Critical Studies in Honor of Wilfrid Sellars. Edited by Hector-Neri Castañeda, 277–93. Indianapolis, IN: Bobbs-Merrill. Sellars, Wilfrid (ISRT). 1976. ‘Is Scientific Realism Tenable?’ In Proceedings of the Biennial Meeting of the Philosophy of Science Association (PSA 1976), vol. II. Edited by Frederick Suppe and Peter Asquith, 307–34, Chicago: The University of Chicago Press. Westphal, Kenneth R. 2015. ‘Conventionalism and the Impoverishment of the Space of Reasons: Carnap, Quine and Sellars’. Journal for the History of Analytical Philosophy, 3: 1–66.

11 Sellars and Wittgenstein, Early and Late Guido Bonino and Paolo Tripodi

1.  The Issue There is a famous passage, at the very beginning of Science and Metaphysics, in which Sellars speaks of the history of philosophy as the lingua franca of philosophy, to be used as a means of communication among philosophers, though not only as a means of communication: The history of philosophy is the lingua franca which makes communication between philosophers, at least of different points of view, possible. Philosophy, without the history of philosophy, if not empty or blind, is at least dumb. Thus, if I build my discussion of contemporary issues on a foundation of Kant exegesis and commentary, it is because, as I see it, there are enough close parallels between the problems confronting him and the steps he took to solve them, on the one hand, and the current situation and its demands, on the other, for it to be helpful to use him as a means of communication, though not, of course, as a means only. (SM, 1; latter emphasis added) Another significant passage concerning this issue can be found in Sellars’s Autobiographical Reflections. This passage is somewhat more personal, in that Sellars tells us the story of how he came to write “Realism and the New Way of Words” (1948), after several years of mostly silent reflections: At last I had found a successful strategy for writing. . . . I soon discovered that spinning out, as I was, ideas in a vacuum, everything I wrote was idiosyncratic and had little direct connection with what others had said. Each spinning required a new web to support it, and the search for fixed points of reference became a struggle for coherence and completeness. . . . I soon came to see that a dialectical use of historical positions is the most reliable way of anchoring arguments and making them intersubjectively available. In the limiting case, this use of history is illustrated by correspondence and controversial exchanges with contemporaries. (AR, 293)

Sellars and Wittgenstein, Early and Late  217 In what follows we would like to examine whether this idea of the history of philosophy as a means of communication can actually be seen at work in Sellars’s dealings with a specific philosopher. Since Sellars suggests that being a means of communication is not the only use of the history of philosophy, we also aim to investigate its other possible uses. The specific philosopher we shall focus on is Ludwig Wittgenstein. Wittgenstein’s case is especially worth investigating, for reasons that pertain both to Wittgenstein’s side and to Sellars’s side. For what concerns Wittgenstein, the interesting point is the rather obvious one that there are (at least) two clearly distinguishable Wittgensteins: the early Wittgenstein (represented by the Tractatus Logico-Philosophicus) and the later Wittgenstein (epitomized in the Philosophical Investigations). For what concerns Sellars, the interesting point is that Sellars wishes to use both Wittgensteins. Those who use Wittgensteinian themes for their own philosophical purposes usually choose one of the two Wittgensteins. On the other hand, philosophers or scholars who are interested in Wittgenstein from a historical-philosophical perspective often focus on the continuity/discontinuity issue between the two phases of Wittgenstein’s philosophy. For these reasons, the present paper can also be read as a contribution to the understanding of the tangled vicissitudes of Wittgenstein in America, and of the role Sellars played in this story. At first, American philosophy neglected or failed to understand the thought of Wittgenstein, especially that of the later Wittgenstein, even when, as in the 1940s and 1950s, a Wittgensteinian philosophical style was becoming dominant across the Atlantic, notably at Oxford and Cambridge. During the 1930s and 1940s the linguistic and cultural distance made it difficult for American philosophers to read, let alone understand, Wittgenstein. It happened quite often that they mistook the philosophy of Wittgenstein for the most similar thing they had stumbled upon, namely, the philosophy of Rudolf Carnap: a relatively well-known example of such confusion is Ernest Nagel’s “Impressions and Appraisals” of ‘analytic’ philosophy, published in 1936, after a sabbatical year in Europe (Nagel 1936). However, the assimilation of Carnap’s thought into American philosophy mainly took place through the criticisms of Quine, Goodman and others. As a consequence, Wittgenstein was allowed entrance into the US as an obscure and less scientific Doppelgänger of Carnap, a great philosopher, yet one who arguably had made epochal mistakes, which were predominantly exposed by American philosophers themselves (for an authoritative reconstruction of this story see Hacker 1996 and Glock 2008; the present account, however, heavily relies also on Tripodi 2009). In the 1960s and 1970s things changed, at least in part, and several influent analytic philosophers in the US took Wittgenstein (especially the later Wittgenstein) into serious consideration. Yet in most cases it was an unsympathetic appraisal: for example, to mention only the most well-known and remarkable cases, Putnam attacked Wittgenstein’s notion of criterion, Davidson criticized the distinction between causes and reasons, and Fodor

218  Guido Bonino and Paolo Tripodi charged Wittgenstein’s philosophy of mind with behaviorism and verificationism (Putnam 1962; Davidson 1963; Chihara and Fodor 1965; Fodor 1975). Only a scant group of philosophers, concentrated around Norman Malcolm at Cornell University, defended Wittgenstein’s philosophy, but they were rather isolated and marginal with respect to the mainstream. There were of course some exceptions, philosophers who knew Wittgenstein well and appreciated his philosophy: from Cavell and Dreben to Stroud and Arrington. Yet, this does not change the overall picture: American philosophy, which in the meantime had become ‘analytic,’ proved to be impervious to Wittgenstein, or even hostile to his style of thought. Given this background, Sellars’s case is uncommon and peculiar. Sellars was a respected philosopher, and his importance and influence were by no means related to some sort of Cornell-inspired Wittgensteinianism. However, he was well acquainted with Wittgenstein’s philosophy, early and late, and he was able to recognize both its elements of continuity and its differences with respect to Carnap. This does not mean that Sellars was completely clear of any confusion between Carnap’s and Wittgenstein’s views, but rather that his degree of awareness was greater than that of most American philosophers belonging to his generation. The present paper provides a contribution to the understanding of Wittgenstein’s alternating fortunes in the US by analyzing and interpreting Sellars’s stance,1 which was characteristically against the tide: for his use of both Wittgensteins for philosophical purposes and, in particular, for his attempt to make Wittgenstein’s and Carnap’s semantic ideas interact with each other. As has just been stated, Sellars wishes to use both Wittgensteins for philosophical purposes. From a philosophical point of view this raises at least two questions: Does it work? That is, is Sellars’s unique combination of motives from both the early and the later Wittgenstein philosophically tenable? Are the uses of motives from both the early and the later Wittgenstein in some way connected, or correlated, to Sellars’s famous distinction between the scientific and the manifest image of the world? Other questions can be asked from a historical-philosophical perspective—that is, from a perspective that focuses on exegetical issues as well as on the more general assessment of Wittgenstein’s place and role in the history of philosophy. Among these questions are: What are Sellars’s interpretations of the philosophies of the early and the later Wittgenstein respectively? What is Sellars’s view concerning the continuity/discontinuity issue? Yet our main question is somewhat different: how does Sellars make the two Wittgensteins interact with each other? Our guess is that it is in these

Sellars and Wittgenstein, Early and Late  219 modes of interaction that the originality of Sellars’s approach is most clearly apparent. In any case, all these questions could be investigated by considering several different aspects of both Wittgenstein’s and Sellars’s philosophies. We choose to focus on one topic—albeit a large one—, that is, semantics, broadly conceived.

2.  From Carnap to the Later Wittgenstein Sellars was in agreement with many of Wittgenstein’s later views: antifoundationalism in epistemology (according to which there is no ultimate and self-justified belief, or set of beliefs, on which our knowledge is based); antiCartesianism in the philosophy of mind (which rejects the idea that we know our own thoughts and sensations directly and with absolute certainty); a sort of anti-reductionism according to which the normative dimension (of logic, epistemology, action, language, morality and so forth) cannot be reduced to naturalistic terms; and even some metaphilosophical views (for example, the idea that philosophy and science ultimately have different goals and work, as it were, on different levels). In this section, however, we shall focus on a specific issue: Sellars’s and the later Wittgenstein’s conceptions of semantics. In Empiricism and the Philosophy of Mind (1956) Sellars makes it clear that there is a strict connection between the Myth of the Given—his main polemical target—and certain conceptions of language. Roughly, the Myth of the Given is the idea that any knowledge we may have is ultimately based on some noninferential, immediate knowledge of basic, self-justified and epistemically independent elements (see de Vries and Triplett 2000, 186 for this formulation). According to Sellars, one way to avoid the Myth of the Given is to realize that what he refers to as psychological nominalism is true, that is, that “all awareness of sorts, resemblances, facts, ect., in short, all awareness of abstract entities—indeed, all awareness even of particulars—is a linguistic affair” (EPM, § 29). Consequently—and, in a way, symmetrically— Sellars also thinks that a special form of the Myth of the Given, formulated as a rejection of psychological nominalism, is the idea that some awareness of abstract entities—or of particulars—is prelinguistic. On this basis, he suggests that the Myth of the Given goes hand in hand with a thesis concerning the acquisition of language, namely, the view that “there is awareness of logical space prior to, or independent of, the acquisition of a language” (EPM, § 31). In this connection, Sellars explicitly refers to Wittgenstein’s later work and, in particular, to the first sections of the Philosophical Investigations, where Wittgenstein criticizes the so-called Augustinian picture of language, i.e., the idea that every word has a meaning, that this meaning is correlated to the word, that the meaning of a word is the object for which the word stands, and that the essential function of words is naming (cf. Wittgenstein 1953, §1). Sellars focuses, in particular, on the Augustinian picture of language acquisition, the view according to which

220  Guido Bonino and Paolo Tripodi the process of teaching a child to use a language is that of teaching it to discriminate elements within a logical space of particulars, universals, facts, etc., of which it is already undiscriminatingly aware, and to associate these discriminated elements with verbal symbols. (EPM, § 30) The reference to Wittgenstein is clear and straightforward: not only in this context does Sellars refer to the case of a child, but he also evokes the case of a “carrier of slabs” (EPM, § 30), which directly calls to mind § 2 of the Investigations. According to Sellars, the Augustinian picture conveys an especially tempting form of the Myth of the Given because it expresses a view “to which even philosophers who are suspicious of the whole idea of inner episodes can fall prey” (EPM, § 30). In fact, even nominalists, who think that there are only particulars, can fall prey to it. Sellars therefore emphasizes the fact that the real test for a theory of language is its account of what, in borrowing the idea from H. H. Price, he calls ‘thinking in presence,’ as opposed to ‘thinking in absence.’ The former expression indicates the Augustinian situation in which the alleged meaning of a name—the named object—is there before the language user (EPM, § 30; cf. Price 1953). Sellars is well aware, however, of the fact that the Augustinian picture of language typically presents itself in the form of a Platonist view, according to which the meaning of a predicate is the abstract entity for which it stands. For example, the word ‘red’ means (that is, designates) the quality (or property) red. This Platonist view is based on two claims: that “the word ‘red’ would not be a predicate if it didn’t have the logical syntax characteristic of predicates”; and that it would not “be the predicate it is, unless . . . we tended to respond to red objects in standard circumstances with something having the force of ‘This is red’ ” (EPM, § 31), in a way similar to that in which a thermometer tends to respond to a certain temperature with something having the force of, say, ‘37 degrees.’ According to Sellars, however, the main argument for the idea that the meaning of a predicate is the abstract entity it stands for is a different one, and depends on what he refers to as a relational conception of means. This is a view of the logic of ‘means’ that depends on a superficial analogy: since the statement ‘(In German) “rot” means red’ superficially resembles such relational statements as ‘Cowley adjoins Oxford,’ then we can assimilate the form ‘ “. . .” means---’ to the form ‘xRy’. This analogy invites us to “take it for granted that meaning is a relation between a word and a nonverbal entity” and to “suppose that the relation in question is that of association” (EPM, § 31). According to Sellars, this analogy is misleading. In fact, one should provide an entirely different picture of semantics. He argues as follows: The truth of the matter, of course, is that statements of the form “ ‘. . .’ means---” are not relational statements. . . . The rubric “ ‘. . .’ means---” is a linguistic device for conveying the information that a mentioned

Sellars and Wittgenstein, Early and Late  221 word, in this case “rot”, plays the same role in a certain linguistic economy, in this case the linguistic economy of German-speaking peoples, as does the word “red”, which is not mentioned but used—used in a unique way; exhibited, so to speak—and which occurs “on the right-hand side” of the semantical statement. (EPM, § 31) Moreover, the absurdity of the Augustinian picture immediately turns out if, for example, one considers the word ‘und,’ rather than the word ‘rot,’ as the meaning of a connective is clearly given by its inferential role, rather than by its being associated with an alleged Platonic entity—or so Sellars seems to think. As is well known, in several papers Sellars formulates his inferentialist view of semantics by employing the so-called dot-quoting device. Consider, for instance, the semantic statement “ ‘Rouge’ in French means red,” as it is typically formulated, for example, in Carnap-style semantics (cf. Carnap 1942).2 According to Sellars this statement should be translated as “ ‘Rouge’ in French means ·red·,” which in turn means “ ‘Rouge’ in French plays the same role that ‘red’ plays in English.” In this way, the Augustinian reading of the semantic statement is explained away, and semantics turns out to be, at least prima facie, an entirely intralinguistic affair (for a general description and discussion of Sellars’s philosophy of language see, for example, O’Shea 2007, Chapter iv, de Vries 2016, as well as Marras 1973). After EPM, Sellars often associates his functional role semantics to Wittgenstein’s later view of meaning as use. For example, in a passage that occurs several times in his works, he writes that 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. (LTC, 116; CC, 181; MFC, 431; NAO, 78; see also BEB, 158, TC, 203, letter to Ausonio Marras, 26–11–1975, in CSM) It is interesting to notice, however, that when he put forward his functional role semantics for the first time (at least in a rudimentary form), Sellars was not primarily concerned with the later Wittgenstein. Rather, his attempt was that of discussing, developing and criticizing Carnap’s semantic views, as Carnap presented them in his Introduction to Semantics (cf. Carus 2004 and Olen 2016 for the relationship between Sellars and Carnap). This implicit dialogue with Carnap on semantics lasted from 1947–48 (“Epistemology and the New Way of Words” and “Realism and the New Way of Words,” ENWW and RNWW) to 1953 (“Inference and Meaning”, IM). Only around 1953 and especially in 1954, with the publication of “Some Reflections on Language Games,” Sellars started to systematically present his

222  Guido Bonino and Paolo Tripodi functional role semantics by using the later Wittgenstein’s vocabulary and conceptual apparatus, framed in terms of games, language games, system of rules, grammar, use, rule-governed activity and so forth (SRLG). As already suggested earlier, an apparent consequence of Sellars’s semantic views is that semantics is entirely intralinguistic, which means that within semantics we do not reach out to the world. One might not believe this to be the case, since in several places Sellars states that functional roles involve not only intralinguistic moves but also what he calls language entry transitions (such as my seeing something red in front of me and uttering ‘This is red’) and language exit transitions (such as my uttering ‘I shall catch a red thing’ and stretching my arm out toward a red apple in front of me). However, Sellars denies that this is enough to guarantee a language-to-world grounding. For example, he writes: Just as an intralinguistic move is not in the full sense an inference unless the subject not only conforms to, but obeys, syntactical rules . . .  so a language entry transition is not in the full sense an observation unless the subject not only (in normal circumstances) tokens ‘This object is green’ if and only if a green object is present to his senses, but is able to infer (in a pragmatic metalanguage) from ‘The thought this object is green occurred to Jones at time t in place s in circumstances c’ to ‘a green object was present to Jones’ senses at t in s’. (ITSA, 313 footnote; see also SRLG, reprinted with extensive additions in SPR, 334–35) The point here is that such behavioral moves and transitions are nothing but conditioned responses and do not involve any semantic/normative connection between language and the world (for a discussion of such issues, with a special focus on Sellars’s multifaceted attitude towards behaviorism, see Tripodi 2011). However, if semantics proper has nothing to do with causal relations between the speakers’ behavior and the world, then once again semantics seems to be an entirely intralinguistic affair, and the danger of linguistic idealism (i.e., “the miserable absurdity” that “the world belong[s] to the linguistic order,” TC, 209) becomes imminent. In sum, on the one hand, Sellars’s main preoccupation is to deny the idea that a predicate such as ‘red’ acquires meaning “because we come to obey ‘semantical rules’ ” (such as the rule: red objects are to be called red), since this idea clearly “presupposes the existence of prelinguistic concepts,” which would be a form of the Myth of the Given (see SRLG, reprinted with extensive additions in SPR, 334). On the other hand, however, Sellars must recognize that his functional role semantics, which is inspired by Wittgenstein’s later view of language, and which avoids the Myth of the Given, has a possible shortcoming; namely, the problem of linguistic idealism, the view according to which in language we never reach out to the world.3

Sellars and Wittgenstein, Early and Late  223

3.  The Tractatus Sellars explicitly addressed linguistic idealism as a problem in “Truth and ‘Correspondence’ ” (1962). Such a problem is addressed by both interpreting and using the Tractatus Logico-Philosophicus, or better, by using some suggestions drawn from the Tractatus, as interpreted by Sellars. According to Sellars the Tractatus—rather unsurprisingly—requires an interpretation that is in accordance with Sellars’s own conception of semantics. In fact, Sellars holds that if a Tarski-Carnap-style semantics is adopted (see endnote 2), then the Tractatus turns out to be incoherent. Let us consider the semantic statement (1) S (in L) means aRb, where L is the object language, and ‘aRb’ belongs to the metalanguage in which the semantic statement is formulated. According to the doctrine of the Tractatus, a statement, or a sentence, can occur within another statement or sentence only truth-functionally (Wittgenstein 1922, 5). But in (1) ‘aRb’ occurs in a non truth-functional way (for one cannot substitute salva veritate ‘aRb’ with another sentence having the same truth-value). Therefore (1) is, strictly speaking, unsayable. This is the famous thesis of the ineffability of semantics, which Sellars regards as a paradoxical and unacceptable result (see TC, 208). Now, the paradox of (1) can be solved if either condition (i) or (ii) is satisfied: (i) ‘aRb’ does not really occur in (1); (ii) (1) is not a statement proper—as Sellars says—that is, it is not a Wittgensteinian Satz, a declarative sentence that may be true or false in that it describes a matter of fact. Both (i) and (ii) are satisfied if one adopts Sellars’s functional role semantics. (i) is satisfied because (1) is reinterpreted as (2) S (in L) plays the same role as ‘aRb’ (in ML), where ML is the metalanguage. What occurs in (2) is not ‘aRb’, but rather “‘aRb’”; in a way, ‘aRb’ is mentioned or, as Sellars says, it occurs only indirectly. To show that (ii) is also satisfied, a different example of semantic statements is taken into consideration, in which the notion of truth is involved. According to Tarski-Carnap semantics, the predicate ‘is true’ can be defined as: ‘Chicago est grande’ (in French) is true iff ∃ p (‘Chicago est grande’ (in French) means p, and p)4.

224  Guido Bonino and Paolo Tripodi According to Sellars’s view of semantics, this formula must be modified in the following way: ‘Chicago est grande’ (in French) is true iff ∃ that-p (‘Chicago est grande’ (in French) expresses the proposition that-p, and that-p is true), where the phrase ‘expresses the proposition that-p’ is equivalent to is a ·Chicago is large·, which in turn must be understood as plays the same role played in English by ‘Chicago is large.’ It is plain to see that in this case, once again, semantics remains entirely intralinguistic. Yet, the crucial question to be answered now is: what do statements such as ‘that Chicago is large is true’ mean? Sellars’s formula ‘Chicago est grande’ (in French) is true iff ∃ that-p (‘Chicago est grande’ (in French) expresses the proposition that-p, and that-p is true) cannot be regarded as a statement providing a genuine definition of the predicate ‘is true,’ since such a predicate occurs both on the left-hand side and on the right-hand side of the biconditional. According to Sellars, a more revealing way to answer the question is by means of the following formula: ‘That Chicago is large is true’ entails and is entailed by ‘Chicago is large’. This is to be understood as a principle of inference, i.e., as a rule that warrants us to do something with language (because ‘inferring is a doing,’as Sellars says in TC, 206). This same view, according to which the meaning of ‘is true’ is entirely captured by such rules of inference finds further confirmation in Science and Metaphysics, where Sellars speaks of truth as assertibility (see especially SM, Chapters iv-v). There are some morals to be drawn. First, it has been shown that semantic statements are to be understood as rules, hence not as statements proper. This also holds for examples such as (1) S (in L) means aRb. Therefore, condition (ii)—i.e., that (1) is not a statement proper—is also satisfied; the ineffability problem is doubly solved; the Tractatus is salvaged and shown to be coherent.

Sellars and Wittgenstein, Early and Late  225 Second, the case of semantic statements involving truth once again shows in what sense semantics according to Sellars’s view never reaches out to the world. Third, it should be noted that this intralinguistic dimension—so to speak—cannot be transcended whenever we have to deal with propositions or facts (in Sellars’s framework propositions or facts belong to the space of reasons and only occur in the context of semantic statements such as those illustrated previously). Sellars explicitly recognizes that this view can be accused of linguistic idealism, and that linguistic idealism is a problem. However, he also thinks that a solution to the problem comes from a suggestion that can be found in the Tractatus itself. The previous characterization of truth in terms of rules of inference holds (according to Sellars and according to the Tractatus as interpreted by Sellars) for all kinds of truth, from mathematical truths to factual ones. Yet, in the case of factual statements (matter-of-factual statements, as Sellars sometimes calls them) there is something more to be taken into account: the picture theory. For Sellars the picturing relation is a highly complex relation of projection, based on natural regularities, holding between two complex (that is, structured) natural objects. It is important to explain what it means that the terms of the picturing relation are natural, and what it means that they are objects, as well as why they must be so. As to the fact that the terms are natural, they are the linguistic sign, which must be conceived of as a purely natural object (an item in the order of causes and effects, in rerum natura), and the pictured object in the world (see TC, 222). The reason why they must be natural is that the causal nature of the picturing relation demands natural terms. But in addition to that, the terms of the picturing relation are and must be objects, rather than facts. The pictured complex object is outside of language, but in order for it to be so, it cannot be a fact, since facts (or propositions) are linguistic entities inhabiting the space of reasons. It should be remarked that the picture theory, conceived of in this way, does not belong to semantics proper. Somewhat equivalently, Jay Rosenberg pointed out (Rosenberg 2007, 126) that Sellars’s picturing is devoid of any epistemological significance. It is clear that Sellars’s view of picturing is quite different from Wittgenstein’s: in the Tractatus the picturing relation has a distinctly non-natural dimension (having to do with the role played by the projective relation and the metaphysical subject), and its two terms—the proposition and the pictured fact—are definitely not complex objects, but rather facts. This of course means that Sellars’s own interpretation of the Tractatus is rather idiosyncratic:5 Carl B. Sachs, in his contribution to this volume, goes so far as characterizing Sellars’s speaking of picturing as an “unfortunate . . . choice of terminology,” which obscures the pragmatistDeweyan connection.

