Wollheim, Wittgenstein, and Pictorial Representation: Seeing-as and Seeing-in 9781317278658, 1317278658

Pictorial representation is one of the core questions in aesthetics and philosophy of art. What is a picture? How do pic

528 79 33MB

English Pages 308 [324] Year 2016

Report DMCA / Copyright


Polecaj historie

Wollheim, Wittgenstein, and Pictorial Representation: Seeing-as and Seeing-in
 9781317278658, 1317278658

Table of contents :
Title Page
Copyright Page
Table of Contents
List of figures
Notes on contributors
PART I Wittgenstein and seeing-as
1 The room in a view
PART II Difficulties with Wollheim’s borrowing from Wittgenstein
2 Seeing aspects and telling stories about it
3 Aspects of perception
4 Aspect-perception, perception and animals: Wittgenstein and beyond
5 Wittgenstein’s seeing as: a survey of various contexts
PART III Benefits from Wollheim’s borrowing from Wittgenstein
6 Leonardo’s challenge: Wittgenstein and Wollheim at the intersection of perception and projection
7 ‘Surface’ as an expression of an intention: on Richard Wollheim’s conception of art as a form of life
8 Richard Wollheim on seeing-in: from representational seeing to imagination
PART IV Rescuing Wollheim’s account without the support of Wittgenstein
9 A measure of Kant seen in Wollheim
10 Seeing-in as aspect perception
PART V Imagination and emotion in Wollheim’s account of pictorial experience
11 Wollheim’s ekphrastic aesthetics: emotion and its relation to art
12 Visions: Wollheim and Walton on the nature of pictures David Hills

Citation preview

Wollheim, Wittgenstein, and Pictorial Representation

Pictorial representation is one of the core questions in aesthetics and philosophy of art. What is a picture? How do pictures represent things? This collection of specially commissioned chapters examines the influential thesis that the core of pictorial representation is not resemblance but ‘seeing-in’, in particular as found in the work of Richard Wollheim. We can see a passing cloud as a rabbit, but we can also see a rabbit in the clouds. ‘Seeing-in’ is an imaginative and perceptual act of the kind employed by Leonardo’s pupils when he told them to see what they could – for example, battle scenes – in a wall of cracked plaster. This collection examines the idea of ‘seeing-in’ as it appears primarily in the work of Wollheim but also its origins in the work of Wittgenstein. An international roster of contributors examine topics such as the contrast between seeing-in and seeing-as; whether or in what sense Wollheim can be thought of as borrowing from Wittgenstein; the idea that all perception is conceptual or propositional; the metaphor of figure and ground and its relation to the notion of ‘two-foldedness’; and the importance in pictorial art of emotion and the imagination. Wollheim, Wittgenstein, and Pictorial Representation: Seeing-as and Seeing-in is essential reading for students and scholars of aesthetics and philosophy of art, and also of interest to those in related subjects such as philosophy of mind and art theory. Gary Kemp is Senior Lecturer in Philosophy at the University of Glasgow, UK. He is the author of Quine versus Davidson: Truth, Reference and Meaning (2012), and What is This Thing called Philosophy of Language? (Routledge, 2013). Gabriele M. Mras is Associate Professor of Philosophy at the WU Vienna, Austria. Her writings include Naturalismus, Reduktion und die Bedingungen von Gedanken (2002) and Wahrheit, Gedanke, Subjekt (2001), and she is co-editor of Conceptus: Journal of Philosophy.

This page intentionally left blank

Wollheim, Wittgenstein, and Pictorial Representation Seeing-as and Seeing-in

Edited by Gary Kemp and Gabriele M. Mras

First published 2016 by Routledge 2 Park Square, Milton Park, Abingdon, Oxon, OX14 4RN and by Routledge 711 Third Ave., New York City, NY. 10017 Routledge is an imprint of the Taylor & Francis Group, an informa business © 2016 Gary Kemp and Gabriele M. Mras, editorial and selection matter; individual chapters, the contributors The right of Gary Kemp and Gabriele M. Mras to be identified as 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. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging in Publication Data Names: Kemp, Gary, 1960 October 15- editor. Title: Wollheim, Wittgenstein, and pictorial representation: seeing-as and seeing-in/edited by Gary Kemp and Gabriele M. Mras. Description: 1 [edition]. | New York: Routledge, 2016. | Includes bibliographical references and index. Identifiers: LCCN 2015044020 | ISBN 9781138123465 (hardback: alk. paper) | ISBN 9781315640983 (e-book) Subjects: LCSH: Representation (Philosophy) | Art–Philosophy. | Aesthetics, Modern. | Wollheim, Richard, 1923–2003. | Wittgenstein, Ludwig, 1889–1951. Classification: LCC BH301.R47 W65 2016 | DDC 701–dc23LC record available at http://lccn.loc.gov/2015044020 ISBN: 978-1-138-12346-5 (hbk) ISBN: 978-1-315-64098-3 (ebk) Typeset in Sabon LT Std by Sunrise Setting Ltd, Brixham, UK


List of figures Notes on contributors Acknowledgements Introduction

vii ix xii xiii

PART I Wittgenstein and seeing-as


1 The room in a view Charles Travis PART II Difficulties with Wollheim’s borrowing from Wittgenstein



2 Seeing aspects and telling stories about it Joachim Schulte


3 Aspects of perception Avner Baz


4 Aspect-perception, perception and animals: Wittgenstein and beyond Hans-Johann Glock 5 Wittgenstein’s seeing as: a survey of various contexts Volker A. Munz PART III Benefits from Wollheim’s borrowing from Wittgenstein 6 Leonardo’s challenge: Wittgenstein and Wollheim at the intersection of perception and projection Garry L. Hagberg







7 ‘Surface’ as an expression of an intention: on Richard Wollheim’s conception of art as a form of life Gabriele M. Mras


8 Richard Wollheim on seeing-in: from representational seeing to imagination Richard Heinrich


PART IV Rescuing Wollheim’s account without the support of Wittgenstein 9 A measure of Kant seen in Wollheim Gary Kemp 10 Seeing-in as aspect perception Fabian Dorsch PART V Imagination and emotion in Wollheim’s account of pictorial experience 11 Wollheim’s ekphrastic aesthetics: emotion and its relation to art Michael Levine 12 Visions: Wollheim and Walton on the nature of pictures David Hills Index

181 183






List of figures

4.1 4.2 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8 6.9 6.10 6.11 6.12 6.13 6.14 6.15 6.16 6.17 6.18 6.19 10.1 10.2 10.3 10.4 10.5

Unknown artist, duck–rabbit Unknown artist, young lady–old witch Paul Cézanne, Château Noir 1904 Pablo Picasso, Factory, Horta de Ebro 1909 Georges Braque, Viaduct at L’Estaque 1908 Pablo Picasso, Self-Portrait 1972 Arnold Newman, photographic portrait of Picasso 1954 Rembrandt van Rijn, The Adoration of the Shepherds 1646 (the smaller) Rembrandt van Rijn, The Adoration of the Shepherds 1646 (the larger) Rembrandt van Rijn, The Adoration of the Shepherds 1646 (the smaller, detail of the fingers with pink edges) Hans Hofmann, Equinox 1958 Jasper Johns, Flag 1966 William Harnett, A Study Table 1882 Frank Auerbach, Study after Titian I 1965 Frank Auerbach, Study after Titian II 1965 Titian, Tarquin and Lucretia 1571 Frank Auerbach, Sketch from Titian’s ‘Bacchus and Ariadne’ 1970–1 Titian, Bacchus and Ariadne 1520–3 Paul Cézanne, The Bathers (also called The Large Bathers) 1898–1905 Paul Cézanne, Bathers 1900–5 Paul Cézanne, Bathers 1894–1905 Kasimir Malevich, Black Circle 1915 Vincent van Gogh, Wheat Field with Cypresses 1889 (black-and-white detail) Unknown artist, A Bowl Unknown artist, A Road with a Median Strip Count Johann, Window Grid 1531

79 85 121 122 122 123 124 127 128 129 132 135 138 143 144 145 147 147 152 153 154 210 214 217 218 219


List of figures

10.6 10.7 10.8 10.9 10.10 10.11 10.12 10.13 12.1 12.2 12.3

Albrecht Dürer, An Artist Drawing a Seated Man 1525 Unknown artist, duck–rabbit figure Unknown artist, Necker cube Ronald C. James, Dalmatian Dog 1965 Rembrandt Harmenszoon van Rijn, Posthumous Portrait of the Preacher Jan Cornelisz Sylvius ca. 1645 Rembrandt Harmenszoon van Rijn, Sketch for the Posthumous Etching of Jan Cornelisz Sylvius ca. 1646 Unknown artist, Kanizsa’s triangle Unknown artist, Ehrenstein illusion Aaron Siskind, Chicago 10, 1948 (photograph) Meindert Hobbema, Mill with the Great Red Roof 1662–4 Pieter Bruegel the Elder, The Hunters in the Snow 1565

220 222 222 228 229 229 232 233 272 291 299

Notes on contributors

Avner Baz is Associate Professor of Philosophy at Tufts University, US. He has written extensively about Wittgenstein, focusing on his remarks on aspect perception and on his method. His book When Words are Called For (2012) defends the Wittgensteinian approach to the dissolution of philosophical difficulties, and brings it to bear on current debates – concerning knowledge and method – within mainstream analytic philosophy. Fabian Dorsch is SNSF Research Professor at the University of Fribourg, Switzerland. He is author of The Nature of Colours (2009), The Unity of Imagining (2012) and the forthcoming Imagination (Routledge 2017) and has published on various topics in the philosophy of mind, epistemology, metaphysics, aesthetics and the philosophy of normativity. He founded the European Society for Aesthetics and is editor of the journal Estetika – The Central European Journal of Aesthetics. Hans-Johann Glock is Professor of Philosophy at the University of Zurich, Switzerland and Visiting Professor at the University of Reading, UK. He has held positions at Oxford and Reading, as well as visiting professorships, research and prize fellowships in Canada, Germany and South Africa. He is the author of A Wittgenstein Dictionary (1996), Quine and Davidson (2003), La mente de los animals (2009) and What Is Analytic Philosophy? (2008). He has edited several anthologies on Wittgenstein and analytic philosophy, and is currently completing a book on animal minds and co-editing The Blackwell Companion to Wittgenstein. Garry L. Hagberg is the James H. Ottaway Professor of Philosophy and Aesthetics at Bard College, US and has in recent years also been Professor of Philosophy at the University of East Anglia, UK. Author of numerous papers at the intersection of aesthetics and the philosophy of language, his books include Meaning and Interpretation: Wittgenstein, Henry James, and Literary Knowledge (1994), Art as Language: Wittgenstein, Meaning, and Aesthetic Theory (1998) and Describing Ourselves: Wittgenstein and Autobiographical Consciousness (2008). He is editor of Art and Ethical Criticism (2010), co-editor of A Companion to the Philosophy of Literature (2009) and editor of the journal Philosophy and Literature.


Notes on contributors

Richard Heinrich teaches philosophy at the University of Vienna, Austria. He is author of, among others, Wittgenstein’s Grenze (1993), Wahrheit (2009) and, recently, ‘Green and Orange – Colour and Space in Wittgenstein’ in Frederik A. Gierlinger and Štefan Riegelnik (eds), Wittgenstein on Colour (2014). David Hills is Associate Professor of Philosophy (Teaching) at Stanford University, US. Gary Kemp is Senior Lecturer in Philosophy at the University of Glasgow, UK. He has published papers on the philosophy of language, the history of analytic philosophy and aesthetics, and he has also edited or written several books, including Quine and His Place in History (with co-editor Frederique Janssen-Lauret, 2015), What Is This Thing called Philosophy of Language? (Routledge, 2013) and Quine versus Davidson: Truth, Reference and Meaning (2012). Michael Levine is Professor of Philosophy at the University of Western Australia. Publications include Prospects for an Ethics of Architecture (with Bill Taylor, Routledge, 2011), Doing Philosophy, Watching Movies (with Damian Cox, 2011), Politics Most Unusual: Violence, Sovereignty and Democracy in the ‘War on Terror’ (with Damian Cox and Saul Newman, 2009), Racism in Mind (edited with Tamas Pataki, 2004), Integrity and the Fragile Self  (with Damian Cox and Marguerite La Caze, 2003), Engineering and War: Ethics, Institutions, Alternatives (with Ethan Blue and Dean Nieusma, 2013) and The ‘Katrina Effect’: Reflections on a Disaster and Our Future (edited with Bill Taylor, Oenone Rooksby and Joely-Kym Sobott, 2015). He is currently working on aesthetics of catastrophe as well as on regret and other emotions of self-assessment. Gabriele M. Mras is Associate Professor at the WU Vienna, Austria and has taught philosophy at the University of Vienna, UC Berkeley and the University of Minnesota, Minneapolis. She is the author of Wahrheit, Gedanke und Subjekt (2001) and Naturalismus, Reduktion und die Bedingungen von Gedanken (2002). More recent publications are on Wittgenstein on colour and on Frege on the unity of judgement. Volker A. Munz is Assistant Professor at the Alpen-Adria University Klagenfurt, Austria. His research interests include Austrian philosophy, philosophy of language, philosophy of mind, Wittgenstein and Viennese modernity. He is the editor (with Bernhard Ritter) of Wittgenstein’s Whewell’s Court Lectures 1938–41 (forthcoming) and the author of Satz und Sinn: Bemerkungen zur Sprachphilosophie Wittgensteins (2005) and numerous essays on the mentioned subjects. Joachim Schulte taught philosophy at the University of Zürich until 2015. He has published a number of articles and four books on the philosophy of

Notes on contributors xi Wittgenstein. He is a member of the Board of Editors of Wittgenstein’s Nachlass and co-editor of critical editions of Wittgenstein’s Tractatus LogicoPhilosophicus and Philosophical Investigations. In recent years he has chiefly worked on Wittgenstein’s remarks on certainty and doubt. Charles Travis is Professor of Philosophy Emeritus of King’s College London, UK and Professor Afiliado of the Universidade do Porto, Portugal. He works on separating the logical from the psychological, the general from the particular and then, in each case, relating first to second, thus on the relations between perception, thought and language. He is writing a book on Frege.


We would like to express our gratitude to Mary Day Wollheim and to Bruno Wollheim for help and encouragement; to the Diebenkorn Foundation for permission to use the picture that appears on the cover; to the Vienna University of Economics and Business for hosting the conference Wittgenstein and Wollheim: Seeing-As and Seeing-In (12–14 December 2013), at which the idea for this volume was hatched; and to Tony Bruce and Adam Johnson of Routledge, not least for their patience. Gary Kemp would like to single out Garry Hagberg for thanks, who is not only a contributor to this volume but introduced Art and Its Objects to him when he was a mere undergraduate at the University of Oregon in 1983–4 (where Hagberg had just finished his PhD); thanks also to Malcom Budd for introducing him, a bumbling postgraduate in 1985, to the Philosophical Investigations, and whose voice has been extraordinarily influential on the subjects of pictorial representation in general, Wollheim’s views in particular and, of course, the later Wittgenstein; and thanks to Chris Belshaw for conversation over the years. Gabriele M. Mras would like to thank Oliver Feldmann, Stefan Riegelnik, Barry Stroud and Florian Wagner for their encouragement and help, and for all the fruitful discussions. And lastly the editors thank each other.


Once upon a time, however painstakingly and perhaps over generations, human beings, or their ancestors, began to make pictures. Richard Wollheim suggests a just-so story of how it might have come about. Surely a caveman or a cavewoman could see things as other things – a cloud as an animal, a torso, a head, or see such things in a rock face, without having yet had so much as an idea of pictorial depiction. But then, all it would take to invent the practice of pictorial depiction, would be for the cave-dweller to begin marking a cave wall in such a way that he could see a face or an animal in it, and to get others to so see it, and then for somebody to copy the skills of that first picture-maker. Simple. What could possibly be wrong with such an account? The key idea – the capacity of seeing-as or seeing-in – seems innocuous; after all, children seem spontaneously to exercise it when they look at clouds or picture puzzles. One might suppose that Wollheim’s use of the notion of seeing-as, or, later, seeing-in, could safely be detached, for these purposes, from that manyheaded hydra that is Wittgenstein’s discussion, which has proven so invigorating yet so perplexing to philosophy. But just as the subject had been in Wittgenstein’s hands so protean, so unamenable to safe interment within a particular department of philosophy – it straddles perception, judgement, meaning, knowledge, representation, intersubjectivity and aesthetics, to name a few – so its various interpretations have profound and sometimes deleterious repercussions for Wollheim’s use of that notion, that form of words. So much so that in Part I of this collection – ‘Wittgenstein and seeing-as’ – we present unapologetically Charles Travis’s attempt to locate and explicate the notion in Wittgenstein’s thought – striving especially to say whether or in what sense the seeing of something as another is an objective or an intersubjective matter – without referring at all to Wollheim. Likewise in Part II – ‘Difficulties with Wollheim’s borrowing from Wittgenstein’ – we present Joachim Schulte’s, Avner Baz’s and Hanjo Glock’s respective cases that Wittgenstein’s treatment of it actually undermines Wollheim’s use of it (along with much else concerning Wittgenstein). For Wittgenstein, according to all three of these figures, when the aspect dawns in seeing the famous drawing as



duck, it is quite an unusual case, one that is fleeting, that lasts only as long as one is ‘occupied with the object in a special way’, in Wittgenstein’s terms – more a transitory event than an enduring state, one that is at the frontier of ordinary perception. It is not a place to found a theory of the ordinary experience of pictures. Volker Munz’s conclusion is not so unfavourable to Wollheim, but he argues that Wittgenstein would not accept there is any one thing that is aspect-perception, a conclusion that chimes with Wittgenstein’s idea of family resemblance and casts some doubt on the idea that his views can profitably be seen as lending support to Wollheim. In Part III – ‘Benefits from Wollheim’s borrowing from Wittgenstein’ – we present cases positively in favour of a Wittgensteinian reading of Wollheim. Both Heinrich and Hagberg show, in very different ways, that Wollheim’s understanding of depiction, or pictorial meaning generally, stands only to gain from the invocation of the complexity of Wittgenstein’s discussion of aspect-perception – from a whole field involving the fine-grainedness, interpenetration and multiplicity of aspects and the entanglement of perception with thought. Gabriel Mras meanwhile focuses on the precise nature of aspect-perception – as Wollheim, taking his cue from Wittgenstein, saw it – to diagnose exactly how Wollheim was motivated in formulating his account of pictorial representation (partly in response to what he perceived as inadequate in Collingwood’s famous ideal theory of art). Any account which separates the duck from the collection of lines, and analogously the meaning of a work of art from its surface, goes wrong. And yet, of course, Wittgenstein’s is not the only or last philosophical word. Perhaps the truth about aspect-perception is not what Wittgenstein said it was, or perhaps there is another notion of aspect-perception to be invoked. True, it is hard or perhaps impossible to maintain such things without committing what for Wittgenstein is a deep if seductive mistake: thinking that here is a phenomenon that one can keep one’s philosophical eye on even where language falls away, where the reality of the phenomenon is not supported by what we ordinarily say about it. Nevertheless in Part IV – ‘Rescuing Wollheim’s account without the support of Wittgenstein’ – Dorsch and Kemp try to support Wollheim’s ideas, trying in different ways to present a Wollheimian account of depiction with the requisite notion aspect-perception at the centre while avoiding the difficulties in seeing Wollheim as feeding off a right interpretation of Wittgenstein. The interlocutors here are not interpreters of Wittgenstein so much as contemporary theorists of depiction or pictorial representation, such as Malcom Budd, Robert Hopkins and Kendall Walton. The two papers in Part V – ‘Imagination and emotion in Wollheim’s account of pictorial experience‘ – call for slightly more in the way of introduction. By the time of the second edition of Art and Its Objects (1980), Wollheim had shifted his account from seeing-as to seeing-in, and connected seeing-in – in his use of the term – to the faculty of the imagination. One perceives the lines and colours constituting the picture, and one relies on



one’s capacity for seeing-in – involving one’s imagination but not identical to it – to see in it what it actually depicts. The reason that all pictures are not like Rorschach tests is that of the possibly many things that can be seen in a picture, the one actually depicted is selected by the artist’s intention – ‘intention’ in a supple and implicit sense if need be. Michael Levine points out that on Wollheim’s own psychoanalytic understanding of emotion – as providing one with an ‘orientation’ to the world – one’s readiness to see a given object or state of affairs in a picture is partly driven by emotion and desire. The uptake of artistic intention relevant to depiction cannot therefore ultimately be disentangled from that relevant to expression. Furthermore, there is a case to be made that Wollheim overestimated the role of seeing-in in the experience of art; the result is that at the centre of Wollheim’s outlook on art are not perceptual notions so much as emotion, intention and imagination as informed by psychoanalysis. The role and capabilities of the imagination can be carried further, perhaps with the effect of blurring the boundaries between representation and expression even more. One natural reaction to Wollheim’s claim about the role of intention is that things simply aren’t so cut and dried as the claim would have it, not even if we allow that such intentions may be implicit or unconscious. Not only is the intention of the artist not always sacrosanct, a right-experience of a picture generally just doesn’t have the character of verifying experientially an hypothesis; it is a process played out in time, with parts that can legitimately be rearranged, with corresponding differences in the meaning or value of the experience. Now as is well known Wollheim rejects Kendall Walton’s ‘make-believe’ account of depiction, according to which pictures present props which we use in our imaginative games, in particular when we imagine seeing the depicted objects. But according to David Hills, mounting a muscular defence of a modified form of Walton’s account, there is no one experience we are meant to have when viewing a picture, just as there is no one experience we are meant to have when playing with a doll. In many cases, there are many equally justifiable ways of exploring the picture; this exploits the flexibility of the imagination, its ability, through variable emphasis and ordering, to do justice to the multifariousness of the legitimate experiences of a given work. Needless to say we have in this very brief introduction barely scratched the surface of these topics. Gary Kemp and Gabriele M. Mras 26 October 2015

This page intentionally left blank

Part I

Wittgenstein and seeing-as

This page intentionally left blank


The room in a view Charles Travis

A thought always contains something which reaches beyond the particular case, through which this comes to consciousness as falling under a generality. (Frege, 17 Kernsätze, Kernsatz 4) That the sun has risen is not an object which sends out rays which arrive in my eyes; is not a visible thing like the sun itself. (Frege, 1918: 61) Between 1946 and 1949 Wittgenstein produced a number of manuscripts which included discussion of seeing-as and what he called ‘seeing aspects’ – among a host of other elements in mental life (see especially 1980a, 1980b, 1982a, 1982b, 1953: IIxi). Neither is remotely the main topic. Nor is there one, conceived as some specifiable part of mental life. Nor is this discussion organised into some connected whole. Rather, it is woven through discussion of other things. No view, or account, of seeing-as is offered or endorsed. These manuscripts read more like daybooks than monographs; daily jottings, over an extended period, of daily (perhaps nocturnal) occurrences of ideas or worries. Wittgenstein himself characterises his aim as follows: The genealogy of psychological phenomena: I am aiming, not for exactness, but for surveyability. (RPP I §895 (158)) Here I will refer to these texts simply as ‘the daybooks’. One main focus will be on a particular sort of surveyability which emerges. Despite their sketchy character, the daybooks contain important lessons. Focus on seeing-as is one way to extract them. One main lesson is how accounts of mental life must, and can, respect the essential publicity of thought – the requirement that, as Frege put it, any thought must be graspable as the same by different thinkers. A related second lesson brings out disjunctivism’s point and worth, particularly in re perception. Not that Wittgenstein was a disjunctivist. But he suggests what is in fact a good reason for being one.


Charles Travis

The Tractatus’ speedy collapse Seeing-as, of one sort, is something on which the Tractatus had confidently, but surely wrongly, pronounced. That misplaced confidence is one by-product of a misplaced confidence in its general account of representation, one which, in 1929, Wittgenstein himself came to see would not do. His project of working out what would do – the rest of his career – brought him, for one thing, closer to Frege, thus to appreciate the depth and importance of what I will call ‘Frege’s challenge’. The Tractatus treatment of seeing-as is a good point of entrée to all this. Discussing the Necker cube, it says this: To perceive a complex means to perceive that its constituents are combined in such-and-such way. This perhaps explains that the figure

can be seen in two ways as a cube; and all similar phenomena. For we really see two different facts. (5.5423) This cannot be right. First, one might, notoriously, see that the Necker contains elements (lines, for example) organised thus and so, while being unable to see them as so arranged. Seeing a Necker one recognises it immediately as a Necker. Knowing his Neckers, he sees immediately that the lines are arranged so as to form an image of a cube oriented such-and-such way. Or seeing the lines and knowing his geometry, or his drawing, well enough, he knows that there must be such a cube depiction present. But today he simply cannot make that depiction come into view for him. Second, conversely, one may see the Necker’s depiction of one of its cubes – call this the A-cube – but miss the fact that the lines in the depiction are organised in some specified way. He may, for example, see the A-cube in a blur, or, while seeing the A-cube, miss a bit of organisation at the back. (He might, say, fail a quiz on this.) Or he might simply be benighted in re the organisation of a cube. Suppose there is a plastic cube – perhaps a recipe box – on the kitchen counter. One might see it without yet recognising it as a cube. One might take it for a half-cube, or recognise it as some sort of rhomboid, without yet quite seeing what sort, etc. – just as one might see a pig snuffling beneath an oak, locating a truffle, while taking it for an aardvark locating ants, or not knowing what sort of animal it might be. Recognising the box as a cube, or as a recipe box, recognising the pig as a pig, what it is doing as snuffling, or trufflehunting, and so on, are not perceptual accomplishments, or not of the sort that seeing may be, but rather exercises of capacities of thought, ones to judge correctly when it is, or when it would be, for example, a pig that one faces.

The room in a view


True, ‘see-as’ has a non-perceptual use, just as ‘see’ does (in ‘see that’). Pia may, for example, see battery power as the future of urban transportation, as she may see that the battery power in her all-electric has run out. So Pia thinks. What she thus thinks has no visual features. Seeing battery power as Pia does is not a visual experience. There is no way what she thus sees looks. Seeing the Necker as a depiction of the A-cube rather than as one of the B-cube, or as neither, is having a visual experience of a certain sort. Such is not to be understood in terms of awareness-that. Verbs such as ‘see’ and ‘look’, which may be used to speak of visual experiencing, seem pretty regularly to have other uses on which they speak, not of that, but rather of ways of standing in thought towards things. Seeing that this year will be a vintage one for port, or that it is the end of an era, or that the fix is in – even seeing that the T-bones now lie on Pia’s new white carpet – are not visual phenomena. Similarly, while for Pia to look like her mother may be for the two to share visual features (quite abstract ones, perhaps, considering Pia’s still girlish mien), for it to look as though Pia’s mother is coming to dinner, or even as though it is Pia’s mother coming down the street, is for there to be things for one, or relevantly informed ones, to think. (Contrast here it looking as though Pia’s mother is approaching with it looking (just) as though she were approaching.) ‘See-that’ speaks of awareness, but not visual awareness – even if visual awareness may sometimes be what makes what one thus realises recognisable. (Sid might also see that the Wildschwein is in Pia’s tulips again by the dismayed look on her face, or by his watch, which tells him it is Wildschwein time.) As Frege notes, that the Wildschwein is in the garden is neither in the garden nor outside the gate. It neither takes long to happen, nor is it momentary. It does not happen. It has no location. Accordingly, it forms no images on retinas. It cannot be an object of sight, nor of sensory awareness, as the Wildschwein trampling the tulips can be. Seeing the A-cube may enable one to see that the lines of the Necker are organised in such-and-such way (though there is no guarantee that it will). Seeing that the lines are organised in that way may help one to see it as a depiction of the A-cube (or to see that depiction in it). But each is a fundamentally different form, or notion, of awareness. The first makes one aware, the second registers what is thus on offer. As each serves a very different function, so, too, each takes a very different sort of object. For good reason, neither entails the other. How could Tractarian Wittgenstein have missed this difference? Perhaps because the Tractatus account of representing-as erases it. That account begins the work: The world is everything that is the case. The world is the totality of facts, not of things. The world is determined by the facts, and by these being all the facts. (1–1.11)


Charles Travis

What is the case, for example, is that Pia is walking a pig on a leash. The world, we are told, is all of that. It is made up of whatever is so. (What, then, might one represent as being thus and so?) The world is, so far, a notion open to be filled in in various ways. But it is plausibly at least that of which we judge in judging either truly or falsely – in judging, that is, that things are thus and so, which makes what we judge and what we judge of at least (perhaps only) superficially the same. Sid judges that Pia is walking a pig on a leash, and that very same thing (that she is leading a pig on a leash) is so (or again not so). One judges that in being all that is the case, the world is, inter alia, this, or perhaps of something in the world that it is that very thing, that Pia is walking, etc. The world, we are told, is everything that is the case. But it is also (one would have thought) everything one judges to be thus and so in judging something to be the case. One smells a regress. Are we at the bottom of the garden path? Have we, to borrow Austin’s image, just nutmegged our own defender, to dribble up smartly before our own goal? It would seem, anyway, that something is already badly wrong. For one thing, as Frege points out, that such-and-such – what is the case – is something invisible, not a possible object of perceptual awareness, whereas we had all hoped that what we judge of might, at least sometimes, be something which we see. But perhaps more needs to be said as to why that such-and-such must be invisible. Perhaps there is only a seeming issue here. In any case, that suggestion that there is just one thing, or category of thing, which is both what we represent as such-and-such when we represent the world as a certain way and what we represent it as, or the representation of it as that, is quite deliberate. For it serves an essential function in the Tractatus story of what representing is. That story of representing begins in section 2 as follows: We make ourselves pictures of the facts. (2.1) The picture is a model of reality. (2.12) The elements in the picture correspond to the objects. (2.13) The elements of the picture stand in, in the picture, for the objects. (2.131, my italics) The picture consists in the fact that its elements are combined with one another in a determinate way. (2.14) That the elements in the picture relate to each other in a certain manner and way represents it as being so that the things so relate to one another. (2.15)

The room in a view


To represent things as being a certain way is thus to depict them in a certain way, on the above model of depiction. A given way for representing things as being is such a depiction. It is thus clear enough what representing truly must be: What the picture represents is its sense. (2.221) Its truth or falsity consists in the agreement, or disagreement, of its sense with reality. (2.222) In the picture, stand-ins for – representatives of – given objects (whatever these are) are structured in a given way. Such representatives are elements in the picture. In it they relate to each other in given ways. If the picture is true (depicts truly), then what those stand-ins stand in for also so relate. Wittgenstein seems to have forgotten here that very elementary point which Frege puts as follows: ‘Clearly one would not call a picture true if an intention did not attach to it’ (1918: 59). One need not take Frege’s term ‘intention’ (‘Absicht’) au pied de la lettre. The point is that, for example, visible arrangements of elements in a depiction of something, if they represent things as any way at all, do so only on a particular understanding which, if that is what they do, they bear. Lichtenstein’s Three Views of Rouen Cathedral is not to be understood as depicting things as such that in Rouen everything, sky included, is, or looks as if, coloured in three-colourdots fashion, though some identical arrangement of dots on canvas might have been understood in this way. (One point stressed by Wittgenstein after the watershed of 1929 is that any picture admits of interpretation.) Bracketing this last point, the Tractatus now quite properly notes that if such is what truth is to be, a condition is thereby imposed on relevant depictings: the structure they give to their elements must be a structure which could occur in what they represent. For which, those depictings and what they depict must both belong to the same category of things, a category of things so structurable. In order to be a picture, a fact must have something in common with what it depicts. (2.16) There must be something identical in order for the one to be a picture of the other at all. (2.161) The idea that the depiction and the depicted are to share a structure must be an idea which at least makes sense: both depiction and depicted must be


Charles Travis

so structurable. But it is just this condition which cannot be satisfied by representing-as. Here is a short account of why. Generality Let grammar be our guide. Suppose that something, A, is, or was, or can be represented as being, B. For example, Sid may represent Pia (to himself, say) as (then) walking her pig in Jezus-Eik. Then something, A, is represented as (being) something; something, B, is what A is thus represented as being. B is then a particular way for something (such as A) to be; A is a particular thing liable to be that way. In such representing A is related to B in a particular way. What way? For a start, for A to be as thus represented is for it, in being as it is, to be the way in question. As one may also put it, A’s being as it is is then a particular case of something being that way, of a thing (being) B – for example, of a thing being such as to be walking a pig in Jezus-Eik. In A we thus find a case of something being a way (B) for something to be. In those two italicised ideas a crucial distinction lies. To begin with the second, a way for something to be is at least a way there might have been for things to be even if things had not been just as they are. If walking a pig in Jezus-Eik is something Pia might intelligibly do, or be thought to do, then it is something she might achieve, or count as achieving, in many ways. It being Jezus-Eik in which she walked her pig, she chose her dress carefully. In Jezus-Eik one does not want to appear a frump – as she did not. But suppose her sense of irony had failed her: she donned a dirndl. Things would then have been different than they were. For all of which she would still be walking her pig. Similarly, she might still have been doing so had the pig’s leash been alligator and not, as it happened, calf. Similarly, she might still have been walking a pig had she done so past empty terrasses. There are indefinitely many ways to walk a pig, even in Jezus-Eik. There are equally many ways of failing to be doing that. For any way for an object to be, one at least of these features of walking a pig in Jezus-Eik holds: many ways of being that way, or many ways of not. Frege puts the point rather elegantly, as follows: A thought always contains something which reaches out beyond the particular case, by means of which this is presented to consciousness as falling under a generality. (1983: 189) But we need not follow Frege here. We need only follow grammar. A way for something to be generalises in a given way over particular cases so as to reach some and/or to miss others, depending in each case on how that case is. A particular case (what is liable to fall under the relevant generalisation, or not) has no such generality. Its being as it is, even if such is in fact for it to

The room in a view


be a case of something being such-and-such way (walking a pig, say) settles nothing as to what else might be one. That what Pia’s leash leashes is a pig decides nothing as to when there would be another. Frege speaks of a thought as containing something which reaches beyond the particular case, which makes that thought represent something as falling under a generality. So far, it is a way for something to be, and not a thought, which has intrinsic generality. But a thought (in Frege’s sense of something to be thought) represents something as some way for it to be. It thus puts a way for things to be to particular work. In representing as it does, a thought makes truth turn in that thought’s own proprietary way on how things are. In representing something as some way for it to be, the thought makes its truth turn on whether that something instances being that way. The thought thus inherits the generality intrinsic to that way for something to be: there is a range of cases in which the thought would be true (casu quo false). These are just those in which relevant things instance the way for something to be which it puts to the work just sketched. Invisibility Generalising is not visible, not the sort of thing to be an object of perceptual awareness. One can grasp how walking a pig generalises over particular cases; how the relevant generalisation would capture cases would miss cases. But one cannot literally see it doing so. One may see how it generalises. But such is an achievement of thought, not vision. By contrast, some particular cases, such as Pia doing what she now is, are visible, at least to an extent, whereas there is nothing to grasp as to how they generalise because they do not. Thus a thought is invisible, not open to sensory observation. Frege, though, has another point in mind, worth noting here. His main point, in both his above-cited remarks, is that a thought must be something invisible; as he puts it in 1918, that ‘all perceptually observable things are excluded from the domain of things by which truth can come into question at all’. For Frege, a thought is, or identifies, just that by which truth is brought into question. How to understand that idea? For Frege, the thought (der Gedanke) is a particular sort of abstraction, designed for a particular purpose. It abstracts from particular instances of representing-as: episodes of authoring such representing, stances in which one represents things to himself in this way. What a thought is meant to capture is a particular question of truth, a particular way to represent things, either truly or falsely, as being: for each such question, its proprietary thought; for each thought, its proprietary question. If thoughts are so to be counted – when one thought twice, when two each once – then a thought is to be identified as the thought it is only by the presence or absence of features which matter to when it would be true, to what would be a case of things being as it represents things. For any identifying feature of a thought, there are thoughts which have it and ones which do

10 Charles Travis not; each sort of thought differs accordingly in re when it would be true. Now the point about invisibility: a visible (perceptually observable) feature of a thought would not matter intrinsically to how it represented things as being. For any visible form a thought might have – as, for example, a sentence may have a spelling – a thought which had that form and a thought which lacked it might for all that represent things as being exactly the same way. Hence a thought can have no visible form. It is just that by which truth can come into question. Therefore there is nothing a thought looks or sounds like. Structurability A thought is a very different sort of abstraction from historical cases of representing-as than is a sentence. One way in which it is has just been noted. But there is a more fundamental point. It lies in Frege’s insistence on placing whole thoughts first (before proper thought-elements) in point of ontology: the idea of a proper thought-element, as Frege saw, is understandable only in terms of a role within a whole thought. A whole thought makes truth outright turn on how things are. This means: whether, in representing something as being as a whole thought represents what it does, one represents truly or falsely cannot depend on which things one so represents. A whole thought may represent things as such that Sid is stoking up his smoker. There are no two thoughts that Sid is stoking up his smoker which differ in who it is who must be such that Sid is stoking up his stoker. Such it is for there to be a way for things to be such that to represent things as that way is ipso facto to represent truly or falsely. One might think of such a way as a way for an n-tuple of things to be (n = 0, for short, a zero-place way). It is part of what it is to be true or false outright that anything which is this represents the same thing as some zero-place way (different ways for different things to thus represent as so). The Tractatus calls this thing Wirklichkeit. Here I will just call it ‘things’, understood so as to bar questions of which ones. A whole thought is decomposable into elements. The thought that Sid is barbecuing ribs is decomposable into an element which makes it about Sid, and another which makes it about barbecuing ribs –an object and a way for an object to be respectively. The whole thought makes truth turn in a certain way on how things are, as is done in representing things as a zero-place way. A thought-element is a partial doing of what the whole thought does. It makes that thought’s truth turn in part on such-and-such – for example, on how Sid is, or on who smokes. For there to be a decomposition is for its elements to do jointly precisely what the whole thought does: that is, make truth outright turn on just what that whole thought does. Any decomposition of a thought assigns the thought some given structure, one defined in terms of the relations between its elements on that decomposition. As Frege notes, there is no need for a thought to have any one decomposition,

The room in a view


nor for there to be any one structure which it would be assigned on all the decompositions it admits of. At the very least, in Frege’s view, such is not the norm. So, while thoughts are structurable (and it is thereby that they enter into the logical relations that they do), they do not have any one intrinsic structure in the sense in which a sentence of a language does. A sentence – of English, say – is generated by the syntax of its language from the relevant vocabulary. The way in which it is thus generated imposes a certain structure on it, which is its absolutely. There is no counterpart to that when it comes to thoughts. Some philosophers, such as Russell, rejecting Frege’s notion of a thought, prefer the term ‘proposition’ for what, in their scheme of things, is that by which truth is brought into question at all. For early Wittgenstein the relevant term was Satz. But ‘proposition’ and ‘Satz’, at least on such philosophical uses, tend to work as waffle words, hovering, or shifting, between something thought-like – something restricted to the task of bringing truth into question at all – and something sentence-like, between something incapable of having a visible form and something to which it may be intrinsic to have one. Perhaps such waffling helps explain the work to which the Tractatus puts the notion ‘picture’ (Bild) in its account of what representing is. Collapse So far, we have identified four ingredients in representing-as: 1

2 3 4

There is, in a given case, that which is represented as some way or other: in a non-zero-place view of the matter, often, though not always, some n-tuple of objects, n ≥ 1 (which n-tuple depending on what non-zero-place view one takes). ‘Object’ is here used in its logical sense: something over which a way for things to be may generalise. This includes, but is not exhausted by, things to collide with, also historical events or episodes, and more. Particular cases of some n-tuple of objects being as they are (n ≥ 0). Ways for an n-tuple of things to be (n ≥ 0) (things for a particular case to instance). Representations of things as some way they may be or not, where in so relating the represented to a way to represent it truth outright comes into question (representations which are thoughts).

Points 3 and 4, we have seen, have a certain sort of intrinsic generality. They range, each in its proprietary way, over particular cases. Points 1 and 2 lack such generality. The Tractatus presents a certain picture of how the represented (in 1) and the representer (in 4) relate. In it the representer represents something which is its structure as the, or a, structure of the represented. What truth thus requires is some identity of structure in some item in 1 and some item in 4. To what sort of structure, or structuring, then, is each susceptible? We can

12 Charles Travis begin with 4. An item in 4, a thought, is decomposable into elements, each of which does part of what that whole does: fixing a particular question of truth, making truth turn in a particular way on how things are. To decompose a thought in some given way is to assign it some particular structure: that is, to identify some particular (non-unique) way in which it is structurable. In the decomposition the elements of the thought jointly do (precisely and only) what the whole thought does. They thus stand to one another, in that decomposition, in particular relations defined in terms of how, in it, they interact with one another. If an element in the thought that Sid smokes, for example, makes truth turn on how Sid is, another element, making truth turn on who smokes, may relate to this first in fixing in what way how Sid is matters. Such relations are ones defined in terms of truth, equivalently, being as represented. A representation, understood as in 4, can have no visual, or perceivable, form. It can have no spatial location or temporal ordering. It cannot be a term in causal chains. Nor can its elements. So these cannot stand in relations defined in such terms. We turn now to the represented, items of type 1. We can start with the zero-place case: things (as they are). By bringing this under generalisations of various sorts we can identify in this various spatially locatable things, each in the stage of its career it is, various historical episodes or occurrences taking place, or having done, and so on. These relate to each other in manifold ways. As it may be, for example, Sid may just have put Pia’s wallaby steak on the grill. Here, then, is a quadruple of objects which relate to each other spatially, causally, temporally (for example, Pia predates the steak) and so on. The quadruple, and thereby things (in my present sense), are structured accordingly. They are thus structured by structuring to which representations (in the sense of category 4) are ineligible. Conversely, they are ineligible to be structured by those relations, defined in terms of truth, by which a thought’s elements stand to one another in some decomposition of it, relations which hold in virtue of those partial doings by those elements of what that whole thought does. An element of a thought is a partial doing of what the whole thought does, whereas neither that wallaby steak, nor Sid, as he plops it on the grill, makes truth (of what?) depend on anything. There is thus no question of a representer and the represented (what it represents as being thus and so) sharing a structure. Each is eligible only for structuring for which the other is not. Asking whether a thought that Sid is grilling wallaby shares a structure with Sid’s grilling of it, or any other bit of history, is like asking which is higher, Sid after his last toke, the Matterhorn or the Queen of the Night’s high C. The Tractatus aims for the identity it posits by insisting that both representer (what it calls a ‘picture’) and represented are ‘facts’. There are various notions of fact. I omit this exploration here. But for the ploy to do the trick, a fact, on the relevant notion, would have to have the generality of a thought, or a way for things to be – actually do the generalising which representing does. And it would have to have that generality on both sides of the

The room in a view


representing-as relation, occurring as the represented as well as occurring as the representer. The represented would then not be a particular case over which representer might generalise – not something for it either to capture or to miss. If one were to represent the fact that Sid is barbecuing wallaby as anything, it would be, for example, as pregnant with implications, or as making him in violation of the law. Or, again, the fact of his barbecuing wallaby may be a fact of law breaking. What we wanted to do, though, was to represent Sid, or what he is doing, as barbecuing wallaby. On the ploy so understood, such is not in the cards. Moreover, it cannot be that all representing is simply representing some generalisations as falling under others. There are no generalisations so to represent unless there are particular cases, devoid of generality, to represent as falling under at least some. The Tractatus loses sight of this. Suppose one modelled a Tractarian Bild on the sort of thing one hangs on a wall – something concrete, replete with visual properties. Now there is generality on neither side of the equation. Frege’s insight about the truth of pictures (modulo the term ‘intention’) is overlooked. One may use such a picture in an act of representing to help achieve his representational ends. He may, for example, make the picture a reference for things being as he represents them: for the cathedral to be as represented, it must look like this (see picture). But to do this he must deploy a generalisation: a particular understanding of that deployed notion, ‘looking like this’, one which captures some determinate range of particular cases, rejects some range of others. The picture may be visible; the generalising thus done is not. Nor, then, is the way the thought thus expressed brings truth into question. Representations, where they are that by which a question of truth arises, remain items in category 4. The present point stands. Objects of visual awareness – a given Necker, say – belong to the represented, to what there is to present as falling under one or another generality. They themselves are devoid of generality. If what there was to represent as such-and-such were itself a fact, and if, consistent with that, we could conceive of a fact as something which is the case – the fact that such-and-such – then the Tractatus account of the Necker might be available. But hélas. Seeing the Necker as an A-cube (or seeing the A-cube in it) cannot be seeing that suchand-such. What might it be? Seeing-as, if (where) a visual experience, as in the Necker, one of things looking a certain way, it must be of the represented, something to instance relevant generalities. Its objects cannot be contents for a representing-as. What is it, then? One question for the daybooks.

Frege’s challenge A thought is just ‘that by which truth can come into question at all’. One thought is distinguished from another only by what matters to whether things are as it represents them. Equally central to what a thought is is its publicity. For there to be a thought is for it to be, in Frege’s words, ‘graspable

14 Charles Travis as the same by different thinkers’. Such belongs to the objectivity of truth. There is, thus, no-one one must be to grasp how any given thought makes truth turn on how things are and hence to recognise which particular cases would count, which not, as things being as the thought represents them. The truth of a thought can only turn on how things stand in that shared environment we all encounter. Things being as they are must consist in no more than things for one to encounter being as they are. Call this principle ‘Publicity’. The present point is not to argue for it, only to consider its significance for what mental life might be. Frege uses the word Vorstellung in a proprietary sense. The most crucial features of a Vorstellung are, first, that it requires a bearer, and, second, that it brooks no two. One has Vorstellungen, Frege insists, one does not perceive them. Wherever there is a Vorstellung, there is someone one would need to be to have it. It is not something for one to encounter. It is thus, by Publicity, not something on which the truth of a thought can turn. Which must be understood as follows. For any given thought, our shared environment’s being as it is – what there is for one to encounter – cannot leave it open for whether that thought is true or false to turn on something extra to this – so not on how some Vorstellung is, if such is extra to this. Publicity needs that understanding, for Frege does not deny that people have Vorstellungen. Indeed, he introduces the notion with a list of types of such case: sense impressions, sensations, feelings, moods, inclinations and wishes (1918: 66). Pia bites down on an overripe cherry and tastes the mould. You would have to be her to do that biting. You would have to be her to do that gustatory experiencing. There may be something that experiencing was like. Such would be what it was like for Pia. For the experience to be what it thus was would be a case of someone (namely, Pia) having a Vorstellung. The idea so far poses no problem for Publicity. For that instance of someone having a Vorstellung is just a particular case, a particular episode – Pia’s being as she then was. There is as yet no generalising over such cases, no thought, to which the Vorstellung Pia had, or how it is (if we may so speak) is to matter. There is no thought – say, that things then tasted to Pia thus and so – whose truth is to turn on how that Vorstellung (in fact, anyway) is. What Publicity tells us is that we cannot conceive of a thought’s truth as so turning, as to be decided by ‘the Vorstellung Pia had being as it was’, or, anyway, not if ‘how that Vorstellung was’ is extra to the shared environment’s being as it is. There are no generalisations over Vorstellungen. Here there are no ways of reaching beyond the particular case so as to capture certain further such cases, miss certain others. A Vorstellung is thus not an object as objects figure in the logical. It is not something on which truth can be made to turn as, in a singular thought, truth turns on how some given object is. It is not something liable to be, just in being as it is, given ways there are for a thing to be. It is not what any object must be, something for which there are (sufficient) definite answers to questions whether a thing is that one. It does not participate in the relation ‘identity’.

The room in a view


But while there are no generalisations to be made over Vorstellungen as such, someone’s having a Vorstellung on an occasion is something over which one may generalise. One may do that in generalising over particular cases of people being as they (then) are. Someone’s having the Vorstellung he did on an occasion may be a case of a person being thus and so. Pia’s having the Vorstellung she did in biting the cherry may, for example, be a case of someone tasting mould. It is just that if there is such a thing as tasting mould, then whether Pia is thus a case of this or not had better not be left open by the environment’s being as it is to be settled by something extra-environmental. Sid’s toast is burning. An acrid scent fills the room. Whether this is so is something we can jointly investigate, or dispute. We can thus agree on what it would be for it to be so. We can recognise of particular cases whether these are what would count as a room being filled with acrid scent. Such is not left open by the room’s, or the environment’s, being as it is, to be settled by what various Vorstellungen are like. The same, Publicity tells us, must go for the room smelling acrid to Sid, or for the cherry tasting mouldy to Pia. The question is how to think of mental life so as to respect Publicity’s demands. If we suppose this to be a main concern of the daybooks, we can then understand something as to why they take the form they do, covering their multitude of topics, often breaking off discussion of one such topic and switching to another. Different cases call for different ways of respecting Publicity – though, as we shall see, there are common threads. Speculatively, I introduce this as a second guiding motive in Wittgenstein’s discussion of seeing-as.

Visual ambiguity Now for seeing-as. Unsurprisingly, ‘see-as’, like ‘see’ itself, acquires nonperceptual uses. Pia (still the ingénue) sees it as an honour to be offered the headship. Sid sees Alf as a rival and a threat. In such uses ‘see-as’ speaks of a way of thinking of things, not a visual experience. There is no particular way being offered the headship must look. Alf may be seen by Sid as all the more a threat for not looking like one, precisely because of his mild and unthreatening manner. Our aim here is to set such uses of ‘see-as’ aside – when possible. We must, further, approach seeing-as, even where a visual phenomenon, in a Butlerian spirit: it is what it is, and not another thing. For example, seeing-as must be kept apart from looking like. To Sid Pia may look just like her sister. Such is not at all for him to see her as her sister (though such may be possible too). Various distinguishable visual phenomena all fall under the rubric seeing-as. This section concerns just one. It is represented in the daybooks by such examples as the Necker, the duck–rabbit and the white cross on black background/black cross on white. In these cases seeing-as is simply a certain sort of case of seeing. The cases highlight one feature of it. They exhibit what one might call ‘visual ambiguity’. The Necker, for example, does depict a cube in one orientation (call this the A-cube) and a cube in another (call this

16 Charles Travis the B-cube). The A-cube (depiction), the B-cube (depiction) and the Necker itself are three different objects of sight. The peculiarity is that (bracketing the Necker itself) one can see one of these objects only at the expense of not (then) seeing the other. What one sees is different when he sees the A-cube from when he sees the B-cube, when he sees the duck in the duck–rabbit from when he sees the rabbit. What is the difference? Since what is in question throughout here is objects of sight, the difference lies, for one thing, in the difference between these objects – for example, between a cube oriented thus and a cube oriented so, or between an image of a duck and an image of a rabbit. Seeing remains what is in question throughout. One thing seen throughout is the Necker, or the duck–rabbit or the black–white cross – the visually ambiguous picture. What, then, is it to be aware of the one object of sight rather than the other? One idea would be: it is to be aware of some object of visual awareness which is a depiction of, say, the A-cube, but is not a depiction of the B-cube. Such an object of awareness would be visually unambiguous. But this unambiguous object could not be any array of lines in the Necker, or in the duck–rabbit, etc. Perhaps, then, an inner object – something for the viewer alone to be aware of, not something for one to view? Such would be (having) a Vorstellung. This idea of an inner picture is one idea Wittgenstein aims to dispose of. He writes, And certainly do not say, ‘But my visual impression is not the drawing. It is this, which I cannot show anyone.’ Of course it is not the drawing, but it is also nothing of the same category which I carry within myself. (1982a: §440) The concept of the ‘inner picture’ is misleading, for the model for this concept is the outer picture, while, however, their applications are not more closely related than those of numeral and number. Someone who wanted to call a number an ‘ideal numeral’ could instigate a similar confusion. (§442) Someone who classifies the organisation of a visual impression with shapes and colours, is thinking of a visual impression as an inner object. Of course the object becomes thereby an absurdity, a peculiar wavering picture. For the similarity with a picture is now disrupted. (§443) The idea was: when I see the depiction of the A-cube in the Necker, I am experiencing, visually, an unambiguous image of an A-cube; but since there is no such image before my eyes, the image in question here, or having it, is a Vorstellung. One thing wrong with this idea is: such an image, if it were something to be experienced visually, could not look like the Necker – could

The room in a view


not, for example, be lines so arranged – for then it would be ambiguous. On the other hand, it must look like the Necker, since the Necker is what I see, and, we may suppose, I do not see it distortedly. And it is the Necker which looks like a drawing of an A-cube when I so see it. For me to have that inner image was for the Necker so to look to me. Further, the idea violates Publicity. Suppose we said: so much the worse for Publicity. Now we are thinking of my inner image (at that moment), or my having it as a particular case which, in fact, instances certain generalities: it is a case of such-and-such an array of lines, one thing there is for an image to be. What remains for me to do, then, is to recognise the image’s instancing of given generalities, thus getting things right which I am liable to get right or wrong, and still so liable even if I grasp perfectly well what it would be for an image to instance the relevant generalities. I may mistake an outer image for one with a line running through it at an angle so, where no line so runs. I may not have looked closely, or slowly, enough, or things may have gone all blurry for me at the crucial moment. To paraphrase Frege (1918: 73), by the step by which we make a Vorstellung – here an inner image – into something which is as it is anyway independent of my recognising its so being, we expose me (its possessor) to liability to error. It is now possible for an inner image to appear to me as it is not. Which, if inner images are what explain here, requires a further inner image. The inner image now shows up as simply no help. What now? Here is an idea. In seeing, there is a scene in view (generally literally before the eyes). What one sees are various objects, or objects of sight, in that scene. But there may be some such object, in plain view, before the eyes – visible – which, for all that, one fails to see. There is, in a scene in view, that which one saw and that which he missed. In an experience of seeing, things are provided for one to be aware of (visually); but for one to see one of these he must respond suitably, or be suitably responsive, to it. Sight affords awareness of visible things in the scene in view; to see one of these is to enjoy (well enough) the visual awareness thus on offer, which is to be suitably responsive to it. In the case of the Necker (or any other case of visual ambiguity) one sees, at one moment, one thing to be seen in it – say, its depiction of the A-cube, and, at another moment, not that, but rather its depiction of the B-cube (or neither). The idea of an inner image was an attempt to locate the difference between seeing and not seeing in such a case in what there was, on the occasion, for the viewer to be visually aware of. That attempt came to grief. Perhaps, then, a better idea is to locate the difference between seeing and not seeing in responsiveness to what there is to see. To elaborate this idea we need to say more about the genre of awareness to which seeing, visual awareness, belongs. We have already noted that this is a different genus from that to which awareness-that belongs. Grammar marks this in the distinctions it draws between the objects of one genus and the objects of another. We have seen one reason for so doing: this marks the distance between the near and the far side of the representing-as relation.

18 Charles Travis There is also another crucial, and, I think, correlative, difference. Visual awareness (and all awareness of its genre) is essentially informative. It reveals how things are, keeps us abreast. We learn, or confirm or investigate things by looking. Awareness-that is not informative. It is a form of registering the sort of thing informative awareness may make recognisable. To see something is to gain acquaintance with it. It is to be positioned to recognise objects of informative awareness as cases of things being ways for them to be. It is to be granted a source of authority as to how things are, to enjoy such authority. How much good this does a given viewer on an occasion depends, for one thing, on that viewer’s conceptual capacities. One is positioned to recognise the object of sight – say, Sid, the burnt toast or some Necker’s depiction of an A-cube – as being whatever ways it can be told by sight to be. Which ways a given viewer can recognise the object to be depends on his particular capacities to tell such things by sight. For another, if seeing is a source of authority as to how things are, it is also a source of illusion of authority. One may think he has seen a smudge in the upper left corner of the facing surface of a certain A-cube, or that Sid’s toast is wholewheat, where nothing of the sort is so. Still, seeing places one in an authoritative position in re how its objects are. To have seen rather than to have failed to see, or to have missed, something visible, in view, the idea is to enjoy enough of this authority, to be made sufficiently authoritative. The A-cube and the B-cube are different objects of sight. Seeing the Necker as an A-cube is just seeing the A-cube (depiction) and not the B-cube. It is thus a source of authority as to how that A-cube, but not as to how that B-cube is. It is distinguished from seeing the other by the authority thus enjoyed. Authority as to how the A-cube depiction is is being placed to answer such questions as whether the lines of its facing surface are smudged or broken. Mutatis mutandis for the B-cube. When you see the B-cube, you may, if you know enough about Neckers, know some things as to how the A-cube must be. Perhaps you can work out that its facing surface must be smudged. But such is a different sort of authority as to how it is. It is not gained by sight as your authority as to the B-cube then is. Whether one sees the A-cube, the B-cube or neither is not independent of whether one takes up – responds to – the opportunities which would thus be put on offer (in present terms, whether one exploits the source – is made authoritative). Seeing the A-cube is a visual experience, hence not seeing that. But for one to be seeing it is for him to be so enabled by sight to see (recognise) things visibly so of it. Such is one thing (only) which authority might be.

Two forms of authority The authority sight may offer over what Sid was up to on the evening of the 17th contrasts with the authority I have over this as author of the story of Sid’s doings on the 17th. There is authority as to such-and-such, and authority over such-and-such, which I will name expert authority and executive

The room in a view


authority respectively. Executive authority is creative. Sid was in the pub with the regulars on the evening of the 17th if that is how I tell the story. Expert authority is not. An expert is one who knows the right answers to questions which have answers. Executive authority can make an answer right. Both sorts of authority play a role in meeting Frege’s challenge. In the case of seeing, expert authority helps meet the challenge: conceiving seeing in terms of it removes the need for inner pictures. This section enlarges on each form and on the contrast. Seeing is a source of expert authority. To see such-and-such is to be placed – equipped – to answers questions as to how what is thus seen is – authoritative as to how those things are. Seeing that pig lying in the straw gains one authority as to how it is. In exploiting what is thus at one’s disposal, one might then, for example, register that object (in fact, the pig) as present in the scene, or as positioned in it thus, or as lying in that straw, or as a pig. One might recognise it to be thus present, or lying on straw, or a pig, given suitable abilities to recognise (by sight) what was a case of something lying on straw as that. Seeing the pig is, inter alia, being furnished with, or enjoying, something so exploitable. Sight makes the visible features of a scene in view (or some of them) thus exploitable in recognising such things about it – it, or things in it, being various ways they are. Seeing enables one thus to exploit such features in, and through, visual awareness of them. In makes one so authoritative as to how things seen in fact are. It is a means by which such visible features can become proof for one that things are thus and so. Pia has a certain scene in view. Suppose there is something in the scene for one to see, and a question whether Pia saw or missed it. Suppose that through sight she was able to register enough of how that item in fact visibly was – made her sufficiently authoritative as to this, granted her expert authority. Then, the idea is, the question is answered. For Pia to have failed to see what was visible – for example, the pig – would be for sight to have failed to make her sufficiently authoritative, for her to fail to be furnished the relevant things to register. This idea leaves it open just how authoritative sight would need to make her in order for her to count as having seen that thing – as it should. The idea, in very broad strokes, is: for us, thinking beings with the capacity of sight, seeing rather than missing something in sight is being enabled, by visual awareness of their being as they are, to register (enough of) how that thing, and/or its surroundings in fact are, what is happening to, or through, them. That capacity – sight – would enable registering enough of what is, in fact, made recognisable, in the circumstances, by sight. Awareness that (for example, that that pig is sleeping) is one form of registering how things are. That it is registering something distinguishes it in one fundamental way from visual awareness, a form of informative awareness. Visual awareness is, intrinsically, what enables such registering. It is a means by which a scene’s being (visibly) as it is may be registered as a particular case of things being various ways for things to be, particular ways things thus are. If seeing the pig, rather than missing it, should require registering

20 Charles Travis that that pig is at least some ways it is, such is not a psychological hypothesis. There need be no route from visual awareness to awareness-that which calls out for psychological explanation. The point would rather be: to count as having seen, rather than missed, the pig, one must have registered, or been enabled to register, what is decomposable into, inter alia, that the pig is some such ways. The psychological innocence of the conceptual point here stands out in the following possible way of understanding awareness-that. One might think of stances, or standings, in which one relates to thoughts – for example, being aware that such-such-and-such, taking it that such-and-such – on the model of Frege’s idea: whole thoughts first. In that idea, a whole thought is what brings truth into question. A whole thought is decomposable – structurable – into elements. But then it is decomposable, so structurable, in many ways, none, as Frege puts it, enjoying objective priority over any other. By analogy, we might think of a person, at a time, as standing towards the world in a certain way, inter alia, as taking it to be a certain way – that is, as he then takes it – as he is guided by its being in his pursuit of the thing to think or do. Such a whole stance towards things may, like a thought, be articulated into intelligible parts for the whole stance, ones of taking, or seeing, things to be this or that way. In the analogy, it is so articulable in many different ways, none able to claim objective priority in an account of what it is that person then thinks. On different occasions for ascribing thoughts to the person, his whole stance may admit of different articulations: what would be a correct articulation in some such circumstances might not be in some others. Whether, on an occasion, a given person’s whole stance is articulable in a given way would, thus, depend in substantial part on considerations which are not psychological – for example, on what would then count as two thinkers agreeing or disagreeing about something – where there is agreement, where not; and/or (closer to Frege’s heart) on for what such-and-such would, or would not, be (a) proof. A thought, Frege tells us, admits of being expressed in many ways, in any of many non-synonymous sentences. In what ways the same thought might be expressed may depend on the occasion for such expression, or its evaluation. Where it did, so too would answers to the question what it is that some given thinker thinks. Such merely sketches a way of conceiving representing things (to oneself) as so. Now is not the occasion to discuss it further. The point is just that if we think of visual awareness as a source of awareness-that, a transition from the one to the other need not be a process. Expert authority provides one way in which respect for Publicity may be achieved in certain cases. If seeing, rather than missing, is a matter of the expert authority enjoyed, then to see whether Pia saw the pig in the straw, or the pattern of the spots on its facing side, we only need see what authority she enjoys over relevant matters in and through the (visual) experiencing she is then doing. Whether her being as she is is someone being suitably authoritative is not something one would need to be Pia to recognise. (In fact, we often mistake

The room in a view


how authoritative we are.) Her so being is part of the environment, something for one to see, to investigate, to discover. What, now, of executive authority – authority over how things are? Executive authority is authority to make something the case. Where exercise of it relates one to a thought, it does not relate him in a way which is representing truly or falsely, even if, given what its exercise creates, there then is something – perhaps that very thought – to which one might stand in representing truly or falsely. Conan Doyle might make it the case that his fictional character, Holmes, lives in Baker Street. He may do so by so writing his stories. Holmes lives there in the stories. That he does is how they go. Conan Doyle’s work done, Eagleton may then announce that Holmes lives in Baker Street. What Eagleton says may be true or false. Again, an umpire in cricket might have the authority to declare someone out LBW. His so declaring makes it the case that, in the game, the man’s leg was before the wicket. We may all see that it was not (in reality). But such does not bear on the correctness of what the umpire makes so. In the sense of LBW in the game of cricket, the man’s leg was before the wicket. The issue is not to be settled by measurement (or replay). Executive authority may be limited authority without ceasing to be executive authority. As it may be, if an umpire is declared non compos mentis, or discovered to be on the take, then perhaps his rulings fall. For all of which, for a player to be out LBW may remain for him to be so declared. Or take another case. In a certain kingdom there may be a minister empowered to appoint judges. So if the minister declares that Bloggs is a judge, then so he is. There may for all that be requirements on judgeship – for example, being over eighteen. If Bloggs is not over eighteen then, ministerial edict notwithstanding, Bloggs is not a judge (similarly if the minister must be sober while declaring, or is prohibited from appointing judges on St. Mamede’s day). There may also be moot cases. Suppose all judges must be baptised. But Bloggs was ‘baptised’ by that notorious saint who also baptised the penguins, or in deuterium. Can the minister exercise his executive authority in such a case? One more general point. In public life executive authority is most familiarly exercisable in saying something. The magistrate in Pahrump declares the couple man and wife. But executive authority at work in meeting Frege’s challenge would be executive authority over one’s mental life. In this realm, at least, its proper exercise need not be an act of representing. For example, it has been proposed (I believe by Thomas Nagel) that the thing about pain is that it is awful. On a charitable enough reading, this may be so. Now Pia gives Sid an ‘affectionate’ tweak of the ear. Was it painful? That Sid said so is compatible with not. But suppose he found the tweak, or feeling thus engendered, awful; to him, it was. For him to do so is for him to hold a stance. There is the inevitable (grammatical) slippage between stances and acts, thus stances and saying something. Calling it awful might be mere petulance. It need not settle anything. For all of which, it may remain pertinent to what being painful is that a sufferer enjoys some measure of executive

22 Charles Travis authority over this. Frege’s challenge is now met if someone’s holding the required stance – finding-awful – is itself an environmental matter, one of his being as he might be encountered being. As Frege conceives mental life, there are two things that must be squared. On the one hand, as he conceives things, people do, sometimes, have Vorstellungen. On the other, a Vorstellung’s being as it is, so far as that goes, is not a particular case over which a way for an item to be might generalise. Someone’s having a Vorstellung which is painful, or convex, or magenta, or just as though he faced a chilidog, cannot be something extra to his being that part of the (our) environment that he is. Executive authority offers one way, among others, for both things to be so. One may see a chilidog on his plate, or a purple tomato on the sideboard. But sometimes, though it looks to him just as though he were doing that, he is not. What he experiences visually is, for him, just as though he were experiencing a chilidog on a plate; but he is merely seeing a ringer-chilidog, or experiencing a ringer for seeing one. Where ringership pertains to what one would need to be him to experience – as it sometimes may – it can, perhaps must, be the work of executive authority. It is, or can be, by such authority that what he experienced was just as though there were a chilidog on his plate, or a purple tomato, or a photic flash. Something – for example, some painted plaster, might look just as though it were a chilidog. For it to do so is for it in fact to fall under a certain generality, be one case of something so doing. Sometimes it is merely a matter of it being for someone as though he were viewing a chilidog on a plate. Here something falls under a generality. But what does so is him. His being as he is is a case of it being for someone as though he viewed a chilidog. What makes this so need only be executive authority. It may well take executive authority to do this. Expert authority and executive authority exclude each other. There is room for exercising executive authority only in matters over which there can be no expert authority – notably, matters of which way some Vorstellung is/was, where such is a fact extra to the environment being as it is. Conversely, in matters to be settled by expert authority, such as what is seen, executive authority is unavailable, and would be otiose. Insofar as (a) visual experience is seeing, it is for expert authority to select, from what there was to see, what was. The ‘phenomenal character’ of the experience, insofar as it is one of seeing, just consists in it being one of seeing that. Executive authority may provide for further facts as to how things were, visually, for the viewer. But it is under no obligation to do so. Its exercise, if possible at all, is optional. We now come to a second variety of seeing-as.

Things not in view Remarks on the Philosophy of Psychology Volume 1 (1980a) begins with this example:

The room in a view


The idea is: this can be seen as an F, or as a mirror-image F (or as a fusion of both, or as a letter in a strange alphabet, or (to an American) as a T written (imagined) European style. This example fits with a number of others: seeing a swastika in a window’s muntins, seeing a line of dots as a line of pairs (or triples), seeing a triangle with one angle rather than the others as the apex, seeing a figure – a vertical line inserted in a V – sometimes as an arrow, sometimes as a bird’s foot. One might also add mistaking – for example, a shadow on a piece of fruit – as spoilage. Wittgenstein also offers some auditory examples – for example, hearing ‘ne. . . pas’ as a negation (rather than as ‘not a step’), hearing ‘Weiche!’ as ‘be off’ (rather than as ‘soft’ (Erda in Der Ring des Nibelungen)). One might debate whether all these examples work to present purpose. In any case, that purpose is meant to be this: these are, or can be seen as, cases not accountable for in terms of the expert authority on offer. They are thus at least ripe territory for work of executive authority. When one sees the as an F rather than as a mirror-image F, or sees the row of dots as a row of pairs, it need not be that seeing is the source of any expert authority proprietary to that way of seeing it. There is that contrast with the Necker, with which belongs another, to be kept in mind here. With the Necker, unlike with most other visually ambiguous figures, to see it as the one thing rather than as the other is literally to see something different. The A-cube is visible, for example, only at the expense of the B-cube’s not being. (As Wittgenstein notes, the Necker is thus related to visual illusions in a way in which other cases (he suggests the black cross/white cross) are not (vide 1982a: §705; PI II xi: 177).) In our present class of cases, no such thing need be so. Seeing the pairs of dots is not at the expense of seeing the line. Seeing the swastika in the muntins need not be at the expense of seeing the cross. It might be that someone cannot find the swastika in the muntins. This might be described as his failing to see it (an application of the idea of missing rather than seeing). But to find it is just to isolate it in what he already does see – again, by contrast with the Necker. Our present class are all cases in which seeing something as one thing or another at least resembles thinking of it in one way rather than another. In thought one organises the dots into pairs. If no more is involved, we have here that non-perceptual use of ‘see-as’ we are trying to avoid (seeing Alf as a threat). But that seeing-as in such cases resembles thinking, or even involves it (if it does), is not yet to say that it simply consists in thinking of what one sees in one way rather than another. Does it just consist of this in cases of the present sort? Or do things here actually look different – have a different look – when seen in one way rather than another, much as the Necker or duck–rabbit? How is this issue to be decided? One might appeal to introspection. In which case, for the very little it is worth, I would be inclined to say that the deviant F does not look different when I see it in different ways; the scalene triangle definitely does. Again, though we all know that there is a swastika

24 Charles Travis to be seen in the muntins of a four-pane window, I find that it can be difficult to make it look that way (again, in comparison with the Necker). Perhaps a number of significantly different cases are lumped together here. But what is introspection worth? Throughout the daybooks Wittgenstein repeatedly reminds us of its inaptness: Do not try to analyse the experience within yourself. (1982a: §548) The greatest danger here is wanting to observe oneself. (1982a: §459) Do not ask yourself ‘How does it work with me?’ – Ask: ‘What do I know about someone else?’ (PI II xi (176)) Our aim, or Wittgenstein’s, is to conduct a conceptual investigation. Does the line of dots really look different when you see it as a line of pairs of dots? Or might it? Do the muntins in the window really look different when you see them as a swastika? Or might they? And so on. If this is to be a conceptual investigation, then what we want to know here is whether the concept ‘looks different’ (where ‘looks’ speaks of a perceptual phenomenon) fits such cases, and similarly for related concepts – for example, whether one’s experience is visually different in seeing the muntins as a swastika from in not so seeing them. How, though, is such an investigation to be conducted? Suppose we want to know whether it is possible for something to count as a chair without counting as an artefact for sitting on. Or we want to know whether in order to believe there is a swastika in the muntins one must know what it is for something to be a muntin. Then we can construct imaginary cases. For example, there is the giant chair/‘chair’ outside the chair factory in Boxtel. A monument. Perhaps also a chair (though much too large to sit in)? Or we consider the case of someone who can point to the place where a swastika should be, can draw it, but does not know her muntins from her mullions. And so on. But what are we supposed to point to here? What are our cases of what may or may not be ‘actually looking different’? What are we all supposed to agree to as to what would, or would not, count as ‘actually looking different’? (Contrast, in Austin’s case, trying to establish whether the wax tomato actually does or does not look different from the actual model for it.) But we can turn our question around. A few paragraphs back I waxed briefly autobiographical. Others might do the same. Some might agree with me on cases, some might not, some might just be at a loss. Suppose those who find the actually to look different when seen as a mirror image (I am not among these) are taken at their word on this. What sort of conceptual

The room in a view


structuring would then apply to them, or to their experience? What would one expect to be so if their experience ‘really’ were as they claimed, if things really were such as for things in view to count as looking different to them? What is it that would count? To describe someone as having experienced things changing their look as he came to see the swastika, or experiencing the window looking different when he sees the swastika from when he does not, is to bring him under a concept which behaves in a given way. If we have settled as much as matters to us what that way is, and a question remains open as to what our informants really experienced, we may anyway reckon that, qua philosophers, our job is done. What remains unanswered at this point can simply remain moot. Or perhaps to that extent one is free to apply the relevant conceptual structure, or refrain from this. Wittgenstein also tells us that seeing is a malleable notion: The concept of a representation of what is seen, of a copy, is very plastic, and with this so is the concept of what is seen. (1982a: §446) There need not be any unique thing which is the right answer to the question of what someone saw when he looked at the window, or came to see when the swastika came into view for him, or to how things looked for him while the swastika was salient. It need not even be, apart from an occasion for applying concepts, that things did look different to him then than they did to someone who did not see the swastika, or to him when he did not see it. Our present task is not to decide that issue, nor, a fortiori, to try to say what looked different to what, tertium non datur, but rather to see what follows if it did look different, or what we would be presupposing in taking informants who insisted so at their word. In introducing the present class of cases I supposed that these are distinguished from ones such as the Necker or the duck–rabbit in that here seeing X in one way rather than another need have no impact on the expert authority thus on offer. By contrast with the Necker, seeing the row of dots in one way rather than another does not per se furnish a different source, nor different objects, of expert authority. It may merely be to notice something one had not yet noticed (as, for example, with the swastika). Suppose that, nonetheless, we continue to suppose the subject of an experience within this class to be especially authoritative, in some way or other, as to what it was that was seen, or how things looked. If we are to respect Publicity, we cannot suppose that such special authority consists in having a special source of information as to how things were, certainly not as to how things were before the eyes, since no such source is on offer in his seeing what he does, and certainly not as to what was not before the eyes, vide Publicity. What remains for this new authority to be is thus executive authority. When Pia saw the as a mirror-image F rather than an F, did things look different to her? Suppose she were familiar with the , but this was the first

26 Charles Travis time she had noticed how, thus that, it could be a mirror-image F, too. Then she gained greater expert authority from a source already on offer. But, as we have already seen, such would not be what it was for the to look different to her. Nor need her experience involve any such revelation. What, then? There is a way in which the muntins look like a swastika. There are ways in which they also look like many other things – a cross, for a start. Pia’s viewing the window – her seeing it as she does – cannot change the looks there are to be noticed in it. For her to see the muntins as a swastika, though, is for it to look a certain way to her. The lead idea now is: she has a certain authority over how things look to her; this is a certain form of executive authority. Executive authority is something we are invested with. At rough first approximation, for Pia to be so invested here, for her thus to hold sway over its relevant object – how things look to her – is for her to have the right to say how they look to her. But this can only be rough, and then in a number of ways. First, executive authority is sometimes exercised, or exercisable, in saying one thing or another (for example, ‘By the power invested in me, Bloggs is now a judge’). But it may be to be exercised in other ways. In our present case, saying cannot be, literally, what is in question. For the muntins to be, at the moment, thus and so for her is (perhaps inter alia) for her to have a certain attitude towards them (or towards her present experience of them). But then there are general remarks about the connections between our attitudes and the things we say. The most obvious is: an attitude is (grammatically) a state (however transient), saying an act – and there is always room for grammatical slippage between act and fact. An act of representing oneself as in a state, even when avowing something over which one has executive authority, is always susceptible in principle to being belied by the state one is in fact in. Pia said that for her Sid is the fair-haired boy – so why that little expression of disgust every time he comes in sight? Thinking Sid the one is part of the pursuit, if not quite of truth, at least of the thing to do. It is, thus, guiding oneself accordingly. In which case calling him that not only cannot make it so, but cannot (psychological law aside) so much as make it so. One would think similarly, in principle, of claiming, even sincerely, to see the as a mirrorimage F). An attitude pervades a life, or at least spreads itself indefinitely far. An act is at best a slice of its career. A more particular feature of the present case of executive authority is that, as in many other cases, the relevant executive authority, or its exercise, must be, so to speak, wrung from one. Though, plausibly enough, our taking ourselves to be obliged to think such-and-such is in some sense constitutive of our so doing (rather than evincing expertise, or lack thereof, about ourselves), at the same time one cannot choose what to believe. If, as Sid sees things, he could choose to believe that Pia is at her mother’s, and also could choose not to believe it, then, ipso facto, he does not believe that Pia is at her mother’s, and no mere choice on his part could make this so, similarly for

The room in a view


seeing the F as a mirror-image F. Though Pia may, in her attitudes towards what she sees, exercise executive authority of whether it is so that she sees the F as a mirror-image F, it is not up to her to choose whether she so sees it or not. The relevant attitudes must be wrung from her much as in the case of taking-to-be-so. One can see a sort of parallel in umpiring. It may be, by the laws of cricket, that if the umpire says ‘LBW’, then, so far as that game was concerned, so it was. But the umpire is also obliged not to do things arbitrarily. So if that ‘LBW’ was not wrung from him (by his goal of pursuing truth) then there is a good sense that he is making incorrect decisions. Even cricket would be a far different game if things were otherwise. Further, there may be conditions on surrounding conditions being apt for exercise of relevant executive authority. For example, it will not do for it to be for Pia as though she is seeing a mirror-image F if her eyes are closed, or it is too dark to see, or, perhaps, if what she is seeing is obviously and clearly a duck–rabbit, a hart pierced by an arrow or a Bentley. It will not do, that is, for making it the case that she saw something as a mirror-image F. This sketches, in barest outline, one sample of how executive authority may figure in respecting Publicity, conforming to Frege’s principle that there are no facts (irreducibly) about Vorstellungen. For someone to have seen something as a mirror-image F is for him, or his being as he is, to fall under a generality. Generalities can only generalise over what there is for one to meet with (whether in experience or in thought). What there is to generalise over in the case at hand is, for example, Pia’s being as she now is, where what matters in that is, notably, how she stands towards what she has in view – that is, towards what there is in that part of the environment which she has in view. With Frege, we may rightly feel ourselves free of pressure to deny that there are Vorstellungen. They just have no role to play in the above story, or any story of its kind. At various places in the daybooks there are remarks such as these: ‘Do you always see this leaf as green, as long as you see it and it doesn’t change colour for you?’ Does this question have a clear sense? (1982a: §720) It would have made as little sense for me to say, ‘I see it now as . . .’ as to say at the sight of a knife and fork, ‘I see it now as a knife and fork.’ One would not understand this utterance.—just as little as this: ‘Now it’s a fork for me’, or ‘That can be a fork too.’ One does not ‘take’ what he recognises on the table as cutlery for cutlery; just as little as one ordinarily tries to move his mouth while eating, or aims at moving it. If someone says, ‘Now it’s a face for me’, one can then ask: ‘what change do you have in mind?’ (PI II xi (166))

28 Charles Travis Seeing-as is not something done routinely. It is not a routine part of perception. Or, perhaps better, there are not routinely things to say, either truly or falsely, as to what, in a given visual experience, someone saw something as. In particular circumstances there may be cause to speak of someone having seen (or currently seeing) something as something. But such circumstances then have a substantial role to play in fixing what it is one will thus have said. There are several things Wittgenstein may have had in mind by this. We are now positioned to appreciate one of these. Among the cases in which it does make sense to speak of someone as having seen (or not) such-and-such as such-and-such are ones in which an answer is to be found in executive authority, along the lines illustrated. Executive authority has something to contribute to some such questions, but by no means all. It need not, often does not, contribute at all to answering a question of what was experienced visually. Nor is its exercise a sine qua non for there having been visual experience. There are two points here. The first concerns the contingent presence of work for executive authority (or any substitute for it) to do. ‘See-as’, where it describes visual experience, picks out a range of special cases in which answers of a special sort are required to some questions as to what was experienced visually. Sometimes – for example, with visual ambiguity – those special questions are still about what was seen and are thus to be settled in terms of the expert authority enjoyed. Sometimes – vide cases of our second kind – some such questions are not answerable in that way. These may call for answers in terms of how things were for the viewer, how they looked, or seemed, to him. Here, for things to have been thus and so for him is for executive authority to have been exercised. In such special cases executive authority contributes to making the experience what it was. But where, or insofar as, expert authority does fix what was experienced, executive authority is not obligatory. There need be no further fruits of it. When expert authority has settled what it settles, there need be no more to say. Sid sits on his deck, longneck in hand. Suddenly he stands up and glances at the space before the rhododendrons at the forest edge. ‘What did you see?’ Pia asks. ‘A rabbit ran by. It was brown, I think. It had long ears.’ ‘Did it have a white breast?’ ‘I’m not sure. I didn’t get a good look. Rabbits are fast.’ Given this, how else did the rabbit, or its jaunt, look to Sid? Pia saw Sid, for the first time, wearing his new summer pullover. What colour was it? Green. A sort of penicillin-mould green. Yes. But what colour did it look to Pia? What Sid saw, and what Pia did, is fixed by the expert authority thus exploited (modulo the usual plasticities in the notion ‘see’). If there is, moreover, executive authority to be appealed to here, such is a contingent fact. Suppose not. Suppose nothing settles any further questions how things were visually for Sid. Still, all that expert authority settles stands. Such is one way a visual experience may be. One might think that, anyway, the neurophysiology of vision, or some other account of ‘the visual system’ – that presumed

The room in a view


battery of processors which enable awareness of scenes before one – will someday provide answers to such further questions, natural ones or not. But, as what was seen (and not missed) is itself a more complex issue than anything a visual system alone might settle, so too further facts as to how things were visually for a viewer. Talk of visual systems is so far (here irrelevant) science fiction. Second, seeing has a certain priority over looking-to-one, as expert authority does over executive in re perceptual experience. Pia sees Sid in his new pine-green linen pullover. Thus far, she simply experiences what is there. The jumper, to her jaundiced eye, looks penicillin-mould green, reminiscent of a ripe Gorgonzola (like Sid himself). How did it thus look to Pia? Penicillinmould green. Executive authority settles this. But how does that look? Go find a lump of Gorgonzola. It looks like that. For Sid’s pullover to look penicillin-mould green to Pia is for it to look like that to her. For this to be understandable, there must be such a thing as what penicillin-mould green, so Gorgonzola, looks like. So, too, for that illustrative lump. Such are the looks to be found in what there is to see – ones to be seen, observed, witnessed. How penicillin-mould green looks is an object for expert authority, of the sort seeing is for gaining. So, too, for Gorgonzola, and for that lump. For Sid’s pullover to look penicillin-mould green to Pia’s jaundiced eye is for it to be for her as though she were enjoying expert authority over something penicillin-mould green, for it to be for her a simulacrum of this. In present matters, expert authority must sometimes hold its sway before there can be any question executive authority might settle. And, in such matters, where there is question of expert authority there is visual experience.

Disjunctivism ‘Disjunctivism’ refers to a family of views, distinguished in not inconsiderable part by their opponents’ incomprehension of them, or of their motives. What seeing-as has taught us so far provides one very clear motive for such a view about perception. The current inspiration for such views – or, anyway, their name – is J. M. Hinton (1967). A disjunctivism might concern, inter alia, knowledge, or perception, or thinking singular thoughts. But it is Hinton’s view on perception which will concern us here. Hinton presents his view in terms of what he calls ‘perception-illusion pairs’ – pairs of statements, that is, of the form ‘N is/I am now seeing suchand-such’ and ‘N is/I am having a perfect illusion of such’. But perhaps there is a simpler way of getting at what matters. Disjunctivism is the rejection of a certain idea. The rejected idea is that wherever someone sees such-andsuch he also, inevitably, experiences visually, something else, Ψ, which is visually thus and so, and which, further, has the following three features: 1

Neutrality. Experiencing Ψ is as such without prejudice to what the person in the case in question saw. One might experience Ψ in seeing

30 Charles Travis



what he did on some occasion. But Ψ is what might also be experienced (visually) in witnessing an illusion of that, or in suffering the illusion of seeing it. If his experience is one of seeing such-and-such, then his then seeing that was, ipso facto, his experiencing (visually) Ψ. But if he were suffering, or experiencing, an illusion of seeing this, for all that, he would still be experiencing Ψ. Plenitude. What our subject experienced visually, on the occasion, is just what he would then experience in, or by, experiencing Ψ. Whatever he saw, if anything, is just what one would then see in experiencing Ψ. Conversely, Ψ is precisely that which he could be said (truly) to have experienced visually without prejudice to whether he saw what he in fact did (or anything) or not. In that sense, seeing something is not experiencing more visually than one would in experiencing Ψ. Ringer-resistance. There is no such thing as a ringer for, or perfect illusion of, experiencing Ψ visually. Or (following Hinton) it is at least very difficult to make sense of that idea. Experiencing Ψ does not admit of ringers as, for example, seeing figs in the bowl on the sideboard does.

What Ψ is depends on the experience in question. For a given such episode, α, call the corresponding Ψ Ψα. What is visually one way or another in an experience of Ψ need only be things being for the subject as they then are visually. Experiencing a Ψ could not, anyway, be experiencing the environment, or any denizen thereof, being visually thus and so. For a start, such could not be ringer-resistant. As Frege notes, ‘By the step by which I win myself an environment, I expose myself to liability to error’ (1918: 73). Whether I see before me a red tomato, though this may be recognisable by, and at, sight, depends on more than just what is visible from where I stand. What I see had better interact otherwise with the environment as a tomato would. If I do see a red tomato, there is thus always, inevitably, room for ringers for what I see, equally for seeing red, or something being so coloured. For there to be something red before me is for what is before me to interact otherwise with the environment as thus required. If, for example, its being red is a matter of how it looks (or what look it has), how it does look depends on more than me, or my experience of it. So, again, seeing red, or something red, in the scene before me cannot be ringer-resistant. But, anyway, independent of this, experiencing a Ψ could not be environmental experiencing if it is to be without prejudice to what, if anything, was seen. We do not need hallucinations to make the point. Suppose for a given tomato, there is another, or a wax copy, just like it. No part of the one is any part of the other. So there is nothing so far to be visually aware of without prejudice to whether one is seeing the tomato or doing something else. One might see something red in seeing either. To do this would be to see something in each case – a particular case of something being as it is – which, in each case, falls under a given generality: for a thing to be red. But in each

The room in a view


case it would be a different thing which so fell. There would be no one thing which one experienced being visually thus and so. Perhaps Ψ need be no more than this. Perhaps experiencing Ψ is meant to be no more than experiencing the instancing, by something, of some specified ensemble of generalities – for example, for a thing to be red. But to experience Ψ one must be seeing something (for example, something red). Again, whether it is red which one thus experienced something as being depends on the place of what was experienced in our environment, which makes experiencing Ψ, on this conception of it, intrinsically susceptible to ringers. Why might anyone feel the call for such a thing as Ψ? There is a reason. Insofar as someone saw such-and-such, what he thus experienced visually can be specified by referring to the scene in view. In the scene before Pia, for example there was, visibly, a sleeping sow and a sleeping piglet pressed against her. The sow was cream-coloured, with a pattern of brown patches. The piglet had not yet got its spots. The sow’s flanks were heaving as she breathed heavily. The piglet quivered. Our subject saw the sow and the piglet, but missed (did not notice, or could not make out) whether the piglet had patches, missed its quivering. To say such things is to identify what Pia experienced visually. She experienced, visually, that sow and her spots. Similarly, she did not experience just that (what she did not notice, or could not make out, etc.). But what was it like, visually, for her to be seeing what she saw? That is what it was like. What did it look like? What is it (or what was it then) like to be seeing all that? Well, what does that sow look like (or did it then) lying like that, snoring away? Such is not just for Pia to appreciate. But there is a problem with this way of identifying what someone experienced visually. There is, for example, that sow on the straw, fast asleep. What is it like to see that? Have a look. There is no one one must be to do so. But just so. Two people viewing the sleeping sow might conceivably differ in what they experienced visually. To one but not the other, the piglet might have appeared to have a patch – a mere trick of the light and shadow. Or the one, but not the other, might have taken in that point on the sow’s flank where those two patches meet. Or the one but not the other might be able to make out the whiskers on the sow’s snout, or to distinguish the colour of the snout from that of her flank. So, it may seem, having said as much as the above as to what a given sow-cum-piglet viewer experienced visually, more remains to be said about this – for example, something to place the viewer on the one side or the other of the distinctions just pointed to. The problem here lies in what the example (seeing that sow sleeping on the straw) seems to exemplify. What has just been said about that case would apply, it seems, to any. Two (simultaneous) viewers of any given scene might always differ somehow in how things appeared to them. So suppose one thought, a) that for any experience of seeing such-andsuch, there is such a thing as that which was experienced visually, a unique right answer to how things were then visually for the subject; b) that the right answer to that question must not be such that different instances of

32 Charles Travis experiencing that visually might differ in what was thus experienced visually. Experiencing what one thus did could then not be (just) seeing such-andsuch. We would, ineluctably, need to look elsewhere than in the scene in view to locate objects of such experiencing. If we needed to specify how things were visually for a subject on an occasion so as to meet the above demand, we would have to look to things for him to be experiencing which were provided just by his being as he then was, independent of how the scene in view was. Feeling compulsion to meet this demand might intelligibly drive one to posit such a thing as Ψ. There are, sometimes, features of a given instance of visual experiencing which are fixed in this last way above. There may sometimes be (true) stories as to what was experienced visually which are, indeed, stories as to how the viewer then was, independent of what was visibly in the scene before him – what was there to be seen. From within a certain philosophical picture (one which makes a and b above seem compulsory), examples like the ones above are what drive us to feel the need. But the lessons we have extracted, above, from seeing-as give reason to think that so viewing the matter cannot be viewing it rightly. The crucial point is this: where what was experienced visually is a matter not of the scene seen, or how it was or looked, but rather of how the viewer then was otherwise, we have left the realm of expert authority and entered that of executive authority. Where the question we want answered cannot be answered by appeal to what there was, then, for one to experience visually – as, by the point above, it cannot – it must be answered by that over which the viewer enjoys executive authority, as in that second set of cases of seeing-as. But the point now is: the exercise of executive authority is optional. Not that it need not be wrung from the viewer if it is to be proper exercise at all. Of course, it is not optional for the viewer in any such sense. Rather, it is optional in the sense that its presence is not needed for there to be a case of visual experiencing. Seeing is itself already such a case. It is a case of visual awareness – awareness which the presence of the right sort of expert authority (authority as to how things are) already is. Executive authority comes into play just where proper objects of expert authority have run out. But it is not part of the bargain in enjoying expert authority that one must exercise executive authority in re what are, as they must be, matters extraneous to that over which executive authority is exercised. Pia sees the sleeping sow. It can be that things thus looked a way to her, where this is extra to her simply taking in how that sow looked. Such is sometimes so. But nothing requires that it always be so. Such is one good reason to reject the idea of Ψ. It would be wrong to call Wittgenstein a disjunctivist: because of his general aversion to isms; because these are daybooks, not a discourse; and because, anyway, placing the topic of disjunctivism in the 1940s would be a bit anachronistic. Still, Wittgenstein’s sense for the variety in cases of seeing-as/seeing aspects, and for the point of attention to these, generates ideas which yield positive, and powerful, considerations which militate in disjunctivism’s favour.

The room in a view


Bibliography Frege, Gottlob, 1882 [1980]: Letter to Marty (?), Gottlob Freges Briefwechsel, Hamburg: Felix Meiner, pp. 117–19. Frege, Gottlob, 1897 [1983]: ‘Logik’, in Nachgelassene Schriften, 2nd edition, H. Hermes, F. Kambartel, D. Kaulbach eds, Hamburg: Felix Meiner, pp. 137–63. Frege, Gottlob, 1918: ‘Der Gedanke’, Beiträge zur Philosophie des deutschen Idealismus, book II, pp. 58–77. Frege, Gottlob, 1919: ‘Aufzeichnungen für Ludwig Darmstädter’, in Nachgelassene Schriften, 2nd edition, H. Hermes, F. Kambartel, D. Kaulbach eds, Hamburg: Felix Meiner, pp. 273–7. Frege, Gottlob, 1983: 17 Kernsätze: ‘17 Kernsätze zur Logik’, in Nachgelassene Schriften, 2nd edition, H. Hermes, F. Kambartel, D. Kaulbach eds, Hamburg: Felix Meiner, pp. 189–90. Hinton, J. M., 1967: ‘Visual Experiences’, Mind, NS v. 76, n. 302, pp. 217–27. Wittgenstein, Ludwig, 1922: Tractatus Logico-Philosophicus, London: Routledge and Kegan Paul. Wittgenstein, Ludwig, 1953: Philosophical Investigations, Oxford: Basil Blackwell. Wittgenstein, Ludwig, 1980a: Remarks on the Philosophy of Psychology, v. I, Oxford: Basil Blackwell. Wittgenstein, Ludwig, 1980b: Remarks on the Philosophy of Psychology, v. II, Oxford: Basil Blackwell Wittgenstein, Ludwig, 1982a: Last Writings, v. I, Oxford: Basil Blackwell. Wittgenstein, Ludwig, 1982b: Last Writings, v. II, Oxford: Basil Blackwell.

This page intentionally left blank

Part II

Difficulties with Wollheim’s borrowing from Wittgenstein

This page intentionally left blank


Seeing aspects and telling stories about it Joachim Schulte

My title mentions the seeing of aspects and the stories people tend to tell about their seeing these aspects. There is, however, a question lurking behind my phrasing the title this way. To put the question very roughly, I might say that I want to find out whether Wittgenstein’s remarks on aspect-seeing, or seeing x as y, can prove useful in discussing certain questions in the philosophy of art. When I say that I want to find out about the potential usefulness of Wittgenstein’s considerations I do not in the least want to suggest that what I have in mind is something approaching an exhaustive answer to this question. The best I can hope to do here is to mention a few examples which may incline us either to explore this question further or to drop it altogether. And in speaking of Wittgenstein’s remarks on the subject of aspect-seeing I do not wish to suggest that what I shall be talking about are ideas or judgements that can with near-certainty be attributed to the historical figure Wittgenstein. In spite of the fact that I shall make a few claims about what I regard as right or wrong or misleading attributions to this historical figure I see no way of supporting these claims in any detail or of making all the cautionary remarks that seem necessary in the context of a chapter like the present one. To bring my introductory observations to an end: let me add that as regards exegetical work on this topic I find the present scholarly situation surprising. On the one hand, virtually everyone with a modicum of familiarity with twentieth-century philosophy has heard of Wittgenstein’s interest in the duck–rabbit figure and is prepared to use, or listen to, phrases like “Wittgenstein’s remarks on seeing-as” or “what Wittgenstein has taught us about aspect-perception” (cf. the recent collection edited by Day and Krebs (2010)). On the other hand, most people will not be able to say much about the possible point of Wittgenstein’s interest in figures of the duck–rabbit type, nor will they volunteer much about the meaning of those remarks on aspect-perception. This situation is all the more surprising if one remembers the fact that we are talking about a comparatively short, and hence surveyable, period of Wittgenstein’s working life, viz. two or three years between 1946 and the spring of 1949, and if one also remembers the further facts that, first, all the relevant manuscripts are easily accessible and, second,

38 Joachim Schulte much of this material is available, not only in print, but even in translations into various languages. Most people’s familiarity with these matters will (for obvious reasons) be restricted to what they have been able to glean from a text originally published as ‘Part II’ of the Philosophical Investigations and now available, in revised form, as ‘Philosophy of Psychology – a Fragment’ (= PPF as in Wittgenstein 2009). But even this fragment comprises a good deal of material and would certainly warrant much more detailed discussions of Wittgenstein’s ideas than we are used to being presented with. And to mention another surprising aspect of the situation: I well remember that, as early as more than 40 years ago, there was nothing unusual in hearing, or reading, people mention the supposed relevance of Wittgenstein’s ideas on aspect-perception to questions of aesthetics. And this was not unusual even though it was plain that there was next to nothing in Wittgenstein’s text that had any direct bearing on questions in the philosophy of art. Outstanding among the early exploiters of Wittgenstein’s ideas for dealing with aesthetic problems was Richard Wollheim, who (in the first edition of his well-known essay on Art and Its Objects (1968)) defends the view that there is a kind of seeing appropriate to representations. Some of these representations would be paintings or other objects generally recognized as objects of aesthetic interest and potential aesthetic value. According to Wollheim’s account, the pertinent kind of seeing can be explained in terms of what he calls seeing-as. As an absolute minimum, this faculty of seeing-as is claimed to be of great use in explaining the kind of seeing appropriate to representations. And in the bibliography due credit is given to Wittgenstein’s work. As we all know, Wollheim changed his mind about this: he not only gave up the idea that seeing-as would supply the most useful terms in which to explain the sort of seeing appropriate to representations; in an important essay appended to the second edition (1980) of Art and Its Objects he also developed a new notion, the idea of seeing-in, which he now claimed to be the idea he had really been after all the time. Here, I shall not say much about this new notion of seeing-in, but I am interested in Wollheim’s later way of characterizing the notion of seeing-as. After all, it is because of this characterization that he finds seeing-as unsuitable for understanding that way of looking at pictures which is supposed to be an essential feature of what makes them potential works of art, and hence potential objects of aesthetic appraisal and appreciation. Now, in his essay on seeing-as and seeing-in added to the second edition of Art and Its Objects (1980) Wollheim emphasizes that by distinguishing these two kinds of seeing he aims at clarifying the difference between what he calls “two distinct perceptual projects.” I must confess that this characterization does not really help me to get a better grasp of the distinction. But I suspect that the idea of distinct perceptual projects involves the idea of fairly unified, circumscribable processes or activities. Moreover, speaking of a project, or projects, seems to suggest that we are dealing with processes or activities that are clearly subject to our will—that is, it is supposed to be up

Seeing aspects and telling stories about it 39 to us to decide whether they will be carried out, and it is also supposed to be up to us to decide whether the carrying-out will be done in this way rather than that way. At any rate, this idea of conceiving seeing-as and seeing-in as distinct perceptual projects may be at the bottom of certain differences between Wittgenstein’s and Wollheim’s views—differences which are surely far from obvious. Wollheim himself underlines the fact that “in developing the distinction” between seeing-in and seeing-as he has “not drawn heavily upon linguistic intuitions.” He adds that he is not “even certain that everything that [he says] about the phenomena conforms to such intuitions as we do have about these phrases” (1980: 209). His use of them may thus be regarded as “quasitechnical” (210), as he says. Wittgenstein, on the other hand, is extremely interested in what we say in situations that appear to warrant being described in terms of seeing-as or related locutions. While he (in contrast to Wollheim) would never have permitted himself to speak of linguistic “intuitions,” he would not have wished to miss any philosophically relevant differences between ways of using words we are inclined to use in the contexts involving the phenomena Wollheim wants to draw our attention to. But Wittgenstein’s desire not to overlook certain diversities in our ways of using words does not depend on any higher-level intention to conform to usage (or the alleged “intuitions” invoked by Wollheim). However, detectable conformity or nonconformity of linguistic behavior may tell us a lot about the human mind and in particular about phenomena we are inclined to speak of in terms of seeing-as or perceiving aspects. Certain differences between their respective approaches may become clearer if we look at Wollheim’s use of phrases that play obviously central roles in Wittgenstein’s remarks. The first phrase I want to discuss is “to see as” or, as I put it earlier, “to see x as y.” At one point in his account Wollheim draws an instructive distinction between the kind of seeing appropriate to painted portraits and the kind of seeing appropriate to photographic portraits. In the former case the standard of correct seeing is, as Wollheim points out, determined by the artist’s intention. And this is the reason why in these cases a distinction can be said to hold between model and sitter for a portrait. For example, the intended sitter’s twin brother may function as the painter’s model, but if a photograph is taken of him, the resulting picture will be of him and not of the intended brother. And it is in respect of this distinction that Wollheim writes that “it must be emphasized that this point applies to the seeing appropriate to photographs, or to seeing photographs as photographs” (1980: 208). Now, to those of us who have read plenty of Wittgenstein on seeing-as a passage like this will give pause. For in various places Wittgenstein points out that seeing x as y will make little sense if y is replaced by x—that is, if you say for example that someone sees a duck as a duck. Such an utterance would, as Wittgenstein says, make as little sense “as to say at the sight of a knife and fork, ‘Now I see this as a knife and fork’” (PPF §122). It must be

40 Joachim Schulte remembered, however, that this is a remark of restricted generality: it is not about the use of these phrases in all conceivable contexts but about the use of given types of phrase in certain types of context of utterance. So, if we have another look at the quoted passage from Wollheim’s essay, we may wonder what it could possibly mean to speak of seeing photographs as photographs. Could it, for example, mean that we are asked to look at photographs in a way characteristic of people who have not yet learned to, as it were, “read” them properly—that is, to see them as representing objects and people, for instance? Such people would only see little slips of paper with gray or colored blots and lines on them. In other words, the standard use of seeing x as y requires that y refer to a thing or type of thing different from what is signified by x. But, as Wittgenstein himself indicates (PPF §121), this description of the standard use is a statement of limited generality: in describing what another person sees I may say, in respect of what is a picture of a duck, that he sees it as a picture of a duck – provided that it is clear to me and my intended hearer that the picture is an ambiguous one and could, for example, be seen as a picture of a rabbit. Analogously, we may reconsider our first response to Wollheim’s statement about seeing photographs as photographs as soon as we learn from what he says later that photographs can under certain conditions serve as pictures of a different kind. If, for example, I take a picture of my friend Jim in fancy dress, it may in certain situations serve, no longer as a photo of my friend Jim, but as a picture of Socrates or as a picture of an (unspecified) ancient Greek citizen. With this contrasting possibility in mind I may then proceed to encourage my interlocutor to see the photograph as a photograph—that is, not as a picture of an ancient Greek citizen but as a picture of Jim. Without the implicit contrast, however, the original words would not work. In his essay, Wollheim mentions a number of reasons why, in specifying the kind of seeing appropriate to representations, he prefers to draw on the notion of seeing-in to relying on the notion of seeing-as. Of course, giving these reasons presupposes that we have an idea of the difference, or differences, between these kinds of seeing. Here, I want to give prominence to a way of distinguishing between these two perceptual projects (as Wollheim calls them) which may help to throw light on certain features of Wittgenstein’s approach. At one point Wollheim writes: The central difference between seeing-in and seeing-as, from which their various characteristics follow, lies in the different ways in which they are related to what I call ‘straightforward perception’. By straightforward perception I mean the capacity that we humans and other animals have of perceiving things present to the senses. Any single exercise of this capacity is probably best explained in terms of the occurrence of an appropriate perceptual experience and the correct causal link between the experience and the thing or things perceived. Seeing-as is directly

Seeing aspects and telling stories about it 41 related to this capacity, and indeed is an essential part of it. By contrast, seeing-in derives from a special perceptual capacity, which presupposes, but is something over and above, straightforward perception. (1980: 217) This passage reveals the extent to which Wollheim’s account relies on the ideas of straightforward perception and an appropriate perceptual experience, where an experience of this kind is understood as a (particular) exercise of the (general) capacity of straightforward perception. Probably, an account of the sort intimated in the quoted passage will strike many readers or hearers not only as familiar but also as more or less obviously along the right lines. So, for these readers or hearers it may come as a surprise that essential elements of the position outlined by Wollheim are caricatured by Wittgenstein or, at any rate, used to identify a view evidently regarded as erroneous. To illustrate this I shall quote from a remark published in ‘Philosophy of Psychology—a Fragment’ (the former ‘Part II’ of Philosophical Investigations), which refers back to a figure that, according to the verbal context in which it is embedded, can be seen as an illustration of various completely different things. The remark begins with the words “Here perhaps one would like to respond . . .”—and there can be no doubt about the correctness of the impression that in Wittgenstein’s opinion there is something wrong about this type of response. He continues as follows: The description of immediate, visual experience by means of an interpretation is an indirect description. “I see the figure as a box” amounts to: I have a particular visual experience which is empirically found to accompany interpreting the figure as a box, or looking at a box. But if it amounted to this, I ought to know it. I ought to be able to refer to the experience directly, and not only indirectly. (PPF §117) For present purposes I shall leave aside the argument indicated by Wittgenstein and concentrate instead on a term central to Wollheim’s account—a term which, as you will see, is connected with a nice problem of translation. The term I mean is “experience.” As you will remember, in the longish passage quoted before, as well as throughout much of his essay, Wollheim freely talks about perceptual experiences as exercises of perceptual capacities. It is this type of use of the expression that Wittgenstein focusses on in his remark, where he proceeds to observe that speaking of such visual experiences does no useful work unless I manage to refer to these experiences, not only indirectly (that is, by way of introducing interpretations of what is seen), but directly (that is, by way of terms directly referring to those supposed experiences—and hence by way of terms we simply do not possess and are ignorant of). The problem of translation I alluded to arises near the beginning of Wittgenstein’s remark: there, he speaks of a Beschreibung der unmittelbaren

42 Joachim Schulte Erfahrung, des Seherlebnisses—and here it is straightaway brought home to us that English has only one word (“experience”) where German has two (Erfahrung and Erlebnis). In this case, the poor translators’ way out consists in simply making do with the one word English has to offer: accordingly, the Beschreibung der unmittelbaren Erfahrung, des Seherlebnisses becomes “the description of immediate, visual experience.” I think this way of rendering Wittgenstein’s words is justified for the simple reason that Wittgenstein does not make much (if anything) of the presence of these two nouns in German and often seems to use them more or less interchangeably. More could be said about this, but there is a much more important point that needs to be made here: as far as I can see, Wittgenstein never uses Erfahrung or Erlebnis as terms for an inferred process or activity of the kind Wollheim supposes to have got hold of unless he is attempting to describe or caricature a position which he himself does not intend to defend. For this reason, it would be misleading to summarize things that Wittgenstein would want to claim about ordinary processes of seeing something, for example, by speaking of perceptual or visual “experiences.” In this sort of context, his use of “experience” (or rather Erfahrung or Erlebnis) is restricted to fairly special or peculiar cases such as sudden changes of aspect: the abrupt switch from duck to rabbit or vice versa would be a case in point. It is an almost physical occurrence often accompanied by, or involving, involuntary gestures and noises. To a bystander, these gestures and noises may reveal what is going on, so he can comment on the situation by saying things like “Ah, now you see that it is rabbit too” or “You see, now you have experienced the switch from duck to rabbit.” Taking his cue from this sort of situation, Wittgenstein makes a great number of relevant remarks. Here I want to emphasize two closely related points lying near the surface of his observations. The first point is that the experiences Wittgenstein is concerned with have natural and often involuntary forms of expression. It is extremely important to note that expressions of these kinds are not in the nature of comments on, or descriptions of, what is going on. Often, we cannot help making these gestures or noises. In many cases of this kind we are quite unable to suppress these natural expressions of what we experience: in this respect they are similar to cases of strong pain or other sensations. The second, closely related point is that there are, not only involuntary gestures and noises, but also certain types of verbal response to these experiences which force themselves upon us—sie drängen sich auf, as Wittgenstein says—even if they are not immediate reactions such as the involuntary gestures mentioned before. So, in a way these verbal responses are very similar to immediate expressions of experiences and share some of the “cannot help it” features typical of those expressions. One characteristic form of expression essentially employs indicators of time. For example, confronted with the duck–rabbit figure, I may say something like “Now it’s a rabbit”, “Now it’s a duck,” and so on. My saying this serves to communicate to other people that I have caught on to the peculiar

Seeing aspects and telling stories about it


switching capacity of the figure, but at the same time it simply is an expression of my experience and as such not an utterance serving this or that (external) purpose. In cases of ordinary, or what Wollheim calls “straightforward,” perception it would be completely out of place to say things like “Now it’s a rabbit” when, for example, I am just watching an unmistakable rabbit in my back yard. In the same way it would be foolish and possibly misleading—or an elaborate kind of philosophical joke—to say that I am seeing it as a rabbit. So, if we think of more or less ordinary situations of perception, we shall find that what is absent are all those features standing out in situations of what Wittgenstein calls perceptual experiences and characterizes by referring to the suitability of certain verbal or non-verbal responses. In other words, these more or less ordinary situations are not what Wittgenstein has in mind when speaking, in his own voice, of visual or perceptual experiences. Accordingly, in making the almost programmatic statement that “We are interested in the concept [of ‘noticing an aspect’, PPF §113] and its place among the concepts of experience” (PPF §115) Wittgenstein is by no means speaking of the whole area of what can be learned by experience in the sense of an exercise of perceptual capacities as guided by our conceptual faculties. No, he is speaking of a much more restricted group of concepts that can come in useful in our attempts at capturing features of experiences in the narrow Wittgensteinian sense.1 Once we are fully aware of the fact that Wittgenstein’s use of the word “experience” is narrowly circumscribed in the way I have indicated, we shall notice further peculiarities of his treatment of cases of aspect-seeing or seeing x as y. One feature of his discussion which is not always accentuated with all the emphasis it deserves is the diversity of cases often grouped together by commentators on Wittgenstein’s work under a label like seeing-as. For Wollheim, for example, there seems to be no doubt about the supposed fact that seeing-as is one thing or one delimitable set of perceptual processes. This idea, that there is one sort of thing that Wittgenstein would be happy to call seeing-as, or seeing aspects, and accept as comprising at least the majority of cases discussed in the relevant context of his writings, is just wrong. This can be seen by looking at central examples well known from ‘Philosophy of Psychology – a Fragment’ and more elaborately examined in other manuscripts. Take the remark where Wittgenstein introduces the notion of noticing an aspect. He writes: I observe a face, and then suddenly notice its likeness to another. I see that it has not changed; and yet I see it differently. I call this experience “noticing an aspect”. (PPF §113) If you compare this example with the points Wittgenstein makes about the changing aspects of the double-cross, you will soon notice that there are numerous differences between these cases. Of course, this is not to deny that

44 Joachim Schulte certain similarities may be found to obtain between them. And what may be the most interesting similarity is the fact that a perceiver will, if duly stimulated, on many occasions be able to report certain experiences—and these will be experiences in what I have called “the narrow Wittgensteinian sense.” The reports of experiences had in perceiving such pictures or in noticing aspects of the kind mentioned are of especial interest to Wittgenstein, and they are what I am interested in in my attempt to examine the potential usefulness of Wittgenstein’s observations for aesthetic investigations. There can be no doubt that questions like “Why should one—or to what extent can one—rely on such reports of experiences?” are themselves eminently important philosophical questions. And they are questions that tend to be in the forefront of Wittgenstein’s mind. To answer them he would invoke the close connections between such reports, or verbal responses in a more general sense, and expressions—and, in particular, involuntary expressions—of experiences, feelings, sensations. The points Wittgenstein makes in this context may well suffice to justify our confidence that in many cases such reports can be relied on, even though they may require a great deal of interpretation if we wish to make use of them. Here, however, questions of justification will be left aside. Instead, I shall concentrate on one type of use that can be made of such reports or verbal responses. Thinking through the passage quoted can show that noticing an aspect— for example, suddenly noticing similarities between faces—is in several respects different from what will prove remarkable about examples of the duck–rabbit kind. They are similar to the extent that in both types of case the seeing involves experiences of certain kinds. The verbal responses, however, will tend to be different. In ordinary cases of noticing a similarity between faces one will simply say “Suddenly I noticed that their eyes are the same” or other words to this effect. There is nothing paradoxical or in other ways particularly striking about this sort of report. But now think of cases of the duck–rabbit kind. Here, experiencing the switch of aspects will tend to evoke certain responses of a much more striking kind. In many cases our reports—the stories we are prepared to tell in response to what we experience– will involve baffling, as involving almost paradoxical or contradictory elements. And what Wittgenstein wishes to emphasize here is that we are not in a position to avoid using these paradoxical elements. The following passage is a good example of Wittgenstein’s considerations: Now, if he alternately sees the picture in one way, then in the other, he is experiencing a change of aspect. What is so amazing about that? – Is it this: that the report “Now I see . . .” can no longer be a report about the object that is perceived. For earlier “I see a cube in this picture” was a report about the object I am looking at. What is incomprehensible is that nothing, and yet everything, has changed, after all. That is the only way to put it. Surely this way is

Seeing aspects and telling stories about it 45 wrong: It has not changed in one respect, but in another. There would be nothing strange about that. But “Nothing has changed” means: Although I have no right to change my report about what I saw, since I see the same things now as before – still, I am incomprehensibly compelled to report completely different things, one after the other. (Wittgenstein 1980 §§473–4) As you see, in this context, too, Wittgenstein emphasizes that there is only one way of phrasing our response to the change of aspects, and that one feels “incomprehensibly constrained” to describe completely different, perhaps even incompatible things. What we need to remember is that these are just two examples of a possibly great variety of cases that one might feel inclined to distinguish in more or less precise ways. This possibility, however, need not detain us. What I want you to bear in mind for the remainder of this discussion are the following two points: •

There are partly instinctive, spontaneous or involuntary, partly acquired, and to some extent conventional ways of responding to, reporting, or describing experiences involved in noticing aspects or seeing x as y. Some of our verbal responses to these experiences include paradoxical formulations. Spontaneous expressions as well as fully verbalized responses permit us to discuss and judge the qualities of our experiences. Conventionality and degrees of coherence give us means to assess the appropriateness or adequacy of given expressions or descriptions.

No doubt these two points are put in what may appear excessively vague terms. But all I need for my purposes are fairly vague terms. Whether they are excessively so, everyone will have to judge for himself. One feature of these verbal responses that Wittgenstein wants us to see is the circumstance that many of them require transposition into a different register in order to become fully intelligible. Thus, as Wittgenstein himself points out, what seems to be a descriptive account of a given situation (as, for example, the words “I saw it quite differently, I’d never have recognized it” (PPF §126)) needs to be understood, not as an ordinary statement of fact but as an exclamation if one wants to make proper sense of it. The very words “Now I see it as a rabbit” may sometimes make better sense if they are understood as an exclamation. And many of our natural as well as conventional responses can be expanded into miniature stories spelling out the background and the point of the utterance in question. Many of Wittgenstein’s remarks contain examples of such ways of expanding, or shifting the register of, our responses to certain kinds of experience. What should be obvious is that the telling of stories is a particularly reliable—inasmuch as intersubjective and often to some extent testable—means of getting across how a specific experience strikes us.

46 Joachim Schulte I think I have reached a point in my discussion where I can, by way of a few concluding comments, apply some of the things I have reminded you of to a question mentioned at the beginning—the question, that is, whether any of this material may prove useful in discussing questions of an aesthetic nature. My answer will be “Yes, it may be useful,” but at the same time much depends on what one means by “aesthetics” or “philosophy of art.” For, if you are a great believer in objective canons of beauty and given, perhaps permanent, standards of aesthetic excellence, you won’t need Wittgenstein’s or my help in appraising works of art. As we have seen, the stories we are inclined, or feel constrained, to tell in response to, or as expressions of, certain experiences can be indicators of how the objects or situations concerned strike us. They may be more or less reliable indicators, and their intersubjective testability may vary a great deal, but the important point is that, first, they can really serve to get across how certain objects or situations strike us, and that, second, they permit discussion of questions like whether the stories we are inclined to tell are adequate. What I do not want to suggest is that the perception or contemplation of works of art is highly similar or almost identical to that of figures like the double-cross or the Rubin vase. It would be ridiculous to claim that musing on essential features of Botticelli’s Primavera or Beethoven’s Fifth is much like contemplating the duck–rabbit. No, the connection I have in mind is a much looser one. It is simply this: that the verbal and non-verbal responses to puzzle pictures and other aspect-invoking devices and our ways of discussing these responses can serve as primitive (but none the less instructive) models of some of our ways of responding to works of art and of discussing these responses. The idea behind this is that if something along these lines is roughly right, then the virtues of the Wittgensteinian account of aspectperception could be seen as carrying over to a similarly inspired account of our ways of responding to works of art. Let me end with a couple of examples. The use of forms of quotation in literary, musical, and other works of art can be an extremely complicated matter to explain. This is especially so when we feel that a line of a poem, say, or a sequence of sounds in a piece of music, or, for that matter, a sequence of images in a film strikes us as a quotation even though we are not in a position to indicate a source. We may even be sure that there is no source: that the sounds or the images only appear to be quoted. In such a case it may be helpful to think of much simpler but in some respects analogous cases which may help to spell out what is peculiar about the effect we have in mind. In this case, one way of going about it would consist in reminding ourselves of the ways in which we respond to and discuss felt quotations in ordinary life—for example, the experience one tends to have when noticing that one has quoted a well-known line even though one did not have the slightest intention of quoting anything. If, for instance, you find yourself saying under the appropriate circumstances of utterance “O Lady! we receive but what we give” or “Gee, baby, ain’t I good to you,” you may all

Seeing aspects and telling stories about it


of a sudden notice that, while you said what you wanted to say, you at the same time quoted a familiar line. Noticing this may then give rise to a discussion of specific features of this experience and our ways of expressing and evaluating it. And this may then, in its turn, shed some light on the probably more complicated case of artistic quotation. My second example harks back to one of Wollheim’s reasons for wishing to appeal to seeing-in rather than seeing-as. He argues that seeing-as, in contrast to seeing-in, does not permit sufficient “simultaneous attention to what is seen and to the features of the medium.” It is in this context that he defends what he calls the “twofold thesis,” according to which in the right sort of seeing representations “my visual attention must be distributed between two things though of course it need not be equally distributed between them” (1980: 213). Now, I do not want to discuss the matter in Wollheim’s own terms. I merely want to remind you that many of the locutions mentioned by Wittgenstein will tend, or are specially designed, to draw attention to two different aspects at once. In particular, this applies to apparently paradoxical formulations like “Everything has changed and nothing has changed.” To be sure, the other side of the coin of paradoxicality is exaggerated obviousness, as in Wollheim’s (above-quoted) invitation to see “photographs as photographs” (1980: 208). Here, we tend to be at a loss on account of our impression that this invitation is supposed to say more than that we should look at some photographs while we do not know, and are not told by the invitation, what in addition to looking at the pictures we are meant to do. One way out of this difficulty would consist in following Wittgenstein’s advice on paradoxical formulations and look for secondary meanings. In the present case this might lead us to exactly the spot where Wollheim wants us to go anyway, viz. to the medium of representation. Thus, paying attention to the words we are inclined to use in situations of seeing aspects may help us to arrive at a place Wollheim wants us to reach by way of his notion of seeing-in.

Note 1 Here, it may be helpful to distinguish cases: (a) if you (as opposed to myself) have never seen the figure as involving different aspects, I (but not you) may speak of your having continuously seen aspect A. (b) If, however, you are aware of different aspects, you will surely find ways of distinguishing between experiences of continuous aspect-seeing and experiences of sudden lighting up (PPF §118): these will be verbal means or verbal-plus-gestural means; if you have no verbal means at your disposal, the problem does become difficult, but might be solved by using an interlocutor’s hints and pictures as a way of eliciting helpful responses from you. See the exchange between Baz (2000, 2010) and Mulhall (2001).

References Baz, A. (2000). “What’s the Point of Seeing Aspects?” Philosophical Investigations 23 pp. 97–121.

48 Joachim Schulte Baz, A. (2010). “On Learning from Wittgenstein, or What Does it Take to See the Grammar of Seeing Aspects?” In W. Day and V. Krebs (eds.), Seeing Wittgenstein Anew. New York: Cambridge University Press. Day, W. and Krebs, V. J. (eds.) (2010). Seeing Wittgenstein Anew. Cambridge: Cambridge University Press. Mulhall, S. (2001). “Seeing Aspects.” In H. J. Glock (ed.), Wittgenstein: A Critical Reader. Oxford: Blackwell. Wittgenstein, L. (1980). Remarks on the Philosophy of Psychology. Volume II. Ed. by G. H. von Wright and Heikki Nyman, tr. by C. G. Luckhardt and M. A. E. Aue. Oxford: Blackwell. Wittgenstein, L. (2009). Philosophical Investigations. Revised fourth edition by P. M. S. Hacker and Joachim Schulte. Malden, MA, and Oxford: Wiley-Blackwell. Wollheim, R. (1980). Art and Its Objects: With Six Supplementary Essays, second edition. Cambridge, New York: Cambridge University Press.


Aspects of perception Avner Baz

The topic of this chapter is what Wittgenstein calls “seeing something as something,” or the seeing of “aspects,” and Richard Wollheim’s (1980) discussion of “seeing-as” in a supplementary essay appended to Art and Its Objects. While I believe I have a fairly clear grasp of what Wittgenstein means by “seeing-as,” or by “seeing aspects,” I suspect, and will try to show, that it is not altogether clear what Wollheim means by “seeing-as”—what phenomenon or set of related phenomena he means to refer to with this expression. And it seems to me that Wollheim’s difficulties are not unique to him. The philosophical topic of seeing-as is difficult. Anyone who wishes to come to a satisfying understanding of that topic must grapple with fundamental and difficult questions about human perception, and at the same time grapple with fundamental and difficult questions about philosophical method—what it is we are after, or ought to be after, in philosophy, and how it may best be pursued. Wittgenstein, who first brought to philosophical attention the topic of seeing-as, is reported to have said not long before death, and after many years of thinking about the topic: “Now try and say what is involved in seeing something as something; it is not easy. These thoughts I am now having are as hard as granite.”1 Over the years I have found myself returning again and again to the topic of seeing-as, prompted in part by a sense of its importance and of the inadequacy of my own understanding of it, in part by the sense that the topic presents us with a particular sort of difficulty that is itself philosophically interesting, and in part by the sense that that difficulty has not been aptly appreciated by some prominent readers of Wittgenstein’s remarks on aspects. Since the above is my topic, I will ignore the broader context of Wollheim’s discussion—namely, his theory of artistic (mostly pictorial) representation, and the distinction he draws between what he calls “seeing-as” and what he calls “seeing-in.” I will begin with a characterization of the phenomenon, or set of related phenomena, that I understand Wittgenstein to be investigating in his investigation of seeing-as, or the seeing of aspects. Wollheim takes himself to be offering an account of essentially the same phenomenon (1980: 209). I will argue that he is not. And the real problem with this is not that Wollheim has lost touch with Wittgenstein’s topic—after all, it is open to

50 Avner Baz him to make clear what phenomenon, or set of related phenomena, he means to refer to by “seeing-as,” and to offer an understanding of it. The real problem is that in losing touch with Wittgenstein Wollheim has rendered his own subject matter—whatever it is he means to be talking about—unclear. Or so I will try to show. At the same time, I think the motivation behind Wollheim’s proposed account of what he calls seeing-as should be taken seriously. Whereas Wittgenstein deliberately refrains from any attempt to offer anything like a comprehensive theory of seeing-as and its relation to human perception more broadly, Wollheim, together with many other readers of Wittgenstein’s remarks on aspect perception, is motivated by the conviction that such a theory, or an anyway comprehensive and unified account, can and ought to be given. More precisely, whereas Wittgenstein characterizes his topic through the phenomenon he calls “noticing an aspect” or the “lighting up (dawning, Aufleuchten)” of an aspect, and though he says at various places things to the effect that “the aspect can only dawn” (RPPI: 1021; see also RPP II: 540) and “lasts only as long as I am occupied with the object in a particular way” (PI, PPF: 237), Wollheim and many others have felt that the dawning of Wittgensteinian aspects is—must be—revelatory of (normal) human perception as such—of what Wollheim calls “straightforward perception.” Specifically, these philosophers have come to hold one version or another of the idea that, over and above the lighting up of aspects, there must also be a continuous version to the perception of aspects, and that all (normal) human perception can, and ought to, be understood as the perception of aspects. It seems to me that all of the attempts (with which I am familiar) to give sense to the notion of “continuous aspect perception” (or some equivalent notion),2 and to use it to characterize (normal) human perception as such, have failed.3 In this paper I will argue that Wollheim’s attempt fails. At the same time, I have come to think that the dawning of Wittgensteinian aspects is revelatory of a fundamental feature of human perception. The problem with previous attempts, Wollheim’s included, to draw a broader lesson about perception from the phenomenon Wittgenstein investigates in his remarks on aspects is that they have over-intellectualized human perception and therefore misidentified that feature. In a word, those attempts identify aspects in terms of determinate concepts, so that, at least in the most basic or paradigmatic case, what something may be seen as is taken to be something it can be judged, or known, to determinately be.4 By contrast, taking my cue from Merleau-Ponty and from Kant’s account of beauty in the Critique of the Power of Judgment, I will propose that the dawning of Wittgensteinian aspects reveals our power to perceive unity and sense that are not aptly thought of as conceptual, but which are nonetheless inter-subjectively shareable. And I will further propose that these unities and senses are, for the most part, importantly indeterminate—open to be perceived, taken up, and responded to in many indefinitely different ways.

Aspects of perception


In proposing a broader lesson about human perception that I think may be drawn from the dawning of aspects, I will be going beyond what may plausibly be found in Wittgenstein’s remarks. In this respect, I will be doing what many other readers of those remarks have done—at the cost, I have argued, of misrepresenting human experience and of failing to make clear sense with their words. I therefore embark on this project with great trepidation, for in no way do I take myself to be immune to the risks of confusion and nonsensicality. But I think there is at least this difference between what has driven others who have written on aspect perception to leave behind Wittgenstein’s ideas and method of inquiry and what drives me to do so: what has driven others away from Wittgenstein are more or less explicit theoretical ambitions that he did not share and moreover considered philosophically harmful. So the drive in their case is not essentially different from that of many others who have either never felt compelled by Wittgenstein’s general approach to the understanding and dissolution of philosophical difficulties or sought to move beyond Wittgenstein in their philosophical reflections on other topics. In my case, the need to move beyond Wittgenstein is internal to the substance of my specific subject matter and proposal. For if, as I will propose, what gets revealed in the dawning of Wittgensteinian aspects is a level of human experience that is pre-conceptual and pre-objective, and which serves as the basis of, but at the same time gets covered up by, everyday discourse—which mostly focuses on the objects of our experience rather than on our experience itself, and thereby hides from us the role we play in constituting the world as perceived—then perhaps it is only to be expected that what the dawning of Wittgensteinian aspects reveals about human perception will never come fully or explicitly to light in a Wittgensteinian grammatical investigation. Its essence will not be adequately expressed by grammar.5

The grammar and phenomenology of Wittgensteinian aspects I begin with what I take Wittgenstein to mean by “seeing (perceiving) something as something” or “seeing (perceiving) an aspect.” The first few remarks of section xi of part II of the Investigations are a good place to seek initial orientation: Two uses of the word “see”. The one: “What do you see there?”—“I see this” (and then a description, a drawing, a copy). The other: “I see a likeness between these two faces”—where the man I say this to may be seeing the faces as clearly as I do myself. The importance of this is the difference in category between the two “objects” of sight.

52 Avner Baz The one man might make an accurate drawing of the two faces, and the other notice in the drawing the likeness which the former did not see. I contemplate a face, and then suddenly notice its likeness to another. I see that it has not changed; and yet I see it differently. I call this experience “noticing an aspect”. (PI, PPF: 111–3, translation amended) The first thing to note, even before we draw on the basis of these remarks an understanding of what Wittgenstein means by “seeing as” or “aspects,” is that he characterizes his subject matter both grammatically—in the Wittgensteinian sense of that term—and phenomenologically. On the one hand, he talks about two uses of the word “see,” and gives an initial and partial characterization of those uses. This is in line with his “later” philosophical practice. At the root of any number of traditional philosophical difficulties, Wittgenstein identified the tendency to suppose that our words—including philosophically troublesome words such as “see,” “understand,” “know,” “think,” “mean,” “intend,” “pain,” and so on— “refer to objects,” or “denote items in the world,” and that the best way to become clear about the meaning of those words, or the concepts they embody, is to identify and study those “objects” directly—that is, not by way of an investigation of the use of those words.6 And since at least many of those “objects” have been taken to be metaphysically private—in the sense that each of us may only directly be acquainted with her or his “objects”—the tendency has been to suppose that such an investigation must either take the form of introspection, or else that of theoretical inference from “mere behavior” to what best explains it. What Wittgenstein tries to get us to see is that the model, or picture, of “object and designation” (PI: 293) is misguided and misleading when it comes to those words, and that what we end up producing, when we attempt to elucidate the nature of the “objects” to which they are supposed to refer, are philosophically constructed chimeras—“structures of air,” as he puts it (PI: 118)—that we erect on the basis of nothing more than “pictures” that we have formed for ourselves of those “objects.” Wittgenstein’s appeal to the use of philosophically troublesome words, or to what he calls their “grammar,” is an antidote to the above tendencies and the philosophical idleness they lead to. In the remarks on aspects, he repeatedly urges his reader (or himself) not to try to understand aspect perception by way of introspection of what happens in or to us when we see an aspect (PI, PPF: 241; Wittgenstein 1980a: 1011). “Forget,” he urges his reader (or himself), “forget that you have these experiences yourself” (RPP II: 531). “Don’t try to analyze the experience within yourself” (PI, PPF: 188; see also PI, PPF: 204). “The question,” he writes, “is not what happens here [that is, when someone tells me: “Now I am seeing this point as the apex of the triangle”—AB], but rather: how one may use that statement” (Wittgenstein 1980a: 315). Wittgenstein reorients his reader’s attention away from his or

Aspects of perception


her own experience and toward the use of relevant words—here, first and foremost, the words with which the experience of noticing an aspect may aptly and naturally be voiced. To attain clarity about the seeing of aspects— or for that matter about any other “concept of experience” (Erfahrungsbegriff, PI, PPF: 115)—we need to do more than just remind ourselves of particular isolated forms of words that may be used to describe or otherwise give voice to our experience. We need also to remind ourselves of “the occasion and purpose” of these phrases (PI, PPF: 311). “It is necessary to get down to the application” (PI, PPF: 165), to ask oneself “What does anyone tell me by saying ‘Now I see it as . . .’? What consequences has this piece of communication? What can I do with it?” (PI, PPF: 176, translation amended). A striking feature of all of the readings of Wittgenstein’s remarks on aspects with which I am familiar, and equally of attempts such as Wollheim’s to offer accounts of seeing-as that are more or less independent from Wittgenstein’s, is that they fail to heed this Wittgensteinian call altogether. The use of the relevant terms—where that, importantly, includes the philosopher’s use of them— tends to be neglected in favor of theoretical commitments and ambitions.7 What can we say about the grammar of (noticing) Wittgensteinian aspects? Taking our initial bearing from the opening remarks of PPF, section xi, cited above, we could say at least the following: first, aspects are contrasted with “objects of sight” of a different “category.” What are these other objects of sight? A red circle over there would be one example (PI, PPF: 121), a knife and a fork would be another example (ibid: 122), a conventional picture of a lion yet another (ibid: 203). Another type of object of sight that Wittgenstein contrasts with aspects is “a property of the object” (ibid: 247). In short, aspects contrast with what is objectively there to be seen, where what is objectively there to be seen may be determined, and known to be there, from a third-person perspective, and independently of any(one’s) particular experience of it. In contrast, someone may look at an object, see everything there is to see about it—in the first, objective sense of see—and yet fail to see (second sense) an aspect that may be seen by another. For this reason, it may aptly be said that aspects “teach us nothing about the external world” (Wittgenstein 1980a: 899). This last remark, while illuminating, has to be taken with caution, however, for it is going to matter what one understands by “teaching something” and by “the external world.” In particular, the tendency to think that if the aspect is not objective (part or feature of “the external world” objectively understood) it must be subjective (“inner,” metaphysically private) needs to be resisted, for it may be that one important lesson of aspect perception is precisely that this traditional dichotomy is at least sometimes misguided and misleading. Given the common philosophical understanding of objective and subjective, the aspect is, importantly, neither. The objects of sight with which aspects contrast may be described and often will be described (or otherwise represented) in order to inform someone else who for some reason is not in a position to see them—in order to

54 Avner Baz teach her, precisely, something about the external world. The other person, in Wittgenstein’s remark, asks, “What do you see there?”; and unless she is testing our eyesight or linguistic competence, she is asking because she cannot, for a more or less contingent reason, see for herself. By contrast, the person with whom we seek to share what we see when we see an aspect would normally be standing there with us and seeing as clearly as we do the object (the two faces) in which we see the aspect (the likeness between the two faces). Indeed, as Wittgenstein says, she could even make an (objectively) accurate representation of the object while failing to see the aspect. Accordingly, in giving voice to the seeing of an aspect, we normally seek, not to “inform the other person” but rather, as Wittgenstein puts it, to come in contact with, or “find,” the other (1980a: 874). In everyday, natural contexts—as opposed to the artificial ones of the lab or classroom—the seeing of aspects makes for a particular type of opportunity for seeking intimacy with the other, or putting it to the test. Like beauty (at least as understood by Kant in his third Critique), Wittgensteinian aspects are importantly characterized by the possibility that a fully competent speaker (and perceiver) may fail to see them even though he sees (first sense) as well as anyone else the objects in which they are seen, and by the particular sense it makes to call upon such a person to see them. This last point is connected with another feature of aspects: their being “subject to the will” (Wittgenstein 1980a: 899 and 976; 1980b: 545). Wittgensteinian aspects are subject to the will not so much, or primarily, in the sense that we can see them at will, but precisely in the sense that it makes sense both to call upon the other to see them and to try to see this or that particular aspect (PI, PPF: 256). Mostly, however, Wittgensteinian aspects dawn on us uninvited, and even, sometimes, against our will (Wittgenstein 1982: 612). They strike us. And yet we know we had something to do with their dawning, for we know that the objective world—the world that may be defined by its independence from any(one’s) particular experience of it—has not changed, and that no element of that world was revealed to us in the dawning of the aspect. So much, for now, by way of grammatical characterization of what Wittgenstein calls aspects. All of this Wittgensteinian grammar notwithstanding, the dawning (or noticing) of a Wittgensteinian aspect—unlike thinking, or knowing, or intending, or understanding, or meaning, or reading . . . this or that—is, first and foremost and essentially, a perceptual experience with a distinct phenomenology. Wittgenstein in no way denies this. On something like the contrary, I think this is one main reason why he found the seeing of aspects so interesting and at the same time so difficult to come to a satisfying understanding of. A striking feature of most of the existing accounts of seeing-as with which I am familiar is that they either neglect altogether or else misrepresent the distinct phenomenology of aspect perception—in favor, once again, of theoretical commitments and ambitions. An important merit of Wollheim’s account of what he calls seeing-as is his insistence that “seeing x as f is a particular visual experience of x” (1980: 223,

Aspects of perception


my emphasis). I will try to show, however, that Wollheim’s theoretical commitments prevent him from doing justice to that experience. The phenomenology of noticing an aspect is fairly easy to give an initial characterization of, though no characterization would be much good to anyone not familiar with the experience, and any form of words with which the experience might be characterized could also be understood in such ways that it would not aptly characterize the experience. When we notice an aspect everything changes and yet nothing changes (Wittgenstein 1980b: 474). We see (in the objective sense of that word, the first of the two uses of it that Wittgenstein speaks of) that the object has not changed, and yet we see it differently (in what Wittgenstein refers to as the second use of “see”). We know, and see (first sense), that the object’s objective features have remained unchanged, but its perceived physiognomy or expression has changed, and changed wholly. In an important sense, the aspect is un-detachable from the experience, or from the object-as-experienced. Another way of putting this point, which will become important for us later on, is that to see an object under an aspect is not the same as applying a concept to it, which, being general, is separate from the particular object and from our particular experience of it. Objects of sight of the first category, Wittgenstein tells us, can be described (or otherwise represented): I may tell you that what I see is a knife and fork, or that the object I see is red, and thereby tell you exactly what I see—in the first sense of “see”; and, if all goes well, you may thereby come to know what I see (first sense) as well as I do, even though you have not yourself perceived the object. Not so with aspects. For illustration, consider the duck–rabbit, keeping in mind that this is just one example, and not in all respects a representative one, of the wider set of phenomena that concerns Wittgenstein in his remarks on aspects. What do you see when, aware of the ambiguity of the figure, you see not merely the duck–rabbit (which is an object of sight of the first category, which may be described and thereby identified geometrically) but, say, the rabbit aspect? What do you see when you see the duck–rabbit as a rabbit? The obvious answer would seem to be “a picture-rabbit” (or maybe “a rabbit”) (PI, PPF: 120–1); and if you were asked what that (i.e. a picture-rabbit, or a rabbit) was, you could point to non-schematic pictures of rabbits, or to real rabbits, etc. (ibid). But note the important sense in which saying “rabbit” or “picture-rabbit,” or pointing to a non-ambiguous rabbit, whether flesh and blood or depicted, as a way of specifying what you saw, would be misleading: it would suggest that you were somehow unaware of the ambiguity of the figure and of the possibility of seeing it differently, and unaware of the active role you play in casting the rabbit aspect onto the figure, so to speak, whereas we are here supposing that this is not the case.8 What you see when you see the duck–rabbit as a rabbit (say) is, therefore, well, this—and now one would like simply to point to the duck–rabbit, perhaps with the addition of hints to help the other see the rabbit aspect, if for some reason she has not yet been able to see it.9

56 Avner Baz The grammatical-phenomenological characterization I have just given of Wittgensteinian aspects is fairly specific; and yet it allows for quite a range of cases that differ from each other in more or less significant ways. Let me mention some of them: seeing a similarity between two faces; seeing the duck–rabbit as a duck or as a rabbit; seeing a figure such as the famous Necker cube as oriented one way or another in space; seeing the doublecross as a white cross against a black background, or vice versa; seeing a triangle—either drawn or “real” (three-dimensional)—as pointing in this or that direction, or as hanging from it apex, or as having fallen over . . . (PI, PPF: 162); seeing a face in a puzzle-picture; seeing a sphere in a picture as floating in the air (ibid: 169); seeing W as an upside-down M and seeing the letter F as facing right or left (Wittgenstein 1980b: 464–5); the aspect we may be said to see when something strikes us in a picture of a running horse and we exclaim “It’s running!” (1980a: 874; see also PI, PPF: 175); hearing a piece of music as plaintive (PI, PPF: 229) or as solemn (ibid: 233), or hearing a bar as an introduction (ibid: 178); the experience in which “everything strikes us as unreal” (1980a: 125–6). . . An important thing to note is that aspects may be seen in non-ambiguous figures: for an aspect to dawn on us, there need not be, and often there is not, two (or more) competing, determinate aspects under which the object may be seen. Thus, for example, there is no clear, determinate aspect that competes with the similarity of one face to another, and which that similarity, when it strikes us, might plausibly be thought to have replaced. Even in cases where it seems that there are two or more determinate aspects under which an object may be seen, this does not mean that we must always be seeing that object under one of them. For example, if you invite me to see, and say, which way the letter F is facing, and I look and it strikes me that it is facing right (say), this does not mean that every time I see the letter F I see it as facing right or else as facing left. This will become important for us later on, when we will ask what sense can be given to the recurrent idea that all (normal) seeing is seeing-as—that everything we see, or at least almost everything we see, is seen under some particular, determinate aspect or another. Another important point is that in some of the above cases, the aspect corresponds to no objective judgment—what the object is seen as is not something that (in a different context perhaps) it could be seen, or known, to be. What would it be, for example, for the letter F to be facing right or left? Moreover, even where we could think of an objective judgment that might be thought to correspond to the aspect—given a suitable context, the duck–rabbit could actually serve as a picture of a rabbit or a duck, and the Necker cube could be (meant to be taken as) an illustration of a cube going this (rather than that) way; a triangular wooden block that stands on its longest side could actually have fallen over (it might be that it is supposed to stand on its shortest side), and a drawn triangle might (be meant to) represent a triangle that has fallen over; there might actually be an objectively

Aspects of perception 57 establishable similarity between two faces; and so on—no such judgment is actually made by the perceiver of the aspect; and in the typical case, the perceiver of the aspect makes it clear that what she sees the object as is not something that she takes it to be. This is why we normally invite the other to see the aspect, and why we do not take her to be mistaken (or literally blind) if she cannot see the aspect we see. This is going to matter when we turn to Wollheim, for whom the typical, or anyway basic case of seeing-as is precisely one in which the object is judged, or believed, to be what it is seen as.

Wollheim on seeing-as According to Wollheim, seeing-as is “an essential part” of the capacity for “straightforward perception,” which he explicates in terms of “the capacity that we humans and other animals have of perceiving things present to the senses” (1980: 217). A little later on he similarly proposes that seeing-as “partially is, partially is a development out of, an aspect of straightforward perception” (ibid: 219). How so? Wollheim explains: Whenever I straightforwardly perceive something, which ex hypothesi is present to the senses, my perception of it is mediated by a concept, or in perceiving it I subsume it under a concept. For any x, whenever I perceive x, there is always some f such that I perceive x as f. But it is crucial to an understanding of seeing-as to recognize that my seeing x as f is not just the conjunction of my seeing x and my judging it to be f. Such a view, which has gained currency amongst perceptual psychologists who talk of perception as hypothesis, errs in that it leaves the judgment external to the perception. It was just this view that Wittgenstein tried to combat when he asked us to consider cases where we switch from seeing something or other as this to seeing it as that. For the relevance of such cases is that they allow us to observe how experience and concept change not merely simultaneously but as one . . . [T]he fundamental point . . . is that, when I see x as f, f permeates or mixes into the perception: the concept does not stand outside the perception, expressing an opinion or conjecture on my part about x, and which the perception may be said to support to this or that degree (ibid: 219–20). The claim that for any perceived object x, whenever I perceive x, there is always some concept f such that I perceive x as f is a very strong claim. So it is worth noting that Wollheim in effect retracts it no sooner than he has made it. For he goes on to describe cases where for one reason or another one cannot at first tell what one sees, and only comes to see the object as f when told by someone else that it is (an) f, or when finally recognizing it to be (an) f “after considerable effort” (ibid: 221). Under what concept are we supposed to be seeing the object before we come to know it to be (an) f?

58 Avner Baz Wollheim does not say. Nor does it seem even remotely plausible that our visual experience (ibid: cf. 223) changes whenever we come to know of a hitherto unrecognized perceived object that it is an f, as happens, for example, when someone tells us that a tree “blurred by the mist” is an oak (ibid: 221) or when we find on closer examination that a tree that has been “damaged, or lopped, or covered with creeper,” and therefore was initially hard for us to recognize, is an oak (ibid). Another immediate difficulty is that Wollheim’s basic claim that “we cannot see something as something it (or its counterpart) could never have been” (ibid: 222) seems to fly directly in the face, not only of some of Wittgenstein’s examples—in what sense could the letter F have been facing right, or left?—but also of some of Wollheim’s own examples. How is this claim supposed to be true of the case of seeing a church as an overturned footstool (ibid: 222) or a mountain range as a naked woman’s body (ibid: 222)? If any sense could be given to the claim that a church could have been an overturned footstool or that a mountain range could have been a naked woman’s body, then, in that sense, anything could have been anything else, and the condition is empty. I set these difficulties aside, and turn to what Wollheim calls “the simplest case” (ibid: 220), for if it turns out that we cannot even make sense of the simplest case as a case of seeing-as, then whether Wollheim’s general account could somehow be made to accommodate the cases he regards as less simple will become significantly less important. The simplest case, Wollheim says, is when the concept arises in the mind along with the perception, and having thus arisen, what it does is to give content to a belief. The concept f enters the mind along with the perception of x, blends with this perception, and stays in the mind to form the belief that x is f. So I look out of the window of a train and see a tree which I straightaway see as an oak, which I thereupon believe it to be. (ibid: 220) So I look out of the window of a train and see a tree which I immediately know—recognize—to be an oak. What sense can be made of the idea that I thereupon enjoy a “particular visual experience” (ibid: 223), which may be described by saying that I “see the tree as an oak”? Of the equivalent idea in the case of a knife and fork, or a conventional picture of a lion, or the letter F, Wittgenstein says that it makes no sense. It makes no sense to say one sees, or tries to see, an object as what one knows it to be, Wittgenstein says (PI, PPF: 122, 203). Once again, it is open to Wollheim, just as it is open to anyone else, to give sense to “seeing something as what we know it to be,” by making clear how he uses, or means it—how his words are to be understood. But as far as I can see, all we get from Wollheim in this respect is the highly metaphorical talk of a concept “entering the mind and

Aspects of perception


blending with the perception” or of a concept “permeating or mixing into the perception.”10 Do we understand this talk, or does it only give us the illusion of sense? In order to even begin to understand Wollheim’s talk, or try to, we need to know what he means by “concept.” Wollheim says nothing to elucidate what he means by that word. He appears to be counting on something like the common understanding of “concept,” but it is not clear what that might be.11 Nor is it clear how any such understanding could serve his purposes. If we follow Wittgenstein (and Austin), and begin by reminding ourselves how the word “concept” functions in ordinary and normal discourse, the first thing we will find is that the word is not used very frequently. When it is used, “the concept of x” is often interchangeable with “the meaning of ‘x’,” and means something like “whatever it is that guides us in our use of the word ‘x’ (and its cognates).”12 Our everyday criterion for “possessing the concept of x,” and similarly for “knowing the meaning of ‘x’ (or what ‘x’ means),” is the ability to employ “x” (and cognates) competently in a wide enough range of contexts, and to respond competently to other people’s employment of it. But if this is roughly what “concept” ordinarily and normally means—what we ordinarily and normally mean by it—then to possess any one concept is to possess very many others and to master a wide range of interrelated practices. And if so, it is not clear what Wollheim’s “the concept enters the mind and blends with (or permeates, or mixes into) the perception” might mean. Here it might be objected that Wollheim is relying not on the everyday use and understanding of “concept” but rather on its more or less technical use in philosophy, as well as in psychology and linguistics. Let us see whether any such understanding of Wollheim’s “concept” could help us understand what he means by “seeing-as.” As commonly used in philosophy, psychology, and linguistics, “the concept of x” means roughly “whatever it is that guides us in classifying items in the world as (belonging to the category of) x, and in distinguishing between correct and incorrect classifications (relative to some concept, of course).” This rough understanding of concept may also be given a linguistic turn: “our concept of x is whatever it is that guides us in applying ‘x’ to cases (or withholding ‘x’ from cases), and in distinguishing between correct and incorrect applications,” where—in stark contrast with Wittgenstein’s understanding of language—the “application” of words to cases is taken to be something that we ought to be able to do, and do mostly correctly, even apart from any context of significant use of those words.13 Most of the academic disagreements in recent years about the nature of concepts, both within and outside philosophy, occur within the framework of this broad characterization.14 The disagreements, in other words, are about what guides our classifications of worldly items, or our applications of words to cases—whether it is rules or necessary and sufficient conditions, or proto-theories, or prototypes or exemplars and ways of measuring an item’s similarity to them, or “family resemblance,” and so on.