226  Guido Bonino and Paolo Tripodi We can thus summarize the two distinct domains of semantics and picture theory according to Sellars. In semantics we have formulas such as ‘aRb’ (in L) means aRb, which is to be understood as ‘aRb’ (in L) plays the same role played by ‘p’ in ML. Here ‘ “aRb” ’ is the name of a meaningful sentence of L; speaking of a role played by ‘aRb’ or by ‘p’ inevitably involves normativity; the semantic statement as a whole is to be understood as a principle of inference. It should be noted once more that in the semantic situation one never goes beyond the linguistic sphere. In the picture theory, on the other hand, we have formulas such as aRb pictures a´R´b´. This is to be understood as aRb is in a certain (highly complicated) relation of projection with a´R´b´. aRb is of course a complex natural object (e.g., a combination of sounds or ink spots), devoid of any normativity. a´R´b´ is a complex natural object as well (e.g., a cat on the table; not that a cat is on the table). The picture theory statement is an object language description of a natural regularity. It is worth observing again that in the picturing situation there is nothing genuinely linguistic or semantic (i.e., nothing involving normativity). It might seem therefore that semantics and picture theory are entirely disconnected. Yet, for Sellars there is a connection between them. More precisely, there is a connection between the truth of a statement and the correctness of a picture. The correctness of a picture is not meant as a normative notion: a picture is said to be correct if it is projected in a way that conforms to certain (highly complicated) causal regularities. The criterion according to which we decide whether a statement is true is whether the same statement, conceived of as a complex object in the realm of causes and effects (i.e., as a natural complex object) is a correct picture. In other words, the statement does not say that the picture is correct, but it implies that the picture is correct, since that is the criterion for its truth (see TC, 223). Thus, in a somewhat indirect way, according to Sellars a connection between language and the extralinguistic reality is finally secured.

4.  Uses and Interpretations The main question of this paper—how does Sellars, in his investigations on semantics, make the two Wittgensteins interact with each other?—can now

Sellars and Wittgenstein, Early and Late  227 be answered, thereby shedding some light on the general topic of ‘Sellars and the history of philosophy.’ First, Sellars used the later Wittgenstein’s vocabulary as a lingua franca to formulate his own view of semantics, which he himself had already developed some years before in a different context— namely, discussing, developing and criticizing Carnap’s relational semantics. At that time, Sellars had already read Wittgenstein’s Blue and Brown Books, and he was also acquainted with Ryle’s critique of the ‘Fido’-Fido conception of meaning (cf. Wittgenstein 1958; Ryle 1949), which in its turn was influenced by Wittgenstein’s critique of the Augustinian picture. Apparently, however, only in the Investigations—published in 1953—did Sellars find an extremely useful vocabulary and conceptual apparatus (i.e., game, language game, system of rules, grammar, use, language as rule-governed activity, rule-following, the analogy between the rules of language and the rules of chess, etc.). It seems that the use of Wittgenstein’s later vocabulary helped Sellars to provide a clearer formulation of his functional role semantics. In other words, the later Wittgenstein was, for Sellars, a good means of communication. And this is particularly interesting in the general context of the history of analytic philosophy in America, especially if one realizes that in that period in the US few people were able to properly understand what was going on in the Philosophical Investigations, let alone to use that work for their own philosophical purposes (see Hacker 1996; Glock 2008; Tripodi 2009). Sellars’s semantics, which originated as an ideal dialogue with Carnap, and then developed employing Wittgenstein’s later conceptual apparatus, had some shortcomings. This was recognized by Sellars himself. In particular, one of the dangerous consequences seemed to be linguistic idealism. Sellars found a solution to this problem in the Tractatus’s picture theory. In this case the history of philosophy is regarded by Sellars as a repository of philosophical tools, rather than as a lingua franca. In particular, the picture theory borrowed from the early Wittgenstein is a philosophical tool which Sellars employed to solve a problem that arose in the context of a semantics inspired by the later Wittgenstein. Sellars’s use of the Tractatus raises a further, interpretative issue: which interpretation of the Tractatus is adopted by Sellars? Sellars’s own interpretation of the Tractatus is rather original and, at the same time, controversial. The need for a new interpretation had an independent motivation, namely, the purpose of solving a philosophical problem Sellars found in the Tractatus itself: the problem of the ineffability of semantics, and the related problem of the incoherence of the Tractatus. This suggests that Sellars oscillates between three options and tasks: interpreting the Tractatus for the sake of understanding (the early) Wittgenstein; interpreting it for the sake of solving a philosophical problem he found in the Tractatus itself (the problem of ineffability); and interpreting it with the purpose of solving a philosophical problem originated outside the Tractatus (in particular, the intralinguistic and somewhat idealistic nature of his functional role semantics). It is worth noting that—as one can see, for example, in the passage taken from Sellars’s

How to formulate a non-relational semantics?

later vocabulary (history of philosophy as a lingua franca)

Using a functional role semantics expressed in later Wittgensteinian terms (history of philosophy as a repository of tools)

A functional role semantics

How to avoid linguistic idealism?

Using the Tractatus picture theory (history of philosophy as a repository of tools)

How to solve the ineffability problem?

The language reaches out to the world

Figure 11.1  Sellars’s semantics and its uses of Wittgenstein

The Tractatus is coherent

Sellars and Wittgenstein, Early and Late  229 Autobiographical Reflections quoted at the beginning of the paper—the three options are not considered by Sellars to be discrete alternatives but, rather, they spread along a continuum. There is also an additional complication, which once more shows that, for philosophical purposes, the way in which Sellars made the two Wittgensteins interact with each other is a particularly interesting and complex one. As we have just stated, Sellars interpreted the Tractatus in order to solve the ineffability problem, in the attempt to make the Tractatus itself coherent. However, he did so by using the later Wittgenstein. Or better, he did so by reading the Tractatus on the basis of a functional role semantics, partly inspired by the later Wittgenstein, rather than on the basis of a Tarski-Carnap semantics (which, in a way, could have appeared as a more natural alternative). Figure 11.1 in the previous page is a schematic representation, or a diagram, of the interesting and complex ways in which, as far as the issue of semantics is concerned, the two Wittgensteins interact in the hands of Sellars.

Notes 1 We believe that it is not a coincidence that, in the US, the current interest in Sellars and Wittgenstein often goes hand in hand in the work of Robert Brandom and John McDowell. 2 Sellars always associates this kind of semantics with the name of Carnap, and sometimes also of Tarski (see, among others, RNWW, TC, SM, letter to Gilbert Harman, 26–2–1970, in CSHT). The historical importance of the publication of Carnap’s Introduction to Semantics in 1942 is difficult to overestimate. After having rejected for many years the feasibility of a semantics along Tarskian lines (Tarski 1933), Carnap gradually changed his mind in the second half of the 1930s, a process which was finally accomplished in 1942 (see Coffa 1991, Chapter xvi). This turn was a disquieting surprise for many people, even for some who had always belonged to the same philosophical milieu. The first criticism came from Otto Neurath, already in the mid 1930s (see Mancosu 2008); then, after the publication of Carnap’s book, came the criticisms by Hall (1944); Bergmann (1944 and 1945); Black (1945) (see Hochberg 1994; Hochberg 2001, 2–10; Bonino 2007). W.V.O. Quine, Nelson Goodman, and even Tarski himself, though for different reasons, were also critical with respect to Carnap’s project (see Frost-Arnold 2013). 3 Sellars takes it for granted that linguistic idealism is flawed. However, he does not elaborate either on the notion itself of linguistic idealism or on the reasons why it is objectionable. Some interpreters, such as Rorty, suggested that Sellars’s solution to the problem of linguistic idealism, that is his theory of picturing, could throw Sellars back into some sort of foundationalist position, in the mire of the Myth of the Given (see Rorty 1979, 91). As we and other interpreters (such as Rosenberg 2007 and Sachs in the contribution to this volume) show, this danger seems to be averted by the fact that the relation of picturing is neither semantic nor epistemological. Yet, some doubts may still linger on in connection with Sellars’s poor elaboration of the concept of linguistic idealism: in order to make a conclusive assessment of this dispute, one should probably provide a definition of the problem of linguistic idealism that is independent of the theory designed to solve it.

230  Guido Bonino and Paolo Tripodi 4 In his Introduction to Semantics Carnap actually uses the verb ‘designate’ rather than ‘mean.’ Here we are adhering to Sellars’s use. 5 The case of picturing is not the only one in which Sellars exhibits an idiosyncratic or at least controversial interpretation of the Tractatus. Another significant case is Sellars’s interpretation of proposition 3.1432 of the Tractatus as suggesting an argument in favor of nominalism, an argument which is developed in ‘Naming and Saying’ and in Naturalism and Ontology, Chapter iv (NS and NAO). For a detailed criticism of Sellars’s argument, see Hochberg (1978, 318–23; 1981; 2001, 108–9; and 2003, 158–9). For an overall assessment of the entire question, see Bonino (2012).

References Bergmann, Gustav. 1944. ‘Pure Semantics, Sentences, and Propositions’. Mind, 53: 238–57. Bergmann, Gustav. 1945. ‘A Positivistic Metaphysics of Consciousness’. Mind, 54: 193–226. Black, Max. 1945. ‘Review of Rudolf Carnap, Introduction to Semantics, and Formalization of Logic’. Mind, 54: 171–76. Bonino, Guido. 2007. ‘The First Station of Gustav Bergmann’s Odyssey’. In Ontology and Analysis. Essays and Recollections About Gustav Bergmann. Edited by Laird Addis, Greg Jesson and Erwin Tegtemeier, 31–50. Frankfurt am Main: Ontos Verlag. Bonino, Guido. 2012. ‘Hochberg and Sellars: Perspicuity and Bradley’s Regress’. In Studies in the Philosophy of Herbert Hochberg. Edited by Erwin Tegtmeier, 51–71. Frankfurt am Main: Ontos Verlag. Carnap, Rudolf. 1942. Introduction to Semantics. Cambridge, MA: Harvard University Press. Carus, André W. 2004. ‘Sellars, Carnap, and the Logical Space of Reasons’. In Carnap Brought Home: The View from Jena. Edited by Steve Awodey and Carsten Klein, 317–55. Chicago: Open Court. Chihara, Charles and Fodor, Jerry. 1965. ‘Operationalism and Ordinary Language: A Critique of Wittgenstein’. American Philosophical Quarterly, 2: 281–95. Coffa, J. Alberto. 1991. The Semantic Tradition from Kant to Carnap: To the Vienna Station. Cambridge: Cambridge University Press. Davidson, Donald. 1963. ‘Actions, Reasons, and Causes’. Journal of Philosophy, 60: 685–700. de Vries, Willem A. 2016. ‘Wilfrid Sellars’. In The Stanford Encyclopedia of Philosophy (Winter 2016 Edition). Edited by Edward N. Zalta. https://plato.stanford. edu/archives/win2016/entries/sellars. de Vries, Willem A. and Triplett, Timm. 2000. Knowledge, Mind, and the Given: A Reading of Sellars’s ‘Empiricism and the Philosophy of Mind’. Indianapolis, IN and Cambridge: Hackett Publishing Company. Fodor, Jerry. 1975. The Language of Thought. New York: Thomas Y. Crowell. Frost-Arnold, Greg. 2013. Carnap, Tarski, and Quine at Harvard: Conversations on Logic, Mathematics, and Science. Chicago: Open Court. Glock, Hans-Johann. 2008. ‘The Influence of Wittgenstein on American Philosophy’. In The Oxford Handbook of American Philosophy. Edited by Cheryl Misak, 375–402. Oxford: Oxford University Press.

Sellars and Wittgenstein, Early and Late  231 Hacker, Peter M.S. 1996. Wittgenstein’s Place in Twentieth Century Analytic Philosophy. Oxford: Wiley-Blackwell. Hall, Everett W. 1944. ‘The Extra-Linguistic Reference of Language. II: Designation of the Object Language’. Mind, 53: 25–47. Hochberg, Herbert. 1978. Thought, Fact, and Reference: The Origins and Ontology of Logical Atomism. Minneapolis, MN: University of Minnesota Press. Hochberg, Herbert. 1981. ‘Logical Form, Existence, and Relational Predication’. In Midwest Studies in Philosophy. Vol. VI: The Foundations of Analytic Philosophy. Edited by Peter A. French, Theodore E. Uehling and Howard K. Wettstein, 215–38. Minneapolis, MN: University of Minnesota Press. Hochberg, Herbert. 1994. ‘From Carnap’s Vienna to Meinong’s Graz: Gustav Bergmann’s Ontological Odyssey’. Grazer Philosophischen Studien, 48: 1–49. Hochberg, Herbert. 2001. The Positivist and the Ontologist: Bergmann, Carnap, and Logical Realism. Amsterdam: Rodopi. Hochberg, Herbert. 2003. Introducing Analytic Philosophy. Its Sense and Its Nonsense. Frankfurt am Main: Hänsel-Hohenhausen. Mancosu, Paolo. 2008. ‘Tarski, Neurath, and Kokoszynska on the Semantic Conception of Truth’. In New Essays on Tarski and Philosophy. Edited by Douglas Patterson, 192–224. Oxford: Oxford University Press. Marras, Ausonio. 1973. ‘On Sellars’ Linguistic Theory of Conceptual Activity’. Canadian Journal of Philosophy, 2: 471–83. Nagel, Ernest. 1936. ‘Impressions and Appraisals of Analytic Philosophy in Europe’. The Journal of Philosophy, 33: 5–24, 29–53. Olen, Peter. 2016. Wilfrid Sellars: The Foundations of Normativity. London: Palgrave Macmillan. O’Shea, James. 2007. Wilfrid Sellars: Naturalism with a Normative Turn. Cambridge: Polity Press. Price, Henry H. 1953. Thinking and Experience. London: Hutchinson. Putnam, Hilary. 1962. ‘Dreaming and “Depth Grammar” ’. In Analytical Philosophy: First Series. Edited by R.J. Butler, 211–35. Oxford: Oxford University Press. Rorty, Richard. 1979. Philosophy and the Mirror of Nature. Princeton, NJ: Princeton University Press. Rosenberg, Jay. 2007. ‘Sellarsian Picturing’. In Wilfrid Sellars: Fusing the Images, 104–26. Oxford: Oxford University Press. Ryle, Gilbert. 1949. The Concept of Mind. London: Hutchinson. Sellars, Wilfrid (ENWW). 1947. ‘Epistemology and the New Way of Words’. Journal of Philosophy, 1944: 645–60. Sellars, Wilfrid (RNWW). 1948. ‘Realism and the New Way of Words’. Philosophy and Phenomenological Research, 8: 601–34. Sellars, Wilfrid (IM). 1953. ‘Inference and Meaning’. Mind, 62: 313–38. Sellars, Wilfrid (SRLG). 1954. ‘Some Reflections on Language Games’. Reprinted with extended additions in SPR, 321–58. Sellars, Wilfrid (EPM). 1956. ‘Empiricism and the Philosophy of Mind’. Reprinted with additional footnotes in SPR, 127–96. Sellars, Wilfrid (ITSA). 1956. ‘Is There a Synthetic A Priori’. Reprinted in SPR, 298–320. Sellars, Wilfrid (NS). 1962. ‘Naming and Saying’. Reprinted in SPR, 225–46. Sellars, Wilfrid (TC). 1962. ‘Truth and “Correspondence” ’. Reprinted in SPR, 197–224.

232  Guido Bonino and Paolo Tripodi Sellars, Wilfrid (SPR). 1963. Science, Perception and Reality. London: Routledge & Kegan Paul. Reissued by Atascadero, CA: Ridgeview Publishing Company 1991. Sellars, Wilfrid (SM). 1968. Science and Metaphysics: Variations on Kantian Themes. London: Routledge & Kegan Paul. Sellars, Wilfrid (LTC). 1969. ‘Language as Thought and as Communication’. Reprinted in EPH, 93–117. Sellars, Wilfrid (BEB). 1970. ‘Belief and the Expression of Belief’. In Language, Belief, and Metaphysics. Edited by Howard E. Kiefer and Milton K. Munitz, 146– 58. Albany, NY: State University of New York Press. Sellars, Wilfrid (CC). 1973. ‘Conceptual Change’. Reprinted in IPH, 172–88. Sellars, Wilfrid (MFC). 1974. ‘Meaning as Functional Classification (A Perspective on the Relation of Syntax to Semantics)’. Synthese, 27: 417–37. Sellars, Wilfrid (AR). 1975. ‘Autobiographical Reflections’. In Action, Knowledge and Reality. Edited by Hector-Neri Castañeda, 277–93. New York: Bobbs-Merrill. Sellars, Wilfrid (NAO). 1980. Naturalism and Ontology. Atascadero, CA: Ridgeview Publishing Company. Sellars, Wilfrid (CSM). ‘Correspondence Between Wilfrid Sellars and Ausonio Marras’. Edited in hypertext by Andrew Chrucky. www.ditext.com/sellars/csm.html. Sellars, Wilfrid (CSHT). ‘Correspondence Between Wilfrid Sellars and Gilbert Harman on Truth’. Edited in hypertext by Andrew Chrucky. www.ditext.com/sellars/ sh-corr.html. Tarski, Alfred. 1933. Pojęcie prawdy w językach nauk dedukcyjnych. Warszawa: Nakładem Towarzystwa Naukowego Warszawskiego. Tripodi, Paolo. 2009. Dimenticare Wittgenstein: Una vicenda della filosofia anali­ tica. Bologna: Il Mulino. Tripodi, Paolo. 2011. ‘Revisiting Sellars: Behaviorism and the Myth of Jones’. History of Philosophy Quarterly, 28: 85–105. Wittgenstein, Ludwig. 1922. Tractatus Logico-Philosophicus. London: Routledge & Kegan Paul. Wittgenstein, Ludwig. 1953. Philosophische Untersuchungen. Oxford: WileyBlackwell. Wittgenstein, Ludwig. 1958. The Blue and Brown Books. Oxford: Wiley-Blackwell.

12 Wilfrid Sellars and Roy Wood Sellars Theoretical Continuities and Methodological Divergences Fabio Gironi

The last decade has seen growing interest in the work of, and the literature on, Wilfrid Sellars—one of the most important American philosophers of the twentieth century. The current state-of-the-art research on Sellars has produced a number of authoritative introductions to Sellars’s thought, edited collections critically exploring various facets of his system in connection with broader philosophical issues (including this one), books on related topics (from the problem of intentionality to that of naturalism) that adopt a Sellarsian outlook, historical monographs placing his thought in context, as well as a vast number of research articles and special issues of journals. However, witnessing this ‘Sellars Renaissance,’ one notes with surprise that no significant work has assessed the crucial philosophical relationship between Wilfrid Sellars and his father Roy Wood Sellars (1880–1973), a preeminent philosopher in his own right1 whose career spanned much of the twentieth century.2 Roy’s thought had a profound—but still mostly unacknowledged—influence on that of his son, and throughout this chapter I will defend the thesis that Roy Wood Sellars’s basic philosophical coordinates and commitments represent the fertile soil from which the seeds of his son’s new philosophical ideas would grow and independently develop. This chapter, then, will offer a panoramic view of the personal and philosophical relationship between father and son, arguing that a closer examination of Roy Wood Sellars’s philosophical output is overdue, not only for its important influence on his son and—by proxy of his son—on contemporary post-Sellarsian philosophy, but also because of its own importance and originality as the most prominent representation of the Critical Realist school of early twentieth-century American Philosophy. To analyze this relationship, then, represents an important task from both an historical point of view (examining an under-explored, transitional chapter of the development of philosophy in the US, one which saw the merging of home-grown philosophical speculations with the new logicolinguistic methods imported from Europe in the 1930s) and as a contribution to present debates of Sellars scholarship and related subfields (conceptualist approaches to the philosophy of perception, emergentist solutions to the hard problem of consciousness, and ‘liberal’, non-reductive naturalist

234  Fabio Gironi stances). The historical/theoretical objectives of this chapter are then to (i) offer an outline of Roy Wood Sellars’s important critical realist and evolutionary naturalist philosophy and (ii) to clarify the origin of Wilfrid Sellars’s currently much discussed philosophical commitments by highlighting the deep commonality of interests and strategies between father and son across the whole spectrum of their philosophical interests: from epistemological and metaphysical themes to metaphilosophical views. It will be shown how Wilfrid carried forward his father’s projects of critical realism and evolutionary naturalism, reformulated through a new philosophical vocabulary, while developing a strikingly original approach to these philosophical problems that merged systematic ambition with analytic methods. The philosophical thought of Roy Wood Sellars evolved in the thenrelatively recent environment of academic philosophy in the US. In the decades between the 1880s and the 1920, this philosophical scene was undergoing a slow but steady transition, moving away from the dominance of religiously minded Idealism in the most important philosophy departments of the country, and towards the development of a few new hotbeds of realist thought— mostly thanks to the influence of William James and his debates with Josiah Royce at Harvard. A new generation of philosophers rebelled against the dominance of idealism on the American scene and took it upon themselves to forge a ‘New Realism.’3 The aim was that of rejecting the spiritualism of their predecessors and of constructing of a realist philosophy more consonant with the natural sciences—and indeed of presenting philosophy itself as a special science. And so, philosophers like William Pepperell Montague, Ralph Burton Perry and Edwin Bissel Holt launched a direct attack against the absolutist idealism of Royce, promoting a direct realism and rejecting any representationalist mediation, any active role of the subject in the perception of the external object, and therefore an ‘epistemological monism,’ according to which the knowledge and the object known would be considered identical, with no distinction between seeming and being. This relatively short-lived, New Realist movement was followed by another group of philosophers who joined forces under the banner of ‘Critical Realism.’4 They shared the New Realists’ anti-idealist commitments, but were also concerned about the tenability of their positive proposal, thus tempering the direct (or presentational) realism of their colleagues with the reintroduction of some form of epistemic mediation (or representation) between object and object in the cognitive process of the acquisition of knowledge. As a founding member of this latter group, alongside preeminent figures like George Santayana and Arthur Lovejoy, Roy Wood Sellars was the philosopher who developed the earliest, most refined and most original form of Critical Realist epistemology in the first four decades of the twentieth century. Roy Wood Sellars comes to the debates on realism as something of an outsider—of Canadian origin and not based at Harvard (born in Seaforth, Ontario, he spent most of his career at the University of Michigan at Ann Arbor)—taking part in the discussion but fiercely preserving his intellectual