60 Avner Baz Given the above, common understanding of “concept,” the following dilemma may be posed for Wollheim. On the one hand, the more cognitive, or even theoretical, one takes concepts to be—the more one packs into one’s understanding of “concept” things (rules, conceptual entailment relations, contextual effects, broad practical considerations, Wittgensteinian grammar . . .) that are not directly perceived, certainly not by the eyes as they lay on a more or less familiar and recognizable object—the harder it should become for one to make sense of the idea of a concept mixing into or permeating the perception of the object and thereby giving rise to a particular visual, or otherwise perceptual, experience. On the other hand, if we go in the opposite direction and take concepts to be (not theoretical or cognitive but) essentially perceptual entities—that is, if we take “the concept of x” to refer to something like a visual (or otherwise perceptual) schema of x-in-general, or what an x should look like—then it is not clear what a concept thus understood could add to our perceptual experience of the particular object. For example, let’s assume that I have in my mind a perceptual schema of dog-in-general, which enables me to recognize dogs as belonging to the particular category of dogs, and to refer to them correctly by means of the word “dog” (or its equivalents in other languages I know). And suppose that here in front of me is Henry, the neighbors’ mixed German shepherd. What could my dog-in-general schema, or even my German-shepherd schema, possibly add to my visual experience of Henry? I should add that the second, perceptual way of understanding “concept” is anyway problematic, for it conflates what Charles Travis (2013: 185–7) calls “recognitional capacities” and what he calls “conceptual capacities”: it fails to distinguish between detectors (of objects of a certain type)—however reliable they might be under normal conditions—and what Travis calls thinkers. The latter, unlike the former, know, at least to some extent and with respect to very many types of objects, what makes objects count, in general or in given contexts, as objects of those types, and are capable of judging that a particular object—for example, an oak tree that has been damaged, or lopped, or covered with creeper (Wollheim 1980: 221), or a table that stands on only one leg—belongs to that type, or may in certain contexts reasonably be taken to belong to it, even though it does not display the usual perceptual features. Conceptual capacities are indefinitely flexible and potentially openended in a way that merely recognitional capacities are not. Now, it may well be that for the most part we relate to objects in our world as detectors, not as thinkers (in Travis’s sense). It may be that much of our talk, for example, is drawn out from us by the world as it presents itself to us—where this refers not just to the world we speak of, but also, and importantly, to the world we speak in—without any prior reflection on our part on the appropriateness of that talk. Not every time that we respond to the world with words, let alone without words, must we be giving voice to what may sensibly be called “a judgment.” But, if so, then Wollheim’s idea

Aspects of perception


that everything we perceive is perceived under some concept or another is a distortion—an intellectualization, so to speak, of human perception that the phenomenon Wittgenstein refers to as seeing-as may actually help us find unsatisfying.

Aspects and concepts Wollheim’s “simple case”—the case of perceiving something that we immediately recognize—know—to belong to some particular type, or, if you will, to fall under some particular concept—cannot be understood as a case of seeing-as. More precisely, seeing-as, as Wollheim invites us to understand it, cannot be made sense of when applied to such a case. And yet Wollheim is not alone among readers of Wittgenstein’s remarks on seeing-somethingas-something who have come to think that there must be a continuous version to seeing-as, and that, somehow, human perception as such must be understood in terms of seeing-as.15 Three ideas have fed into this widespread conviction. The first is the idea—which some have attributed to Kant, incorrectly it seems to me—that our perception is necessarily and always “conceptualized,” or, as Wollheim puts it, “mediated by concepts” (1980: 219).16 As I have begun to show in the previous section, it is not clear what this idea comes to, exactly; but it has nonetheless appealed to many and has exerted much influence in modern Western philosophy up until the present. Since I have criticized that idea elsewhere, I will not focus on it here. Some of what I will say, however, and some of what I have already said, has more or less immediate and clear bearing on it.17 The second idea is that the dawning of aspects, as described by Wittgenstein, could only be understood as occurring against the background, so to speak, of a state describable as continuous aspect perception. Differently put, the idea is that the aspect that dawns must be replacing some other determinate aspect that had been perceived up until the dawning of the new aspect. The third idea, which bridges the first two, is that what dawns on us when a Wittgensteinian aspect dawns on us may be identified with, or in terms of, a concept. This third idea may be expressed by saying, as Wollheim does, that what something may be seen as must be something that it could be judged to be. In the later section “Perceptual indeterminacy” I will briefly discuss the second idea, which has an important grain of truth in it but has hitherto been spoiled by the tendency to overlook the indeterminacy of the perceived, or phenomenal, world. In this section and the next one I will discuss the third idea. I will argue that Wittgensteinian aspects may not aptly be identified with, or in terms of, what may sensibly be called “concepts.” Echoing Wollheim’s metaphorical talk, I could also put the point by saying that what permeates or mixes with or blends into our perception of an object when a Wittgensteinian aspect dawns on us may not plausibly be taken to be a concept. This is not to deny, however, that the power that we have to give unity and sense, or significance, to what encounters us in experience—a power that comes out in the

62 Avner Baz dawning of aspects—is not the power that, ultimately, makes it possible for us to perform, when we do, what may sensibly count as acts of judgment, or the application of concepts. For the sake of clarity and avoidance of repetition, I will be using as my stalking horse Wittgenstein’s example of being struck by the likeness of a face we are looking at to another. Though there is a great variety of cases of aspect dawning that differ from each other in significant ways, if the idea that dawning aspects may be identified with, or in terms of, concepts can be shown to be misguided in this case, then I think we will have a good reason to be generally suspicious of it. The reader is invited to extrapolate to other cases. One important advantage of this case is that, unlike cases of seeing aspects of schematic drawings and ambiguous figures, being struck by the likeness of one face to another is something that may naturally happen to us in the course of everyday experience. And this is important, for if we wish to learn something general about human perception from the phenomenon of aspect dawning, we had better take into account the artificiality of some of the cases Wittgenstein discusses, and the differences between them and the more natural cases. The candidate empirical concept in the case of the dawning of a similarity of one face to another is, I suppose, that of bearing (some) visible similarity to a particular, given face. Being a concept, it is general: it allows for indefinitely many instantiations that differ from each other in any number of ways; and it transcends any finite set of instantiations: for any particular face, and for any finite set of faces that may all correctly be judged to bear visible similarity to that face, there could always be another face that is visibly distinguishable from all of those faces and yet may correctly be judged to bear visible similarity to the first face. One could go a step further and argue that any two faces may, in some contexts, correctly be judged to bear some visible similarity to each other. This leads us to the further point that, as Charles Travis has taught us to recognize, the concept of bearing visible similarity to a particular face, just like any other empirical concept, is “context-sensitive”: for any given face, and for a wide variety of faces that in some contexts would correctly count as bearing visible similarity to it, there could be other contexts in which those same faces would not correctly count as bearing visible similarity to that face. This means that in judging that one face bears (or does not bear) visible similarity to another, we are beholden, not just to the two faces, but also to the context in which we make the judgment. And if someone were to ask us, apart from a context suitable for fixing what “bear visible similarity” means (in that context), whether two given faces bear visible similarity to each other, the correct response would be “yes, or no, depending on what you mean.” I think the above reminders should already give pause to anyone who wishes to claim that what blends with or mixes into our perception of a face, when its likeness to another face strikes us, is a concept, not because they show that the claim might be mistaken, but because they show that it is not

Aspects of perception 63 even clear what exactly is being claimed. But let’s move closer. Concepts, at least as commonly thought of in Western and contemporary Analytic philosophy, are paradigmatically applied to cases in objective, truth-evaluable, judgments. As noted in the previous section, however, it is important that the case Wittgenstein describes is not one of judging that the one face is similar to the other. In fact, it seems essential to at least many of the cases Wittgenstein discusses that what we perceive something as is something we are not taking it, let alone claiming it, to be. This might be thought to be accommodated by Wollheim’s allowing for possible “developments” of seeing-as (1980: 220), beginning with “the simple case” and moving, in one dimension, along a series of “declining degrees of assent, diminishing from belief through likely supposition, informed guess, outside bet, to the case where there is no commitment at all to the satisfaction of the concept by the object and imagination or make belief takes over” (ibid: 221). But where exactly in this series should we place Wittgenstein’s case of being struck by the similarity between two faces? The case he describes fits nowhere in Wollheim’s series. The person struck by the similarity is not imagining (let alone supposing or believing) that the face she is looking at satisfies the concept of bearing some visible similarity to the particular other face. In being struck by the similarity between two faces, I am not imagining that they are similar. Nor am I imagining a counterfactual state of affairs in which they would be. Wollheim’s invocation of imagination and make-belief is doubly misleading. First, the aspect is not something we merely imagine (to be there). If it were, calling upon others to see what we see, as we characteristically do when an aspect strikes us, would not make sense. (By this I do not mean to say that no role is played in the dawning of an aspect by something we may call “imagination.” I actually suspect that a certain equivocation on “imagination” has led Wollheim, and perhaps some of his readers, to miss the inaptness in this case of what he means, or must mean given his overall account, by “imagination.”) Second, the aspect may not aptly be identified with, or in terms of, an empirical concept. If a concept is something that may contribute to the content of judgments or Fregean “thoughts,” however hypothetically or even counterfactually entertained; if, in other words, the application of the concept of f to a case is what may be expressed by asserting, or even just hypothesizing, that the case is (a case of) f, then what dawns on us when a Wittgensteinian aspect dawns on us is not a concept, nor may it be identified in terms of one. The empirical judgment that something is f, and so, if you will, a subsumption of a case under the empirical concept f, situates the object and its property of being f in the objective world—within what Charles Travis calls “networks of factive meaning” (2013: 91). A particular face’s being similar to some other particular face, for example, factively means certain objectively establishable things (and indicates or makes likely certain other things), where “factively” here means: if those other things do not hold, then either the similarity of the one face to the other does not mean them or the similarity

64 Avner Baz does not hold. Because empirical judgments, and more broadly Kantian “cognitions” (Erkenntnisse), are interconnected and form a system—the system Kant calls “nature”—they commit those who make any one judgment to indefinitely many other Kantian cognitions, or Fregean thoughts. They also commit them practically. Empirical concepts, understood as constituents of empirical judgments (or cognitions), or as what those judgments apply to cases, may accordingly be thought of as individuated or defined by those commitments, regardless of whether particular applications of them are committed or somehow uncommitted (hypothetical, counterfactual).18 If I judge that one face is visibly similar to another, for example, then I am committed to expecting all normal and competent people who are suitably positioned to recognize this; and I am committed to holding those who deny the similarity to be mistaken, and to be liable to err practically as a result; and I am committed to taking it that each of the two faces, or some feature(s) of it, may be pointed to as a way of giving someone (some) information about the other face (“The escaped suspect’s face is similar to so and so’s face”); and I am committed to there being certain objectively establishable features of the faces that are responsible for the similarity; and I am committed to being able to identify those features—to specify in what the similarity consists (“They have the same pointy and slightly crooked nose”); and so on and so forth.19 If I am not thus committed, I have not thus judged. It is true that when I merely imagine that, or a state of affairs in which, one face bears a visible similarity to another I do not commit myself as I do when I judge that it does. But what I imagine may still be defined or specified in terms of the same set of commitments:20 what I imagine is, precisely, a state of affairs in which there is an empirically establishable visible similarity between the faces, where that means a situation in which the two faces are such that normal and competent perceivers who are suitably positioned may rightfully be expected to find them similar to each other (given a suitable context, but I will not always be careful to add that); and in which those who denied the similarity would be mistaken and would be liable to err practically as a result; and in which each of the two faces, or some feature(s) of them, would be such that it could be pointed to as a way of giving someone some information about the other face; and in which the faces have certain objectively establishable features—identifiable by normal and competent perceivers who are suitably positioned—that make them alike; and so on and so forth. A Wittgensteinian aspect, by contrast, is not similarly situated within a network of factive meanings; it is not a feature of the objective world. While my being struck by the similarity between two faces is an objectively establishable fact, and as such means, factively, any number of objectively establishable things (mostly things having to do with me), the similarity between the faces that strikes me does not (factively) mean, or indicate, or make likely, anything objectively establishable. It is not part of the objective (“external”) world. Nor, as noted, is it imagined to be part of the objective

Aspects of perception 65 world. But, if so, then it may not aptly be identified in terms of the empirical concept of (visible) similarity. And yet the aspect is not merely subjective. It is there, in the perceived face. And though I cannot objectively establish its presence, or describe it geometrically, or prove wrong those who fail to see it, I still take it that others could be brought to see it there too, and I take it that they are missing something about the face if they don’t. Though the (Kantian, objective) “I think” could not sensibly accompany our experience of the aspect, neither the experience nor the aspect are “nothing to us” (contra Kant 1998: B131). A long tradition, beginning with Kant’s Critique of Pure Reason (at least in one popular reading of it) and exerting its strong influence all the way to the present, would have us suppose that only the subsumption of what presents itself to us in our experience under concepts—thought of as systematically interrelated rules for the unification and organization of the “sensible manifold”—could enable us to move from the merely “inner” or “subjective” succession of Vorstellungen to a world sharable with others (Kant: 1998, A196–7/B242). Part of what Wittgenstein (or, in some readings, already Hegel) has taught us to recognize is that what may sensibly be called “the application of concepts to cases” could itself only truly be understood in terms of inter-subjectively shared practices into which we are initiated, and in which we participate, in a world that is, to some degree, always already shared with others. This is a point of deep agreement between Wittgenstein and phenomenologists such as Heidegger and, especially, Merleau-Ponty. But as those phenomenologists have taught us to recognize, it is extremely difficult to describe without distortion our perceptual relation to the notyet-objective and yet inter-subjectively sharable, and largely shared, world. In particular, it is extremely difficult to resist the temptation to objectify the perceived world, and to think of our relation to it in terms of the very same empirical concepts whose application may only be understood, if Wittgenstein and the phenomenologists are correct, against the background of that very relation. And it is here that the dawning of Wittgensteinian aspects could prove philosophically enlightening.

Aspects as perceived internal relations Wollheim, as we saw, complains about the tendency to “leave judgment external to perception” (1980: 219). The whole point of Wittgenstein’s asking us to “consider cases where we switch from seeing something or other as this to seeing it as that,” he says, is that those cases “allow us to observe how experience and concept change not merely simultaneously but as one” (ibid: 220). I find the main interest of aspect dawning to lie in its showing something like the opposite of what Wollheim takes it to show. Far from bringing out the inseparability of judgments (or Kantian cognitions) and perception, it brings out the important distinction between those two—a distinction that the tradition of Western philosophy has tended to obscure.

66 Avner Baz “What I perceive in the dawning of an aspect,” Wittgenstein writes, “is not a property of the object, but an internal relation between it and other objects” (PI, PPF: 247). The notion of internal relation (interne Relation) is drawn from Gestalt psychology and is, importantly, a perceptual notion, not an objective, third personal notion.21 Among elements of the objective world only external relations may hold. And these are the elements, and relations, that empirical concepts enable us to grasp, or constitute. Two (or more) perceived things (objects, elements) stand in an internal relation to each other when their perceived qualities are not independent of the perceived relation between them. Here is a passage from Kurt Koffka that illustrates the notion: “Two colors adjacent to each other are not perceived as two independent things, but as having an inner connection which is at the same time a factor determining the special qualities A and B themselves” (2007: 221). According to Gestalt psychology, what we perceive, at the most basic level, is not atomic sensations that we must then somehow synthesize into significant wholes, but rather unified, significant wholes, where the perceived qualities of the elements of a perceived whole, and so the specific contributions those elements make to the overall perceived significance of that whole, are not perceptually independent from that perceived overall significance. The duck–rabbit provides a simple illustration of this. When you see it as a rabbit, say, you see the two “appendages” as ears; but your seeing them as ears is not independent from your seeing the whole thing as a rabbit. Perceptually, the ears are (seen as) ears only when the whole thing is (seen as) a rabbit. One important thing this means is that your seeing the duck–rabbit as a rabbit cannot be explained as the outcome of your seeing this portion of the drawing as ears, that portion as mouth, another portion as the back of the head, and so on. The rabbit aspect is not synthesized from elements that have their “rabbit-parts” significance independently of being elements of that overall aspect. On the other hand, if you took the basic elements of our perception of the duck–rabbit to only have objectively establishable, geometrical properties, and so devoid of any rabbit (and equally duck) significance, then you would never be able to explain, on that basis, why those elements got synthesized into the rabbit aspect, say, rather than the duck aspect. This shows that the perception of significant wholes should be taken as primary. Wittgenstein gives clear, if also characteristically non-theoretical, expression to this fundamental feature of human perception, in the following remark: Look at a long familiar piece of furniture in its old place in your room. You would like to say: “It is part of an organism”. Or “Take it outside, and it is no longer at all the same as it was”, and similar things. And naturally one isn’t thinking of any causal dependence of one part on the rest. Rather it is like this: . . . [I]f I tried taking it quite out of its present context, I should say that it had ceased to exist and another had got into its place.

Aspects of perception 67 One might even feel like this: “Everything is part and parcel of everything else” (external and internal relations [my emphasis, AB]). Displace a piece and it is no longer what it was. (1980a: 339). In the world as perceived, prior to reflection and objectification, everything is internally related to everything else. Another case of Gestalt perception, which is at the heart of Wittgenstein’s understanding of philosophical difficulty, is that of linguistic meaning, or sense. In Wittgenstein’s view, which may be seen as a development of Frege’s “context principle,” the basic unit of linguistic sense is neither the isolated word, nor the isolated string of words, but an utterance—a human act performed against the background of the history of the language, the culture, and of the individual participants. Phenomenologically—from the perspective we all occupy as speakers engaged in discourse (as opposed to theoreticians reflecting on it academically)—analysis presupposes synthesis: the contribution made by each word to the overall sense of an utterance is not independent from, and therefore cannot analytically explain, that overall sense. “In understanding others, the problem is always indeterminate,” Merleau-Ponty writes, “because only the solution to the problem will make the givens retrospectively appear as convergent” (1996: 184, 408–9). It is important to note that internal relations hold not just among the perceived elements of perceived things but also, and equally fundamentally, between perceived objects and the background against which they are perceived. This is likely to be missed by those who mostly focus on the examples of schematic drawings and deliberately ambiguous figures. These objects are typically encountered in the artificial context of a psychology lab or philosophy classroom. They are therefore “cut off,” as Merleau-Ponty puts it, from our perceptual field, with its temporally structured and extended personal and cultural “horizons”; and this is what makes it possible for us to give them significances, or project different physiognomies on them, more or less at will (1996: 282). Even here, however, the perceived objects stand in internal relations to other objects, as Wittgenstein suggests; but the way in which foreground and background are internally related in normal perception, and therefore change together, does not come out clearly in their case. It comes out far more clearly in the more natural cases of aspect dawning. (That the analysis of perceptual experience presupposes its synthesis and therefore cannot explain it was one of Kant’s fundamental insights and his most basic objection to empiricist-mechanical accounts of how unity arises in our experience. Kant saw that we must play an active role in bringing about—“constituting,” as the phenomenologists later came to say—the unity of our experience. What Kant missed, the phenomenologists have argued, is the possibility of a synthesis that, while in some clear sense intelligible and inter-subjectively shareable, is not (aptly thought of as) conceptual. The dawning of Wittgensteinian aspects brings out especially clearly the distinction

68 Avner Baz between what we perceive and what we objectively think (or know), and the reality of inter-subjectively shareable, non-conceptual, perceptual synthesis.) Now go back to the experience Wittgenstein describes of being struck by the similarity between two faces. A similarity understood as an objective property of the faces is an external relation between them: each face has its objective properties, which one may come to know without knowing anything about the other face, and those properties determine whether, and, if so, to what extent, the two may count (context-dependently) as bearing some objective similarity to each other. And so you may look at a face and see (first sense), or have someone point out or demonstrate to you, that there is some visible similarity between it and another, where seeing that need not involve, or bring about, any change in how you visually experience the face you’re looking at: its perceived Gestalt (physiognomy, expression) need not change at all. By contrast, in the experience Wittgenstein describes, the perceived Gestalt of the face you’re looking at changes; and what dawns on you here is an internal relation between the one face and the other, precisely because the perceived relation—of similarity—is inseparable from the perceived change in the overall physiognomy or expression of the face. The perceived qualities of each of the two faces that make them bear a similarity to each other are not independent, perceptually, from our perception of the similarity. (Again, they could be: we could recognize an objectively establishable similarity between the faces—a similarity that may simply be known to be there, and which does not depend on anyone’s visual experience of the face. But that would not be the seeing of a Wittgensteinian aspect—the seeing of one thing as another. As Wittgenstein notes, even the person he calls “aspect-blind” and defines as someone “lacking in the capacity to see something as something” should be able to recognize objective similarity and “execute such orders as ‘Bring me something that looks like this’” (PI, PPF: 257)).

Perceptual indeterminacy As we have seen, judging that one face is similar to another, or otherwise conceiving of a similarity between them, is one thing, having the similarity between them perceptually dawn on one another. This distinction shows itself as well in the less natural cases of Wittgensteinian seeing-as. Thus, for example, it is one thing to take, or consider, the duck–rabbit to be a picture of a duck (say), or know it to be meant to serve as such a picture, and quite another thing to see it as a duck. Similarly, it is one thing to (cognitively) take the Necker cube as meant to represent a cube going this rather than that way, and quite another thing to be able to see it as going this or that way. Again, the former is something that even the “aspect blind,” as Wittgenstein describes him, could do. If he could not, his handicap would be even severer than aspect-blindness. “Ordinary experience,” Merleau-Ponty writes, “draws a clear distinction between sense experience and judgment” (1996: 34). He appeals to cases

Aspects of perception


where we know, or think, one thing about what we perceive, but perceptually experience something else. One of the examples he appeals to is the Necker cube: A cube drawn on paper changes its appearance according as it is seen from one side and from above or from the other and from below. Even if I know that it can be seen in two ways, the figure in fact refuses to change its structure and my knowledge must await its intuitive realization. Here again one ought to conclude that judging is not perceiving. (ibid: 34) As Wittgenstein notes, seeing something as something requires that you attend to the object in a particular way. It therefore could not be one’s ordinary or habitual relation to the object, could not be continuous—not even if we willed it to be, for what we attend to and how is not normally subject to the will, and is never subject to the will for long. What could be continuous is, precisely, a cognitive relation to an object—cognitively taking it to be one thing, or type of thing, or another (Wittgenstein 1980a: 524). Those who, like Wollheim, have taken the continuous seeing of aspects to be unproblematic, have invariably conflated the question of how you see something—how it organizes itself under your gaze, so to speak—and how you conceive of it, or what you cognitively take it to be. They have taken what we know (or take ourselves to know) we perceive—that is, objects of sight of Wittgenstein’s first category, determinate objects determinately situated in the objective world—to determine what we actually perceive, in the sense of how things in fact present themselves to us in our experience. They have taken the objective world—the world about which science has ultimate authority—to be, or determine, the world as perceived. They have thus committed what Merleau-Ponty, following Husserl and Gestalt psychologists, calls “the experience error” (1996: 5). The dawning of Wittgensteinian aspects reveals the experience error to be an error. It shows that there is a perceived physiognomic unity, or overall sense, that is importantly different from the unity, or overall sense, capturable in Kantian Erkenntnisse. And it thereby shows that the “constancy hypothesis” is false: there is no one-to-one correlation between the world we objectively know we perceive and the world as perceived.22 And, arguably, it shows even more than this: that the perceived unity or sense of the world as perceived is importantly indeterminate. The dawning of a Wittgensteinian aspect is not normally the replacement of one determinate physiognomy by another determinate physiognomy. Rather, it is the necessarily passing replacement of an indeterminate physiognomy with a relatively determinate one: a particular way of taking hold of the object perceptually replaces, for a time, not (normally) some other particular way of taking hold of it but, rather, no particular way of taking hold of it. Early on in the Phenomenology of Perception Merleau-Ponty says that “We must recognize the indeterminate as a positive phenomenon” (1996: 6)—that

70 Avner Baz is, not as due to some kind of contingent limitation of our cognitive or perceptual powers. This is one of the most difficult ideas in his account of perception—difficult both to understand and to accept. I will not here try to explicate and defend the idea, which is tied to Merleau-Ponty’s characterization of the internal relation between foreground and background in perception, to his discussion of the inherent ambiguity in human experience between the “personal” and the “impersonal,” or “anonymous,” and to his understanding of the temporal structure of perception. I do want to propose, however, that the dawning of Wittgensteinian aspects may be seen as an illustration of, and at the same time as lending support to, Merleau-Ponty’s idea. This is likely to be missed, and has in fact been missed, by those who mostly focus on the dawning of aspects in the artificial cases of ambiguous figures. In the case of the duck–rabbit, for example, it just seems obvious that the determinate aspect that dawns replaces another, equally determinate aspect under which the object had been seen up until the dawning of the new aspect. Here it would help to remind ourselves of some of Wittgenstein’s other examples of aspect dawning, such as the case we discussed in which one is all of a sudden struck by the similarity of one face to another. Here there does not seem to be any plausible candidate for the competing aspect under which the face had been seen up until right before the dawning of the new aspect. We had been seeing the face all right, but not as having some particular, determinate overall expression or physiognomy—not unless we had been struck, for some time and up until the dawning of the similarity, by some such expression or physiognomy. Nor would it help to insist here that we had been seeing the face continuously as a face, for, as Wittgenstein notes, that insistence makes no (clear) sense, and, in any case, that alleged aspect does not disappear when the “new” aspect dawns. So the phenomenon of aspect dawning, far from showing that everything we perceive is perceived under some determinate aspect or another, should actually make us suspicious of that idea. Those who take the idea of continuous aspect perception to be clear and unproblematic, Wittgenstein suggests, invariably conflate how we see something and how we conceive of it; and then they attribute to the former the relative determinacy and stability that characterize the latter.23 Thus, when we say “I’ve always seen it in this way” what we really mean to say, Wittgenstein suggests, is “I have always conceived it this way (Ich habe es immer so aufgefasst), and this change of aspect has never taken place” (Wittgenstein 1980a: 524). And if we find ourselves tempted to say that there is some particular aspect under which we’ve always seen a face, he further suggests, then we should try to say what that aspect is—how we’ve always seen the face, for as soon as we describe the aspect in some way, Wittgenstein says, it will become clear to us that we have not always seen the face under that aspect (ibid: 526). But is the dawning aspect determinate? We have already noted that it is necessarily passing: it only lasts as long as we are “occupied with the object in a particular way” (PI, PPF: 237; see also Wittgenstein 1982: 14–15); it

Aspects of perception


“presents a physiognomy that then passes away” (PI, PPF: 238). For it to last indefinitely, the aspect would have to turn into a piece of knowledge, or Kantian “cognition.” It would then become, for us, an objective feature of the face, and thereby cease to be dependent on how we perceive the object. And then it would no longer be a Wittgensteinian aspect. Is the dawning aspect determinate while it lasts? That would depend, of course, on what one means by “determinate.” It is undeniable that in some cases we are readily able to describe the dawning aspect well enough to get other people to (see whether they can) see it. However, it is important to note, first, that this is not always the case. Sometimes aspects dawn on us for which we have no readily available description: something strikes us all of a sudden about the mood of a party or the spirit of a time, for example, and we struggle to put it into words, and perhaps even find that someone else is better able to do so than we are. This is one reason for feeling gratitude toward literature and poetry.24 As Juliet Floyd correctly notes, there are cases of aspect-perception [in which] there is a more open-ended range of significance: What is to be discerned is not an object or fact or concept, but a world, a human being, an expression or gesture, a total field of significance. (2010: 324) I note that cases of this kind are more clearly telling than others against the idea that whenever an aspect dawns on us there is some particular concept that corresponds to it. But let us consider those cases in which we do seem to have a readily available description of the dawning aspect. Thus we may say that we see a similarity between one face and another, for example, or that we see the duck–rabbit as a duck or as a rabbit. Surely, however, “a similarity to some other, given face” does not capture the particular physiognomy that has dawned on us. And even the two aspects of the duck–rabbit, for all of their schematicity, have physiognomies—“quite particular expressions,” as Wittgenstein puts it in the Brown Book—that go beyond anything capturable by “duck” and “rabbit.”25 (I would go as far as to propose the following: normal human perceivers cannot attend perceptually to a face, however schematic and however unlike a human face, without seeing it as expressive, as having a particular—however passing and indeterminate—expression.) We could try to describe the dawning physiognomy further. The duck, we might say, looks serious and somewhat self-important, like a general posing for a portrait. The rabbit too looks pleased with itself, but in a more naïve or less pompous way, like a teenager driving an open-roofed convertible for the first time, taking pleasure in the feeling of freedom and speed and the wind in his hair, as well as in the thought of the (imagined) envious gazes of onlookers. Similarly, we could try to describe the similarity we see between the faces, to say how they are similar. Or it could happen that the similarity strikes us,

72 Avner Baz we call upon someone else to see it too, and then we find that the other is better able than we are to describe or articulate the similarity. But I wish to propose that no description would exhaust and finally capture the dawning physiognomy of a face whose similarity to another has struck us, or even that of the duck or the rabbit. Someone else, or us at a later moment, could see the duck as loyal and eager to please but not too intelligent, and the rabbit as stunned and taken aback by something it faces.26Any description of the aspect would be improvable, or contestable. In this and other respects, Wittgensteinian aspects—at least those that strike us in the course of natural, everyday experience—are akin to Kantian beauty.27

Concluding remark: aspects and beauty As we have seen, it is one thing to see something as x and quite another thing to conceive of it as x, or judge it to be x. And seeing something as x—I mean, the perceptual phenomenon Wittgenstein investigates under that title—cannot be continuous. The dawning of a Wittgensteinian aspect, especially when it happens in the natural course of everyday experience, is the momentary emergence, more or less willed or invited, of relative determinacy—a particular way of momentarily taking hold of what encounters us in our experience, a particular way of passingly orienting ourselves toward it. Wittgenstein’s investigation of aspect perception, far from showing, or trying to show, that everything we see is seen under some particular concept, as Wollheim proposes, rather suggests that the more or less indeterminate unity of the perceived world is neither brought about nor secured by the application of concepts. And this, interestingly enough, is an insight that Merleau-Ponty, in the preface to the Phenomenology of Perception, credits to the author of The Critique of Judgment (Merlau-Ponty 1996: xix). For beauty, as Kant characterizes it phenomenologically, is precisely a perceived meaningful unity that is not, and cannot, be captured by any available concept or set of concepts, is in this sense indeterminate, and yet for all that is experienced as genuinely perceived and as inter-subjectively sharable (Kant 2000: 240–1, 287, 292).28 What the natural dawning of Wittgensteinian aspects suggests is that Kantian beauty is perceptually prior to Kantian Erkenntnisse, and is to be found everywhere.

Notes 1 Quoted by Monk (1990: 537). 2 Beyond the very narrow sense that Wittgenstein gives it in one remark of the Investigations (PI, PPF: 118). In Baz (2000) I say what I understand Wittgenstein to be talking about in that remark. Wittgenstein is talking about something far more specific, and far less central for him, than Stephen Mulhall (1990, 2001) has made it out to be. In my reading, Wittgenstein is setting that case aside— distinguishing it from the primary phenomenon he is investigating—rather than singling it out as the true object of his interest. “Continuous seeing” of an aspect,

Aspects of perception



5 6


8 9



as Wittgenstein here uses the term, refers to the state of someone who sees an ambiguous figure—say, the duck–rabbit—but is unaware of its ambiguity. If we then asked him, with reference to the duck–rabbit, “What’s that?”, he would say simply “a duck” (say); and then it would make sense for us, who know that the picture can be seen in more than one way, to say about him that he is continuously seeing the duck aspect of the duck–rabbit. Such a person, Wittgenstein says, would simply be describing or reporting his perception (PI, PPF: 121, 128), whereas about what he calls “seeing-as” he says that it “does not belong with perception” (ibid: 137). The latter remark can only be made plausible, it seems to me, if by “perception” he means something like “objective perception”; but, either way, what matters for our purposes is that he distinguishes what he is calling “continuous seeing of an aspect” from what he elsewhere calls “seeing-as.” I make this point in Baz (2000) by saying that even the “aspect-blind”—defined by Wittgenstein as someone who lacks the capacity to see something as something—should be perfectly capable of “continuously seeing an aspect” thus understood. In Baz (2000 and 2010) I argue against Stephen Mulhall’s (1990, 2001) influential claim that what he calls “continuous aspect perception” characterizes our normal perceptual relation to the world and is the focus of Wittgenstein’s interest in his remarks on aspects. This characterization does not quite capture Mulhall’s account of aspect perception. Mulhall’s account has the important merit of emphasizing, even if ultimately mischaracterizing, the distinction between seeing—in the sense in which Wittgensteinian aspects are seen—and (mere) knowing. I am suggesting that what Juliet Floyd has insightfully called Wittgenstein’s “grammaticalizing our talk of the intuitive” (2010: 316), while it may help us dissolve any number of philosophical difficulties, may have its limitations, too. In Baz (2015a and 2015c), I argue, following Wittgenstein and Merleau-Ponty, against this prevailing understanding of language, which (I argue) has underwritten the philosophical “method of cases” and hence much of the work produced within mainstream analytic philosophy in the past 50 years or so. As a result, “aspect” as used by philosophers who present themselves as interpreting Wittgenstein has come to mean, literally, just about everything and anything one might be said to perceive. Thus, for example, Severin Schroeder writes: “[W]henever something is seen (and not only looked at inanely or absent-mindedly) some aspect of it must be noticed, be it only certain shapes or colors” (2010: 366). But how exactly, or in what sense, is the color of an object or its shape an aspect? Surely not in Wittgenstein’s sense. And why are aspects, thus understood, philosophically interesting? That other case is what Wittgenstein calls the “continuous seeing” of an aspect. See note 2. Moreover, as I will later point out, despite the schematicity of the duck–rabbit, its duck aspect and its rabbit aspect each have a physiognomy, or expression, that goes beyond anything capturable by “duck” or by “rabbit,” and which defies any attempt to put it completely and finally into words. I should say that in a couple of his remarks Wittgenstein also speaks metaphorically about how, in seeing an aspect, “we bring a concept to what we see” (1980a: 961) or how the aspect is “the echo of a thought in sight” (PI, PPF: 235). But, first of all, Wittgenstein is here trying to characterize the experience of noticing an aspect, not our ordinary and normal perceptual relation to just about everything. Second, in contrast with Wollheim, he makes it clear that the talk is not only metaphorical, but also tentative—something that “one would like to say” (ibid): it does not by itself constitute an account or explication of anything. And, third, it may be that these remarks of Wittgenstein’s, even if taken in context and with

74 Avner Baz



13 14 15 16

17 18 19 20


22 23 24

a grain of salt, are still potentially misleading or problematic in how they invite us to understand the seeing of aspects. In a recent paper Joseph Rouse points out that “a remarkable sequence of prestigious John Locke Lecturers”—he mentions John McDowell, Jerry Fodor, Frank Jackson, and Robert Brandom—“have presented and defended very different accounts of concepts or the conceptual domain.” “The disconnection among these views is substantial enough,” he suggests, “that an observer might wonder whether we philosophers have any idea (or at least any one idea) of what we’re talking about when we talk about concepts” (Rouse 2013: 250). “Concept” may also mean something like an approach to, or a way of looking at and doing things, as in “the management of the company has come up with an altogether different concept of marketing.” But that could not possibly be what Wollheim means by “concept,” or what he must mean by it given the overall story he wishes to tell. I discuss this fundamental difference between Wittgenstein and both the tradition of Western philosophy and mainstream Analytic philosophy in Baz (2012) and Baz (2015a and 2015c). As is evidenced in Margolis and Laurence (1999). All of the contributors to this comprehensive volume think of concepts from within the framework of that broad characterization. The earliest version of this idea is found in Strawson (1974). Later versions may be found in Mulhall (1990 and 2001), Johnston (1994), and Schroeder (2010). John McDowell, in Mind and World (1996) and subsequent work, has played central role in promoting the notion that for Kant all perceptual experience is conceptualized. This notion seems to me incompatible with Kant’s account of beauty. Lucie Allais has recently argued powerfully against the notion (2009). See Baz (2003). A powerful and detailed critique of the idea that human perception is “conceptualized” may be found in Travis (2013). This connects with Kant’s saying that the modality of a judgment “contributes nothing to the content of the judgment” (1998: A74/B100). This list of commitments is not meant to be complete; and it does not even matter whether it is accurate (as far as it goes). What matters for my purposes is that an accurate (even if still incomplete) such list may be given. This is really just the basic Fregean truth, emphasized by Geach (1965), that (what may count as) the same thought may be, on the one hand, asserted or otherwise committed to and, on the other hand, merely “entertained” or “considered” (or imagined to be true). Schroeder muddles his discussion of aspect perception by speaking of the similarity that strikes us as at once “an internal relation” (2010: 359) and “an objective feature of the object, namely a relation of likeness between it and some other object” (2010: 360). But a similarity thought of as an objective feature cannot, conceptually, be an internal relation. See Merleau-Ponty (1996: 7–8), who credits the notion of “constancy hypothesis,” as well as the idea that Gestalt changes refute the hypothesis, to Wolfgang Köhler. I say “relative” because even the determinacy of the objective world is historically conditioned and context-dependent. But that’s a topic for another occasion. In Dora Bruder Patrick Modiano talks about how the German occupation of Paris and its horrors have been covered up in various ways, and for the most part forgotten. “Nobody remembers anything anymore,” he writes. “And yet,” he goes on to say, “from time to time, beneath this thick layer of amnesia, you can certainly sense something, an echo, distant, muted, but of what, precisely, it is impossible to say” (Modiano 1999: 109).

Aspects of perception


25 See Wittgenstein (1958: 135). See also a note attached to PI: 165. 26 This illustrates the way in which the perceived physiognomy an object presents may change in accordance with its perceived, or imagined, background, which is one important source of perceptual indeterminacy. 27 I work out the affinity between Wittgensteinian aspects and Kantian beauty in Baz (2015b). 28 Where Merleau-Ponty goes beyond Kant, and beyond virtually everyone else in the tradition of Western philosophy, is in bringing out the way in which this preconceptual and largely indeterminate unity of the world is a unity for and in relation to not our disembodied cognitive powers but, precisely, our body. The perceived world is a field of actual and potential embodied engagement. This is why the way to try to see an object under some particular aspect is to adopt a bodily attitude toward it that fits that aspect.

References Allais, L. (2009). “Kant, Non-Conceptual Content and the Representation of Space.” Journal of the History of Philosophy 47: 383–413. Baz, A. (2000). “What’s the Point of Seeing Aspects?” Philosophical Investigations 23: 97–121. Baz, A. (2003). “On When Words Are Called For: Cavell, McDowell, and the Wording of Our World.” Inquiry 46: 473–500. Baz, A. (2010). “On Learning from Wittgenstein, or What Does It Take to See the Grammar of Seeing Aspects?” In Day, W. and Krebs, V. (eds), Seeing Wittgenstein Anew. New York: Cambridge University Press, pp. 227–48. Baz, A. (2012). When Words Are Called For. Cambridge, MA: Harvard University Press. Baz, A. (2015a). “On Going (and Getting) Nowhere with Our Words: New Skepticism about the Method of Cases.” Philosophical Psychology (forthcoming). Baz, A. (2015b). “The Sound of Bedrock: Lines of Grammar Between Kant, Wittgenstein, and Cavell.” European Journal of Philosophy (forthcoming). Baz, A. (2015c). “Recent Attempts to Defend the Method of Cases, and the Linguistic (Re)turn.” Philosophy and Phenomenological Research (forthcoming). Floyd, J. (2010). “On Being Surprised: Wittgenstein on Aspect-Perception, Logic, and Mathematics.” In Day, W. and Krebs, V. (eds), Seeing Wittgenstein Anew. New York: Cambridge University Press, pp. 314–37. Geach, P. (1965). “Assertion.” Philosophical Review 74: 449–65. Johnston, P. (1994). Wittgenstein: Rethinking the Inner. New York: Routledge. Kant, I. (1998). Critique of Pure Reason, Guyer, P. and Wood, A. (trans.). New York: Cambridge University Press. Kant, I (2000). Critique of the Power of Judgment (including the ‘First Introduction’), Guyer, P. and Matthews, E. (trans.). New York: Cambridge University Press. Koffka, K. (2007) [1927]. The Growth of the Mind, an introduction to ChildPsychology, second edition, Ogden, M.R. (trans.). New York: Kessinger Publishing. McDowell, J. (1996). Mind and World. Cambridge, MA: Harvard University Press. Margolis, E. and Laurence, S. (eds) (1999). Concepts: Core Readings. Cambridge, MA: MIT University Press. Merleau-Ponty, M. (1996). Phenomenology of Perception, Smith, C. (trans.). New York: Routledge. Modiano, P. (1999). Dora Bruder, Kilmartin, J. (trans.). Berkeley, CA: University of California Press.

76 Avner Baz Monk, R. (1990). Ludwig Wittgenstein: The Duty of Genius. London: Vintage. Mulhall, S. (1990). On Being in the World: Wittgenstein and Heidegger on Seeing Aspects. New York: Routledge, pp. 246–67. Mulhall, S. (2001). ‘“Seeing Aspects.” In Glock, H.J. (ed.),Wittgenstein: A Critical Reader. Oxford: Blackwell, pp. 246–67. Rouse, J. (2013). “What is Conceptually Articulated Understanding?” In Schear, J. (ed.), Mind Reason and Being-In-The-World: The McDowell-Dreyfus Debate. New York: Routledge, pp. 250–71. Schroeder, S. (2010). “A Tale of Two Problems: Wittgenstein’s Discussion of Aspect Perception.” In Cottingham, J. and Hacker, P.M.S. (eds), Mind, Method, and Morality: Essays in Honour of Anthony Kenny. New York: Oxford University Press, pp. 352–71. Strawson, P. (1974). “Imagination and Perception.” In Freedom and Resentment. London: Methuen, pp. 50–72. Travis, C. (2013). Perception: Essays After Frege. New York: Oxford University Press. Wittgenstein, L. (1958). The Blue and Brown Books. Oxford: Blackwell. Wittgenstein, L. (1980a). Remarks on the Philosophy of Psychology, Volume I, Anscombe, G.E.M. and von Wright, G.H. (eds), Anscombe, G.E.M. (trans.). Oxford: Basil Blackwell. Wittgenstein, L. (1980b). Remarks on the Philosophy of Psychology, Vol. II, Anscombe, G.E.M. and von Wright, G.H. (eds), Luckhardt, C.G. and Aue, M.A.E. (trans.). Oxford: Basil Blackwell. Wittgenstein, L. (1982). Last Writings on the Philosophy of Psychology, Vol. I, von Wright, G.H. and Nyman, H. (eds), Luckhardt, C.G. and Aue, M.A.E. (trans.). Chicago: University of Chicago Press. Wittgenstein, L. (2009). Philosophical Investigations. Anscombe, G.E.M., Hacker, P.M.S. and Schulte J. (trans.). Oxford: Blackwell. Wollheim, R. (1980). Art and Its Objects: With Six Supplementary Essays, second edition. Cambridge, New York: Cambridge University Press.


Aspect-perception, perception and animals Wittgenstein and beyond Hans-Johann Glock

The aim of my contribution is to shed light on the connection between perception, aspect-perception, recognition, evaluation, appreciation and judgement, in three respects. In ascending order of importance, these are: first, sketching the historical background of the idea of aspect-perception; second, making a case for a particular interpretation of Wittgenstein’s discussion of the topic; and, third, addressing some central philosophical questions raised by his discussion, notably the contrast between experiential and conceptual dimensions of perception and the issue of belief and perception in animals. As regards both Wittgenstein exegesis and substantive issues I reject two popular views: • •

the ubiquity thesis according to which all cases of perception involve aspect-perception; reductionism, i.e. the view that aspect-perception holds the key to a proper explanation of other forms of perception, including common-orgarden perception and pictorial depiction.

Concerning the ubiquity thesis, I shall defend my interpretation of Wittgenstein against the objections raised by Mulhall, partly by presenting new evidence, partly by placing the passages he invokes in their proper context. But I shall also meet his legitimate challenge, namely to explain Wittgenstein’s obsession with aspect-perception without imputing the ubiquity thesis to him. I shall dissent from both Mulhall and Schroeder as regards the optimal way of meeting that challenge. Concerning the philosophical merits of the ubiquity thesis, I shall distinguish various options, arguing for two claims: not all perception is aspect-perception; and not all aspect-perception involves interpretation. To this end I shall explore the important differences between the concepts of seeing, seeing-as, being visually aware of and recognizing-as. This will prepare the ground for a discussion of reductionism. Reductionism about perception in general is a stronger claim than the ubiquity thesis and can hence be ruled out along with the latter. But there are more restricted and subtle versions, notably Wollheim’s attempt to explain depiction by reference to aspect-perception.

78 Hans-Johann Glock Here I shall only consider very briefly the criticism of Wollheim that Hyman derives from Wittgenstein. The final part returns to the concept-ladenness of perception, though only in the context of the philosophy of animals’ minds. That animals can believe things would seem to be an obvious implication of the fact that they can perceive. But it has been denied by some proponents of the idea that genuine perception must be seeing-as and hence concept-laden, notably Heidegger, Sellars and McDowell. I shall take issue with this denial, in particular with the idea that a creature bereft of concepts is at most capable of a kind of quasi- or as-if perception.

Historical context The phenomenon of aspect-perception has been recognized since the eighteenth century, mainly for the case of picture puzzles (German: Kippfiguren, Vexierbilder; Jantschek 1995). From this time onwards, visual perception has taken central stage; and, for better or worse, neither Wittgenstein’s discussion nor this article constitute an exception. Aspect-perception became a subject of systematic research towards the end of the nineteenth century, following the establishment of psychology as a separate academic discipline. The main psychological contribution to the understanding of aspect-perception hails from Gestalt theories, and in particular from the seminal work of Wolfgang Köhler. His main contention was that the change in how we perceive aspect figures, and especially picture puzzles, cannot be explained atomistically, since the elements of what is seen before and after a Gestaltswitch remain constant. Aspect-perception was also an important theme within the phenomenological tradition, even though its members did not employ the terms ‘aspectperception’ or ‘aspect-seeing’. The ‘as-structure’ (als Struktur) of perception was thematized by Husserl (1973: §5). He was keen to distinguish between the mere complex of sense-data and the object of perception, a mental phenomenon, yet one unified by an intentional structure. Heidegger maintained in Sein und Zeit (1977: §32) that ab initio (immer schon) everything is seen as being something (of a certain kind). Accordingly, even the humdrum perception of everyday life is a process in which we interpret the world, albeit for the most part without being aware of this activity. By far the most important and influential treatment of aspect-perception is that of Wittgenstein, notwithstanding the fact that it is unsystematic and sometimes enigmatic. To this day it is common among commentators to complain that this part of Wittgenstein’s work has received relatively little attention. In view of a spate of recent publications, however, this complaint has become somewhat obsolete. The real problem is rather that the debate has remained unfocused. The root cause is that interpreters have turned to Wittgenstein’s discussion out of a wide variety of motives. Some of them – for example, linking Wittgenstein and Heidegger or treating his remarks as

Aspect-perception, perception and animals


a contribution to a therapeutic conception of philosophy – have little to do with Wittgenstein’s own interest in the topic (this malaise is particularly pronounced in many contributions to Day and Krebs 2010). Others, notably Wollheim’s ambition to put his ideas in the service of an account of pictorial representation, are more congenial. Nevertheless, they raise a host of complex issues that tend to lead away from Wittgenstein’s own focus. The latter, I shall argue, lies precisely where one would expect, namely in the philosophy of perception.

Wittgenstein’s preoccupation with aspects For Wittgenstein, ‘aspect-perception’ denotes a gamut of interrelated perceptual phenomena. The paradigmatic case is what he calls aspect-dawning or change of aspect (Aufleuchten eines Aspekts or Aspektwechsel): certain objects, especially schematic drawings – ‘picture objects’ (PI II 194; LW I §489) – can be seen under more than one aspect. An aspect ‘dawns’ on us when we notice such a hitherto unnoticed aspect of the object we are looking at, come to see it as something different. Thus we may pass from seeing a picture puzzle as a mere collection of lines to seeing it as a face; or from seeing Jastrow’s duck–rabbit (Figure 4.1) as the picture of a duck to seeing it as the picture of a rabbit. Picture puzzles like the Necker cube briefly appear in the early work (O’Sullivan 2015). Wittgenstein starts using the label ‘seeing-as’ in the 1930s; and from 1935 onwards his philosophy of psychology returns to the topic with increasing frequency (TLP 5.5423; NB 9.11.14; BrB II.16–22; PI II.xi; RPP I & II; LW I; LW II.i). Between 1947 and 1949 it even dominates his work, partly under the influence of Köhler’s Gestalt psychology.

Figure 4.1 Unknown artist, duck–rabbit.

80 Hans-Johann Glock Perceptive readers have remarked that Wittgenstein’s obsession with aspect-perception stands in need of explanation (Mulhall 2001; Schroeder 2010). Several candidates have or could be put forward. These are by no means exclusive; all of them may go some way towards making his intense interest intelligible; and all of them invite and repay closer scrutiny for both exegetical and substantive reasons.

The indeterminacy of the mental A first potential explanation is that Wittgenstein regarded aspect-perception as essential to third-person mental statements. He is famous (or infamous) for holding that the relationship between mental phenomena and their behavioural manifestations is not a causal one to be discovered empirically, through theory and induction, but a criterial one: it is part of the concepts of particular mental phenomena that they have a characteristic manifestation in behaviour (LPE 286; LSD 10). And it is part of mental concepts in general that they have some such manifestation. We would have no use for these expressions if they were not bound up with behavioural criteria. If we came across human beings who used a word which lacked any connection with pain-behaviour and the circumstances in which we display it, we could not translate it as ‘pain’. Putnam’s claim (1975) that there might be ‘supersuper-spartans’ who are in constant agony without ever showing it is as incoherent as describing as soulless human beings who behave exactly like us (LPP 281). However, while sensation-terms like pain are ascribed to others on the basis of straightforward behavioural criteria, intentional verbs (‘believes’, ‘desires’, ‘intends’, etc.) and expressions for moods cannot be applied simply on the basis of an individual’s momentary behaviour. The occasions for their use form a highly complex syndrome. For example, what counts as a manifestation of sadness on one occasion may not on another. And what someone is disposed to do as a result of holding a particular belief depends not just on the context but also on his other beliefs, desires, intentions, etc. (PI II 174, 229; LW I §§862–9, 942, 966; LW II 42–3, 55–6, 84; Z §567–9; RPP I §§129, 314; Z §492; Geach 1957: 8; contrast Ryle 1980: 128–30). In his last writings, more particularly in material adjacent to discussions of aspect-perception, Wittgenstein qualified the idea of criterial support even further. There can be no proof of third-person ascriptions of emotions, and we may often be unable to decide whether someone is, for example, annoyed. This ‘indeterminacy’ and ‘unpredictability’ is constitutive of some of our concepts of the inner. Moreover, often those who are closely acquainted with a person can make even the most subtle emotional ascriptions with certainty, without being able to specify conclusive criteria, since their evidence is ‘imponderable’, i.e. it consists of a syndrome of behaviour, context and previous events (PI II 227–8; LW II 70, 87, 90–5). And it is here that aspectperception comes in.

Aspect-perception, perception and animals


We describe a person as placid or anguished on account of certain features of her face. But these features are not always related to the mental state in an obvious way, one easily accessible to normal human observers. Sometimes one needs to recognize the anguish or placidity in the face of another – see it as manifesting anguish or placidity. Recognizing highly complex mental states may also involve aspect-perception of a different kind. It may require an acquaintance with the subject which is based on recognizing ‘patterns’ in her past behaviour (Glock 1996: 97, 129). This explanation has a fundamentum in res. On the other hand, while aspect-perception plays a prominent role in some of Wittgenstein’s discussions of third-person mental statements, these are the ones pertaining to the ‘indeterminacy of the mental’. The latter is an important case, as Wittgenstein gradually came to realize. Yet it is nonetheless remains the exception rather than the rule. More seriously still, while pertinent remarks about recognizing faces as manifesting a mental state and of detecting ‘patterns in the weave of life’ hail from roughly the same period as the extensive discussions of aspect-perception, the latter do not refer to the former topic.

Secondary meaning A second explanation invokes another topic that began to grip Wittgenstein towards the end of his life, namely ‘experiences of meaning’. He claims that words have a ‘familiar physiognomy’: they are associated with other words, situations and experiences, and can assimilate these connections. Thus one may feel that names ‘fit’ their bearers, and words turn into mere sounds, if these connections are severed – for example, when they are mechanically repeated several times (PI II 214–5, 218; LPE 283). This diagnosis stands on even firmer ground. Wittgenstein detects a close connection between two phenomena: ‘meaning blindness’ and ‘aspect blindness’. He illustrates aspect-seeing through aspect-blindness: the inability to experience aspect-dawning (PI II 213–4; RPP I 877; RPP II 42, 479 490; LW I 174, 492–3). An aspect-blind person could apply a new description to a picture object – for example, use the schematic drawing of a cube as picture of a three-dimensional object. But he would not experience this as seeing something differently, and would not recognize the incompatibility with treating it as a two-dimensional complex of three parallelograms. His defect is not one of sight but of imagination. Now, a special kind of aspect blindness is meaning-blindness, the inability to experience the meaning of a word (PI II 175–6, 210; RPP I §§189, 202–3, 243; see also James’s ‘soul blindness’, 1918: chapter 2). Some passages declare outright that the importance of aspect-perception lies in its connection with experiences of meaning (PI II 214; LW I §784); but these declarations do not settle the question of how to explain his obsession, for other passages insist that aspect-perception is not essential to the concept of meaning but only to what he calls ‘secondary sense’ (RPP I §358; RPP II §§242–6). Experiencing

82 Hans-Johann Glock meaning underlies the ‘secondary sense’ of terms: some people are inclined to say things like ‘“e” is yellow’, ‘“u” is darker than “i”’ or even ‘Tuesdays are lean, Wednesdays are fat’. This secondary sense (PI II 216; LW §§797–8) – – –

differs from the primary sense: obviously ‘e’ is not yellow in the sense flowers are – it cannot be compared with a sample of yellow; presupposes the primary sense: it can only be explained by reference to the primary one, but not vice versa; is not a matter of ambiguity or of metaphor: while we can disambiguate ‘bank’ by introducing a new term and can paraphrase metaphors, we cannot express secondary senses in any other way.

Secondary sense also explains sylleptic ambiguity: the fact that we speak of deep sorrows and wells, of plaintive melodies, etc. The meaning-blind person uses and explains words correctly but has no ‘feel’ for their physiognomy, a lack which is comparable to the lack of a musical ear. For this reason, he is barred from important forms of aesthetic discourse, or from understanding puns. Nevertheless, experiences of meaning are not essential to linguistic understanding. The latter requires nothing more than the ability to apply and explain an expression. Wittgenstein does not intend to reinstate the idea that understanding a word is a mental episode. And while experiences of secondary meaning raise intriguing philosophical puzzles, they hardly constitute the core of linguistic meaning and understanding.

Köhler, Wittgenstein and objects of perception The role of aspect-perception in Wittgenstein’s later thought derives from its being a kind of perception rather than from ‘secondary’ connections to the philosophy of mind and the philosophy of language. An obvious suggestion is that Wittgenstein felt an urgent need to criticize Köhler’s theory of this kind of perception. In so far as his discussion mentions or alludes to any authors, these are Gestalt psychologists; and, while he draws extensively on their examples, most of his remarks about their claims about aspect-perception are negative. It is by no means easy, however, to pinpoint what his critique targets or ought to target. According to Köhler, what we perceive immediately is not a mosaic of discrete and unorganized stimuli (dots and coloured surfaces, sounds), as empiricism and behaviourism have it, but Gestalten, circumscribed and organized units, such as material objects or groups of objects (1975: chapter 5). We do not see three dots but see them form a triangle; we do not hear a chaotic array of sounds but detect a melody. This is congenial to Wittgenstein (not to mention Austin) in rejecting the reductionist view according to which we construct perceptual objects out of raw sense data. But it may appear from a Wittgensteinian perspective that Köhler’s treatment of aspect-dawning both reifies and privatizes Gestalten. He claims that in aspect-perception we

Aspect-perception, perception and animals


do not see one and the same object under different aspects but two different ‘visual objects’ or ‘visual realities’ (1975: 82, 107, 148–53). In his attempt to do justice to the idea that we see the picture object differently, Köhler runs the risk of turning an aspect (Gestalt) into a private mental entity. At the same time, one is entitled to query whether Wittgenstein’s consistently avoids this danger. The discussion of aspect-perception in PPF (PPF xi p. 203ff.) kicks off with an influential distinction Two uses of the word ‘see’. The one: ‘What do you see there?’–‘I see this’ (and then a description, a drawing, a copy). The other: ‘I see a likeness between these two faces’—let the man I tell this to be seeing the faces as clearly as I do myself. The importance of this is the difference of category between the two ‘objects’ of sight. Wittgenstein here seems to reckon both with a physical object to be seen and an object which is accessible only to those noticing a specific aspect. And wouldn’t the latter have to be merely mental, like one of Köhler’s visual realities? Admittedly, Wittgenstein tries to guard against the threat of reification by placing the ‘objects’ of sight in scare-quotes. It is in this spirit that his distinction was later taken up by Anscombe (1965) and then elaborated by Scruton for the case of pictures (1981). The latter distinguishes • •

the material object of sight, i.e. the object to be seen (in his case, a painting which consists of an arrangement of colour-patches on a canvas); the intentional object of sight, i.e. the object as defined by the subject’s experience (for example, experiencing a depicted object as a god rather than a warrior; it is the object as given ‘under a description’, to use Anscombe’s terminology); the represented object, determined by the painter’s intentions.1

In the present context, we can leave aside the represented object. Picture puzzles or picture objects are standardly created with an intention, notably to allow seeing them under several aspects; but this plays no role in Wittgenstein’s discussion of aspect-perception. His two-fold distinction seems to run between material object and an ‘intentional’ object determined by the subject’s experience, in particular by how the latter would describe what it sees. While the former remains unaltered during an aspect-change, the latter changes: first I see (it as) a rabbit, then I see (it as) a duck. As the scare-quotes indicate, Wittgenstein’s two ‘objects of sight’ are not bona fide objects that differ from material objects only in being mental or abstract but are constituted by (or shorthand for) how the material object strikes the subject. As regards Köhler, his general perspective may be congenial to treating the two visual objects as mental entities that display the privacy of ownership criticized by Wittgenstein. This means that if a perceiver A has, i.e. experiences, such a visual object then it is impossible that any subject other than A should

84 Hans-Johann Glock have it (Glock 1996: 304–9). Nevertheless, in Köhler’s treatment of aspectperception in particular, there is no suggestion that he regards these objects as epistemically private, i.e. such that only A herself can know that A experiences the visual object. In particular, there is no suggestion that Köhler precludes the possibility of A communicating to others what she experiences. In that case, however, why not extend the courtesy of a deflationary reading from Wittgenstein’s objects of sight to Köhler’s objects of vision? Wittgenstein seems to have thought that the latter involve a reification which is more than terminological but integral to his account of aspect-perception. The two ‘visual objects’ are said to differ in their organization, which is as much a feature of them as their colour and shape. Accordingly what changes in aspect-perception is not the colour or shape of the elements of the visual impression but the way in which they are arranged or organized. Wittgenstein rejects this explanation. It suggests that what changes in cases of aspect-dawning is the way we perceive the (spatial) relationships between the elements of the picture. But he regards this implication as mistaken. When asked to depict faithfully what we see, i.e. the picture object, before and after an aspect-change, there is no more difference in the organization of the elements than in their shape or colour. A subject may go about producing such a depiction in a different manner – for example, a different sequence. But the end product remains the same. The characteristic of aspect-dawning is precisely that no specific feature of the visual field changes, whether it be the components or their arrangement. This complaint has to be handled with care, however. The spatial organization of the material object remains the same, of course; but the spatial organization of the intentional object need not. A subject seeing in Figure 4.2 the face of an old witch looking right to left will describe the spatial relations (lying in front, lying in the back, pointing away from the observer or not) differently from one seeing in it a young lady looking backward over her right shoulder. What does remain the same, then? It is the spatial relations between the parts of the material object as described in purely geometric terms, as when one characterizes a two-dimensional visual field in terms of an arrangement of patterns and shapes. It is in those austere terms that the alleged change in organization of what is seen cannot be specified. And that means that as regards that purely geometric level, Köhler’s difference of ‘organization’ is either spurious or could only refer to ineffable features of a private object, which the private-language argument excludes as chimerical (PI II 196–7; RPP I §§ 536, 1113–25; LW I §§444–5, 510–12). Whether Köhler should be interpreted as treating his visual objects as geometric arrangements in the way ruled out by this line of reasoning is an interesting question, but it will not be broached here. As regards Wittgenstein, my contention is this: whether justified or not, his critical interest in Gestalt psychology only goes some way towards explaining his fascination with the subject. Leaving speculations about Wittgenstein’s personal motives aside, an objective reason for thinking so is that his treatment of the topic has a

Aspect-perception, perception and animals


Figure 4.2 Unknown artist, young lady–old witch.

much wider scope than pinpointing weaknesses in Köhler’s claims would require or suggest. Let us therefore consider one final rationale.

The paradox of aspect-perception Wittgenstein’s immediate aim was to dissolve the paradoxical appearance of aspect-dawning: when looking at a picture object we can come to see it differently, although we also see that the object itself remains unchanged. And this can lead to a characterization that is downright paradoxical: it seems to be the case both that what I see has changed and that what I see has not changed (PI II 193–5; LW I §493; see also Ahmed forthcoming). One way of dealing with aspect-dawning is to point out that perceiving alternative aspects is caused by different patterns of eyeball movements. Wittgenstein was aware of such correlations but denied that they resolve the paradox (PI II 193, 203, 212; LW I 795; TLP 5.5423), for even if they explain how and why the phenomenon occurs, they do not provide a description of it which avoids the air of paradox.

86 Hans-Johann Glock Wittgenstein attached enormous importance to aspect-perception, since he thought that in this phenomenon ‘problems about the concept of seeing come to a head’ (LW I §172). Presumably this is because the paradox exemplifies in a precise form the tension between two apparent aspects of perception: that it is, on the one hand, a basic cognitive faculty and hence non- or pre-conceptual and that, on the other, in at least some of its variants it is concept-saturated or concept-laden.2 At a non-conceptual level we see one and the same thing – for example, a person’s face; yet at a conceptual level we may see it differently – for example, as placid or anguished. Schroeder (2010: 357) has denied that being exercised by the paradox explains Wittgenstein’s interest in aspect-perception. His rationale is that the paradox is easily resolved by the aforementioned distinction between two objects of perception Wittgenstein draws at the very beginning of his discussion: while the material object remains the same, the intentional object changes. While he is right to insist that the distinction is pivotal to Wittgenstein’s attempt to defang the paradox, the distinction by itself does not suffice to do the job. We can appreciate this by taking a closer look at the paradox. Obviously, there is nothing paradoxical about an object x having distinct features F1 – Fn at time t. But some features are mutually exclusive in that an x cannot possess them both at the same time. For instance, even if an organism could be transformed from a duck into a rabbit, it cannot be both at the same time but at most a mongrel of the two (for example, with the body of a duck and the head of a rabbit). If an object x – for example, the picture object – were to have incompatible features F and G, this would have to be due to a change – for example, from x being F at t1 to being G at t2. But the picture object which occasions a change of aspects does not change. What is more, the perceiver is cognisant of this fact. And, as explained above, the content of his visual field also remains the same. What then does change? The object of perception in another sense – namely the intentional object. This ‘object’, however, is a projection from how the material object strikes me. We talk of an intentional ‘object’ of A’s perceiving not on account of having identified a genuine object either within or beyond space and time. Rather, we do so on account of the fact that the question ‘what does A perceive?’ is to be answered by a noun-phrase: a noun or noun-clause. If A possesses language, that nounphrase is determined by the answer that A could proffer in response to the question; if A lacks language, it is determined, though less precisely, by A’s potential behaviour with respect to things that lie within its field of sight (Glock 2013). Accordingly, the resolution of the paradox does not take the straightforward form of noting a change from property F to G in one object x but not another y. As argued above, one of these objects is simply a construction from how things strike a subject. As a result, resolving the outright paradox through the distinction between two objects commits one to tackling a puzzle about the relation between conceptual and non-conceptual dimensions of aspect-perception. More specifically,

Aspect-perception, perception and animals 87 the question is why the way the material object strikes me should be a matter of (non-conceptual) seeing rather than (conceptual) interpretation or thought. This puzzling feature of aspect-dawning is brought into sharper relief by noting a contrast to cases of optical illusion. Someone cognisant of being subject to such an illusion would say things of the form: 1



x looks (like an) F but really x is not F (x is G).The qualification ‘but really x is not F’ is not an option for a subject who can see both aspects of a picture puzzle. At a time at which she sees the duck rather than the rabbit in the picture she cannot say: The picture puzzle looks like a duck but really it is a rabbit, for she knows that the picture puzzle is neither a duck nor a rabbit. But neither can she say: The picture puzzle looks like a picture of a duck but really it is a picture of a rabbit, for she knows that it is both the picture of a duck and the picture of a rabbit.

Aspect-dawning is not simply a matter of intellectually recognizing that visual appearances are delusive. The perceiver herself experiences the aspectchange as a visual difference – she sees the object differently rather than thinking about it in different terms. But how is that possible? Therein lies the key to Wittgenstein’s preoccupation: to place aspect-perception between the poles of immediate perception and intellectual thought.

Wittgenstein’s solution to the puzzle His first point is that types of aspect-perception differ according to the degree of thinking involved (PI II 207–12; LW I §§179, 530, 582, 588, 699–704; RPP I §§1, 70–4, 970; RPP II §496, 509; see also Ahmed forthcoming). At one end lie ‘conceptual’ aspects like the duck–rabbit, which cannot be expressed solely by pointing to parts of the picture-object but require possession of the relevant concepts. At the other lie ‘purely optical’ cases, such as the double-cross, in which we can express our seeing of the aspect by following certain lines of the picture-object, without using concepts (yet even here concepts like background and foreground may be involved). His second point is that the concept of seeing an aspect lies between that of seeing, which is a passive state or process that the subject is in, and that of interpreting, which is an action, something the agent actively does. It is closer to the latter in the following respects (PI II 212; RPP I §§27, 169, 1030; RPP II §§544–5; LW I §§451, 488, 612): – –

the ‘optical’ or ‘visual’ picture remains the same, as we have seen; aspect-seeing, unlike most cases of seeing, is subject to the will: although we may not always succeed in noticing an aspect or keeping it in focus, it always makes sense to try to do so, and we often succeed;

88 Hans-Johann Glock –

in noticing an aspect of the conceptual kind we do not just focus on properties of the object perceived but realize certain ‘internal relations’ between it and other objects, relations of similarity and dissimilarity such as that between two human faces.

Aspect-perception is closer to seeing in the following respects (PI 203–4; 212; RPP I §§8, 1025; RPP II §§388, 547): –

– –

if a subject detects both aspects – for example, in the duck–rabbit picture – there is no possibility of her being mistaken about which of the two aspects she currently sees;3 aspect-seeing is a state; in particular it has genuine duration, i.e. it has a beginning and an end which can be clocked, can be interrupted, etc.; there is no more direct expression of the experience than the report of aspect-perception ‘I see it as a rabbit’, i.e. there is no sharp contrast between the ‘interpretation’ and the uninterpreted data.

It may seem that Wittgenstein creates an artificially stark contrast between seeing and thinking by restricting the latter to interpreting, conjecturing what a picture represents (RPP II §390; RPP I §§8–9, 20; PI 193, 197, 212; LW II 14). But the paradox of aspect-dawning does not depend on this. I may know about the duck–rabbit picture and think ‘One can see a rabbit here’, without being able to see it. Schroeder is right on one important point. Wittgenstein suggests that the paradox trades on an equivocation: what I see in the ordinary sense has not changed, while what I see in the sense of ‘seeing’ closer to thinking has. But given the extent to which he has laboured the paradox even after introducing the distinction between material and intentional object, this one solution is more like a whimper than a bang. Nevertheless, it features important insights. Reports of aspect-dawning are not descriptions – either direct or indirect (interpreting) – of an inner experience which accompanies ordinary perception but avowals, spontaneous reactions to what we see (PI 197, 212; RPP I 13, 20). Moreover, what changes in aspect-dawning is ultimately neither what we perceive nor its ‘organization’ but our attitude to and abilities concerning the picture object, how we react and what we can do with it. Suddenly we copy or explain the picture puzzle differently, change the way we play a piece of music or recite a poem (PI II 197–8, 208; RPP I §982; LC I). One important thing we do in noticing an aspect is placing what we perceive in another context; we detect new connections or draw fresh comparisons. This is why changing the context of an object may alter our perception of it (PI II 212; RPP I 1030; LW I §516). Let us finally note that Wittgenstein’s resolution of the paradox, such as it is, is connected to his characterization of aspect-blindness. His diagnosis of the root cause of the latter – lack of a certain kind of (visual) imagination – should call to mind Kant’s idea that imagination is a key factor in bringing

Aspect-perception, perception and animals 89 together the intuitions of the experiential faculty of sensibility and the concepts of the intellectual faculty understanding. Many details of Kant’s transcendental psychology are alien to Wittgenstein. But even a stern critic of the latter like P. F. Strawson has applauded and developed his idea that imagination plays a key role in the combination of purely experiential and conceptual factors characteristic of perception more generally (1974).

The ubiquity thesis from an exegetical perspective At this juncture it is tempting to take my explanation of Wittgenstein’s preoccupation one step further: for Wittgenstein aspect-perception was integral to all perception. This seems supported by his distinction between aspect-dawning and ‘continuous seeing of an aspect’. Yet Wittgenstein denied that seeing-as is typical of all experience. Seeing-as requires a contrast between two different ways of perceiving an object, but under normal circumstances it makes no sense to say that one sees the cutlery as a knife and fork (PI II 194–5). Accordingly, Wittgenstein confines continuous aspect-perception to objects like pictures. These have been produced with the intention that they should be perceived as depicting something, with the intention that a subject should see something in them, to use Wollheim’s vocabulary. And picture puzzles have been produced with the more specific intention that it should be possible to see more than one intentional object in them. In these cases no special circumstances are needed for a contrast between relating to what is perceived either as the depiction of something else or as an object in its own right. The key to my exegetical claim is PI II 195b-c: It would have made as little sense for me to say [of the picture-rabbit] ‘Now I am seeing it as . . .’ as to say at the sight of a knife and fork ‘Now I am seeing this as a knife and fork’. This expression would not be understood.—Any more than: ‘Now it’s a fork’ or ‘It can be a fork too’. One doesn’t take what one knows as the cutlery at a meal for cutlery; any more than one tries to move one’s mouth as one eats, or aims at moving it. Mulhall has taken issue with my reading. According to him, the passage concerns ‘Now I’m seeing it as . . .’. Wittgenstein associates this phrase with aspect-dawning rather than the continuous seeing of an aspect. So he is not denying that ordinary perception involves continuous aspect-perception. This objection fails. First, it does not accommodate 195c, which concerns a phenomenon that is not necessarily episodic, namely taking what one knows (erkennt) as cutlery for cutlery. Second, it is difficult to see how Wittgenstein’s qualms about a ubiquitous employment of seeing-as could

90 Hans-Johann Glock apply to aspect-dawning without also applying to normal perception, for, with respect to the latter, it is even less plausible to regard it as a case of constantly recognizing something as than with respect to the former. Third, there is further specific evidence against the ubiquity thesis. Do I always see a thing as something, although only puzzle pictures bring this out? . . . Suppose I show it to a child. The child says ‘It’s a duck’ and then suddenly, ‘Oh it’s a rabbit’. So he recognizes it as a rabbit. This is an experience of recognition. So if you see me in the street and say, ‘Ah, Wittgenstein’. But you haven’t an experience of recognition all the time. The experience only comes at the moment of change from duck to rabbit and back. In between, the aspect is as it were dispositional. Geach: Couldn’t I say at any time how I see it—not just when it changes? Wittgenstein: Only if you are concentrating on it. (LPP 103–4) Fourth, the ubiquity is in tension with Wittgenstein’s strategy for tackling the paradox. The concept of seeing an aspect lies between that of seeing and that of interpreting; but the latter is an intermittent intellectual activity rather than a permanent experiential state (PI II 207–12; RPP I §§1, 27, 70–4, 169, 970, 1030; RPP II §§5496, 509, 44–5; LW I §§451, 488, 612).

The ubiquity thesis from a substantive perspective The ubiquity thesis is that all perception involves aspect-perception or something like it. It unites otherwise diverse philosophers like Heidegger, Vesey, Hanson, Strawson and Mulhall. In order to assess the ubiquity thesis, it is helpful to distinguish four possible stances on the relation between perception, aspect-perception and interpretation. 1 2 3 4

All aspect-perception involves interpretation and not all perception is aspect-perception (Schroeder; early Wollheim). Not all aspect-perception involves interpretation and not all perception is aspect-perception (Wittgenstein as interpreted here). Not all aspect-perception involves interpretation and all perception is aspect-perception (later Wollheim, Mulhall).4 All aspect-perception involves interpretation and all perception is aspect-perception, hence all perception involves interpretation (Heidegger; Vesey 1956; Hanson 1958; Strawson; McDowell 1996).5

Only (3) and (4) imply the ubiquity thesis. But the other parameter, the involvement or absence of interpretation, is important for assessing the different rationales for either accepting or rejecting the ubiquity thesis. It is also crucial to assessing the possibility of animal perception.

Aspect-perception, perception and animals


Wittgenstein’s claim that outside of cases of a sudden recognition ‘the aspect is as it were dispositional’ (LPP 104) suggests the following critical diagnosis of the ubiquity thesis: it starts out from the continuous possibility of aspect-dawning and infers from this the continuous presence of aspectperception. In that case, however, it would be based on the mother of all philosophical fallacies, namely from a potentiality to an actuality. The fact that when A looks at x it is always possible that A should come to perceive a hitherto unnoticed aspect of x does not entail that when no aspect dawns on A she is all the while perceiving x under a familiar aspect. The ubiquity thesis would need a different kind of backing, therefore. One option is suggested by Geach’s intervention. Even without any aspectchange, at any time a subject can say how she sees the duck–rabbit. According to Wittgenstein’s response, A can say how she sees x only if she concentrates on x. In my view this restriction is mistaken. I can tell you that I see the distant object as a castle even if all the while I have focused on our conversation. At this juncture, the difference between being able to say what one sees x as and simply seeing x as something becomes important. Even if A says what she sees x as not in response to a query but spontaneously, A is taking a stand on the question of what feature x possesses. But her being able to take such a stand by saying that x is F is perfectly compatible with her not having seen x as F before then. She may only come to see x as F in considering the question of whether x is F. Accordingly, although Geach is right to maintain that a subject in possession of language can always say what she sees x as, this does not suffice to establish the ubiquity thesis. His observation does not preclude the possibility that seeing-as is confined to the two cases for which Wittgenstein rightly grants it: sudden recognition or aspect-dawning on the one hand, looking at picture puzzles, etc. on the other. Another way of precluding that possibility and thereby establishing the ubiquity thesis starts out from the claim that all seeing involves interpretation anyway, independently of whether all seeing involves aspect-seeing. If that is correct, this line of reasoning continues, all seeing also involves subsuming x under a concept F, as opponents of the idea of ‘nonconceptual content’ like McDowell (1996) have been urging. And that in turn draws in its wake seeing x as instantiating F. A limitation of this defence of the ubiquity thesis is this: neither seeing x nor even seeing x as F implies seeing x as being of a certain kind or type and hence subsuming x under a sortal concept. For instance, A does not have to see (i.e. recognize) the stone in front of A as an Acheulean biface in order to see it. Nor does A have to see the stone as being of a more general type. In seeing any object x A perceives a visibilium, standardly a material object. But why should it follow that A must see x as a material object? The question is all the more acute because it is unclear what the latter involves, whether it presupposes mastery of the concept of a material body, and what such mastery requires (Strawson 1992: chapter 2). Even

92 Hans-Johann Glock such qualms aside, neither ‘material object’ nor ‘visibilium’ are sortal or kind terms, since there are no principles for individuating and counting things that fall under them. The rejection of ‘nonconceptual content’ may contain a kernel of truth. For a subject A that possesses concepts, and for every object x which A can identify perceptually, it is arguable that the following holds: A can subsume x under some concept, though perhaps only under demonstrative, non-sortal concepts like ‘over there’, ‘in the upper left-hand corner of my visual field’, etc. The moral for the ubiquity thesis remains predominantly negative nonetheless. A subject A does not need to recognize x as being of a certain kind in order to see x. All that is needed is that A be visually aware of x. And that only requires noticing x in A’s visual field. Finally, those versions of the ubiquity thesis that also regard all aspectperception as conceptual, i.e. positions of type (4), face an additional hurdle. Even aspect-perception does not per se amount to recognition, i.e. noticing that an object falls under a concept. Indeed, seeing x as F does not entail recognizing x as F. Adding, ‘as’ cancels the factive (and thereby the success, or achievement) implication of ‘sees’, whereas ‘recognizes as’ has such an implication. This marks yet another sharp contrast between seeing in the standard, factive or achievement sense on the one hand and seeing-as on the other. It is prima facie attractive to cling to a different conceptualist principle: if A sees x as F, then A believes that x is F. But even this principle does not hold generally if ‘A sees x as F’ means ‘x visually appears as F to A’. I may perceive the object in front of me as a red tomato, in the sense that it visually appears to me to be a red tomato, yet know it to be a realistic imitation.

A brief note on reductionism It turns out that the ubiquity thesis is untenable. Now, reductionism is a stronger claim than the ubiquity thesis. It requires not just that all perception is aspect-perception but two further points: first, that aspect-perception is a simpler or more tractable phenomenon than perception; second, that it be capable of explaining the latter but not the other way around. Given the failure of the ubiquity thesis, reductionism concerning perception in general can be ruled out. Perhaps, however, the prospects of a restricted version are less bleak. It seems prima facie attractive to explain the much more specific phenomenon of depiction by reference to seeing-as. That was Wollheim’s position in his seminal Art and Its Objects. The ‘representational seeing’ involved in seeing what a picture represents is a matter of seeing x – the ‘medium of representation’ in Wollheim’s terms, the material object in our idiom, i.e. the surface of the picture – as y – the represented object or intentional object in our scheme. Wollheim later retracted this analysis. But it is worth mentioning a criticism of Wollheim’s original account that Hyman (2006: 255n8) derives

Aspect-perception, perception and animals 93 from Wittgenstein. Concerning aspect-change in the case of picture objects like the duck–rabbit, Wittgenstein raises the question: Wie wäre diese Erklärung: ‘Ich kann etwas als das sehen, wovon es ein Bild sein kann’? Das heisst doch: Die Aspekte im Aspektwechsel sind die, die die Figur unter Umständen standing in einem Bild haben könnte. (PI II xi; §166) Hyman starts out from the Anscombe translation of the first sentence: ‘How would the following account do: What I can see something as, is what it is a picture of?’ He goes on to quote a sentence a few lines further down: ‘Could I say what a picture must be like to produce this effect? No.’ And he seems to regard this as a rejection of reductionism about depiction. But the first sentence was mistranslated by Anscombe. The passage should be rendered instead the way it is in the Hacker/Schulte translation (with one minor modification): How would the following account do: ‘I can see something as whatever it can be a picture of?’ This means after all: the aspects in a change of aspects are those which, in certain circumstances, the figure could have permanently in a picture. Wittgenstein does not even consider reductionism à la Wollheim, the attempt to explain depiction through aspect-perception. Instead, he considers and rejects an explanation running in the opposite direction, namely of aspectperception or change by reference to depiction. This once again counts against ascribing the ubiquity thesis to Wittgenstein. He does not even consider the possibility of employing aspect-perception as a fundamental phenomenon that can be used as the basis for a reduction. Instead, he wonders how aspect-perception itself might be explained. Although Hyman’s interpretation of Wittgenstein is based on a faulty translation, his argument against reductionism inspired by it is promising: 1 2

aspect-perception of a constant kind concerns ambiguous depictions, pictures that can be seen either as (depicting) x or as (depicting) y; but then there is no more chance of explaining depiction in terms of seeing-as than there is of explaining linguistic meaning in terms of ambiguity.

Some elaboration is required concerning the analagon. For in some accounts the question of what linguistic meaning is should be resolved or replaced by questions relating, among other things, to questions of identity and difference of meaning. But although a satisfactory explanation of meaning must include an account of synonymy and ambiguity, Hyman is right to point out that the latter cannot function as a reduction base for the former. It remains

94 Hans-Johann Glock a truism that one needs to explain what it is for an expression to have a meaning before one can set out to explain what it is for an expression to have more than one meaning.6 Mutatis mutandis, one needs to explain what it is for a picture to depict an object before one can set out to explain what it is for a picture to depict more than one object. Note finally that premise (1) of this objection to reductionism about depiction presupposes that the ubiquity thesis is wrong. Rightly so, given the flaws of that thesis exposed in ‘The ubiquity thesis from an exegetical perspective’ above.

Aspect-perception and animal perception After this brief glance at the philosophy of representational art I turn to another subject in which the idea of seeing-as is vitally important, namely the philosophy of animal minds. Whereas assimilationists regard the differences between the mental capacities of humans and animals as gradual or quantitative, differentialists detect a sharp categorial chasm. One central bone of contention between these warring camps is whether animals can be in intentional states like believing, knowing and desiring. The latter are standardly characterized as propositional attitudes, relations between a subject and a propositional content. I reject this representationalist orthodoxy in Glock 2013. In order to believe or know that p, animals do not have to stand in a relation to a bona fide entity, whether mental or abstract, conceptual or non-conceptual. Even if we set aside this misguided way of conceiving the issue, however, we can still raise the question of what type of intentional state, if any, animals are capable of. The most important differentialist answer is what I have called the ‘lingualist master-argument’: no intentional states without concepts (concept thesis); no concepts without language (language thesis); ergo, no intentional states without language (lingualist conclusion); animals lack language (dumbness thesis); ergo, no intentional states in animals (differentialist conclusion). Aspect-perception is relevant to this argument, since perception presents a formidable stumbling block to the lingualist argument, casting doubt on the conjunction of concept and language theses. How could animals be capable of perception without also being capable of believing and knowing things?

The univocality of perception A radical differentialist reaction to this challenge rejects its presupposition. According to several representatives of the ‘Pittsburgh School’ inaugurated by Sellars (1957), animals simply cannot perceive the world in a sense of ‘perception’ that applies univocally to humans as well (for example, Brandom 1994 part 1; McDowell 1996: 108–26, 181–7; Rödl 2007). According to

Aspect-perception, perception and animals 95 this position, genuine perception is conceptual and propositional. Therefore it is beyond the capacities of non-conceptual animals. This stance is liable to induce exasperation about ‘unscientific’ anthropocentrism among ethologists and paroxysms of righteous indignation among pet-owners. More importantly in our context, it is also vulnerable to philosophical objections. As mentioned above, we may want to concede that every object we humans can identify perceptually can somehow be described conceptually by us, if only with the additional aid of demonstratives. It does not follow, however, that there is a list of propositions which captures precisely and completely what A perceives – the content of A’s visual field. And even if a propositional description of A’s visual field were available, it would transcend what even a subject endowed with language is capable of verbalizing. Nor does it follow that creatures without concepts cannot perceive what creatures with concepts can perceive. It only follows that they cannot understand or characterize what they see in conceptual terms. In terms of Wittgenstein’s characterization of aspect-perception, it is a privation of thinking rather than of perceiving. Furthermore, even if there were a sharp difference in type between propositional and non-propositional perception, neither of the following two claims would follow: – –

animals only perceive in an attenuated, ‘second class’ (Brandom 1994: 150) sense rather than literally speaking; ‘perception’ is equivocal.

There is a case for maintaining that perception means, i.e. amounts to, something different in conceptual and non-conceptual subjects. It has different preconditions and implications. But that does not imply that ‘perception’ linguistically means something different in the two cases. In both it means roughly something like the following: the capacity to gather information about the proximal and distal environment with the aid of sense organs – organs dedicated to this purpose. Irrespective of whether this explanation is correct, however, it is difficult to see on what grounds one could deny the univocality of ‘perception’. One possible rationale is an extreme conceptual holism, according to which any significant difference in the context of a phenomenon per se constitutes a distinct phenomenon or concept. But such holism reduces to absurdity, at least in the case of animal perception. That, at any rate, was suggested by Plutarch in his criticism of the radical differentialism of the stoics: As for those [stoics] who foolishly say of animals that they do not feel pleasure, nor anger, nor fear, nor do they make preparations, nor remember, but the bee only ‘as-if’ remembers, and the swallow ‘as-if’ makes preparations, and the lion is ‘as-if’ angry, and the deer ‘as-if’ afraid:—I do not know how they will treat someone who says that they do not see

96 Hans-Johann Glock nor hear either, but ‘as-if’ see and ‘as-if’ hear, and do not give voice, but ‘as-if’ give voice, and, in general, do not live, but ‘as-if’ live. For these last statements, I believe, are no more contrary to plain evidence than their own. (Sollertia 961 E–F). However, why should holistic as-ifness about perception or emotion imply as-ifness about forms of behaviour like acoustic signalling or about life? Because, by the holists’ own lights, it is partly constitutive of the concept of perception that the information perception provides can guide activities like signalling, running, swimming, etc. Similarly, it is partly constitutive of that concept that such information can be put in the service of biological functions, and hence, roughly speaking, of life. By the same token, it is partly constitutive of those behavioural and biological concepts that the activities and functions they express are capable of being guided by perception. But now, if what is partly constitutive – of running, for example – differs between humans and animals, then what is constituted must differ as well. The moral is manifest: all capacities interact, in the behavioural repertoire of a subject that possesses them, not just in causal but also in conceptual ways. Therefore, if lack of rational powers barred animals from anything other than as-if perception, it would also bar them from anything other than as-if movement, as-if digestion, as-if life, etc. Anyone willing to bite that bullet incurs the consequence of lead-poisoning.

Appreciating as Most differentialists grant the obvious, namely that animals perceive in the same sense of ‘perception’ as we do. This leaves open the question of whether there are important types of perception that are beyond them. Here aspectperception once more plays an important role, for mainstream differentialists often deny that they can perceive something as something. Animals, the story goes, can see, hear, feel, smell, taste a thing (object, event, creature) x of kind F; yet they cannot perceive x as (being of kind) F. This segregation is hard to reconcile with the discriminatory capacities of higher animals: how could animals discriminate between Fs and Gs, Fs and non-Fs, if they could not perceive an x as F? How, in particular, could they discriminate in the deliberate and controlled manner displayed in feats such as the nut-cracking of chimpanzees. The latter are capable of selecting stones that are hard enough to crack nuts of a certain kind, while at the same time being light enough so that they can be carried to the site of the nut-trees (Glock 2010). With respect to certain properties or kinds, this general qualm can be sharpened. Animals can recognize or perceive a physical phenomenon x as having certain conative – utilitarian, affective, emotional – qualities. They can recognize or perceive things – for example, as painful or stimulating,

Aspect-perception, perception and animals


dangerous or harmless, edible or inedible, interesting or dull, appealing or repugnant. In application to animals as much as to humans, ‘recognizing x as F’ is factive, whereas ‘perceiving as’ is non-committal. I shall refer to the latter as appreciation. Accordingly, the following scheme applies to some animals: (Appreciation): A can appreciate x as F. Intelligent higher animals can also recognize the instrumental significance, not just of inanimate objects but also of other creatures, whether conspecifics or human caretakers. (Instrumental Appreciation): A can recognize x as a way of attaining y. X can acquire significance for an animal A, for instance when A adopts a previously neutral feature of the physical environment as a landmark – for example, as the border of its territory. (Appreciation) looms large in Gibson’s (1979) celebrated theory of ecological perception. According to Gibson, animals can directly perceive so-called affordances. These are properties of the environment that are relevant to the possibilities for action and reaction on the part of the perceiving organism. A water surface, for instance, affords support for a water bug but not for a polar bear. And what is prey to one animal is predator to another. So one and the same x can be appreciated as F by A and as non-F or G by B. The lesson has not been lost entirely on philosophers either. Even commentators of differentialist proclivities have acknowledged both (appreciation) and (instrumental appreciation), although without paying sufficient attention either to the differences or the connections between them. MacIntyre (1999: 47–8) observes that, Heidegger notwithstanding, animal perception does feature as-structure. A dolphin or a gorilla ‘recognizes individuals, notices their absences, greets their returns, and responds to them as food or as source of food, as partner or material for play, as to be accorded obedience or looked to for protection and so on’. And Lovibond insists, against McDowell (Lovibond 2008: 116), that animals act for a reason in that certain things have a point or value in their view. Ultimately these concessions lead to the assimilationist conclusion that intelligent animals cannot just perceive an object x as F but also that x is F. But to make a compelling case for that contention is beyond the remit of this paper.7

Notes 1 Christoph Pfisterer alerted me to the possibility that Wittgenstein had in mind not a distinction between material and intentional object but between two types of linguistic construction—seeing x and seeing that p. But he characterizes both uses in terms of noun-phrases—seeing this and seeing a likeness. Whether the latter can be paraphrased in terms of seeing that x is like y in certain respects is a

98 Hans-Johann Glock



4 5

6 7

moot question. Note also that one might interpret the ‘this’ of the first use not as taking the place of a description, drawing or copy but as referring to one of these items. In that case the first kind of object would be a linguistic or non-linguistic representation, just like a grammatical object, a notion that Anscombe uses to characterize the intentional object. This tension also plays a role in Hanson’s idea (inspired by Wittgenstein) that scientific observation is ‘concept-laden’, as well as in contemporary debates about ‘nonconceptual content’ and ‘cognitive penetration’. But these lie beyond the remit of this paper. At the same time, a subject may be mistaken in claiming to have noticed an aspect in the first place—e.g. when she is persistently unable precisely to pinpoint the contours of the figure she avows to see in a picture puzzle. In that respect aspect-perception is linked to certain capacities closer to thinking. The appearance that aspect-perception is generally subject to first-person authority derives from the fact that ‘A sees x as (an) F’ often means something like ‘x appears to be (an) F to A’ rather than ‘A notices the aspect F of x’. The former, but not the latter, is immune to error and doubt. Similarly for ‘x looks (like an) F to A’ and ‘A sees x under the aspect F’ vs. ‘A recognizes x as F’. For ‘noticing’ and ‘recognizing’ are achievement verbs that preclude first-person authority. Finally, ‘x looks F’ does not carry first-person authority because how a thing looks, by contrast to how it appears to a particular observer, is an objective or at any rate intersubjective matter. Commenting on ‘straightforward perception’, the capacity shared by humans and animals to perceive things present to the senses, Wollheim (1980: 217) maintains: ‘seeing-as is directly related to this capacity, and indeed an essential part of it’. See Hanson (1958: 105, 172, 110): ‘different scientists are aware of the same but see different things’. ‘Almost everything we usually call seeing involves . . . seeing as’. ‘Seeing is a composite of visual experience and what we know. Hence seeing as and seeing that’. Searle (2015: 74, 110) also contends that all seeing is seeing-as and seeing that. But whereas Hanson obviously exemplifies option (4), since he regards seeing as theory-laden and, a fortiori, as conceptual, this is less clear for Searle. For a parallel criticism of the attempt to explain meaning by reference to interpretation see Glock 2003: 204–5. I have tried to make that case in Glock 2012. In making the current case I was aided and abetted by the editors of this volume, Meret Polzer, John Hyman and Christoph Pfisterer, as well as by audiences at Vienna and Zurich. Many thanks!

Bibliography Ahmed, Arif M. forthcoming. ‘Wittgenstein on Aspect Perception’. In H. J. Glock and J. Hyman (eds), The Blackwell Companion to Wittgenstein. Oxford: Wiley. Anscombe, G. E. M. 1965. ‘The Intentionality of Sensation: A Grammatical Feature’. In R. J. Butler (ed.), Analytic Philosophy. Oxford: Blackwell, pp. 158–80. Brandom, Robert B. 1994. Making It Explicit. Cambridge, MA: Harvard University Press. Day, William and Krebs, Victor J. 2010. Seeing Wittgenstein Anew. Cambridge: Cambridge University Press. Geach, Peter T. 1957. Mental Acts. London, New York: Routledge and Kegan. Gibson, James J. 1979. The Ecological Approach to Visual Perception. Boston: Houghton Mifflin. Glock, Hans-Johann 1996. A Wittgenstein Dictionary. Oxford: Blackwell.

Aspect-perception, perception and animals 99 Glock, Hans-Johann 2003. Quine and Davidson on Language, Thought and Reality. Cambridge: Cambridge University Press. Glock, Hans-Johann 2010. ‘Can Animals Judge?’. Dialectica 64: 11–33. Glock, Hans-Johann 2012. ‘Thought, Judgement and Perception’, Grazer Philosophische Studien 86: 207–21. Glock, Hans-Johann 2013. ‘Animal Minds: A Non-Representationalist Approach’. American Philosophical Quarterly 50: 213–32. Hanson, Norwood R. 1958. Patterns of Discovery. Cambridge: Cambridge University Press. Heidegger, Martin 1977 [1927]. Sein und Zeit. F. W. von Hermann (ed.), HeideggerGesamtausgabe, Vol. 2. Frankfurt am Main: Vittorio Klostermann. Husserl, Edmund 1973. Ding und Raum: Vorlesungen 1907. U. Claesges (ed.), Husserliana 16. The Hague: Martinus Nijhoff. Hyman, John 2006. The Objective Eye: Color, Form, and Reality in the Theory of Art. Chicago, London: University of Chicago Press. James, William 1950 [1918]. The Principles of Psychology. Vol.1. New York: Dover Publications. Jantschek, Thorsten 1995. ‘Sehen als; Aspektsehen’. In J. Ritter and K. Gründe (eds), Historisches Wörterbuch der Philosophie Vol. 9. Basel: Schwabe, pp. 162–5. Köhler, Wolfgang 1975 [1930]. Gestalt Psychology. New York: Mentor. Lovibond, Sabina 2008. ‘Practical Reason and Its Animal Precursors’. In J. Lindgaard (ed.), John McDowell: Experience, Norm, and Nature. Oxford: Blackwell, pp. 112–23. McDowell, John 1996. Mind and World. Cambridge, MA: Harvard University Press. MacIntyre, Alasdair C. 1999. Dependent Rational Animals: Why Human Beings Need the Virtues. London: Duckworth. Mulhall, Stephen 2001. ‘Seeing Aspects’. In H. J. Glock (ed.), Wittgenstein: A Critical Reader. Oxford: Blackwell, pp. 246–67. O’Sullivan, Michael 2015. ‘Judgement and Aspect: Tractatus 5.5423’. In M. Campbell and M. O’Sullivan (eds), Wittgenstein and Perception. London: Routledge, pp. 152–64. Putnam, Hilary 1975. ‘Brains and Behavior’. In H. Putnam (ed.), Mind, Language and Reality. Philosophical Papers Vol. 2. London: Cambridge University Press, pp. 328–42. Rödl, Sebastian 2007. Self-Consciousness. Cambridge, MA, London: Harvard University Press. Ryle, Gilbert 1980 [1949]. The Concept of Mind. London: Penguin. Schroeder, Severin 2010. ‘A Tale of Two Problems: Wittgenstein’s Discussion of Aspect Perception’. In J. Cottingham and P. M. S. Hacker (eds), Mind, Method, and Morality: Essays in Honour of Anthony Kenny. Oxford: Oxford University Press, pp. 352–71. Scruton, Roger 1981. ‘Photography and Representation’. Critical Inquiry 7 (3): 577–603. Searle, John R. 2015. Seeing Things as They Are. Oxford: Oxford University Press. Sellars, Wilfrid 1957. ‘Intentionality and the Mental. A Symposium by Correspondence with Roderick Chisholm’. In H. Feigl, M. Scriven and G. Maxwell (eds), Minnesota Studies in the Philosophy of Science, Vol. 2. Minneapolis: University of Minnesota Press, pp. 507–39.

100 Hans-Johann Glock Strawson, Peter F. 1974. ‘Imagination and Perception’. In P. F. Strawson, Freedom and Resentment and Other Essays. London: Methuen, pp. 45–65. Strawson, Peter F. 1992. Analysis and Metaphysics. Oxford: Oxford University Press. Vesey, G. N. A. 1956. ‘Seeing and Seeing As’. Proceedings of the Aristotelian Society 56: 109–24. Wollheim, Richard 1980 [1960]. Art and Its Objects. 2nd edition. Cambridge, MA: Cambridge University Press.


Wittgenstein’s seeing as A survey of various contexts Volker A. Munz

Points of contact: Wollheim, Gombrich and Wittgenstein When one starts to look a little closer at what Wollheim says about representational painting, and the relation between the spectator, the surface of the painting and its pictorial content, it becomes obvious that the twofoldness thesis is a reply to Ernst Gombrich. Gombrich argues that looking at a painting involves two different experiences, one of the surface and the other of the representational content of the painting. Where Wollheim originally thought that the viewing of a painting involves two experiences occurring simultaneously, Gombrich, in contrast, points out that my looking at a painting requires that I direct my attention either to the pictorial content or to the marks on the canvas, but never simultaneously to both. Otherwise the so-called ‘alternating illusion view of depiction’ would not be possible – for example, my looking at a portrait and my attending to the content would be experimentally indistinguishable from seeing the person portrayed (Winters 2003: 2). Gombrich illustrates this point by telling a story about Kenneth Clarke, a ‘master of introspection’, who described a recently experienced situation in which he was defeated when he attempted to stalk an illusion. Looking at a great Velazquez, he wanted to observe what went on, when the brush stripes and dabs of pigment on the canvas transformed themselves into a vision of transfigured reality as he stepped back. (Gombrich 1956: 6) While stepping backwards and forwards, Clarke realised that he could never hold both visions at the same time. To support his view, Gombrich also refers to Jastrow’s duck–rabbit head. Here, too, we can switch between the two aspects, of a duck and a rabbit, but never see the two aspects simultaneously. So, Gombrich tries to model, for instance, seeing a portrayed figure and seeing the portrait as a painted surface on the aspect change between duck and rabbit. He does, however, concede that we do not have the illusion that we perceive a real duck or a real rabbit, since the shape on the paper does not resemble either of the two animals very closely (ibid: 5). Interestingly enough, here Gombrich speaks in terms of ‘switching from one interpretation

102 Volker A. Munz to the other’ (ibid). On the other hand, he talks in terms of a ‘shape’ that transforms itself in some subtle way ‘when the duck’s beak becomes the rabbit’s ears and brings an otherwise neglected spot into prominence as the rabbit’s mouth’. Therefore, we are forced ‘to look for what is “really there”, to see the shape apart from its interpretation’ (ibid). However, according to Gombrich, this is not really possible. Of course, we can switch from ‘one reading’ to the other, as he puts it, or remember the rabbit while looking at the duck. But we simply cannot experience the two alternative ‘readings’ at the same time. We cannot see both the duck head and the rabbit head simultaneously, just as we cannot realise the two possibilities of viewing the Necker Cube at the same time. It is, however, also quite obvious that the analogy is pretty flawed. The duck–rabbit case is a case of two aspects belonging to the pictorial content, whereas Gombrich’s ‘aspect change’ concerns the depicted content on the one hand and the pictorial surface on the other. This disanalogy is also pointed out by Wollheim. The analogy only holds between seeing the drawing line that forms the ambiguous figure and ‘whichever of the duck-picture or the rabbit-picture I see the black line as’ (Winters 2003: 2). I cannot see the black drawing as both a duck-picture and a rabbitpicture, but it does not follow that I cannot see the black lines and either of the two pictures together. One could even say: I see the duck by seeing the black lines. For Gombrich, in contrast, only the impossibility of simultaneously viewing both the surface and the content aspect allows for the ‘alternating illusion view of ’. Wollheim, however, points out that when we look at pictures as pictures, we do not suffer the effects of a visual illusion. In his paper ‘On Pictorial Representation’, Wollheim talks about the special skill called ‘seeing in’, which ‘furnishes for each representation, its appropriate experience. For that is the experience of seeing in the pictorial surface that which the picture is of’ (Wollheim 1998a: 221). Hence, when looking at a surface of a painting, we are ‘visually aware at once of the marked surface and of something in front of or behind something else’ (ibid). This phenomenological feature Wollheim calls ‘twofoldness’. Originally, he also identified this twofoldness with two simultaneous perceptions, i.e. one of the surface and one of what is represented. According to this view, seeing the canvas and seeing the pictorial content both correspond to genuine perceptual claims. And these two experiences are somehow bound together in the perception of the painting. In Painting as an Art, Wollheim changes his view and argues that there is only one experience, which divides into two aspects. These two aspects he calls ‘configurational’ and ‘recognitional’ aspects (1998b; see also 1998a: 221). In order to see any enlightening parallels between Wollheim’s ‘seeing in’ and what Wittgenstein calls ‘seeing as’, one has to be clearer first of all about what fascinated Wittgenstein so much about his concept of ‘seeing as’ and its connection to aspects and aspect changes. Generally speaking, as Eilan points out: There are, according to Wittgenstein, many different types of aspects, introduced largely recessively, via different requirements on noticing

Wittgenstein’s seeing as 103 changes in aspect. Some such noticings require concepts, others do not; some require familiarity with past instances, others do not; some require imagination, others do not, and so forth. Given this variation, there is no reason to expect there to be a single, uniform account of the role of aspects in seeing, nor that what is interesting about aspects is the same in all cases. It patently isn’t. (Eilan 2013: 3) This remark precisely points to what is crucial for Wittgenstein’s whole discussion of aspects and aspect seeing, namely that there is no general moral that could be drawn from all the various different examples, except that there is none. And this is what this paper hopes to bring out. As already pointed out, the duck–rabbit case is one between two changing pictorial contents and therefore a misleading analogy to Gombrich’s case of a portrayed figure and the painted surface. There are, however, also cases where the relation between the marks and what we see these marks as plays a role in Wittgenstein. Furthermore, when Wittgenstein talks about the concept of an ‘organisation’, both in connection with Köhler as well as in a rather different context in his Lectures on Philosophical Psychology, we also find remarks that are connected to Wollheim’s concept of ‘configurational aspects’. And when Wittgenstein speaks about the resemblance relation between the picture-object and other objects, it seems obvious that recognition plays an essential role. Finally, Gombrich’s use of the terms ‘interpretation’ and ‘readings’ can be related to Wittgenstein’s discussion of seeing and interpreting a picture and the role of the intentions of a subject in such cases. It might therefore be helpful to try to give some kind of survey of the various contexts in which Wittgenstein discusses the phenomenon of ‘seeing as’. One way of doing this is to roughly divide the following discussion into two parts. The first will point at the role of the perceiving subject, the second at the perceived object. Within this rough separation, we will have to look at both the various types of drawings involved and the various concepts Wittgenstein discusses in those contexts. Such concepts are, for example, ‘private inner pictures’, ‘reaction’, ‘attitude’, ‘will’, ‘imagination’ and ‘interpretation’, which are concerned with the subject, as well as ‘organisation’, ‘internal relation’ and ‘recognition of similarities’, which are related to the object.

Aspect changes and the paradox of sameness and change Generally speaking, different experiences of aspects show various family resemblances. Such experiences can be: seeing similarities, seeing aspects of a puzzle picture, imagining something as something, hearing something as something, experiencing the meaning of a word, etc. What is common to all these various aspect experiences is the form in which all these experiences are expressed, i.e. ‘Now, I see it as that’, ‘Now I see it so’, ‘Now it is this – now that’, ‘Now I hear it as’, ‘I heard it before as’, etc. (LW I §697). These expressions describe the experience of an aspect change. But our explanation of the

104 Volker A. Munz ‘that’s and ‘so’s is quite different in different cases (ibid). To put it in other words: ‘The difference is contained in the description that one uses in reporting the aspect’ (LW I § 696). Accordingly, the different descriptions of ‘aspect change’ point to a paradox, which is inherent in the form of the expression of aspect experiences: ‘The expression of a change of aspect is the expression of a new perception, and at the same time of the perceptions being unchanged’ (PI II: 196). Asking about the difference, Wittgenstein answers that ‘I describe the alteration like a perception; quite as if the object itself had altered before my eyes’ (ibid: 195). The notion of description of what is seen plays a crucial role in Wittgenstein’s discussion of aspects (for example, PI II xi: 193, 195, 197–200, 202–4, 207, 209). In experiencing an aspect change, one becomes ‘conscious’ of the aspect (LW I §169), which presupposes that the spectator has to occupy herself with the object (ibid § 555). And it seems as if the aspect experience changes the former seeing experience in such a way as to become inclined to talk in terms of a new seeing experience, as if the object had changed. However, the point of the new experience is not that we look at the object because we notice that it has actually changed, but rather that it has changed in spite of the fact that it has not changed in itself. ‘I see that it has not changed; and yet I see it differently’ (ibid: 193). And, again, both change and unchange come together in the expression of the seeing experience. Or as Wittgenstein puts it (LW I § 174): ‘What is strange is really the surprise; the question: How is it possible! It might be expressed by: The same – and yet not the same.’ Here, too, Wittgenstein’s use of the term Ausdruck (‘expression’) in grasping the paradox shows that the aspect change takes place in language. This is an important point. When we look at all the various contexts and examples Wittgenstein introduces, it also becomes obvious that the seeing experience depends upon both the perceiving subject and the object perceived.1 Let us therefore briefly look at some aspects that concern the role of the subject.

The perceiving subject: inner picture, primitive reactions, attitude, will and interpretation One way of making different subjective contexts explicit is to look at the role the subject plays in the various examples Wittgenstein gives. First, what seems obvious when we look at the expressions of aspect changes is that they are not descriptions of any kind of inner pictures or mental states. Wittgenstein discusses this erroneous assumption of an inner picture that changes as an answer to the paradox, in various passages, particularly in PI II: xi and LW I §§ 516 onwards. What leads to the misleading concept of an inner picture is the assumption of organising the visual impression by means of colours and shapes, or, as Wittgenstein puts it: If you put the ‘organization’ of a visual impression on a level with colours and shapes, you are proceeding from the idea of the visual

Wittgenstein’s seeing as 105 impression as an inner object. Of course this makes this object into a chimera, a queerly shifting construction. For the similarity to a picture is now impaired. (PI II xi p. 196) In RPP I, II and LW I, II, Wittgenstein discusses this point in connection with his criticism of Köhler, whose concept of ‘organisation’ can only refer to a kind of inner picture, since the object itself has not changed (see, for example, Glock, 1996: 37–8). In other words, if one sees the concept of organisation on the same level as colours and shapes one would have to assume two different visual impressions, one before and one after the aspect change, since one and the same perception cannot have two distinct shapes or colours at the same time. This seems to force Köhler to assume two different inner objects, since the paradox requires that the external object itself has not changed. The inner objects would then be the subjects of the changed colours, shapes or organisational aspects (Eilan 2013: 6). But this would turn the object into a ‘chimera’. In other words, Köhler wants to get rid of the paradox by introducing inner objects, whereas the real demand lies in trying to explain the paradox, which itself is a description of our experiences and cannot therefore be argued away. In RPP I, Wittgenstein makes the following remark in the context of Köhler’s example of the Mediterranean Sea: When one fails to recognize the Mediterranean on the map with different colouring, that does not show that there is really a different visual object before one (Kohler’s example). At most that might give a plausible ground for a particular way of expressing oneself. For it is not the same to say ‘That shows that there are two ways of seeing’ – and ‘Under these circumstances it would be better to speak of ‘two different objects of sight’. (RPP I §1035) One reason for pointing out the difference is precisely the paradox itself, which demands that in one sense the object of perception does not change at all; a point, however, which the introduction of a new object would not allow. More generally speaking, the very idea of an inner object has to do with the misleading idea of a representational conception of perceiving objects in the first place. At one point Wittgenstein remarks: What I really see must surely be what is produced in me by the influence of the object. Then, what is produced in me is a sort of copy, something that in its turn can be looked at, can be before one; almost something like a materialization. (PI II xi: 199) Second, Wittgenstein sometimes argues that the expressions of aspect changes have the form of avowals or primitive reactions, quite similar to crying

106 Volker A. Munz ‘Ouch!’ in the case of feeling a sudden pain. In some cases, the primitive reaction of pointing with my finger can be called an expression of ‘seeing it so’ (RPP I § 1048). Another primitive reaction would be the recoiling from a strangely written word (ibid § 1087) or a context in which one sees a painted horse running. In such a case, I do not just want to say that I know that this represents a running horse. It is much more like a primitive ‘Tally ho!’ or an exclamation like ‘Oh, it’s running’ or ‘See how it runs!’. And obviously this does not have the character of a piece of information (ibid §874). Interestingly enough, in the context of primitive reactions, Wittgenstein also draws a distinction between the cases of the duck–rabbit and the double cross. In the case of the duck–rabbit aspects, one has to be familiar with the shapes of the animals. Here, Wittgenstein uses the word innehaben (LW I § 700, § 702).2 The fundamental aspects of the double cross, however, ‘could express themselves in primitive reactions of a child who could not yet talk’ (LW I § 700). This might happen by simply pointing, alternatively, to an isolated white cross and an isolated black cross. Wittgenstein explicitly emphasises that this might be a primitive reaction of a child who does not yet have a concept of a cross or cannot speak. The report of the two aspects would therefore just consist in pointing to the relevant parts of the double cross. No further familiarity seems involved here. The duck–rabbit case, in contrast, could not be described in a similar way (ibid § 701). Here we have to presuppose a familiarity with the shapes of both a duck and a rabbit, which is not necessary in the case of the double cross. Wittgenstein, does not, however, explicitly say whether innehaben also includes having the relevant concepts ‘duck’ and ‘rabbit’. Also, showing the child a photograph of her mother could lead to a primitive pointing or to a smile in the child’s face, even before he or she is able to talk. This shows that the identification of similarities does not necessarily presuppose the possession of the relevant concepts. Third, when looking at more complex contexts, we find another quite familiar Wittgensteinian notion, namely our attitude towards a particular painting, photograph, drawing, etc. For instance, when a person describes what she sees by saying ‘to me it’s an animal pierced by an arrow’ (§ 665), this could be an expression of how she treats the figure. In other words, it expresses a particular attitude towards the figure. This, Wittgenstein argues, is ‘one meaning in calling it a case of seeing’ (ibid). In this sense, one could say that the person only sees it this way as long as she has this particular attitude towards it (LW I § 670). Additionally, we also expect different things from someone who sees the picture as that particular animal, or in cases of drawings of a human face, or ‘picture faces’, as Wittgenstein calls them (PI II xi: 197). Here we stand in a relation to it, as we do towards a human face, in so far as we react to it as to the expression of such a face (ibid: 194). In Philosophical Investigations II, section xi, Wittgenstein further points out the various roles that pictures like paintings, as opposed to, for example, working drawings, play in our lives. One way of putting this is

Wittgenstein’s seeing as


to say that we regard the photograph of our beloved or the painting on our wall as the very object or subject that is depicted. And that may also be something Gombrich had in mind, although erroneously, when he spoke about the ‘illusion’ when treating a portrait as the portrayed person herself. In the context of dealing with photographs and pictures on the wall, Wittgenstein introduces the alternative concept of ‘viewing’ (betrachten) as opposed to seeing as (so Sehen). We view the photograph of a particular person as the very person herself, which, however, does not have to be so in cases where people would have a different attitude towards photographs – for example, being repelled by them (PI II xi: 205).3 The answer to the question ‘When and how long do we view a portrait as a human being?’ – ‘Always, if we see it at all (and do not, say, see it as something else)’ (ibid, my italics) – would determine the concept of ‘viewing’ as opposed to ‘seeing as’, when I am actually engaged in dealing with a picture. This grammatical distinction obviously shows that ‘permanent aspect perceptions’ are not genuine cases of seeing aspects. The opposite view (see, for example, Mulhall 2001: 246–55; Schroeder 2010: 353–4) is commonly defended by reference to Wittgenstein’s following remark: ‘How would the following account do: ‘I can see something as whatever it can be a picture of?’ What this means is: the aspects in a change of aspects are those which, in certain circumstances, the figure could have permanently in a picture’ (PI II xi: 201). However, as Wittgenstein later tries to capture with his notion of ‘viewing’ (betrachten), one need not be actually engaged in seeing a portrait as a person portrayed. Aspect seeing requires attention; viewing does not. When I see a change of aspect I have to occupy myself with the object. I occupy myself with what I am now noticing, with what strikes me. In that respect, experiencing a change of aspect is similar to an action. (LW I §§ 555–6) It is, again, our attitude and the way we react towards a portrait that is of relevance here, or, as Wittgenstein puts it when elaborating on the idea of aspects being permanent: ‘This is how we react to the picture’ (PI II xi: 201). Generally speaking, ‘You need to think of the role which pictures such as paintings (. . .) have in our lives’ (ibid). Moreover, the cases of portraits seem utterly different from those of imaginary ones, when seeing, for example, a triangle as a hole, an aspect one could hardly argue the triangle might permanently have. Fourth, in the context of the duck–rabbit and the double cross, Wittgenstein now draws a further distinction between aspects related to a double cross as opposed to a triangle, a distinction he regards as fundamental. Here, another subjective feature takes place: our human capacity to imagine. In order to see the triangle in a particular way, what is needed is our imagination, in contrast to the case of the double cross. Wittgenstein offers various alternatives of how to see the triangle – for example, as half of a parallelogram, as

108 Volker A. Munz fallen over or as a triangular hole. In this case, one must be familiar with the various concepts involved: ‘hole’, ‘parallelogram’, etc. Furthermore, it becomes quite obvious that we use the verb ‘to see’ in a rather different sense, by employing our imaginary capacities. If someone does not see the rabbit in the duck–rabbit figure, or the black and white cross, she is simply overlooking something in the picture. This, however, does not apply to the triangle. Here it is not that you can see (or overlook) an object that has fallen over: it is rather something to imagine. Besides, it would be quite odd to speak in terms of a relation of similarity between a triangle and, say, a hole, let alone an internal one. We can of course take the duck–rabbit for the picture of a rabbit or the double cross for the picture of a black cross, but we cannot take the triangle for the picture of a hole, or of something that has fallen over. To see this aspect does indeed require imagination (LW I 703; PI II xi p. 207). Additionally, you would not and need not point to, for example, a triangular hole in order to point to one aspect you see the triangle as. The imagination does not even seem to presuppose that there is such a thing you see it as. At least, it hardly happens that we speak in terms of fallen-over triangles or triangular holes. It would also be difficult to teach someone ‘seeing a triangle as fallen over’ or ‘black triangular hole’ by showing her a triangle with a long hypotenuse or by showing a picture of a black triangular hole, if you could find one in the first place. When Wittgenstein talks about ‘seeing an aspect’ and ‘imagining’, he remarks that these concepts are akin. So ‘now seeing as’ and ‘now having this image’ resemble each other. For instance, one needs imagination to hear a particular piece of music as a variation of a theme, but at the same time one also perceives something in hearing it this way. Wittgenstein then goes on to argue that both concepts are subject to our will, and tries to make this point clear by drawing parallels between an order such as ‘Imagine this’ and also ‘Now see the figure like this’; but not ‘Now see this leaf green’ (PI II xi: 213). Here we see a fundamental difference between seeing something and seeing something as something. In order to see the duck–rabbit as, for example, a duck you must first of all see the duck in the picture, like in other cases of veridical perceptions. If I saw the duck-rabbit as a rabbit, then I saw: these shapes and colours (I give them in detail) – and I saw besides something like this: and here I point to a number of different pictures of rabbits. – This shows the difference between the concepts. ‘Seeing as . . .’ is not part of perception. And for that reason it is like seeing and again not like. (ibid: 196–7) In this sense, you cannot see the duck–rabbit as, for example, a turtle. This is again different in the example of the triangle where it would not make sense to look for the triangular hole in the picture if you cannot ‘see’ it at first sight.

Wittgenstein’s seeing as


Accordingly, seeing and imagining are different phenomena. ‘I see’ is used differently from ‘I imagine’. An order to see is different from an order to imagine, and ‘I try to see it’ is different from ‘I try to imagine it’, etc. One point that comes out here is the difference between the subject being either passive or active, i.e. that the perceptions of things we normally experience are somehow imprinted upon us, which is covered by the term ‘impression’.4 In this sense, ‘seeing’ is different from ‘imagining’ because our imaginative powers are not restricted in the same way as particular variations of seeing aspects are. As already mentioned, it would make little sense to give someone the order ‘See the duck–rabbit head as a turtle’ or ‘See it as the shape of the German country with its capital marked with a dot’. In order to see the duck–rabbit either as a duck or as a rabbit, it must contain the very possibility to be seen this way. Here, as we shall see later, the concept of an internal relation comes in between the picture and other pictures or objects – for example, rabbits or ducks – or of the pictures and part of the pictures. The words ‘trying to see X as Y’ must somehow presuppose the possibility of seeing it in a veridical sense, in contrast to ‘trying to imagine X as Y’. This is, however, a requirement on the side of the object not the subject. In this sense, our will is limited in a way, which does not hold for the case of our imagination, as pointed out in the triangle case. Fifth, I want to briefly mention Wittgenstein’s discussion of the relation between seeing and interpreting, or seeing and thinking respectively. Here, in order to dissolve the paradox, we could assume that it is not our inner picture that has changed but rather our interpretation of the perceived object. We could then ask whether noticing an aspect is either a seeing or a thinking. Wittgenstein argues that someone who has a visual experience expressed by a primitive avowal is also thinking of what she sees. It is for this reason that the flashing of an aspect seems half visual experience, half thought (PI II xi: 197). At least, it is obvious that there is quite a range of cases with quite different degrees of seeing and thinking involved, as we can see in the case of the double cross, the duck–rabbit and the triangle: for example, primitive pointing to the black or white cross versus the innehaben of shapes in the case of the duck–rabbit, recognising similarities being subject to the will or imagining in the case of the triangle. And Wittgenstein also draws a distinction between ‘seeing’ as a state with a particular duration as opposed to ‘thinking’ as an action, i.e. something we deliberately do. In the case of the drawing of a cube, however, we get the impression of a very close connection between seeing and interpreting when Wittgenstein speaks about the accompanying textbook that sometimes refers to it as a glass cube, sometimes as an inverted open box, a wire frame, etc. Here, he argues that the text supplies the interpretations of the illustration, but that we could also see the drawing as one or the other thing. ‘So, we interpret it, and see it as we interpret it’ (PI II xi: 193) This, however, does not mean that we somehow try to fit, for example, the triangle into a form it does not really fit into or, as Wittgenstein puts it: ‘No squeezing no forcing took place here’

110 Volker A. Munz (ibid: 200). Accordingly, the point seems to be that we cannot distract the interpretation from the perceiving of the ‘raw material’, so to speak. Therefore, the description of an immediate experience is in no way an indirect description of it. To see X as Y is having a particular visual experience, which goes experientially hand in hand with interpreting it as Y. In the case of recognising an acquaintance in a crowd, Wittgenstein asks whether it is a special kind of seeing, or even both seeing and thinking, or, as he puts it, ‘an amalgam of the two’ (ibid: 197). However, as already mentioned in the context of ‘permanent aspect seeing’, this does not mean that every seeing is to be understood as a seeing as, which Wittgenstein points out in various passages. There is no seeing the fork as a fork, the lion as a lion, the letter F as the letter F or the photograph as the photograph of a human being (see, for example, PI II xi: 195; RPP II, § 515).

The perceived picture: internal relations and organisation Finally, let us look at two points that belong to the object side. One is Wittgenstein’s already mentioned concept of an internal relation, the other is related to so-called ‘organisational aspects’. Already in a notebook entry from November 1914, Wittgenstein notes: ‘Puzzle pictures and the seeing of state of affairs’ (Sachverhalte). In Tractatus 5.5423, we find another interesting remark in connection with the Necker cube. Wittgenstein writes: To perceive a complex means to perceive that its parts are combined in such and such a way. Perhaps this also explains the fact that one can see the figure as a cube in two ways; and all similar phenomena. Because we quite truly [eben wirklich] see two different facts. Here it is interesting to see that Wittgenstein explicitly speaks of two different facts, which is obviously due to his picture-theoretical approach. In reply to Wittgenstein’s ‘For we really see two different facts’, Charles Travis writes: But surely not. Seeing the Necker in one way or another, or switching back and forth, is a visual phenomenon. A change in aspect is a change in visual awareness. So that one can perfectly well be aware (see) that the lines of the Necker are an image of a cube in orientation A while seeing the Necker as an image of a cube in orientation B. (. . .) Once again, ‘see that’ is not a perceptual use of ‘see’. (Travis 2014: 48)5 However, Wittgenstein’s remark about the two facts immediately follows his discussion of the phrase ‘A believes that p’ (TLP 5.542–5.5423). At least in this context it is quite clear that the psychological subject A plays no role. One could object that we should ignore the context. When we look at the

Wittgenstein’s seeing as 111 Necker alone, we will see that Wittgenstein is wrong. But according to the Wittgenstein of the Tractatus, to perceive a complex means to perceive how the constituents are related to one another and that they are related in this way (ibid: 5.5423). In one case, the As appear in front, in the other the Bs. And I think it is precisely for this reason that Wittgenstein chooses a threedimensional representation. Depending upon the different spatial relations, we would also give a different description of the Necker. This shows that indeed we see two different facts and that the Necker case is one that belongs to the side of the object and not to the awareness of the perceiving subject. In 3.1431, Wittgenstein asks us to imagine the propositional sign as a complex of spatial objects, such as tables or chairs, i.e. as a fact, and writes: ‘Then the spatial arrangement of these things will express the sense of the proposition’. The only similar case in Wittgenstein’s discussion of aspect seeing I know is his example of a special figure representing a convex step (PI II xi: 203). And even here Wittgenstein remarks that if one finds ‘that the flat aspect alternates with a three-dimensional one, this is just as if I were to show him completely different objects in the course of the demonstration’. Furthermore, to be able to see an object like the Necker in two different ways seems to imply that it must be possible to see it in two different ways, or, to put it differently, we can only make ourselves pictures if they are pictures of possible states of affairs in the first place. In his paper on Pitcher’s The Philosophy of Wittgenstein, Rhees discusses Pitcher’s misleading idea that for Wittgenstein the correlation of pictorial elements with elements of reality is a mental act (Pitcher 1964: 88). Rhees remarks: The question ‘How is the picture connected with the fact it pictures?’ can only mean: ‘How does it have the role of a picture at all?’ If ‘I’ were to perform the mental act or intention of correlating the marks or sounds with elements of a fact this would not make the complex a picture. I cannot ‘correlate’ unless I am working with a picture. And then the correlating means simply that the syntax of the picture shows the possible structure of the fact. The relations of logical syntax are internal relations, and they make it possible to introduce conventional signs. (Rhees 1970: 40) This is also a very good case of what Wittgenstein meant by the concept of internal relations in the Tractatus. And in the case of seeing aspects, too, it seems that, at least for particular kinds of drawings, the aspects must somehow already be contained in the picture. In this sense, we could also take Rhees’s remark as a criticism of Köhler. For example, in a puzzle picture it would make no sense to look for the face in the tree if there were not anything that could look like a face, and similarly in the case of the duck–rabbit. These may also be examples where Wittgenstein’s idea of an internal relation comes in. Interestingly enough, at one point Wittgenstein remarks that

112 Volker A. Munz ‘by noticing the aspect one perceives an internal relation, and yet noticing an aspect is related to forming an image (Vorstellen)’ (LW I § 733). Here, the more immediate context is related to the question whether noticing a likeness – for example, between a face on a street and a friend of mine whom I have not seen for ages – is a seeing or not. There are, of course, other examples, such as the triangle, where it would be difficult to show in what sense the aspect of a hole must somehow be in the picture of the geometrical drawing. In LW I § 506, Wittgenstein hesitantly asks whether ‘seeing an aspect’ means perceiving an internal relation. A few paragraphs later (§ 516), he argues that what I perceive in the dawning of an aspect is not a property of the object, but an internal relation between it and other objects. This could be the relation between a face on the street and my old friend, a similarity between two faces or between the duck head and real ducks. But, nota bene, the dawning of the aspect itself is not the perceiving of the internal relation. The internal relation is rather what I perceive in the dawning of an aspect. It is through the noticing of an aspect that I perceive an internal relation. (This also reminds us of the famous ‘saying – showing distinction’). The talking in terms of internal relations seems, however, to indicate, that seeing something as something cannot be solely explained in terms of the perceiving subject. Again, only what already is a picture, as opposed to a mere bunch of lines and dots, can stand in an internal relation to the object depicted. In what sense the drawing of a triangle must allow for seeing it as a hole is certainly a different question. At least we could say that it has to be a closed figure and not just one or two lines, for example. To come to the last point, we find similar ideas in Wittgenstein with respect to the organisation of a drawing: the misleading idea of a different organisation of my private visual impression in the context of Köhler, belonging to the subject, I have already mentioned. In the context of the perceived object, however, Wittgenstein remarks that one kind of aspect could be called ‘aspects of organisation’: ‘And when the aspect changes, parts of the picture go together, which before did not. In the triangle I can see now this as apex that as base, now this as apex that as base’ (PI II xi: 208). This point becomes even more obvious in the case of a puzzle picture, where I suddenly realise that certain lines do not belong to the leaves of the tree but instead form the body of a mouse: ‘I suddenly see the solution of a puzzle-picture. Before there were branches there; now there is a human shape’ (ibid: 196). The same holds for the double cross, being seen as a white cross with a black background or vice versa. In the 1946–7 lectures on Philosophical Psychology (Wittgenstein 1988: 101), Wittgenstein points out that the word ‘organisation’ fits cases like seeing a row of five dots ‘.. . ..’ as 2 + 1 + 2 or 2 + 2 + 1 dots, as opposed to the standing and hanging aspects of the famous triangle. Similarly, the different arrangement of background and foreground could be regarded as an organisational aspect.

Wittgenstein’s seeing as


Again in response to Köhler, Wittgenstein remarks: E.g., it is, I believe, misleading when Köhler describes the spontaneous by saying: the lines which belong to the same aspects of the figure arm in one aspect, now belong to different arms. That sounds as if what were in question here were again a way of taking these radii together. Whereas, after all, the radii that belonged together before belong together now as well; only one time they bound an ‘arm’, another time an intervening space. (RPP I § 1117) Here, I think, Wittgenstein wants to make the point that you cannot say, as Köhler does, that now these radii belong together and now those, depending upon whether we choose to refer to three big arms or three small arms. It is rather that the same radii still belong together in both cases as boundaries, only either as delimiting an arm or an intervening space. It is in this sense that we are already dealing with ‘arms’ and ‘intervening spaces’ and not just with meaningless lines which we plainly count and group differently, whereas ‘The picture does not organize itself under our gaze’ (RPP I § 1125). If I get Wittgenstein’s point, Köhler’s approach could not explain the aspects involved. In this very same lecture on Philosophical Psychology, where Wittgenstein discusses the seeing of a row of dots, he raises the question: ‘Can we find some road through all the phenomena of aspects?’ (Wittgenstein 1988: 105) What I have been trying to show is that there is no such single road but rather, as in many other of Wittgenstein’s topics, a number of ‘sketches of landscapes, which were made in the course of these long and involved journeyings’ (PI: IX).

Notes 1 For further considerations about the role of the subject and object see Hiltmann 1998, particularly chapters 2 and 3. 2 Translations such as ‘thoroughly familiar’ and being ‘conversant’ are only paraphrases of innehaben. 3 Another case of having a certain attitude towards a drawing is that of an aspectblind person who does not see, for example, the spatial variations of a cube diagram. She would therefore not be able to see the drawing as a drawing of a spatial object. This would show that she had a different attitude from ours. It could rather resemble an attitude we have towards a blueprint, which means that the aspect-blind person may use it as an instruction to form a particular material object. A similar case would be that of one who misses absolute pitch. These are, however, only two examples that deal with the reactions and attitudes of a perceiving subject. 4 Hume, for instance, draws the distinction between ideas and impressions by emphasising the ‘force and liveliness’ of impressions, ‘with which they strike upon the mind, and make their way through into our thought or consciousness (. . .)’, entering ‘with most force and violence’ (Hume 1985: 49). 5 This quotation as well as many fruitful discussions I owe to Bernhard Ritter.

114 Volker A. Munz

References Eilan, Naomi 2013. ‘On the Paradox of Gestalt Switches: Wittgenstein’s Response to Köhler’. The Journal for the History of Analytical Philosophy 2(3): 1–20. Glock, Hans-Johann 1996. A Wittgenstein Dictionary. Oxford: Blackwell. Gombrich, E. H. 1956. Art and Illusion. A Study in the Psychology of Pictorial Representation. Princeton: Princeton University Press. Hiltmann, Gabrielle 1998. Aspekte sehen. Bemerkungen zum methodischen Vorgehen in Wittgensteins Spätwerk. Würzburg: Königshausen und Neumann. Hume, David 1985. A Treatise of Human Nature. E. Mossner (ed.). London: Penguin. Mulhall, Steven 2001. ‘Seeing Aspects.’ In H.-J. Glock (ed.), Wittgenstein. A Critical Reader. Oxford: Blackwell, pp. 246–67. Pitcher, George 1964. The Philosophy of Wittgenstein. Englewood Cliffs, NJ: PrenticeHall Inc. Rhees, Rush 1970. ‘The Philosophy of Wittgenstein’. In R. Rhees, Discussions of Wittgenstein. London: Routledge, pp. 37–54. Schroeder, Severin 2010. ‘A Tale of Two Problems: Wittgenstein’s Discussion of Aspect Perception’. In J. Cottingham and P. M. S. Hacker (eds), Mind, Method, and Morality: Essays in Honour of Anthony Kenny. Oxford: Oxford University Press, pp. 352–71. Travis, Charles 2014. ‘Suffering Intentionally?’. In M. Campbell and M. O’Sullivan (eds), Wittgenstein and Perception. London: Routledge, pp. 45–62. Winters, Edward 2003. ‘Pictures and Their Surfaces: Wollheim on “Twofoldness”’. http://www.um.es/logica/Winters.htm. Wittgenstein, Ludwig 1982. Last Writings on the Philosophy of Psychology, vol. 1. G. H. von Wright and H. Nyman (eds), C. G. Luckhardt and M. A. E. Aue (trans.). Oxford: Blackwell. Wittgenstein, Ludwig 1992. Last Writings on the Philosophy of Psychology, vol. 2. G. H. von Wright and H. Nyman (eds), C. G. Luckhardt and M. A. E. Aue (trans.). Oxford: Blackwell. Wittgenstein, Ludwig 1953. Philosophical Investigations. G. E. M. Anscombe and R. Rhees (eds), G. E. M. Anscombe (trans.). Oxford: Blackwell. Wittgenstein, Ludwig 1980. Remarks on the Philosophy of Psychology, vol. 1. G. E. M. Anscombe and G. H. von Wright (eds), G. E. M. Anscombe (trans.). Oxford: Blackwell. Wittgenstein, Ludwig 1980. Remarks on the Philosophy of Psychology, vol. 2. G. H. von Wright and H. Nyman (eds), C. G. Luckhardt and M. A. E. Aue (trans.). Oxford: Blackwell. Wittgenstein, Ludwig 1961. Tractatus Logico-Philosophicus. D. F. Pears and B. F. McGuinness (trans.). New York: Humanities Press. Wittgenstein, Ludwig 1988. Wittgenstein’s Lectures on Philosophical Psychology 1946–1947. P. T. Geach (ed.). New York: Harvester. Wollheim, Richard 1998a. ‘On Pictorial Representation’. The Journal of Aesthetics and Art Criticism 56(3): 217–26. Wollheim, Richard 1998b. Painting as an Art. Cambridge, MA: Harvard University Press.

Part III

Benefits from Wollheim’s borrowing from Wittgenstein

This page intentionally left blank


Leonardo’s challenge Wittgenstein and Wollheim at the intersection of perception and projection Garry L. Hagberg

Leonardo famously advised young artists to sit in front of a heavily cracked plaster wall for enough time and with enough visual imagination to see landscapes and figures in it. Because plaster evidently does not crack with representational intentions, we cannot appeal to any criterion of correctness of perception that is based on intentional content. Yet the imaginationenriched perception is not arbitrary; we cannot see just anything in the plaster cracks at will. There must obtain a relation between the appearance of the lines formed by the cracks and the lines of the scene as seen within them—but this need not conform to any real scene in the world either. Yet it must be possible in the real world: what we know of the appearances of its possibilities thus informs, and sets constraints upon, what we can see in the wall. Leonardo set the imaginations of his pupils free—but still with the requirement that they creatively interact with materials external to their minds, materials that can be made representational. Materials that can be seen-as, lines that can be seen-in. The full explanation of this phenomenon is more intricate than one might initially imagine: one might begin with a stark distinction between (a) brute perception on the one hand—seeing things rightly and as they actually are— with (b) imaginative projection on the other—seeing, or “seeing,” as children do, things in clouds, and seeing things, as Leonardo’s understudies did, in cracked plaster. But this distinction starts to break down or blur when we look closely both at numerous examples of imaginative seeing and at the influences on our thought that would lead us to oversimplify the real complexity these cases present. Closely looking is precisely what Wittgenstein did with unprecedented acuity in section xi of part II of his Philosophical Investigations, which was a central influence on Richard Wollheim’s thinking about aesthetic perception. Wollheim developed his concept of “seeing-in” out of Wittgenstein’s notion of “seeing-as,” or aspect-perception; seeing-in is for Wollheim a two-fold yet single perceptual experience, in which we are aware both of the representational content and the non-representational markings within which we see that content. It is, in short, what Leonardo was talking about as essential to the development of the eye and the mind of the artist: we see the cracks and we see the landscapes in them. Yet in art,

118 Garry L. Hagberg unlike the wall, we add intentional content, which plays its role in ways that turn out also to be more complex than we might initially think. In Section I of this chapter, I begin with the perception/projection dichotomy and its foundational presupposition. In Section 2, I will discuss Wittgenstein’s distinction between two uses of the word “see” and its special significance for aesthetic perception. In Section 3, I turn to Rembrandt and the philosophical work his paintings can do in connection with Wittgenstein’s distinction. In Section 4, I consider the first and foundational stage of Wollheim’s contribution to our understanding of aesthetic perception, followed by an examination of his second-phase contribution (this progression is seen in connection with the abstractionist Hans Hofmann, a late Picasso self-portrait and Arnold Newman’s great photographic portrait of Picasso, and a Jasper Johns flag painting). Then in Section 5, we move through Wollheim’s third and most mature phase of his inquiry into representational seeing, now explicitly incorporating both the intentional constraints on interpreted content and the central position of what he calls “cognitive stock” in terms of the representational content having such stock enables us to recognize, circling back (after a consideration of a pair of Frank Auerbach’s paintings) to Wittgenstein. Finally, in Section 6, I ask if there might remain an important step not taken. And there also we will see there that throughout this discussion— perhaps despite the appearances along the way—I have been pursuing one specific articulation of an underlying analogy between art and language.

1 I begin with a seemingly odd—but as we will see interestingly instructive— asymmetry that resides within our verbal usage. To say that someone is “reading in” to a text is to suggest that the read-in content is false, misleading, prismatic, or distorted either by intent or by unwitting prejudice or presupposition. To use the same formulation but in visual terms—what they see in this, to see into the issue at hand—does not (or at least does not necessarily) carry those negative connotations of motivated misperception or of falsely superadded content. The question “What have you read into this?” is a challenge, a criticism, and the beginning of a correction-via-reduction back to legitimate content. The question “What do you see in this?”, if not always, is very often a solicitation, an invitation to say more fully what one sees, an exploration of the content of one’s perception. Why does our language, as a repository of wise subtlety, house this distinction? “Perception” is a word that carries within its grammar the possibility of correctness; one of its contrast terms, “misperception,” shows this. And this possibility presumes that there are criteria for the adjudication of perceptual correctness: directness is often found among, or functioning within, these criteria (for example, “I saw it with my own eyes”; “She was standing right in front of me”). Courts of law are filled with such statements, just as they are filled with attempts to demote such statements to the lower epistemic

Leonardo’s challenge


station of the indirect: “Did you see the defendant directly, or only in your rear-view mirror?” Trustworthiness in the content of the reported perception can thus be apportioned accordingly—that is, as we tend to think on this straightforwardly simple model of perceptual accuracy. The exposure of false testimony further reinforces this model, where a perceptual statement is repudiated by showing that the alleged witness was not present at the scene at the right moment and so could not have seen the act in question directly. This simple model can have some—if, as we shall see, very limited or basic—application in the arts: consider the example of the Viennese music critic whose career was ruined when his review of a concert (by a conductor and orchestra he knew well) that he secretly did not attend was published the following morning, but unbeknownst to him the conductor collapsed of a heart attack en route to the concert, which was cancelled at the last minute. So as a basic criterion for the veracity of perceptual reportage this kind of direct acquaintance has to be in place, and its absence is immediately and wholly epistemologically disqualifying. But aesthetic perception requires much more, layered atop this basic foundation; and as we will see, the meaning-content of this distinct usage of “more” is what Richard Wollheim did so much, throughout his work, to articulate. But “projection,” in discussions of interpretation, serves as the polar antithesis to “perception”: one might characterize this distinction in terms of oppositely pointing arrows, perception going from the object or action to the eye of the beholder, projection going from the eye of the beholder to the object or action. Projection, on this simple dichotomy, reaches out to, and then layers over, what is in the extreme case the blank screen we make of the actual object in question, where this process pretends to be perception. Projection disguises itself by falsely reversing its arrow. And so regarding trustworthiness: the extent to which we expose projection for what it is and remove it—peel it back, as it were, thus allowing the arrow to come to us, not from us—is precisely the extent to which (again according to this simple dichotomized model) we see the object in question aright. It is, as we say, to see it for what it is. But even at this early stage one wants to ask: what would it be to see an object, an action, a person, an artwork, just for what it (or he or she) is? What would seeing simpliciter be? Is it the content of direct perception as just discussed above? But then is there such a thing as contextinvariant directness—indeed directness itself simpliciter? And if we could find answers to these questions, would those answers provide a reliable basis for aesthetic perception and judgment, thus convincingly answering the threat of perceptually and interpretively reckless subjectivism? These questions, I believe, flowed as subcurrents beneath Wollheim’s entire intellectual project,1 coming to the surface in varying ways across the span of his work. And as he frequently indicated, he owed his background thought on this—the cognitive stock that prepared him for his excursions into these issues—to Wittgenstein.

120 Garry L. Hagberg

2 Wittgenstein begins section xi of part II (now rightly changed to “Philosophy of Psychology: A Fragment”) of his Philosophical Investigations with this remark: Two uses of the word “see”. The one: “What do you see there?”—“I see this” (and then a description, a drawing, a copy). The other “I see a likeness in these two faces”—let the man to whom I tell this be seeing the faces as clearly as I do myself. What is important is the categorial difference between the two ‘objects’ of sight. (Wittgenstein 2009a: section 111) Rightly understood, this has a power (in terms of the depth and reach of its implications) comparable to the opening statement of Beethoven’s Fifth Symphony: it may seem simple, but what it contains within is anything but. Given the intuitively plausible (and in some contexts practice-enmeshed) simple dichotomy just described above, we would expect subjectivity and objectivity to divide themselves along the same lines as projection and perception, or reading-in versus rightly seeing. So what Wittgenstein’s two viewers have in common would be more objective, more epistemologically trustworthy, than what separates them, i. e. the latter seeing the likeness, the former not. But this is quite evidently not the case: both see what is before them clearly, and the former draws or fully describes what he sees. The latter recognizes the veracity of that representation of perception (so far as that representation goes), but then she also sees the likeness, and while this is in a sense dependent upon her perception, it is nevertheless there to be seen—and the former does not see it. Hence the degree of veracity in reporting what is seen, what is actually visible, does not (here the long implications begin) correlate to what we might here call basic perception, i.e. what the two viewers have in common. Indeed, a fuller account of what is perceptually present will include—and not in a secondary, derivative, or more subjective and thus less epistemologically respectable way—what the second viewer sees. Similarity, likeness, is not arbitrary, whimsical, or merely a projected feature. So how might we encapsulate and sharpen the difference recorded in our language by the two uses of “see” in play here? Wittgenstein continues: The one man might make an accurate drawing of the two faces, and the other notice in the drawing the likeness which the former did not see. (Wittgenstein 2009a: section 112.) This is to say that, now refocusing the difference, the first viewer may be able to replicate everything that is visible as it stands in plain view before the two

Leonardo’s challenge


Figure 6.1 Paul Cézanne, Château Noir 1904.

and yet not see what the second viewer sees in that replication.2 So the accurate replication—although everything visible is there—still leaves out for the first viewer what is included in the perceptual content of the second viewer. The difference Wittgenstein is pursuing is illustrated in his example by the case of faces with a resemblance, but the “categorical difference” reaches throughout aesthetic perception. The first viewer could see Cézanne’s Château Noir of 1904, (Figure 6.1) and see it next to Picasso’s early Factory, Horta de Ebro of 1909 (Figure 6.2), and yet not see the likeness—here a likeness that shows how Cézanne opened the door to, or expanded representational possibility in order to make possible, cubism. (It is of interest that if one did not initially see the likeness, one could be shown Braque’s Viaduct at L’Estaque of 1908 (Figure 6.3) and then on looking back to the Cézanne and Picasso suddenly see the likeness.) By contrast, this meaning-determinative likeness-content is precisely what the second viewer does see. Everything visible lies before both of them—both see all the pigment on the canvas, yet the latter sees the extremely significant aesthetic likeness linking the two works and without which one foundationally important element of Cézanne’s contribution to modern painting remains indiscernible. While neither reads-in anything, what the second sees in the Cézanne is profoundly different from the first.

Figure 6.2 Pablo Picasso, Factory, Horta de Ebro 1909.

Figure 6.3 Georges Braque, Viaduct at L’Estaque 1908.

Leonardo’s challenge


Wittgenstein continues: I observe a face, and then suddenly notice its likeness to another, I see that it has not changed; and yet I see it differently. I call this experience “noticing an aspect”. (Wittgenstein 2009a: section 113) It has not changed, and yet I see it differently. One might hint at what Wollheim saw in Wittgenstein that was of such profound importance to our understanding of aesthetic perception by saying: to see is not to see. Or we might use the phrase “to just see, rather than really see.” Examples, as so often, can do much philosophical work here. Keeping Wittgenstein’s remark on faces in mind, consider Picasso’s late and haunted self-portrait of 1972 (Figure 6.4). Then look at Arnold Newman’s famous 1954 photographic portrait of Picasso (Figure 6.5). And then look back, with Wittgenstein’s phrase “and yet I see it differently” particularly in mind. And then—and, importantly, in terms of the work this example does—look anew, as though you had not seen either, looking at the photograph first, and then to the painting. With examples like this before us, it is again the paradoxical-sounding

Figure 6.4 Pablo Picasso, Self-Portrait 1972.