Wilfrid Sellars and Roy Wood Sellars  235 independence and the rebellious originality of his thought, a trait that characterizes his entire philosophical production.5 Roy Wood Sellars was certainly the most interesting of the new generation of realists (as well as the most recognizably materialist of them) most notably because of his articulation of two core commitments—his critical realism6 in epistemology and his evolutionary naturalism in ontology—as a systematic philosophical vision (at times called his physical realism,7 which is not to be confused, as we will see later, with a straightforward physicalism). The three terms indeed appear in the titles of his three major books which together—using different forms and slightly different vocabularies—explore the totality of his system.8 Moreover, Roy was also profoundly interested in social issues, a concern that led to the publication of monographs on democracy and on a humanist approach to religion, and, most notably, to the first draft of the well-known Humanist Manifesto, a programmatic text signed by dozens of influential American intellectuals and published in 1933 that outlined a form of secular religion, promoting science at the service of human needs and a broadly socialist outlook for the fair distribution of wealth.9 This complex background is all part of Wilfrid’s philosophical inheritance, an intellectual legacy rarely explored by Sellars’s interpreters. In this chapter, I will mostly focus on Roy’s epistemological thought and on his approach to naturalism, because it is there that the clearest parallels with Wilfrid’s approach can be discerned. Nonetheless, a careful survey of the—admittedly rare—occasions in which Wilfrid comments on non-philosophical themes (like his confession to be “a second generation atheist” [AR, 281] and, probably, his political placement “on the left” [AR, 287]) seems to suggest that even regarding these broader issues, he was very much his father’s son. Since readers of this volume will be well acquainted with Wilfrid’s philosophy, I will focus mostly on reconstructing the cardinal points of Roy’s thought, highlighting the commonalities of outlook with his son. However, I will also insist on a crucial difference between the two philosophers, to be sought on the methodological plane rather than in terms of substantive theoretical commitments.10 Wilfrid himself seems to endorse just this kind of interpretive approach when he writes—in the opening one of the two papers that directly engage with his father’s work—that [a] discerning student of philosophy, familiar with the writings of Sellars pere, who chances to read Sellars fils, and is not taken in by the superficial changes of idiom and emphasis which reflect the adaptation of the species to a new environment, will soon be struck by the fundamental identity of outlook. The identity is obscured by differences of terminology, method and polemical orientation, but it is none the less an identity. How natural, then, and, in a sense how true to say that Critical Realism, Evolutionary Naturalism, and all that they imply, are part of my paternal inheritance. (PR, 13)11

236  Fabio Gironi And, opening his Naturalism and Ontology lectures, he recalls how [w]hen I was coming to philosophical consciousness, the great battles between the systems which began the Twentieth Century were drawing to a close, although the lightning and the thunder were still impressive. I cut my teeth on issues dividing Idealist and Realist and, indeed, the various competing forms of upstart Realism. I saw them at the beginning through my father’s eyes, . . . After striking out on my own, I spent my early years fighting in the war against Positivism—the last of the great metaphysical systems; always a realist, flirting with Oxford Aristotelianism, Platonism, Intuitionism, but somehow convinced, at the back of my mind, that something very much like Critical Realism and Evolutionary Naturalism was true. (NAO, §1, 2) In what follows, then, I will try to highlight those aspects of Critical Realism and Evolutionary Naturalism that Wilfrid found most congenial by exposing Roy’s core commitments. In so doing, I will occasionally use a vocabulary that belongs more to Wilfrid’s generation, but that nonetheless can precisely describe his father’s tenets. First, however, it’s worth mentioning another crucial metaphilosophical commitment that Wilfrid inherits from his father (one very much relevant to this volume): the great value assigned to the history of philosophy. Time and again throughout Roy’s work, we can find him espousing the belief that philosophical investigation should always begin with an awareness of the historical genesis of whatever philosophical problem one is trying to tackle (in particular, considering his interests, epistemological problems from Descartes onwards). To be sure, Roy is very clear that such historical work should not be pursued for its own sake, and that it remains a means to novel philosophical production. Nonetheless, he was firmly convinced about the necessity of critically assessing previous (and often erroneous) historical beliefs in order to then formulate a clear statement of a philosophical problem, since the individual thinker who has seriously adventured himself on the path of philosophy can usually find at least the preliminary steps which his mind is inclined to take already examined by past thinkers. What extremely able men have so carefully done cannot help but be of assistance to those who come after. (EOP, 42) Without such preliminary historical training, Roy believes, no new philosophy of value can be produced, and indeed it would be a methodological mistake to see the two as separate enterprises: “Philosophy cannot be successfully separated from the history of philosophy” (EOP, 42). Considering

Wilfrid Sellars and Roy Wood Sellars  237 how uncharacteristic Wilfrid’s attention to the history of philosophy was— as part of a generation often adopting an explicitly anti-historicist stance—it seems reasonable to postulate that here too his father’s influence came to play an important role on his intellectual development, possibly being the root of his well-known belief that “philosophy without the history of philosophy is, if not blind, at least dumb” (SM, 1).12 In the next section, I will begin by sketching a picture of the personal relationship between the two men, which was characterized by a strong affective bond as well as by a constant and productive intellectual exchange. After this brief vignette—useful to give a human face to these intellects—I will move to the two pillars of Roy Wood Sellars’s thought, Critical Realism and Evolutionary Naturalism, the two clusters of philosophical tenets shared and re-elaborated by his son.

Father and Son Before examining more properly philosophical topics, it is worth briefly exploring what we can gather about the evolution of the personal and philosophical relationship between the two men. The first mention of Wilfrid in Roy’s published work is in the context of a 1932 book review of Friedrich Nietzsche by George Burman Foster. Roy opens his review writing that “[n]ot long ago in Germany the reviewer was helping to introduce his son to the prose magnificence of Zarathustra. We read in turns page after page out loud” (RFN, 131).13 We cannot know if these kinds of readings were a regular event or not, but thanks to Wilfrid’s own recollections, in his 197514 Autobiographical Reflections, we can suppose that they indeed were. Remembering the year he spent studying in Paris (where he attended the Lycée Louis Le Grand in the school year 1929/1930, one year before Munich) Wilfrid writes: It was here that I had my first encounter with philosophy. I say ‘my first encounter’ in all seriousness for I scarcely knew that there was a subject called philosophy, let alone that there was such a subject. It had never come up as such in any conversation with my father, at least that I can remember. . . . [The philosophy course he took at the Lycée] was thin stuff. But it did give me a sense of how philosophical issues were classified and an acquaintance with some of the major philosophers (in French perspective). It suddenly hit me that my father was a philosopher and that I knew nothing about this dimension of his existence. What my mother was able to tell me whetted my curiosity, and, by the time he joined us in February, I was eager to explore this unexpected goldmine. My father and I had always gotten along well together, but we had found little to talk about besides topics that arose per accidens from the events of the day. Indeed, he had been a distant figure who almost daily disappeared either into the University or into his study

238  Fabio Gironi in the attic, where he turned out book after book. (The muffled sound of his typewriter in almost continuous operation—or so it seemed— was later to haunt me, when I began my own attempts at publication). Thus, there is a sense in which father and son first met that spring in Paris through the good offices of philosophy. Needless to say, I found his views congenial from the start and quickly sloughed off the pseudoHegelian jargon of Marxist Naturphilosophie. More resistant (as they should have been) were the Hegelian overtones of Marxism as a schema of historical explanation. In any event, a dialogue was initiated which has continued for some forty-two years. (AR, 279, 280) And this dialogue did indeed continue, mostly in the form of an intense epistolary exchange between the two, until just a few months before Roy’s death on September 4th, 1973. In these letters15 aside from a few personal and family issues (Wilfrid’s mother’s health, the tragic death of Wilfrid’s sister Cecily in a car accident, and Wilfrid’s divorce, among other things) the two men mostly discussed philosophical issues, with Roy often offering comments on drafts of some of Wilfrid’s papers. Perhaps the most interesting— and somewhat moving—of these can be found in a letter written after having read a draft of Empiricism and the Philosophy of Mind (which he refers to as a ‘book’ probably because it was the longest thing Wilfrid had yet written, and was still unpublished). In it, Roy writes: Dear Wilfrid, Your argument is magnificent and, I think, entirely convincing. It should, I believe, represent an important step in the clarification of issues. Technically, I think, it stages an advance in philosophy of a rather epochmaking sort—It is so detailed and careful in its advance from point to point and so well phrased that I was enthralled. And closing the letter [b]ut let me come back to your book which I propose to read over and over. You have an assured competence in the mode of expression of your generation which is fascinating to me; and I know how much thought and labor has gone into its attainment. I am proud of you. Love, Dad Aside from the heartfelt fatherly pride that clearly emerges from these lines, this passage validates the main message that I want to convey: the two philosophers were brought together by a profound agreement regarding substantial philosophical issues, but set apart by a difference in language and method, caused by Wilfrid’s training in, and adherence to, the new tradition of analytic philosophy.16

Wilfrid Sellars and Roy Wood Sellars  239

Critical Realism As it was mentioned earlier—and unlike many of his contemporaries—Roy Wood Sellars often adopted an historical approach, reconstructing the history of epistemology with the aim of showing its progress as naturally leading to his own systematic—or one could even say synoptic17—philosophical vision. In his opinion, the inability to discern a middle way between direct realism and copy-theory representationalism was the great mistake of modern philosophy, which engendered the sundry epistemological mistakes of naïve realism, idealism and pragmatism. He found critical Realism (his own brand of Critical Realism, set apart from that of his colleagues, mostly George Santayana, Arthur Lovejoy and Durant Drake, and constituting what was known as the ‘essence wing’ of Critical Realism) a reformulation and correction of representationalism, or a representationalism made coherent with contemporary psychological and physiological scientific knowledge. It therefore represented a development and improvement (a ‘reflective development’ as he puts it) of naïve realism, since Roy saw the epistemological process as starting with the physical object, not with an idea that resides in the mind. In a nutshell, Roy wanted to argue that we can infer object from content, but only because content is caused by (and is about) the object in the first place. Roy’s revision of both direct realism and classical representationalism, then, means moving away from the subject-object duality towards a triad composed by subject-percept-object, where the middle term, the percept, is not to be considered an ‘entity’ nor a ‘copy’ of the object of which presents its likeness. Roy wants his Critical Realism to represent a dialectical progression beyond the stalemate between Idealism and (direct) Realism, since “only the realism which passes through idealism can hold its ground” (CC, 238). From this perspective, the New Realists had one thing right: with knowledge, we know the external world, and not mental entities; the idealists, on the other hand, were correct in thinking that knowledge has conditions, that it is not a transparent process of exchange with the external world, and that it involves a ‘personal element.’ But both had the mistaken idea that we directly intuit reality, a physical object in the first case, and an individual experience in the latter. Roy found it true that, as the idealist James Ward held, “[t]he ‘object’ of science is a construction in which conceptual elements dominate, but the possessor of this construct is still the concrete individual” (CR, 47), although according to Roy this does not undermine the independent existence of the external, physical object. He explains that it is the movement of the inner, personal sphere upon the outer, common sphere which we shall call the Advance of the Personal. The Advance of the Personal does not necessarily lead to idealism, but it does result in the recognition of the personal element in knowledge and raises

240  Fabio Gironi questions which cannot be answered without a thorough analysis of the individual’s experience. (CR, 49) Roy is adamant that the mediating elements ‘between’ (a term to be used carefully) subject and external object are ‘mental’ only to the extent that we use the term as opposed to ‘ideal’ (like essences or universals are ideal). But they are not ‘mental’ entities as they would feature in a Cartesian picture, where mental stuff is opposed to physical stuff. These contents (or percepts, or characters, or idea-objects, or sensations . . . Roy never quite settles his terminology) have no ontological status, but only a functional one. Contents are packets of information about the object, not ideal replicas of the object (instantiating a relation of likeness, like in classical representationalism) nor universals instantiated in the object (a relation of participation, like Santayana’s doctrine of essences). In a passage clearly reminding us of his son’s concerns, and exemplifying what he took to be his scientific approach to epistemology, Roy explains [H]ere is the specific problem of perception with which epistemology has always concerned itself. What is it that is given? What is the unit of perception? The naïve realist says, ‘a thing’. The critical realist says that, while this seems to be the case, reflection cannot accept it at face value. The hypothesis which best accounts for all the facts is that perception consists of the directed response to a thing and the assignment, in connection with this response, of certain characters to the things as its qualities. This is a hypothesis to explain what seems to be given in perception. The flaw in much of past theory of perception was the stress laid on the causal relation from the thing to the percipient organism, and the neglect of the response with its construction of meanings, its sense of externality and location. (CRC, 394) Contents of perception, then, are features of the cognitive process considered as a whole, interrelated organic process. They are ‘features of the field of consciousness,’ or a ‘character-complex’ that we assign to the external object, and which arises in the organism in response to the external object. There is an epistemological relation between percepts on the one hand and external objects on the other, but not a metaphysical one, because the likeness between the two is one of order and not of stuff—in Wilfrid’s terms a structural isomorphism.18 This is the meaning of Roy’s motto that there is ‘no cognitive relation’ in perception: “[t]he preposition, ‘of,’ in the phrase ‘idea of’ is not symbolic of any actual relation, but of a distinction between two spheres with different characteristics. These spheres are considered existentially distinct” (ITCR, 231; emphasis original). It is no surprise that Wilfrid considers his father’s understanding of this ‘of-ness’ as

Wilfrid Sellars and Roy Wood Sellars  241 not expressing a relation, but rather as a “ ‘subjective genitive’ which classifies the sensation” to be a “most important insight” (DKMB, 277). Indeed, this is very likely the source of Wilfrid’s own belief that “the ‘of-ness’ of sensation simply isn’t the ‘of-ness’ of even the most rudimentary form of thought” (AR, 285). For both father and son, the sensation of a blue Lego brick is the structural reproduction of an order or structure, located in the brain and produced according to environmental inputs, and not a ‘copy’ that would somehow be ‘like’ the blue plastic brick, and the content/percept ‘blue Lego brick’ does not enter in a (non-natural, semantic) cognitive relation with the plastic object but it depends on the referential directedness of an organism towards a denoted, or mediatedly knowable (yet noninferentially observable, in Wilfrid’s language) mind-independent entity located in the surrounding environment. Roy’s arguing for “an identity of order and not of material” between external object and internal representation, and his insistence that “[a] thing is an ordered material and it is this order that may arise elsewhere under its control with no identity of material” (CTB, 36) is a clear anticipation of Wilfrid’s treatment of sensations, as exemplified in his passage: [T]he pinkness of a pink sensation is ‘analogous’ to the pinkness of a manifest pink ice cube, not by being a different quality which is in some respect analogous to pinkness. . . , but by being the same ‘content’ in a different categorial form. (FMPP, 111.47) This continuity also allows us to appreciate the (often inexplicit) historical lineage of Wilfrid’s own ideas, since Roy’s stance is a dialectical evolution deriving from his engagement with the debates between idealists and new realists. In a 1912 paper Roy writes: [I]s there, indeed, a cognitive relation either external or internal? I am of the opinion that there is no such relation. . . . It seems, then, that the subject-object relation is a dogma which has been an article of faith in the philosophic world. The nearest approach, hitherto, to heresy has been the doctrine of external relations. But such a doctrine is halfhearted. We need the complete and final heresy; there is no cognitive relation. (ITACG, 225) Here Roy is referring to a long-standing epistemological debate about relations: are they internal to the relata (as the idealists held) or external (as the New Realists claimed?). Roy wants to side-step the entire conversation, because “[r]elation is a treacherous word. It is better just to say that we know things, than to say there is a cognitive relation between mind and things” (CRC, 392). Sensations—as he often puts it—are not terminal:

242  Fabio Gironi they are rather an intermediate passage in a complex process of knowledge acquisition, causally produced by the external object impinging on the organism. They are used as packets of information about, and guide the behavioral response of the organism towards, the external object. So “mere subjective occurrences, call them sensations and images or sense-data, are not ideas in the cognitive sense. It is the cognitive use of these subjective events which makes them ideas” (EN, 28). Sensations, taken in isolation, play no epistemic role: on the contrary, taking sensations as terminal leads to either phenomenalist subjectivism or naïve direct realism. The material of knowledge is not the same as the object of knowledge. Once again, Roy’s critical realism grows from, and proceeds beyond, these epistemological debates, rejecting both sides of the debate for their shared insistence on an immediately present object of knowledge—or, we could say, a given. Another novelty in Roy’s approach, as compared to most of his contemporaries, is his constant reference—guided by his conviction that epistemology needs be approached as a science, and informed by it—to psychological, physiological, biological and evolutionary explanations. As he insists time and again I have unmistakably stood for the view that the content of perception and of judgmental knowledge is intra-organic in its existential locus. . . . Things seem to appear or to reflect themselves in the contents which they control in the organism: and on the theory of realism they would do just this. (CRC, 379) [e]xistentially, [sensations] might well turn out to be intracortical. But the operation of perceiving itself would be tied in with response and concern itself with the objects to which the organism was adjusting itself. (DMV, 289) The idea is that these sensations are “systems of neurons” (EOP, 80) that exhibit a structural isomorphism with the external object, but that our perception pertains to the external object which stimulated the creation of these intracortical states, and not to these internal representations. Roy therefore speaks of a ‘to-and-fro’ process of perception: a feedback circle starting from the reception of stimuli from the environment, passing through conceptual knowledge of this environment grounded on the sensations that these stimuli produce in us (sensations thereby causally dependent on the external object), and leading to actions upon the environment guided by our knowledge of it. While what is ‘present’ to us is not the external object directly (as the new realists would have it), what we know is the object itself, and not our ‘representation’ of it (as classical representationalism suggested). In a characteristically phrased expression, Roy

Wilfrid Sellars and Roy Wood Sellars  243 explains that “perceptual knowing is concerned with the things environing the organic percipient” (NNO, 692). In sum, his naturalistically motivated insistence on a mediated and complex process of perception, rejecting the direct realism of the new realists, mirrors his son’s polemical stance against sense-data empiricism. Indeed, he clearly foreshadows Wilfrid’s ideas about the epistemic inertness of mere sensory data and his conception of perceptual experience as always conceptually mediated.19 And in a passage like the following he even seems to prefigure the crucial role given to language by Wilfrid (educated, as we have observed, within an analytic tradition which put far more emphasis on the analysis of language than Roy’s generation ever did): [M]ay it not be that the perceptual experience was oversimplified in traditional empiricism because of the too dominant causal approach? What if perceiving is responsive, symbolic, interpretive, judgmental? What if it must be regarded as a higher-level activity than sensing? Would it not then be probable that language, as a social affair, would, on the whole, reflect perceiving more than sensing? (CP, 535)

Evolutionary Naturalism Hostile to idealism and to theism, Roy was nonetheless far from being orthodox in his adoption of naturalism, another trait that later emerged in Wilfrid’s own “naturalism with a normative turn” (to use O’Shea’s now-canonical phrase). Roy in fact sees himself as opposing the ‘old naturalism,’ by which he means either the crudely biologistic naturalism of Huxley and Spencer, supported by a reductionist matter-in-motion kind of materialism, or a positivist stance, agnostic about scientific entities: both positions, he thought, ignore the complexity of contemporary science. Roy’s guiding idea is that naturalism as a worldview must evolve by tracking the development of the sciences, particularly the (then more recent) biological ones: past naturalisms “did not take evolution seriously nor did it take mind seriously” because they “founded themselves upon the results of the exact sciences alone, leaving out the levels of the organic and social behavior” (EN, 16, viii). Ultimately Roy’s core belief is that “[t]o attempt to solve the basic queries as to the nature of life in the light of physics alone is to challenge failure or a resort to sophistry” (EN, 6). The signature move of the ‘old naturalism’ was analysis; the new naturalism—Roy’s evolutionary naturalism—(and, in fact, his philosophy as a whole) must instead strive for synthesis. A synthetic effort to be applied on different levels: an inter-disciplinary methodological synthesis of science and philosophy; an intra-scientific synthesis of physics, biology, chemistry, psychology and sociology; a naturalist/ontological synthesis of mind and body;

244  Fabio Gironi and finally a naturalist/cosmological synthesis of nature as an evolutionarily interrelated whole. Hence the stress placed on the concepts of organization and emergence (the latter term employed without any supernatural or spiritualistic connotation): different configurations of physical elements— different structures or forms of organization—naturally give rise to new functions, irreducible to their components. Roy’s naturalism, like his critical realism, rejects any form of dualism and entails a world of immanence: [A]ll events are in the one world, and there are currents passing back and forth with nothing alien and imported from outside. Evolution and devolution, the higher and the lower, the simple and the complex, are components of the one great physical theatre. (NM, 223) This understanding of naturalism needs to be contextualized in the larger debates of the time, where early forms of physicalist naturalism were challenged by the still influential defenders of Idealism, on the grounds of their inability to account for the experiential realities of everyday life. The new naturalism, then, was keen to reject these accusations: as Roy’s fellow critical realist, George Santayana, put it in an article addressing Dewey’s naturalism in Experience and Nature, [W]e are not compelled in naturalism, or even in materialism, to ignore immaterial things; the point is that any immaterial things which are recognized shall be regarded as names, aspects, functions, or concomitant products of those physical things among which action goes on. A naturalist may distinguish his own person or self, provided he identifies himself with his body and does not assign to his soul any fortunes, powers, or actions save those of which his body is the seat and organ. (1925, 674) The common shift then was from a synchronic mechanistic naturalism of matter-in-motion to a more markedly Darwinian (and diachronic) one, where function and organization play an ineliminable role (a stance that is far more obvious today than it was in the early decades of the twentieth century). In Roy’s case, just as with critical realism, evolutionary naturalism is meant to be a synthesis of the best insights of scientific naturalism and idealism: in a certain sense, it would then be a naturalization of idealism.20 Such a non-reductive evolutionary naturalism (sometimes called by Roy a ‘reformed materialism’)21 does not reduce higher sciences to (mathematized) fundamental physics and, in consequence, does not reduce the human (the person) to a mere physical assemblage. In other words, according to Roy we need to find a place, in nature, for the characteristics of persons that are not strictly reducible to physics: “The whole of man must be included in nature, and nature so conceived that his inclusion is possible” (EN: 20). This clearly

Wilfrid Sellars and Roy Wood Sellars  245 reminds us of Wilfrid’s own interest, made explicit in Philosophy and the Scientific Image of Man, in articulating a stereoscopic vision through the fusion of the scientific and the manifest images of man-in-the-world, where he reminds us that “the very fact that I use the analogy of stereoscopic vision implies that as I see it the manifest image is not overwhelmed in the synthesis” (PSIM, 9). The nature described by evolutionary naturalism is one wherein the mode of organization of a system has a causal efficacy (such that different modes can produce different effects), different from the mere physical impact kind of causation included in a purely Newtonian naturalist picture. Roy insists that function is an intrinsic feature of physical structure, not a super-added property; evolution is the process which, by mutating the underlying structures, allows for new functions of a system to emerge The crucial outcome of the ontological worldview afforded by evolutionary naturalism (supplementing the epistemological approach of critical realism) is the solution of what Roy considered the single most important problem of philosophy. As he clearly puts it: [I]t is still—as it has always been—my opinion that the adequate handling of the mind-body problem represents the synthetic stage of any philosophy and is at one and the same time a supreme test and an indication of its power. Epistemology, ontology, and science must be marshalled together and all the essential terms of the problem must be defined and reintegrated. (AAMB, 461) The sapient human being cannot be reduced to its physical properties (as the crude mechanistic naturalism attempted to do) but rather it exhibits properties that are non-physical while still natural—first of all consciousness. In Wilfrid’s language, Roy wants to acknowledge that consciousness is a central and ineliminable item in the conceptual scheme of the manifest image, as the name for our first-personal, private awareness of contents/ mental states. On the other hand, the physicalist conceptual scheme of the scientific image seems to have no space for such an item. A stereoscopic fusion is therefore needed, where consciousness exists (and in particular, has a kind of causal power over us) but is not anything non-natural: “[C]onsciousness is an event and not a thing or stuff; and it is an event adjectival to the brain. It is in this latter that we must find enduring patterns and principles of integration” (PPR, 408). Roy then promotes a double-knowledge approach to the mind-body problem (or, more accurately, to the consciousness-brain problem). According to this view we have a double knowledge of ourselves: (i) an external (scientific, factual) knowledge of mind, deriving from sense-perception and observations of the causal interaction of mind with the world and (ii) an internal, introspective knowledge on our own private and qualitative