124 Garry L. Hagberg

Figure 6.5 Arnold Newman, photographic portrait of Picasso 1954.

formulation that seems to especially fit: the images have of course not changed—and yet they have. To say that this is not in any case real perception would be to make, by mere stipulation, an exclusion that in truth reveals only the extent to which we are in the grip of a simplifying picture. Of his much discussed (indeed, overworked) duck–rabbit, Wittgenstein writes: Had I replied “It’s a rabbit”, the ambiguity would have escaped me, and I would have been reporting my perception. (Wittgenstein 2009a: section 128) It is clearly not an ocular disorder of any kind that prevents this viewer—for us the first viewer—from seeing the duck, and thus from seeing that this is an ambiguous image, i.e. seeing it for what it is. (We will see Wollheim offer

Leonardo’s challenge


a criterion for the determination of correctly seeing it for what it is below.) What we have here, as Wittgenstein knew and Wollheim understood, is aesthetic perception in microcosm: one can have this second kind of perceptual content that is not reducible to projective content escape one’s perception, and yet what there is to see (in the first sense) lies plainly before one’s eyes. And, by contrast, as does the second viewer, one can see it all, while not being able to precisely point (in the way one can point to a brushstroke) to anything present that the first viewer does not see. But I want to suggest that there is a mistake, a powerful one, in saying (as many have done) that all there is to see lies plainly open to view, and here is our next step forward. In aesthetic perception, it does not lie plainly open to view. To say that it does is to implicitly suggest that the real perception is ocular, is the first sense of “see,” is physical—physical in both the object of sight and the visual apparatus with which we see it.3 This wrongly places the second sense of “see” into a subordinate position. One could be metaphor-blind: on hearing “No man is an island” one would reply, “Well, obviously!”; on hearing how Romeo describes Juliet one would reply, “That’s obviously false—she’s a person, not the gravitational center of the solar system.” One heard all of the words, one did not hear any words one did not know, yet one did not hear what was said. (This is the linguistic analogue of the rabbit-only viewer.) One could say here: well, one heard everything physically; metaphor-blindness is not an auditory disorder. What I will suggest below, in connection with Wollheim, is that there is such a thing as meaning-blindness in aesthetic perception, that it can be incremental ranging from slight to severe, and that it is not properly addressed by increased physical acuity or physical amplification. It is addressed, as we shall see, by aspect-revealing “cognitive stock,” as Wollheim identifies it. So if there is a mistake in saying that everything lies plainly open to view, what is that mistake, precisely? Because one can have content in the second sense escape one’s perception (aesthetic misunderstanding can often be explained in precisely these terms), we may well feel an inclination to describe the perceptual situation in terms of an additive model: we superadd the second-sense content to the first-sense content, and we have, as the perceptual sum, the full content. Wittgenstein, ever vigilant about the misleading power of underlying too simple and overly generalized philosophical pictures, writes: ‘Seeing as . . .’ is not part of perception. And therefore it is like seeing, and again not like seeing. (Wittgenstein 2009a: section 137) But this is easy to misunderstand, and the matter is becoming intricate. To say that a thing is like another thing is to implicitly say that it is not that second thing: a fork is not like a fork. The phenomenon under scrutiny here—what Wittgenstein too generically calls “seeing-as” (although he goes

126 Garry L. Hagberg to appropriately great pains to show the subtle differences within this category) is once again paradoxically put: “like seeing, and again not like seeing.” It is, I want to suggest, like seeing in that one can see an aspect or fail to do so; one can take it in or not. But it is unlike seeing in that the terminology we use to describe perceptual failures is borrowed from seeing (the metaphorical usages are like the literal ones) and is never the result of a literal blockage. However, to say that seeing-as is not part of perception is (and it is surprising that Wittgenstein does not qualify this assertion in a way that captures what must be its intended restricted scope) is indeed to almost invite misunderstanding. To say that seeing-as is not part of perception leads us to picture the larger perceptual situation as consisting of two parts: (1) the real (with connotations of objectivity, provability, and plain sensation-based presence) and (2) the fanciful, the speculative, the whimsically imaginative (with connotations of subjectivity, unprovability, and an ethereal and non-sensation-based kind of quasi-presence). This is the fundamental mistake to make in these waters, and this superimposed dichotomy falsifies, by template-forced misdescription, far more than it captures, both of human perception broadly understood and of aesthetic perception more narrowly.4

3 Rembrandt will help. We might consider his The Adoration of the Shepherds of 1646 (there are two paintings of this title in that year; I will refer primarily to the smaller of the two that is in the National Gallery, London (Figure 6.6), but will include a few comparisons to the larger in the Alte Pinakothek in Munich (Figure 6.7). In this compact masterpiece we see, as John Drury brilliantly describes it, Rembrandt’s fascination with “St John’s myth of the descent of the divine light” precisely because he was “a poet of the light and the dark” (Drury 1999: 66). Similarly, “St Luke’s story of peasants coming to the manger appealed to him as a dedicated recorder of ordinary folk, gathered at home or in the street” (ibid). In the larger slightly earlier (and ultimately very different— despite strong pigment-level similarities) painting (Figure 6.7), the light is, as Drury explains, very bright on the child to the point of being miraculous (the lamp as held by Joseph is far dimmer than the light it casts on the central subject, and the light it casts in all other directions is gentle and normal). In the smaller painting, however, Rembrandt “had some rather different religious ideas which he wanted to work out: ideas about light and dark, God and people, and how in each of these opposites there was a belongingtogether” (ibid). But more important (and leading to the most profound aspect of this painting) is the fact that, in this smaller painting, Rembrandt ingeniously “combines obscurity and dazzle so realistically that we have to train our eyes to see things, as if we were actually there” (Drury 1999: 67). Rembrandt remarkably places his viewers—us—inside the perceptual world of the subjects within the picture; we can try to view it from the outside, but,

Leonardo’s challenge


Figure 6.6 Rembrandt van Rijn, The Adoration of the Shepherds 1646 (there are two paintings of this title in that year; this is the smaller of the two, in the National Gallery, London).

as Drury perceptively observes, we cannot view all of the representational content of the painting as though it is all and equally plainly there. “The second cow, the third infant-in-arms and the face looking in through a window at the back do not disclose themselves instantly” (ibid). Or I would say, we look on the central scene just as does the painted viewer, almost wholly obscured by darkness, in the window opposite us, making out figures by concerted acts of focusing our attention in the areas where the light recedes into dark. This establishes the aspect of the special placement of the viewer (not at all the same in the larger painting, although the represented pictorial space is nearly identical) and the perceptual demands placed on the viewer as a way of giving the viewer a sense, however literally impossible, of active participation—indeed of communal belongingness—with the others in the painting. But this does not exhaust, but only begins to describe, what we see in this painting: Drury next captures perfectly a sense that a viewer might to a certain extent feel, might sense, without fully realizing what it is that

128 Garry L. Hagberg

Figure 6.7 Rembrandt van Rijn, The Adoration of the Shepherds 1646 (the larger, in the Alte Pinakothek, Munich).

Rembrandt is doing and how masterfully powerful a work of art this is for this special reason. Drury writes: “His next step was to work on the dark figure of the kneeling shepherd with his back to us. So this humble man was Rembrandt’s own way into his picture” (Drury 1999: 67–9). This self-created point of psychological entry for the artist into his own work is the kind of phenomenon to which Wollheim was especially sensitive, but it is what Drury identifies next that is of still greater importance both to the Wittgensteinian issues considered presently and Wollheim’s psychological aesthetics to follow. Drury continues (this is a profound insight into this painting): He can aptly be ours too. Because he is the nearest figure to us, because he is almost a silhouette and because he is the only person whose face

Leonardo’s challenge


we cannot see, we can, as it were, slip into him. Or, more circumspectly, we might advance into the space to his left, indicated by the line of his stick on the floor and the well lit broom at the picture’s edge. (ibid) To bring this aspect out more strongly, we can compare this feature to the larger painting: there is no such magnetic pull on the viewer there; indeed, although there is space to the kneeler’s right, the nearly closed circle of hands makes this thought (which we would not have by looking at that painting alone, so it is alien to this larger painting) seem an intrusion rather than a humanly warm response to a strongly welcoming invitation.5 And then one final feature: although Drury calls attention to the kneeler’s hands, and he rightly points out the subtle way in which Rembrandt depicts the fingers with pink edges (Figure 6.8), creating the pink-translucent look one gets by holding one’s hands up next to a candle in an otherwise dark room, he speaks of that light as coming from a hidden (miraculous) source behind the kneeler, also brightly illuminating the child. That is perfectly plausible, but one can also see it as a manifestation in the fingers of an inner light, a light source within the praying figure. Of the figures present, only he is praying, and only he is illuminated in this way; but, still, it is a light they all share, if each differently. To join the painting is to step into the atmosphere of, and share, that inner light, and the space reserved for this psychological entry is reserved, after all, immediately next to the kneeler. As Drury suggests, it can be miraculously bright light emanating from without; but I suggest it

Figure 6.8 Rembrandt van Rijn, The Adoration of the Shepherds 1646 (the smaller, detail of the fingers with pink edges).

130 Garry L. Hagberg can also be a warming spiritual glow radiating from within. Like the duck– rabbit, it cannot be seen as both at once, but the fingers can be seen, I think with equal plausibility, as one or the other. And so, with this example behind us, back to the perception/projection dichotomy and the attempt to generically demarcate what is, and is not, “part of perception.” If we make a brief catalogue of the elements operating within this reading (not reading-in; I’ll come to this shortly) of this painting, we find: (a) the precisely accurate description of Rembrandt’s painterly mission as “a poet of light and dark”; (b) some religious ideas that he was (as Drury importantly puts it) working out in the continuum of light to dark in this painting—those ideas being the presence of God among and within people, their suffusion by religiosity in the form of light, the way that God– man and light–dark oppositions or polarities belong together in synthesized unions and how they interpenetrate within the canvas; (c) the special way that Rembrandt places us so as to have to strain, squint, and focus our eyes to see what is there in the darkness, and to use this as a technique to put us inside the visual atmosphere of the painting; (d) the fact that not all representational content is equally there, but where distinctions of degrees of “thereness” do not correspond to the dichotomy, or fall on a continuum, between the directly perceptual in ocular terms on the one pole and imaginatively superadded projection on the other; (e) how perceptual demands placed on the viewer function as a powerful invitation to the point of requiring communal participation; (f) the hint—something like a visual whisper, if there can be such a thing—of an intention on Rembrandt’s part to instill in the viewer a gripping sense of inclusion and belonging, even (or perhaps especially?) where one cannot articulate it; (g) Rembrandt designing and exactingly executing his own portal into his painting’s depicted life, ingeniously connected to (h) his creating, and then in his unmatched way magnetizing, a space for us; and (i) his depiction in the fingers of miraculous light either bathing from without or shining from within. If we now ask which of these are contained within the category of direct perception, and which contained within the category of imagination-dependent projection, we will instantly and appropriately be more than a little puzzled—and that is precisely the philosophical work Rembrandt’s painting does here. No one doubts that “a poet of light and dark” is a perfect—and perfectly true— expression of Rembrandt’s distinctive idiom. But if asked to point to what shows that, we cannot point to it in the way that we point to the cow or the partly hidden face. But that difference is not equivalent to, nor does it reduce to, the difference between objective perceptual content and subjective imaginationadded content—this is true, and it is there. We do see Rembrandt working out ideas in this painting (made clearer by comparison to the larger painting), but to what exactly do we point for a verification of this creative fact? We can see synthesized oppositions—but are the oppositions in the painting in the way that the faces are? (One wants to say: what are you expecting in asking that?) We see how Rembrandt places us and our eyes—but that is a

Leonardo’s challenge 131 fact, although true of the painting, about us, and yet it is not subjective to us. Degrees of “thereness” are not, as mentioned, equivalent to degrees of perception or projection. We have a feeling of being swept-in, a participatory sense that we might initially (under the influence of the to-that-point unchallenged picture) loosely call projective, but on comparison to another work without this sense we will call it, unmistakably, objectively present (or better, we will just say, “Think about this for a moment, and then look”). We see an intention to instill a gripping sense of inclusion (again, Wollheim was particular good at discerning and elucidating this sort of phenomenon); if an aestheticphysicalist skeptic says, “Well, where is it? Point to it?”, we will explain the category mistake. (The fact that Gilbert Ryle could not point directly and singularly to the university did not mean that the university was not there.) We cannot point to Rembrandt’s portal by pointing only to the kneeling figure, but it is there as much as is pigment. Against the dichotomized categories, a work of art like this makes one feel one is condemned (if one tries for a time to stick with the categories and reduce all aesthetic features to one or the other without remainder or exclusion) to work on the proverbial Swiss watch with a jackhammer. And if one said that there are imaginative elements here, and that those are projective, one needs another comparison, I think of this kind: “Rembrandt is deliberately showing that religion, as the opiate of the masses, proves pathetically attractive to the downtrodden, who are more vulnerable to this better-world-beyond myth because of the unrelenting oppression they suffer as a consequence of the private ownership of the means of production.” This is reading-in; this is (anachronistic) projection. Seeing-in, as Wollheim developed the concept as the centerpiece of his entire aesthetic project, is an interestingly different matter from projection, and it is built on the foundations we have now seen Wittgenstein lay down.

4 In the course of introducing (Wollheim 1980b: 12–22) his conception of representational perception, Wollheim discusses the case of Hans Hofmann, the great New York abstractionist, instructing the students in his painting studio to place a black mark on a white canvas and then stand back and consider how the black stands on the white. (Hofmann used this principle and it extends throughout many color-relations in his paintings.) This captures the two senses of “see,” and, like the Rembrandt, it shows that the distinction between them does not correspond to a distinction between what is objectively present and what is fancifully projected. Here this is so for two interlinked reasons: Hofmann was obviously not asking his students to stand back and recall that, a moment before, they put the black on the surface of the white and so it was “on” in that simple physical sense (although it is, and this is the simpler or first-viewer sense of “see”). Hofmann wanted them to see that the black stood off the canvas and was forward, while the white receded. Wollheim accentuates and clarifies the point by changing the example

132 Garry L. Hagberg

Figure 6.9 Hans Hofmann, Equinox 1958.

to blue on a white canvas, where the blue recedes behind the white. (Of course, in the first viewer’s sense, it does not, and could not, unless it somehow punctures the canvas and falls or drips through.) Wollheim says at this point that this captures the nature of representational seeing and what it is “for something to have representational properties” (ibid: 15). Again, an example does the most powerful work in showing what is being said: see Hofmann’s Equinox of 1958 (Figure 6.9). For Wollheim, this species of representational seeing is anything but rare in human experience: “it would be little exaggeration to say that such seeing is co-extensive with our seeing of any physical object whose surface exhibits any substantial degree of differentiation” (ibid: 16). And it is here, in Wollheim’s first discussion of representational seeing, that he invokes the master’s instruction: In a famous passage in the Trattato Leonardo advises the aspirant painter to “quicken the spirit of invention” by looking at walls stained

Leonardo’s challenge 133 with damp or at stones of uneven colour, and find in them divine landscapes and battle scenes and strange figures in violent action. This passage has many applications both for the psychology and for the philosophy of art. Here I quote it for the testimony it provides to the pervasiveness of representational seeing. (ibid) At this stage, Wollheim, after removing a few objections to the idea of explaining representational seeing in terms of seeing-as, links representational seeing to the central role intention plays in the determination of representational content. In closing his early and foundational discussion of representational seeing, Wollheim points out the difference between representational properties and the physical properties of an object and how these can conflict: we attribute depth to the painting, while the canvas is flat; there is a recessive void in the middle of a fresco while the wall upon which it is painted is intact (ibid: 21). But this leaves us very close indeed to what I identified above as the dichotomy mistake: it appears that these are entirely separate modes of perception and that thus in aesthetic perception one is in either the one or the other (where we will then follow one or the other of the incompatible lines of description—flat or with depth, etc.). Wittgenstein referred, in the course of his discussion (Wittgenstein 2009a: section 140) to the lighting up of an aspect seeming half visual experience and half thought. The important word, for present purposes, is “seeming”: Wittgenstein was investigating the buried presuppositions that would lead us to see this matter as neatly categorized within the simple bifurcated categories—his extensive examples lift us out of this and into a conceptually clarifying complexity well beyond the reach of a falsely “clarifying” oversimplification. I think Wollheim, if he did not quite directly say it, saw this too, and he thus needed to find a way to capture the complexity of representational seeing at the heart of aesthetic perception without falsifying it with the dichotomy error. In his later essay, “Seeing-as, Seeing-in, and Pictorial Representation,” Wollheim writes: The nature of the perceptual genus of which the seeing appropriate to representations is a species is, I have said, harder to characterize, and the only suggestion I made in the original text—to the effect that representational seeing . . . , if not identical with, can be elucidated through, seeing-as—I now regard as wrong. The suggestion had much to recommend it. In addition to its immediate appeal it invoked a phenomenon of which, through the initiative of Wittgenstein, we seemed to be gaining good understanding. However, I now think that representational seeing should be understood as involving, and therefore best elucidated through, not seeing-as, but another phenomenon closely related to it, which I call

134 Garry L. Hagberg “seeing-in”. Where previously I would have said that representational seeing is a matter of seeing x (= the medium of representation) as y (= the object, or what is represented), I would say now that it is, for the same values of the variables, a matter of seeing y in x. (1980a: 213) To a cursory glance, this is not what it can seem, i.e. a small grammatical difference of minimal substance. Wollheim’s discussion is appropriately detailed, but to my mind the essence of the “as-versus-in” difference is this: seeing-as requires, and implicitly reserves room for, a perceptual oscillation between what we see and what we see-as. Or, to put it another way, we can always revert back to the seeing “beneath” the seeing-as, and we can always point to the specific features of what is seen that allow or make possible the seeing-as. However, in those cases where there are sustaining features of my seeing y in x, then seeing-in contrasts with seeing-as in that I can simultaneously be visually aware of the y that I see in x and the sustaining features of this perception. (ibid) And that is to say: “the seeing appropriate to representations permits simultaneous attention to what is represented and to the representation, to the object and to the medium” (ibid). Wollheim calls this his “twofold thesis,” and the way that I would put (very briefly) what it accomplishes is that it remains true to the character of representational perception while effecting a reunification such that the dichotomy error is avoided. The twofold thesis also captures the difference between seeing Picasso in person and seeing Newman’s photographic portrait of him. We do not see Picasso: we see Picasso in the medium of the photographic plate, of which we remain aware alongside or indissolubly interwoven with our awareness that this is Picasso. (And so, to look back for a moment, if one said, “I see this dark patch as Picasso’s forehead, I see that dark line as the outline of his finger,” etc., one could reply, “And what else can you see those as?”, where only puzzlement or disorientation would ensue. Seeing-in does not generate this problem.) Wollheim continues: It is a misfortune of Wittgenstein’s exposition of his argument that he chose as examples of alternating perception cases of alternating perception of representations: notably, the duck-rabbit drawing. For such cases introduce additional complexities, which can be the source of confusion. But the fundamental point in Wittgenstein’s argument, which remains, is that, when I see x as f, f permeates or mixes into perception: the concept does

Leonardo’s challenge


not stand outside the perception, expressing an opinion or conjecture on my part about x, and which the perception may be said to support to this or that degree. (ibid: 220) Thus while the “as” and the “in” are substantially different, with Wittgenstein rightly understood on this point, one can see how Wollheim is building on, and extending, rather than sharply departing from, Wittgenstein’s investigation of representational perception as we considered it above. In closing this second stage of his inquiry into aesthetic perception, Wollheim asks if there are aesthetic perceptions that include both seeing-as and seeing-in. In connection with this question, he considers the example of a Jasper Johns flag painting (Figure 6.10). These works, he rightly says, are posed as visually embedded philosophical problems: a number of people have seen this problem as a materialized form of the question “Are they representations of flags or are they flags—or does the distinction between represented and representation collapse here?” But the question Wollheim sees is “Do we see-as and see-in simultaneously in this kind of artistic case?” (Johns was a long-time reader of Wittgenstein.) His answer (I compress it here) is: there are elements of the perceptual experience of a Johns flag painting that, consistent with his larger account of seeing-in above, invite seeing the painting as a flag; but to the extent that we

Figure 6.10 Jasper Johns, Flag 1966.

136 Garry L. Hagberg maintain an awareness of the handled artifactual material within which that flag is manifest, we see the flag in the painting. But then also: there is no part of the painting that is not the flag; the fact that Johns paints the flag to the edge of the work promotes this philosophical tension (which Wollheim also rightly says we as viewers are meant to experience, not to resolve). This is thus deliberately unlike the experience of, say, seeing the flag as contained within the larger work in Delacroix’s Liberty Leading the People, and so the possibility of the two modes of seeing in our perception of the Johns seems real. But a lesson we learned from Wittgenstein comes back into play here: to see a fork is not to see a thing as a fork (against the perception-as-hypothesis psychologist’s picture), and so here the seeing the flag in the painting is not equivalent to, or interchangeable with, seeing the painting as a flag.6 As above, one may initially think these are small grammatical differences, but such differences are what allow us, as both Wittgenstein and Wollheim knew, to exactingly capture the phenomenology of the representational core of aesthetic perception.

5 We have already seen the value, and the meaning-determining power, of examples. We know that we can see Cézanne as both finding and initially wedging open the gates of the path that led to cubism, and what one sees is instructively indirect; he himself did not go in, but what one sees is how he opened the possibility, with Picasso and Braque then realizing and extending what, with their work in mind, we can see he saw.7 We can see the late Picasso as mortality-obsessed. We can see Rembrandt as the visual counterpart to a religious thinker. We can see Hofmann as the master of over-and-under, of foregrounding and backgrounding color relations. But how shall we now both sharpen and give fuller and more “exampled” meaning-content to Wollheim’s phrase “seeing-in”? Answering this will lead us to both Wollheim’s third and final stage of his investigation into representational seeing, and, closing the circle, back to Wittgenstein and the perception/projection dichotomy with which we began. It is in his magnum opus, Painting as an Art, that Wollheim marks the precise point at which seeing-in becomes representation. An encapsulation of the experience [W]hat is unique to seeing in is the kind of visual experience in which it manifests itself. For when (say) I see a face in a picture, the experience that I have has two aspects, which are distinct but inseparable. On the one hand, I recognize the face: on the other hand, I am visually aware of the surface of the picture. I call this all-important characteristic of the experience “twofoldness”. (1987: 101; see also 48–51)

Leonardo’s challenge 137 is followed by an observation concerning the priority of our capacity for seeing-in I say “when I see a face in a picture” by way of example. But a picture of a face is, of course, a representation, and my claim has been that seeing-in is prior to representation: prior to it, both logically and historically. Logically, in that I can see objects in things that neither are nor believed by me to be representations—such as clouds, or damp stains on walls, or the silhouettes of cast shadows: and historically, in that I am sure our remotest ancestors could do this before they thought of adorning their caves with the images of animals they hunted. (ibid) which is followed, with this preparation, by the dividing line and its attendant criterion of correct perception: Representation comes into being when someone—an artist, for short— marks a surface intending that a spectator should see something in it: say, a face. The difference that representation makes to the natural capacity of seeing-in is that it imposes on it a standard of correctness and incorrectness. For, if the artist succeeds in his intention so that a face can be seen in the surface, then the spectator sees the surface correctly if he sees a face in it: otherwise he sees it incorrectly. What a particular picture represents can now be defined in terms of seeing-in plus a standard of correctness, where this standard invokes the intentions of the artist in so far as they are fulfilled. (ibid) Leonardo would not have said, “That’s not there” in reply to a pupil in front of that cracked plaster wall who said, “I’ve been looking for a while, and now I see a Tuscan hillside, a meandering row of tall poplars, soft clouds above, and a gentle brook with a stream flowing out of the frame of the landscape.” (In such cases—and this is what Leonardo profoundly saw— there is no line separating reading-in from seeing-in.) But if Leonardo had said, “Now paint it,” and the student had done so, it would have been right to say, “Yes, well done, that’s all there.” Viewers would then see-in all those landscape elements, and they would see those elements aright in correspondence to the painter’s intentional content. Wollheim is able to concede that either side of the twofold experience can evaporate as a result of thoroughgoing focus upon the other—we can see just the cracked wall (and if we continue to see only that, we are the visual analogues of the metaphor-blind person above) or we can, by letting our imagination take over the experience, see only the figures (as he notes, trompe-l’oeil paintings try to do this to us). There is the moment of letting go of seeing-in,

138 Garry L. Hagberg

Figure 6.11 William Harnett, A Study Table 1882.

when we see ourselves (this is the moment of fascination in this kind of painting) seeing the thing and not the thing in the painting. Consider in this light William Harnett’s A Study Table of 1882 (Figure 6.11). But this monodimensional visual experience will, as he says in a way that I think is true to our experience in front of paintings, be fleeting; “Seeing-in will probably reassert itself: such is its pull” (Wollheim 1987: 47); that is the moment when we reclaim seeing-in, recalling that it is (parallel to what we whisper to frightened children at the movies: “It’s only a movie!”) only a painting. And then we can fleetingly (and, like Proust,8 playfully and amusingly) overrule the intention of the artist: to my companion I can mock-willfully say of a painting that reminds me of my grandfather, “No! That is not Winston Churchill, it’s my grandfather!”, but I do so knowing full well that this is severe reading-in and only a jocular, deliberate override. Intentional content, like twofoldness, fights back; either side alone is informatively perceptually unsustainable. It is also here that Wollheim makes the argument that most abstract paintings are still representational, owing to the seeing-in of spatial relations (recall the Hofmann painting). This is important to an understanding of his third stage because it casts in highest relief the fact that he is giving enormous power to the phenomenology of this species of representational seeing9 in expanding what is admissible as representational content: the fact that abstract artists do not depict objects or persons or landscapes is not sufficient reason to conclude (as many have done) that representational seeing is thereby rendered impossible. Indeed, the representation of three-dimensional space not only invites, but requires, this species of perception.

Leonardo’s challenge


But the fundamental achievement of Wollheim’s third phase of his investigation is, as we have glimpsed above, (a) the clarification of the role of intentional content on the part of the artist as it informs what should be seen-in, (b) the necessity of the recognition of that intentional content on the part of the viewer in seeing depicted representational content, and (c) the circumscription of correct seeing-in as determined by the correspondence of (b) to (a). There is an observation Wollheim makes here that serves to powerfully remind us of the great importance we actually give to the artist’s intentional determinations of representational content and its subsequent correct identification (despite what one might think under the influence of some varyingly fashionable over-general theories of representation): he notes that even if the identity of a person depicted in an ancient portrait is long lost to humanity, we still know that it is a portrait of a particular person—that is, it has not thereby devolved into a non-specific portrait. It is just, rather, than we know we will never know the intentional content that it does, in permanently undeniable fact, have.10 The artist, like that work’s contemporary viewers who could recognize the identity of the person depicted, is many-centuries gone, but the work still has its (now undeterminable) intentional content. And then there are in some cases gradations of intentional recovery: we initially see a person, and then later a woman, in the stone; then still later we may learn, through archeology, that it is a sculpture of a queen, but not know which one. After that we discover by dating and location that it is one of a possible three queens; and so forth. There is a truth toward which we are progressing and that truth is intentional in nature. But like some remarks of Wittgenstein that we considered above, this— for strikingly similar reasons (i.e. the drive for schematic oversimplification)—is easy to misunderstand. The easiest mistake to make is to regard Wollheim as a narrow (or what we might call “one-to-one”) intentionalist, where the intentional content is pictured, simplistically, as a fully articulated mental predecessor of the artifact to follow; the success of the realization would then be determined by the fidelity—the one-to-one matching—of matter to mind. Wollheim holds no such view; the one-to-one picture does as much damage to any nuanced understanding of real intentional11 content as does the picture of simple mental or ghostly-prototype mental-object predecessors of intentional content in language. Just as we often work out in language (in speaking or in writing) what it is we wanted to say, to capture, to express, to convey, to record, so in the creation of art the painter works as a shaper of interactive materials and not as an assembler of individual static parts with predetermined locations. The next-easiest mistake to make is to regard Wollheim as an all inclusive intentionalist, i.e. one who claims that anything in the life and mind of the artist is or can be relevant to the interpretation of a painting—what we see in it can be anything we associate from the artist’s life. But this is too projective and thus insufficiently supported; Wollheim defends the “thoughts, beliefs, experiences, emotions, commitments that the broader understanding of artist’s intention bring in

140 Garry L. Hagberg train” (Wollheim 1987: 52), but in a way that is distinctively disciplined. And the fundamental point here, central to his third-stage achievement in understanding representation, is that this discipline will be exemplified in our taking the middle way between narrow one-to-one intentionalism on the one hand and a too broad free-for-all of projectionist biographical or culture-wide associations on the other.12 And the criterion for correctness, accuracy, and interpretive persuasiveness is what we are able to convincingly justify or rationally support as the representational content that we see in the work. (This, I want to add, stands parallel to the ways in which we justify the interpretation of a text and what we can see in it in a responsible, disciplined, and acute way.) We have a developed vocabulary with endless phrases that mark stations along the continuum from brute and limited perception to too fanciful projection: a few suggestive examples moving from near one end of the continuum to the other would include phrases such as, “Yes, but beyond those rectangles, one can also see . . . ”; “Ah yes, but the dog is a deliberate symbol of fidelity; this is allegorical painting”; “Yes, a strong, sharply defined, and formally clear colonnade, but also an allusion, like an intertextual reference in literature, to Greek classicism”; “Yes, it does make functional sense, but this cathedral is also a cross made to be seen from above”; “As a decisive break with tradition and any discernible precedent, this is meant to seem culturally amnesiac”; “Perhaps, but I’m not sure Caliban in this painting of Shakespeare’s The Tempest was intended as a representation of the economically oppressed”; “Well, yes, but Goya did not read Freud”; “Alright, but if this is nothing but a random play of arbitrary signifiers, then everything is”; and so on. The strength of Wollheim’s via media is in the fact that he does not attempt to falsely reduce all distinctions made along this continuum to a single simply applicable litmus test; we have to use judgment and bring into play the relevant background and knowledge. And, with Wollheim working well within the tradition of connoisseurship,13 that background knowledge will be cultivated over time—indeed across a lifetime of serious aesthetic engagement. It is precisely here that his fully developed conception (I have mentioned it above) of “cognitive stock” comes in, and the cultivated possession of this is what he means when he refers (as he does frequently throughout his writing) to “a suitably informed and sensitive spectator” (Wollheim 1987: 77). It is cognitive stock that, as an intellectual-imaginative precondition, allows the second viewer to see—and consistent with all that we have seen above, not to merely project—what he sees. Without it, the first viewer remains, like the metaphor-blind speaker above, aesthetically meaning-blind. Wollheim considers the criterion for the admissibility of items of cognitive stock (for short, “information,” as he uses the term—although I think there are problems with that14) that first suggests itself to virtually anyone, i.e. if we are able to see the feature or aspect in the picture just by looking at what the item of information makes visible, then it is legitimate. But the interesting problem, one that complicates the picture for Wollheim just as it did for

Leonardo’s challenge 141 Wittgenstein, is that one then has to immediately ask, “but seen just by looking in connection with what other cognitive stock?” And there we must answer either the too narrow or infinite-regress-generating “in connection with any information that we can see directly in the picture by looking at it,” or the too wide, indeed interpretively reckless “in connection with any information whatsoever.” Or we could try to delimit acceptable information by claiming that only that information describing aesthetic features internal to the painting is legitimate, but Wollheim as quickly mentions Michael Baxandall’s famous example of the very high market price of lapis lazuli (the most expensive pigment), its use in painting the cloak of St Francis in Sasetta’s St Francis Giving his Cloak to a Poor Knight (1437–44), and the fact that, as a contemporary viewer would have known of this material’s monetary value (cognitive stock that Sasetta presupposed), this item of information would have molded “our perception of the picture,” where this information “enhances the liberality” and “ensures the grandeur of the saint’s gesture” (Wollheim 1987: 91). So no neat and clean line separating kinds of information seems convincingly available; as Wollheim says, prima facie irrelevance can in truth be information that we vitally need to grasp the content of the painting. And so Wollheim writes: Indeed there seems to be only one limitation that should be placed upon what information can be drafted into the spectator’s cognitive stock. It relates, not to the source from which the information derives, nor to its content, but to the use to which it is put. The information must be such that by drawing upon it a spectator is enabled to experience some part of the content of the picture which otherwise he would have been likely to overlook. (ibid) But then this too is not quite as simple as it may seem—and there is a clear line, another one, that Wollheim wants to draw. If a person looks at a painting considers it without knowing its attribution, then is told it is a Rembrandt, and then says he now sees it as a Rembrandt, “we have every reason to distrust him” (ibid: 93); what this spectator says will change, while what he actually sees will not. Wollheim reminds us here of children’s visual puzzles with figures—animals, fish, people, objects—hidden in the drawing. We may get the list of hidden figures, or we may get the number of hidden figures, but the puzzle is not solved until, taking surprised delight in each discovery, we see all of them in the drawing. The empty, language-only version of this would be to read out the list of each hidden figure, saying confidently that we now know they are in there. But of any such hidden figure, if we do not see it, we cannot, with the same sense and the same authority that legitimates the cognitive stock that led to that new vision, say it. Did we see all of the drawing at the beginning? Yes—no part of it was concealed under a magazine cover, etc. Did we see all of the drawing at the beginning? No—we

142 Garry L. Hagberg did not see any of the ten hidden figures (the fish, the old man, the camel, the rabbit, etc.). Wittgenstein writes: I meet someone whom I have not seen for years; I see him clearly, but fail to recognize him. Suddenly I recognize, I see his former face in the altered one. I believe that I would portray him differently now if I could paint. (Wittgenstein 2009a: section 143) This is wholly unlike the circumstance, parallel to Wollheim’s “reason to distrust” case, that may masquerade for this one, i.e. the circumstance in which we meet the person, are then told he is that same person from before, and then say he is that person from before, but do not see him differently— we do not see the younger face in the older one standing before us. What we know does not transform what we see. We would paint him as we see him, not as we (if we truly saw the younger face we once knew in the older face before us) see him now. As if working in concert with Wollheim’s discussion (in truth of course Wollheim has insightfully put to new work much from Wittgenstein’s investigation), Wittgenstein continues: Imagine the duck-rabbit hidden under a tangle of lines. Now I suddenly notice it in the picture, and notice it simply as the head of a rabbit. At some later time, I look at the same picture and notice the same outline, but see it as a duck, without necessarily realizing that it was the same outline both times. If I later see the aspect change—can I say that the duck and rabbit aspects are now seen quite differently from when I recognized them separately in the tangle of lines? No. But the change produces a surprise not produced by the recognition. (Wittgenstein 2009a: section 152) Like the children’s visual puzzle, we can (a) see the rabbit in the tangle of lines, (b) later see the duck in that same tangle, and (c) later still recognize that it was the very same line in which we saw the duck and the rabbit. The surprise of (c) is produced by seeing, by recognizing (meaning here: re-cognizing), that in (a) we could also have seen (b) and vice versa but that we did not; our first viewer sees (a), and later (b), but not (c). Our second viewer here again sees it all—including the intentional content, that the artist-draftsman meant to produce an ambiguous line-drawing in which a viewer could both see two things and recognize that it was so intended, where the realization upon seeing that it was so intended is the cognitive stock that enables us to see its true identity.15 In Wittgenstein and in Wollheim, knowledge does not merely inflect, it rather transforms, what we see. (If Wittgenstein had drawn children’s visual puzzles, they would have had the duck–rabbit, the old lady–young lady, the two faces or one vase, etc.,

Leonardo’s challenge


with the list of hidden items exactly twice as long as the number of hidden line drawings that are required therein to depict them.) But although these cases capture much, they are children’s games. What about real art, now with these particular perceptual issues in mind? Bringing the three stages together, to my mind the best representation of what we might now encapsulate as Wollheim’s “twofoldness with recognized intentional content plus perceptually legitimated cognitive-stock” view—indeed in front of these paintings one has a hard time believing that they were not painted as Wollheim sat alongside and wrote with the painter reading what was written before continuing—is found in Frank Auerbach’s work. And Auerbach’s pair of deeply powerful paintings, Study after Titian I (Figure 6.12) and Study after Titian II (Figure 6.13) (along with a related drawing and its source, we will see shortly) brings together and integrates all of the perceptual themes under discussion here. Perhaps the first thing we notice about Auerbach’s work is the thickness of the oil paint on the canvas; we see swirls of deep oil paint almost heaped upon the supporting canvas, a canvas which we know to be there but never see as a flat surface carrying pigment upon it. We feel that it would take excavation, and massive damage to the work (and not merely abrasion, as it would, for example, on an eighteenth-century English portrait), to get down to it. So at once, we are seeing the one side of the twofold experience powerfully and directly. And then, particularly in the first painting, we see the outlines of the human figures in the oil, and we see that they are in motion.

Figure 6.12 Frank Auerbach, Study after Titian I 1965.

144 Garry L. Hagberg

Figure 6.13 Frank Auerbach, Study after Titian II 1965.

This allows us to make out an interaction between the two figures, one that conveys a sense of violence or attack or an overpowering against resistance. We make out facial features, lines of the two bodies suggesting male and female, and we see all this in the swirl of thick paint—always perfectly aware that it is thick paint in which we see these things. It places us perfectly at Wollheim’s twofold, or bifocal, point. If we then look to the second painting, we see at a glance a more abstract approach to the representational content, and yet, having seen the first painting (and so the first picture becomes cognitive stock for the second), we see the same male and female figures, in the same antagonistic or violence-conveying relation, as we did in the first—and yet we see that they are different. (And so here again—I will come back to this—we see that a paradoxical formulation seems to fit: they are the same, and they are different.) It is important to recognize here that, had we not seen the first painting first, we would not have been able to recognize the central figures in the way that we do—it would take more searching, a moment longer, to find them in the paint. If we then compare the two, looking back and forth, we see differences of brushstroke, differences in the application of paint and its literal depth, differences of color, of hue, of shade, and of the exact placement of lines and alterations in those lines’ continuity and resolution within the larger swirl. And then one exceptionally

Leonardo’s challenge


powerful feature can suddenly strike us, and particularly so in the second painting (which we are enabled to see more clearly by the back-and-forth comparison): we see that in the forearm of the male figure there is a deep gash, a violent sideways slash in the paint that the literal depth of paint makes possible. Suppose we now look away from Auerbach’s canvas to its model, its predecessor, its artistic progenitor: Titian’s Tarquin and Lucretia of 1571 (Figure 6.14). We see here the initiating moment of the rape of Lucretia by Sextus Tarquinias, a sixth-century-bc Roman Prince.16 We see in this paint (actual pigment of which we are, in the ratio-balance between the two sides of the twofold experience, far less directly aware) the threatening and raised knife blade, we see the terror on Lucretia’s face, we see her protesting arms, we see the forced intrusion of Tarquin’s right knee, we see the tight grasp of her right arm, we see the sheet falling down the leg. We see terror-inducing violent brutality in the act. Now, we look back to the second of Auerbach’s

Figure 6.14 Titian, Tarquin and Lucretia 1571.

146 Garry L. Hagberg works and we see what we might have taken as swirls of abstraction surrounding the outlines of figures as what they really are and what we are intended to see in them—depictions of all those things as organized together into a coherent whole. And then it can strike us: the gash, the violent sideways slash, is itself representation, through a kind of painterly re-enactment, of the act. It is still an arm; it is flesh represented. But it is also a violent, forced, literally penetrating slash into the paint.17 We thus see Auerbach standing in relation to his work as Tarquin stands to Lucretia—but we then realize that it is after all Tarquin’s arm being gashed, Tarquin being violently attacked, within its representation. And so we also suddenly see in this: an eye for an eye—transmuted into paint. By seeing these paintings together, by looking back and forth, we see complexities of representational content that are indisputably in the paintings, that are there to be seen, and yet, while standing before them and just attending to what we might, prior to Wittgenstein’s and Wollheim’s work, regard as content that is in a brute-factual sense truly there, we see, as we would say in retrospect, almost nothing. Wittgenstein wrote (employing in his second remark here yet another paradoxical formulation): “Ah, now I see this”, I might say (pointing to another picture, for example). This has the form of a report of a new perception. The expression of a change of aspect is an expression of a new perception and, at the same time, an expression of an unchanged perception. (Wittgenstein 2009a: section 130) I said that Auerbach’s work brings all the themes in play into focus within the experience of his work. His two Titian studies interweave much, but not all, of this. So before moving to the concluding section of this chapter, let us briefly consider Auerbach’s remarkable Sketch from Titian’s ‘Bacchus and Ariadne’ of 1970–1 (Figure 6.15). At first untutored glance, we can see a figure, center-left, appearing in a stance resembling that of a pitcher who has just thrown a baseball, and two figures, front and lower right, appearing in the stance, perhaps, of runners in close formation. And we see a swirl of lines surrounding this of no determinate representational suggestiveness. But if here, too, we look to the progenitor, Titian’s Bacchus and Ariadne of 1520–3 (Figure 6.16), and then back to the Auerbach sketch, a rapid series of representational identifications follows, and these occur following a very similar perceptual path that children take with their hidden-figure illustrations—except that here, in the intertextual artistic case, the texts are both visual to start with and not names of hidden items leading us to look for drawings of them. Our seeing-in, in the progenitor, is the cognitive stock carrying the specific intentional content required to see (and therein justify beyond the reach of mere projection) the content in the other. I have asked above: did we see the

Figure 6.15 Frank Auerbach, Sketch from Titian’s ‘Bacchus and Ariadne’ 1970–1.

Figure 6.16 Titian, Bacchus and Ariadne 1520–3.

148 Garry L. Hagberg lines initially? And here also, of course, the answer is yes. But in this particular case, the distance between what we are initially able to recognize, and what we subsequently identity, is by far the greatest. So my question now—in our final return to the perception/projection dichotomy—is: does this greater distance correspond to a greater distance from objective or legitimated perception—does it, because of this greater distance, move along our initial continuum ranging from perception to projection? The striking and reorienting answer is no. Only a simplifying picture would make us say that, and never the facts of aesthetic perception as we derive them from their positions in situ here. What we see here, as in the cases considered above, is content in the work. And what this work-to-work comparative perceptual experience discloses is real content (no question reminiscent of the fully projecting Marxist interpreter above arises, or can be made to arise, here). This content is beyond what is discernible to the retina at first glance, but yet it is content that is there in a way far more secure than projection alone. This is the intersection, or the overlap, or really the interweaving of perception and projection: it is in its way dependent upon both but reducible to neither. And I want to add: there is a strong sense in which Auerbach’s Sketch from Titian’s ‘Bacchus and Ariadne’, with the ground we have now covered behind us, itself serves well as a representation of the very concept of representational seeing. Wittgenstein writes: The concept of seeing makes a tangled impression. Well, that’s how it is. — I look at the landscape; my gaze wanders over it, I see all sorts of distinct and indistinct movement; this impresses itself sharply on me, that very hazily. How completely piecemeal what we see can appear! And now look at all that can be meant by “description of what is seen”! There is not one genuine, proper case of such description—the rest just being unclear, awaiting clarification, or simply to be swept aside as rubbish. (Wittgenstein 2009a: section 160) If we think that there will be one single determinate, bounded, exacting, final, and uniquely true description of what we see—i.e. the fixed content of our perception, neatly captured in uniquely fitting words—and if we correspondingly think that this description will on closer analysis always be of either perception or projection simpliciter, we let ourselves become the ventriloquial dummies of a blinding dichotomy, or, in Wittgenstein’s special sense of the term, of an underlying philosophical picture. We began with the distinction we make between what we see in a thing (where we think of this as probably or at least potentially true to what it actually is) and what we read into a thing (where we think this is always false to what it actually is), but, in light of all the clarifying complexities we have now considered, we would do well to not expect the world of representational perception to fall without remainder and without resistance into these categories. I mentioned above that I would return to the issue of the paradoxical formulations that repeatedly arise in these matters: like seeing and yet not like seeing; what we

Leonardo’s challenge 149 see has changed and yet has not changed; is part of perception and is not part of perception; is an expression of a new perception and at the same time an expression of an unchanged perception; the spatial relations between color patches are there and yet are not there; we see a flag and do not see a flag. I suggest that the arrival at repeated and seemingly unavoidable paradoxical formulations is a manifestation of an underlying residual allegiance to the ultimate rigidity of the perception/projection dichotomy—indeed it seems that Wittgenstein himself did not fully escape this, despite his intricate dichotomy-undercutting observations and his last remark above concerning the tangled impression and “how it is.” Indeed, if we now look back over all the examples discussed above, consider what we see in them, consider how we see in them, consider what we see across them (from one work to another and back), and then try to categorize them all into one or the other (or determine which way the arrow is pointing, as discussed at the outset), one can only get the impression that the categories with which we are working are too blunt for the task at hand, or, again, the hopeless feeling of trying to work on the proverbial Swiss watch with a jackhammer. (Think of what motivates the phrase “Well, it’s like projection, and yet not like projection,” i.e. an allegiance to a pair of categories too blunt to capture the subtle phenomena they are falsely expected to neatly contain.) In any case, for Wollheim, all recognition of representational content in a painting is of the kind we have just seen dramatized in Auerbach’s work; it is just that cases such as Auerbach’s bring out this essential feature of our aesthetic perception (in this case both literally and figuratively) in higher relief. (Francis Bacon and Lucian Freud are also, in this “higher-relief” respect, “Wollheimian” painters; we see in their work the convergence, into one doubleaspected entity, of heavy corporeal flesh and heavy material paint). We see the content in the pigment; we see the richer content of the second viewer in the material the first viewer sees. One could express the matter in terms of a final reunification after an initial bifurcation: for Wollheim, Wittgenstein’s two uses of “see” in the end—with the mind and the eyes both fully engaged in front of a painting—re-converge in a unitary yet twofold act of perception. Or: he stands at the intersection—looking both ways.

6 For this final section I would now like to ask: while Wollheim, building out from the foundations Wittgenstein set down in his remarks on aspectperception, took a major step in the right direction, is there an element remaining that is essential to our understanding of aesthetic perception still to be uncovered in connection with Wittgenstein’s work and its broader significance? I believe that there is a step remaining, and an important one. (My intention here is not to fully do the work that this step would require— that would take a book—but to demarcate its location and convey a sense of its content.)

150 Garry L. Hagberg One might think that in all of the foregoing, as we have been discussing aesthetic perception and modes of seeing, we have departed quite widely from the theme (mentioned as a central theme of this chapter at the outset) of the analogy between language and art. In truth, we have not done so for a moment. It has proven easy—too easy—for interpreters to regard Wittgenstein’s philosophy of language as one thing and his remarks on perception in (what was) part II of Philosophical Investigations quite another. But they are intertwined in indissoluble ways. Throughout his writings on the philosophy of psychology and perception we see countless references to words, to what we say, to the understanding of language, and to the description of experience. Let one example, for now, stand for many: “And is it really a different impression?”—In order to answer this, I’d like to ask myself whether there is really something different there in me. But how can I ascertain this?—I describe what I see differently. (Wittgenstein 2009a: section 172) That is, given a certain picture of visual experience as inner content only contingently connected to language, we would like to introspect upon that content and use introspective content as the criterion for sameness or difference of impression. But what do we actually do, he asks? We describe differently. In all the cases or paintings considered above, one might ask: how would we isolate and specify the visual experience we have, in a way wholly prior to and independent of language? The disorientation one feels in facing this question itself, I want to suggest, argues for an essential inseparability (and it argues for a relation between words and the world that is far more intimate, and intertwined, than any ostensive naming picture could accommodate).18 Words, in truth, are inseparable from the representational experiences both Wittgenstein and Wollheim are investigating. Those experiences take place in and among, and not prior to, our life in language. But that itself is not the point (nor could it be convincingly made so briefly), not the remaining step; it is only prefatory to the remaining step. Consider Wittgenstein’s observation: Take as an example the aspects of a triangle. This triangle

can be seen as a triangular hole, as a solid, as a geometrical drawing; as standing on its base, as hanging from its apex; as a mountain, as a wedge, as an arrow pointer, as an overturned object which is meant, for example, to stand on the shorter side of the right angle, as a half parallelogram, and as various other things. (Wittgenstein 2009a: section 162)

Leonardo’s challenge


We can see the figure as a hand-drawn triangle, but when we do (and see only that), we will ask a question concerning what it is, what it stands for, what it represents, what it means. Or (rightly connecting pragmatics with a question of meaning): “What’s that doing there?” What, indeed, is it doing?19 That is: in what language-game does it have a role to play; what is it being used for; what—in briefest scope—is its point? And until we grasp that, we do not know, or see, what it is. To grasp this, to see this, is to bring out one aspect and foreground it among possible but now irrelevant others; or it is to see in to the geometrical figure so that we see, not what we project onto it, but what it is (where this content is determined by the role, the point, the use)—even though, to connect back to the beginning of this discussion, what makes it what it is is not brutely perceivable, not visible in Wittgenstein’s first sense of “see.” Wittgenstein’s triangle, out there in life, works like that (and it works here as an illustration of what it is to work like that), showing in microcosm how so much of representational seeing works, how the role or use or point of a thing gives the sense of what it is. So with these reflections in mind (as cognitive stock), what precisely is the important step not taken? There has been, in philosophy’s history, a linguistic mythology born of scientism: as mentioned at the outset of this chapter, that words carry their contextually invariant meanings within their own shells, and thus the meaning of a sentence will be the added sum of its words, the meaning of a paragraph its added sum, and so forth. The idea was that this intellectual project, when completed, would fulfill a dream of exhaustive atomistic analysis. This is not the place20 to discuss this view other than to merely mention it as the contrasting position to Wittgenstein’s mature view of linguistic meaning— one much more sophisticated, much more encompassing, and much more— well—real (where “real” means an account of language in which we can see ourselves and our actual practices). Wittgenstein, having worked into, through, out of, and beyond the atomistic picture, sees in language something very different—and everything discussed here under the heading of aesthetic perception is central to that vision. Stanley Cavell expressed the point (in connection with his own Wittgenstein-inspired efforts) in this way: A striking idea among Wittgenstein’s remarks about seeing aspects is his saying that the importance of the concept lies in its connection with experiencing the meaning of a word and with our attachment to our words. This is an idea I emphasize in The Claim of Reason about Wittgenstein’s discussion of seeing-as, in my speaking about our relation to words as an allegory of our relation to others. (Cavell 2010: 85) “Its connection with experiencing the meaning of a word”: how do we see a word, a phrase, a remark, a statement, an assertion, a declaration? What do we see in it? And again back to the beginning: what do we see in (not read in)

152 Garry L. Hagberg the text? What was a person doing with those words? What are we doing with them? What do we know of the site, the circumstance, of their words’ work? The answers to all these questions will be site-specific, and they will require—this is what Wittgenstein saw and Wollheim (for all that he did investigate and clarify) did not pursue—precisely the kind of attention, sensory engagement, intellectual and imaginative engagement, multiple aspect-perception, seeing-as, and seeing-in, all in layered and interwoven ways,21 exactly as we have seen these modes of perception interact in our perception of art in all of the cases above. And then what of Cavell’s phrases “our attachment to our words” and “our relation to words as an allegory of our relation to others”? When we look at Cézanne’s The Bathers (also called The Large Bathers) of 1898–1905 (in the Philadelphia Museum of Art) (Figure 6.17), we see in the oil, in the result of intentional brushwork, the depiction of the women, the landscape, the water, the distance, the sky, the clouds, and the trees; all of that is consistent with what we have now considered. But if we shift focus to a larger scale, we see in those other representations a portrait of a woman’s face: the trees become strands of hair curling around the face to each side, foliage becoming eyes, eyelashes, and eyebrows, the distant tallest tree becoming the line of a small nose, the nearest water line becoming the lower chin line, lips (if more ambiguous) just above, the tallest

Figure 6.17 Paul Cézanne, The Bathers 1898–1905 (also called The Large Bathers, in the Philadelphia Museum of Art).

Leonardo’s challenge


nearer tree on the other side of the water becoming the line of the woman’s left cheek. Or do we see this? Are we, as human animals, hardwired to see faces (as potential threats on Darwinian grounds), and so cannot help but project alien content onto Cézanne’s work (in the way, for example, that one can see little surprised, round-mouthed double-stacked faces in grounded American electrical wall outlets—we can see them, humorously, as that, but we of course know they are not intentionally depicted there22). Perhaps: but when we compare this painting to another version of a similar name, Bathers 1900–5 (in the Art Institute of Chicago) (Figure 6.18), we see a larger scale similar portrait, but with the face now saddened, now downcast. Does this strengthen the belief that Cézanne produced double-paintings, or does it strengthen the belief that we humans with our evolutionary history are hardwired to see faces? And if we then consider a third Bathers, of 1894–1905 (in the National Gallery, London) (Figure 6.19), given the cognitive stock we now have we can in a sense feel our eyes straining to see a woman’s face but we at the same time feel this is forced projection—she is not really there. So does this third painting decrease the plausibility of the double-painting hypothesis? With these paintings, we are asking: what are those lines really doing there? Do they mean what they can appear to mean, or is this merely para-representational pseudo content? Here we want immediately to know (and this is information we find we cannot have) about the intentional

Figure 6.18 Paul Cézanne, Bathers 1900–5 (in the Art Institute of Chicago).

154 Garry L. Hagberg

Figure 6.19 Paul Cézanne, Bathers 1894–1905 (in the National Gallery, London).

content that may or may not make it right for us to see a large-scale portrait hidden in a smaller-scale scene. We can have the visual experience, but is it legitimated according to the criteria Wollheim has articulated? What, indeed, was Cézanne’s cognitive stock here? Or is this emergent aspect, this seeing-in, just plainly untrue of the work? To adapt Cavell’s phrases about words, what was Cézanne’s attachment to his lines here—and, more pointedly, what was his relation to his lines as an allegory of his relation to others or to another? If she is there, who was she? Standing before this great painting in Philadelphia, we might say, “I would like to know what was in his heart”; but this means “I would like to know whether this was a half-concealed expression of his soul or a representational coincidence,” to be answered only by gaining a grasp of his relation to his lines, his brushstrokes. We have words, phrases, sentences, declarations, things said in life, that can remain open to interpretation in precisely the same way; and it is precisely the same criteria to which we appeal (or try to or in some cases fervently wish we could do so) in settling questions of content; and the way we perceive them, with all attendant complexity of the kind Wittgenstein uncovered, is the same across the worlds of visual and linguistic meaning. In closing, there remains one final observation to be made about the ubiquitous presence of seeing-in and its central facilitation of the interactions between visual and verbal meaning. In a beautiful passage of Painting as an Art, Wollheim, in discussing the sources of visual delight, quotes Proust, who wrote: If, in looking at a Chardin, you can say to yourself: This is intimate, this is congenial, this is full of life like a kitchen, then you will be able to say

Leonardo’s challenge 155 to yourself, walking round a kitchen: This is strange, this is grand, this is beautiful like a Chardin. (Wollheim 1987: 98) Wollheim discusses this on pages featuring Chardin’s painting of a gentle, soft, perfect, and quiet beauty, The White Tablecloth of 1727, juxtaposed with Eugene Atget’s quietly atmospheric photograph La Nappe (“The Tablecloth,” early twentieth century). The interconnecting experience of reading these pages is, I want to suggest, (even) more important than what they explicitly say. One finds oneself contemplating, savoring, Proust’s words (as recently added cognitive stock) while carrying on reading Wollheim on how what he calls the transfer of pleasure “must, in a lover of painting, go both ways.” “It cannot simply go from domesticity to Chardin, it must also go from Chardin to domesticity,” and Wollheim—if we forgive him for anthropomorphizing and attributing intentional interest to a sensation—captures the engaged experience perfectly in words: “Pleasure now seeks a Chardinlike quality in domesticity, or a quality which can be discerned only by having looked at Chardin” (Wollheim 1987: 98–9). As one here reads Wollheim, as one recalls Proust while reading Wollheim, as one less exactingly yet nevertheless quite fully recalls all the ground now covered by both Wittgenstein on seeing-as and Wollheim on seeing-in, one finds oneself glancing back and forth between the Chardin and the Atget, and then one glances over and back to one’s own tablecloth-covered dining-room table next to one’s study while perhaps remembering (to abbreviate) candlelit others. In this experience—a complex experience of (a) seeing-in in written words (and words behind words, Proust behind Wollheim, Wittgenstein behind Wollheim) about painting, (b) seeing-in in photography where that photography becomes in its way, in this particular juxtaposition or combination of works, about painting, (c) seeing-in, in painting, aesthetic content that we then look across and see in photography, (d) seeing in painting a larger way of seeing the world that then carries over into life—these are all in meaning-determining ways intertwined, where each one inflects, and augments, the others. Here one wants to say there can be no simplification without oversimplification. And note (and fundamental to my discussion beyond where Wollheim left the matter of seeing-in) that Proust’s words were, “If, looking at a Chardin, you can say to yourself . . . , then you will be able to say to yourself.” What you can say is intimately interwoven with what you can see and what you have seen, how you can see, what you can see-as, and what you can see-in: language is not a separate realm only contingently added after the fact. And so in this connection, recall Wittgenstein’s criterion of perceptual difference above: “I describe what I see differently.” We can see words of poetry in the light of life’s experience, and we can see life through the words of poetry. But that is only a microcosm of what these reflections bring out: like Proust’s words reaching into our perception of Chardin, an experience which then stands as cognitive stock behind reading Wollheim in conjunction with looking at Atget, which then reaches across into our dining

156 Garry L. Hagberg room and our recollected images—we see our words in the light of other words, and as Wittgenstein captures it (having mentioned how a poet’s words can pierce us), “We let our thoughts roam up and down in the familiar surroundings of the words” (Wittgenstein 1967: section 155). If the kind of perception involved in aesthetic seeing is also active at the very core of our using, exchanging, and understanding words, then our aesthetic perception roams there (one might say: roams in and out, from the visual into the verbal and back, here Proust to Chardin to Wollheim to Atget to our words in our private perceptual worlds) as well. These are precisely the perceptual-linguistic and seeing-in-the-second-sense connections that I believe Caillebotte had in mind when in a letter to Monet he explained that, against some long-entrenched conceptions of representation as necessarily introducing distance, painting does not (or need not) remove us from the world, but on the contrary places us ever more intimately into it; he stated this most succinctly with the words, “I imagine that the very great artists attach you even more to life” (Caillebotte 2015). So bringing in language, and the intertwining way it works in these respects, somewhat complicates the picture Wollheim painstakingly painted; this linguistically encompassing view would take in the multiform phenomena of seeing-in in three dimensions, not only two. But it is precisely that categorically unwieldy complexity—the kind that brings Wittgenstein, Wollheim, Cavell, Proust, Chardin, Atget, a memory of the provocation of Leonardo, and our dining room together—that I believe increases the mimetic fidelity of this larger discussion of aesthetic perception as it manifests itself in art, in language, and in life.

Notes 1 I remain grateful for stimulating conversations with Richard Wollheim on issues of aesthetic perception and artistic meaning (some in the late 1970s in London, during the time he was completing the expanded edition of Art and Its Objects, and more in the mid 1980s in New York, during the time he was completing Painting as an Art); I would like also to acknowledge deeply helpful conversations on these matters with Michael Podro (in the same years) that have also informed this discussion. 2 This is not the place to consider in detail the connections to the work of Stanley Cavell (although I will note one below), but he insightfully discusses the difference under investigation here: I said that one aspect is hidden by another aspect. Suppose we ask: What is my relation to an aspect which has not dawned upon me, is in that sense hidden from me, but which is nevertheless there to be seen? What don’t I see when everything is in front of my eyes? I find that I want to speak of failing to see a possibility: I do not appreciate some way it might be—not just some way it might appear, but might be. (Cavell 1979: 370) 3 One can consider in this light Wittgenstein’s much discussed comment: “Not empiricism and yet realism in philosophy, that is the hardest thing” (Wittgenstein,

Leonardo’s challenge

4 5


7 8 9 10 11 12



1978: 325). As Cora Diamond has rightly observed, this was directed to F. P. Ramsey, but at the same it served as a larger point about philosophy (Diamond 1991a). In the present context, to reduce what we can say we know to what we can see in the first sense and to dismiss or subordinate everything else would be to allow an empirical picture of verification to direct our thought; to transcend this false limit, but to still preserve the distinction concerning mere “reading-in,” would be to maintain realism. I should note that I have found the acute study by Malcolm Budd (1989) consistently helpful and clarifying on these matters. The word “invitation” resonates throughout Wollheim’s extensive contribution to aesthetics more than one might initially realize; he was strongly interested in and influenced by the writings of Adrian Stokes. Stokes (1978: 270) briefly describes what he focuses on extensively as “the invitation to identify empathetically” with the work. One could say that seeing-as and Wollheim’s conception of seeing-in in fact always collaborate, for the simple reason that one has to see the representation as a representation before one sees anything (aesthetically speaking) in it. Wollheim discusses this briefly (1980a: 226), noting that while this may be said to be the case, the perceptual content of what is seen-as (that this, generically, is a representation before me) will always be different from the content of what is seen-in (I see, specifically, a flag in this paint, in these brushstrokes). I offer a discussion of this kind of aesthetic advancement, on a linguistic model, in Hagberg 2015c. Wollheim 1987: 51. Wollheim 1987: 77, where he writes, “Those views which do not ground representation in visual experience disqualify themselves on the spot.” Wollheim 1987: 71. I offer a way in which this richer intentional content might be characterized in Hagberg 2015b: 45–63. In this connection we can reconsider the two Picasso cases above (Picasso’s late self-portrait and Newman’s earlier photographic portrait of him). Wollheim is of course suggesting that the intention of the artist functions as determinant of interpretive correctness, but in a way not reducible to simple mind-to-matter matching. We can say that we better understand Picasso’s own intentional content in his self-portrait by seeing Newman’s photograph. And looking to the painting and then back to the photograph, we see better what Newman saw in Picasso as the person standing before him and that he brilliantly captured in his photograph—yet of course Newman had not seen that (much later) painting at the time of the photograph. So an understanding of, or a gaining of cognitivestock information about, Picasso as person was constitutive of the content of Newman’s intention in making that photograph; we, with both works before us, see in the portrait to a heightened degree the aspect of Picasso that Newman saw earlier and captured, and conversely in seeing Newman’s portrait we see better into Picasso’s intentional content. This is all relevant intentional material but not in any oversimplified way. Through such comparative and layered visual experience we see, as we say, more of “what they were going after.” It is important to bear in mind, in understanding Wollheim, that he is clear on the legitimate usage of this term and what it should and should not connote: he writes: the most traditional mode of art-history remains to my mind the most profound connoisseurship, the study inaugurated by Jonathan Richardson, advanced by Giovanni Morelli, and temporarily discredited by Bernard Berenson, who made it serve profit and the self-aggrandizement of the rich. (Wollheim 1987: 89)

158 Garry L. Hagberg 14 The fundamental problem being that the term “information” can suggest a list of cognitive items, all equally presently available to consciousness and all equally amenable to succinct propositional encapsulation. I would say that some are, and some, although part of the sensibility we bring to the perception of a work of art, are not. I offer a discussion of this in Hagberg 2016. 15 On this point see Wollheim’s remarks on the “bootstrapping” character of this perceptual operation (1987: 89). 16 The notes on this painting in the Fitzwillliam Museum, Cambridge, have been of help to me. 17 I have been reassured on this point by the remarks on this painting of the curator in the Tate Gallery, London. 18 I offer a discussion of this in connection with artistic meaning in Hagberg 2015d. 19 And to see what it is doing may involve the imagination in some cases, not in others, and to varying degrees where imagination is called for. But here again, and tellingly, this does not correspond at all to distinctions or gradations of subjectivity. This point is, I think, implicit in Wittgenstein’s remark following the above: Someone can take the duck–rabbit simply for the picture of a rabbit, the double cross simply for the picture of a black cross, but not the bare triangular figure for the picture of an object that has fallen over. To see this aspect of the triangle demands imagination. (Wittgenstein 2009a: section 217) 20 I extend this point in Hagberg 2015d. 21 These are the topics I pursue in detail in Hagberg 2015a. 22 I owe this imaginative perception to Laurie Anderson, who called attention to this in a presentation at Dartmouth College, 1984.

Bibliography Budd, Malcolm. Wittgenstein’s Philosophy of Psychology. London: Routledge, 1989. Caillebotte, Gustave. “Letter to Monet,” in the exhibition Gustave Caillebotte: The Painter’s Eye, National Gallery of Art, Washington, D. C., 2015. Cavell, Stanley. The Claim of Reason: Wittgenstein, Skepticism, Morality, and Tragedy. Oxford: Oxford University Press, 1979. Cavell, Stanley. “The Touch of Words,” in Seeing Wittgenstein Anew, ed. William Day and Victor Krebs. Cambridge: Cambridge University Press, 2010. Diamond, Cora. “Realism and the Realistic Spirit,” in The Realistic Spirit: Wittgenstein, Philosophy, and the Mind. Cambridge, MA: MIT Press, 1991a, pp. 39–72. Diamond, Cora. The Realistic Spirit: Wittgenstein, Philosophy, and the Mind. Cambridge, MA: MIT Press, 1991b. Drury, John. Painting the Word: Christian Pictures and Their Meanings. New Haven: Yale University Press, 1999. Hagberg, Garry L. “A Person’s Words: Literary Characters and Autobiographical Understanding,” in Philosophy and Autobiography, ed. Christopher Cowley. Chicago: University of Chicago Press, 2015a. Hagberg, Garry L. “Implication in Interpretation: Wittgenstein, Artistic Content, and ‘The Field of a Word’,” in Mind, Language, and Action: Proceedings of the 36th International Wittgenstein Symposium, ed. Daniele Moyal-Sharrock, Volker Munz, and Annalisa Coliva. Berlin: De Gruyter, 2015b, pp. 45–63.

Leonardo’s challenge 159 Hagberg, Garry L. “Wittgenstein, Verbal Creativity, and the Expansion of Artistic Style,” in Wittgenstein and the Creativity of Language, ed. Sebastian Greve and Jacob Macha. London: Palgrave, 2015c, pp. 141–76. Hagberg, Garry L. “Word and Object: Museums and the Matter of Meaning,” Philosophy, Supplementary Volume: Philosophy and Museums. Cambridge: Cambridge University Press, 2015d. Hagberg, Garry L. “On Patience (After Sebald): Documentary as a True Portrait of Sensibility,” in The Philosophy of Documentary Film, ed. David LaRocca. Rowman and Littlefield, 2016. Stokes, Adrian. The Critical Writings of Adrian Stokes, Vol. III, ed. Lawrence Gowing. London: Thames and Hudson, 1978. Wittgenstein, Ludwig. Zettel, ed. G. E. M. Anscombe and G. H. von Wright, trans. G. E. M. Anscombe. Oxford: Blackwell, 1967. Wittgenstein, Ludwig. Remarks on the Foundations of Mathematics, ed. G. H. von Wright, R. Rhees, and G. E. M. Anscombe. Oxford: Blackwell, 1978. Wittgenstein, Ludwig. “Philosophy of Psychology: A Fragment,” in Philosophical Investigations, revised 4th edition, ed. P. M. S. Hacker and Joachim Schulte, trans. G. E. M. Anscombe, P. M. S. Hacker, and Joachim Schulte. Malden: Wiley-Blackwell, 2009a. Wittgenstein, Ludwig. Philosophical Investigations, revised 4th edition, ed. P. M. S. Hacker and Joachim Schulte, trans. G. E. M. Anscombe, P. M. S. Hacker, and Joachim Schulte. Malden: Wiley-Blackwell, 2009b. Wollheim, Richard. “Seeing-as, Seeing-in, and Pictorial Representation,” in Art and Its Objects: With Six Supplementary Essays, 2nd edition. Cambridge: Cambridge University Press, 1980a, pp. 205–26. Wollheim, Richard. Art and Its Objects: With Six Supplementary Essays, 2nd edition. Cambridge: Cambridge University Press, 1980b. Wollheim, Richard. Painting as an Art. Princeton: Princeton University Press, and London: Thames and Hudson, 1987.


‘Surface’ as an expression of an intention On Richard Wollheim’s conception of art as a form of life Gabriele M. Mras

1 Art is not what one sees is a surprising, provocative claim. Would we not say, to the contrary, that it lies in the very nature of art to be perceivable? What do we admire in a painting, a sculpture or a musical composition if not its perceivable qualities? Strange as it may seem, just that claim, which appears to run counter to everything we assume in explaining our practice of appreciating art, has been put forward by not just one, but at least two philosophers. One philosopher Richard Wollheim was concerned with throughout his life – both in criticism and in defence of him – is the Oxford philosopher of logic, history and art R. G. Collingwood. Collingwood held the view that what we are aware of when we listen to music exists in something beyond the material sounds out of which the music is made. It is because of the imagination of a human being that sounds become tones with meaning, or a surface with paint on it becomes something more than wallpaper (Collingwood 1958). Collingwood’s point is clearly not that works of art are objects that cannot be perceived at all. It is because his aim was to explain how an objet d’art can really be understood that he was led to regard the perceivable features of an object as nothing more than the basis for an aesthetic appreciation of the object. Wollheim’s response to Art is not what one sees is also not to be understood just as the simple negative claim that sounds so paradoxical ‘out of all context’. He is convinced, too, that for a satisfactory explanation of aesthetic experience one must answer the question of how the viewers of art through the centuries can be said to have understood what the person who produced the artwork meant the object to be. The difference between Collingwood and Wollheim is not that one of them doubts the perceivability of the ‘meaning’ of an objet d’art while the other insists on affirming it. What Wollheim agrees with in what Collingwood says does not imply what Collingwood thinks it does. Wollheim would deny that accounting for the significance that works of art have for all of us

‘Surface’ as an expression of an intention


requires equating the work of art with something that the painting is seen ‘as’, something the painting is ‘of’ or something ‘in’ the painting. What Wollheim meant in introducing his notion of seeing-in is that the meaning of the work of art is what is seen in the painting.

2 What Wollheim calls ‘seeing-in’ (or ‘seeing y in x’) (Wollheim 1980: 209) – ‘the seeing appropriate to representations’ (ibid: 207), ‘a phenomenon’ (ibid: 209), ‘the nature of the perceptual genus of which the seeing appropriate to representations is a species, a certain species of seeing’ (ibid: 205) – is essential to the task of giving ‘a psychological account’ of the appreciation of art. Wollheim says he invokes ‘psychological factors’ in order to answer fundamental questions in the philosophy of art: a ‘psychological’ account, and by that I mean an account that invokes, in addition to the visible surface of the picture, only psychological factors. . . . It invokes the visual experiences of the spectator and the fulfilled intentions of the artist. (Wollheim 1993: 188) This is intended to reveal how those views that regard the surface of a painting as the only basis or evidence for its meaning have failed. Basically, such views exist in two ways – either as insisting on the artist’s intention, or what the artist meant as being most significant, or not. Wollheim treats both of them, interestingly, in a manner which even suggests how one could arise out of being dissatisfied with the other. I suggest reconstructing such a ‘development’, making use of Wollheim’s ‘The Art Lesson’, where he argues for a ‘transposition’ of Wittgenstein’s method from the Philosophical Investigations (on)to the philosophy of art.1 But let us look first at a view – actually giving rise to Collingwood’s theory – which turns to the relation between the painting and an object in accounting for our ability to see what is ‘on’ the canvas as meaningful. It seems to be a natural thought in Western culture that paintings have a determinate content in virtue of the relation they have to other objects. Of what a painting is – whereby Wollheim regards ‘being of something’ as an ‘essential feature of representation’ (Wollheim 1993: 159) – is hereby understood to be as ‘real’ thing, or as kinds of types of thing. As a paradigmatic assertion of this view one could take the one which claims that ‘for every representation . . . there must exist instances of the kind that the something or other represented is represented as belonging to’ (ibid: 161) but also ‘for every representation . . . there must exist the object that the something or other represented is represented as belonging to’ (ibid). A painting then is taken to be understandable as what it is a painting of – either as of an animal,

162 Gabriele M. Mras or a saint, or a situation or an event – depending on someone’s art (τε´χνη) to reproduce the features of the real object he or she meant to depict on the canvas. Wollheim finds inadequate any kind of ‘realism’ that attempts to explain how it is possible for a painting to represent something by appealing to something on the ‘canvas’ which is believed to reproduce the features of the things of which a painting is said to be. One way to show why this must leave us unsatisfied is to imagine or remind ourselves of other kinds of paintings. The restricted character of this kind of realism then becomes apparent. It is this restricted character of this kind of realism that I believe Wollheim finds in the calls ‘portrayal’ and ‘existential’ and the ‘figurative’ thesis (Wollheim 1993), and what he sometimes objects to in ‘seeing-as’ accounts. Let me remind you here what all such ‘natural’ or ‘representation by duplication’ accounts imply: the only things we become aware of in paintings are the particulars or kinds of things that are meant to be depicted. To understand what a painting is a painting of all that is required is the ability to recognize the kind of things depicted – animals, human beings, cities, etc. One of the things Wollheim has said many times against the view that the meaning of a painting is the object depicted is that the particularity of the way something is depicted gets completely lost. I take him not to be primarily concerned with modern art, where questions about the subjectivity of the artist’s depiction of something became more significant. When Wollheim criticizes this plausible-seeming account of how representation is possible, he does so from a philosophical point of view. I take him to be concerned with how art can be understood as an expression of the artist’s subjectivity. The ‘representation by duplication’ view fails in this respect. Wollheim’s ‘psychological account’ is meant to reveal the position we have in the world that makes it possible for us to see a painting as having a determinate content and makes it possible for us to share that same content with others. If all we had to be ‘equipped’ with, or all we had to share with others, when we engaged in art as viewers, producers or critics, were simply a list of objects of different kinds, this would perhaps look like a promising explanation of the similarities in our responses. But that would be so only under an unacceptable restriction on what we can look for, and what we can see, when we look at objects of art. One reaction to this dissatisfaction would then be to appeal to a second factor in addition to the painting’s relation to an object which is said to be responsible for the possibility of representation: the intention of the person responsible for there being a painting at all – the artist. The move towards the artist’s intention seems the right way to go. We do not wander through museums reminding ourselves, or inquiring, what the objects are that are depicted in the paintings. We have learnt to appreciate the different ways that artists place objects in situations or spaces and the perspectives and colours they employ. But this multiple-conditions view – grasping what a painting means or represents or is of, going beyond a list of

‘Surface’ as an expression of an intention


its depicted ingredients – does not really compensate for what proved to be unsatisfactory in the first place. Although it is true that without the artist’s intention a painting would not have been brought into existence, it is not enough to reduce what the artist produces to some ‘dominant’ idea that is then taken to be ‘the idea’ that made the artist paint the particular painting she or he did. In book I chapter VII §§6–7 of The Principles of Art (1958) Collingwood suggests such a reduction when he offers among the prior intention(s) something like an ‘imaginary painting’. Wollheim quotes his famous remarks – ‘A work of art . . . may be what we call an imaginary thing’ or ‘A work of art may be completely created when it has been created as a thing whose only place is in the artist’s mind’ (Wollheim 1973b: 250) – and objects: in internally producing a work of art-particular – say, in painting a picture to oneself (or imagining painting a picture) one is ex hypothesi not engaged in internally producing a token of a type which, in turn, is a work of art. (ibid: 258) He insists that ‘what the artist had in mind’ or what ‘one paints to oneself’2 needs to be understood in a different way for it to be of explanatory relevance. A mere appeal to the artist’s intention never in itself provides an adequate explanation of the painting. In appealing in the right way to an artist’s ‘intention’ one must appeal to the expression of what is taken to be what the painting represents. If we try to understand something as ‘an expression of an intention’ without having understood the expression itself, or the mode of expression, we lack all basis on which to identify what the intention really is. If we were to grant that an art object is the expression of an intention, while at the same time having no account of what the expression actually means, we would have a statement of an ‘identity’ between intention and ‘its’ expression that leaves us with nothing to say about the painting.

3 At the centre of Wollheim’s ‘psycho-methodological account’ lies the conviction that this disappointing conclusion is the result of a confusion or misunderstanding. The way to answer the question ‘how can we understand a painting as representing something (definite)?’ is not by any appeal to an ‘external’ relation between the painting and something else. We take ourselves to be capable of appreciating pictorial art while being unburdened by questions concerning who commissioned a painting, what exactly the situation depicted is meant to represent – what role some of the objects and figures in a painting really have – and in which competition the artist stood or stands with his companions or with his colleagues. So far it

164 Gabriele M. Mras seems that – at least as ‘ordinary spectators’ – we do not take questions about what a painting really ‘means’ to be something that is ‘behind’ or ‘in’ something in addition to what a painting represents; and if it were otherwise, given what was said before, our attitude towards paintings would then have to be essentially different from the one we are familiar with now – possibly we would treat paintings as welcomed sources of inspiration, or as making us think more about ourselves, or just as some colour patches that encourage meditating. When Wollheim writes that seeing-in is ‘the seeing appropriate to representations’ which ‘separates itself from other species of the same perceptual genus . . . in that . . . there is . . . a standard of correctness’ (1980: 207, my italics), we see Wollheim clearly being concerned with how we can settle the question of whether spectators recognize what the painter meant to make them see in a painting. That there is a measure of correctness includes the thought that we might miss it – miss, for example, what the depicted objects in a painting ‘stand’ for, or miss the respect in which the use of some material is intertwined with the situation represented. The fact that we only sometimes believe we see what we are meant to see is what makes us raise questions about how to approach questions concerning the representational content of a painting. It is here where questions like ‘What is it we really see in a painting?’ or ‘What is on the surface of a painting?’ become pressing. It is also here where Wollheim thought the significance of his work in the philosophy of art and mind lies. What corresponds to his belief that the various theories of depiction are ‘misunderstandings’ is that to have a ‘broader’ conception of ‘representational content’ does not pose serious difficulties. Wollheim believes that we do see what we understand a painting to have as representational content. Contrary to Collingwood, Wollheim believes that leaving ‘as the bearer’ of pictorial meaning ‘the visible surface of the picture “on” the painting’ does not imply a conception of seeing a surface as a kind of seeing that could be isolated (completely) from imagining. When, for example, we see some of the depicted objects in a painting as symbols of beauty or perfection – as in Giovanni Bellini’s sacra conversazione or in Chardin’s still lives – this is so because both painters succeed in using the actual material in a way that gives us (visually) the impression that the ‘relation’ these objects have to each other is of a ‘self-absorbed’ quality (Wollheim 1973d: 155–76). Or: in the late self-portraits of Lucian Freud what we see his face to be is (partly) determined by the very quality of the strokes and the amount of layers of paint Freud uses. For our experience of the (representational) content of these paintings the property of the surface material is relevant for our having a particular impression of his face even though this property has nothing to do with our being able to perceive what skin looks like. ‘Seeing x in y’ for Wollheim is the right ‘form’ of the ‘phenomenon’ or ‘the experience’ of paintings having meaning, for what we see when viewing a

‘Surface’ as an expression of an intention 165 painting is what we see when we see something in a painting; the relevant ‘is’ is constitutive. Sometimes it sounds as if Wollheim’s emphasis on ‘surface’ suggests that it is a means in order to understand the representational content of a painting. Following what was said earlier, however, this cannot be the case. But we do find Wollheim writing that ‘my account of pictorial meaning . . . leaves as the bearer of such meaning the visible surface of the picture’ (Wollheim 1993: 189) and of ‘perceptions, one of the pictorial surface, the other of what it represents’ (Wollheim 1998: 222). The metaphor of a ‘surface’ and something ‘behind’ it, or the suggestion that ‘we are visually aware of the marked surface and of something in front of or behind something else’, is indeed easy to be taken in the wrong way. Let us briefly look back at the very first ‘thesis’ discussed here. Wollheim reminds us in connection with the ‘portrayal and resemblance-thesis’ that when resemblance is invoked as what makes us able to identify the object the painting is of, then this always rests on the spectator’s having some knowledge of the objects said to be identifiable via the surface features the object and the painting share. It goes without saying that what the object and its picture ‘share’ as properties are visual properties. This has its expression in the fact that we understand resemblance – we cannot say, except in a special setting, ‘Napoleon is exactly like this drawing’ or ‘Napoleon resembles this drawing’: which seems to throw some light on how the ‘this’ in the first sentence is to be taken. (Wollheim 1980: 18) – not to be symmetrical. But if not, when we point to some features as similar to aspects of a person, what we point to is some area ‘on’ the canvas as (expressing) an aspect of a face or a figure. What is ‘on’ a canvas is taken as a case of resemblance only relative to what we take the surface of a canvas to be an expression of – and this clearly cannot be what is suggested as the external object, which the painting is believed to be related to by resembling it. Wollheim finds Wittgenstein making – in what he says about the duck– rabbit picture puzzle – the same point. It might seem as if the ‘it’ or ‘this’ to which we point when we seek to demonstrate something about a picture is ambiguous: ‘I see that it has not changed; and yet I see it differently’ (Wittgenstein 2009 §§111, 113, 191). As ‘expression of our experience’ it is suggested that we use ‘it’ in a manner such that on the one hand it stands for a configuration of lines and on the other for a depicted object (§118: 194, §116: 193). Believing, however, that we could isolate the seeing of the arrangement of what are purely lines as what stays the ‘same’ – only one time as standing for the subject expression the other time for the predicate expression – is a mistake. Since if in ‘I see that it has not changed; and yet I see it differently’ these lines are understood as what is predicated and what something is predicated of, and if ‘it’ is taken to stand for ‘the same’

166 Gabriele M. Mras nonetheless, we end up saying that we do not know what this is that is able to change its position. A ‘surprise’, expressed by ‘The same – and yet not the same’ (§111), which we react to best by giving up the idea that the relation – speaking in Wollheimian terms – between what is ‘on’ the surface and what we see ‘in’ a painting as that what is ‘on’ the surface, can be conceptualized on the model of a subject-predicate relation. That a subject-predicate analysis of representation is ultimately unforthcoming emerges clearly when we recognize that, in the cases where such an analysis is plausible, as with a proposition, there is no limit to what I can predicate (truly or falsely) of a subject. (Wollheim 1973c: 302f.)3 For Wollheim there is no conflict between being a representation and being seen or being perceivable. The impression that this is not so, i.e. we only believe that we see, has so far nothing in its favour. As seeing something ‘in’ a painting ‘cannot be accounted for by the mere appeal to anything like an intention’, it could not possibly be understood according to a model where the connection between ‘in’ and ‘as’ or ‘on’ is thought to reproduce the relationship between the subject and predicate of a sentence. This is ultimately also mirrored in the way ‘content’, ‘form’ and ‘surface’ of a painting are related. Of a content one could say that it is presented in a certain way, yet ‘surface’ and ‘representational content’ are not even cognate aspects of a painting belonging to the same categorical kind. A content has no surface or features; a surface has no meaning or content. The surface features used, on the other hand, could be said to partly determine the content of a painting, so that ‘form’ or ‘style’ are regarded as constitutive for a representational content to be what it is. All this means, however, is that in the case of a painting what we say about the (pictorial) content of a painting is underdetermined as to its visual realization. The same is true when we talk of ‘style’ or ‘form’ – as if one could simply compare paintings or painters according to aspects which could be visualized without their relata. Given this criticism, it only seems to be that it is the surface – or the appreciation of the surface – that is suggested by Wollheim as that by virtue of which we arrive at an adequate understanding of the ‘meaning’ of a painting. For him a painting is essentially not a means by which we come to understand what it is. It would thus be a mistake to read Wollheim as affirming some epistemological relation between the surface features of the painting and the intention of the artist – as if the ‘surface’ now were the reason we see the intention expressed ‘in’ it – the ‘content’.

4 At the beginning of this paper I suggested that Wollheim’s calling his account of ‘seeing-in’ a psychological approach should be taken seriously. If in seeing, one simply sees what is so, i.e. that the object in front of us has

‘Surface’ as an expression of an intention


certain perceivable characteristics, then the difficulties we face in accounting for the necessary conditions of our perceiving what we do in art would be intertwined with all our (perhaps mistaken) conceptions of perception in general. If all Wollheim had aspired to show was that a capacity for conceptual thought is exerted in perception, the significance of his introduction of ‘seeing-in‘ would be largely diagnostic. But what he came to appreciate about ‘seeing-in’ – and in doing so left behind primarily conceptual investigations – is much more than that. According to both views described at the beginning of this paper, the connection between artist and specator requires a general practice of the appreciation of pictorial art. On the first – more ‘realist’ – view, it is of little significance that what we see represented in a painting is a product of the subjectivity of the artist. What matters is that we are connected as spectators with the artist by finding ourselves inhabiting the same world of objects. The second and more dominant view in the philosophy of art holds that in coming to understand a painting in front of us we become acquainted with and linked in a particular way with the subjectivity of the artist. We find in successful pictorial depiction a particular expression of a conception of human beings as persons – a recognition of something essential to our nature as human beings. But, if so, we do not just recognize that in seeing a painting we share perceptions. We recognize that we are like her or him by perceiving what is put in front of us as painting.

Notes 1 ‘to transpose Wittgenstein’s finding from the language lesson to the art lesson’ (Wollheim 1973a:144). Let it then be supposed that my procedure is correct: suppose that it is legitimate to transpose Wittgenstein’s finding to the domain of art: suppose that a proper understanding of the art lesson will give us an insight into the nature of art – where does this lead us? (ibid: 145) 2 ‘the identity of a painting that one paints to oneself and some painting that, at a later moment in time, one actually paints is notoriously difficult to determine’ (Wollheim 1973b: 258). 3 Wollheim sometimes takes ‘seeing-as’ – insofar as it rests on a sharp distinction between that ‘in virtue of which the picture picks out . . . a certain object’ and that in virtue of which it ‘describes that object’ – as suggesting such a subject predicate model (1973c: 302).

Bibliography Collingwood, R. G. 1958 [1938]. The Principles of Art. New York: Oxford University Press. Hopkins, Robert 2003. ‘What Makes Representational Painting Truly Visual?’. In Aristotelian Society Supplementary Volume 77.1. Wiley Online Library, pp. 149–67.

168 Gabriele M. Mras — 2010a. ‘Inflected Pictorial Experience: Its Treatment and Significance’. In Philosophical Perspectives on Depiction. Ed. by Catherine Abell and Katerina Bantinaki. The Mind Association Occasional Series. Oxford: Oxford University Press, pp. 151–80. Kemp, Gary 2003. ‘The Croce–Collingwood Theory as Theory’. The Journal of Aesthetics and Art Criticism 61.2: 171–93. Wittgenstein, Ludwig 2009 [1953]. Philosophical Investigations. Ed. by P. M. S. Hacker and Joachim Schulte. 3rd edition. Oxford: Wiley-Blackwell. — 1958. The Blue and Brown Books. Oxford: Blackwell. Wollheim, Richard 1969 [1959]. F. H. Bradley. London: Penguin Books. — 1973a [1971]. ‘The Art Lesson’. In On Art and the Mind. Cambridge, MA: Harvard University Press, pp. 130–51. — 1973b [1972]. ‘On an Alleged Inconsistency in Collingwood’s Aesthetic’. In On Art and the Mind. Cambridge, MA: Harvard University Press, pp. 250–60. — 1973c. ‘Nelson Goodman’s Language of Art’. In On Art and the Mind. Cambridge, Massachusetts: Harvard University Press, pp. 291–314. — 1973d [1970]. ‘Walter Pater as a Critic of the Arts’. In On Art and the Mind. Cambridge, MA: Harvard University Press, pp. 155–76. — 1980. Art and Its Objects: With Six Supplementary Essays. 2nd edition. Cambridge: Cambridge University Press. — 1987. Painting as an Art. Princeton: Princeton University Press. — 1993. The Mind and Its Depths. Cambridge, MA: Harvard University Press. — 1998. ‘On Pictorial Representation’. The Journal of Aesthetics and Art Criticism 56.3: 217–26.


Richard Wollheim on seeing-in From representational seeing to imagination Richard Heinrich

Introductory remarks ‘Ambiguity – rabbit or duck? – is clearly the key to the whole problem of image reading’. This quotation from Ernst Gombrich’s Art and Illusion (1972: 238) is as good a starting point as any when taking up questions concerning ‘seeing-in’. Gombrich’s ideas about the interplay between theories of perception on the one hand and historical developments in the representational arts on the other hand opened up many important issues which, given my title, could reasonably be discussed in detail1; but for the moment I will use the quotation only in a superficial way, as a signpost pointing in two directions. The one is Richard Wollheim’s conception of ‘seeing-in’ as fundamental for our understanding of representational art – and the origins, revisions and refinements of that conception; in this context I am going to mention objections raised against Wollheim by Michael Podro and Wollheim’s reaction to them. The other direction is given with the reference to Wittgenstein by Gombrich in a footnote when he brings up the duck–rabbit figure on page five of his book. I cannot resolve the question of whether this footnote was actually meant to refer to Wittgenstein’s remarks about the duck–rabbit – it could as well just have been the easiest way to supply the formally required quotation of the source of the illustration. Anyhow, the illustration is definitely not the drawing in the Anscombe edition of the Philosophical Investigations (Wittgenstein 1958: 194)2 referred to in the footnote; neither is it – as Gombrich erroneously suggests in the main text – taken from the German magazine Fliegende Blaetter in 1892. It is a drawing by Joseph Jastrow, copied from a print in the American magazine Harper’s Weekly (1892) – which in turn was based on the Fliegende Blaetter version. Jastrow originally published his drawing in 1899, and a second time in his book Fact and Fable in Psychology (Jastrow 1900); that is indeed the reference given by Wittgenstein. However in the text Wittgenstein provides a drawing of his own, which differs from the one by Jastrow significantly in that it is totally schematic and lacks, above all, the hatching of the interior of the figure (Brugger 1999). That Gombrich does not quote the appropriate source for the illustration he used, referring instead to

170 Richard Heinrich Wittgenstein (where he may have come across the figure originally) could be taken as an indication that he indeed wanted to direct attention to the ideas expressed there. I tend to see these ambiguities as a reminder that in the development of Gombrich’s and Wollheim’s ideas about the relationships of seeing and representation, Wittgenstein’s remarks about the duck–rabbit figure are a constant but not always clearly visible thread. Conversely, when asking what in Wollheim’s concept of ‘seeing-in’ has become of the issues of aspect-seeing as treated by Wittgenstein, it might occasionally be helpful to recall Gombrich’s theories.

Seeing-as One reason to keep this complex situation in mind is this: when Wollheim, in his review of Art and Illusion, criticizes Gombrich’s use of the figure, his main point is that Gombrich concentrates on its character as a Kipp-Bild, a flipflop-image, taking this as an analogy helping to illuminate his thesis of the mutual exclusion of two (equally necessary) attitudes towards a picture, i.e. seeing the material object vs. seeing the represented subject: For Gombrich, to see something as a picture of an object is to see it sometimes as a picture and sometimes as that object. To admire something as a good or naturalistic picture of an object is to say something about the speed or facility with which we can move between these two different ways of seeing it. (Wollheim 1974b: 278) Wollheim thinks that this illuminates the issue from the wrong side and that those two attitudes towards pictorial representation are not (and cannot be) mutually exclusive and that, consequently, if the duck–rabbit figure is a valuable analogy at all, then it is certainly not because of its flipflop character. And then in the further development of his own ideas Wollheim saw no need to reflect on the peculiarities of flipflop-images and more or less disregarded this side of the issue. Like Wollheim, Wittgenstein thought it important to distinguish between ‘continuous seeing’ and ‘dawning’ (aufleuchten) of an aspect,3 but he seems to have been equally convinced that significant insights about ‘seeing-as’ in general could be gained from the phenomena of ‘change of aspect’ (Aspektwechsel) – his choice of the duck–rabbit-figure to illustrate this was certainly not a misunderstanding. But what exactly was Wollheim’s criticism of Gombrich? Wollheim does not explicitly say this, but I think it basically comes down to the following: that Gombrich simply presupposes his incompatibility-thesis, expressed in laconic formulations like: ‘Ambiguity cannot be seen’ (1972: 329); or in a more elaborated form like: But is it possible to see both the plane surface and the battle horse at the same time? If we have been right so far, the demand is for the impossible.

Richard Wollheim on seeing-in


To understand the battle horse is for a moment to disregard the plane surface. We cannot have it both ways. (1972: 279) That appeal to the duck–rabbit figure does not in the least explain why and how this should be the case but simply lures us into the illusion of a better understanding by giving a striking example of a situation where indeed we cannot have it both ways. But to that end no example was necessary – the resolution in Wollheim’s words: But why does Gombrich assume that we can no more see a picture simultaneously as canvas and as nature than we can see the duck-rabbit figure simultaneously as a duck and as a rabbit? Because canvas and nature are different interpretations. But, if this is Gombrich’s argument, it is invalid. For the reason why we cannot see the duck-rabbit figure simultaneously as a duck and as a rabbit is not because ‘duck’ and ‘rabbit’ are two different interpretations, it is because they are two incompatible interpretations. (Wollheim 1974b: 280) So nothing in favour of Gombrich’s incompatibility-thesis can be obtained from the comparison with the duck–rabbit figure. But Wollheim is very cautious as to further conclusions or conjectures: he does not, for instance, say that in the case of a picture ambiguity can be seen, provided that the respective interpretations are not incompatible and fulfil certain additional conditions. He does not say that an interpretation as canvas and an interpretation as nature is an example of ambiguity. Although he uses the very phrase ‘seeing-as’ to characterize theories which understand seeing as interpretation (1974b: 285), this use is limited to his criticism of Gombrich’s position, namely his particular version of ‘seeing as interpretation’ as progress by trial and error or, more specifically, schema and correction;4 no independent statement regarding this type of theories in general is articulated, and ‘seeing-as’ is not used by Wollheim to express any theoretical position of his own. But in the first edition of Art and Its Objects the concept of ‘seeing-as’ is given positive content, although Wollheim still proceeds in an extremely guarded manner: He now says that ‘there is no incompatibility between seeing one mark on the canvas as behind another and also insisting that both the marks and the canvas [ . . . ] are physical objects’ (1980: 16), which more or less amounts to the position that what was earlier described as ‘ambiguity’ can be seen; but he does not say this explicitly, nor does he offer an argument. Instead he appeals to our familiarity with phenomena of seeing something as having properties which it cannot have due to its being the material object it is. And that is what he calls ‘representational seeing’, or ‘seeing-as’ (1980: 16). At the end of the chapter he admits to having said nothing substantial in favour of this rather weak position, and that is true. But before that he has explicitly stated his opinion that seeing-as is a concept needed to explain representation in general, and (second) that to explain representation one needs the

172 Richard Heinrich further concept of intention. Both of these strong claims have remained cornerstones of his later theories, the second specified as providing a standard of correctness (of perception).

Development Only when he sketched his conception of ‘seeing-in’ in the fifth addendum to the second edition of Art and Its Objects (1980: 205sqq) did Wollheim systematically confront the issues connected with seeing-as, or representational seeing. Before directly addressing this conception I will try to illustrate how the need for an independent and complex theory of that kind had grown over the years by comparing two passages: the first is from the 13th chapter of the first edition of Art and Its Objects; the second is from an earlier lecture. In chapter 13 of Art and Its Objects Wollheim argues against the ‘resemblance theory’ of representation: that is, against the idea that resemblance between the material object and the depicted subject is the basis or explanation of representation. He says (1980: 18) that in a situation where we look at a portrait and exclaim ‘Oh, how exactly like NN this looks . . . if we expand the pronoun “this”’ we will rather say ‘this person’ than ‘this configuration’. That is very plausible indeed. But it raises further questions: what is the exact status of the notion ‘person’ in the explanation of the representation? The relevant configuration is here before our eyes, NN is standing over there in the same room – where is the ‘person’?5 Or, more aptly: what is the link between the two possible resolutions of ‘this’, the configuration and the person? And, second, what is the configuration? What is ‘it’ that we see as? The usual answer to this question is something like: strokes, patches of colour, lines, etc. But is this really satisfactory? What is it that we see as a line of this or that length?6 Only a comprehensive theory of representation can answer such questions, and the fifth supplement will be the sketch of such a theory. There is a passage in Wollheim’s Inaugural Lecture to University College London from 1964, ‘On Drawing an Object’ (1974a), which gives a more accurate impression of how his awareness of these problems grew. In this text he quotes extensively from Wittgenstein’s Philosophical Investigations, II, xi. Some of his remarks contain extremely valuable insights, but in one footnote he criticizes Wittgenstein on a strange point, namely the introduction of the concept ‘picture-object’ right after the presentation of the duck– rabbit figure. Relying on this concept of a picture-object, Wittgenstein says (1958: 194sq.) that when I am exposed to the duck–rabbit figure and see a rabbit (and nothing else), I could answer the question ‘What do you see?’ by saying that I see a picture-rabbit. And that in that case I would certainly not answer: ‘Now I am seeing it as a picture-rabbit’; as I would not say, when holding a fork in my hand: ‘I see this as a fork’. And Wollheim takes this as an indication that Wittgenstein intended to speak of seeing-as only in cases

Richard Wollheim on seeing-in


where an alternative reference (in the sense of: incompatible depicted object) is at hand – a bit as if he saw Gombrich at work in Wittgenstein’s thoughts. And then Wollheim says that this view is wrong, because obviously even if he had not noticed the duck, Wittgenstein could have seen the drawing as a configuration (of a curved line and a point). And then he says that this failure ‘seems to be concealed by Wittgenstein’s introduction of the clumsy phrase ‘picture-object’, which he uses to cover the configuration, the representation, and the object represented’ (Wollheim 1974b: 25). In my view this is a misunderstanding of what Wittgenstein wanted to say. I cannot, however, clear up that misunderstanding in the sense of putting Wittgenstein’s intentions straight. I will just try to explain, in a few words, why it is significant for the difficulties Wollheim himself had at that time to articulate a positive understanding of seeing-as. When he says that Wittgenstein’s fictional interlocutor could very well have said ‘I see this as a rabbit’, because there was the alternative possibility that she could have seen it as a configuration (interpretation in a different, but not incompatible way), then what would his answer have been when asked what he had in mind as that which is being seen-as a configuration? As already observed, in contexts of this type (where it is assumed that when not stated otherwise we talk of pictures) the object answering to that question is usually supposed to be the material configuration on the canvas. That is ruled out here. But regardless of that, it is quite obvious what Wollheim’s answer in this situation would have been: the drawing. It is the drawing which, in his opinion, can be seen-as a rabbit (representationally) and can as well be seen as a configuration – ‘even though there is not something else of which it can be seen as a representation’, as he says (Wollheim 1974b: 25). But that amounts to the same as what was brought out by Wittgenstein – not clumsily, but on the contrary very elegantly – when he introduced his ‘picture-rabbit’. When asked ‘What do you see?’ in the situation described by Wittgenstein, ‘picture-rabbit’ is the same answer as ‘rabbit’, plus the qualification that it is a picture, a drawing, which I see-as rabbit. If we imagine a situation where it is understood that we have to do with ‘pictures as pictures’, as when someone is shown a series of animal depictions, each on its own card, and asked to identify them – if we imagine such a situation, then ‘rabbit’ could perfectly well stand in for ‘picture-rabbit’. One could make this point even more clearly by saying: ‘plus the qualification that it is in a drawing that I see the rabbit’. And that is indeed how Wittgenstein puts it sometimes – for example, when he speaks of ‘seeing the old face in the altered one’ or (right at the beginning of II, xi) ‘seeing a likeness in two faces’ (Ich sehe eine Aehnlichkeit in diesen beiden Gesichtern; Anscombe translates: ‘between these two faces’). I do not want to suggest that Wollheim’s later use of the phrase ‘seeing-in’ can be directly linked to Wittgenstein’s parlance here. I just want to draw attention to the fact that four years later, in Art and Its Objects, when Wollheim speaks of the natural expansion of the word ‘this’ into ‘this person’ rather

174 Richard Heinrich than ‘this configuration’, he openly does the same as Wittgenstein did with the introduction of his Bildgegenstand (picture-object): identify something which can be seen depicted but different from (and possibly similar to) the referent or subject of the picture.7 To repeat: Wollheim gave no hint as to the status of this ‘person’ and the whole remark must be seen in its negative context, which is an argument against the resemblance-ideology.

Seeing-in Art and Its Objects, 2nd edition Now when I come to Wollheim’s sketch of seeing-in as the key to an understanding of representational seeing, I will concentrate on two points only: the so-called ‘twofold-thesis’, and the problem of the (special) additional perceptual skill he postulates for seeing-in. So I will not say much about the important distinction he makes between representational seeing on the one hand and ‘seeing appropriate to representations’ (1980: 213) on the other hand (the latter including a standard of correctness referring back to an intention on the side of the artist); and, of the three reflections which lead to the concept of seeing-in, I will not comment on the negative features of category constancy and localization which characterize seeing-as, and only underline that the third – ‘simultaneous attention to what is represented and to the representation, to the object and to the medium’ (1980: 213) – now, for the first time, positively states and sort of explains what was from the beginning (1974b) his conviction (against Gombrich). There is a special competence responsible for this twofoldness, called seeing-in, which he analyses structurally in opposition to other perceptual competences, and of which he postulates that it be phenomenologically comprehensible in some way, that it should, as it were, manifest itself phenomenologically. This means that seeing-in is in fact his first independent and autonomous theoretical approach towards what until then he had always called seeing-as; and that what he now calls seeing-as is a competence underlying (in his view) all perception of objects: that seeing an object always includes the interpretation of the perception by a covering concept, but in a way which cannot be understood as the interaction of two mutually external agencies (perception and judgement), as some sort of inferential process or process of trial and error, but must be understood as a permeating or mixing of perception and a general concept or attitude. Seeing-as, understood this way, excludes simultaneous attention to the object seen and those features of the visual experience which sustain our seeing of that object. So when Wollheim now says that he was in error when he, in the first edition, tried to explain representational seeing by means of seeing-as, he retrospectively applies his new and more narrowly defined notion of seeing-as.8 What he really does is to refine and clarify his old notion of seeing-as, to baptize it ‘seeing-in’ and to apply the notion ‘seeing-as’ to something which

Richard Wollheim on seeing-in


under the most important aspect (namely: simultaneous attention to subject and medium) is the exact opposite of what he until then had intended with it. But what then is the competence, the skill characteristic for seeing-in? The key is the peculiar way in which it is related to what he calls ‘straightforward perception’ (1980: 216). Seeing-as is directly related to straightforward perception and is part of it; ‘by contrast seeing-in derives from a special perceptual capacity . . . [that] allows us to have perceptual experiences of things that are not present to the senses’ (ibid). The way from this stipulation to an explanation of the appropriate perceptual attitude towards representational pictures is in steps: first there is a level where dreams and hallucinations can be referred to as examples, then comes the stage where such experiences are had through the seeing (perception) of things present (Leonardo),9 and the last stage is the intentional bringing about of such experiences (by the artist; Wollheim 1980: 220–7). I will not comment on details of this conception but only address one striking feature of his presentation of the theory. Obviously, it is built around the nucleus of the ‘special perceptual capacity . . . [that] allows us to have perceptual experiences of things that are not present to the senses’; this constant core is integrated into more complex contexts (independently perceived things, independently developed skills of manipulating and shaping objects). Now isn’t it strange how Wollheim throughout his exposition consciously avoids calling this perceptual capacity by the name it has had for thousands of years: imagination? He simply does not speak of images, he seems to avoid the word at all costs and cautiously speaks of ‘visions’: visual experiences in the mind’s eye – and not only in the mind’s eye but originating from the seeing of present things. Or instead of images he speaks of visions of things not present in things present. One reason for this could have been to avoid any appearance of convergence with or closeness to traditional theories from Aristotle onwards which see imagination as a necessary element within direct perception itself. Those theories were, for the most part, motivated by questions like: how is the mixing of intellectual and sensual factors in direct perception (now ‘seeing-as’ for Wollheim) to be explained? There have been different answers, of course, but the keyword used has, basically, always been the same: imagination. Now take into account that also for Wittgenstein aspect-seeing was the topic where the interaction of conceptual and perceptual competences should be discussed in the first instance. This is well brought out by Budd: One reason for his [Wittgenstein’s] interest in the idea of seeing an aspect was undoubtedly its close relationship with the idea of experiencing the meaning of a word. But it would be mistaken to think of his interest in aspect perception as being entirely derivative from his interest in the experience of meaning. On the contrary, the experience of meaning owes its philosophical importance to its kinship with the perception

176 Richard Heinrich of an aspect. For the interest of the experience of meaning is dependent upon the range of similar psychological phenomena which in general have nothing to do with word-meaning and the perception of an aspect is a principal member of this related class. The independent philosophical importance of the concept of noticing an aspect is due to its location at a crucial point in our concept of the mind – a point from which lines radiate in all directions across the field of psychological phenomena. This point, as we shall come to recognize, is the juncture of the sensory and the intellectual. (1987: 1) Given this common focus, we can reasonably presume that a strong reason for Wollheim to avoid speaking of images or imagination lies in Wittgenstein’s general reservations concerning inner pictures or mental images (innere Bilder). The classic passage comes early on in Philosophical Investigations II, xi: My visual impression has changed; what was it like before and what is it like now? If I represent it by means of an exact copy – and isn’t that a good representation of it? – no change is shewn. And above all do not say ‘After all my visual impression isn’t the drawing; it is this – which I can’t shew to anyone’. Of course it is not the drawing, but neither is it anything of the same category, which I carry within myself. The concept of the ‘inner picture’ is misleading, for this concept uses the ‘outer picture’ as a model; and yet the uses of the words for these concepts are no more like one another than the uses of ‘numeral’ and ‘number’. (And if one chose to call numbers ‘ideal numerals’, one might produce a similar confusion.) (Wittgenstein 1958: 196) The point is put more generally by Wittgenstein here: What is the criterion of the visual experience? – The criterion? What do you suppose? – The representation of ‘what is seen’. – The concept of a representation of what is seen, like that of a copy, is very elastic, and so together with it is the concept of what is seen. The two are intimately connected. (Which is not to say that they are alike.) (ibid: 198) The ‘elasticity’ of the concept of representation (Darstellung) is important: it covers a wide range of possibilities of expression, from drawings via verbal descriptions to pantomime: ‘If someone sees a smile and does not know it for a smile, does not understand it as such, does he see it differently from someone who understands it? He mimics it differently, for instance’ (ibid). This – Wittgenstein’s insistence on the external criterion for inner states, and

Richard Wollheim on seeing-in


the implications for the concept of an inner or mental image – Wollheim had understood from the beginning. One cannot put it more clearly than he did in On Drawing an Object: Of course, if a man draws what he has seen, there is something that is prior to his doing the drawing and which is also that on the basis of which he does the drawing. And that is the visual experience itself. If the visual experience had not been as it was, the man would not have drawn as he did [ . . . ] That is indubitable. But the fact that the visual experience can be in this way operative after it has passed, should not lead us into the view that it somehow persists in the form of a lingering image, which we then try to reproduce when we set ourselves to represent what we have seen. For there is no reason either in logic or in experience to believe in the existence of such an image. We do not need it in order to explain the facts of the case, nor do we have any independent evidence for its existence in our actual consciousness. It seems a pure invention conjured into being to bridge the gap between one event – our seeing as we do – and another event, when what we have seen asserts its efficacy. (1974a: 9) Painting as an Art So if we look at Wollheim’s conception of seeing-in sympathetically, leaving aside the question of terminology and the more difficult or more special or more complex issues, we will find interesting affinities with Wittgenstein’s fragment throughout, and one could say that it covers at least an important part of what for Wittgenstein the issue of aspect-seeing meant. In particular the last quotation convincingly shows the ‘Wittgensteinian’ origins of Wollheim’s reservations concerning any unguarded use of the concepts of image and imagination. Now, as is well known, Wollheim gave a third presentation (if not to say: third version) of his conception of representational seeing (after the two editions of Art and Its Objects) in his book Painting as an Art (1987). It could be called a third version on the basis of his own remark in a footnote (1987: 360) that, following critical objections by Malcolm Budd and Michael Podro, he had changed his view on a significant point, namely that he does not understand ‘twofoldness’ anymore as the simultaneity of two different kinds of attention but rather as two aspects of one state or activity of attention. That was indeed a point in Podro’s review, where he found Wollheim still too close to Gombrich: For what the argument has done is to separate representational seeing or seeing-in by definition from scrutiny or curiosity about the object which is the bearer of the representation. This does not make them perceptually incompatible, as Gombrich has assumed, but as in

178 Richard Heinrich Gombrich’s theory it divides concern for what is really there . . . from concern for what it depicts or suggests. (Podro 1982: 100) Two qualifications are in order here. First, this is not the most important point Podro makes in his review. In fact, he sees it as dependent on Wollheim’s more principal failure to distinguish ‘between the artist’s material, the marks and surfaces he employs, and his medium in the sense of those materials used with the purpose of “raising the ideas of bodies in the mind”’ (ibid). Podro elaborated this criticism in remarks about Wollheim in a later article: a line in a drawing is not just a mark on a surface but is ‘being perceived as the material in use within the representational procedure’ (Podro 1987: 18). Second, I do not think that the change from the ‘two-kinds version’ to the ‘two-aspects version’ amounts to much. But even if I would, consequently, rather not speak of a third version of the theory, it certainly is a distinct presentation. One distinction, when compared to Art and Its Objects, is the much more detailed elaboration of the process which leads from the ‘special perceptual skill’ to the varieties of producing, seeing and understanding of pictures – what he calls the cultivation of this special kind of visual experience. I cannot go into this, but certainly the ingenuity of his conceptualizations and the details of his interpretation of pictures are what make for the immense richness of the book. But then at an early stage of this reconstruction there is a new idea which can be treated independently, and it is to this I am going to devote a few final remarks. In the opening chapter Wollheim introduces (for the first time systematically) the concept of an image. He invites us in this passage about ‘Ur-Painting’ to watch an imaginary artist, a painter, in the process of inventing, as it were, representation. We are at a point where the artist has already ‘thematized’, as Wollheim calls it, the mark, the surface (in distinction to the support) and the edge, and has, in addition, already had conscious experiences of seeing-in: He has had them on looking at stained walls or at clouds and seeing in them fighting horsemen, or whales, or camels. But now we are to imagine that the thought of things that can, through such experiences, be seen in front of other things comes to guide the way he marks the surface. He now marks the surface, he forms motifs, so as to produce such experiences, so that, when the surface is looked at, something will be seen in front of something else. In doing so, he thematizes the image. It is important to recognize that thematizing the image is a step beyond thematizing the motif. For, though the front end of the deposited motif, or the edge closest to the agent, will indeed be raised above the surface on which it lies, this is far too fine a point for the agent to notice, at any rate standardly. No, what he sees in front of something else is something that he sees through, [as a result of] looking at the motif. And it is this which I call the image. (Wollheim 1987: 21)

Richard Wollheim on seeing-in


And Wollheim then proceeds to say that this more or less amounts to the discovery of representation. Among other things, this concept of an image obviously clears a debt incurred with the introduction of the ‘person’ (seen in a picture) in the 1980 example. But the main point to be made in connection with Wittgenstein is certainly the careful placing of the concept ‘image’ at a specific stage of a highly specific process, in which a certain perceptual skill, seeing-in, is being cultivated; this process can be described as a development and re-structuring of intentions and practices of a considerable degree of sophistication. There is no obstacle to calling this, given a sufficiently described context, a ‘visual image’; confusion with an illusory ‘mental image’ (inneres Bild), a replica of external images provided by a mysterious independent ‘faculty of the mind’, is excluded. It is instructive to compare these ideas with a different attempt to understand mental images in the light of more or less Wittgensteinian assumptions undertaken by Hide Ishiguro (Dilman and Ishiguro 1967: 50ff). There, the commitment to an independently existing ‘picture in the mind’ is avoided by stipulating that what we see in the mind’s eye is simply the object itself. This is a clarification on a purely conceptual, analytical level, whereas Wollheim offers a realistic description of the context which lends significance to the concept of a visual image. As a further consequence, Wollheim now has room for an independent explanation of imagination10 as such. And, indeed, perhaps the most fascinating aspect of his theory is contained in the third chapter, entitled ‘The spectator in the picture’ (Wollheim 1987), where, complementary to the introduction of the image but separate from it, he develops an ingenious conception of imagination: imagination as a process of distancing oneself, of freeing oneself in a sense from the restriction of perception, by entering the vision of another seeing subject. When seen against the background of its careful and cautious development over a very long time, Wollheim’s conception of seeing-in presents itself as truly Wittgensteinian in spirit – perhaps most so here, with his ideas about image and imagination, where it goes beyond the range of Wittgenstein’s own reflections.

Notes 1 Particularly with reference to Wollheim’s Painting as an Art (1987). 2 In the following, all quotations from Wittgenstein will refer to this edition of the Philosophical Investigations, part II, xi. 3 ‘And I must distinguish between the “continuous seeing” of an aspect and the “dawning” of an aspect. The picture might have been shown me and I never have seen anything but a rabbit in it’ (Wittgenstein 1958: 194). 4 Wollheim speaks of ‘the unacceptable consequences of analyzing perception in terms of schema and correction’ (1974b: 285). 5 To be sure, the ‘person’ asked for must not be identified with the person who NN, standing over there, is – it has to be looked for in the portrait. 6 That these are not pedantic questions can be learned from James Elkins (1998: 14ff). See also my later remarks on Michael Podro’s review of Art and Its Objects.

180 Richard Heinrich 7 In the sense of a possibly existing independent object (or, as in the case of a portrait, person), which is referred to by the Bildgegenstand. Recently, an elaborate theory of pictorial representation, based on these distinctions between configuration, picture-object and reference has been proposed (Pichler and Ubl 2014). 8 ‘Seeing-in permits unlimited simultaneous attention to what is seen and to the features of the medium. Seeing-as does not’ (1980: 212). 9 As when Leonardo describes his seeing dancers in a stained wall (Wollheim 1987: 40). 10 An explanation which does not take imagination as the cognitive faculty stipulated for the purpose of explaining the possibility of images (in the mind).

References Brugger, Peter 1999. ‘One Hundred Years of an Ambiguous Figure: Happy Birthday, Duck/Rabbit’. Perceptual and Motor Skills 89: 73–7. Budd, Malcolm 1987. ‘Wittgenstein on Seeing Aspects’. Mind 96: 1–17. Dilman, Ilham and Ishiguro, Hidé 1967. ‘Imagination’. Proceedings of the Aristotelian Society, Supplementary Volumes 41: 19–56. Elkins, James 1998. On Pictures and the Words That Fail Them. Cambridge, New York: Cambridge University Press. Gombrich, Ernst H. 1972. Art and Illusion. Bollingen Paperback Edition. Princeton: Princeton University Press. Jastrow, Joseph 1900. Fact and Fable in Psychology. Boston: Houghton Mifflin. Pichler, Wolfram and Ubl, Ralph 2014. Bildtheorie zur Einführung. Hamburg: Junius. Podro, Michael 1982. ‘Art and Its Objects by Richard Wollheim’. The Burlington Magazine 124: 100–2. — 1987. ‘Depiction and the Golden Calf ’. In Andrew Harrison (ed.), Philosophy and the Visual Arts. Dordrecht: Reidel, pp. 3–22. Wittgenstein, Ludwig 1958. Philosophical Investigations. 2nd edition, tr. by G.E.M. Anscombe. Oxford: Blackwell. Wollheim, Richard 1974a. ‘On Drawing an Object’. In On Art and the Mind. Cambridge, MA: Harvard University Press, pp. 3–30. — 1974b. ‘Reflections on “Art and Illusion”’. In On Art and the Mind. Cambridge, MA: Harvard University Press, pp. 261–89. — 1980. Art and Its Objects: With Six Supplementary Essays. 2nd edition. Cambridge, New York: Cambridge University Press. — 1987. Painting as an Art. London: Thames and Hudson.

Part IV

Rescuing Wollheim’s account without the support of Wittgenstein

This page intentionally left blank


A measure of Kant seen in Wollheim Gary Kemp

It is tempting to think that Wittgenstein’s remarks about aspect-perception – seeing-as and seeing-in – were intended somehow to support a certain hypothesis that Kant so impressively advanced: that the objects of normal perception – perception that is capable of underlying everyday empirical judgement – are experienced as inherently permeated by thought. By citing examples such as the duck–rabbit, which seem to show that there is a conceptual underpinning of perception that operates beneath the threshold of awareness, it might be supposed that Wittgenstein shows the psychological inevitability of that phenomenon, if not its necessity in the Kantian sense. But this runs manifestly against Wittgenstein’s antipathy towards philosophical theory, analysis and generalisation; I am persuaded by Avner Baz, Charles Travis, Joachim Schulte and Hanjo Glock (chapters in this volume) that this would be entirely to misread Wittgenstein. Richard Wollheim, however, despite his occasional remarks to the contrary, is best seen as borrowing not from Wittgenstein in formulating his theory of pictorial depiction, but, at least implicitly and loosely, from Kant. By emphasising its Kantian affinities, then, what follows is an attempt to extricate Wollheim’s basic position on pictorial depiction from certain recent challenges.1 I shall begin with various difficulties with Wollheim’s preference of seeing-in over seeing-as in his account of the experience of pictorial seeing.2 I think that Wollheim’s preference was indeed a mistake. But contrary to what Wollheim himself held, I urge that his notion of ‘two-folded perception’ – which he held to be essential to pictorial representation – does not provide anything like a decisive reason for the preference. I then shift momentarily to the account advanced by Robert Hopkins. In a forceful exchange with Wollheim, Hopkins has argued for his account of pictorial representation in terms of perceived resemblance: x pictures the object y only if a suitable spectator sees y in x; but that in turn is to be analysed as the spectator’s perceiving x as resembling y.3 I argue that seeing-as and seeing-in are psychologically more basic than Hopkins’ notion of subjective resemblance, and that it is more plausible to regard them as jointly at the root of pictorial representation than it is Hopkins’ notion. Then I consider some remarks of Wollheim’s aimed at substantiating his view of the distinction between seeing-as and seeing-in.

184 Gary Kemp His attempt to conceive seeing-in as operating at, so to speak, a psychologically higher level than seeing-as is incompatible with the compelling and relatively modest Kantian thesis that awareness of states-of-affairs must be conceived as complementary to the awareness of objects; there is no one without the other. In closing I use the revised Wollheimian account to address the problems of ‘integration’ and ‘inflection’ – the fact that pictorial experience is not disjointed in the way apparently entailed by Wollheim’s theory and the fact that one’s awareness of certain details of pictorial content cannot be identified independently of both aspects of two-folded perception. In passing I will offer some brief remarks on the way this use of the terms ‘seeing-in’ and ‘seeing-as’ departs from Wittgenstein’s own use of them.

Preliminaries If one sees y in x, or sees x as y, x can be the surface, or part of the surface, of a drawn, etched or painted picture. But x also can be a cloud or the clouds; a mountain or other feature of the landscape; a tangle of grasses, branches or trees; a swarm of insects, fish or birds; a rock, a fire or a flame; the surface of a pond, a river, the sea or a dried lake-bed; staining in the plaster, a person’s hair . . . virtually anything one can see. For clarity and ease in what follows, I shall sometimes speak of the primary object – x, that which is seen as y or in which one can see y – as opposed to the secondary object – y, that which x is seen as, or that which is seen in x. Sometimes I will find it convenient to refer to seeing-in and seeing-as indifferently as varieties of aspect-perception. The question of what sorts of entity are fit to be the secondary objects is by comparison much more complex and contentious. The following seem at first blush to be acceptable examples of seeing-as (leaving tacit the presence of the seer): A: Seeing-as 1 2 3 4

Seeing a cloud as a camel. Seeing the clouds as Snoopy. Seeing a landscape as gloomy. Seeing the cornice as a woman wearing a ceremonial headdress.

The following seem at first blush to be acceptable examples of seeing-in: B: Seeing-in 1 2 3 4

Seeing a camel in the clouds. Seeing Snoopy in the clouds. Seeing gloominess in a landscape. Seeing a woman wearing a ceremonial headdress in the cornice.

A measure of Kant seen in Wollheim


I shall assume that both A4 and B4 indicate a state-of-affairs (type or tokens) as the secondary object. Thus we seem to have as secondary objects (1) kinds or types, (2) particulars or individuals, (3) properties and (4) states-of-affairs or ways-that-things-could-be. (In the formal as opposed the material mode of speech we have (1) indefinite descriptions, (2) singular terms, (3) open sentences and (4) sentences or nominalisations of same.) Wollheim typically discusses properties as secondary objects only implicitly in what he has to say about the depiction of states-of-affairs; I will follow him in this. Events and processes are interestingly capable of appearing as both objects and states-ofaffairs; but I will assume that they do not constitute a separate category. Most if not all the fun is in the contrast between seeing ordinary objects and seeing states-of-affairs. Differences between seeing-as and seeing-in appear sometimes to be highly nuanced, and the view for which I will be arguing has it that the two are indeed psychologically very close – closer than Wollheim thinks.4 One might suppose that there is no more to the distinction than that seeing-as is more specific or explanatory than seeing-in – for example, one might say something like ‘I see an elephant in the clouds because I see the bit on the left as its trunk’ – or that it is merely grammatical, just a case of swapping grammatical places, as in the distinction between the active and passive voice.5 As we will ultimately come to see, the former contains some truth, the latter not; in any case it is by no means clear that Wollheim’s way of making out the distinction is successful, as I shall now argue.

Wollheim In his supplementary essay on pictorial representation, written after the first edition of Art and Its Objects, Wollheim claims three differences between the concept of seeing-in and the concept of seeing-as, differences from which he infers that the former and not the latter is at the perceptual heart of pictorial representation. This might strike one immediately as odd, for then Wollheim will have either to deny that the skill of the boy who says of a cloud that it is a camel is exercising the same skill as when he identifies a camel as the subject of a drawing, or, say, somewhat idiosyncratically, that what is really going on in the cloud case is not seeing-as, or not merely seeing-as, but that the boy sees a camel in the clouds. I will raise some initial doubts, suggesting that the matter is messier than is required for Wollheim’s reasoning, but then argue that the essentials of his account of pictorial representation do not require seeing-in to the exclusion of seeing-as. First, Wollheim holds that whereas in the case of seeing-as the secondary object must be a particular (as in Charles V) or a kind or type (as in a soldier), the secondary object in the case of seeing-in may be a state-of-affairs. Unlike seeing-in, seeing-as is restricted to types 1 and 2. One sees, for example, in a picture that the barmaid is preoccupied (I take it that the context

186 Gary Kemp created by ‘ . . . sees in . . . that . . . ’ is unlike that created by ‘ . . . sees that . . . ’ in that it is not factive, enabling one knowingly to describe non-existent states-of-affairs). But no item is seen as the barmaid’s being preoccupied. Seeing-in thus commands a much richer range of content than seeing-as; ‘scenes’, writes Wollheim, ‘may be represented as well as persons and things’ (1980: 211). What then of the prima facie examples A3 and A4? It is tempting to accept Wollheim’s thought that one cannot simply transform without further ado an example of seeing-in a case of a picture to its corresponding seeing-as form: one can see in the picture a soldier’s being angry, but one cannot see the picture as a soldier’s being angry – rather, one can see the picture as depicting a soldier’s being angry, or a soldier depicted in the picture as angry, both of which would render the analysis of depiction circular (Wollheim 1980: 210–11). But one can simply resist Wollheim’s thought. If one sees in a picture that a soldier is angry, or a soldier’s being angry, what exactly is to stop the transition to ‘I see it as a soldier’s being angry’ or ‘I see it as an angry soldier’, as opposed to ‘I see it as depicting a soldier’s being angry’ or ‘I see it as depicting an angry soldier’ – or the halfway houses ‘I see it as a soldier’s being depicted as angry’ or ‘I see it as a soldier depicted as angry’ (or even ‘I see it as the anger of a soldier’ or ‘I see it as a soldier’s anger’ instead of ‘I see it as depicting the anger of a soldier’)? One may feel a tug either way, but I am not sure that the grammar of English is determinate on the issue. And, for what it is worth, to convey Wollheim’s generic idea that pictures depict scenes, it seems indifferent whether we say ‘pictures are seen as scenes’ or ‘We can see scenes in pictures’ (1980: 211). Can Wollheim not appeal beyond these broadly linguistic or grammatical considerations to the phenomenology of seeing objects versus that of seeing states-of-affairs? In fact his take on the relevant phenomenology tends to work against what he says on this issue. Central to his account of pictorial meaning in general is the phenomenon of prompting: that sometimes one cannot see a certain state-of-affairs in a picture without having information of various kinds, or without being coached, does not imply that the thing is not, strictly speaking, seen, once one is able to bring the relevant information to bear (Wollheim 2001: 24; see also 1993: 132–43). What this shows is the ‘permeability to thought’ of the ‘central phenomenological feature’ of seeing-in: such cases often require that the spectator be prompted by thought – meaning that the relation is causal, rather than conceptual – where the content is nonetheless perceptual. But this applies also to seeing-as. Prompting can bring it about that we see certain marks as a white cross floating above a black background, or others as a frightened stag. The phenomenon of prompting makes it more difficult to resist accepting that, say, the tempestuous future that lies in wait for the infant Christ in Bellini’s Madonna of the Meadow is indeed an object suitable for seeing-as as well as seeing-in. One can see the smudge of paint as Icarus as readily as one can see him in the Bruegel – indeed, perhaps given coaching, one can see the smudge as Icarus,

A measure of Kant seen in Wollheim 187 who has been flying with wings made of feathers and wax, but who flew too close to the sun, which melted the wax, causing him to fall into the sea. One naturally thinks that there must be limits to this – that the relative clause in that sentence does not plausibly pertain to perceptual content, that not even in an extended sense is it something seen (Wollheim 2001: 23–4; 2003: 10–14). But it is much less certain that that response addresses A3 and A4 as well as the angry soldier, the stag and cross examples; they call for a more heavy-duty theory of perceptual content to adjudicate than anything in play here. Wollheim does offer something more to the purpose, but I shall defer discussion of it to the final section of this chapter. Second is Wollheim’s localisation requirement. Whereas for seeing-as it is always possible to answer the question of which parts of the primary object correspond to which parts of the secondary object – one may point to the parts of a drawing of a dog that one sees as the dog’s ears, for example – there is no such demand in the case of seeing-in. The dreariness of the weather, for example, need not show up in any particular feature of a picture, and hasn’t parts in the requisite sense. But – as Wollheim recognises – seeing-in does often permit localisation, in the sense that if I see a dog in the picture, the question ‘Where do you see the dog’s ears?’ typically has an answer. Indeed, if we restrict the question to type-1 and type-2, it looks as if an opponent of Wollheim could insist that the localisation requirement holds for both seeing-as and seeing-in, for he can insist on re-construing the apparent type-1 or type-2 counterexamples as type-3 or type-4: ‘the dreariness of the weather’ becomes ‘the weather was dreary’ and so on. Do seeing-as and seeing-in fare differently with respect to the localisation requirement for type-3 and type-4? An opponent of Wollheim could point out that for at least many statements of seeing-in, there is an equivalent statement of seeing-as and vice versa: for example ‘sees in the picture that Icarus is F’ becomes ‘sees the picture as Icarus being F’ and vice versa. Either both are subject to the localisation requirement or neither is, the reply runs. Third, in Wollheim’s view – and this he regards as crucial – whereas seeingas prohibits simultaneous attention to primary and secondary objects, seeing-in permits it. But – and this is crucial for the alternative that I shall be arguing for – I think that in the sense in which seeing-in permits simultaneous attention, seeing-as does so as well. Of course, as Gombrich (1960: 5) stressed, simultaneous attention to mutually exclusive secondary objects is impossible (one either sees the duck or sees the rabbit but not both at the same time; cf. Wollheim 1972: 280). But it is the essence of aspect-perception that we are able, for example, to see the cloud as a camel, but at no point interrupt our perception of the cloud (Gombrich 1960: 278–9 famously missed this point). So none of these points, individually or collectively, serve cleanly and categorically to make the case that seeing-in is more suitable for use in a theory of pictorial depiction than seeing-as; the remaining difference under the second point is too thin.

188 Gary Kemp Now I said above that the failure decisively to distinguish seeing-as from seeing-in as he wants would not be detrimental to Wollheim’s account of pictorial representation. The central pillar of Wollheim’s account of pictorial representation is the concept of two-foldedness. In seeing a representation as a representation, we have a single experience with two aspects: we see the picture-object – a certain configuration of charcoal or paint, etc. on a surface – and simultaneously we are visually aware of – do not merely infer or have the thought of – another object, the represented object. The capacity for perceiving in a two-folded manner, Wollheim holds, is a real perceptual phenomenon, one to be explored in psychology or phenomenology, not linguistics or the philosophy of language. Indeed Wollheim thinks our capacity for two-folded perception as incapable of phenomenological analysis, that it is bedrock phenomenologically: it ‘is a distinct kind of perception . . . theorists of representation consistently overlook or reduce this phenomenology with the result that they garble representation’ (1987: 46); we must not expect . . . a description from which someone who never had the experience could learn what it would be like to do so . . . the philosophical point of phenomenological description . . . [is] to see how some particular experience can, in virtue of what it is like, do what it does. (2001: 20; more on this later, but for acute challenges to it see Budd 1992)6 Two-foldedness is necessary but not sufficient for representation. In addition, a picture represents a given object only if the maker of the picture has their intention fulfilled – perhaps in an implicit or elastic sense of ‘intention’ – that the secondary object be seen in the primary object – the marked paper, for example – by the intended spectators. So a picture represents y if and only if the maker successfully intends the envisaged audience to have an experience of two-folded perception with the picture as the primary object and y as the secondary object. Call this the foundational account. The foundational account makes no mention of seeing-in or seeing-as. I suggest we can retain the foundational account, and regard both seeing-in and seeing-as as necessary for the further explication of two-foldedness. Crucially, both – not just seeing-in, as Wollheim thought – allow for twofolded perception. Whatever else may be true of them, they are surely very closely related if not identical phenomenologically. I will say a lot more to substantiate their connection and their difference below; for the moment we can just observe that, this way, we retain all the advantages of either.

Hopkins Hopkins is a prime advocate of the (subjective) resemblance theory of pictorial depiction (others are Budd 1992, 2008: chapter 2; Peacocke 1987; for the distinction between subjective and objective resemblance see Newall 2010: 92).

A measure of Kant seen in Wollheim 189 His difference with Wollheim centres on his particular account of seeing-in. He agrees that it is the key to pictorial representation, but thinks that it admits of substantive elucidation – indeed, in a sense, of outright analysis: it is an experience of visual resemblance, where this is the seeing of the marks composing the picture as resembling the depicted object (in point of ‘visible figure’ or ‘outline shape’). So seeing-in is a special case of seeing-as. I think that Hopkins is wrong about this, but interestingly so. I shall begin by rehearsing and commenting upon parts of Hopkins’ exchange with Wollheim in the Proceedings of the Aristotelian Society, before advancing another way of looking at the issue that favours Wollheim. There are two particular areas of dispute between Hopkins and Wollheim. Both concern Hopkins’ assimilation of seeing-in to seeing the depicting object as resembling the depicted. Wollheim first objects to Hopkins’ assimilation of seeing-in to seeing-as by pointing out an example which fails the localisation requirement on seeing-as: the aftermath of a storm in a Turner. We can see the aftermath in the picture, but we cannot see it anywhere in particular in the picture. Hopkins replies that this does not show that the object has no location. Consider the similar case where one is looking face-to-face at the aftermath of a storm. ‘If your visual field expanded considerably’, Hopkins writes, ‘by your changing your location, or lifting some blinker to reveal more of the surroundings, then the aftermath might lie somewhere in particular in the visual field’ (Wollheim and Hopkins 2003: 154). So saying ‘everywhere’ to the location question is not vacuous or uninformative. But what sort of thing, one may ask, is an aftermath? Is the example one of type-1 or 2, or is it really of type-3 or 4, as discussed above? If forced to choose, I would say that it is an object – type-1 or type-2 – only in a grammatical sense; it is really a type of state-of-affairs. We can give the appearance of speaking of many such entities via nominalisation of predicates, as in ‘The coming of the storm’ via ‘The storm is coming’, or ‘The goodness of Theresa’ via ‘Theresa is good’. Although ‘aftermath’ is not a nominalisation of any extant predicate, we could, if we were so minded, invent an (ugly) adjective – ‘aftermathy’ – giving us such predicates as ‘ . . . is storm-aftermathy’, enabling us to transform certain examples of type-1 or type-2 to type-3 or type-4. A comparison might be made with the case of ‘sake’ as in ‘for his sake’, for which there are only the vaguest of identity criteria and no serious prospect of entities corresponding to the word in the domain of quantification. It is most unlikely that the language stripped of corresponding nouns would be impoverished. So I submit that the Turner case is really type-3 or type-4. Hopkins’ view is not just a view about type-1 and type-2 cases but must also cover type-3 and type-4. Hopkins’ view struggles with types 3 and 4. Consider the spiritual desolation of van Gogh’s Night Café (1888) or the complicated state-of-affairs of Salome’s sexual exultation over the freshly severed head of John the Baptist in Cranach the Elder’s picture (the one thought to have been painted in 1530). True, we can point to things in these pictures to which their representational

190 Gary Kemp powers are due, but we cannot correlate parts of each canvas with parts of their represented objects as the localisation requirement demands; neither properties nor states-of-affairs, and especially not ones of such complexity as these, admit of decomposition into spatial parts. It seems right and natural to say we see those states-of-affairs in the pictures, but not that we see the elements of the pictures or the pictures themselves as resembling those state-of-affairs. They are not the sort of things to be resembled by anything. The second area of dispute concerns Wollheim’s characterisation of Hopkins’ view that experienced resemblance is sufficient for seeing-in. Wollheim supposes that Hopkins is committed – implausibly – to its being sufficient for experienced resemblance that one (1) see the primary object, (2) think of the secondary object and (3) know the appearance of the secondary object. Hopkins replies by simply denying that his view is committed to (1)–(3) being sufficient for experienced resemblance (Wollheim and Hopkins 2003: 155). (1)–(3) serve partially to explain or explicate experienced resemblance, but experienced resemblance is perhaps too primitive to admit of outright analysis, of philosophical definition. ‘Its ability to enlighten stems . . . from its being a phenomenon with which we are all familiar’, writes Hopkins (ibid: 156). But there are cases which challenge the necessity of Hopkins’ conditions (1)–(3). In particular, depictions where no one has anything but a vague idea what the subject looks like seem to be failures of condition (3). Pythagoras is thought to be the red-sleeved figure at the front left of Raphael’s School of Athens. Suppose he is. Neither Raphael nor any person contemporary to Raphael or after him could have known what Pythagoras looked like, in which case it can’t be a constraint on the viewer’s experience that they experience the marked region as resembling Pythagoras. Yet all the same the picture depicts him; it is a picture of Pythagoras, however infelicitous it may in fact be. Hopkins does offer some relief on the point, saying that what matters is really the ‘object as it is depicted as being’ (ibid: 159). But still condition (3) is vacuous or nearly so in this case; one wants to say that resemblance of the patch of paint that represents the figure to the subject – or the object as depicted – has little or nothing to do with the fact that the picture depicts Pythagoras. Instead, the viewer must rely on other facts, some of which will involve what is elsewhere visible in the picture, and others not. What Hopkins evidently has to say is that, just as Wollheim has to allow that such knowledge may legitimately prompt the viewer as to the sort of thing that may be seen in the patch of paint, so it is with his own view, except that his theory goes on to characterise the experience as one of resemblance-as. There is as far as I can see nothing inconsistent in that reply, but it now looks as if the appeal to perceived resemblance is in danger of being a mere free-spinning wheel, an optional add-on. I shall mention one final problem with Hopkins’ answer to Wollheim. It was raised by Wollheim himself (Wollheim and Hopkins 2003: 140–1) but is not discussed by Hopkins as far I know. It concerns the objects normally compared when we say of one that it resembles, is similar to or looks like another.

A measure of Kant seen in Wollheim 191 The contrast between ‘I see y in x’ and the grammatically more complex ‘I see x as resembling y’ is brought to the fore. If one says of a certain man that he resembles Napoleon, one is saying that that man resembles another man; in the normal sort of case, one may so express one’s experience as the one resembling the other. If one says of a photograph of a man that it resembles Napoleon, one expresses one’s experience of the man resembling another man. Likewise, it seems, if one says it of a pictorial depiction of a man. But then the fact that the surface depicts or represents the man is presupposed; the account is circular. Now a likely reply is that the objection simply misunderstands the view; in the formula x resembles y, x is not an object depicted but a region of the picture – a patch of paint, for example – which is experienced as bearing a similarity to the outline shape of y. But then, that is surely not, perhaps despite Hopkins’ protestations, a matter of ordinary experience: it is a matter of its interpretation, a matter of what the philosopher says upon reflecting on its phenomenology. One can be understood by a five-year-old if one asks ‘What do you see in the picture?’, whereas presumably one would not be if one asked ‘To what do you see this region of the picture as being similar in shape?’; it is a strength of Wollheim’s theory that it stays so close to common sense. I don’t claim, of any of these objections or quibbles with Hopkins’ theory, that Hopkins cannot answer it. What I will claim is that a variant of Wollheim’s theory, formulated partly in response to these difficulties, is a cleaner, more compelling theory.

Aspects and schemata I claim that the notion of perceived resemblance is  not  a kind of seeing-as, at least not any more than any visible quality F is a kind of seeing-as (because we can always form the phrase of the form ‘sees as F’: ‘sees as blue’, ‘sees as silky’ and so on). One naturally assumes that the phrase ‘looks like’ shows that we have here only one phenomenon; but – so I claim – not only are the notions conceptually separable; pertaining to distinct phenomena, operating at different cognitive levels, they are not materially equivalent: one can see the stones one has collected at the beach, or two persons, as resembling, or looking like, but one does not thereby see one as the other. A clue is the subtly different logic of the two relational expressions: they differ in whether or not they are graded (a), and over transitivity (b) and symmetry (c). (a) As Jerrold Levinson (1998: 223) points out, perceived resemblance is a matter of degree, but seeing-as is categorical or discrete, like throwing a switch, either on or off: just as one can be more or less tall or handsome, but one cannot be more or less a mother, one can experience resemblance in various degrees, but one either sees the famous drawing as a duck or one does not. The experience of exact resemblance is merely the asymptote of the experience of degrees of resemblance or similarity. One’s experience when

192 Gary Kemp looking at an emu might be that its appearance is roughly similar to that of an ostrich; that of a jackdaw closely similar to that of a rook; and that of a crow exactly similar to that of another. The discreteness of the experience of aspects is especially evident in the case of experiences of puzzle-pictures, where suddenly the secondary object leaps out of the lines – the aspect ‘dawns’, as Wittgenstein memorably put it. (b) Perceived resemblance or perceived exact resemblance is not a fullblooded transitive relation. But it comes close. Suppose that we are speaking only of particular objects x, y and z, not of kinds of object. If one perceives x as exactly resembling y (in outline shape), and if one perceives y as exactly resembling z, then one can perceive x as closely (perhaps exactly) resembling z. If one perceives x as closely resembling y, and y as closely resembling z, then one can perceive x as resembling (perhaps exactly or closely) z. In general, let Rxy be the degree of perceived resemblance of x and y (where 0 ≤ Rxy ≤ 1). Then – at least plausibly – if for some values μ and ν, Rab = μ and Rbc = ν, then Rac ~ > μν. The resemblance of x with z is roughly at least the product of those of x to y and y to z. There is, I take it, no such temptation to think that seeing-as or seeing-in – whether the secondary is singular or general – exhibits anything like such transitivity or near-transitivity. (c) Again, we speak only of particular objects, not kinds. If one perceives x as resembling y, then one can perceive y as resembling x. Seeing-as exhibits no such symmetry. If one sees a cloud as Fido, it is quite rare that one can see Fido as a cloud. What accounts for these differences? It is plausible that they are connected to their different though closely related roles in the human cogito-perceptual apparatus. Now Wittgenstein did not take the seeing of an aspect to be our usual way of perceiving things – that, for example, when I see a chair and recognise it in the usual way as a chair, I am seeing it as a chair.7 As I said at the beginning, I am convinced that this was not Wittgenstein’s view; but for reasons which I will get to below I think it is open to employ both concepts of aspect perception.8 Suppose I look across the river on a midday walk and see what looks for all the world like a curled-up fox. Yet it is surely too exposed, I think further, for a fox to be sleeping in broad daylight. I conclude that it is not a fox but a tree stump. I try seeing it as a fox and am successful; I try seeing it as a stump and am equally successful. My judgement is taken on other grounds, having to do with knowledge of the habits of foxes and so on. In fact, it was a fox, as I was surprised to learn a moment later when I saw it move. Now I am back to seeing a fox as a fox. What is the difference between that state and the state a moment earlier, when I was trying out the hypothesis that what I was looking at was a fox? An obvious difference is that in the later state there is something such that I judged it to be a fox. Not so in the earlier state, when it was a mere hypothesis, a case of withholding assent. This difference is not captured in

A measure of Kant seen in Wollheim


terms of aspect-perception; I was seeing the object as a fox in both cases, indeed my perceptual state is the same in the two cases – so, at any rate, it seems to me. Let us call the two states non-judgemental seeing-as, and judgemental seeing-as. Now of course ordinary perception is not two-folded perception. When one sees a rabbit as a rabbit, one is not seeing a rabbit in something else. If we use the earlier nomenclature, it is the case when the primary object is identical to the secondary object. Nor is a case of non-judgemental seeing-as thereby a case of two-folded perception: for example, in the fox case it is not possible to fixate on the primary object and entertain hypotheses about the secondary object that are not about the primary object, for the two are really one; there is nothing analogous to the clouds or the lines-on-the-page, no room for the distinction between primary object and secondary object. All the same, it is a case of non-judgemental aspect-perception, as is evinced by the possibility of switching aspects. And the fox case can become a case of two-folded perception if, once I learn that the object is a fox, I self-consciously undertake to see what I now know to be a fox as a stump. More generally, two-folded perception – the battle-scenes-in-the-plaster, the cloud-as-rabbit, a drawing of a rabbit – is what we can call negative judgemental aspect-perception: there is an x such that I see x as Φ, but I judge x to be not-Φ. Positive judgemental aspect-perception is the normal case: there is an x such that I see x as Φ, and I judge x to be Φ.9 Judgemental aspect-perception thus comprises the negative and positive variety, and either variety can be veridical or non-veridical as the case may be. The reason we do tend to talk of aspect-perception (seeing-in, seeing-as) only in cases like the camel-in-the-clouds or the duck–rabbit figure, and not in ordinary cases of perception, is that only in the former sort of case is the phenomenon of aspect-perception brought to the fore. It is only then that we are made aware of its being subject, within limits, to the will, and of the potential for perceptual error. These cases interest Wittgenstein, but they are only a tiny sub-class of examples of the phenomenon of aspect-perception. Normally it operates silently and automatically.10 Wittgenstein says of one who says of pieces of cutlery that ‘I am seeing it as a knife and fork’ (PI Fragment 6 p. 195e) that one would not be understood. We can interpret the point as concerning pragmatics, the propriety of speechacts. Saying such a thing, in Grice’s terms, implicates that one would not assert ‘I see a knife and fork’; that if one makes the weaker claim it is because one is not in a position to make the stronger claim. It would violate conversational maxims (in particular, of Quality). Nevertheless, the weaker claim does not entail the falsity of the stronger claim; it is not ruled out that ‘I see a knife and a fork’ is true. According to this picture, aspects enter into perception at a basic level, upstream from the activities of thought and judgement (although presumably they cannot be entirely encapsulated from them, immune to their effects). It happens beneath the threshold of consciousness. The aspect is given to the

194 Gary Kemp perceiver as part of the visual field, not as something super-added to it without changing the field’s original identity or character. Aspects play the role of Kant’s schemata: an aspect is a bridge between the brute sensory impression and the understanding; they are part of what enables us to think about empirical objects.11 The act of organising the visual field by means of an aspect is in Kant-speak spontaneous. They are proto-conceptual, not fullblooded concepts: they play a role in the recognition of objects or kinds of objects, but they do not themselves supply criteria of individuation of objects. And they are only partly descriptive of the visual field: if one sees a face in the rocks and remarks ‘A face!’, that term and not others expresses what I see but only part of it; the rest has to be supplied or ‘filled in’ by the hearer, in perceiving the face (Isenberg 1949). This Kantian reading of the notion of aspect-perception is not Wittgenstein’s. It rules out of court the conceivability of aspect-blindness and rules in the phenomenon of continuous aspect-perception, both evidently contrary to Wittgenstein’s view. But I do not think the two concepts are incompatible. Let us temporarily call the two seeing-asw and seeing-ask. Whereas cases of seeing-asw are comparatively rare – the camel in the clouds, etc. – seeing-ask is perfectly normal: whenever one sees something, there is almost always some concept or other, and some item, such that one sees the item ask an instance of that concept (at least subliminally). Unlike seeing-ask, seeing-asw is irreflexive: one can see a duck ask a duck, but one cannot see a duck asw a duck. The revision to Wollheim issues in a flat material difference. Wollheim supposes that since one does not undergo two-folded perception in the case of trompe de l’œil pictures, they are not pictorial representations in his sense (Wollheim 1987: 62). The position comes under added strain from the evident fact that the distinction between the trompe de l’œil and ordinary pictures is matter of degree, not of kind (Lopes 1996: 50; Hyman 2006: 132; Gaiger 2008: 61): at one end we have perfect trompe de l’œil pictures, at the other we have stick-figure drawings and the like. Wollheim’s response is closed off to the revised position, but I think the revision does better. Consider one’s looking at the Evert Collier picture Letter Rack, where one is not taken in by it. As was implied above by the discussion of the fox, the difference between that state and one’s perception of an ordinary picture is not that, unlike the latter, it fails to be a case of twofolded perception. In both cases one is in a state of negative judgemental aspect-perception. The main difference is that one knows that if the surroundings were different – say, a room with real newspapers, combs and letter openers mounted on the walls – then, since the pictured objects are of the right scale, one would be deceived: it would be a case of positive (erroneous) judgemental aspect-perception. For some pictures other than the trompe de l’œil, what prevents such illusions is similar: the absolute stillness of people, the pictures being hung on a wall (giving improbable glimpses of Venice), matters of scale (usually too small), matters of three-dimensional perspective (it is impossible to see round an object in the foreground by

A measure of Kant seen in Wollheim


moving to the side) and so on. The truth is simply that it is very difficult, perhaps impossible, in certain cases to see, at the expected viewing distance, the picture as achieved by paint, even if one knows it is achieved by paint. Even if one is not deceived, one is ‘baffled’ by the picture, as Wollheim says (1980: 217). But the right conclusion is that one should not assume a too narrow conception of what awareness of a configuration of marks can consist in; looking at a photo-realist picture of some pebbles, one can describe some of the picture’s configurational properties – one can speak of a round patch of sepia paint there, surrounded by grey, mottled with black and white, and so on, even if one relies primarily on background knowledge that what one is looking at is paint (Gaiger 2008: 58–62; cf. Levinson 1998: 228–9). In the case of the maximally successful trompe de l’œil, one’s awareness of the configurational properties is schematic and minimal but not absent. One is not forced to say that because it is not a case of two-folded perception it fails to be a pictorial representation. Again, it is a case of negative judgemental aspect-perception, hence of two-foldedness.