246  Fabio Gironi field of consciousness. But these two perspectives entail neither dualism nor monism, since they represent two ways of knowing the same reality: together they offer a synoptic vision of humans as both physical organisms in a natural environment and as conscious persons in a social environment. Yet here we find the only (public) occasion in which Wilfrid had a (admittedly minor) disagreement with his father. Although he believed that Roy’s functional interpretation of consciousness—according to which “consciousness is not a physical system but a qualitative dimension of the existential content of a highly evolved physical system, . . . [it] is in the brain after the manner that an event, or state, is in that of which it is a state” (AAMB, 476; emphasis original)—was, “fundamentally correct” (DKMB, 284), Wilfrid wanted to push further the identification of consciousness with neurophysiological processes. Taking issue in particular with passages like “physical science, that is, science which deciphers nature in terms of the revelatory capacity of sense data, must ignore consciousness altogether” (PPR, 421) (because, as we have seen, physical science would only be able to offer an ‘external’ description), and “[consciousness] can have no causal significance for science, since this is always dealing with the brain-mind and its states as physical events” (PPR, 424), Wilfrid insists that it is a mistake to limit the causal power of internal states, and of consciousness itself, and to assume that neuroscience would in principle be unable to capture those qualitative (‘internal’) features (and not just the quantitative/structural ones) of these states. So Wilfrid—rhetorically—asks: [W]hy could not concepts of sensible redness, etc. be introduced into a theory of the visual cortex as concepts of certain qualitative features of neurophysiological events which play specific roles in the functioning of the visual system centers, and in the discriminative behavior of the larger system which is the organism as a whole? (DKMB, 287) This objection demonstrates Wilfrid’s greater consideration for (and perhaps understanding of) neurophysiology, and how his—perhaps speculative— ideas for a future science are in principle capable of defining fully determinable counterparts of the qualitative concepts which Roy limited to the realm of consciousness. Although Roy’s skepticism might be justified in practice, “this practical scepticism must not be confused with an impossibility in principle of such a theory” (DKMB, 287). This stance is consistent with Wilfrid’s introduction22 of the distinction between physical1 and physical2, and his insistence that (in agreement with his father) any attempt to account for qualitative features of experience in terms of the latter (mere microphysical causal relations) is destined to fail, but it also moves beyond his father’s stance by stating that there can indeed be causal explanations couched in physical1 terms that will be able to do justice to qualitative experience. This is a deep-seated commitment that will eventually lead to his adoption of the

Wilfrid Sellars and Roy Wood Sellars  247 terminology of process ontology (a development that Roy would never witness, since Wilfrid started using these terms to describe “the non-particulate foundation of the particulate image” (PSIM, 37) only in the early 1980s in his Carus Lectures, several years after his father’s death).

A Methodological Change Throughout his philosophical career, but particularly towards its end, Roy Wood Sellars was acutely aware of, and profoundly embittered by, what he perceived as the lack of appreciation of his ideas among his peers. While this was true even in the 1920s and early 1930s (the peak of the popularity of the critical realist movement, during which he felt that his stance was overshadowed by the more popular ‘essence wing’ of critical realism, spearheaded by George Santayana),23 it was the sudden popularity of the new British/ European import—analytic philosophy—that most severely bothered Roy. Writing in 1969 and looking back at the 1930s and 1940s, he recalls how [i]t seemed to me that the so-called analytic philosophy which got quite a vogue was ambivalent. In one sense, I liked its emphasis. In another sense, it did not seem to me very creative in either epistemology or ontology. American addiction to it and disregard of its own momentum struck me as a form of neo-colonialism. (RAPW, 5) Aside from personal resentment, it is undeniable that the new language, methods and emphasis brought about by analytic philosophy (via the popularity of Russell and Moore first across the Atlantic, and via the immigration of numerous Austrian and German logical positivists later) produced a radical change in American philosophy, creating a new environment where the kind of style and systematic approach to philosophy favored by Roy became very quickly outdated and outmoded. The man who often insisted that the reconstruction of the history of philosophical beliefs was a necessary step towards the formulation of new philosophical statements found himself, and his tradition, forgotten by younger—and often militantly anti-historical—philosophers. But, as unfortunate as this was for Roy (and for his entire philosophical generation), this also makes Wilfrid’s role in the history of American philosophy all the more important and, indeed, utterly unique. The connection to his father afforded him a direct association with the vigorous home-grown epistemological debates of the 1910s, 1920s, and 1930s, a tradition of which most (if not all) of the major American philosophers of his generation— whose philosophical training was mostly rooted in logical positivism or, at best, in some variety of pragmatism—showed very little knowledge. Albeit fully convinced, unlike his father, of the clear methodological superiority of the new formal tools and strategies offered by analytic philosophy, Wilfrid

248  Fabio Gironi was also capable of placing his own thought in a precise historical stream of philosophical problems, with the aim of reformulating those problems in a clearer fashion rather than abandoning or forgetting them as quaint curiosities. And Roy was well-aware of the greatness of his son’s accomplishments, even if the new tradition to which he clearly belonged had eclipsed his own work.24 As he somewhat painfully confesses in a letter dated February 1973 (written only a few months before his death), and offering some comments on a draft of Wilfrid’s Autobiographical Reflections: [I]t is clear that you belonged to an era in which I existed—how shall I put it?—largely as an ignored unbranded background figure, as Delany gently puts it in his recent article in The New Scholasticism.25 Sir Alfred Ayer, Lewis Price, etc. got the limelight. I just had to go ahead as well as I could. It required some fortitude. I am too old now to care much. But the men you mention26 I still think knew little about what Kurtz calls the heroic age of American Philosophy.27 Neither the positivists nor the English were cognizant of it. I don’t believe I tried to indoctrinate you. You had to work out your own view. And I am proud of what you have done. Wilfrid indeed managed to forge his own path, merging the new philosophical style with the concepts and concerns proper to his father’s generation and their debates (although, like his father, never quite achieving as much success with his peers as he would have wished). In one of his earliest papers, published in 1948, when he was still a young and relatively unknown philosopher, Wilfrid refers precisely to the debates about realism in which his father took part writing: [T]he acts of the tragedy (though not always performed in this order) were Naive Realism, New Realism, Critical Realism, Idealism, Pragmatism, and Epistemological Solipsism of the Present Moment. It has become increasingly clear, in the course of the past decade28 that this particular tragedy was based on a mistake; on an asking of the wrong, or better, of a confused question. This suggests immediately, . . . that the curtain is being rung down on this particular cluster of controversies, and that new dramatis personae are moving to the center of the stage. This is true; but those considerations also suggest that while the newer questions may be clearer, they will none the less be in essence the same, and that consequently the new play will be the old, cut and adapted to modem dress. The empirical and the formal, the psychological and the epistemological will be more clearly distinguished, yet the competing points of view will be found capable of translation into the new frame of reference, if only to be curtly dismissed. (RNWW, 602–3)

Wilfrid Sellars and Roy Wood Sellars  249 There’s no clearer statement of Wilfrid’s approach to this complex philosophical heritage, a legacy that is often ignored when approaching his thought. Indeed his own philosophical trajectory could be reconstructed as initially parting ways with his father (a product of his infatuation with the formal analysis promoted by the ‘new way of words’),29 and then slowly coming back to it, with the gradual abandonment of his early formalist metaphilosophy.30 Roughly put, his youthful passion for formal tools slowly gave way to a return to an empirical-psychological approach which more closely resembles his father’s interest in explanations grounded on the best insights of psychology and evolutionary biology—something of a (minor) vindication of his father’s opinions regarding the overtly narrow nature of analytic philosophy. As he writes in a letter to Wilfrid in 1954 It has always seemed to me that this school tried to get objectivity through some linguistic hocus-pocus instead of getting a sound basis in the epistemology of perception which the realists over here were trying to do. For me, language behavior develops within the setting of pointing and social action connected with the objective intent of perceiving.

Conclusion This chapter has attempted to demonstrate (in all too brief a fashion) the profound human and philosophical connection between Sellars father and Sellars son. On these grounds, I would indeed like to make the (bold) claim that the intellectual trajectory and systematic thought of Wilfrid Sellars cannot be fully appreciated unless the critical realism and the evolutionary realism defended by his father are studied and comprehended. Roy Wood Sellars should be considered a key philosophical influence on Wilfrid—just as much as Carnap or Wittgenstein (to cite but two of the most proximate philosophical sources) were. Unfortunately, due to decades of neglect of his work (today rather quaint in appearance),31 and to the relative unfamiliarity of contemporary philosophers with the 1890 to 1930 philosophical period in the US, today Roy is considered as more of an historical curiosity— whose role is compressed in the standard disclaimer: “Sellars’s father, a significant philosopher in his own right”—than an interesting author worth investigating. And yet, as I hope I have shown, Roy Wood Sellars was a crucial influence on his son, particularly regarding deeply rooted shared commitments such as: (i) the insistence on a close dialogue between science and epistemology; (ii) a metaphilosophical vision interpreting philosophy as a systematic humanist enterprise; (iii) the development of a critical realist epistemology wary of immediacy or givenness; (iv) the adherence to a nonreductive evolutionary naturalism capable of acknowledging the reality of human qualitative experience and values without explaining them away in physicalist terms; and finally (v) a keen awareness of the importance of the

250  Fabio Gironi history of philosophy. To close I would turn to Wilfrid’s own words, when, commenting on the state of ‘realism’ in the early 1970s, he complained that such a movement “seems to be largely unaware of its historical antecedents, and one can wonder whether a philosophy can be truly critical which lacks the awareness without which, historians assure us, one is doomed to make old mistakes anew” (DKMB, 270).

Notes 1 During his long philosophical career (from the first years of the 1900s to 1973, the year of his death), Roy Wood Sellars published over ten books, between monographs and collections, and over 100 articles. He was President of the Western Division of the American Philosophical Association in 1924, and in 1971 his friend and biographer Preston Warren established a lectureship in his name at his home institution (Bucknell College): the Roy Wood Sellars Lectureship in American Philosophy. Warren himself was the first lecturer of the series in 1971, Wilfrid Sellars in 1972, and in the decades to follow several preeminent American philosophers have been Roy Wood Sellars lecturers, including Roderick Chisholm, Stanley Cavell and Hilary Putnam. 2 Indeed, even Roy’s work in isolation has not been the object of much scholarly attention in the last several decades. The most recent monograph solely dedicated to Roy’s work, published in 1975, is Warren’s Roy Wood Sellars, (Warren 1975) while the other major monograph is Melchert’s Realism, Materialism, and the Mind, published in 1968. Relatively recent journal articles (though still over twenty years old) dedicated to Roy Wood Sellars are Wright (1994) and Slurnik (1996). 3 See Holt et al. (1910 and 1912). 4 See Drake et al. (1920). 5 As William Frankena recalled “[u]ntil 1930 at least idealism and pragmatism dominated in philosophy, theism in religion, and capitalism in social theory. He was bucking them all” (1973/1974, 231). 6 He is indeed the first philosopher to adopt this name for his position, which appears in the title of one of his earliest papers in 1908, as well as serving as the title of his 1916 monograph (CR) 7 As he explains, “physical realism” should be seen as “a shortened expression for critical realism and evolutionary naturalism taken together. Thus it symbolizes the integration of epistemology and ontology” (CS, 2) 8 Not the most brilliant stylist, with the exception of a few particularly enlightening passages, Roy also tends to repeat his arguments throughout his publications, complicating the interpretive work by changing his terms as time goes by. For example, the terms ‘content,’ ‘sense-datum,’ ‘percept,’ ‘appearance’ and ‘presentation’ (among others) are used as near-synonyms across his texts. 9 Roy himself summarizes his thought—insisting on its systematicity—looking back to his contribution to the 1930 volume Contemporary American Philosophy: Personal Statements (Adams and Montague 1930), recalling how “I made my contribution to this project and I allowed myself the luxury of bringing it under three headings: realism, naturalism, and humanism. Realism stood for my epistemology. Naturalism for my ontology and cosmology. And humanism for the domain of valuation. It was thus comprehensive in import” (RAPW, 57). 10 Frankena observes that Roy could “enjoy the pleasure of being the first part of the most successful father and son combination so far in the philosophical scene, especially since his son’s views are so similar in substance though very different in form” (1973/1974, 232).

Wilfrid Sellars and Roy Wood Sellars  251 11 Playing on his father’s terminology and interests, Wilfrid goes on to write “[h]ow tempting to explore the network of stimulus and response, or perhaps the depth—psychological forces that gave one mind the shape of another. This, however, I am not going to do, at least on the present occasion” (PR, 13). Unfortunately, he never did explore this connection in depth, nor did anyone else in the decades to follow. This chapter and Gironi (2017) are attempts to do just this. 12 Although their opinions diverged about which historical figures were most important: like all realist philosophers of his generation, Roy had no time for the German idealist tradition, Kant and Hegel above all. 13 Thanks to the mention of a stay in Germany, we can gather that here Roy was probably thinking of the family’s short stay in Munich, in 1931, where Roy taught at the University of Munich and Wilfrid, then 19 years old, studied for a year and learned German. 14 1975 is the date of publication of the collection dedicated to Wilfrid’s thought (Castañeda 1975) which includes his Autobiographical Reflections. However, this was probably written a few years earlier, since in a letter dated February 1973, Roy tells Wilfrid how he had just read it. 15 Much of the Roy-Wilfrid correspondence is available in the Sellars Archive at the University of Pittsburgh (Box 162, Folders 4–11) and on-line at http://digital. library.pitt.edu/cgi-bin/f/findaid/findaid-idx?type=simple;c=ascead;view=text;su bview=outline;didno=US-PPiU-asp199101. I will quote from a few of these letters Roy wrote to his son. As the decades went by, the decline of Roy’s steadiness of hand and typewriting skills are evident, and the later letters from the early 1970s contain a large number of misspellings and punctuation errors, which I have endeavored to correct in order to aid clarity. 16 Something else should also be noted: the lectures that later became Empiricism and the Philosophy of Mind, were delivered at University College London in March 1956, originally titled ‘The Myth of the Given: Three Lectures on Empiricism and the Philosophy of Mind.’ It is safe to assume that these lectures were the result of a long period of philosophical reflection spanning the immediately preceding years. It is then interesting to note that in 1954 Wilfrid participated in a symposium organized in his father’s honor, and published on the pages of Philosophy and Phenomenological Research. His contribution, the paper ‘Physical Realism’ (PR) is an engagement with what Wilfrid considered the most mature and systematic exposition of his father’s stance, the book The Philosophy of Physical Realism (PPR). In particular, this paper is an exegesis of his father’s critical realism, a position (as I will explain later) whose polemical target has some clear connections to Wilfrid’s own objectives in his attack to the Myth of the Given. That is to say, it seems to me very reasonable to assume that the time spent going back to his father’s work—just around the same time that he was developing his own canonical form of the critique of the Myth of the Given— had an influence on the way he tackled these epistemological problems, thus representing a very direct example of Wilfrid re-elaboration, in a new key, of positions he shared with his father. 17 Indeed, the very expression ‘synoptic vision’—so familiar to Sellarsians because of its employment in PSIM, describing an alternative to the purely analytic conception of philosophy—can be found in one of Roy’s earliest papers. In the opening lines of his ‘An Important Antinomy’ he writes that “[i]t not seldom occurs that the presence in an individual’s mind of cognate problems leads by slow process of assimilation to a new viewpoint in which what was before separate, gathers fresh import in a larger whole. Such is, undoubtedly, the genesis of all constructive inference; the hewing of wood and drawing of water being but preliminary to the synoptic vision without which the weary task—work avails little” (IA, 237).

252  Fabio Gironi 18 Wilfrid pretty much restates in his own words his father’s position when he comments that “[o]ur problem, of course, is how this ‘likeness’ is to be construed, if the propositional character of the ‘idea’ is taken seriously; that is to say, if we are to preserve the essence of Hume’s contention, while avoiding his mistake of thinking of ‘ideas’ as likenesses in the sense of duplicates. This essence is the contention that the ‘likeness’ between elementary thoughts and the objects they picture is definable in matter-of-factual terms as a likeness or correspondence or isomorphism between two systems of objects, each of which belongs in the natural order.” (TC, 219) 19 See Gironi (2017) for a more in-depth discussion of how Roy and Wilfrid disagree regarding the importance of Kant, and the role of the categories. 20 On several occasions Roy mentions the influence that James Ward’s Naturalism and Agnosticism had on his thought, since Ward’s critique of naturalism made clear to him that it needed stronger foundations. See for example R. W. Sellars AAMB, 462. 21 R.  W.Sellars NMO. 22 In CE. 23 As he complains, “too much stress was laid on the essence wing of critical realism. Probably, Santayana’s prestige had something to do with this outcome” (RAPW, 47). 24 To an extent, Wilfrid is indirectly responsible for the institutional neglect of his father’s work and of the tradition to which he belonged. The landmark collection Readings in Philosophical Analysis—which he co-edited in 1949 with Herbert Feigl, his then colleague at the University of Minnesota—became the standard introductory textbook for an entire generation of American philosophers and, explicitly intended as a collection introducing the new philosophical style, it did not include any article from the debates in which Roy participated (Feigl and Sellars 1949). The one article included in the collection related to the realism debates of the 1910–30 period was W. T. Stace’s ‘The Refutation of Realism.’ And indeed, in a letter to his son Roy complains about this addition: “[B]y the way, how did you and Feigl come to include Stace’s Refutation of Realism in your analytic book? It is merely directed against the new, or presentational, realism. Whipping a dead dog” (July 1954). 25 I believe that Roy was here misquoting a passage where Delaney talks of him as “[s]omewhat of a maverick in the history of American philosophy, he was an unabashed epistemologist when such was not fashionable” (Delaney 1971, 470). 26 Among the many philosophers Wilfrid mentions as having had an influence on his thought (and probably those Roy is referring to here) are G. E. Moore, C. D. Broad, H. A. Pritchard, H. H. Price, Herbert Feigl and Gustav Bergmann. 27 Here Roy is probably misquoting the phrase “The Golden Age of American Philosophy” used by Paul Kurts in the introduction to his collection (Kurtz 1966), referring to the period from 1880 to 1940. 28 So, roughly the ten years between 1937 and 1947, precisely the period of everincreasing fame of the numerous logical positivists (including Feigl, Carnap, Reichenbach, Bergmann, Tarski and Hempel) who fled from Europe to the US between 1930 and 1939. 29 A clear statement of his early stance can be found in his 1947 PPE, where he writes that “[t]he analytic movement in philosophy has gradually moved towards the conclusion that the defining characteristic of philosophical concepts is that they are formal concepts relating to the formation and transformation rules of symbol structures called languages. Philosophy, in other words, tends to be conceived of as the formal theory of languages” (PPE, 181). 30 A process carefully traced in Olen (2016). 31 And occasionally hard to find: none of the major monographs of Roy Wood Sellars has ever been reprinted.

Wilfrid Sellars and Roy Wood Sellars  253

References Adams, George P. and Montague, William Pepperell. (Eds.). 1930. Contemporary American Philosophy: Personal Statements. London: George Allen & Unwin Ltd. Castañeda, Hector-Neri (Ed.). 1975. Action, Knowledge and Reality: Critical Studies in Honor of Wilfrid Sellars. Indianapolis: Bobbs-Merrill. Delaney, Cornelius F. 1971. ‘Recent Work on American Philosophy’. The New Scholasticism, 45, 3: 457–77. Drake, Durant, Lovejoy, Arthur O., Pratt, James B., Rogers, Arthur K., Santayana, George, Sellars, Roy Wood and Strong, Charles A. (Eds.). 1920. Essays in Critical Realism: A Co-operative Study of the Problem of Knowledge. New York: Palgrave Macmillan. Feigl, Herbert and Sellars, Wilfrid. (Eds.). 1949. Readings in Philosophical Analysis. New York: Appleton-Century-Crofts Inc. Frankena, William K. 1973/1974. ‘Roy Wood Sellars 1880–1973’. Proceedings and Addresses of the American Philosophical Association, 47: 230–2. Gironi, Fabio. 2017. ‘A Kantian Disagreement Between Father and Son: Roy Wood Sellars and Wilfrid Sellars on the Categories’. Journal of the History of Philosophy, 55, 3: 513–36. Holt, Edwin B., Marvin, Walter T., Montague, William P., Perry, Ralph B., Pitkin, Walter B. and Spaulding, Edward G. (Eds.). 1910. ‘The Program and Platform of Six Realists’. The Journal of Philosophy, Psychology and Scientific Methods, 7: 393–401. Holt, Edwin B., Marvin, Walter T., Montague, William P., Perry, Ralph B., Pitkin, Walter B. and Spaulding, Edward G. (Eds.). 1912. The New Realism: Cooperative Studies in Philosophy. New York: Palgrave Macmillan. Kurtz, Paul. 1966. American Philosophy in the Twentieth Century: A Sourcebook from Pragmatism to Philosophical Analysis. New York: Palgrave Macmillan. Melchert, Norman P. 1968. Realism, Materialism, and the Mind: The Philosophy of Roy Wood Sellars. Springfield: Charles C. Thomas Publisher. Olen, Peter. 2016. Wilfrid Sellars and the Foundations of Normativity. Basingstoke: Palgrave Macmillan. Santayana, George. 1925. ‘Dewey’s Naturalistic Metaphysics’. The Journal of Philosophy, 22, 25: 673–88. Sellars, Roy Wood (CC). 1908. ‘Consciousness and Conservation’. The Journal of Philosophy, Psychology and Scientific Methods, 5, 9: 235–8. Sellars, Roy Wood (IA). 1908. ‘An Important Antinomy’. The Psychological Review, 15, 4: 237–49. Sellars, Roy Wood (ITCR). 1912. ‘Is There a Cognitive Relation?’ The Journal of Philosophy, Psychology and Scientific Methods, 9: 225–32. Sellars, Roy Wood (CR). 1916. Critical Realism: A Study of the Nature and Conditions of Knowledge. Chicago: Randy McNally & Co. Sellars, Roy Wood (EOP). 1917. The Essentials of Philosophy. New York: The Macmillan Company. Sellars, Roy Wood (CTB). 1922. ‘Concerning “Transcendence” and “Bifurcation” ’. Mind, 31: 31–9. Sellars, Roy Wood (EN). 1922. Evolutionary Naturalism. Chicago: Open Court. Sellars, Roy Wood (CRC). 1924. ‘Critical Realism and Its Critics’. The Philosophical Review, 33, 4: 379–97.