Further ideas from Wollheim The above section makes little of the distinction between seeing-in and seeing-as. In the section ‘Wollheim’ above I considered Wollheim’s main reasons for favouring one over the other; now, to articulate further the revision that I am advancing, I consider some further attempts by Wollheim to distinguish them.12 Wollheim generally agrees with view I have just advanced about seeing-as (1980: 219–20) but argues that seeing-in stands apart, that it derives from a special perceptual capacity, which presupposes, but is something over and above, straightforward perception . . . In order to attain to seeing-in a crucial development has to occur . . . visions of things not present now come about through looking at things present. (ibid: 217–18) He first considers dreams, daydreams and hallucinations, which are the ‘most primitive experiences’ (ibid: 217) with which seeing-in is ‘plausibly connected’. The question of what I am seeing as a face is an expression of curiosity about an object, whereas the question of what I can see in a forest is more a question about me, about my psychology or imagination. But given the way I have characterised seeing-as, those primitive experiences are equally connected with seeing-as. The famous example of Leonardo’s advising the aspirant painter to look at damp-stained walls to see battle scenes is plausibly a case in which Leonardo means to cultivate the pupil’s ability to free up his capacity for negative judgemental aspect-perception – which incorporates both seeing-in and seeing-as.

196 Gary Kemp Wollheim also advances a phenomenological explanation of why it is that, unlike seeing-in, seeing-as is incompatible with two-folded perception. A crucial passage is this: there is no room when seeing x as counterfactual F also to be visually aware of those properties of x which would have to change if x were actually to be or become F. In other words, the properties that sustain my perception of x as F would have themselves to be perceptually masked if I am – as I have put it – to try out the appearance of F upon x. So twofoldness in the case of seeing-as is ruled out. (1980: 223) But suppose I am seeing a cloud as a camel. Can I not be visually aware of those properties of the cloud which are incompatible with its actually being a camel, such as the cloud’s vaporousness? The vaporousness of the cloud plays no positive role in my perceiving it as a camel, but it is difficult to see how its being vaporous suddenly becomes invisible, or un-seen or perceptually masked. What is true is just that that property is discounted – into the penumbra of attention – as not being of special interest for present purposes. Properties that do play a central role in the project – primarily the shape of the cloud – are just those that interest me in the project, that sustain two-foldedness; but they are not the only ones of which I am visually aware. Attention is not exclusivity. The terms ‘seeing-as’ and ‘seeing-in’ do indicate distinct phenomena, even if the way in which the grammar of those expressions tracks the phenomena is messy, as indicated in the section ‘Preliminaries’ above. Of course in practice we do not normally speak of seeing-in in the case of normal perception of a state-of-affairs, a case of veridical, positive judgemental aspect-perception; rather than employing that locution, we just say we see that so-and-so. We do reserve seeing-in for cases where we exercise our capacity for non-judgemental aspect-perception. But, as before, the gap between our use of the locutions seeing-in and seeing, according to this line, is readily bridged by Gricean means. If we wanted to clean up ordinary language for theoretical purposes in the manner of Carnap, then we would restrict ‘seeing-as’ to types 1 and 2 – appertaining to individuals and types of individuals – and we would restrict ‘seeing-in’ to type-4 – states-of-affairs (I am not sure where to put type-3, properties). Then I can put my difference with Wollheim as that he was right to separate the aspectual character of perceiving objects from perceiving states-of-affairs – seeing-as as explicated from seeing-in as explicated – but wrong to suppose that the two operate at distinct psychological levels. The reason is that, in either way of speaking, awareness of states-ofaffairs, according to Wollheim, must take place only optionally and at a different level than awareness of objects. This violates the second component of the Kantian dictum that thoughts without intuitions are empty, intuitions without concepts are blind (B75); of course it is more plausible than

A measure of Kant seen in Wollheim


it is easy to analyse, but I take it as compelling that there is no conscious awareness of objects without some representation of how their parts are spatially configured and of the properties they are experienced as having – if not always of the larger states-of-affairs in which they are experienced as embedded. At the minimum, an object of experience must have some location, some colour, some shape, some size (or relative size). Thus the capacity for seeing-in is best conceived as a complement of the capacity for seeing-as: both operate at the same primordial level of perception – indeed they are psychologically indissoluble – but they pertain to different levels or categories of feature. Whereas seeing-as concerns particular elements of the field, understood as singular, seeing-in pertains to the field as whole, or to a significant portion of it that is understood as complex. The capacity for two-folded seeing is then explicable as a further downstream or higher-order process adverting to judgement, as explained above.

Integration and inflection By these words I mean another pair of challenges to a theory of depiction; the first is problematic for Wollheim’s own account, but I believe that both are got round by this revised Wollheimian account. John Kulvicki, in his piece ‘Heavenly Sight and the Nature of Seeing-In’ (2009), points to a place where Wollheim’s view seems to fall short. In a late expression of the view, in which he discusses Manet’s Émilie Ambre, Wollheim writes: When I look at the Manet, my perception is twofold in that I simultaneously am visually aware of the marked surface and experience something in front of, or behind, something else – in this case, a woman in a hat standing in front of a clump of trees. These are two aspects of a single experience. (2003: 3) Exactly what relation between the aspects or ‘folds’ did Wollheim intend? What relation did he envisage the experience comprises between an awareness of the surface of the marked canvas – the ‘configurational’ content – and simultaneous awareness of its three-dimensional ‘recognitional’ content? Of course, that one is having them both is a relation – a relation of having them both in a single perceptual experience – but is there nothing more to be said about it? Wollheim was apparently committed to its not being experienced as a spatial relation: he seemed to think of the experience as including two spaces which are entirely disjoint (the woman is in front of the trees, not in front of the tree-depicting part of the canvas13). But if it is a matter of the experience of two spaces which do not stand in a spatial relation to each other, then Wollheim cannot find room for the evident reality of the integration of pictorial experience: that particular features of the experience of the

198 Gary Kemp representational space substantially depend on particular features of the experience of the configurational space, as if one system of spatial coordinates rested on another (Kulvicki terms this ‘inflection’, but I shall reserve the term for a slightly different phenomenon described in detail by Podro (1998), discussed below). As a clue, Kulvicki describes the experience of seeing a scene through a window, which at the same time partially reflects a different scene: in the same direction, one sees two scenes which are spatially incompatible. Normally this happens only when the reflected scene is perceived as at least partly transparent. But he then cites empirical (though perhaps anomalous) evidence for the possibility of experiencing two scenes when neither is transparent, and concludes: ‘The proposal is that seeing-in is a perceptual state in which an opaque object is experienced as being in front of another opaque object even though neither object is obscured by the other’ (Kulvicki 2009: 394). There is no doubt that Wollheim’s remarks on the problem of integration are unsatisfying. But whatever the empirical support for the reality of the sort of experience Kulvicki describes, I remain sceptical, as I imagine others will be, of its capacity to contribute to an explanation of integration, and thereby to fill out an explanation of two-foldedness. In fact it seems to do just the opposite. It strikes one immediately that in the cases Kulvicki cites – two spaces in a window, as mentioned above, for example, or one’s thumb held up and the scene beyond – one is aware that the two objects, one of which partially occludes the other (one’s thumb and a distant bicyclist, say), can move apart without interrupting one’s visual awareness of either. This is emphatically not so in the case of depiction. Indeed one is aware that if one’s thumb were an inch to the left, so that it would not at all occlude the bicyclist, one would still be visually aware of the bicyclist and one’s thumb; but one does not think of the patch of paint as somehow spatially independent in this way of what it depicts, as if one could see Botticelli’s Venus in one location and the marks and strokes of paint which represent her in another (let alone without seeing the marks and strokes of paint at all). Of course the dependence is sometimes much more complicated and is not in general predictable; like the movement of a butterfly’s wing causing a tsunami, a little difference can utterly transform the experience of the represented object or scene. But, in any case, Kulvicki’s view violates the evident spatial dependence of the secondary object on the primary object (for a slightly different argument taking issue with Kulvicki, see Hopkins 2012: 656). The fact that the localisation requirement holds for the revised Wollheimian account shows that it at least partially explains integration (that pictorial experience is not disjointed, that the two phenomenological spaces are unified). It shows that matters of representational orientation, direction and spatial relation are constrained, if not wholly determined, by the relative location of the marks on the physical canvas. If, looking at a canvas, I see two men in it playing table tennis, this is normally because when I look at the canvas, I see two bits of paint as the players, a larger quadrilateral between those bits of paint as the table and so on. And that dependence is something I see.

A measure of Kant seen in Wollheim


The spatial intimacy of the two folds stands forth more conspicuously if we consider depiction in sculpture for a moment. Consider a Henry Moore reclining figure. Sometimes, one sees such a sculpture without immediately seeing anything in it or without seeing it as something else. But then the aspect dawns; we see it as a reclining woman or see a reclining woman in it (unlike the case of something like Louise Nevelson’s wall sculptures, where the attempt at seeing-in or seeing-as is strongly provoked but always frustrated). This is a case where the spatial relations in the secondary object depend entirely on the spatial relations in the primary object – and we see this dependence, can readily point it out. There is no mystery of integration in such a case. Nor should there be when we shift our attention to relief-sculptures, a revealing halfway house between sculpture in the round and painting or drawing. Inflection in a more substantial sense, made prominent by Michael Podro (1998), presents a different problem. Despite its success in facilitating integration, the revised Wollheimian account, as well as Wollheim’s original, looks as if it divided pictorial experience into two disjoint classes of experienced property: the configuration of marks on the surface and the represented properties. The supposed problem for Wollheimians is that inflected properties are those that straddle the gap between configurational, or design, properties and represented properties: a specification of certain represented properties must refer to the picture’s design properties (Hopkins 2010: 158). Simple examples include the ‘movement lines’ in cartoons, which are incapable of being seen face to face but which are part of the design and which are also part of the representational content of the cartoon. Once we appreciate that there are such examples, we will be open to a lot more, such as much of de Kooning’s Woman series: one might describe these figures not so much as people we could encounter face to face as ones literally made of or emerging from paint; but of course they are not representations of women made of paint. If this is right, then pictorial experience has an irreducible complexity: it cannot be divided into ‘seeing configurations’ and ‘seeing represented objects’ without remainder. They interpenetrate (for more see Lopes 1996; Bantinaki 2010; Nanay 2010). The answer to this problem as I see it is that no reason has been given why such examples cannot after all be assimilated to our model or indeed to Wollheim’s original model. Wollheim himself does not think that representational properties are generally ones that might be encountered in another way, such as in a real face to face encounter; nor is his view that one seems to see such properties (Hopkins 2012). Again, two-foldedness is not to be explained as the sum of seeing the canvas non-representationally and seeing (or some other independently understood form of vision such as visualising) the depicted objects. The representational fold does not admit of occurring on its own.14 Wollheim: it is true that each aspect of the single experience is capable of being described as analogous to a separate experience. It can be described as though it were a case of simply looking at a wall or a case of seeing a

200 Gary Kemp boy face-to-face. But it is an error to think that this is what it is. And we get not so much into error as into confusion if, without equating either aspect of the complex experience with the simple experience after which it can be described, we ask how experientially like or unlike each aspect is to the analogous experience. We get lost once we start comparing the phenomenology of our perception of the boy when we see him in the wall, or the phenomenology of our perception of the wall when we see the boy in it, with that of our perception of boy or wall seen face-to-face. Such a comparison seems easy enough to take on, but it proves impossible to carry out. The particular complexity that one kind of experience has and the other lacks makes their phenomenology incommensurate. (1987: 47) One knows that there aren’t any women with shapes such as those depicted in Les Demoiselles d’Avignon, that there aren’t women’s faces with green lines in the middle as in Matisse’s picture. Yet they are women all the same. And remember, too, the camel in the clouds, the battle scenes in stained plaster, displaying the rife suggestibility of our powers of perception. Seeing-in does not demand that the experience be assimilated to, or even that it resembles, the face to face (or other visual) experience of the depicted objects; ‘is capable of being described as analogous’ or ‘can be described as though’ is a very weak requirement, as Wollheim stresses the moment his pen had written those words. Once we appreciate this, and add to the mix not only context but imagination and convention, and the miscellaneous subconscious triggers of recognition such as those described by Gombrich – noting that these are largely unconscious and are a subject for psychology, not philosophy – we see the potential, the logical elasticity, of seeing-as and seeing-in. It is thus unclear whether de Kooning’s women genuinely pose a problem for Wollheimians (or the pen sketch by Rembrandt described by Hopkins, which seems to represent a man whose hand is made of ink).15

Conclusion The revised view then is Wollheim’s Foundational Account, with an altered substructure: two-folded seeing is defined – not merely explained – as negative judgemental aspect-seeing, where this is explained – not defined – in terms of both seeing-in and seeing-as; or, more to the point, the seeing of aspects of both objects and states-of-affairs. In closing I wish to underscore what is an admittedly deep difference between Wollheim’s own theory and what I am calling its revision. Wollheim again: The central difference between seeing-in and seeing-as, from which their various characteristics follow, lies in the different ways in which they are related to what I call ‘straightforward perception’. By straightforward

A measure of Kant seen in Wollheim


perception I mean the capacity that we humans and other animals have of perceiving things present to the senses. Any single exercise of this capacity is probably best explained in terms of the occurrence of an appropriate perceptual experience and the correct causal link between the experience and the thing or things perceived. Seeing-as is directly related to this capacity, and indeed is an essential part of it. By contrast, seeing-in derives from a special perceptual capacity, which presupposes, but is something over and above, straightforward perception. (1980: 217) According to the revised theory, all ordinary perception involves both seeingas and seeing-in: it involves both the carving-up into objects – seeing-as precisified – and awareness of the properties and larger states-of-affairs in which they figure – seeing-in precisified. If that seems strange, to run against what we ordinarily say, remember that the queerness of the idea can be explained by the fact that perception is almost always veridical; we reserve explicit talk of aspect-perception for anomalous cases or cases of representational seeing (and perhaps only if we are philosophers).

Notes 1 A basic commitment that I do not discuss is Wollheim’s acceptance that an explanation of depiction must run through the human mind, thus differing from a view which deals only in matters of optics, occlusion and perspective – a venerable view which John Hyman resurrects impressively (2006) in a recent book. 2 In addition to Hopkins (Wollheim and Hopkins 2003), there are various pieces on Wollheim’s use of the distinction in Hopkins and Savile (1992), van Gerwen (2001) and Abell and Bantinaki (2010). 3 Hopkins doubts that seeing-as is anything but a grammatical variant of seeing-in. I don’t think he is right on this point, as will emerge. See Hopkins (Wollheim and Hopkins 2003: 167). 4 It is not clear that Wittgenstein himself thought of the two as distinct. Sometimes, when Wittgenstein is discussing aspect-perception, he does find it more natural to speak of one’s seeing a secondary object in the primary object (RPPII 552, 556; cf. PI I 193d; RPPI 1042), such as one’s seeing a father’s face in the child’s face. But he does not develop the point, at least as far I know. 5 Actually this doesn’t quite work in any case; one can see a face in a fire, where, rather than seeing the fire as a face, one sees a part or a region of the fire, some configuration of flames or coals in particular, as a face. 6 I think an especially important criticism of Budd’s can be got round (1992: 270–3). Wollheim has his two folds – the configurational and the recognitional – happening only in the context of two-folded seeing; and they are equally ‘incommensurate’ with normal non-representational visual experience. But how then does the configurational fold differ from the straightforward visual inspection of the surface? It does not seem very different. And how does the recognitional fold differ (and in what respects does it match) a straightforward visual encounter with the object face to face? Budd points out that Wollheim says next to nothing about the first and is disappointing in regard to the second. As far as I can see, Wollheim ought simply to have said that the first fold is identical to the straightforward visual inspection of the surface, not that it is special to the seeing of representations;

202 Gary Kemp





11 12


two-folded perception can still retain its distinctiveness by virtue of the added recognitional fold. For the second, I do not see any reason to take the bait. If seeing-in really is a distinct and irreducible kind of experience, then there may be similarities one can point to with the face to face encounter, and with visualisation, but one should not demand more than the facts can deliver. Perhaps Wollheim erred in using the (ambiguous) word ‘incommensurate’; the two are alike in some ways, different in others. For more, see Dorsch (Chapter 10, this volume). Mulhall (1990) takes ‘continuous aspect seeing’ for our normal human attitude not only towards language and the human form but, for example, towards cutlery. As for Wittgenstein exegesis, this will call for special interpretation of PI I 195b, 206b, 210h; RPPI 1028; LWI 478, 518, 534–6, 686, 692; among others. Mulhall faces the issue on pp. 126–37. See Baz (2000: 97–122 and in this volume) for further discussion. Wittgenstein is concerned precisely to distinguish aspect-perception from ordinary perception (PI II 195c, 206b; RPPII 412, 520ff, 1028). Two differences are important. First, aspect-perception is, at least indirectly, subject to the will (PI II 206b, 213f; RPPI 899, 971; RPPII 545; LWI 452, 612). It always makes sense to suggest to someone looking at a cloud, ‘See it as a rabbit’. But it is near unintelligible if we say, of someone looking at a rabbit, ‘See it as a rabbit’. The second difference – not stressed by Wittgenstein – is that when I see a cloud as rabbit, there is no question of its being correct or incorrect to do so. When I take myself to see a rabbit, I thereby form the perceptual judgement that that is a rabbit and am objectively incorrect if it turns out to be a squirrel. But if, under typical circumstances, I see a cloud as a rabbit, there is no room for my being mistaken, in which case no question of correctness arises. This is consistent with Wollheim’s just-so story of Ur-painting or Ur-drawing (1987: 19–25). Caveman Og notices that he can see in the sand things that are not present in the sand; more generally, that he has a capacity for negative judgemental aspect-perception. Og gets the idea of bringing such experiences about by manipulating the sand with his finger or a twig; thus drawing is born. Often too much is made of ambiguous cases like the duck–rabbit, as if ambiguity is essential to what we are supposed to learn from them (McFee 1999: 262–84; for an example, see: 275; cf. PI II 193b, 195d, 196c, 198i, 201f, 204c–f; RPPI 207, 316, 379; RPPII 398, 487, 513; LWI 598; a defender of McFee’s interpretation might look to PI II 195i). I think it clear that this is not so. When we come to see a passing cloud as a rabbit, we do the same sort of thing we do when we see a collection of drawn lines as a rabbit; it is to see one thing as another. For more on this, see Schwyzer (1990). Hyman downplays these remarks in his crucial discussion of Wollheim (2006: 127–42). Wollheim does not think it possible to ‘explain how pictures represent’ (ibid: 136) – to ‘explain depiction’ (ibid), to define ‘the kind of perception he calls “seeing-in” (ibid: 139) or to ‘find some other kind of predicate that ties the marks on the surface of a picture to the kinds of object they depict’ (ibid: 140). Wollheim’s view is that seeing-in is phenomenologically basic. Nevertheless, there is a great deal to say about the experience, in particular how it differs from the experience of seeing-as – which the remarks are concerned to articulate. Wollheim says in Painting as an Art: ‘I discern something standing out in front of, or (in certain cases) receding behind, something else’ (1987: 46). Hyman thinks that the view is subject to counterexamples (2006: 133–6). For example, stick-figure drawings lack this feature. I cannot see that Wollheim’s view cannot tolerate such cases, of the figures being relatively schematic, as much as hieroglyphs or icons as pictures. There is no reason not to accept that the recession or precession in seeing one thing in another, or one thing as another, can be indeterminate or indeed absent. Integration is not absent in such cases, however: it is merely limited to two-dimensions.

A measure of Kant seen in Wollheim


14 See note 6. 15 Thus I don’t think Wollheim simply ‘ducked’ the issue as Hopkins says (2010: 173 fn 10), given that he is prepared to grant a lot of scope to seeing-in. He would not have been sympathetic to the strain of reductionism now popular, which strives for a theoretical explanation of seeing-in.

References Abell, Catherine and Katerina Bantinaki 2010. Philosophical Perspectives on Depiction. The Mind Association Occasional Series. Oxford: Oxford University Press. Bantinaki, Katerina 2010. ‘Picture Perception in Twofold Experience’. In Catherine Abell and Katerina Bantinaki (eds), Philosophical Perspectives on Depiction. The Mind Association Occasional Series. Oxford: Oxford University Press, pp. 128–50. Baz, Avner 2000. ‘What’s the Point of Seeing Aspects?’. Philosophical Investigations 23(2): 97–122. Budd, Malcolm 1992. ‘On Looking at a Picture’. In James Hopkins and Anthony Savile (eds), Psychoanalysis, Mind and Art: Perspectives on Richard Wollheim. Aristotelian Society Monographs. Oxford: Blackwell, pp. 259–80. Reprinted in Malcolm Budd, 2008, Aesthetic Essays. Oxford: Oxford University Press. Gaiger, Jason 2008. Aesthetics and Painting. London: Bloomsbury. Gerwen, Rob van (ed.) 2001. Richard Wollheim on the Art of Painting: Art as Representation and Expression. Cambridge: Cambridge University Press. Gombrich, Ernst 1960. Art and Illusion: A Study in the Psychology of Pictorial Representation. The A. W. Mellon Lectures in the Fine Arts, 1956. Princeton: Princeton University Press. Hopkins, James and Anthony Savile (eds) 1992. Psychoanalysis, Mind and Art: Perspectives on Richard Wollheim. Aristotelian Society Monographs. Oxford: Blackwell. Hopkins, Robert 2010. ‘Inflected Pictorial Experience: Its Treatment and Significance’. In Catherine Abell and Katerina Bantinaki (eds), Philosophical Perspectives on Depiction. The Mind Association Occasional Series. Oxford: Oxford University Press, pp. 151–80. Hopkins, Robert 2012. ‘Seeing-in and Seeming to See’. Analysis 72(4): 650–9. Hyman, John 2006. The Objective Eye: Colour, Form, and Reality in the Theory of Art. Chicago: Chicago University Press. Isenberg, Arnold 1949. ‘Critical Communication’. Philosophical Review 58(4): 330–44. Kulvicki, John 2009. ‘Heavenly Sight and the Nature of Seeing-In’. The Journal of Aesthetics and Art Criticism 67(4): 387–97. Levinson, Jerrold 1998. ‘Wollheim on Pictorial Representation’. The Journal of Aesthetics and Art Criticism 56(3): 227–33. Lopes, Dominic 1996. Understanding Pictures. Oxford: Oxford University Press. McFee, Graham 1999. ‘Wittgenstein on Art and Aspects’. Philosophical Investigations 22(3): 262–84. Mulhall, Stephen 1990. On Being in the World: Wittgenstein and Heidegger on Seeing Aspects. London: Routledge. Nanay, Bence 2010. ‘Inflected and Uninflected Experience of Pictures’. In Catherine Abell and Katerina Bantinaki (eds), Philosophical Perspectives on Depiction. The Mind Association Occasional Series. Oxford: Oxford University Press, pp. 181–207. Newall, Michael 2010. ‘Pictorial Resemblance’. The Journal of Aesthetics and Art Criticism 68(2): 91–103.

204 Gary Kemp Peacocke, Christopher 1987. ‘Depiction’. The Philosophical Review 96(3): 383–410. Podro, Michael 1998. Depiction. New Haven: Yale University Press. Schwyzer, Hubert 1990. The Unity of Understanding. Oxford: Oxford University Press. Wittgenstein, Ludwig 1951. Philosophical Investigations. Oxford: Oxford University Press. Noted as PI. References to this work are to section numbers in part I, and page numbers with paragraphs in part II. Wittgenstein, Ludwig 1980a. Remarks on the Philosophy of Psychology, Volume I. G. Anscombe and G. von Wright (eds), tr. by G. Anscombe. Oxford: Blackwell. Noted as RPPI. References to this work are to section numbers. Wittgenstein, Ludwig 1980b. Remarks on the Philosophy of Psychology, Volume II. G. Anscombe and H. Nyman (eds), tr. by C. Luckhardt and M. Aue. Oxford: Blackwell. Noted as RPPII. References to this work are to section numbers. Wittgenstein, Ludwig 1982. Last Writings on the Philosophy of Psychology, Volume I. G. von Wright and H. Nyman (eds), tr. by C. Luckhardt and M. Aue. Oxford: Blackwell. Noted as LWI. References to this work are to section numbers. Wittgenstein, Ludwig 1992. Last Writings on the Philosophy of Psychology, Volume II. G. von Wright and H. Nyman (eds), tr. by C. Luckhardt and M. Aue. Oxford: Blackwell. Noted as LWII. References to this work are to page numbers followed by paragraph numbers. Wollheim, Richard 1972. On Art and the Mind. Harvard: Harvard University Press. Wollheim, Richard 1980. Art and Its Objects: With Six Supplementary Essays. 2nd edition. Cambridge, New York: Cambridge University Press. Wollheim, Richard 1987. Painting as an Art. Princeton: Princeton University Press. Wollheim, Richard 1993. The Mind and Its Depths. Cambridge, MA: Harvard University Press. Wollheim, Richard 2001. ‘On Pictorial Representation’. In Rob van Gerwen (ed.) Richard Wollheim on the Art of Painting: Art as Representation and Expression. Cambridge: Cambridge University Press. Wollheim, Richard 2003. ‘In Defense of Seeing-In’. In Heiko Hecht, Robert Schwartz and Margaret Atherton (eds), Looking into Pictures: An Interdisciplinary Approach to Pictorial Space. Cambridge MA: MIT Press. Wollheim, Richard and Robert Hopkins 2003. ‘What Makes Representational Painting Truly Visual?’, Proceedings of the Aristotelian Society 77: 131–67.

10 Seeing-in as aspect perception Fabian Dorsch

1 I shall argue that seeing-in, the central element of pictorial experience, is a form of aspect perception. The argument thus connects Wollheim’s main contribution to the philosophy of depiction with one of the central themes in Wittgenstein’s philosophy of psychology.1 This Aspect View of pictorial experience that I would like to put forward is designed to be a direct competitor to the Experienced Resemblance View and the Imagination View, and to improve on both of them by incorporating some important elements of either. More specifically, the Aspect View claims that seeing-in involves the imperfect illusion of the picture’s surface as possessing the aspect of having the visual appearance of a three-dimensional arrangement of objects (i.e. the depicted scene). And, as part of this aspect perception, we both experience the picture’s surface as resembling the depicted scene in twodimensional shape and have a non-perceptual awareness of the depth and volume of that scene which is similar to, but not quite like imagining. The argument proceeds in five steps. The first (Sections 2–7) is primarily concerned with discussing certain important features of pictorial experiences, including their relation to pictures. I start by distinguishing two important aspects of pictorial experience, namely that it makes us visually aware both of a surface’s status as a picture and of the latter’s specific pictorial content. Then, after introducing Wollheim’s idea that these two properties of pictures (i.e. their status and their content) are response-dependent, specifically on experiences of seeing-in, I spell out in more detail what seeing-in is like, with a special focus on its twofoldness. Following Wollheim in identifying the central element in recognising something as a picture with the visual awareness of the depth of the depicted scene, I conclude the preparatory first step of the argument by raising two crucial questions – the Question of Content about how we recognise that a given picture depicts X rather than Y; and the Question of Depth about how we recognise the depth of the depicted scene. The second step of the argument (Sections 8–12) establishes the distinction between divisive and unitary accounts of seeing-in and finds fault with Wollheim’s endorsement of a divisive view, as well as with his conception of

206 Fabian Dorsch the precise difference between the two folds of seeing-in, the configurational and the recognitional fold. The third (Sections 13–16) and the fourth steps (Sections 17–21) critically discuss the two most prominent unitary accounts of perceptual experience, the Experienced Resemblance View and the Imagination View respectively. While the main difficulty for the first view is that it has no resources to answer the Question of Depth, the second view fails to provide a proper reply to the Question of Content and, moreover, misidentifies our awareness of what is depicted as imaginative. As a viable alternative to both views, I present and begin to defend the Aspect View in the fifth and last step of the argument (Sections 22–6), not least by spelling out in a bit more detail what aspect perception generally involves and why seeing-in should be understood as an instance of it. In particular, the Aspect View is in a position to answer both the Question of Content and the Question of Depth.

2 A pictorial experience is one in which we are visually aware of a flat surface as a picture. This means that, whenever one of our pictorial experiences is veridical, we succeed in visually recognising some picture for what it really is, namely a two-dimensional depiction of some objects and their visible features – and not just a flat surface.2 In addition, merely judging and coming to know that something is a depiction of something else is not an instance of pictorial experience because we are not aware of the object’s status as a picture in a visual manner.

3 Visually experiencing a picture as a picture necessarily involves at least two aspects of awareness. First, we have to be visually aware of the experienced object’s property of being a picture. That is, we should visually experience it as being a picture, rather than merely as being a flat surface, say. Second, we have to be visually aware of its pictorial content, of the specific objects and features that it happens to depict. That is, we should have a visual experience of X rather than Y, say, if the picture depicts X rather than Y. In other words, to visually recognise something as a picture means to visually recognise both that it is a depiction and what it is a depiction of. It is impossible to enjoy a pictorial experience, in the sense specified, that includes only one of the two elements. On the one hand, we cannot experience something as a picture without being aware of (at least part of) its pictorial content. For the only perceivable feature by virtue of which I can visually recognise something as a picture is precisely its possession of a pictorial content. Pictures do not show any other perceivable characteristics which indicate their pictorial nature. For instance, just being aware of a flat surface with marks of colour on it

Seeing-in as aspect perception 207 does not suffice for experiencing the surface as depicting something. We also have to be aware, at least to some extent, of what it depicts. Therefore, although it is not required that one has to be aware of the whole pictorial content of an object in order to recognise it as a depiction, awareness of something as a picture implies the simultaneous awareness of (part of) what is depicted. The same applies to the specificity of pictorial content. While we might not always succeed in recognising the particular nature of the depicted object when having a pictorial experience, such an experience always involves the recognition of at least some determinable aspect of its nature – minimally, an object or figure distinct from the depicting surface. On the other hand, as already hinted at, if we are merely aware of what a picture depicts but do not recognise that it is indeed a picture, then we end up having a (potentially misleading) perceptual experience, rather than a pictorial experience. There are two possibilities here, depending on whether the picture concerned depicts three- or two-dimensional objects. Trompe-l’oeils belong to the first kind of pictures. When we look at them and fail to experience them as pictures, it (wrongly) seems to us as if we really see a threedimensional scene, while this scene is in fact merely depicted. Good examples of pictures of the second kind are actualist pictures: that is, pictures which depict flat surfaces that look exactly like the pictures themselves (Jasper Johns’s Target Paintings are representative cases). Again, when we look at such pictures, we simply see nothing but the visual appearance of the twodimensional object that is depicted (which also happens to be the visual appearance of the picture itself). In fact, in contrast to trompe-l’oeils, which we can normally recognise as such just by changing our point of view on them, we can arguably never visually recognise actualist pictures as pictures, given that their visual appearance is identical with, and hence cannot be visually discriminated from, the visual appearance of what is depicted.

4 Nonetheless, the two aspects of pictorial experience concern two distinct facts about the nature of depictions, which show a certain kind of independence of each other. First of all, that something is a picture has no influence on what it depicts. This suggests that those properties of a picture that determine its status as a depiction are, taken on their own, not sufficient for the determination of its pictorial content. It is difficult to see how, otherwise, it could be explained that pictures with very different subject matters – and with very different degrees of specificity and complexity of pictorial content – still count as instances of the same kind (i.e. as depictions). The best explanation seems to be that some shared feature determines their status as pictures, but not – or not on its own – their pictorial content. But it also seems true that what specifies the pictorial content is, at least to some extent, independent of what determines the status as a depiction. In particular, we can easily change what a picture depicts (for example,

208 Fabian Dorsch by applying more paint), without influencing its property of being a picture. So there appear to be features of pictures that are central to the determination of what they depict but not of their status as depictions. The idea is thus that we can distinguish two sets of properties of a picture, one of which is responsible for there being a picture, while the other determines what it depicts; and the two sets of properties are to some degree independent of each other, even though they presumably overlap and cannot be instantiated completely independently of each other, given that nothing is a picture without a minimally determinate pictorial content, and vice versa.

5 One of Wollheim’s important contributions to the philosophy of depiction was his emphasis of the fact – already noted by Gombrich (1960) – that the properties in question are response-dependent (though he did not put it this way): both that something is a picture and what it depicts depend on how we normally respond to it.3 More specifically, something is a picture of something else only if, and partly because, we normally visually recognise it to be a picture of that other object (assuming that we are looking at the picture). If we act on our intention to produce a depiction of a certain object, but the resulting marks on the surface concerned are such that no one (us included) can become visually aware of the surface as a picture, or identify what it is supposed to depict, then it is not a depiction of the object in question, and we have failed in our attempt to produce one. In other words, both being a picture and having a certain pictorial content are not strongly objective properties in the sense that they could be instantiated in the world without any (human) beings ever possessing the capacity to visually recognise things as pictures of something else. Instead, they are objective properties only in the sense that they allow for intersubjectivity and misrecognition.4 But the two properties are also not strongly subjective in the sense that whenever we visually experience something as a picture of something else then it is such a picture. For we may err about whether something is a picture, as well as about what it specifically depicts if it is a picture. This is why what matters for the presence of depiction are our normal responses, which happen under conditions that are suitable for veridical pictorial experience. It might not be easy to characterise these conditions, but they are likely to be very similar to those for visual perception. Of course, like most, if not all response-dependent properties, being a depiction and depicting something specific are features that are partly grounded in the lower-level properties of the pictures concerned. Given that we can modify what a picture shows us by applying or removing paint, the colour, texture and shape of marks of paint plays some role, yet to be specified, in the determination of what is depicted. Similarly, some aspect of the marks on the surface should also be relevant for whether it is a depiction in the first place.

Seeing-in as aspect perception


6 Now, the fact that the properties of being a depiction and of having a certain pictorial content are response-dependent raises the question about the nature of the normal response(s) concerned. Given that, as just argued, we cannot be aware of one of the properties without being aware of the other, the response by means of which we recognise a thing as a picture and the response by means of which we recognise what specifically this picture depicts should be one and the same. This is indeed what Wollheim proposes. The core of his view is the claim that we visually recognise a flat surface as a picture of some object or scene by means of ‘seeing’ the object or scene ‘in’ the picture. This visual experience of seeing-in is, minimally, characterised by five features (Wollheim 1986: section 2f.); 1987: 46f., 62; 1998: 221). First, it involves the visual awareness of the flat surface and its visible features, including its flatness and the colour, shape and arrangement of the marks on it (i.e. what Wollheim calls the ‘configurational fold’ of seeing-in). Second, seeing-in involves the visual awareness of what is depicted: that is, of a certain spatial arrangement of two- or three-dimensional objects and their visible features (i.e. what Wollheim calls the ‘recognitional fold’ of seeing-in). Third, our visual awareness is always distributed between the configurational and the recognitional fold, meaning that we are never only aware of one of them. In other words, we are always aware both of the depicting surface and of what it depicts. This third feature is what Wollheim calls the ‘twofoldness’ of seeing-in.5 Fourth, the two folds of awareness do not constitute two independent experiences that merely happen to occur simultaneously, but form a single, unified experience. That is, the configurational and the recognitional fold could not exist in isolation from each other, or at least not in exactly the same experiential way in which they form part of seeing-in (for example, they could not have the same phenomenology if they were to occur on their own (Hopkins 1998: 20f.)). After all, there is an experiential difference between seeing a person face-to-face and seeing her in a portrait. Similarly, switching from seeing a surface merely as a surface to seeing it as a picture, or vice versa, also comes with a noticeable experiential difference. This may happen, for instance, when we move very close to the surface of a painting and look just at a small section of the canvas: we may stop seeing this portion of the surface as depicting anything. Fifth, the recognitional fold involves the awareness of depth. To get clearer about this observation of Wollheim’s, it is helpful to consider the case of non-figurative paintings, such as Malevich’s Black Circle (Figure 10.1), for cases like this illustrate well how Wollheim conceives of the central element of the recognitional fold and, hence, of our experience of a flat surface as a picture (or ‘representation’): To see something as a representation is intrinsically bound up with, and even in its highest reaches is merely an elaboration or extension of, the

210 Fabian Dorsch

Figure 10.1 Kasimir Malevich, Black Circle 1915 (State Russian Museum, St Petersburg).

way in which, when the black paint is applied to white canvas, we can see the black on the white, or behind the white, or level with it. (Wollheim 1974: section 25) [W]hen seeing-in occurs, two things happen: I am visually aware of the surface that I look at, and I discern something standing out in front of, or (in certain cases) receding behind, something else. (Wollheim 1987: 46) These passages make clear that Wollheim takes not only figurative paintings but also non-figurative ones to be pictures, which normally give rise to seeing-in (see also Wollheim 1987: 62). Indeed, he suggests that figurative seeing-in is just a less basic and more complex form (i.e. an ‘elaboration or extension’) of nonfigurative seeing-in. Moreover, and more important for our current purposes, the two passages reveal that Wollheim takes the recognitional fold to be, at its core, an awareness of depth: that is, an awareness of that dimension of space relative to which things are in front of – or, indeed, level with – other things. He seems perfectly right to claim that the recognitional fold of seeing-in always involves a visual awareness of a three-dimensional arrangement of objects, irrespective of whether these objects are two-dimensional figures (for example, circles and planes), or whether they are themselves three-dimensional objects (for example, trees or animals), and irrespective of whether the objects are experienced as actually being at different distances relative to the point of view inherent to the picture, or instead as being at the same distance, but with the potential of being at different distances. This awareness of depth is precisely what distinguishes experiencing a flat surface as a picture from experiencing it merely as a flat surface. If, as

Seeing-in as aspect perception


part of the recognitional fold, we were not to experience one thing as being in front of another (or as being level with it, but in such a way that it could also be in front of it), then we would not enjoy an experience of something depicted over and above our awareness of the flat surface in question. In particular, Hyman’s claim to the contrary is unconvincing (Hyman 2003: 681f; 2006: 135f.). When we look at stick-figure drawings or the silhouettes of people (two of his favourite counterexamples), we usually do not see merely two-dimensional bodies in the surfaces. And even when we do experience the various body parts as being flat and on the same plane, they are at least visually given to us as being in a space which would allow them to be at different distances (for example, if they were to be rotated, or presented from a different point of view). Similarly, even if depictions of the starry sky stay neutral on whether any of the stars are closer or further away than the others, we are still aware of the depicted sky as allowing for such differences in distance. Seeing-in thus always involves some awareness of depth; and the corresponding pictures are always depictions of depth.

7 The idea that pictorial experience consists in a unified experience of twofold seeing-in, with the five features just listed, is today widely accepted.6 It also helps us to answer the question about the nature of the normal responses which determine whether something is a picture and what it depicts. Accordingly, something possesses the status of being a picture (rather than that of a mere surface) only if, and partly because, we normally see something in it when looking at it (i.e. we normally have a unified experience of it that involves both a configurational and a recognitional fold). Similarly, a picture is a depiction of X (rather than Y) only if, and partly because, what we normally see in it is X (rather than Y). These considerations about the nature and role of seeing-in raise two important questions, each of which is concerned with one of the two highlighted properties of pictures, their status and their content. The first question is simply about our access to the specific pictorial content of pictures: how are we aware of the fact that a given picture depicts X (rather than Y)? The second question, by contrast, addresses our access to the status of pictures as depictions. Given that the awareness of depth is central to our recognition of a flat surface as a picture, the question deals more specifically with this kind of awareness: how are we aware of the depth of the depicted scene? For ease of reference, I label these two questions the Question of Content and the Question of Depth respectively. Both questions are primarily concerned with the recognitional fold, emphasising the relative importance of this fold for seeing-in. As we will see, the most prominent accounts of seeing-in are unable to answer these two questions in a fully satisfactory manner, generating the need for an alternative elucidation of seeing-in.

212 Fabian Dorsch

8 There has been much debate about the precise nature of twofold seeing-in and, indeed, about how much we can actually say about it. Wollheim is famous for insisting that the configurational and the recognitional fold could not occur on their own, and that we cannot describe their phenomenology by reference to other, independent experiences, notably visual perceptions of surfaces that are not experienced as pictures, and visual perceptions of objects and scenes that are experienced face-to-face rather than by means of depictions (Wollheim 1987: 46f.). As a result, Wollheim concludes that our capacity for seeing-in is a ‘fundamental perceptual [capacity]’ (ibid: 45f.), which is likely to be innate (Wollheim 1986: 46) and cannot be much further elucidated. Others have been critical of Wollheim’s conclusion and his arguments in favour of it, and consequently also more optimistic than Wollheim about our prospects of providing a substantial and illuminating account of pictorial experience in terms of seeing-in (Budd 2008b; Hopkins 1998: section 1.4; Walton 1992). I turn to two such accounts – the Experienced Resemblance View and the Imagination View – in a minute. There are two ways in which the configurational and the recognitional fold may be related to each other, assuming that twofold seeing-in constitutes a unified experience. Either the two folds may constitute two different instances of awareness or they may together form a single instance of awareness. One way of spelling out this distinction is in terms of content. According to the first option, the unified experiences of seeing-in may involve two contents, one of which presents the surface of the picture as being a certain way (for example, ‘there are many white areas separated by many thin black lines’) and another which presents the depicted object and its features (for example, ‘there is a figure reclining on a bed’). Hopkins (2010b: 167) calls such a view ‘divisive’ since it takes the two contents to be fairly independent of each other.7 The second option is to maintain that the configurational and the recognitional fold are nothing but abstractions from a single, complex content which presents the surface of the picture and the depicted object as standing in a certain relation to each other. Paradigms of this ‘unitary’ view, as Hopkins (ibid) labels it, are the two views just mentioned. While the Experienced Resemblance View argues that a pictorial experience perceptually presents the surface of a picture as looking like the depicted object (for example, ‘the lines on the surface look like the contour of a reclining figure’), the Imagination View insists that a pictorial experience imaginatively presents the surface as being the depicted object (for example, ‘the black lines are the contour of a reclining figure’) – or, indeed, that a pictorial experience presents our perceptual experience of the surface as being our perceptual experience of the depicted object (for example, ‘my seeing the black lines is my seeing the contour of a reclining figure’).8 Wollheim’s view on the matter is divisive, in the sense just described. That is, he takes the configurational and the recognitional fold to constitute

Seeing-in as aspect perception


different instances of awareness or content. This gives rise to the question of how the two folds are supposed to differ in awareness and, in particular, what kind of awareness is assumed to be involved in our visual recognition of what is depicted.

9 The two passages from Wollheim’s writings quoted earlier indicate that he believes that the configurational fold consists in the visual awareness of a flat surface, while the recognitional fold involves the visual awareness of depth. In addition, he explicitly speaks of ‘seeing’ the black circle in front of the white square and takes seeing-in, as a whole, to be a form or instance of perception (Wollheim 1987: 45f.). This suggests he assumes that not only the configurational but also the recognitional fold are perceptual and thus do not differ in the kind of visual awareness that they involve. Taken together, these two elements of Wollheim’s view imply that he conceives of the difference between the two folds of seeing-in solely in terms of depth: they differ because only our visual awareness of what is depicted involves the visual awareness of a third dimension.9 But this characterisation of the difference cannot be right. First of all, the surfaces of pictures need not – and often are not – completely flat. Indeed, with many paintings, the relief-like texture of brushstrokes, say, is crucial for depiction (for example, when van Gogh, in his painting Wheat Field With Cypresses (Figure 10.2), uses them to portray the leaves and branches of cypresses in motion). In addition, our perception of the surfaces of pictures locates them in our three-dimensional environment (for example, in a certain distance from ourselves). Hence, the configurational fold often, if not always, involves awareness of depth as well. In addition, the proposed characterisation misidentifies the true location of the difference between the two folds. For the configurational and the recognitional fold do not (or not merely) differ in whether they make us visually aware of a third dimension, but instead (or also) in whether they make us perceptually aware of the spatial dimensions concerned. While the configurational fold of seeing-in is clearly perceptual, the recognitional fold is not. When looking at Malevich’s painting, we do not literally see a black circle as being in front of a white square (or a larger white area that extends beyond that part of the non-figurative scene that is depicted). Nor do we feel any inclination to judge that this is really the case. So, perhaps, a better characterisation of the difference between the two folds is that, while we are perceptually aware of the depicting surface, we are non-perceptually aware of what is depicted. This, however, leaves unanswered the question of what kind of visual awareness is actually involved in the recognition of the depicted objects and features, if not perception – an issue to which we have to return to later on.

214 Fabian Dorsch

Figure 10.2 Vincent van Gogh, Wheat Field with Cypresses 1889 (black-andwhite detail, Metropolitan Museum of Art, New York).

10 There is also a second problem for Wollheim’s view which does not concern his conception of the difference between the two folds but instead the fact that he adopts a divisive view on the relation between the two folds, namely that they constitute two different instances of awareness or content. The challenge for Wollheim is to identify the two kinds of visual awareness that the configurational and the recognitional fold respectively involve. It should be clear that the two kinds are distinct. One of them makes us aware of something that exists in our environment, right before our eyes, while the other does not. This is, indeed, partly why one of the folds counts as configurational and the other as recognitional. Wollheim is also right to insist that the configurational and the recognitional fold cannot be independent of each other in the sense of being able to occur on their own, for, otherwise, the two folds would constitute two distinct experiences that just happen to exist at the same time, rather than one unified experience. So, the challenge is to single out two kinds of visual awareness that are distinct from each other, but nonetheless cannot be instantiated independently of each other. Moreover, if the two folds cannot occur on their own, then they cannot be of the same kind as instances of visual awareness that do occur on their own, notably experiences of seeing, recalling or visualising. Thus, when we are

Seeing-in as aspect perception


looking at a painted portrait of a friend, say, our visual awareness of the depicting canvas cannot be of the same kind as a visual perception, recollection or imagination of a normal, non-depicting surface (for example, a painted wall).10 Similarly, our visual awareness of the person portrayed cannot be of the same kind as a standard (i.e. ‘face-to-face’) visual perception, recollection or imagination of such a person. The two kinds of visual awareness in need of identification therefore have to be different from seeing, recalling, visualising or any other kind of visual awareness that enjoys distinct existence. This means that the two folds of seeing-in must constitute two distinct, but interdependent sui generis kinds of visual awareness, which can occur only in the context of pictorial experience. The plausibility of this idea is closely linked to the issue of whether something more illuminating can be said about the nature of these two kinds of visual awareness. Their postulation seems already to be rather ad hoc, given that it is motivated solely by the need to account for the twofoldness of seeing-in, and given that the two kinds of visual awareness are assumed to play no role outside of seeing-in. But if, in addition, their nature remains completely unelucidated, their introduction loses not only any explanatory power attributed to it but also becomes very difficult to assess. If it remains ill-defined which two kinds of visual awareness a divisive view like Wollheim’s actually postulates, we should in fact reject this view for being unclear.

11 Wollheim seems to think that this charge is misguided because it overlooks the fact that twofoldness introduces such a complexity into seeing-in that, as a result, the configurational and the recognitional fold turn out to be ‘incommensurate’ with simpler visual experiences that enjoy independent existence (Wollheim 1987: 47). What this means is that, while there may very well be certain experiential similarities between seeing-in and other visual experiences, we are not in a position to inquire into and discover their nature. In particular, it is futile to try to describe the two folds of seeing-in by comparing them to standard visual perceptions (or recollections, imaginations, etc.) of surfaces and objects. So, according to Wollheim, it is the very nature of seeing-in that prevents us from saying more about the two kinds of visual awareness involved in it. However, Wollheim’s claim about the incommensurability of seeing-in does not enjoy much plausibility. First, as Budd (2008b: 200) notes, that seeing-in is a complex whole does not show that none of its parts (i.e. its folds) can be compared with other visual experiences. Second, assuming that we are generally able to introspect and describe similarities in awareness between different kinds of experience, that we could not report any similarities between seeing-in and other experiences would be a good sign that there are no such similarities. But this conclusion would seem to be too strong, given that seeing-in shares at least some properties of other visual

216 Fabian Dorsch experiences – notably their visual character. Third, it appears that we are actually able to introspectively note and characterise some of the similarities between (the two folds of) seeing-in and the other experiences. We do not seem to have difficulties recognising that the configurational fold comes very close to a standard perception of a painted surface (for example, in making us visually aware of a flat expanse with marks of paint on it) or that the recognitional fold is much like a visual experience face-to-face (for example, in making us visually aware of some voluminous object or scene).

12 But even if it is possible to identify and describe the various similarities (for example, in content, attitude, functional role, etc.) between the two kinds of visual awareness involved in seeing-in, on the one hand, and standard kinds of visual experience, on the other, this is unlikely to fully remove the challenge to divisive views in general, and to Wollheim’s version in particular. Knowing that the configurational fold resembles a normal perception of a flat surface in visually presenting a marked surface as really being a certain way and eliciting a corresponding perceptual belief does not tell us anything about the nature of that fold. That it is like standard instances of perception might suggest that it itself is a standard instance of perception. But this conclusion would be wrong precisely because the configurational fold cannot, unlike normal perception, occur without the recognitional fold. And the recognition of the resemblance does not suggest any better conclusions. Similarly, noting that the recognitional fold is very similar to visualising in that both visually present some objects without (necessarily) making any claim about their existence or real nature and without (normally) giving rise to matching beliefs might at best provide some evidence for believing that the fold is an instance of visualising.11 But this verdict would, again, be wrong because of the essential dependence of the recognitional fold on the configurational one. Unless the challenge of providing a proper elucidation of the nature of the two kinds of visual awareness said to be involved in seeing-in is met, we should prefer unitary accounts over the divisive views of Wollheim and others. The two main candidates on the market are, as already mentioned, the Experienced Resemblance View and the Imagination View. However, neither of them can satisfactorily account for both of the two facts discussed at the beginning. While the Imagination View fails to explain our recognition of what a picture depicts, the Experienced Resemblance View cannot explain our awareness of flat surfaces as pictures, notably pictures of voluminous objects or scenes.

13 Let us begin with the Experienced Resemblance View. Proponents of this account maintain that seeing-in has to be elucidated in terms of an

Seeing-in as aspect perception


experience of resemblance. Its main claim is that recognising surfaces as pictures essentially involves experiencing them as resembling some other objects, namely the depicted ones. Accordingly, an object counts as a picture only if it gives rise to such an experience of resemblance and it depicts those entities which it is experienced as resembling. The resulting view is unitary since it interprets the two folds of seeing-in as being combined in a single instance of awareness, with something like ‘the surface looks like X’ (where ‘X’ describes the depicted objects or scene) as its content. And the view postulates merely an experienced resemblance, rather than a real one, since it does not require that the depicting surface and the depicted objects are really similar to each other. Indeed, the depicted objects need not even exist, in which case there could not be any relation of resemblance in the first place. Different philosophers have presented different accounts of the resemblance relation that we are assumed to experience between depicting surfaces and depicted objects.12 But their views have in common that they focus on two-dimensional shapes as the focus of similarity. In other words, they argue that pictures and what they depict are experienced as resembling each other with respect to certain two-dimensional shapes.13

14 The two relevant sets of two-dimensional shape may be labelled surface shapes and outline shapes.14 Surface shapes pertain to flat surfaces, such as those surfaces that depict something. They consist in the shapes formed by the lines and coloured patches arranged on the surface in question. For example, the black lines drawn on a white sheet of paper may delineate an elliptical or a triangular area (Figures 10.3 and 10.4). Outline shapes, by contrast, pertain to voluminous objects, such as those objects that pictures depict. They are identical with the shapes that one would draw on a transparent plane (for example, a pane of glass), aligned orthogonally to the line of sight, in order to trace the outlines of such objects as seen through that plane. Paradigm examples of apparatuses for

Figure 10.3 Unknown artist, A Bowl.

218 Fabian Dorsch

Figure 10.4 Unknown artist, A Road with a Median Strip.

tracing outline shapes are the various perspective frames and machines devised by Leon Batista Alberti, as shown in the two woodcuts by Count Johann and by Albrecht Dürer respectively (Figures 10.5 and 10.6). When the circular rim of a bowl is seen from the side through such an apparatus, its outline shape turns out to be elliptical, while the outline shape of a road that disappears into the distance will be triangular. We can thus understand two-dimensional outline shapes as abstractions from three-dimensional volumes that discard the dimension of depth.15 Because of this abstraction from the third dimension, outline shapes (unlike surface shapes) turn out to be relative to a certain point of view (i.e. to a combination of a certain location in space and a certain direction of gaze). For instance, when the rim of the bowl is seen from above, its outline shape will be circular, rather than elliptical.16 Now, according to the Experienced Resemblance View, what is central to seeing-in is the complex visual experience of the surface shapes of the depicting surface as looking like the outline shapes of the depicted objects. For example, when we look at a drawing that depicts a round bowl from a certain point of view, we visually experience the relevant elliptical shape on the sheet of paper as resembling the elliptical outline shape that the rim of such a round bowl would possess relative to the point of view in question (Figure 10.3). Similarly, in the case of a picture of a road disappearing into the distance, the relevant triangular shapes formed by the lines on the picture’s surface are visually experienced by us as being similar to the triangular outline shapes that the lanes and the median strip of a straight road possess relative to the perspective in question (Figure 10.4).

Seeing-in as aspect perception


Figure 10.5 Count Johann, Window Grid 1531.

15 There are some open questions about the nature of this visual experience of resemblance. The only thing that is clear is that we see the surface shapes but not the outline shapes, for we see the picture but not what is depicted, given that only the former and not the latter is experienced as being there before our eyes. As things appear to us, there is only a picture of a bowl in front of us and no bowl. Seeing-in is thus said to be more like seeing a person and recognising her resemblance to another one who is far away or does not even exist than it is like noticing a resemblance between two people while seeing both at the same time. This raises the question of how we are aware of the outline shape of the bowl or of whatever is depicted, if not perceptually. It is surprising how difficult it seems to provide a satisfactory answer to this question. Hopkins, for instance, maintains that our awareness of the depicted consists in a ‘thought’ about its visual appearance, ‘in the light of which’ we see the picture’s surface, but he also acknowledges that ‘there

220 Fabian Dorsch

Figure 10.6 Albrecht Dürer, An Artist Drawing a Seated Man 1525.

are limits to what I think can illuminatingly be said about experienced resemblance itself’ (Hopkins 2003: 156). Apart from the problem that it is unclear whether ‘thinking’ may indeed be understood in such a wide sense as to include visually experiencing, we are not really any closer to an answer to our question until it is specified what kind of ‘thought’ is at issue. Perhaps, when the picture of a bowl depicts our own bowl at home and we recognise this fact, our awareness of the bowl’s outline shape might be a product of memory. But looking at the picture does not include episodic memory: that is, the recollection of one particular of our past perceptions of our own bowl. We do not have the appearance of our bowl before our mind as we saw it in the past. Hence, the form of memory involved could at best be generic: that is, the kind of non-occurrent visual memory that comes with, or consists in, the capacity to re-identify our particular bowl on the basis of visual perception. Such generic memory might indeed give rise to, inform or become manifest in our visual awareness of the bowl’s outline shape as part of seeing-in. But unless we know more about the nature of this kind of awareness, the Experienced Resemblance View is still incomplete and difficult to assess.

Seeing-in as aspect perception


In addition, pictures need not be of any objects that existed in the past or exist in the present. Hence, the picture in our example may just depict a bowl of a certain type (i.e. one with certain qualities) without depicting any particular bowl (i.e. one with a specific numerical identity). In this case, the idea of recalling the bowl and its outline shape relative to a certain point of view would not be an option. As several proponents of the Experienced Resemblance View have noted, we also cannot treat our awareness of the bowl as an instance of the visual imagination. Pictorial experience is not a matter of seeing a picture’s surface and simultaneously visualising the depicted, partly because the resulting experience would lack the unity characteristic of seeing-in and, in particular, would not include experiencing the depicted by means of, or on the basis of, seeing the picture.17 Moreover, even if we get a sufficient grasp of the awareness of outline shape supposed to be involved in seeing-in, it is still undecided whether our awareness of the relation of resemblance between the two sets of shapes is perceptual or not. That we see the two people whom we experience as resembling one another does not necessarily mean that we see them as being similar to each other. We might also just imagine there being a resemblance between the two on the basis of seeing them. Again, proponents of the Experienced Resemblance View are not very explicit on this issue. True, as observed earlier, they do not assume that the depicting surfaces and the depicted objects resemble each other. But claiming that this resemblance is merely experienced leaves it open whether it is experienced perceptually or in a different manner. Perhaps it is best to reject the perceptual option, simply because this could mean that seeing-in turns out to be systematically illusory, given that there may very well be no real similarity.

16 In any case, the most important issue for our current purposes is whether the Experienced Resemblance View can provide satisfactory answers to the two questions introduced in Section 7. With respect to the Question of Content, this seems to be the case (with the proviso that the issues raised in the last section can be resolved). We recognise that a given picture depicts X (rather than Y) by experiencing its surface shapes as resembling the outline shapes of X (rather than those of Y). In other words, which pictorial content we experience a picture as having is constrained by what we experience to be similar to the picture’s surface with respect to two-dimensional shape. Of course, there are ambiguous pictures – such as the duck–rabbit drawing included in Wittgenstein (2009: part II § xi) or the Necker-cube line drawing (Figures 10.7 and 10.8) – whose surface shapes can be experienced as looking like the outline shapes of several, often incompatible things. But even then, there is a strict limitation on what we can recognise the pictures as depicting; and this limitation is explained by our capacity to recognise similarities between surface and outline shapes. For instance, we cannot see the

222 Fabian Dorsch

Figure 10.7 Unknown artist, duck–rabbit figure.

Figure 10.8 Unknown artist, Necker cube.

duck–rabbit picture as a depiction of a horse, say, because the shape of the outline on the drawing’s surface does not look like the outline shape of a horse. More generally, among the kinds of object that we are capable of visually identifying, hardly any share exactly the same outline shape relative to a given point of view.18 The Experienced Resemblance View fares, however, much worse with respect to the Question of Depth (which, it must be added, it was not explicitly designed to answer). Since, as noted earlier, pictures are always pictures of three-dimensional arrangements of objects, recognising a flat surface as a picture requires an awareness of depth with respect to what is depicted. But the proposed experience of resemblance is limited to the two-dimensional outline shapes of the depicted objects. Hence, it cannot constitute or contain the awareness of depth that is involved in the recognitional fold of seeing-in. As a

Seeing-in as aspect perception


result, the Experienced Resemblance View – at least in its current form – does not have the resources to capture this central aspect of the recognitional fold and is thus unable to fully capture the nature of pictorial experience.

17 The Imagination View does much better on this count. At the heart of this view is the idea that our awareness of what is depicted is imaginative. After all, the recognitional fold is like imagination and unlike perception in that we do not experience the depicted objects as really being there in front of our eyes. The kind of imagining in question, however, can be neither visualising nor propositional imagining (i.e. imaginative thinking). The main reason for this is, as argued earlier, that a unitary account of seeing-in is to be preferred over a divisive one. But while seeing-in should thus be taken to involve two interdependent folds, instances of visualising and propositional imagining can occur independently of other mental episodes, notably perceptions of flat surfaces. In addition, the fact that seeing-in constitutes a single, unified experience already speaks against the identification of the imaginative element with visualising or propositional imagining, given that instances of the latter do not form such unities together with simultaneous perceptions of suitable flat surfaces.19 Besides, propositional thought – even if it is about visual appearances – does not show the visual and spatially perspectival character of our awareness of what is depicted. Instead, Kendall Walton, the most prominent defender of the Imagination View, has proposed understanding seeing-in as the experience of imagining one’s perception of the picture’s surface to be a perception of what is depicted. Accordingly, when we have a pictorial experience of a portrait of a friend, say, we are said to imagine our actual perception of the portrait to be a perception of our friend. As part of imagining this identity of two perceptions, we also imagine the identity of the respective objects of perception. That is, in our example, we imagine the portrait to be our friend. This becomes evident, for instance, in the fact that it would be perfectly natural for us to say, while pointing to the portrait, ‘there is our friend’ or ‘here we see our friend’ (Walton 1990: 215ff.). This imagination likewise extends to the features of the objects concerned. For instance, when we are scanning the relevant marks of paint on the canvas with our eyes, we imagine this very experience to be a scanning of the face of our friend. And we imagine certain features of the canvas that we thereby see (for example, some colours or shapes) to be (closely related to) features of that face. Walton argues that, in this way, the visuality of our awareness of the depicted can be preserved, given that the ‘vividness’ – for example, the continuity and detailedness – of our real perception of the canvas transfers to the imagined perception of the friend (ibid: chapter 8, especially 296, section 8.2). Moreover, imagining this identity between the real perception of the picture’s surface and the imagined perception of the depicted is meant to ensure

224 Fabian Dorsch that the two folds of seeing-in are united in a single experience with a single (though complex) content, with respect to which the configurational and the recognitional fold are merely abstractions.20 More specifically, Walton’s idea is that, in imagining our perception of the portrait to be a perception of our friend, the imaginative thought of our friend penetrates our perception of the portrait and changes its character and content. In other words, seeing-in involves, for Walton, the cognitive penetration of the visual perception of the picture’s surface by an imaginative thought about what is depicted.21

18 One problem with the resulting Imagination View is, however, that it is difficult to motivate the classification of our awareness of the depicted as an imaginative thought. First of all, there are good reasons to believe that imagining is generally susceptible to the will (McGinn 2004:12ff.; Dorsch 2012: chapter 13f.). But, apart from ambiguous pictures, we cannot actively bring about changes in what we take a given picture to depict: that is, what we imagine ourselves as seeing when looking at the picture. Hence, our awareness of the depicted cannot really be imaginative. In addition, it is questionable whether it can be an instance of thought, for the reason already mentioned, namely that our awareness of the depicted is visual and spatially perspectival, while thought is not. But even if we accept that seeing-in combines perception and imaginative thought, this combination could not involve cognitive penetration. Cognitive penetration changes the perceptual experience concerned with respect to how we perceive the object in question as being (Stokes 2013: section 1). When it suddenly strikes us that the object that we are currently seeing is an engine, say, this thought is likely to enter into our perception of the object. As a result, we begin to experience it as being an engine, which we did not before. But we do not, in any way, experience the canvas before us as being our friend when having a pictorial experience of the portrait concerned. We might perhaps entertain the thought that the canvas is our friend. But this does not influence our perception of the canvas in such a way that we visually experience it as having the property of being our friend. In fact, how we perceive the canvas as being (for example, which colours, shapes and textures it possesses) has not changed. Part of the underlying explanation of this difference is that, while cognitive penetration involves only one object of awareness, seeing-in involves two distinct ones (i.e. the picture and the depicted). For there is no reason for us to expect that thinking of one object as being a certain way has any impact on how we see another object as being.22

19 With respect to our two questions, the Imagination View has, finally, the complementary problem to that faced by the Experienced Resemblance View.

Seeing-in as aspect perception 225 The Imagination View promises a rudimentary answer to the Question of Depth – at least as long as we ignore the problems that have already been noted in relation to the idea of understanding seeing-in as a combination of seeing and imagining. The view claims that we are aware of something as a picture by imagining our experience of it to be an experience of the depicted, which includes imagining the picture’s surface to be the depicted. As part of this imaginative identification, we imagine the depicted arrangement of objects to be threedimensional in one way or another – say, as consisting of a black circle in front of a white plane or as involving some voluminous objects, such as our friend and his surroundings. In other words, we are aware of the depth of the depicted scene by imagining it to be three-dimensional; and we imagine it to be this way as part of imagining (our perception of) the two-dimensional surface of the picture to be (a perception of) a three-dimensional arrangement of objects. By contrast, the Imagination View has difficulties in answering the Question of Content. The challenge is to explain why we imagine perceiving X rather than Y. For example, when we look at the portrait of our friend and scan the canvas with our eyes, why do we imagine this to be a perceptual experience of our friend’s face and not, say, of some animal or a tree? Walton’s answer to this question is twofold. First, he proposes that there are certain principles or rules (called ‘principles of generation’ by Walton) which determine what a given picture depicts (Walton 1990: 110). These principles are very similar to those that fix what words or sentences in a certain language mean, or what props in games of make-believe stand for. In particular, they are, at least to a considerable extent, intersubjective (ibid: 23, 68, 301f.), as well as natural and non-arbitrary, rather than conventional and stipulated (ibid: 23, 301, 351f.). Besides, given that they determine what a picture depicts, they also dictate what it is appropriate for us to imagine when looking at the picture and recognising it as a depiction (ibid: 215ff., 293ff.). Second, we normally imagine the depicted in accordance with these principles because we are prompted to do so by perceiving certain properties of the picture’s surface (Walton 2002: 32). Typically, this transition from the perceptual awareness of the picture to the imaginative awareness of the depicted is involuntary and non-inferential (Walton 1990: 216, 311). Moreover, it need not – and usually does not – involve explicit knowledge of, guidance by or reflection on the principles in question (ibid: 139, 185f., 216). Instead, we implicitly understand and accept the principles simply by being disposed to imagine whatever is depicted in response to seeing the relevant properties of the picture’s surface (ibid: 216). As Walton writes: Often it just strikes us that, given the words of a novel or the paint on a stretch of canvas, such and such is fictional (i.e. to be imagined). (Walton 1990: 139)

226 Fabian Dorsch Even assuming that our spontaneous understanding of what is to be imagined can be explained in terms of rule-following, this reply to the Question of Content is at best a partial answer. For there still remains a set of crucial questions: which properties of the surfaces of pictures determine, and prompt us to imagine, what they depict; and what kinds of link, expressed by the principles, obtain between these properties and what is depicted?

20 Walton leaves these questions open and, furthermore, stays neutral on whether our implicit knowledge of the principles is innate or acquired and on whether the principles are open to changes (ibid: 301f.). But he seems to be sympathetic to the natural idea that what is at work here is the awareness of visual similarities between the picture and the depicted, notably in two-dimensional shape (Walton 2002: 31). In fact, the suggestion that the principles are concerned with similarities in surface and outline shape fits also very well Walton’s observation that, as part of seeing-in, we often scan the outlines on the canvas and imagine of this experience that it is an experience of scanning the outlines of whatever is depicted.23 But if it were indeed the case that our imaginative awareness of the depicted is informed by a prior awareness of the resemblance between the picture and the depicted, then it seems that all the elements needed for seeing-in are already there before we start to imagine anything, for although the awareness of resemblance might not be all that there is to seeing-is, it is at least sufficient for the presence of the configurational and the recognitional fold. In particular, being aware of how what is depicted resembles the marks on the picture’s surface in the relevant respects is already sufficient for being aware of the depicted and its features; and there is no need for any further kind of awareness, imaginative or otherwise. In short, imagining the depicted seems already to presuppose seeing-in.

21 Walton might reply that the awareness of resemblance in question is merely subpersonal, while the recognitional fold occurs on the personal level. However, it should be clear that it is often, if not always, possible that we come to personally experience the relevant similarities. In particular, people enjoying a pictorial experience are generally able to report, or point to, the similarities in question when prompted to do so, which again requires that they become aware of those resemblances on the personal level. But if they do start to experience the similarities between the picture and the depicted, there is once more no need for some subsequent imaginative awareness of the latter. Hence, the Imagination View could at best insist that seeing-in involves imagining only with respect to those cases in which our awareness of resemblance actually remains subpersonal. But, apart from the fact that

Seeing-in as aspect perception


such cases would be rather rare (if they occured at all), this would lead to the implausible consequence that there is more than one kind of seeing-in, namely one involving imagining and another involving the experience of similarity in two-dimensional shape. Unless the Imagination View comes up with a better explanation of how what we imagine while having pictorial experiences is directly constrained by what the surfaces of the pictures concerned are like, it is unable to answer the Question of Content without giving decisive ground to the Experienced Resemblance View. Hence, it is best, at least for the time being, to conclude that whether we imagine seeing X or Y when looking at a picture is normally determined by whether we see X or Y in the picture (which again may – but need not – depend on whether we experience the picture as resembling X or Y) and not the other way round. As Walton rightly notes, ‘it seems undeniable’ that someone looking at the picture of a ship, say, typically imagines himor herself looking at a ship (Walton 2002: 217). But this observation is compatible with the view that this form of imagining is not part of, but instead based on and informed by seeing-in.

22 Both the Experienced Resemblance View and the Imagination View fail to satisfactorily answer one of the pair of questions introduced in Section 7. But their shortcomings and, more importantly, their strengths are complementary. In particular, while the Experienced Resemblance View has a satisfactory reply to the Question of Content, the Imagination View suggests an answer to the Question of Depth. So, perhaps it is possible to combine these two elements and maintain that seeing-in involves both an experience of resemblance in two-dimensional shape and a non-perceptual awareness of depth. This latter awareness should thereby not be understood as imaginative, for the reasons spelled out earlier – most notably because imagining is subject to the will and cannot cognitively penetrate a perceptual experience of a flat surface in such a way as to give rise to an experience of seeing-in. Accordingly, seeing something three-dimensional in a two-dimensional surface does not amount to imagining the third dimension of the objects, the outline shapes of which we experience to be similar to the respective surface shapes. Instead, I suggest that we conceive of seeing-in as being an instance of seeing an aspect.24 The aspect in question is the property of having the visual appearance of a three-dimensional arrangement of two- or three-dimensional objects (i.e. the visual appearance of the depicted scene). As part of seeing this aspect, we experience the outline shapes of these objects as looking like certain surface shapes on the canvas. But seeing this aspect also involves being visually aware of depth. More specifically, we see the aspect by seeing the marks on the picture’s surface – or, perhaps more neutrally, the points on, or

228 Fabian Dorsch

Figure 10.9 Ronald C. James, Dalmatian Dog 1965

parts of, that surface – as being visually organised in a certain three-dimensional manner. In the well-known picture of a Dalmatian dog (Figure 10.9), for instance, we perceptually experience the two-dimensional black areas on the white sheet of paper as being the three-dimensional shape of a dog in an autumn landscape. How we see the marks as being visually organised depends in part on the relevant surface shapes. That is, we see the marks as forming a three-dimensional arrangement of objects, the outline shapes of which are experienced by us as resembling the surface shapes. In the example, some of the surface shapes suggested by the black areas look like the outline shapes of the dog. Other properties of the marks on the picture’s surface may also contribute to the visual organisation of the aspect, notably colours and textures. Indeed, examples of inflection can perhaps be accounted for by reference to the fact that certain material qualities of the surface may contribute to the threedimensional visual appearance that constitutes the aspect noticed in seeing-in – for example, when we see the hands of Jan Cornelisz Sylvius, depicted in his two portraits by Rembrandt, as being made of paint or ink rather than flesh (Figures 10.10 and 10.11).25 However, experienced resemblance and experienced inflection are likely to be only two elements among many that are involved in seeing the aspect under consideration.

23 The resulting account – the Aspect View of perceptual experience – gets support from the fact that our awareness of what is depicted shows many of

Figure 10.10 Rembrandt Harmenszoon van Rijn, Posthumous Portrait of the Preacher Jan Cornelisz Sylvius ca. 1645 (National Gallery of Art, Washington).

Figure 10.11 Rembrandt Harmenszoon van Rijn, Sketch for the Posthumous Etching of Jan Cornelisz Sylvius ca. 1646 (British Museum, London).