254  Fabio Gironi Sellars, Roy Wood (NM). 1927. ‘Why Naturalism and Not Materialism?’ The Philosophical Review, 36, 3: 216–25. Sellars, Roy Wood (PPR). 1932. The Philosophy of Physical Realism. New York: The Macmillan Company. Sellars, Roy Wood (RFN). 1932. ‘Review of Friedrich Nietzsche by G. B. Foster and C. W. Reese’. The Journal of Religion, 12: 131–32. Sellars, Roy Wood (AAMB). 1938. ‘An Analytic Approach to the Mind—Body Problem’. The Philosophical Review, 47, 5: 461–87. Sellars, Roy Wood (CS). 1943. ‘Causality and Substance’. The Philosophical Review, 52, 1: 1–27. Sellars, Roy Wood (CP). 1944. ‘Causation and Perception’. Philosophical Review, 53: 534–56. Sellars, Roy Wood (NNO). 1944. ‘Does Naturalism Need Ontology?’ The Journal of Philosophy, 41: 686–94. Sellars, Roy Wood (DMV). 1968. ‘In Defense of Metaphysical Veracity’. In The Philosophy of C. I. Lewis. Edited by Paul Arthur Schilpp, 287–308. La Salle: Open Court. Sellars, Roy Wood (RAPW). 1969. Reflections on American Philosophy from Within. Notre Dame: University of Notre Dame Press. Sellars, Wilfrid (PPE). 1947. ‘Pure Pragmatics and Epistemology’. Philosophy of Science, 14, 3: 181–202. Sellars, Wilfrid (RNWW). 1948. ‘Realism and the New Way of Words’. Philosophy and Phenomenological Research, 8: 601–34. Sellars, Wilfrid (PR). 1954. ‘Physical Realism’. Philosophy and Phenomenological Research, 15: 13–32. Sellars, Wilfrid (PSIM). 1963. "Philosophy and the Scientific Image of Man". In Science, Perception, and Reality. Atascadero, CA: Ridgeview Publishing Company. Sellars, Wilfrid (SM). 1968. Science and Metaphysics: Variations on Kantian Themes. Routledge & Kegan Paul. Sellars, Wilfrid (DKMB). 1971. ‘The Double—Knowledge Approach to the Mind— Body Problem’. The New Scholasticism, 45: 269–89. Sellars, Wilfrid (AR). 1975. ‘Autobiographical Reflections’. In Action, Knowledge and Reality: Critical Studies in Honor of Wilfrid Sellars. Edited by Hector-Neri Castañeda, 277–93. Indianapolis, IN: Bobbs-Merrill. Sellars, Wilfrid (FMPP). 1981. ‘Foundations for a Metaphysics of Pure Process (The Carus Lectures)’. The Monist, 64: 3–90. Sellars, Wilfrid (NAO). 1996. Naturalism and Ontology. Atascadero, CA: Ridgeview Publishing Company. Sellars, Wilfrid and Meehl, Paul E. (CE). 1956. ‘The Concept of Emergence’. In Minnesota Studies in the Philosophy of Science, Volume I: The Foundations of Science and the Concepts of Psychology and Psychoanalysis. Edited by Herbert Feigl and Michael Scriven, 239–52. Minneapolis, MN: University of Minnesota Press. Slurink, Pouwel. 1996. ‘Back to Roy Wood Sellars: Why His Evolutionary Naturalism Is Still Worthwhile’. Journal of the History of Philosophy, 34: 425–49. Warren, W. Preston. 1975. Roy Wood Sellars. Boston: Twayne Publishers. Wright, Edmond. 1994. ‘A New Critical Realism: An Examination of Roy Wood Sellars’ Epistemology’. Transactions of the Charles S. Peirce Society: A Quarterly Journal, XXX: 477–514.

Conclusions

13 Thinking With Sellars and Beyond Sellars On the Relations Between Philosophy and the History of Philosophy Dionysis Christias 1. Introduction In this paper it will be suggested that Sellars’s metaphilosophy can provide support for the thesis that the history of philosophy is internally related to what philosophy is, in that the former is implicated in the very content of philosophical problems. In order to show this we shall, first, sketch a postKantian ‘critical’ conception of philosophy according to which the latter can be practiced adequately only if it is self-critical and self-reflective, i.e., only if it makes a conscious effort to make explicit and justify its own methods and presuppositions, and identify its limits. The connection between this post-Kantian critical conception of philosophy and the metaphilosophical issue of concern here emerges if we realize that a philosophical framework can be self-reflective and self-critical in the above sense only if, among other things, it incorporates its history in its own subject-matter, and attempts to understand and critically engage with it. It will be suggested that support for the above conception of the relations between philosophy and the history of philosophy can be found in Sellars’s rejection of the Myth of the Given in all its forms (including what might be called the Myth of the ‘Philosophical’ Given) as well as in Sellars’s well-known distinction between the manifest and the scientific image, understood as operating at a meta philosophical level. As is well known, unlike most analytic philosophers, Sellars had the highest respect for the history of philosophy, offering novel interpretations of major figures in the history of philosophy, such as Plato, Aristotle, Descartes, Locke, Berkeley, Hume, Leibniz and, above all, Kant. And he is also known for suggesting that “the history of philosophy is the lingua franca which makes communication between philosophers, at least of different points of view, possible. Philosophy without the history of philosophy, if not empty or blind, is at least dumb” (SM, 1). Yet, I take it that, for Sellars, the importance of the history of philosophy is not exhausted in its facilitating communication among different philosophical perspectives. The history of philosophy is more intimately connected with philosophy proper: it can be suggested that the history of

258  Dionysis Christias philosophy is implicated in the very content of contemporary philosophical problems. However, I will not here attempt to derive Sellars’s views on the relations between philosophy and the history of philosophy from his own (scattered) remarks about this issue. Instead, I shall first sketch a metaphilosophical view according to which the internal relation between philosophy and the history of philosophy follows from a certain way of understanding philosophy proper—namely by understanding philosophy as an essentially self-reflective enterprise—and, then, argue that the Sellarsian rejection of the Myth of the Given in all its forms (including what might be called the Myth of the ‘Philosophical’ Given) as well as the Sellarsian distinction between the manifest and the scientific image, understood as operating at a meta philosophical level, can throw light on the specific way in which philosophy can be self-reflective by incorporating the history of philosophy. Finally, I will attempt to show that the previous Sellarsian metaphilosophical distinction can be used for purposes of self-critique, namely as a tool for identifying the limits of Sellarsian philosophy itself, thereby rendering it capable of ‘transcending’ these limits from within.

2. Philosophy as an Essentially Self-Critical and Self-Reflective Enterprise Before we start developing our interpretation—and extension—of Sellars’s views on metaphilosophical issues concerning the relation between philosophy and its history, it is instructive to say a few things about the philosophical tradition which first explicitly thematized the relation of philosophy with its history in the manner that, as will be argued, is most congenial to Sellars’s own views on the matter.1 This is the Kantian critical conception of philosophy, which was further developed in Hegel’s self-reflective view of philosophy. The critical turn in philosophy, inaugurated by Kant, and his Critique of Pure Reason, and ‘completed’ by Hegel, developed an essentially self-critical and self-reflective conception of the philosophical enterprise. Kant was the first to argue that traditional metaphysical doctrines of the past, by supposing that that there are some eternal (necessary) entities out there in the world ‘as it is in itself’ corresponding to the laws (necessity) or reason, were essentially forms of ‘transcendental illusion’ produced by the self-hypostatization of reason. Building on this Kantian critique of metaphysics, Hegel went even further and took Kant to task for not seeing that the ‘transcendental illusion’ in question could find expression even in philosophy’s conception of the source of its own doctrines. There was no reason to suppose that the doctrines of Kantian critical philosophy were themselves a product of pure reason (as Kant believed). Another way to make this point is by pointing out that Kant’s insight that we do not have any privileged insight into the fundamental nature of reality also applies to Kant’s own system: we do not have any privileged insight into the fundamental nature of the mind

Thinking With Sellars and Beyond Sellars  259 (thought, reason), either. Kantian critical philosophy seemed in this way to fail to live up to its own standards of self-critical reflection. For only a philosophy aware of its own methods, presuppositions and limits could be a truly self-critical and self-reflective one. And this, pace Kant, seems to demand that philosophy be aware of the genesis, context and development of its own doctrines. Only if philosophy were historicized could the problem of transcendental illusion be fully eradicated. Correlatively, from a Hegelian point of view, it is only in this way that a philosophy could aspire to be truly presuppositionless: not by abstracting from the historical process that brought it into being and simply relying upon one’s individual reason (as Descartes, and even Kant, effectively believed), but by incorporating its own history within itself (see also Beiser 1993, 270–73). According to this line of thought, the most pervasive source of dogmatism in philosophy (its ‘transcendental illusion’) lies in forgetting the origin, context and development of its own ideas. And this is because, if a philosophy lacks this special kind of self-awareness (i.e., the awareness of the genesis, context and development of its own doctrines), it thereby lacks the resources to uncover—and hence overcome—its own deeply immersed unquestioned background assumptions, which organize and coordinate the worldview expressed in the philosophical system in question. These background assumptions seem virtually unchallengeable precisely because, being the organizing principles for a wide range of practices in which we think, act and deal with the world, they become so embedded in our manner of thinking and acting that they are never thematized as such, not even at the level of philosophical thinking (whose aim is precisely to make explicit the presuppositions of thought). In this way, those background assumptions seem to be characterized by a certain deeply ingrained naturalness, a presumption of uniqueness which goes without saying, thereby creating the impression that a certain philosophical worldview is ‘the only game in town,’ and the illusion that genuine alternatives to that worldview are not even thinkable (see also Taylor 1984). Hence, by not being able to even identify the unstated assumptions that stand in the background of its explicitly formulated theses and arguments, a philosophical system is thereby unable to identify its own limits and incapable of conceptualizing alternative possible philosophical worldviews. And the problem with a system’s being oblivious of its contingent status and limitations is not only that it lacks self-consciousness, but, more importantly, that it thereby lacks the conceptual resources that alone could make it possible for it to overcome or go beyond its limitations, toward a more comprehensive picture of the world and ourselves.2

3.  The Rejection of the Myth of the ‘Philosophical’ Given Now, I take it that the previous conception of philosophy as an essentially self-critical and self-reflective enterprise, and as something which is internally related to the history of philosophy would be congenial to Sellars.

260  Dionysis Christias Consider, for example what Sellars says in his well-known article “Philosophy and the Scientific image of Man” (1963): “It is the reflection on the place of philosophy itself in the scheme of things which is the distinctive trait of the philosopher as contrasted with the reflective specialist; and in the absence of this critical reflection on the philosophical enterprise, one is at best but a potential philosopher” (PSIM, 3).3 Moreover, direct corroboration for the claim that the self-critical character of philosophy is intimately bound up with taking a stand on issues in the history of philosophy can be found in his (not well-known) article “The Double-Knowledge Approach to the Mind-Body Problem” (1971) where Sellars explicitly says that a philosophy which is largely unaware of its historical antecedents cannot be truly critical, and, as a result, it is bound to make old mistakes anew (DKMB, 270). But, for the time being, I will explore a more roundabout way in which Sellars’s philosophy can be considered as ‘truly critical’ in the above sense, which, though not found in Sellars’s explicit metaphilosophical writings, is perhaps one of the most important and pervasive features of Sellarsian philosophy as a whole. More specifically, it can be suggested that Sellars’s rejection of (a certain form of) the Myth of the Given, among other things, goes hand in hand with a rejection of an ahistorical, time-transcendent conception of the structure of the mind, thought, reason or, for that matter, philosophical argumentation. And if this time-transcendent conception of mind, thought or reason is rejected in all its forms (including certain transcendental conceptions of mind or reason, which equate the notion of the ‘a priori’ with that of ‘categorial unrevisability’), the way is clear for recognizing that philosophy itself, to the extent to which it is an investigation through argumentation of the relation of the mind, thought or reason with the world, cannot be an ahistorical discipline, but has an essentially historical dimension. But how exactly does the rejection of the Myth of the Given lead to the corresponding rejection of an ahistorical conception of mind, thought or reason? We can see this if we reflect on the Sellarsian notion of the categorial Given. To fall prey to the myth of the categorial Given is to believe that “if a person is directly aware of an item which [in fact] has categorial status C, then the person is aware of it as having categorial status C” (FMPP I, §44), or to hold that “the categorial structure of the world—if it has a categorial structure—imposes itself on the mind as a seal imposes an image on a melted wax” (FMPP I, §45). Now, the interesting thing to note in these general definitions of the categorial Given is that, despite first appearances, the ‘Given’ element ‘we are aware of as having categorial status C’ or which ‘imposes its categorial structure on the mind ‘as a seal imposes an image on a melted wax’ need not be something radically external to the mind (or some kind of sensory or perceptual experience), but may well be thought or reason itself, transcendentally considered—at least to the extent to which we are thereby supposed to be able to reveal its necessary categorial structure by the sole use of a decidedly non-empirical and non-hypothetical

Thinking With Sellars and Beyond Sellars  261 transcendental investigation. But this means that even some versions of transcendental philosophy—namely, those according to which there is a sui generis level of transcendental description of the necessary categorial features of our mind/thought/reason which is impervious to reconceptualization on the basis of empirical-scientific theorizing or hypothetical philosophical concept-formation about the world and the mind—fall prey to the myth of the categorial Given. According to this Sellarsian line of thought, since we are bereft of intellectual intuition (‘the natural light of reason’), the categorial structure of the world or of the way in which we reason about the world do not imprint themselves on our minds as a seal imprints itself on wax. Hence, we do not have transparent epistemic access to (or knowledge of) the categorial structure of reality or of the mind (thought, reason). Extending this line of thought, Sellars in essence holds that knowledge of the world and knowledge of our categorially structured means of accessing the world (mind/ thought/reason) are on the same footing, advance hand in hand and are subject to historical development through a critically controlled process of mutual self-correction (see also Brassier 2016). Now, it can be suggested that an instance of the myth of the categorial Given is what might be called the myth of the ‘philosophical’ Given, that is, the view according to which philosophy, due to its essential reliance on ‘pure reason’ has transparent epistemic access to its essence or subjectmatter. That is, ‘pure reason,’ through its modes of ‘proper reasoning,’ is supposed to provide us with an a priori insight into the modal structure of philosophical discourse, thereby fixing the space of possibilities within which philosophical questions are meaningfully raised. Yet, one might wonder, what is precisely problematic about the philosophical Given? Why exactly is it a mythical? Very briefly and schematically, it can be suggested that the basic problem with this version of the Given (a kind of intellectual intuition on the essence of the content of fundamental philosophical problems) stems from its ‘practice-transcendent’ nature. More specifically, in order to function as a semantic or categorial Given, thought or reason must be characterized by epistemic and semantic independence. That is, their epistemic and semantic content must be determined independently of any (formal or material) inferential relations they may stand in with other such contents as the latter figure in our practices of ‘giving and asking for reasons’—in our case, the practice of ‘giving and asking for philosophical reasons’ throughout the history of philosophy. However, by effectively severing all internal relations between the semantic content of a philosophical thesis/problem/question and a public, historically evolving practice in which it finds overt expression (i.e., the diachronic practice of ‘giving and asking for philosophical reasons’) the proponent of the philosophical Given is ultimately unable to provide criteria which distinguish between cases in which the semantic content in question really is what we take it to be (in ‘intellectual intuition’) and cases in which the content in question is not

262  Dionysis Christias what we take it to be. The only such criterion could be provided by the supposed (semantic, epistemic or categorial) transparency of the content—in our case, of the logical space of possible philosophical questions—given in ‘intellectual intuition’, but in such a case, as Wittgenstein puts it “whatever is going to seem right to me is right. And that only means that here we cannot talk about right” (Wittgenstein 1958, §258; emphasis added). As we saw, the notion of philosophical Given, far from securing the objectivity of the semantic and epistemic content of philosophical propositions (theses/questions/problems) in fact ultimately undermines it. In order to secure the determinacy and objectivity of the practice of ‘giving and asking for philosophical reasons’ we have to abandon the myth of philosophical Given. And to reject this specific version of the myth amounts to realizing that the power of reason and the tools of argumentation that mankind now possesses, and philosophy makes explicit, are not given to it by ‘the natural light of reason,’ but have—and can—only been acquired through centuries of effort, through the cultural inheritance and occasional revision of the categorial structure of language (including philosophical language) from generation to generation. Recall, in this connection, that the rejection of the myth of the categorial Given goes hand in hand with the rejection of all forms of Platonism about meanings, the mind, thought or reason: the object (subject-matter) of reason is not a timeless form, complete in all its ‘meaning’ prior to our reflection upon it, and the reflection upon it is not an eternal contemplation, a passive intuition or timeless perception of these ‘forms.’ Instead, the object of reason (or thought) is not given to our thinking, but is ‘discovered’ only through its own self- positing.4 In good Hegelian fashion, we can say that we can discover the meaning of the object of reason or thought only through the very ‘act’ of positing it. The ‘act’ of positing in question should not be understood as something static or momentary but as a dynamic temporally extended process of extending and revising our concepts and conceptions of thought or reason in order to better accommodate incompatible commitments existing in former versions of theirs (see also Brandom 2009, 100–1). But, among other things, this means that the mind (thought, reason) discovers itself though its historical development and transformation. Correlatively, to reject the myth of the ‘philosophical’ Given amounts to realizing that philosophy, to the extent to which it is essentially an investigation of the relation of mind/thought/reason to the world, is internally connected to the history of philosophy, in that the history of philosophy is precisely the place where this self-discovery of mind/thought/reason is explicitly articulated. In other words, the history of philosophy is an expression or ‘crystallization,’ in a specific time in history, of mind’s/thought’s/reason’s conception of themselves through their historical development. Hence, bereft as we are of any transparent access (intellectual intuition) into the essence of mind/ thought/reason, it seems that the only way in which we can know our way around with respect to those things and the way in which they relate to the

Thinking With Sellars and Beyond Sellars  263 world, is to thematize, understand and critically engage with the history of philosophy, which is precisely what makes the conceptual development of these notions (and the content of our commitment in applying them) explicit. The modal structure of philosophical discourse—i.e., the logical space of possible philosophical questions and answers—is not somehow permanently fixed by the categorial structure of the world or of our reasoning capacities, patiently waiting for us to discover the precise way in which it is actualized in the real world. Rather, philosophical discourse discovers itself and its modal structure through its own explicit historical realization and articulation in the history of philosophy. And it does that through a dynamic, never-ending process of constructing, challenging, revising and reconstructing its conception of its modal structure, ultimately guided by reason’s own ‘desire’ to reflectively understand itself so as (per impossibile) to be actualized in reality.

4. The Manifest and the Scientific Image Understood as Critical Meta philosophical Concepts Linking Philosophy With its History I will now argue that, at a more specific level, Sellars’s metaphilosophical construal of what he famously calls the ‘manifest image’ and the ‘scientific image’ of ‘man-in-the-world’ is the key to understand the way in which philosophy proper is internally related to the history of philosophy. To see this, notice, for example, that Sellars explicitly says that the notions of the ‘manifest image’ and the ‘scientific image’ are idealizations which are designed to illuminate the inner dynamics of the development of philosophical ideas. Sellars, interestingly, also adds that, from a somewhat different point of view, the notions of the manifest and scientific image can be compared to the ideal types of Weber’s sociology (PSIM, 5). Both of these points suggest that the distinction between the manifest and the scientific image is metaphilosophical and is designed to illuminate (i.e., provide a hypothetical explanation5 of) the historical development of philosophical ideas. This means that those metaphilosophical conceptual tools, among other things, provide the framework in which we can make sense of the history of philosophy. Recall, for example, that Sellars, in his “Philosophy and the Scientific Image of Man,” argues that the major strands of what has been called the perennial tradition in philosophy, i.e., the Platonic and Aristotelian tradition broadly understood, (including Plato and Neo-Platonism, Aristotelian Scholasticism, major strands of medieval philosophy, contemporary neoThomism etc.), can be construed as attempts to understand the structure of the manifest image. Indeed, according to Sellars, the defining feature that unites the perennial tradition is that it construes the manifest image as ultimately real. And precisely because an essential feature of the manifest image is that the connection between conceptual/rational thinking and reality

264  Dionysis Christias cannot be understood in more elementary (nonconceptual, non-rational, prepersonal) terms, i.e., in terms which do not already, in some way or another, presuppose the conceptual framework of persons, the perennial tradition was limited to understanding the relation between the intelligible worldly order and individual minds which use the conceptual framework of persons in terms which bear an essential analogy to (the capacities and abilities of) persons (PSIM, 15–16). For example, the relation between the conceptual and the real order was understood in terms of an illumination of the mind by intelligible essences, forms, God, as the social-historical unfolding of Absolute Spirit, as an intentional correlation sustained by a ‘transcendental subject,’ etc. All these notions are understood by analogy to the conceptual framework of persons, as is shown, e.g., by the fact that, exactly like the latter, they are all conceptually and explanatorily irreducible to more elementary (nonconceptual, non-rational, pre-personal) processes. That is, at bottom, each of these philosophical conceptions characteristic of the perennial tradition is, in its own way, the expression—and absolutization—of the following fundamental ‘intuitive’ picture which animates the manifest image framework: we cannot build a person—its capacities, abilities and the whole (perceptual, practical) way in which he is related with the world—out of assemblages of parts, however complex the latter might be, if they do not already exhibit the right form of ‘unity’ that characterizes persons. Moreover, Sellars attempts to show that modern and contemporary forms of epistemological skepticism are the result of an immature attempt by early modern philosophy to replace the categorial ontology of the manifest image, and the ensuing categorial irreducibility of the framework of persons to more elementary terms, with that of the emerging scientific image, by attacking the manifest-image conception of nature. In this way, perceptual and affective qualities are ultimately construed as subjective states of the human mind projected onto an external world now understood in terms of relations (of mechanistic causation) and properties of imperceptible particles, and hence devoid of qualitative content. This pervasive philosophical tendency finds expression in the—otherwise very different—systems of Descartes, Spinoza, Locke and Hume. Sellars also sees the ‘transcendental turn’ of Kant’s philosophy as an attempt to salvage the sound core of the manifest image (the rationality, freedom and autonomy of persons) and insulate it from the onslaught of the scientific image, while simultaneously accounting for the objectivity of the latter, i.e., for the fact that the scientific image gives us an objective representation of nature. Kant’s attempt, however, remained unsuccessful and skeptical or ‘dualist’ in its form, mainly due to Kant’s overly restrictive picture of the scientific image. Kant conceived the scientific image along the lines of its early modern predecessors. According to this conception, nature is a system governed by deterministic mechanistic laws, bereft of any qualitative content. But, as a consequence, two irreconcilable dualisms emerge: that between qualitative content and structural/relational form (perceptual and

Thinking With Sellars and Beyond Sellars  265 affective qualities are understood as subjective states of the human mind projected onto an objective world governed by natural laws and bereft of any qualities whatsoever) and that between deterministic relations of mechanistic causation (sufficient for the objective description and explanation of nature) and human freedom, understood as the capacity of persons to make choices based on reasons, rather than mere external, a-rational causes (desires, impulses, social pressures, tradition). Moreover, although Sellars himself does not make this point explicitly, I take it that this Sellarsian metaphilosophical framework, which interprets the history of modern philosophy in terms of the conflict between the manifest and the scientific image, can also find application in the post-Kantian reaction against Kant’s radical separation of the realm of reason (or ‘realm of freedom’) from the realm of nature. Specifically, the reaction of philosophers such as Marx, Nietzsche and Freud (also known as ‘philosophers of suspicion’) against Kant’s hypostatization of reason and their view to the effect that what we take to be a purely autonomous, insulated realm of ‘pure reason,’ is at best a rationalization in which we unconsciously engage in order to infuse our life with meaning and purpose, and can be adequately and exhaustively explained in terms of (sub-personal) natural and (prepersonal) social/historical causes, can be interpreted as an—again premature— attempt of the nineteenth-century’s scientific image to attack the manifest image at its very conceptual core: the category of personhood, and the view that persons can have unproblematic categorial and epistemic access to their own beliefs, thoughts, desires and intentions (see also Brandom 2012). But this is not the sole metaphilosophical function of this fundamental Sellarsian distinction between those two basic ways of understanding the world and ourselves: As is well known, Sellars believes that this distinction also illuminates the content of many different and seemingly disparate contemporary philosophical views or systems. That is to say, according to Sellars, the very formation of the ideological space or ‘battlefield’ of contemporary philosophical ideas or frameworks can also be understood in terms of the manifest-scientific image distinction. Recall, e.g., Sellars’s conviction to the effect that such seemingly unrelated or disparate philosophical frameworks as the major schools of continental thought (presumably including phenomenology, existentialism and hermeneutics) and American and British philosophies of ‘common sense’ (including Strawson’s Oxford Aristotelianism and later Wittgenstein’s philosophy) are, at bottom, sophisticated philosophical systematizations of the manifest image categorial framework of man-in-the-world, which take the manifest-image conception of the world as ultimately real. For recall that, in Sellars eyes at least as I understand him, all those seemingly unrelated continental and ‘analytic’ philosophical systems of thought are united in believing that the categorial framework of persons and its relation to the world cannot be understood in terms that do not already presuppose the basic categoral concepts and distinctions of the framework. And in this way, I take it that, by Sellarsian lights, all these

266  Dionysis Christias schools of thought effectively absolutize what is perhaps the most basic (and in a certain sense sound) ‘intuition’ which stands behind the basic categorial distinction of manifest image: that we cannot build a person—its capacities, abilities and the whole way in which he is related with the world—out of assemblages of parts, however complex the latter might be, if they do not already exhibit the right form of ‘unity’ that characterizes persons. In this sense, continental phenomenology and existential hermeneutics as well as Strawsonian Aristotelianism and the ‘practice-based’ philosophy of the later Wittgenstein, all prove, in their own distinctive way, to be heirs of philosophia perennis. Furthermore, it can be suggested (though Sellars himself is less clear on this point) that the manifest- scientific image distinction also serves as a metaphilosophical tool which throws light on the very emergence of philosophical problems in general, and, more radically, to the emergence of philosophy itself as a discipline. Now, as was shown earlier, the relevance of the manifest-scientific image distinction for explaining the emergence of philosophical problems in modernity (from Descartes onwards) is, I think, obvious. For Sellars, at bottom, contemporary philosophical problems in the fields of metaphysics, epistemology and ethics—and the characteristically modern skeptical anxieties aroused in all these fields—spring from the conflict between our manifest-image conception of ourselves-in-the-world and the radically disenchanted image of ourselves-in-the-world offered by the scientific image. In effect, the very distinction between the manifest and the scientific image is, at bottom, the theoretical expression—internal to the discipline of philosophy—of the cultural process of what Weber called the disenchantment of the world (and, eventually, of ourselves), i.e., of the gradual overcoming of ‘magical’ or ‘anthropomorphic’ ways of understanding the world and ourselves, through the relentless depersonalization of explanation.6 It is exactly this basic feature of modernity, i.e., the process of the disenchantment of nature that stands behind the characteristically modern radical disconnection of the Sellarsian ‘logical space of reasons,’ which represents our manifest-image conception of ourselves as possessors of distinctively human characteristics, such as freedom, rationality, deliberation, from the ‘logical space of causes and effects,’ which articulates our modern, scientific-image conception of the world and ourselves as complex physical systems characterized by all kinds of habits (related to perception, inference and action) yet devoid of normativity (rationality, freedom, deliberation).7 In this way, the connection between the manifest-scientific image distinction and the emergence of distinctively modern philosophical problems does seem plausible enough. Distinctively modern philosophical problems emerge as a result of the separation and conflict between the manifest and the scientific image of humanity-in-the-world, and much of modern and contemporary philosophy can be understood as an attempt to bridge the gap in our self-understanding opened by the separation and conflicting nature of those two basic categorial images of ourselves-in-the-world.

Thinking With Sellars and Beyond Sellars  267 But, one might wonder, what is the explanatory value of the manifestscientific image distinction in the case of the emergence of problems in premodern (e.g., ancient or medieval) philosophy, or, more radically, in the emergence of philosophy itself, as a discipline, in antiquity? Although Sellars himself did not delve into this issue in detail, I suggest that a case can be made for thinking that philosophical discourse was itself born as a result of the emergence of the manifest image proper out of what Sellars calls the ‘original’ image—where all kinds of objects were understood on the model of persons and their capacities, abilities and impulses. Recall that, for Sellars, the manifest image is itself in its way a scientific image (it is selfcritical and uses explanatory methods or correlational induction to refine and extend the framework) and emerges out of a gradual depersonalization of objects other than persons. Natural objects are thereby eventually deprived of the full range of capacities and abilities of persons. The former, unlike persons, do not now count as self-determining beings (i.e., beings capable of doing things with an end in view by exercising their rational (deliberative) powers).8 Specifically, in the early stages of the development of the manifest image, nature is understood as including ‘truncated persons’ which aremere creatures of habit, acting out routines, broken by impulses, in a life which never rises above what ours is like in our most unreflective moments. Now, on the basis of this model, we can begin to understand the gradual transition from the original image to the manifest image. Instead of mythically construing natural phenomena on the model of persons and its capacities, abilities and impulses, humanity, precisely by depersonalizing nature, gradually distinguishes itself from it, and begins to construct the image in terms of which man first becomes aware of himself as man-in- theworld, where nature is considered now as the realm of non-sentient (nonliving) or sentient yet non-rational existence. I suggest that this transition was the result of the rise of a very early and primitive form of the scientific image,9 and, in particular, (1) of the Jonesean ‘scientific revolution’ of the postulation of thoughts, sensations, intentions, desires and emotions conceived as inner mental states (as opposed to semantically significant states existing ‘out there’ in the external public world, occasionally ‘taking control’ of us), and (2) of ancient Greek astronomy and mathematics. Here, we will focus on this latter development of the ‘ancient’ scientific image.10 Astronomy with its mathematical discourse of the motion of the planets ceases to explain a phenomenon with the exemplary narratives of myth (a classic mode of explanation of natural phenomena and human behavior in the original image) and utilizes essentially abstract, idealized and non-perceptible modes of explanation—thereby making possible a substantial distinction between what is ‘really real’ and what merely appears to be so.11 In this way—and also in the wake of the Jonesean ‘scientific revolution’—a gradual and unthought scission is produced between the new science of ‘being’ (based on the mathematical description of the motions of celestial bodies), which is now considered to be what is ‘really real,’ and man’s understanding of himself as a source of values and norms within a

268  Dionysis Christias perceptible, concrete and non-exact ‘manifest’ reality, which is now relegated to the realm of mere ‘appearances.’ This scission (between value and being) gradually makes its presence felt, among other things, because the discourse of values, unlike the new science of imperceptible ‘being’ described by astronomy and mathematics, preserves the structure of narrative and myth, as if by inertia,12 thereby gradually severing a hitherto taken for granted ‘mythical’ unity between values and being (see also Meillassoux 2011, 197–9). Thus, it might be claimed that the very manner in which the manifest image comes into existence produces a gap in man’s understanding of nature and his place in it (registered by the sophists), which, in turn, creates the need for the emergence of a form of thinking whose constitutive aim is precisely to erase the gap in question and restore the disturbed coherence in our understanding of the world and ourselves: and philosophy as a discipline, from Plato onwards, emerges precisely as an attempt to do just that, albeit in a distinctive (non-religious, non-scientific, non-ordinary) way. Notice that although there are important differences between the ‘ancient’ and the ‘modern’ version of the scientific image, each in their own distinctive way, challenged the hitherto natural way of understanding nature and ourselves in it, by creating a gap between value and being—which philosophy, since then, aims to bridge. In Sellarsian terms, we could perhaps understand this whole historical process, where the manifest image challenges the original image and is, in turn, challenged by the (‘modern’) scientific image, as characterized by two waves of disenchantment (of nature and ourselves), and corresponding ‘scientific revolutions’: the first wave of disenchantment took place already at the dawn of history and antiquity—through the ‘scientific revolutions’ of the postulation of thoughts, intentions, emotions, desires as inner mental states and the invention of mathematics and astronomy—and gradually transformed the original image into the manifest image (albeit without challenging its basic category, i.e., that of persons). The second ‘Weberian’ disenchantment occurred in early modern times— through the modern Copernican revolution all the way to Galilean and Newtonian physics—and challenged the categorial authority and adequacy of the manifest image itself (including the fundamental category of personhood) from the standpoint of an image emerging as the ‘determinate negation’ of the latter, namely the ‘modern’ scientific image of man-in-the-world. The first wave of disenchantment brought philosophy proper into being while the second gave it its distinctively modern form.

5. A Sellars-Inspired Account of the Relation Between Philosophy and Its Internal and External History We suggested above that the Sellarsian concepts of the manifest and the scientific image are in effect metaphilosophical tools which can illuminate issues about the nature of philosophical theses and problems, at least from

Thinking With Sellars and Beyond Sellars  269 three different explanatory angles. They can illuminate metaphilosophical issues regarding (1) the content of different contemporary philosophical views and systems, (2) the development of philosophical ideas in the course of the history of philosophy, and (3) the origin or emergence of philosophical problems as such, and of philosophy itself as a discipline. And it can be argued that the very fact that Sellars attempts to explicate those three different yet intimately related metaphilosophical issues with the use of the same metaphilosophical conceptual tools, suggests that those issues are intimately related in a way that makes them distinguishable yet inseparable from one another. (1) and (2) show that philosophy is intimately related to its ‘internal’ history (the history of philosophy), while (3) shows that philosophy is also related (in another sense) with its ‘external’ history, i.e., with nonphilosophical events such as scientific advances or revolutions.13 In section 3 we saw the way in which the rejection of the Myth of the ‘Philosophical’ Given shows us, among other things, that philosophy is internally related to its history, and that, through this relation, mind (thought, reason) discovers itself though its historical development and transformation. We can utilize this account to put more flesh on the bare bones of what was stated above in points (1) and (2), and vice versa: Specifically, if we combine the Sellarsian insights discussed in section 3 (the rejection of the philosophical Given) and section 4 (the notions of the manifest and the scientific image construed as metaphilosophical tools that can account for the content of contemporary philosophical problems as well as the inner dynamics of the development of philosophical ideas through the history of philosophy), we can perhaps venture to suggest the following: for Sellars, the ‘space of philosophical discourse’—i.e., the logical space of possible philosophical questions and answers—is first constituted by a rupture of the manifest image from the original image of man-in-the-world, and is then extended, challenged, revised and reconstructed in such a way that enables each ‘successor’ philosophical theory to construct a conceptual space from which it can explain the explanatory failures and critically accommodate the partial explanatory successes of its ‘predecessors’ in the light of the ‘successor’ theory’s conception about the nature of the (epistemological, ontological, semantic, practical) relation between the manifest and the scientific image categorial frameworks. In this way—i.e., as a ‘living record’ of explanatory anomalies and partial explanatory successes on the issue of the relation between the manifest and the scientific image of ourselves-in-theworld that itself demands explanation, the history of (modern) philosophy becomes directly relevant to the formulation of the space of possibilities within which contemporary philosophical questions and answers get their very meaning. And it can be argued that this is because this ‘space of possibilities’ derives its content partly from its being ‘designed’ to be a proper response to the challenges and dead-ends of older historically significant rival philosophical systems. That is, the very formulation of philosophical problems and attempted solutions within a given philosophical system,

270  Dionysis Christias among other things, is designed—whether an individual who uses the system realizes this or not—to explain why rival ‘predecessor’ philosophical theories faced insuperable explanatory anomalies and to account for (what from the standpoint of the given philosophical system in question are) their partial explanatory successes.14 Moreover, in regard to point (3), it must be pointed out that the fact that the manifest-scientific image distinction provides a conceptual framework in the context of which we can understand the origin or emergence of philosophical problems and of the field of philosophy as such, by intimately relating the internal philosophical problematic with its external history (non-philosophical events such as scientific, political, social, technological advances or revolutions), shows that philosophy is, at best, only relatively autonomous as a discipline. Indeed, from this point of view, philosophy can be understood as a practice whose raison d’être is to respond to major ideological and conceptual upheavals that profoundly affect how we understand the world and our place in it (i.e., radically change the way in which we conceive ‘things’ and how they ‘hang together’ in the broadest sense of those terms). Philosophy is precisely the conceptually organized response which, in the course of its historical development, tries to make sense of the changes and ruptures in our understanding of the world and ourselves produced by historically accumulated empirical knowledge about the world and our natural, social and individual existence (see also Baltas 2012, 65–6).

6. Conclusion: Applying Sellarsian Metaphilosophical Tools to Sellarsian Philosophy Itself Recall that, as was mentioned in section 2, according to a Kantian-Hegelian point of view, philosophy worthy of the name would have to be selfconscious, self-reflective and self-critical. This said, I think that the bottom line of the argument of section 4 is that the Sellarsian notions of the manifest and the scientific image, in their metaphilosophical use, can be understood as precisely the concepts through which a philosophical system can become self-aware, self-reflective and (hence) self-critical—that is, aware of its own methods, presuppositions and limits. But in what precise way can the metaphilosophical concepts of the manifest and the scientific image serve as tools for fulfilling the ultimate purpose of a properly self-reflective philosophy, namely to make it aware of its own presuppositions and limits, thereby rendering it capable of ‘transcending’ them from ‘within,’ as it were? Can those Sellarsian metaphilosophical tools be applied to Sellarsian philosophy itself, thereby providing it with a means to ‘immanently transcend’ itself? I suggest that indeed there is— irrespectively of whether Sellars himself was actually aware of this possibility. Specifically, it can be proposed that the scientific image of man-in-theworld, applied to anthropology, psychology and the social sciences (including economics, social psychology, social anthropology and social-cultural

Thinking With Sellars and Beyond Sellars  271 history), could throw light on the specifics of the very complex way in which ‘external’ history, i.e., non-philosophical events such as scientific or sociopolitical advances and revolutions, mediated and theoretically reconstructed throughout ‘the history of philosophy,’ are related with the very content of contemporary philosophical problems. For example, in this way we could perhaps throw light on the complex interrelation between the ultimate practical sources of the cultural process of Weberian disenchantment (a candidate here, for example, could be a scientific-image successor concept of what Horkhemer and Adorno (2002) call ‘the domination of physical and human nature’), the way in which the major practical consequences of this cultural process are theoretically reconstructed and comprehended throughout the history of philosophy (e.g., as indicating an essential bifurcation between the ‘space of reasons’ and the ‘space of causes’), and the way in which this latter theoretical understanding finds expression in the formulation of purely philosophical problems (and attempted answers thereof) in contemporary philosophy (e.g., the problems of accounting for the place of modality, morality, meaning and intentionality within a naturalistically described world). Note here that the scientific image, e.g., as applied to anthropology, economics, social psychology and history, goes beyond a manifest-image rendering of these disciplines in that it postulates unobservable and impersonal or pre-personal states, ‘forces’ and mechanisms to account for observable anomalies within the ‘personalistic’ manifest-image account of the social sciences. Might we not then suggest that the scientific image, as applied in anthropology, economics, social psychology and history can, in principle, be used as a meta philosophical tool for retrospectively throwing light on the way in which non-philosophical events such as scientific, economic and political advances or revolutions in the twentieth century (i.e., in Sellars’s time), mediated and theoretically reconstructed by twentieth-century’s ‘history of philosophy’, can illuminate those crucial practical, cultural and theoretical presuppositions which formed the unquestioned background for the formulation of Sellarsian philosophy itself? If this were possible—and I do not see why it should not—we would be in a position to identify the social-historical (and ultimate conceptual) limits of Sellars’s own philosophical system, while at the same time transcending them, precisely through the very recognition of these limits. This could be done, in turn, by gradually incorporating the scientific-image (postulational, impersonal and ‘factual’) knowledge about the decisive external social-historical factors that influence the internal history of philosophy, in our evolving manifest-image (nonpostulational, first-personal and normative) conception of ourselves-in-theworld. Importantly, this ‘transcendence-through-immanent-recognition’ of the social-historical limits of Sellars’s system (perhaps expressed in a sedimented way into his very concepts) would be (retrospectively) possible by means internal to Sellarsian philosophy itself. And this fact would reveal a dynamic self-critical dimension lying at the heart of the Sellarsian system: the

272  Dionysis Christias ability of this philosophy to revise itself and expand its horizons through the use of its own meta philosophical principles.

Notes 1 There are some references to Kant’s and even Hegel’s metaphilosophical views in Sellars, but, so far as I know, Sellars does not explicitly develop his views on the relations of philosophy with its history in connection with Kant or Hegel. What I shall claim in what follows is only that Kant’s critical turn and Hegel’s self-reflective conception of philosophy represent a precursor of Sellars’s own metaphilosophical views on the issue of the relation of philosophy with its history. These latter views, with some exceptions (see section 3, PSIM, 3; IRH, 103; DKMB, 270), are for the most part implicitly conveyed or expressed rather than explicitly stated in his works. One of the main purposes of this article is precisely to make Sellars’s views on these matters explicit. 2 The purpose of this section is not to offer an argument in favor of the position that sees philosophy as an inherently self-reflective enterprise that is internally connected with its history, but rather just to provide a rough sketch of the view in question, as a precursor to Sellars’s own view. Explicit arguments for the view that this is a legitimate precursor of Sellars’s metaphilosophical position, as well as for the plausibility of the critical/self-reflective conception of philosophy itself, will be offered in sections 3, 4 and 5. Moreover, in section 6, we shall attempt to apply this Sellars-inspired self-critical, self-reflective view of philosophy to Sellarsian (‘first-order’) philosophy itself. 3 See also IRH. As he puts it there, commending (favorably) on Everett Hall’s metaphilosophy: “[Everett Hall’s philosophy] is in a most important sense selfconscious or self-referential in that it includes as an essential component a theory of the philosophical enterprise, a theory which faces up to the ultimate challenge which any systematic philosophy must face: What is the status of your philosophical claims, and what are the criteria by which you distinguish them as true from the false and unacceptable claims made by rival philosophies?” (IRH, 103). 4 In Sellars, this view is paradigmatically expressed in his ‘myth of Jones’ (EPM), where mythical genius Jones and his peers discover their own thoughts, desires, intentions as inner objects to which they have privileged first-person access not by just noticing or paying attention to them (as if they were already categorially in place), but by positing them as in a manner akin to theoretical entities to explain anomalies in the verbal behavior of himself and his peers in his ‘Rylean’ community. Ryleans may well think, desire or intend to do thus-and-so but they do not thereby (that is, independently of theoretically positing the categories of inner thinking, desiring or intending) classify or categorize these semantically significant states as inner mental states to which they have first-personal privileged access. To believe so is to fall prey to the Myth of the Given. 5 It must be emphasized that the metaphilosophical concepts of the ‘manifest image’ and the ‘scientific image’ are hypothetical concepts designed to provide an illuminating explanation of the historical development of philosophical ideas. They are not revealed through ‘intellectual intuition,’ conceptual analysis or ‘phenomenological’ reflection into the essence of philosophy in its historical unfolding. 6 Brandom gives the following concise description of Weberian disenchantment: “The meanings and values that had previously been discerned in things are stripped off along with the supernatural and are understood as projections of human interests, concerns, and activities onto an essentially indifferent and insignificant matter” (Brandom 1994, 48). An interesting thing to note in this connection is that in this relentless disenchanting process, our human interests,

Thinking With Sellars and Beyond Sellars  273 concerns and activities themselves are gradually purged of their normative significance and are eventually understood in non-normative terms, as adaptive processes ultimately serving navigational ‘purposes’ of certain kinds of complex physical systems (human organisms). 7 Sellars was fully aware of this role of the disenchantment of nature in the occurrence of specifically modern philosophical problems concerning the place of a value-laden human experience within a norm-free nature (although he did not use the specifically Weberian term, but rather spoke of the ‘depersonalization’ of explanation). See, e.g., PSIM, 9–14, 25–32. 8 Thus, in the original image to say of the wind that it blew down one’s house would imply that the wind eitherdecided to do so with an end in view, and might, perhaps, have been persuaded not to do it, orthat it acted thoughtlessly (either from habit or impulse), or, perhaps, inadvertently, in which case other appropriate action on one’s part might have awakened it to the enormity of what it was about to do. In the early stages of the development of the manifest image, the wind was no longer conceived as acting deliberately, with an end in view; but rather from habit or impulse. Nature became the locus of ‘truncated persons’; that which things could be expected to do, became nature’s habits; that which exhibits no order, its impulses. Inanimate things no longer ‘did’ things in the sense in which persons do them—not, however, because a new category of impersonal things and impersonal processes has been achieved, but because the category of person was now applied to these things in a pruned or truncated form (PSIM, 13). 9 Recall that the manifest image is itself in a sense a scientific image, as opposed to the original image which is completely unscientific. The main difference between the pre-modern and the modern version of the scientific image lies in the categories they postulate as basic: the basic category of the pre-modern scientific image remains that of persons, applied in a truncated form in the case of objects other than full-blown persons, whereas the basic category of the modern version of the scientific image is that of an impersonal thing or, ultimately, of an impersonal process. 10 The ideas that are presented in what follows about the basic structure of the ‘ancient’ scientific image are not thematized by Sellars, nor do they seem to be implicitly endorsed by him. I present them as a friendly amendment, or reasonable extension of his metaphilosophical views in this area. 11 Note that one of the most distinctive features of mythical description and explanation, characteristic of the ‘original’ image, is that no substantial distinction is made between appearance and reality. As Cassirer (whom Sellars had read) puts it in The Philosophy of Symbolic Forms (1929): “In myth there is no thingsubstance lying at the basis of the changing and fleeting appearances . . . as a stable and enduring something. Mythical consciousness does not infer from the appearance to the essence, but it possesses the essence, it has the essence in the appearance. . . . The demon of the rain itself is that which is living in every water droplet and which is palpably and bodily there. . . . The essence . . . manifests itself in the appearance as a whole, as an unbroken and indestructible unity” (Cassirer [1929] 1957, 67–8). 12 Thus, it was still possible, e.g., for a cultivated Greek to explain courage by narrating the feats of Achilles even after ceasing to believe in such mythical narratives as regards the movement of celestial bodies or the nature of becoming. In this way, myth continued to serve as the legitimating source of values, while at the same time being itself gradually delegitimized through its incapacity to explain the realm of ‘being’ in general. 13 In Realizing Reason (2014) Macbeth also uses modified versions of the ‘manifest image’ and the ‘scientific image’ as metaphilosophical conceptual tools for showing the intimate connection between philosophy and its (internal and external)

274  Dionysis Christias history (see also Macbeth 2018). Yet, Macbeth differs from Sellars in the way she conceives these terms for her purposes, and this is especially clear in her understanding of the scientific image. Specifically, Macbeth argues that the categorial structure of the ‘mature’ scientific image is the outcome of two scientific revolutions (in mathematics and physics), both of which took place in modernity. The first ‘early modern’ one took place in the seventeenth century, with the central transformative figures being Descartes (who abandoned ancient diagrammatic reasoning in mathematics, replacing it with his own algebraic method of problem solving) and Newton (who rejected the traditional observationalcorrelational explanations of physical phenomena, offering in their place explanations based on the postulation of unobservable entities). The second scientific revolution, which is not generally recognized as such (not even by Sellars) took place in the nineteenth century, with the central figures here being Riemann (who abandoned constructive algebraic problem solving in mathematics, replacing it with deductive, logical reasoning directly from the content of concepts) and Einstein (who provided explanations of physical phenomena directly in terms of mathematical structures, without postulating unobservable physical entities or processes) (see also Macbeth, this volume). According to Macbeth, this second scientific revolution in mathematics and physics calls for a corresponding revolution in philosophy, which is only now beginning to take shape. Unfortunately, we do not have the space here for comparing and contrasting (our version of) Sellars’s and Macbeth’s conception of the scientific image and the corresponding appropriate philosophical response to it. This would at the very least demand the development of a properly Sellarsian philosophy of mathematics, something which Sellars only hints at (but see Sicha 1974, 1978). At any rate, notwithstanding their differences, it is important to stress that both Sellars (as I understand him) and Macbeth agree in that the history of philosophy is essentially related to the very content of contemporary philosophical problems. The latter cannot be properly solved absent an account of their emergence, historical development and transformation. 4 Consider, for example, the way in which Kant formulates his own system of phil1 osophical questions and answers in an attempt to solve the problems he inherits from his philosophical ‘predecessors,’ empiricists (the justification of induction, accounting for the motivational power of moral values) and rationalists (the coherence of their conception of ‘real’—i.e., metaphysical—necessity). To this end, Kant replaces the traditional ontological distinction between physical and mental entities—which, operating as an unquestioned background assumption, sets the philosophical agenda for both empiricists and rationalists—with the deontological distinction between the realm of nature (of causal relations) and the realm of freedom (of normative relations) (Brandom 2009, 114–16). Note that with this ‘revolutionary’ theoretical move, Kant attempts to offer a successful response to deep explanatory anomalies characteristic of both rationalism, whose basic problem is its ‘dogmatic,’ i.e., unjustified, character, and empiricism, whose problem is that it leads to skepticism. And in this way, Kant—and, I would argue, virtually every major figure in philosophy—does not just offer a different answer to an already understood philosophical question (posed by the rationalist-empiricist tradition) but changes the form of the question itself: ontological questions about how mental entities interact with extra-mental reality are transformed into deontological questions about the binding character (validity, objectivity) of concepts. And recall that this transformation is the only way to explain the explanatory dead-ends to which both rationalists and empiricists were inevitably led. This is because the ultimate cause of the dead-end in question, namely the problematic character of the very shape of philosophical discourse that both rationalists and empiricist took for granted, could not even be thematized by rationalists and empiricists alike. Now, I take it that this account

Thinking With Sellars and Beyond Sellars  275 of the relevance of the history of philosophy and the revolutionary transformation of its ‘space of possibilities’ to the content of contemporary philosophical problems would be congenial to Sellars. For example, Sellars hints in this direction when he says that “the historical development of philosophy is more truly conceived as the periodic formulation of new questions, than as a series of attempted answers to an enduring body of problems. . . . An essentially similar point of view, which, however, cuts a little deeper . . . suggests that the evolution of philosophical thought is accurately conceived neither as a series of different answers to the same questions, nor as a series of different sets of questions, but rather as the series of approximations by which philosophers move toward the discovery of the very questions they have been trying to answer all the time” (RNWW, §1, 2).

References Baltas, Aristides. 2012. Peeling Potatoes or Grinding Lenses: Spinoza and Young Wittgenstein Converse on Immanence and Its Logic. Pittsburgh: University of Pittsburgh Press. Beiser, Frederick. 1993. ‘Hegel’s Historicism’. In The Cambridge Companion to Hegel. Edited by Frederick C. Beiser, 270–300. Cambridge: Cambridge University Press. Brandom, Robert. 1994. Making It Explicit: Reasoning, Representing and Discursive Commitment. Cambridge,MA: Harvard University Press. Brandom, Robert. 2009. Reason in Philosophy: Animating Ideas. Cambridge, MA: Harvard University Press. Brandom, Robert. 2012. ‘Reason, Genealogy and the Hermeneutics of Magnanimity’. www.pitt.edu/~brandom/currentwork.html. Brassier, Ray. 2016. ‘Transcendental Logic and True Representings’. Glass Bead Journal, Site 0: ‘Castalia: The Game of Ends and Means’. www.glass-bead.org/ article/transcendental-logic-and-true-representings/?lang=enview. Cassirer, Ernst. [1929] 1957. Philosophy of Symbolic Forms. Volume Three: The Phenomenology of Knowledge. New Haven: Yale University Press. Horkhemer, Max and Adorno, Theodor. 2002. The Dialectic of Enlightenment. Stanford: Stanford University Press. Macbeth, Danielle. 2014. Realizing Reason: A Narrative of Truth and Knowing. New York: Oxford University Press. Macbeth, Danielle. 2018. ‘Sellars and Frege on Concepts and Laws’. In Sellars and the History of Modern Philosophy. Edited by Antonio Nunziante and Luca Corti. London: Routledge. Meillassoux, Quentin. 2011. ‘Excerpts from L’Inexistence divine’. In Quentin Meillassoux: Philosophy in the Making. Edited by Graham Harman, 175–238. Edinburgh: Edinburgh University Press. Sellars, Wilfrid (RNWW). 1948. ‘Realism and the New Way of Words’. Philosophy and Phenomenological Research, 8, 4: 601–34. Sellars, Wilfrid (PSIM). 1963. ‘Philosophy and the Scientific Image of Man’. In Science, Perception and Reality, 1–40. London: Routledge. Sellars, Wilfrid (IRH). 1966. ‘The Intentional Realism of Everett Hall’. Southern Journal of Philosophy, 4, 3: 103–15. Sellars, Wilfrid (SM). 1967. Science and Metaphysics: Variations on Kantian Themes. Atascadero, CA: Ridgeview Publishing Company.

276  Dionysis Christias Sellars, Wilfrid (DKMB). 1971. ‘The Double Knowledge Approach to the Mind— Body Problem’. New Scholasticism, 45, 2: 269–89. Sellars, Wilfrid (EPM). 1997. Empiricism and the Philosophy of Mind. Cambridge, MA: Harvard University Press. Sellars, Wilfrid (FMPP). 1981. ‘Foundations for a Metaphysics of Pure Process’. The Monist, 64, 1: 3–90. Sicha, Jeffrey. 1974. A Metaphysics of Elementary Mathematics. Amherst, MA: University of Massachusetts Press. Sicha, Jeffery. 1978. ‘Logic: The Fundamentals of a Sellarsian Theory’. In The Philosophy of Wilfrid Sellars: Queries and Extensions. Edited by Joseph C. Pitt, 257– 86. Dordrecht: D. Reidel Publishing Company. Taylor, Charles. 1984. ‘Philosophy and Its History’. In Philosophy in History: Essays in the Historiography of Philosophy. Edited by Richard Rorty, J.B. Schneewind and Quentin Skinner, 17–30. Cambridge: Cambridge University Press. Wittgenstein, Ludwig. 1958. Philosophical Investigations. Oxford: Wiley-Blackwell.

Index

aboutness 84, 199, 202 ‘above the line’/‘below the line’ 126, 127, 133, 135 abstraction 16, 25, 31, 155; abstractionist 26, 30 Adorno, Theodor 271, 275 agency, human 135, 181 – 3, 185 – 6, 194; account of 181 – 2, 186, 194; behavioristic explanation of 185; neurophysiological pictures of 189; normative depiction of 182; understanding of 183, 185 algebraic problem solving 149, 150, 274 Allison, Henry 56, 83, 84, 94 analogy 2, 28, 31, 47 – 8, 63, 70, 77, 85, 90, 163, 212, 220, 227, 245, 264; analogical concepts 88, 164; analogical thinking 89; analogous to/ conceived by analogy 81, 85, 89, 90, 106, 164, 166, 200, 241 analysis 7 – 8, 18, 48, 53, 55 – 6, 80, 83, 93, 162, 171, 179 – 81, 185 – 9, 193 – 7, 201, 208, 210, 214, 221, 230, 240, 243, 249, 252 – 3, 272; behavioral 179 – 80; causal 185, 188, 189; conceptual 179, 187, 272; functional 185, 188 animal learning 180, 182 anthropology 158, 171, 173, 270 – 1 anti-Cartesianism 27, 219; in philosophy of mind 219 anti-reductionism 219 appearances 48, 56, 79, 80 – 4, 86, 89, 90 – 1, 93, 141, 209, 211, 260, 268, 273 Aristotle 31, 74, 140, 151, 165, 257 assertibility 162, 171, 224 Augustinian picture 219 – 21, 227

Bedeutung/designation 138, 143, 146, 149 – 50, 152, 201 – 3, 210,  231 behavior: external 176, 186 – 7; free 194; habituated 194; linguistic 187; non-verbal 187, 193; observable 179, 184; pattern governed 187, 189; purposive 161, 184, 196; verbal 27, 178 – 81, 185 – 8, 191, 193, 195 – 6,  272 behaviorism 8, 34, 115, 174, 176, 178 – 84, 186 – 9, 191 – 6, 218, 222, 232; analytic 183; history of 179; logical 182, 191; methodological 191; philosophical 182, 194; radical 194 – 5; scientific 183; substantive 194; verbal behaviorism 188 – 9 Bennett, Jonathan f. 10, 94 Bergmann, Gustav 36, 57, 194 – 5, 229 – 31,  252 Berkeley, George 10, 34 – 6, 55, 58, 61, 81, 113, 155, 257 Black, Max 155, 229 – 30 Brandom, Robert 80, 103, 120, 135, 137, 212 Broad, Charlie Dunbar 38, 44, 53, 252 Carnap, Rudolf 8, 120, 128, 174 – 5, 185, 197 – 9, 201 – 15, 217 – 19, 221, 227, 229 – 31, 249, 252 Cassirer Ernst 35, 96, 115, 273, 275 category/ies 11, 16, 56, 87, 95, 100, 133, 159 – 61, 170, 186, 201, 211, 214 – 15, 252 – 3, 265, 268, 272 – 3; behavioristic 186; Kant’s conception of 56, 87, 100, 211; of personhood 265, 268, 273 causality 43, 254 causation 65, 67, 69, 72, 78, 245, 254, 264 – 5; mechanistic 264 – 5

278 Index causes: natural 71; physical 81, 85, 180; social/historical 265 Cavell, Stanley 218, 250 Chisholm, Roderick 250 Coffa, j. Alberto 212 – 13, 229 – 30 cognition 74, 87, 91, 120, 137 – 8, 146, 151, 157, 165, 167, 169 – 70, 172 – 4,  198 cognitive access 152 – 3 cognitive science 164 – 5, 169, 171, 174 cognitive significance 151 – 2 coherentism 46, 120, 125, 139 community 89, 103 – 5, 124, 160, 162, 173, 181, 183, 185, 187, 191, 207, 272; linguistic/language 104, 124, 185; membership 187; social 103 concepts: behavioral/behavioristic: 181 – 3, 191, 194; cognitive 179 – 81, 186; complete 37, 44, 53; formation of 123, 143, 182, 190, 210; Fregean 143, 152 – 3, 138; functional theory of 2, 100, 110; hypothetical 272; individual 37, 51; intensional 187; Kantian 151; knowledge of 148; mental/mentalistic 28, 178 – 9, 181 – 4; physical 33, 182, 205, 207; postulational 184; psychological 184, 193, 202 – 3; publicly accessible 186; semantic 184, 187; theoretical 27, 185 conceptual analysis 179, 187, 272 conceptual capacities 25, 87, 107, 127, 134, 179, 187, 191, 272 conceptual content 80, 85, 87, 100, 103 – 4, 107, 111, 119, 122, 125 – 6, 132, 136 – 7; non-conceptual content 84 – 5,  87 conceptual episodes 100, 106 – 7, 189 conceptual meaning 101, 144 – 7, 149, 153, 160, 164, 172 – 3 conceptual representation 17, 84, 87 – 8, 126, 129 – 30, 154, 204 conceptual synthesis 86 consciousness/conscientia 15, 17, 19 – 26, 30 – 3, 57, 79, 92 – 4, 112, 163, 169, 176, 230, 233, 236, 240, 245 – 6, 253, 273; self-consciousness 4, 95, 111, 114, 259 correctness 104 – 5, 166 – 7,  226 critical: philosophy 56, 258 – 9; realism 193, 234 – 7, 239, 242, 244 – 5, 247 – 54 Darwinism 158, 161, 165, 175 Davidson, Donald 15, 85, 217 – 18, 230

deductive-nomological account 61 definitions 20 – 4, 36, 75, 100 – 1, 143 – 4, 153, 180, 224, 229, 260; explicit 143 – 4, 153; of mathematical concepts 143; molar 180; molecular 180; operational 180 Descartes, René 2 – 3, 5 – 10, 12, 15 – 37, 40, 48, 55 – 8, 113, 150 – 1, 154, 236, 257, 259, 264, 266, 274 descriptions 50 – 1, 57, 74, 78, 111, 124, 146, 179, 204, 209, 211, 214, 221, 226, 246, 261, 265, 267, 272 – 3; definite 50, 146; mythical 273; objective 265 Devries, Willem A. 9 – 10, 55, 81, 94, 100, 107, 110, 112, 135 – 6, 176, 178, 194 – 6, 200, 210 – 11,  213 Dewey, John 34, 110, 112, 157 – 62, 165, 167, 171 – 6, 214, 244, 253 diagrammatic 149 – 50, 274; practice 150; reasoning 149, 274 disenchantment 266, 268, 271 – 3 dot quotation/quote/quoting 191, 200, 221 dualism 17, 31, 132, 209, 213, 244, 246, 264; Platonistic 31; of substance 17 Einstein, Albert 82 – 3, 91, 150, 154, 274 emergence/emergentism 169, 198, 209 – 10, 244, 254, 266 – 70,  274 empiricism: classical 84, 123, 131; logical 197 – 8, 203 – 5, 208 – 9, 211; nominalistic 199; scientific 99 Entity: abstract 1, 36, 39, 58, 187 – 8, 198 – 203, 207 – 8, 210, 215, 219 – 20; mental 151, 239, 274; objective 138, 144, 148, 153; theoretical 63, 191, 212, 272 epistemic 25, 30, 34, 39, 45 – 9, 56, 115, 123 – 4, 129, 131, 167, 207, 234, 242 – 3, 261 – 2, 265; access 261, 265; fact 43, 123; justification 167 epistemology 36, 40 – 1, 55, 113, 119 – 25, 134 – 5, 156, 158 – 61, 164 – 5, 167, 173 – 4, 177, 215, 219, 221, 231, 234 – 5, 239 – 40, 242, 245, 247, 249 – 50, 254, 266; early modern 36, 55; epistemological monism 234 essence 6, 28, 31, 55, 131, 149, 151, 239 – 40, 247 – 8, 252, 261 – 2, 264, 272 – 3; Santayana’s doctrine of essences 240

Index  279 ethics/ethical/ethos 114, 158, 171, 175 – 6, 195,  266 Euclid 142 Evans, Gareth 85, 94 evolutionary 129, 158 – 9, 165, 169 – 70, 173, 182, 234 – 7, 242 – 5, 249 – 50, 253 – 4; biology 249; explanations 242; naturalism 234 – 7, 243 – 5, 249 – 50, 253 – 4; realism 249; theory 158 – 9, 165, 173, 182 existence 7, 17 – 18, 37, 40 – 3, 46 – 7, 49, 51, 53, 55, 60, 67, 75, 77, 83, 132, 138, 151, 156, 199, 201 – 2, 204 – 5, 207, 211 – 12, 215, 222, 231, 237, 239, 267 – 8, 270; formal 17; objective 17 – 18, 40, 43, 49 experience: perceptual 91, 99, 106, 109, 120, 136 – 7, 243, 260; sense/ sensory 119 – 20, 121 – 2, 127, 131; theory of 36, 40, 58, 79 – 80, 90 – 1, 93, 95; tribunal of 138 explanation: behavioral/behavioristic 182, 185; causal 67, 69, 70 – 3, 183, 188, 190 – 1, 246; depersonalization of 266, 273; mythical 273; observational–correlational 274; physiological 179; psychological 180; reductive 180; scientific 61 – 4, 73, 100, 126, 165, 190; theoretical 63, 95, 192 Feigl, Herbert 34, 95, 155, 197, 214, 252 – 4 Fodor, Jerry 9 – 10, 217 – 18,  230 foundationalism 125, 131, 134, 139; anti-foundationalism 219 Frankena, William 250, 253 freedom 57, 89, 115, 181, 194, 214, 264 – 6, 274; human 181, 194, 265; realm of 265, 274 Frege, Gottlob 2, 4, 8, 138, 140 – 4, 146 – 55,  275 Freud, Sigmund 265 Friedman, Michael 154 – 5, 211, 213 functionalism 99, 110, 113; normative 99, 110, 113 functional role 2, 100, 106, 138, 148, 166, 178, 187 – 9, 190 – 3, 221 – 3, 227 – 9; functional role semantics 2, 178, 187 – 93, 221 – 3, 227 – 9 Geist (spirit) 35, 99, 102 – 4, 106 – 8, 110, 132, 158; absolute spirit 264 Glock, Hans-Johann 4 – 5, 9 – 11, 217, 227, 230

Goodman, Nelson 36, 217, 229 Gray, Jeremy 154 – 5 habit 70, 122, 129 – 31, 132, 134, 140 – 1, 145, 174, 179, 185, 266 – 7,  273 habituation 187 Hacker, Peter M.S. 217, 227, 231 Hall, W. Everett 229, 231, 275 Hanna, Robert 85 – 6, 94 Harman, Gilbert 3, 229, 232 Hegel, Georg Wilhelm Friedrich 8, 89, 94, 97 – 9, 101 – 5, 107 – 15, 136 – 7, 158 – 74, 251, 258, 272, 275; Hegelian serpent of knowledge 99 Hempel, Carl Gustav 197, 252 history of philosophy 1 – 12, 34, 36, 54, 58, 78, 97, 112 – 14, 137, 153, 155, 194, 216 – 18, 227 – 8, 232, 236 – 7, 250, 253 – 4, 257 – 63, 269, 271, 274 – 5; contextualist 6; external 268; internal 271 Holt, Edwin Bissell 234, 250, 253 Horkheimer, Max 271, 275 Humanism/humanist enterprise 235, 249 – 50 Humanist manifesto 235 Hume, David 3, 8 – 10, 36 – 7, 57, 59 – 78, 94, 113, 123, 129, 209, 252, 257, 264 Huxley, Thomas Henry 243 hylomorphism 17, 31 – 2 idea: material 16 – 17, 19; objective 16, 19 – 20, 22, 32; platonic 49; sensory 28 idealism 11, 55 – 7, 79 – 80, 83, 91, 93 – 5, 101 – 2, 111 – 15, 136 – 7, 222 – 3, 225, 227 – 9, 234, 239, 243 – 4, 248, 250; anti-idealism 234; linguistic 222 – 3, 225, 227 – 9; religiously-minded 234; transcendental 56, 79 – 80, 83, 91, 93 – 5,  102 identity 37, 42, 72 – 3, 75, 136, 140, 173, 235, 241; law of 140 image, manifest 74, 77, 80 – 1, 91, 93, 132, 154, 164 – 5, 174, 205, 208 – 12, 218, 245, 263 – 9, 271, 273 image, scientific 34, 61, 74, 81 – 2, 91 – 2, 94 – 5, 115, 154, 156, 176, 190, 192 – 3, 196, 203, 205, 208 – 11, 245, 254, 257 – 8, 260, 263 – 75 imagination 16, 28, 31, 65 – 6, 69 – 70, 80, 86, 89, 95; productive 86

280 Index imagism 59, 61, 74 immanent object 17 – 21, 23, 28, 30 individuals 37, 39 – 41, 45, 48, 51, 53, 55, 66, 79, 96, 104 – 6, 154, 183, 186, 192 induction 1, 267, 274 ineffability 186 – 7, 223 – 4, 227 – 9 inference 2, 19, 102 – 3, 122, 137 – 8, 141 – 9, 155, 159, 221 – 2, 224 – 6, 231, 251, 266; formal 102, 146; habits of 141; material 102 – 3, 144, 149, 151; rules/licenses 141 – 2, 144 – 6 inferential 100 – 3, 107 – 9, 111, 121, 123, 129, 132 – 3, 142, 144 – 7, 149, 153, 168, 170, 221, 261; articulation 111, 144, 147, 153; content 146, 168 inferentialism 74, 78, 94, 120 – 2, 133, 136 – 7, 175; hyper-inferentialism 134 innate/innatism 26 – 7,  33 inner episodes 27, 123, 126, 178 – 9, 181, 183, 185 – 8, 194, 220 instrumentalism 76, 95, 181, 184, 188, 214 – 15 intellect/intellective 16, 18, 26, 28 – 31, 113, 163 – 4,  172 intentionality 2, 5, 17 – 18, 30, 35, 56, 112, 163 – 5, 168, 174 – 6, 187, 199, 233, 271; of action 187; of thought 187 introspection 122, 174, 181 intuition 40, 74, 86 – 9, 107, 109, 111, 122, 127, 133, 135, 147, 153, 169, 199 – 200, 261 – 2, 266, 272; intellectual 261 – 2, 272; sensible 86 – 7 isomorphism 163, 240, 242, 252 James, William 1, 11, 57, 79, 94 – 6, 112 – 14, 157, 159, 176, 214, 231, 234, 239, 252 – 3 judgment 23, 39, 72, 86 – 7, 101, 111, 122, 127 – 31, 134, 142, 151, 209; perceptual 10, 80, 84 – 5, 136, 122, 128 – 31,  176 Jumblese 149, 151, 167, 174 Kant, Immanuel 2, 4, 5, 8, 10, 18, 31, 35 – 7, 40 – 1, 46 – 9, 54 – 61, 74, 78 – 80, 82 – 97, 99 – 100 – 3, 106 – 7, 109 – 11, 113, 115, 120, 127, 133, 137, 150 – 6, 165, 195, 204, 211 – 13,

216, 230, 251 – 2, 257 – 9, 264 – 5, 272, 274 Kenny, Anthony 19, 32, 35 knowledge: empirical 47, 57, 123, 125, 132, 138 – 9, 155, 198, 270; non-inferential 27, 123; perceptual 10, 80, 84 – 5, 136, 176; a priori 99; representative 49; scientific 198, 203, 212, 239; theory of 83, 94, 175, 197, 209 language: acquisition 180, 186 – 7, 191 – 2, 219; learning 187, 192; mathematical 152 – 3; natural 144, 152 – 3, 168; physiological characterization of 187; reductionist understanding of 182 language game 100, 114, 146, 155, 168, 176, 185, 195, 221 – 2, 227, 231; intralinguistic moves 146, 222; language departure 100; language entry transition 92, 100, 145, 188, 222 laws 38, 48, 61 – 2, 73, 77, 80, 82, 95, 99, 138, 140 – 4, 149 – 53, 155, 182, 209, 258, 264 – 5, 275; causal 80; of motion 73; of physics 48 Leibniz, Gottfried Wilhelm 3 – 4, 8, 10, 35 – 58, 78, 84, 88, 113, 257 Lewis, C. I. 93, 157 – 65, 172 – 6, 209, 248, 254 lingua franca 1, 6, 36, 54, 97, 101, 109, 216, 227 – 8,  257 Locke, John 10, 34, 61, 66, 78 – 9, 84, 88, 95, 113, 153, 156, 257, 264 logic 36, 43, 58, 98, 101 – 2, 108, 111, 113 – 15, 136, 138, 140 – 3, 150 – 6, 159 – 60, 164, 173, 175, 186, 197, 202, 209, 213 – 14, 219 – 20, 230, 275 – 6; symbolic 159; transcendental 150, 275 logical positivism 61, 73, 74, 78, 120, 160, 177, 193, 196, 209, 213, 247, 258, 252 Lovejoy, Arthur 234, 239, 253 Malcom, Norman 218 manifest image cf. image maps 166, 171, 173 – 5, 184, 196; cognitive 174, 184, 196 Marras, Ausonio 178, 191, 195, 221, 231 Marx, Carl 95, 215, 265

Index  281 materialism 58, 213, 243 – 4, 250, 253 – 4; reformed materialism 244 mathematics 138, 142 – 4, 149 – 53, 155 – 6, 230, 267 – 8, 274,  276 McDowell, John 37, 57, 80, 85 – 6, 92 – 4, 102, 105 – 7, 111, 113 – 14, 120, 122, 125 – 7, 129 – 37, 174 – 5,  229 meaning: conception of 99, 101, 178, 195, 227; functional concept of 27; Platonism about 262; reference theory of 185; as use 221 mental: acts/activity 1, 16 – 19, 22 – 3, 25 – 6, 28, 30 – 2, 40, 100, 107, 112, 199; concepts 28, 178 – 9, 182; entity 151, 239, 274; events 35, 95, 176; ideas 25 – 6, 30, 75; images 60, 187; life 179, 181, 183, 186; physics 150, 159, 244; representation 18, 59, 61, 74 – 5, 78; states 26, 61, 245, 267 – 8,  272 metalanguage 182, 184, 222 – 3; semantic, pragmatic 182, 222 metaphilosophy/metaphilosophical 3 – 4, 9, 54, 136, 219, 234, 236, 249, 257 – 8, 260, 263, 265 – 6, 268 – 70, 272 – 3 metaphysics: analytic/analytical 36, 170, 172, 174; metaphysical concept of reality 205, 207; metaphysical realism 207; transcendental 35, 58, 80, 96, 115 mind: Cartesian picture of 15; divine 47; mind-body problem 114, 195, 245, 254, 260, 276; mind-independent entity/object 86, 241 mindedness 158, 163, 165, 168 modality 167, 198, 271; modal structure 261, 263 model(s) 18, 27, 33 – 4, 38 – 40, 46, 52, 55, 62 – 4, 69 – 71, 73 – 4, 76 – 7, 79, 91, 100, 105 – 6, 122, 127, 143, 150 – 1, 179, 181, 267; perceptible models 63, 69 – 70, 73 – 4 monad 40 – 1, 55; monadic 41, 48, 150; monadic individuals 41, 48 Montague Pepperell, William 234, 250, 253 Moore, G.E. 110, 113, 247, 252 myth of the given 15, 24 – 7, 30, 61, 85 – 6, 107, 120 – 3, 126, 129, 131, 133, 139, 157 – 8, 162, 165, 173,

198, 202, 219 – 20, 222, 229, 251, 257 – 8, 260,  272 myth of jones 27, 54, 61, 137, 186, 191, 222, 232, 272 naturalism 10 – 11, 34, 57, 78, 93 – 4, 109, 112 – 13, 115, 120, 132, 174 – 6, 195, 213 – 14, 230 – 7, 243 – 5, 249 – 50, 252 – 4; biologistic naturalism (or old naturalism) 243; evolutionary naturalism (or new naturalism) 234 – 7, 243 – 5, 249 – 50, 253 – 4; liberal 112; mechanistic naturalism 244 – 5; scientific naturalism 93, 109, 244 nature: conception of 264; disenchantment of 266, 268, 273; human nature 62, 64, 70, 73 – 4, 76 – 8, 94, 271; laws of 99; objective representation of 264; philosophy of 98; realm of 265, 274; second nature 105; social nature 104, 160, 180 Neurath, Otto 229, 231 neuroscience 164, 168, 174 – 5, 189, 192, 246 neurosemantics 172 Newton, Isaac 40, 48, 62, 82 – 3, 91, 150, 274 Nietzsche, Friedrich 237, 254, 265 nominalism 11, 36, 40, 42, 55, 57, 132, 188, 199, 201 – 2, 213, 219, 230, 199; psychological nominalism 11, 57, 132, 199, 202, 219 non-representational: awareness 33; consciousness 24 norm 100, 102, 107, 108 – 9, 142, 149 normativity 10 – 11, 80, 94, 102 – 5, 107, 109, 111 – 13, 132, 136 – 7, 169, 174, 176, 195, 226, 231, 253, 266 norm-governed 102 objective idea cf. idea objectivity 16, 49, 83, 93, 105, 127, 152 – 3, 176, 249, 262, 264, 274; of concepts 153 observation(s) 17, 27, 72 – 3, 74, 76 – 7, 92, 125, 136, 138, 146, 163, 168, 184, 192, 207, 222, 245; direct 27, 77; observation reports 125, 138; perceptual 92 ontology 31, 34, 36, 45, 50, 57, 63, 75 – 7, 81, 93, 115, 137, 156, 176, 201 – 2, 204, 206 – 8, 210, 213 – 15,

282 Index 230 – 2, 235 – 6, 245, 247, 250, 254, 264; categorical 81, 264; of facts 50; manifest image ontology 82; ontological categories 56; platonic ontology 202; process ontology 247; scientific image ontology 83; of truth 45 particulars 36, 39, 53 – 4, 57 – 8, 62 – 4, 73, 77 – 8, 88, 144, 152, 219 – 20 Peirce, Sanders Charles 8, 119, 122, 127 – 37, 157, 159, 161 – 2, 175, 177, 211, 254 percept 122, 127 – 31, 135, 239, 240 – 1,  250 perception 16, 20 – 2, 24 – 5, 29, 31, 34, 40, 43, 55, 58, 64, 77 – 8, 84 – 6, 92 – 3, 95, 106 – 7, 109, 113 – 15, 119 – 22, 125 – 37, 156, 163, 174, 176, 195 – 6, 199, 214 – 15, 232 – 4, 240, 242 – 3, 249, 254, 262, 266, 275; petites perceptions 41; sense perception 16, 25, 29, 31; two-ply theory of 121, 127, 133 perceptual: belief 134; cognition 86, 91; content 106; judgment (cf. judgment); taking 86, 87 perennial philosophy 7, 263 – 4, 266 Perry, Ralph Burton 209, 234, 253 personhood 67, 265 phenomena 62 – 4, 73, 76 – 7, 93, 98, 150, 160, 169, 199, 205, 211, 267, 274 phenomenalism 83 – 4, 93, 130, 215 phenomenology 11, 102, 109, 112 – 13, 126, 131, 265 – 6, 275 physicalism 85, 198, 213, 235 picture theory 226 – 7, 225 – 8 picturing 87, 135, 157, 161 – 74, 176, 212, 225 – 6, 229 – 31; pictorial representation (cf. representation); vs. signifying 170 Pinkard, Terry 101 – 2, 107, 110 – 11,  113 Pippin Robert 101, 103, 105, 107 – 8, 110 – 11,  114 Pitkin, Walter B. 253 Plato 164, 172, 230, 257, 263, 268 Platonic: idea 49; ontology 202 Platonism 43, 199, 202, 210, 236, 262; Neo-Platonism 263 Popper, Karl 98, 114

positivism 74, 78, 105, 177, 193, 196, 209, 213, 236, 247; logical (cf. above; logical positivism) pragmatism 3, 8, 57, 95, 110 – 12, 120, 133, 136 – 7, 157 – 62, 165, 172 – 3, 175 – 7, 195, 239, 247 – 8, 250, 253; American 8, 175; analytic pragmatism 120 – 37; conceptual 159, 167; history of 157, 160, 162; neopragmatism 136 – 74 pragmatist: social 164 – 5, 168; theory of cognition 165; theory of concepts 161; theory of meaning 129, 134; tradition 98, 157, 172 Price, H.H. 220, 231, 148 Price, Huw 157, 165, 170, 171, 172, 174, 175, 211, 214 a priori 79, 87, 99, 155, 168, 231, 260, 261 Pritchard, H.A. 252 Psychological: nominalism (cf. nominalism); notions 180, 193; theory 178, 181 psychology 79, 158, 173, 179, 181, 189, 194 – 6, 199, 243, 249, 253 – 4, 270 – 1; American 194; transcendental psychology 79 Putnam, Hilary 9, 11, 15, 217 – 18, 231, 250 Qu, Hsueh 74, 78 qualia 91, 119, 121 – 2, 136 quality(ies) 29 – 31, 48, 60, 61, 63 – 5, 66 – 78, 81 – 3, 91 – 3, 133, 199, 210 – 11, 220, 240, 241, 264 – 5; primary 29, 30, 48; secondary 29, 31, 48; sensible 61, 70, 81 – 3, 91 – 3 Quine, Willard Van Orman 3, 157, 173, 175, 177, 198, 203, 206, 209, 214 – 15, 217, 229 – 30 rationalism 10, 56, 58, 120, 124, 274 rationality 11, 28, 103, 112, 123, 131, 138 – 40, 144, 153, 206, 264, 266; common-sensist account of 123; Sellars’ account of 131 realism 3, 19, 43, 46, 48, 74, 76, 80, 82, 84, 92, 93, 122, 129, 130 – 2, 134, 193, 202, 203, 208, 212, 216, 221, 234, 235, 239, 242, 244 – 5, 247 – 8, 250 – 2, 275; American realism 8; anti-realism 207; critical realism 193, 234, 235, 239, 242,

Index  283 244, 245, 247, 248, 249; direct realism (or presentational) 122, 129, 130, 131, 134, 234, 242, 243; empirical realism 84, 93; evolutionary 249; indirect realism 19; naive realism 43, 48, 248; New Realism 3, 193, 234, 248; physical realism 251; scientific realism 48, 76, 80, 82, 84, 92, 93, 132, 208, 208, 212; transcendental realism 46, 74 reality 2, 12, 34, 40, 55, 57 – 8, 78, 83, 95, 114 – 15, 124, 131, 136 – 8, 144, 152 – 3, 156, 162, 173, 176, 184 – 5, 193, 195 – 6, 201, 204 – 8, 212, 214 – 15, 226, 232, 239, 246, 249, 253 – 4, 258, 261, 263, 268, 273 – 5; extra–mental 274; formal 40, 55, 57; mental 274; objective 2, 40, 55 – 7, 144; psychological 184 receptivity 88, 97, 106, 107, 111, 120, 127, 154; sheer receptivity 88, 127, 154 Redding, Paul 110, 111, 120, 121, 126, 128, 130, 134 reductionism 179, 181; anti-reductionism 219 Reichenbach, Hans 197, 252 Reid, Thomas 19, 129, 130 relations 37, 39, 41 – 3, 45, 47, 66, 67, 69, 87, 88, 89, 90, 142, 149, 150, 152, 164, 171, 199, 201, 202, 241, 246, 264, 265, 274; causal 67, 142, 202, 222, 246, 265, 274; contiguity and causation 65, 67, 69; external 241; inferential (cf. inference); Leibnizian theory of 37, 39, 41 – 3, 45, 47; logic of, 150, 164; maps-like 171; resemblance relations 66; spatial 88, 149 representable 32, 47, 48, 49, 81, 211 representation 18, 19, 28, 39, 41, 44, 49, 55, 59, 60, 61, 64, 66 – 9, 74, 75, 77, 84, 87, 88, 91, 126, 130, 151 – 2, 154, 166 – 72, 174, 180, 234, 242; act\content model of, 40; cognitive 172, 174; concept of, 39; conceptual 17, 84, 87, 91, 126, 130; environmental and external (e-representation) 170 – 1; first-order 41; general 65, 66, 67, 69, 76, 77; Hume’s theory of 66 – 9, 74, 75, 77; inferential and internal (i-representations)170 – 1; internal 242; Kant’s division of representation

152, 154; Kant’s notion of 49, 55; Leibnitian 41, 44; linguistic 168; mental (representation) 18, 19, 59, 61, 174, 180; neurophysiological theory of 189; nonconceptual 84, 88; non-linguistic 169; pictorial 167, 168, 169; realm of 47; second-order 41; sensory 28, 84, 88, 91; singular 151; of a substance 64, 65 representational: act(s) 16, 18, 22, 23, 31, 91; awareness 33; (character of) ideas 23, 31, 32; consciousness 34; content 61, 77, 84, 85, 86, 87; function 28; intermediaries 30; model of knowledge 46; object 55; representational system (cf. system); representation of a manifold; state(s) 29, 166; theory of mind 36; thoughts 23 representationalism 133, 170, 171, 174, 234, 239; anti-representationalism 169; classical representationalism 239 representationalist account of content 120 representings 32, 46, 80, 88, 172, 212 Riemann, Georg F. Bernhard 150, 274 Rorty, Richard 2, 4, 5, 6, 9, 19, 32, 92, 157, 162, 165, 167, 171, 173, 174, 229 Rosenberg, Jay 2, 80, 81, 89, 90, 91, 92, 102, 157, 165, 166, 167, 170, 172, 225, 229 Royce, Josiah 110, 159, 161, 162, 173, 234 rule 56, 99, 100 – 1, 102, 106 – 9, 111, 141, 142 – 3, 145, 147, 148, 154, 169, 173, 187, 188, 222, 225, 227; inferential 100, 141, 142, 146, 225; linguistic rule 187, 188, 227; of logic 140, 143, 145, 146; ought-to-be 44, 55, 82, 85, 107, 173; ought-to-do 55, 145; rule following 134; semantic 145, 222; syntactical 146, 187, 222; system of 222, 227 Russell, Bertrand 36, 53, 54, 98, 143, 247; Russell’s paradox 143 Ryle, Gilbert 178, 191, 192, 194, 227, 272 Santayana, George 234, 239, 240, 244, 247, 252 Schliesser, Eric 4

284 Index Seibt, Johanna 157, 165, 167, 168, 169, 170, 172, 174 self-reflective conception (of philosophy) 239, 257 – 60, 270 Sellars, Roy Wood 233 – 52 semantics 1, 2, 18, 97, 99, 100, 101, 103, 107, 109, 111, 121, 133, 145, 164, 173, 187 – 8, 189, 191 – 3, 199, 202, 210, 212, 219, 220, 222 – 3, 224 – 7; analysis 208; assertibility 162, 171; concepts 183, 184, 187; conceptual role 97, 99, 100, 101, 106, 109; content 172, 261, 262; descriptive 210; discourse 171, 185; formal 18; functional role 2, 179, 187 – 8, 189, 190, 191, 192, 193, 222; intensional 173; metavocabulary/metalanguage 170, 182; neurosemantics 174; notions 164; rules 145, 222; statement 149, 221, 223; Tarski-Carnap semantics 202, 212, 223, 229; terms 180; vocabulary 190; Wittgensteinian 219 sensation 18, 25, 26, 28 – 9, 30 – 1, 60, 81, 89, 106, 123, 125, 126 – 7, 129 – 31, 135, 160 – 1, 163, 181, 219, 240 – 2,  267 sense-datum 123, 129, 135, 210, 242, 243; theories of 61, 139, 250 sense impressions 59, 61, 76, 81, 85, 87, 89, 90, 93, 106, 128, 129, 131 sensing 28, 33, 34, 81, 85, 87, 89, 93, 139, 151, 243 Sinn/(Fregean) sense 143, 144, 146, 148, 152 skeptical: anxieties 266; method 25 skepticism 160, 164, 246, 274; epistemological 264 Skinner, Burrhus Frederic 5, 6, 9, 126, 179, 181 – 5, 186, 187, 188, 192, 193, 194 Smart, John Jamieson Carswell 84 soul 16, 31, 33, 66, 89, 244; monism 16, 31; sensory 16, 31; vegetative 16, 31 space 40 – 1, 46, 48, 50, 60, 82, 87, 88, 89, 90, 133, 147, 172, 203; of causes 266, 271; conceptual 165, 168, 269; of implications 141; intuitional representation of 88; logical space 106, 219, 220, 262, 263, 269; of reasons 106, 110, 119, 122, 124, 126, 128, 133, 135, 225, 266, 271; social space 102

speech 33, 52, 171, 172, 174, 186, 187, 188, 201; external 186; fictional 52; figures of 188; inner 33; material mode of 201; private 54; speech acts 171, 172, 174 Spencer, Herbert 243 Spinoza, Baruch 264 stimulus-response 121, 131, 133, 179, 181, 182, 251; analysis 179; model 181 Strawson, Peter, Frederick 4, 36, 79, 80, 83, 93, 165, 265, 266 Stroud, Barry 110, 218 subjectivism 79, 242; phenomenalist subjectivism 242; transcendental 79 substance 17, 18, 21, 37, 38, 39, 40, 41, 42 – 5, 48, 49, 50, 51, 52, 62, 64 – 5, 67, 68 – 73, 75 – 8; Cartesian 18; dualism of 17; idea of 64, 65, 75, 76; individual 37, 39, 43, 44, 48, 49, 50, 52; material 18; natural 37; nature of 38, 43, 44, 50; unobservable 62 system(s) 2, 27, 40, 41, 44, 45, 47, 49, 56, 59, 63, 77, 81, 85, 101, 106, 108, 142 – 3, 145, 147, 159, 164, 166 – 9, 172 – 3, 188 – 9, 204 – 5, 211 – 12, 222, 233, 235, 236, 242, 245, 246, 252, 258, 259, 264, 266, 270; cognitive 168, 169, 172; of concepts 145; conceptual 159, 189, 204; inferential 101, 147; linguistic 106; logical 159; of mathematics 142; (micro)physical 27; monadic 40 – 1; nervous 81, 85, 164, 242; orientation system 169; physical 246, 266, 273; representational 81, 166, 189 solar 63, 77; of states of affairs 47; visual 246 Tarski, Alfred 223, 229, 252 theism 243, 250 thing-in-itself 211 thinking-out-loud 26 – 7,  33 thought(s) 20 – 2, 23, 24, 27, 28, 30, 31, 32 – 3, 37, 46, 49, 52 – 4, 59 – 60, 70, 73, 83, 86, 87, 91, 98, 99, 100, 102, 104, 107, 109, 121 – 6, 128, 132 – 4, 138 – 9, 141, 150 – 2, 159, 162, 164 – 5, 169, 181, 183 – 8, 190, 193, 198, 219, 222, 233 – 5, 237, 241, 248 – 9, 252, 259, 260 – 2, 265, 267 – 8, 272, 275; classical picture

Index  285 of 187; cogitationes (thoughts) 16, 17; conceptualized 86; confused 28; empty 83, 91; erroneous thoughts 128; first-person thought 181; lingua franca of, 6, 54; linguistic 169; non-ideational 16, 21, 22, 23, 24, 31, 32 – 3; non-verbal 27; postulation of, 267, 268; practical attitudes of, 159; private 186; representational 23; thought episodes 27; thought experiment 121, 123, 178; thought-responses 87 Tolman, Edward 179, 180, 181, 184, 186, 187, 188, 192 – 3, 194, 195 transcendental 46, 49, 56, 74, 79, 80, 83, 86 – 7, 90 – 1, 100, 102, 107, 110, 211, 218, 258 – 9, 260 – 1, 264; deduction 86 – 7, 107; idealism 79, 80, 83, 90 – 1, 102; illusion 258 – 9; imagination 95; linguistics 100; philosophy 261; psychology 79; realism 46, 56, 74; theory of experience 79; turn 264 truth 23, 36, 39, 40, 42, 45, 46, 47, 48, 50, 56, 91, 98, 119, 127, 140 – 2, 162, 170, 173, 174, 180, 208, 211 – 12, 223, 225, 226; correspondence theory of 40, 45, 46; Fregean conception of 140; Kant’s conception of 48; norm of 142; Pierce account of, 162, 211; truth-apt 119, 127; truth-evaluable 23; truth-value 153, 223 truthfulness 46, 48, 205, 207

value 24, 25, 46, 90, 153, 207, 223, 236, 240, 249, 267, 268, 272, 273, 273; explanatory 267; moral 274; truth-value 153, 223; value laden 273 Ward, James 239 Warren, W. Preston 250 Watson, John B. 179, 180, 181, 182, 183, 186, 187, 189, 192, 193, 194 Weber, Max 263, 266, 268, 271, 272, 273 Wittgenstein, Ludwig 56, 100, 101 – 2, 106, 110 – 11, 134, 162, 171, 185, 216 – 29, 249, 262, 265 world 26, 37, 41, 43, 45 – 7, 49 – 53, 55, 61, 63, 74, 80 – 3, 92, 111, 119 – 21, 132, 113, 135n15, 135, 142, 144, 154, 158, 162 – 5, 167, 169, 170 – 3, 176 – 7, 182 – 3, 203 – 5, 208, 209, 212, 222, 225, 228, 239, 244, 245, 258 – 60, 262 – 8, 270 – 1; actual world 46, 52, 55n9, 63, 142, 182; animal 158; of appearances 48, 203, 209; common sense 82, 91, 135, 203, 209; disenchantment of the (p. 413); empirical 144, 165; external 135n18, 170, 208, 239, 267; manifest image of (cf. manifest image); natural 81, 85, 163 – 4, 270; noumenal 208, 258; perceptual 87, 80, 83, 90, 205; philosophic 241; physical 48, 45, 49, 79, 82; possible worlds 47, 49, 52, 142; real 263 Wright, Edmond 250