230 Fabian Dorsch the features characteristic of experiences of aspects. Indeed, the idea that seeing-in involves seeing an aspect is not new:26 When we are looking at an object we sometimes see that it has not changed while we have been looking at it and yet the way in which we see it has changed: we see it differently, although we see that it is no different from how it was. For example, we might pass from seeing a puzzlepicture as mere lines to seeing it as containing a depiction of a face, [ . . . ] from seeing the ambiguous duck-rabbit figure as a duck-picture to seeing it as a rabbit-picture, from seeing [a certain drawing] as an arrow (a sign in the form of an arrow, pointing in a certain direction) to seeing it as depicting a bird’s foot [ . . . ]. In each case we can be said to notice an aspect of what we are looking at. (Budd 1987: 1) In this passage, Budd represents Wittgenstein’s view on the matter, which is likely also his own. According to this view, both the switch from seeing something as a mere surface to seeing something as a depiction (or vice versa) and the switch from seeing something as a depiction of X to seeing it as a depiction of Y are changes in whether, or which, aspects are seen by us. This might seem at odds with my suggestion that the aspect involved in seeing-in is the property of having the visual appearance of a three-dimensional scene rather than the property of being a depiction. But there is no real tension here. It is part of the twofoldness of seeing-in that we recognise a flat surface as a picture by becoming visually aware of a three-dimensional scene over and above, and on the basis of, perceiving the two-dimensional surface. The perception of the surface (i.e. the configurational fold) is concerned with the visible features of the marks on the surface, such as their locations, shapes, colours, textures and so on.27 As a result, this visual awareness stays constant during the various switches just described.28 That is, it does not make us aware of the aspect which we notice when coming to recognise something as a picture. Instead, the awareness of this aspect pertains to the visual awareness of the depicted (i.e. the recognitional fold). And visually experiencing what is depicted just consists in visually experiencing the appearance of a three-dimensional arrangement of objects. The conclusion should therefore be that recognising something as a picture involves seeing an aspect only in so far as, and because, it involves visually experiencing a three-dimensional scene (over and above seeing a flat surface). What changes during the switches in seeing-in is whether we are aware of such a scene or which one we are aware of. The aspect that we notice by means of seeing-in is thus the property of having the visual appearance of a three-dimensional scene.

24 These considerations show that seeing-in satisfies the basic characterisation of seeing an aspect, which the passage quoted from Budd’s text starts off with.

Seeing-in as aspect perception 231 The experience of seeing an aspect is defined by the fact that, while switching from seeing one aspect to another, from seeing no aspect to seeing one, or from seeing an aspect to seeing none, we do not see the object in question as changing, although our experience of it changes. This is exactly how it is with pictures and seeing-in. We do not experience the picture’s surface as changing when shifting from seeing it merely as a flat surface to seeing it as a picture. Nor do we see any changes in the lines of the duck–rabbit drawing while moving back and forth between seeing a duck in it and seeing a rabbit in it. This is due to the fact that our perception of the locations, shapes, colours and textures of the marks on the surface stays constant throughout, irrespective of whether we are also aware of the visual appearance of a threedimensional scene. In other words, the change that we experience during the switches pertains to the recognitional fold, not the configurational one. Indeed, it is a change that occurs in our mind, not in the object – which is in line with the fact that being a picture and having the visual appearance of a three-dimensional arrangement of objects are response-dependent properties. In addition, seeing-in possesses other important features that are central to aspect perception. To start with, seeing-in is an experience which is both visual and immediate and, as such, does not involve any conscious form of reasoning or inference. The same is true of other instances of aspect perception, such as the experience of grouping phenomena. Then, seeing-in allows for the possibility of error. We may fail to recognise something as a picture; we may hallucinate a depicting surface; and our experience may mislead us about what a given picture depicts (for example, whether it depicts a round or an elliptical object). Similarly, we may fail to notice other aspects; we may hallucinate them; and we may err about their nature. Finally, just as it is generally not up to us which aspect we see a given object as having, it is generally not up to us which surfaces we recognise as pictures, and what we thereby see in them. True, we may be able to stop seeing a picture as a picture (for example, by attending to it in a certain way); and we may have some control over whether we see the duck–rabbit drawing as a depiction of a duck or as a depiction of a rabbit. But we cannot bring ourselves to see a surface as a picture if the surface is not such as to give rise to seeing-in in normal viewers under suitable conditions (for example, a random wall of a building); and we cannot see a horse in the duck–rabbit drawing. Indeed, once we have recognised a flat surface as a picture, it may no longer be possible for us to switch back to seeing it as a mere surface. Again, the same applies to other aspects that we see objects as having. As noted earlier, this lack of subjection to the will is part of the reason why both seeing-in and seeing an aspect should not be taken to involve imaginative awareness.

25 The Aspect View can also answer both the Question of Content and the Question of Depth. The answer to the first question is roughly the same as

232 Fabian Dorsch the one given by the Experienced Resemblance View. The reply to the second question points to the fact that the aspect seen is three-dimensional. But, of course, we do not really see depth (there is none), nor do we come to falsely believe that there is a three-dimensional scene before us, or that the picture does possess the visual appearance of such a scene. Rather, our visual awareness of the three-dimensional arrangement of objects is an imperfect visual illusion. Imperfect illusions are illusions that we are able to introspectively distinguish from veridical perceptions and thus normally do not take at face value (for example, by endorsing them in belief or action). Seeing-in involves such an imperfect illusion because we can tell from the inside that it differs from seeing a three-dimensional scene face-to-face, especially with respect to the dimension of depth. We may still sometimes – particularly when concerned with very realistic depictions that only just fall short of being a trompe-l’oeil – feel an inclination to judge that there is a three-dimensional scene before us. And this might perhaps explain why some people have thought that pictorial experience involves, or comes very close to, a perceptual illusion of whatever is depicted (for example, Gombrich 1960: 3ff.). But, unless we are concerned with trompe-l’oeils, such inclinations are typically undermined or outweighed by our recognition of the illusory character of our awareness of depth and volume. Nor is it more generally unusual that aspect perception is illusory with respect to the experienced aspect (which is why it may not always be appropriate to speak of seeing an aspect). Contour illusions like the Kanizsa and Ehrenstein illusions are good examples (Figures 10.12 and 10.13). We see

Figure 10.12 Unknown artist, Kanizsa’s triangle.

Seeing-in as aspect perception


Figure 10.13 Unknown artist, Ehrenstein illusion.

the incomplete black circles as being grouped together in such a way as to form a white triangle, and the thick black lines as being organised in such a way as to circumscribe a white circle. But there is really neither a white triangle nor a white circle before us. Hence, our perceptual experiences of these shape aspects are illusory. The same is true in the case of seeing-in, only that here the illusion is more widespread and concerns most aspects of the visual appearance of the three-dimensional scene in question, notably its dimension of depth.

26 Finally, the Aspect View has the resources to capture the characteristic features of seeing-in. First, the kind of aspect perception under consideration involves the visual awareness of the picture’s surface and its visible features. Second, it also makes us aware of the three-dimensional scene that is depicted. In particular, this awareness is visual in the same way as many other experiences of aspects. Third, our awareness – and probably also our visual attention – is always divided between the surface and the aspect, given that we cannot stop being aware of one of them without also ceasing to have an experience of seeing-in. Fourth, our awareness of the surface and our awareness of the aspect are not independent of each other but form a single, unified experience. In particular, we cannot experience the aspect without also seeing the surface; and our experience of the surface differs introspectively relative to whether we notice the aspect. Fifth, as already mentioned, experiencing the aspect in question involves experiencing depth.

234 Fabian Dorsch Furthermore, the Aspect View is unitary in that it claims that seeing-in has a single but complex content, according to which the visible features of the marks on the surface and the visual appearance of the depicted scene stand in a certain relation. This relation – which I have described, in want of a better word, as a relation of organisation – is an instance or determination of the relation that, more generally, holds between any aspect and the relevant visible lower-level features. Evidently, more needs to be said in elucidation and defense of the Aspect View. But the preceding considerations should have made clear that this view has at least one advantage over its unitary rivals: it can accommodate both the experience of resemblance in two-dimensional shape and the experience of the third dimension of depth by understanding seeing-in in terms of aspect perception.29

Credits With the exception of Figures 10.4 and 10.9, all images have been made available by Wikimedia Commons and are included under the Creative Commons Public Domain Mark 1.0 licence. Figure 10.4 is from wikiHow and is used under the Creative Commons Attribution-NonCommercial-ShareAlike 2.5 Generic licence. Figure 10.9 was originally published in Life Magazine 58 (7) 19th February, 1965.

Notes 1 The idea that pictorial experience is an instance of seeing an aspect can also already be found in Wittgenstein’s writings, albeit not yet combined with an understanding of pictorial experience in terms of seeing-in in Wollheim’s technical sense (see, for example, Wittgenstein 2009: part II §xi; 1958, 163ff.; Budd 1987, 1991: chapter 4) for discussion. 2 My notion of pictorial experience (and seeing-in) is thus narrower than Lopes’s (2005) and Newall’s (2009) notions, which are also meant to apply to our standard experiences of illusionist and actualist pictures (i.e. experiences which are introspectively indistinguishable from potential face-to-face perceptions of the threeand two-dimensional objects that those pictures depict). Besides, I concentrate in what follows exclusively on two-dimensional pictures. See Hopkins (2004), 2010a) and Martin (2012) for discussions of whether there are also three-dimensional pictures. 3 According to Wollheim, at least the status as a depiction also depends on the artist’s intention to produce a picture. Clouds in the sky or marks on a wall do not depict something, even if we can ‘see’ something ‘in’ them, because they have not been intended to give rise to such an experience (Wollheim 1987: 47ff.). 4 See McDowell (1998) for this distinction between two senses of objectivity. 5 At least Wollheim’s earlier view seems to be even stronger, in that he appears to claim that twofoldness concerns the distribution of visual attention and not only mere visual awareness (Wollheim 1980: 216). 6 In particular, this idea is accepted not only by the proponents of the Experienced Resemblance View and the Imagination View (both to be discussed later) but also – at least to some extent – by philosophers who are more critical of the orthodoxy,

Seeing-in as aspect perception


8 9


11 12

13 14



17 18


such as Dom Lopes, who argues that seeing-in is only one form of pictorial experience among others in order to be able to capture our standard experiences of illusionistic, naturalistic and actualist pictures (Lopes 2005: chapter 1), or John Kulvicki, who likens pictorial experience to perceptual experience as conceived by the sense-data theorists (Kulvicki 2006: chapters 8, 10). Hopkins uses this term to also describe views – like the theory defended in Gombrich (1960, especially 3ff.) – that deny that pictorial experience consists in twofold seeing-in and instead construe pictorial experience as the conjunction or simultaneous occurrence of two distinct experiences. See Sections 13 and 17 for further discussion and references. Budd (2008b: section 5) too seems to read Wollheim in this way, given that he does not consider the option of accounting for the difference between the two folds in terms of a difference in kind of experience (i.e. perceptual vs. non-perceptual), rather than in terms of a difference in content (i.e. two- vs. three-dimensional, or one kind of depth vs. another). Walton (1992) seems to maintain that the configurational fold can be of the same kind as a normal perception of a flat surface. The underlying thought appears to be that even normal perception is always penetrated by the thought ‘I experience X’, thus ensuring that it involves the same kind of cognitive penetration as seeing-in. See Sections 17–18 for a discussion of Walton’s view. I do believe that visualising, taken on its own, never makes a claim about what reality is like and never leads to belief in a rational person but, for the current argument, it suffices if this claim holds in typical circumstances. See Peacocke (1987), Budd (2008a), Hopkins (1998) and Hyman (2006). Note, however, that the last of these philosophers does not account for depiction in terms of experienced resemblance and thus does not endorse the Experienced Resemblance View. Hyman (2006) argues that the two are also experienced as being similar in certain aspects of colour. Here, it suffices to focus on experiences of resemblance in two-dimensional shape because of their close link to depth awareness. In light of the earlier observation that painted surfaces are hardly ever entirely flat and two-dimensional, the surface shapes are, strictly speaking, also outline shapes, relative to a frontal point of view on the surface. But this complication can be safely ignored in what follows. See Hopkins (1998: 115). With some qualifications, this characterisation of outline shapes captures the core idea of Hopkins’s own notion of ‘outline shape’ (the plane is located between object and viewer), of Hyman’s notion of ‘occlusion shape’ (the plane is located behind the object) and even of Budd’s and Peacocke’s notion of ‘visual field shape’ (the plane is identical with what is left over if we abstract the dimension of depth from the three-dimensional scene that we experience (Budd 2008b: 208; 2008a: 221). In this respect, outline shapes resemble the apparent or variable shapes that figure in our perceptual experiences of the constant shapes of objects. A coin perceptually appears to look round (i.e. to have a certain constant shape) irrespective of which perspective we look at it. But our perceptual awareness of the coin and its constant roundness also involves, in some way or another, the variable property of being elliptical, which changes from one point of view to another. This variable shape property is merely apparent since the coin does not perceptually appear to possess it and since, relatedly, our perceptual experience of the coin neither inclines, nor entitles us to ascribe it in belief to the coin. See (Budd 1992) and (Hopkins 1998, 20ff.) for this and other arguments against the Imagination View, which I discuss in more detail in Section 17. See also the discussion in Hopkins (1998: section 5.7), which argues that, even if many different things share outline shapes, this does not undermine the success of

236 Fabian Dorsch

19 20



23 24

25 26

27 28 29

the Experienced Resemblance View in explaining what the difference is between seeing X and seeing Y in a given picture. See Budd (1992) and Hopkins (1998: 20ff.) for discussion. See (Walton 1990, 295 and 301). In fact, Walton acknowledges that the configurational fold can occur on its own (see note 10). But it is still true for him that the recognitional fold requires the configurational fold, given that the imagining involved in seeing-in is de re with respect to the perception of the picture (Dorsch 2015b). Moreover, as already mentioned in Section 6, the experience of a surface as a mere surface and the experience of the same surface as a picture are very unlikely completely indistinguishable from the inside. See Walton (1990: 295, 300). Indeed, Walton thinks that visual experience – including perception – is in general cognitively penetrated by the self-referential thought ‘I am experiencing X’ (ibid: 295 (note 3)). And he likens his own view of seeing-in to Wittgenstein’s notions of ‘seeing as’ and ‘aspect seeing’ as fusions of seeing and thinking (ibid: 295 (note 2), 351), which suggests that his view might be closer than initially recognisable to the one that I will put forward in Section 22ff. In any case, Walton’s conception of seeing-in as consisting in cognitively penetrated perception supports his repeated insistence that seeing-in is an instance of perception, albeit of imaginative perception (Walton 1990: 302; 2002 (note 37)). As I discuss in Dorsch (2015b), the account of pictorial experience defended in O’Shaughnessy (2003: 346ff.) is very similar to Walton’s (see also O’Shaughnessy 2012). There is also the problem that cognitive penetration seems to require sameness in attitude or commitment. Accordingly, while judgemental thoughts are limited to entering into perceptual experiences, imaginative thoughts are restricted to merging with imaginative experiences. For instance, when we notice the visual resemblance between a bouncer at a club and a gorilla and, as result, see him under the concept of a gorilla, we do not see him as being a gorilla (because we do not believe him to be a gorilla) but rather see him as looking as a gorilla (because we do take him to be visually similar to a gorilla). See Dorsch (2015b) for further discussion. See Walton 1990: section 8.2) for a detailed description of how our imaginative engagement with the depicted may be grounded in, and correspond to, our perceptual engagement with the picture. Voltolini (2012; 2015) also defends the idea that seeing-in involves seeing an aspect, although he maintains that aspect perception already occurs on the level of the configurational fold. A comparison of his view with mine has to wait for another occasion. See, for instance, Hopkins (2010b) for a general discussion of inflection. The omitted parts of the passage make clear that Wittgenstein (and Budd) also think(s) that experiencing a similarity amounts to seeing an aspect. This strengthens my suggestion that the experience of resemblance in two-dimensional shape is part of seeing the aspect of a three-dimensional scene. It may also involve the perception of more higher-level properties, most notably grouping phenomena and other Gestalt properties. But these aspects are different from the aspect central to the recognition of something as a picture. Or at least almost constant, given that what changes is whether the awareness in question serves as the base for a dependent second kind of awareness or not. I am very grateful to Malcolm Budd, with whom I had the great pleasure of discussing in much detail a previous version of this chapter, as well as to Robert Hopkins and the two editors of this volume. My research on this chapter was generously funded by the Swiss National Science Foundation (PP00P1139004) and the Fundación Séneca (18958/JLI/13).

Seeing-in as aspect perception 237

References Budd, Malcolm. 1987. ‘Wittgenstein on Seeing Aspects’. Mind 96 (381). JSTOR: 1–17. —. 1991. Wittgenstein’s Philosophy of Psychology. London: Routledge. —. 1992. ‘Review of “Mimesis as Make-Believe: On the Foundations of the Representational Arts” by Kendall L. Walton’. Mind 101: 195–8. —. 2008a. ‘The Look of a Picture’. In Aesthetic Essays, 216–38. London: University College London. —. 2008b. ‘On Looking at a Picture’. In Aesthetic Essays, 185–215. Oxford: Oxford University Press. Dorsch, Fabian. 2012. The Unity of Imagining. Berlin: De Gruyter. Gombrich, Ernst H. 1960. Art and Illusion: A Study in the Psychology of Pictorial Representation. London: Phaidon. Hopkins, Robert. 1998. Picture, Image and Experience. Cambridge: Cambridge University Press. —. 2003. ‘What Makes Representational Painting Truly Visual?’ In Aristotelian Society Supplementary Volume 77: 149–67. Wiley Online Library. —. 2004. ‘Painting, Sculpture, Sight, and Touch’. The British Journal of Aesthetics 44 (2): 149–66. —. 2010a. ‘Sculpture and Perspective’. The British Journal of Aesthetics 50 (4): 357–73. —. 2010b. ‘Inflected Pictorial Experience: Its Treatment and Significance’. In Philosophical Perspectives on Depiction, edited by Catharine Abell and Katerina Bantinaki, 151–80. Oxford: Oxford University Press. Hyman, John. 2003. ‘Subjectivism in the Theory of Pictorial Art’. Monist 86 (4): 676–701. —. 2006. The Objective Eye: Color, Form, and Reality in the Theory of Art. Chicago: University of Chicago Press. Kulvicki, John V. 2006. On Images: Their Structure and Content. Oxford: Oxford University Press. Lopes, Dominic. 2005. Sight and Sensibility: Evaluating Pictures. Oxford: Oxford University Press. Martin, M. G. F. 2012. ‘Sounds and Images’. The British Journal of Aesthetics 52 (4): 331–51. McDowell, John. 1998. ‘Values and Secondary Qualities’. In Mind, Value, and Reality, 110–29. Cambridge, MA: Harvard University Press. McGinn, Colin. 2004. Mindsight. Cambridge, MA: Harvard University Press. Newall, Michael. 2009. ‘Pictorial Experience and Seeing’. The British Journal of Aesthetics 49 (2): 129–41. O’Shaughnessy, Brian. 2003. Consciousness and the World. Oxford: Oxford University Press. —. 2012. ‘Seeing an Aspect and Seeing Under an Aspect’. In Wittgenstein and the Philosophy of Mind, edited by Jonathan Ellis and Daniel Guevara. Oxford: Oxford University Press, pp. 37–60. Peacocke, Christopher. 1987. ‘Depiction’. The Philosophical Review 96 (3): 383–410. Stokes, Dustin. 2013. ‘Cognitive Penetrability of Perception’. Philosophy Compass 8 (7): 646–63. Voltolini, Alberto. 2012. ‘Towards a Syncretistic Theory of Depiction’. In Perceptual Illusions. Philosophical and Psychological Essays, edited by Clotilde Calabi. London: Palgrave Macmillan, pp. 166–94.

238 Fabian Dorsch —. 2015. A Syncretistic Theory of Depiction. London: Palgrave Macmillan. Walton, Kendall. 1990. Mimesis as Make-Believe. Harvard: Harvard University Press. —. 1992. ‘Seeing-in and Seeing Fictionally’. In Psychoanalysis, Mind and Art, edited by Jim Hopkins and Anthony Savile, 281–91. Oxford: Blackwell. —. 2002. ‘Depiction, Perception, and Imagination: Responses to Richard Wollheim’. The Journal of Aesthetics and Art Criticism 60 (1): 27–35. Wittgenstein, Ludwig. 1958. The Blue and Brown Books. Oxford: Blackwell. —. 2009. Philosophical Investigations, edited by P. M. S. Hacker and Joachim Schulte. 3rd edition. Oxford: Wiley-Blackwell. Wollheim, Richard. 1974. ‘On Drawing an Object’. In On Art and the Mind. Cambridge, MA: Harvard University Press. —. 1980. Art and Its Objects: With Six Supplementary Essays. 2nd edition. Cambridge: Cambridge University Press. —. 1986. ‘Imagination and Pictorial Understanding’. Proceedings of the Aristotelian Society, Supplementary Volumes 60. JSTOR: 45–60. —. 1987. Painting as an Art. London: Thames & Hudson. —. 1998. ‘On Pictorial Representation’. The Journal of Aesthetics and Art Criticism 56 (3): 217–26.

Part V

Imagination and emotion in Wollheim’s account of pictorial experience

This page intentionally left blank

11 Wollheim’s ekphrastic aesthetics Emotion and its relation to art Michael Levine

The role of emotion is to provide . . . the person – with an orientation, or an attitude to the world . . . That emotion rides into our lives on the back of desire is a crucial fact about emotion, as well as a crucial fact about us. – (Richard Wollheim 1999: 15) If artworks can arouse emotions, it is not likely that these emotions will have a history anything like the one Wollheim proposes. – (James Young 2001: 337)

Introduction After revisiting Wollheim’s account of ‘seeing-in’, this chapter examines Wollheim’s (1999) theory of emotion in relation to aspects of his aesthetics – including imagination, expression, appreciation and value. Wollheim’s account of emotion, embedded in psychoanalysis, came after much but not all of his writings on art.1 One would expect to find that his aesthetics and interpretation of paintings are intimately or indissolubly linked to his general theory of the emotions. Yet neither Wollheim himself, nor those who focus on his aesthetics, particularly on the phenomenology of seeing-in and twofoldness, as well as his account of intentionality, interpretation, and imagination, has made anything like a satisfactory effort to trace these. Or so I shall argue. The first section, ‘Seeing-in and aesthetics’, revisits Wollheim’s account of seeing-in. He articulates the experience of seeing-in, which he calls an innate visual capacity, as important and even central to art. However, the significance of this capacity and its place in Wollheim’s aesthetics is questioned. Minimally, he has overgeneralised the experience. ‘Wollheim: aesthetics and emotion’ makes the case for a close connection between Wollheim’s aesthetics and his theory of the emotions. ‘Representation is perceptual’, according to Wollheim, but perception is suffused with an orientation to the world that only emotion can provide. This is different from the mapping of the world that belief and presumably knowledge provide. If so then we can speak of emotion as a condition of representation and interpretation in art. Furthermore, emotion in this view is not just any condition or even a necessary condition or such representation. It is central.

242 Michael Levine ‘Wollheim’s aesthetics and psychoanalysis’ gives an account of the psychoanalytic basis of Wollheim’s aesthetics and its importance. He draws from both Freudian psychoanalysis and Melanie Klein’s object-relations theory. Wollheim, like Freud, emphasises the artist’s intentions in interpreting works of art, and he too holds the view that art is orectic, driven by desire and wish-fulfilment and functioning so as to satisfy certain psychological needs of artists and audiences. This is what makes the artist’s intentions – psychoanalytically understood, so crucial for interpretation. Wollheim, like Freud, sees art as deeply integral to a life. In the final section, ‘Music cannot convey emotion? I don’t think so’, I argue that although Wollheim’s focus in aesthetics is on painting, his theory of emotion in relation to aesthetics can be applied more widely. It can help address the question of whether music can make one a morally better person. An aural counterpart to seeing-in is postulated. On Wollheim’s view, emotion colours perception and influences what is seen-in the object. If so, can it also influence what is ‘heard-in’ the object? This section argues for a connection not only between Wollheim’s account of the emotions and aesthetics, but also between aesthetics and ethics. Throughout this chapter I argue that seeing-in is neither particularly interesting nor central to Wollheim’s views – or to what is valuable and, I think, lasting in his aesthetics. The role of emotion and the ways in which art can only be understood via psychoanalysis are more important to his outlook. This is explicit in Painting as an Art, The Thread of Life and On the Emotions. The second to fourth sections are intended to support these claims and are therefore important to the tale.

Seeing-in and aesthetics In Noël Carroll’s 2010 Richard Wollheim Memorial Lecture, ‘Art Interpretation’ (Carroll 2011), there is no mention of Wollheim’s theory of emotion (the word ‘emotion’ is not to be found) let alone how it is indispensably related to Wollheim’s theory of the interpretation of painting, as well as to his actual interpretations. Given the connection between his account of emotion and psychoanalysis, and the centrality of psychoanalytic theory to Wollheim’s aesthetics2 and to his theory of emotion, it seems odd that seeing-in has garnered a far greater amount of philosophical attention than his more aesthetically (and philosophically) significant account of emotion.3 This disconnect between integrally related streams of Wollheim’s work is significant. If, unlike Jerrold Levinson (2011: 33)4 and many other analytic aestheticians, one thinks that, on the one hand, the phenomenological analysis of seeing-in – ‘twofoldness’ – as an account of how we perceive representations is aesthetically uninteresting, on the other, connections between his aesthetics (indeed, aesthetics generally) and his psychoanalytically grounded theory of the emotions are both interesting and essential to understanding the most important aspects of his aesthetics. Even if it is at times phenomenologically

Wollheim’s ekphrastic aesthetics


accurate, seeing-in is, in my view, largely inconsequential to the important issues of aesthetics. Levinson says: I agree with Wollheim that the concept or pictorial representation, or depiction, cannot be explicated without appeal to a characteristic sort of experience, the sort of experience Wollheim has denominated ‘seeing-in.’ Sustaining an appropriate seeing-in experience, that is, a seeing-in experience that conforms with the artistic intention governing a given picture, is what is criterial of such representation and not anything else. (2001: 28) Suppose it is ‘criterial’ in some sense. Is it also aesthetically important? Why? If the value of a work of art is to be explained in terms of the experience of a work of art, and part of that involves seeing-in, then seeing-in is phenomenologically relevant. Nevertheless, while (arguably) artistic intention is crucially important to issues like understanding and appreciation, it is less clear why the phenomenology of the experience – which is what seeing-in is about – is important (or so important). It isn’t even clear that in Wollheim’s account seeing-in is supposed to be particularly or centrally relevant to important issues (below) in aesthetics.5 Although there is a great deal of literature criticising seeing-in and some supporting it, there is a scarcity of literature discussing its significance. This supports my view. The important issues are about the nature, purpose and value of art – its significance – as well as how art should be judged, understood, interpreted, and the attraction it holds for us. What role does the phenomenology of seeing-in have in relation to any of these? Wollheim’s account of emotion is crucially relevant to his treatment of these issues and of course if he is right, to the issues themselves. Like Freud, Wollheim’s concern is to understand and interpret art as well as its significance, function and value – as well as how these are related (see ‘Wollheim’s aesthetics and psychoanalysis’). Nevertheless, Freud had a narrower conception of understanding art (though not artists) than Wollheim. Freud was mostly interested in the unconscious content in a work and, so, silent on music, pictorial qualities, evaluation, etc. The idea that Wollheim’s account of seeing-in is aesthetically uninteresting will strike those who see it as central to his aesthetics as either extravagant, mistaken or both. It does need to be modified somewhat given that Wollheim (Savile and Wollheim 1986) discusses seeing-in in relation to issues in aesthetics including imagination, intention, representation and resemblance and understanding and interpretation. Still, the claim that his account of seeing-in is relatively unimportant is meant to suggest that it plays an insignificant role with regard to these other issues as well. But it is not just that the thesis of seeing-in is uninteresting, it is also that it is mistaken. And although an account of why it is mistaken is largely irrelevant to the connections between his aesthetics and

244 Michael Levine his theory of emotion, given the focus of this volume some explanation as to why Wollheim’s account is unconvincing as a general account of visual experience is worth giving. He has the phenomenology wrong or at least overgeneralised it. Wollheim gives the following account of the ‘experience of “seeing-in”’: [W]e have a perfectly good explanation of how we perceive representations . . . in terms of a very specific visual capacity which we humans have-innately . . . and which is distinguished by the special phenomenology which the visual experiences that manifest it display. I call the visual capacity ‘seeing-in’, and I call the phenomenological feature distinctive of experiences of seeing-in ‘twofoldness’, and by ‘twofoldness’ I mean this: When I look at, say, the representation of a woman, it is to be expected that my visual experience will have two aspects to it: on the one hand, I recognize or identify a woman, and, on the other hand, I am aware of the marked surface, and both the recognition or identification and the awareness are visual. When my experience satisfies this description, I may be said to see a woman in the marked surface or in the picture . . . . my account of what it is to see representations specifically distinguishes itself from accounts that posit not one experience with two aspects but two experiences . . . On the account of seeing-in that I favour, seeing-in-the experience, of course, not the capacity-has two aspects to it. I dub them the recognitional aspect and the configurational aspect, and it is my belief that no systematic account can be given of how the two aspects correlate or how the marked surface has to be or to seem for a given thing or event to be perceived in it.6 (Savile and Wollheim 1986: 46–7) Wollheim does not say why he claims ‘no systematic account can be given of how the two aspects correlate’, but we can postulate a reason. His account of seeing-in precludes any correlation simply because there is no deeper, more fundamental explanation for what he claims is phenomenologically the case. It is, as he says, an innate visual capacity, presumably one in which the two aspects are a fundamental (phenomenological) given and for which no further explanation, other than perhaps some biochemical and physical one, can be given. But if Wollheim’s phenomenological account of how we perceive representations is descriptively and theoretically mistaken (it is certainly overgeneralised), then another explanation may be in order. A systematic correlation of how the two aspects of seeing-in is precluded because, depending on the perceiver’s focus, their conscious awareness (or lack thereof) and how, phenomenologically speaking, they happen to be experiencing the representation (the picture), any correlations will vary. There will be times when the experience that Wollheim describes as essentially ‘two-fold’ cannot – on phenomenological grounds – properly be

Wollheim’s ekphrastic aesthetics 245 described as such because, for example, one won’t even be aware of the configurational aspect. One simply sees the picture of the cat without at the same time being aware, at least not consciously, of experiencing the paper on which it is drawn or how it is drawn. Of course one could answer that if one is not even tacitly or unconsciously aware of the configuration, then this is just a case of illusion. But this is to overtheorise the experience. The only proof or argument one needs to show that Wollheim’s account fails as a general account, though descriptively it is phenomenologically correct in some cases, is introspection. This is surely a principal reason why many disagree with Wollheim’s account of seeing-in. Although Levinson (2001: 29) says he is largely in agreement with Wollheim’s account (‘the picture of picturing that I uphold remains, in broad outline, a recognizably Wollheimian one’), he goes on to question whether ‘the experience of seeing-in [is] in fact necessarily characterized by twofoldness, that is simultaneous awareness of medium and of subject, such that seeing-in has always a configurational as well as a recognitional aspect?’ Furthermore, he asks: What can be said about the recognitional awareness that is arguably at the core of seeing-in, especially if configurational awareness, or awareness of medium, is not always present as well . . . Is seeing-in really the same phenomenon or mental state across all the sorts of things it is said can be seen in pictures? (2001: 29) Levinson says: [I]t would seem reasonable to include within the ambit of pictorial seeing, that is, seeing of the kind normative for pictures understood and appreciated as pictures, both seeing where there is simultaneous awareness of design and content (or twofoldness), and seeing in which there was alternation back and forth between phases of simultaneous awareness of design and content and phases of exclusive or nearexclusive focus on one or the other. It seems that our knowing engagement with pictures does in general display an alternation between phases of simultaneity, often sustained without deliberateness, and ones of switching, often occasioned by deliberate reflection on what one’s experience is like. (2001: 33) In other words, and contrary to Wollheim, twofoldness is just part of the story of pictorial seeing. Sometimes it is the part that is significant and sometimes (mostly) it is not.

246 Michael Levine These and the other questions he raises (2001: 29–30), and the answers he gives, leave one wondering how Levinson can describe his views as even ‘broadly’ Wollheimian. Levinson concedes as much: There is a real question whether the experience I have continued to refer to as seeing-in to as simple seeing-in should in fact be so called . . . [First] the association of seeing-in-which is, after all, a term of art introduced by Wollheim – and twofoldness is so entrenched that an experience of seeing-in sans twofoldness sounds almost oxymoronic. . . . [Second] conceiving such experiences as the ‘seeing-in’ label encourages one to do, as a matter of seeing things in surfaces, does undeniably occasion strain where trompe l’oeil pictures are concerned since in such cases the surfaces are, by hypothesis, neither seen nor seeable. One might thus concede that the visual experience of pictures I have been calling simple seeing-in, and that is present even when twofoldness is not, might in certain cases with more justice be called seeing-from.7 (2001: 37 (note 13)) There are however difficulties with Levinson’s critique, his phenomenology, as well. He says I agree with Wollheim, as against Budd, that seeing-in is generally prior to, and not to be analyzed in terms of, the perceiving or resemblance as such, whether between objects or experiences. The fundamental rationale for so insisting is this. Though perception of resemblance, or more narrowly, structural isomorphism, between object aspects or visual fields, may be a concomitant, trigger, or consequence of seeing-in, it is not equivalent to seeing-in. Seeing-in can occur without such perceptions, and vice versa, and so there can be no identifying them. The experiences of perceiving resemblances and seeing-things-in-other-things are different, and irreducibly so; the former is inherently relational and comparative, the latter not.8 (2001: 28) Granted that perceiving resemblances is not equivalent to seeing-in; but even if seeing-in generally precedes perceiving resemblance, how does this provide a reason for supposing that seeing-in rather than perceiving resemblance gives the correct phenomenological account of visual perception in paintings in most let alone all cases? Introspection with regard to my visual experience does not confirm that I have an experience of seeing-in (though I understand what it is as Wollheim describes it) rather than the experience of perceiving resemblance when I view a painting. Nor do I see the perception of resemblance or structural isomorphism as being ‘a concomitant, trigger, or consequence of seeing-in’. Why suppose it is?

Wollheim’s ekphrastic aesthetics 247 Levinson says: We may observe, in addition, that were seeing-in to be identified with perception of structural isomorphism, then since the latter is clearly a degree notion one would expect the former to be as well. But seeing-in is not evidently a notion of degree, nor is that of depiction, which seeing-in underwrites; seeing-in and depiction are closer, if anything, to being on-off or all-or-nothing affairs. (2001: 28) While I don’t see why one would want to identify seeing-in, as Wollheim describes it, ‘with perception of structural isomorphism’ or perception of resemblance, I also don’t see why seeing-in cannot also be understood as a matter of degree, a degree either of awareness of the two aspects (or two experiences) in the one experience of seeing-in, or else of one’s recognition of the representation for example. And why isn’t ‘depiction’ to be understood as a matter of degree (a ‘degree notion’)? It is quite ordinary to say x depicts y to some extent. Indeed, depiction is quintessentially a matter of degree – even when it comes to photographs. Either my phenomenological introspection is not as finely tuned as Wollheim’s and Levinson’s, or else they are restricting themselves to certain cases in which they are phenomenologically aware of just what they claim to be in those cases. But then why generalise what may be true in some cases of perception of representational painting to all cases? Continuing, Levinson says: What is likely true in this matter is that a nonzero degree of structural isomorphism between a representation and its subject is required for seeing-in to take place, that is, that some such isomorphism may be a causal precondition of seeing the subject in the representation; the mechanisms whereby seeing-in – a kind of seeing after all - is enabled to occur seem to require as much. But even if that is so, the perception of such isomorphism, as opposed to its mere existence, remains strictly unnecessary to the occurrence or the distinct experience of seeing-in. (2001: 28–9) How does Levinson know this? On what basis does he conjecture that structural isomorphism ‘may be a causal precondition of seeing the subject in the representation’ and ‘is required for seeing-in to take place’? The claim is an empirical one and cannot be based on introspection only grounded in the relevant science, a field of psychology or visual physiology and cognition. Structural isomorphism may have nothing to do with such recognition. Puzzling too is Levinson’s claim that even if structural isomorphism is ‘a causal precondition of seeing the subject in the representation’, then nevertheless

248 Michael Levine ‘the perception of such isomorphism, as opposed to its [the isomorphism’s] mere existence, remains strictly unnecessary to the occurrence or the distinct experience of seeing-in’. Levinson’s caveat (2001: 37 (fn 10)) may be relevant: ‘To say it does not involve visual awareness off the medium is not, of course, to say that it does not involve visual processing of the information embodied in the medium’.9 But, again, whether the perception of such isomorphism is necessary or not seems an empirical (scientific) issue. If perception of structural isomorphism is ‘a causal precondition’ rather than the mere presence of such an isomorphism, then such perception is at least causally necessary to seeing-in. It would seem likely furthermore that if such isomorphism is causally necessary to seeing-in then so too would be the perception, on some level, of such isomorphism. Ordinary language, reflecting introspection, may also be employed as evidence against Wollheim’s account of ‘twofoldness’ as an accurate phenomenological account of the experience of ‘how we perceive representations’, or what he terms ‘a very specific visual capacity which we humans have-innately’, that of ‘seeing-in’. When I see a picture (or photograph) of a cat, I simply say I see a cat – and that is because that is what I see. I know it is not a ‘real cat’ but a representation (or photograph, which is a different kind of representation than a painting). But I don’t say anything about the ‘twofoldness’ of my perceptual experience or usually give any indication that I am aware of simultaneously perceiving – in the same overall experience – the paper on which it is drawn. I may be, but I may not be. When I look at the painting of a cat I see a cat – though I will not try to either feed or brush it. Wollheim’s insistence that the two aspects of seeing-in are contained in the one experience is a way of supressing the phenomenological fact that in many cases at least ‘twofoldness’ is not present but is instead assumed or ascribed. Wollheim’s account (Savile and Wollheim 1986: 47–8) of how the ‘permeability of perception by cognition applies to both aspects of the experience of seeing-in’ also strikes me as either prescriptive or commonsensical. After all, one needs the concept of a cat to experience some representation as a cat. But is it true, always true, as Wollheim claims (Savile and Wollheim 1986: 48), that how one ‘sees the marked surface and how he sees what he sees in the marked surface are determined partly, of course, by how the surface is, but partly . . . by the concepts and beliefs that the spectator has and mobilizes’? (I can, for example, imagine beliefs sometimes playing no part.) And if it is always true, is this due to the twofold nature of the single experience of seeing-in or has Wollheim made it an analytic truth determined by his account of what ‘seeing-in’ means? Wollheim’s account may be more accurate if it is meant as an account not merely of conscious perception, or of what it is we are aware of when we perceive a depiction, but also of perceptions that we are not consciously aware of. But even here I am not convinced. I do not think that when I consciously perceive (recognise) the cat in the painting of a cat I must at least be unconsciously aware of what it is painted on or its configurational aspects.10

Wollheim’s ekphrastic aesthetics 249 Moving to the question of the role that imagination may or may not play with regard to perception of representation in pictures – an issue also relevant to the role of emotions in Wollheim’s aesthetics – Wollheim addresses a question raised by Savile. He asks: might we not, at any rate with certain pictures, be required at some moment to stop looking at them and to start imagining something about their content so that then, at a later point, our perception of what we see in them will be enriched, deepened, and made more secure, by what we have imagined? (Savile and Wollheim 1986: 48) Both Savile and Wollheim answer affirmatively though their reasons differ – and yet the discussion seems unnecessary. The answer is not simply ‘yes’ – but ‘of course’. Why would the case of pictures be any different, say, to those of fiction, drama or film, where imagination suffuses, shapes and even alters narrative or pictorial representation, as the case may be? Why would one think that imagination does not significantly play such a role? Both Wollheim and Savile think that imagination should be constrained by the representation in the work or the author’s intentions or both, if it is to have some meaningful role in understanding the work. The question, however, is how is the imagination to be constrained, in what way and to what degree? Why can’t a free-ranging (or free-to-range) imagination, one inspired by but not immediately connected to the work, other than causally, give added value to an interpretation of the work, creatively enhancing an understanding of the work though active spectatorship? It might take the meaning of a work and the significance further than the artist themselves intended or understood. In short, and related to both of the above points about imagination, I see no reason for supposing that a spectator or critic might not, and often does, better understand a work of art than an artist. Artists may after all repress centrally meaningful aspects of their work to themselves – albeit not to others. And, according to Freud, this may be a necessary condition of the work of art being produced at all. Artistic production is sometimes, and usually in part, a way of coping with such repression. A spectator may have a better understanding than the artist of the mythology incorporated into a work or access to historical material that is relevant to the interpretation of a work. Additionally, a spectator may have a better understanding of those aspects of an artist’s intentions, beliefs and motivations than the artist themselves – even if the grounds for his understanding are precarious. Given our orectic natures – in which desire, wish-fufillment and phantasy play a prominent role, in the generation of beliefs, for example – that which is opaque to an artist may be clear and transparent to an observer. Why shouldn’t one be better able to understand and interpret Anna Karenina than Tolstoy himself? In a similar vein that again underscores significant differences between his own account and Wollheim’s, Levinson (2001: 29–30)

250 Michael Levine asks ‘Is the artist’s fulfilled intention to depict such and such an apt criterion of what it is correct to see in a picture, and so of what it depicts?’ Wollheim (Savile and Wollheim 1986: 53) says: ‘We can be required to come to understanding of a picture along some such route [that is, aided by the imagination] only if following this route concurs with the fulfilled intentions of the artist’. But Wollheim’s psychoanalytic credentials should be kept in mind here – and also his examples (1987b). We can agree that an artist’s fulfilled intentions may be (may at times be) paramount to an adequate interpretation of the meaning of a work. But some of these intentions may be unconscious or semi-conscious. Some may even be sincerely denied. Wollheim is of course well aware of this and the fact that its plausibility rests on the validity of psychoanalytic commitments. Works often reveal intentions relevant to meaning and interpretation that artists do not, indeed could not, consciously intend. Sometimes – often, in fact – unintentional meaning, in the sense of unconsciously intended meanings, may be essential to interpreting a work of art adequately. Works may reveal aspects the artist is neither consciously aware of and would in any case refuse to intentionally incorporate and reveal in the work if they were aware in any consciously and controllable sort of way just what it is they were saying. Why, too, suppose an adequate interpretation must ‘concur with the fulfilled intentions of the artist’? Suppose the artist isn’t very good. Can’t one hypothesise what their intentions might have been? And, again, why aren’t those things that an artist may reveal unintentionally – could never reveal intentionally – the crucial things needed to interpret and understand a painting? Wollheim’s (1987b) psychoanalytic interpretations of Ingres and Picasso, for example, do not rest on an understanding of their fulfilled intentions, unless those intentions are taken to be something the artists themselves are not consciously aware of and might well disavow when confronted with such an account. Wollheim must of course, agree with this. Wollheim makes some questionable phenomenological claims in relation to interpretation as well. He says, for example, Now the endowment of the picture with an internal spectator [essentially a projected perspective on the picture] offers distinctive access to the content of the picture in that the external spectator can – and must if he is to reach full pictorial understanding of the picture – imagine this figure from the inside. He must identify with him. He must, and will if he follows the intimations of the picture, centrally imagine this figure interacting with the represented scene as his repertoire allows or constrains him to, and then the outcome will be that the spectator of the picture will find himself in the same residual state as the spectator in the picture might be expected to be on doing whatever he is imagined to have done. (Savile and Wollheim 1986: 54) While something like this may occur on rare occasions, Wollheim assumes it as a norm with regard to certain pictures. Yet the claim seems false even with

Wollheim’s ekphrastic aesthetics


regard to those pictures. Why would he think that ‘full pictorial understanding’ requires such imagination and the projection of, and experiential identification with, a spectator internal to the picture – whether that spectator really is depicted or just imagined? Does he think such imagining occurs often (generally?), and why does he think, if he does, that such imagining can generally be successful? He says (Savile and Wollheim 1986: 54): ‘I believe this account to hold true only for certain paintings, and it is a hard matter to see which’. He continues in a note: ‘Cases of which I am certain are most of the single-figure paintings of Manet, the nature-pictures of Caspar David Friedrich, the group portraits of Hals and probably Rembrandt, and, amongst non-figurative paintings, the all over paintings of Jackson Pollock’. I don’t see this as obvious (or, as Wollheim says, ‘certain’) with any of these examples, least of all Pollock. What constitutes ‘full-pictorial understanding’ according to Wollheim and can it ever be achieved? (Surely not.) Wollheim continues: [W]hat we must consider is whether it could ever be legitimate that we should, in concurrence with the artist’s intentions, cease looking at the picture, engage in some imaginative project that involves acentrally imagining [imagining that is done from no-one’s point of view], and then return to a perception of the picture into which is then fed the fruits of this project which we have taken time off to engage in. I think that the answer to this question is, No, but it depends upon yet deeper considerations, or on being able to show that acentrally imagining a represented scene can only subvert, it cannot supplement, the perception of the representation. Two considerations here need to be exploited. The first is that acentrally imagining a represented scene is, unlike centrally imagining it, likely to be purely visual. It will have no affective aspect. This is because, since the imagining is done from no-one’s point of view, the imagined visual experiences are disembodied. And this is novel, because in the case which we have so far been considering where imaginative access occurs through centrally imagining some figure, the contribution that such imagining makes to the understanding of the picture comes almost exclusively from what this figure is imagined to feel-in response, of course, to what he is imagined to see. But this is not all there is to it. It is not only that the fruits of acentrally imagining are purely visual: it is unclear how they are visually relevant to the understanding of the picture. (Savile and Wollheim 1986: 60) Is acentral imagining, understood as imagining from some point of view, but no point of view in particular, even possible?11 (Think of Locke’s critique of the claim to imagine a dog in general but no dog in particular.) Suppose it is possible. Why must the ‘the fruits of acentrally imagining’ be ‘purely visual’? And whether or not acentral imagining is purely visual, why can’t such imagining be ‘visually relevant to the understanding of the picture’, or relevant in

252 Michael Levine other ways to understanding the picture? What is frequently needed to discern what is visually relevant to understanding a painting, let alone to understanding a painting in ways that have little to do with the visual, is just the kind of wide, across-the-board imaginative engagement that Wollheim claims, for reasons that remain unclear, inhibits understanding and appreciation of the work of art. Wollheim’s constraints on the role that a spectator’s imagination can or should play in understanding representation in painting ‘in concurrence with the artist’s intentions’ appear to be grounded in phenomenological inaccuracies and are stipulative. Why, at least sometimes, a spectator cannot take a walk around the block, go home, imaginatively engage with a painting and then, whether they see the painting again or merely picture it mentally, better understand the representation and the artists intentions as a result of their imaginative engagement eludes me (so much so, in fact, that I may be mistaken in attributing such a constraint to Wollheim). It seems to me that having the painting in front of one may often prevent the kind of imaginative engagement with a painting and the artist’s intentions needed to better see and understand the painting. Not only would the constraints Wollheim claims pertain to a spectator’s engagement with representation prevent understanding, they would also undermine one’s affective reaction. The enjoyment (or otherwise) of a work of art, and aspects of its value, depends upon just the kind of free imaginative engagement with representation and one’s perceptions, occurrent and remembered, with respect to figures both centrally and acentrally imagined, including those that aren’t represented (in the picture) at all, that Wollheim, for reasons unclear to me, claims can only hinder our perception, genuine engagement and understanding of the work of art. It is not just the phenomenology that I think Wollheim (generally) has wrong. What is interesting is, given Wollheim’s understanding of psychoanalysis and its significance to philosophy, how inconsonant and out of touch his account is with what psychoanalysis tells us about understanding a work of art in relation to the artist and the artist’s intentions. Here, to a degree, and contrary to both the details and wider gist of Wollheim’s contentions, imaginative speculations on the artist’s motivations, intentions, desires, neuroses, etc. are essential for an understanding and appreciation of the work of art. Of course imagination should, ideally, in the service of genuine engagement with the work of art as such, be constrained in certain ways by one’s perceptions. But the fact that imagination can lead us astray, does lead us astray and can result in misinterpretation, self-deception and wish-fulfilling phantasy on the part of the spectator does nothing to show that it is not essential to a more or less correct perception, appreciation and understanding of a painting. Wollheim (Savile and Wollheim 1986: 60) says ‘visual experiences upon which imagination from the outside will concentrate will surely be experiences that are simply discrepant with the experience that the spectator of the picture has from looking at the picture’. How does he know this? What does ‘discrepant’ even mean here? In any narrow sense of ‘discrepant’ (contradictory, antithetical, conflictual) it is hard to see how

Wollheim’s ekphrastic aesthetics 253 ‘visual experiences upon which imagination from the outside’ may be brought to bear need be discrepant. Wollheim (Savile and Wollheim 1986: 60) says ‘They will be in a different perspective, and they will not be experiences of seeing-in: they will lack twofoldness’. In any broader sense of ‘discrepant’ there seems even less reason to suppose any necessary discordance. Wollheim continues: ‘In so far as such experiences can be thought to have a contribution to make to the understanding of a painting, does this not simply constitute a reflection upon the adequacy of painting?’ But why would it do this? Why would it call into question the adequacy of painting as such or of a painting in particular? Why not suppose that such imaginative experiences, causally generated in art by the visual perceptions of the painting, help us better understand, affectively experience and, at times, better perceive the painting? Perhaps the phenomenological analysis itself reveals something of interest and importance – but what? Is it meant to teach us how to look at, experience, better understand or value art? One thing that is important to Wollheim is that representation in art should be explicable in terms of visual experience. So long as visual experience is part of the experience of the spectator, as it surely is, then the fact he has overgeneralised the experience of seeing-in and its significance (if he has) may be relatively inconsequential to his more general theory of art in terms of psychoanalysis (see ‘Wollheim’s aesthetics and psychoanalysis’ below). With regard to the imagination’s effect on perception, Wollheim (Savile and Wollheim 1986: 48) asks ‘even if the perception of representations is not necessarily imaginative, does it necessarily become imaginative when concepts and beliefs in excess of a certain limit are brought to bear upon it?’ But shouldn’t he be asking, or also asking, something similar about emotion? What effect does emotion have on perception of representation? Let’s turn to his theory of emotion as it relates to aesthetics.

Wollheim: aesthetics and emotion Wollheim succinctly describes the function of emotion – what emotion does, as well as, very broadly here, how it does it. The role of emotion is to provide the creature – or, as we might now get used to saying, the person – with an orientation, or an attitude to the world. If belief maps the world, and desire targets it, emotion tints or colours it: it enlivens it or darkens it, as the case may be. (1999: 15) If this is right, and if (1) understanding an artist’s intentions is necessary to understanding an artwork, and (2) understanding an artist’s orientation or attitude to the world is necessary for understanding their intentions, then a secure and significant connection between aesthetics and emotion has been established. It is furthermore a connection that even on the surface far

254 Michael Levine outweighs whatever the importance of the innate capacity of ‘seeing-in’ – a capacity surely not exercised in all cases of perceiving – to Wollheim’s (or indeed most other) aesthetics. Wollheim’s claim that emotion provides a person with an orientation or attitude to the world might be seen as an illustration of Wittgenstein’s (1922: 6.43) aphorism (and vice versa) that ‘the world of the happy man is a different one from that of the unhappy man’. The connections and influences between Wollheim’s theory of emotion and his aesthetics can be approached from either side and go both ways. It is clear from reading Wollheim’s Painting as an Art, for example, that his theory of emotion was substantially influenced by his aesthetics and analyses of paintings. This is not surprising given the role that artists’ intentions plays in deciphering the meaning of works of art, according to Wollheim. Though the connections go both ways, the direction the influence takes is more pronounced and, I think, arises from his aesthetics and efforts to interpret paintings with his later theory of emotion. It is psychoanalysis, however, that underlies both his account of emotion and his aesthetics – including his account of representation, interpretation, the significance of the artists’ intentions, imagination and other areas. Wish-fulfilment and phantasy, for example, play an important role in Wollheim’s account of emotion, and he says (1987b: 265), ‘Of Ingres’s own family circumstances we know enough to recognize how the subject-matter of these various compositions, to which he returned so tenaciously, could easily have summarized some of his earliest wishes’. There is no direct connection here between Wollheim’s interpretation of Ingres’s paintings and an account of emotion. But how he viewed the role of emotion in these detailed and heavily psychologised interpretations was bound to be, and was, taken up in his theory of emotion. Compare his account of Ingres with his discussion of the moral emotions, shame and guilt, in the third and final lecture of On the Emotions (based on lectures given in 1991). Wollheim (1987b: 271), again referring to Ingres, says, for example, ‘Invariably under the shadow of the wish, massive condensation, massive displacement, massive associative thinking, obscure, indeed destabilize, the object of the desire’. A central claim in Wollheim’s account of emotion is that desire gives rise to much, though perhaps not all, emotion. As Lovibond (2002: 153) says, ‘Wollheim sees emotion as flowing from an attitude that develops in us toward something we identify – rightly or wrongly, actually or prospectively – as responsible for the satisfaction or frustration of one of our desires’. In his review of Wollheim’s book, Prinz (2002), like other reviewers, seems to make much of the fact that Wollheim’s account in terms of desire cannot account for all (each and every) emotion. It would be amazing, given the various ways in which emotion terms are used, if any reactively circumscribed account of what an emotion is could be universal (see, for example, Rorty 1986:121). She says, ‘There is a set of psychological attitudes – love, joy, perhaps some sorts of desire – that are individuated by the character of the subject, the character of the object, and the relation between them’.

Wollheim’s ekphrastic aesthetics 255 These psychological attitudes may (and typically do) include emotions, but they are different from emotion. In Rorty’s account, love is not an emotion but a psychological attitude. But this is unduly prescriptive. Sometimes the term ‘love’ is used to describe an emotion, occurent or dispositional as the case may be, and there is little point in suggesting that such usage is incorrect. That Rorty’s account of love as a psychological attitude is insightful and describes something important about how we may at times think of ‘love’ does nothing to change this. The fact is, emotion terms are used in various contexts in different ways. In any case, what is important to Wollheim is less whether one calls love an emotion, but whether or not it provides one with an orientation or attitude to the world. But it is odd why Prinz and other reviewers see the exceptions as so problematic or as undermining anything other than the universality of Wollheim’s claim relating emotion to desire. If what Wollheim says about emotion is significant and correct with regard to a great deal of emotion, then so what if his theory is not universally true? Wollheim goes on to say tempting though it is to deepen the meaning of these paintings by arguing from their representational meaning to desires and wishes that they express – along the lines made familiar to us by the interpretation of dreams, systematically replacing manifest with latent content - I think that we should resist the temptation to do so: at any rate, in any except the broadest terms. By and large we lack what is essential to such an interpretation – the artist’s associations to what the pictures represent. (1987b: 266) Despite the disclaimer, Wollheim immediately employs psychoanalytic theory to hypothesise connections between paintings’ representational meanings and the various desires, defences, wishes and phantasies of the artist – all aspects of the interpretation of dreams. He says (1987b: 276), for example, ‘Identification on the artist’s part with an overestimated figure is an invariable feature underlying the instrumental conception of painting’. These are, after all, things that have to be taken into account to talk about artists’ intentions, psychoanalytically understood, which for Wollheim are essential to understanding an artwork and the nature of representation. There is no shortage of examples. Wollheim says: Painting credited with an instrumentality which it does not, which it could not, have, presupposes . . . as in the case of the wish, an attitude that the person who resorts to it has towards himself . . . it presupposes an opinion which overestimates . . . the person’s powers, and without such an opinion the artist would not think of his painting as having the efficacy he attributes to it. In this connection Freud borrowed from a patient the term ‘omnipotent’ . . . meaning by this, not, of course,

256 Michael Levine someone who actually was all-powerful, but someone who experienced or conceived of himself and his thoughts as all-powerful . . . In origin omnipotence of thought is an instinctual feature of the primitive or infantile mode of experience. (1987b : 275) It is the greater part of what makes his accounts of various paintings interesting and insightful. And it is this, along with his at times magisterial knowledge (historical, biographical, etc.) of the paintings, rather than his overgeneralised account of seeing-in, that constitutes Wollheim’s enduring contribution to and (one can just tell) love of aesthetics. He relishes interpreting paintings in terms of the psychology of the artist. This is so despite the fact that ‘seeing-in’ has unfortunately garnered most of the philosophical attention. Wollheim’s account of the ‘internal spectator’ in connection with a spectator’s imagination and interpretation of painting has been discussed above. But when he takes the issue up again fifteen years later, two years after the publication of On the Emotions, the focus, though present in Painting as an Art (1987a), is on a far more Freudian role of emotional resonance between the imagined internal spectator and the external spectator in terms of its significance for interpretation and perception. The focus is now on the role of ‘a psychological repertoire: a repertoire of beliefs, desires, attitudes, responses’ seemingly required to adequately experience the painting. Wollheim says: I too find a place for imagination in my account of representational meaning, but it is a place that is ancillary to seeing-in, and is relevant only to certain paintings . . . These are paintings in which the suitable spectator is offered a distinctive form of access through the presence in the represented space – though not in that part of it which is represented – of a figure, whom l call the Spectator in the Picture. The Spectator in the Picture has, amongst other things, a psychological repertoire: a repertoire of beliefs, desires, attitudes, responses. What then happens is that the suitable spectator, the suitable external spectator we might say, starts to identify with the internal spectator: that is, to imagine him centrally, or from the inside, interacting with the represented scene as the repertoire assigned to him allows or constrains him to. The net result will be that the external spectator will find himself in a residual state analogous to that of the internal spectator, and this state will in turn influence what he sees in the picture when he reverts from imagination to perception. (2001: 25) The artist’s intention remains central to Wollheim’s interpretation, but there has been a development from Painting as an Art (1987a) in how one is able to grasp such intention. Wollheim (1987a: 8) says, ‘At least in the context of art, intention must be taken to include desires, beliefs, emotions, commitments, wishes, that the agent has and that, additionally, have a causal influence on the way he acts’.

Wollheim’s ekphrastic aesthetics


In Wollheim’s ‘On Pictorial Representation’ (2001), imaginative identification is seen as necessary for interpretation, perception and adequate experience with respect to ‘certain paintings’. Why, however, he restricts the scope of his claim in this way (I doubt Freud would) isn’t clear, even if such identification is most clearly illustrated with Wollheim’s examples. Of course, prima facie, such imaginative identification seems difficult with non-representational paintings. Wollheim says: If all the suitable spectator can do is to pick up on the artist’s intention, and interpret the work accordingly, and there is no register of this in his experience of the picture, the conditions of representation have not been satisfied. Representation is perceptual. (2001: 27) What kind of register in experience does Wollheim have in mind if not psychological – experience affected by emotion? ‘Representation is perceptual’, according to Wollheim, but such perception is, it would seem, suffused with an orientation to the world that only emotion can provide. This is different from the mapping of the world that belief and presumably knowledge provides. If so, then we can speak of emotion as a condition of representation and interpretation. Furthermore, emotion in this view is not just any condition or even a necessary condition or such representation. It is central. Wollheim says: the corresponding emotion, once invoked, should not stand apart from the perception through which it is invoked. It should not be a mere association to what is perceived. The emotion should flood in on the perception. In expressive perception it is not enough that what is perceived invokes the corresponding emotion: the emotion must effect how we perceive what we perceive. Expressed emotion and perception fuse. (1987a: 82) If one is looking for a connection between Wollheim’s aesthetics and his later theory of emotion, one can see a nascent version of the role of emotion here. Wollheim’s description of emotions as providing a person with an orientation or attitude to the world is consonant with or reflects the claim that, as psychoanalytically conceived, individuated and originating in infantile phantasy, emotions or emotion types may not correspond in a direct way (cannot be isomorphically mapped) to emotion terms as actually used.12 Though the point is somewhat different from Gardner’s (1992: 42–4), who discusses this issue, the idea that our emotions do not neatly line up with emotion terms is illustrated in this passage from a novel by Jeffrey Eugenides: Emotions, in my experience, aren’t covered by single words. I don’t believe in ‘sadness,’ ‘joy,’ or ‘regret.’ Maybe the best proof that the language is patriarchal is that it oversimplifies feeling. I’d like to have at my

258 Michael Levine disposal complicated hybrid emotions, Germanic train-car constructions like, say, ‘the happiness that attends disaster.’ Or: ‘the disappointment of sleeping with one’s fantasy.’ I’d like to show how ‘intimations of mortality brought on by aging family members’ connects with ‘the hatred of mirrors that begins in middle age.’ I’d like to have a word for ‘the sadness inspired by failing restaurants’ as well as for ‘the excitement of getting a room with a minibar.’ I’ve never had the right words to describe my life, and now that I’ve entered my story, I need them more than ever. (2002: 130) In Gardner’s (1992: 44) account, the mismatch is accounted for by the fact that ‘[T]he consciously recognized and self-ascribed state of emotion is a historical descendant of an earlier kind of state, which is one of phantasy’.13 Does this reflection on emotions and emotion terms – that they are different and at times mismatched – lend some support to Wollheim’s view of the emotions? If so, then it is also suggestive of a reason why we do or should hold aesthetics to be such a valuable and central part of our lives. It is a reflection, a true reflection, of experience that even language and conceptualisation may or even must miss. Aesthetic experience is a unique way of experiencing and interacting with the world, and intimates aspects of ourselves beyond the reach of cognition.

Wollheim’s aesthetics and psychoanalysis I have mentioned the psychoanalytic underpinnings of Wollheim’s aesthetics and account of interpreting art. But what are these? They include Freudian psychoanalysis bolstered in particular by Melanie Klein’s object-relations theory. Wollheim’s aesthetics are closely related to Freud’s views on art, though there are of course differences and disagreements as well (Wollheim 1970). Like Freud, Wollheim places great emphasis on the artist’s intentions in interpreting works of art, and he too holds the view that art is orectic, driven by desire and wish-fulfilment and functioning so as to satisfy certain psychological needs of artists and audiences. This is what makes the artist’s intentions – psychoanalytically understood – so crucial for interpretation. Wollheim, like Freud, sees art in terms of psychoanalysis, as deeply integral to a life – rather than shallowly in terms of the isolated concepts representation and expression that analytic philosophers tend to employ.14 It is evident that in Wollheim’s view any account of interpretation or representation in aesthetics – indeed any account of the nature, function and value of art that is not psychoanalytically informed – must be radically incomplete. Here then, again, is the explanation why analytic aestheticians have focused on seeing-in rather than emotion in Wollheim’s aesthetics. Phenomenological introspection can address seeing-in, but psychoanalysis is needed to discuss emotion in Wollheim’s aesthetics and the deeper aspects of his aesthetics altogether. Analytic philosophers largely ignore and eschew psychoanalysis.

Wollheim’s ekphrastic aesthetics 259 Decades earlier Sterba described the source of artistic impulse as psychoanalytically conceived as follows: At the basis of artistic creation we find the same instinctual forces which are effective as components of intrapsychic conflicts . . . The work of art is therefore the product of psychic forces that are in opposition to each other, such as desire and inner prohibition. It represents reconciliation between these conflicting forces and has therefore the character of a compromise, as have also those psychopathological formations – errors, dreams and neurotic symptoms – which are well known. The fundamental dynamic force at the root of a work of art is an unfulfilled wish of the artist; just as in dreams and fantasies, the work of art represents this wish as fulfilled . . . The artist is an introvert. Since he is unable to satisfy his overpowering instinctual needs in the world of reality, he is obliged to turn away from the real world to the realm of phantasy thus taking the way which leads to neurosis. But it is here that the creative process sets in, enabling him through discharge of instinctual energy and the effect of the work of art on the outside world to save himself from neurosis and to regain contact with reality. He is able through the creation of a work of art to obtain sufficient gratification of his intense childhood wishes which he represents as fulfilled in his creation . . . His particular method of representation, possible to the artist on account of his talent, enables him in a certain measure to find a way back from fantasy to reality, obtaining in this roundabout way a means of gratifying actual wishes and of achieving success for which, in a direct way, his forces would never have been adequate. (1940: 257–60) It is worth considering or comparing some of Wollheim’s interpretations of paintings (1987b) with Sterba’s account. Notice how necessary it is to understand the artist’s intentions as well as the artistic process if one is to interpret their artwork. On the point about art being used as a way of returning to reality – or satisfying wishes in reality – Wollheim says that: In a number of celebrated places Freud equated art with recovery or reparation or the path back to reality: in the case of Leonardo, as we have seen, he attempted to document this moment. But nowhere did he indicate the mechanism by which this came about. (1970: 224) Just what sort of mechanism Wollheim is looking for, or what he means by it, is unclear, and Wollheim never tells us. Is he referring to the ‘creative process?’ Why, for instance, and in what way, is Sterba’s psychoanalytic account of the process (or mechanism) insufficient? Perhaps it has to do with the fact that while some people produce only phantasies and symptoms, artists produce art.

260 Michael Levine And that no-one, certainly not Freud or Sterba, has really explained why some can and others can’t. It is unclear to me, however, that this can be explained except in rather general terms. Towards the end of his essay Wollheim appears to suggest that infantile wish-fulfilment does not play the central role in understanding and interpreting art or the artist that most psychoanalysts claim it does, or at least that the part it does play is not as straightforward as in Sterba’s account. Wollheim asks: Is there, according to Freud, anything parallel in the work of art to the purpose that finds, seeks, expression in the work of art? An understanding of psychoanalysis should warn us that this question is unanswerable. For outside the comparative inflexibility of the neurosis, there is no single unchanging form that our characters or temperaments assume. There are constant vicissitudes of feeling and impulse, constant formings and reformings of phantasy, over which it is certain very general tendencies pattern themselves: but with a flexibility in which, Freud suggests, the artist is peculiarly adept. The artist expresses himself in his work – how could he not? But what he expresses has not the simplicity of a wish or impulse. (1970: 223–4) What sort of ‘parallel’ thing in a work of art – that is, parallel to the purpose that finds expression in a work of art – is Wollheim referring to? Isn’t it Freud’s view that the work of art results from, expresses and in a sense embodies such purpose (for example, wish-fulfilment)? Wollheim claims that because ‘there is no single unchanging form that our characters or temperaments assume’ what the artist expresses in a work of art, and what the work satisfies, may frequently not involve infantile wish-fulfilment. He may mean only – and if so, there is no disagreement with Freud here – that what the artist expresses cannot ‘simply’ be reduced to such wishes and impulses. Along with the idea that wish-fulfilment may at times play virtually no role in an art work, this is worth keeping in mind. But not only may variations in the modes of expression of infantile wish-fulfilment (variations in forms of art themselves) take account of such ‘vicissitudes of feeling and impulse’: accommodating such changes is in line with the view that works of art express such wishes. Wollheim may also be suggesting that there is more to character than a set of desires and that these other things also get expressed, and perhaps may be expressed, without the desires. What is ‘parallel in the work of art’ to the purpose that finds expression in the work of art is the (artistic) embodiment of what may be a quite complex wish-fulfilling expression (or the expression of a wish-fulfilling phantasy) itself – one (or more) such expressions tinged with changeable feelings and impulses. This does not make them anything other than wish-fulfilling phantasies. It does, however, suggest that understanding a work of art may also, or even sometimes primarily, require understanding and otherwise

Wollheim’s ekphrastic aesthetics


accounting for these other feelings and impulses involved – biographically perhaps. (A parallel might be the dream in which there is always the fulfilment of some infantile wishes but compounded with the day’s residues and other circumstantial material.) Far from denying the significance of context and individual circumstances to understanding a work of art, Wollheim, like Freud, emphasises it. Freud not only accounts for the function of art (in terms of the satisfaction of unfulfilled wishes), but also how and why a work of art is able to elicit the audience response it does. Sterba says: The disclosure of the fact that unfulfilled wishes, originating in the unconscious, are satisfied in the work of art, and that it is possible through analysis of the work of art to discover these; that, furthermore, the enjoyment of the work of art [on the part of spectators] derives principally from the satisfaction of these infantile wishes, is Freud’s most important contribution to the psychology of the work of art. (1940: 262) Wollheim must surely have agreed. It is, after all, central to his aesthetics as well. According to Freud: [a]n artist is originally a man who turns away from reality because he cannot come to terms with the renunciation of instinctual satisfaction . . . [The artist] finds his way back to reality, however, from this world of phantasy, by making use of special gifts to mould his phantasies into truths of a new kind, which are valued by men as precious reflections of reality. (quoted in Sayers, 2007: 23) In Art and Its Objects (1980: section 50), Wollheim makes the very same point. Relatedly, Sterba (1940: 266) notes that, ‘[t]he mutual permeation of the fields of domination of the pleasure and reality principles led Freud to call art a kind of reconciliation between the two principles in mental functioning’.15

Music cannot convey emotion? I don’t think so. Wollheim’s focus in aesthetics is on painting. But his theory of emotion in relation to aesthetics, combined with a psychoanalysis which underpins both his aesthetics and his theory of emotion, can be applied more widely – to music, for instance. I will give an example – a particularly easy target.16 Before doing so, it is worth noting a connection or analogy with seeing-in. In Wollheim’s view of emotion, emotion colours perception; presumably, then, this may influence what is seen-in the object. But if emotion can do this, doesn’t it also influence what is ‘heard-in’ the object?17 Without pursuing the

262 Michael Levine point, it could perhaps be argued that ‘hearing-in’ is an even stronger (clearer) aesthetic example, a more prevalent and so more generalisable phenomenon than seeing-in. Kivy (2009: 225) claims that ‘absolute music cannot arouse the gardenvariety emotions’. He bases this assertion on his adopted theory of emotion. He says, according to the cognitive theory of emotion, the ‘garden-variety emotions are typically aroused by the forming of beliefs appropriate to the emotions, which take intentional objects, and then eventuate in some appropriate mode of behavior . . . . The all too obvious point of the story is that absolute music simply does not have the resources to do this sort of thing . . . How could it? What beliefs could music elicit in us to make us afraid, or angry, or sad? What would the intentional objects of these emotions be? And what would be the appropriate actions? (2009: 225–6) There are two plausible responses to Kivy. The first is to accept the cognitive theory and look for the beliefs elicited by music as well as the intentional objects of the emotions aroused and appropriate actions encouraged. That neither the beliefs elicited nor the intentional objects are always obvious in the case of music should come as no surprise. Even in the case of ordinary emotion the beliefs and objects are not always obvious and are frequently misidentified. Freud, for example, thought that while every case of emotion (including those that seem apparently objectless) does have an object, and one appropriate to that emotion, one is not always consciously aware of what it is.18 In any case, if music does arouse emotion – and it certainly seems to – then one can only maintain the cognitive theory of emotion if one gives some account of what the requisite beliefs (or other cognitive states) involved are, as well as what the intentional objects of the emotions aroused are. Perhaps the context, tradition and history of a piece of music can supply, on occasion, just such states and objects. Kivy (2009: 226–7) critiques Jenefer Robinson’s (2005: 400–12) affective appraisal theory. He says that in Robinson’s account what music does if it does it, is to stimulate the listener to produce a garden-variety emotion in her self, some time after the event. Surely, though, the claim we are investigating, that music possesses moral force through emotive arousal, is the claim that musical masterpieces possess the power of themselves, in their own character as art works, to directly arouse gardenvariety emotions and, thereby, alter moral behavior (for the better). (2009: 227) Leaving aside whether Robinson’s account is correct, Kivy’s view about the claim we are investigating is strange. The claim is that music arouses

Wollheim’s ekphrastic aesthetics


emotion – causally. It is not that music does so, or must do so, immediately and directly with nothing else occurring in the casual chain. Who knows what else is going on in the casual chain commencing with hearing a piece of music? Why would the fact, if it is a fact, that the music arouses certain associations which then causally give rise to emotions (emotions which in turn can lead to moral reflection and motivation on occasion) make music any the less powerful as a cause of emotion – or not a cause at all? The second response is to shed the cognitive theory of emotion while not necessarily denying that its central claims hold in some cases. Consider again these few lines from Wollheim: The role of emotion is to provide the creature – or, as we might now get used to saying, the person – with an orientation, or an attitude to the world. If belief maps the world, and desire targets it, emotion tints or colours it: it enlivens it or darkens it, as the case may be. That emotion rides into our lives on the back of desire is a crucial fact about emotion, as well as a crucial fact about us. The colour with which emotion tints the world is something to be understood only through the origin of emotion in desire. (1999: 15–16) If emotion provides us with ‘an attitude to the world’, then surely music can be said to elicit such attitudes. These attitudes in turn colour perception of the music, and they may be ethically charged. They may affect our characters, how we live and even provide, if not direct moral insight (propositions), then the conditions that make such insight and receptivity to insight possible. Kivy critiques the alleged character-building force of absolute music and asks whether music has ‘the power to build moral character; to make someone, as it were, a better human being’. But if music, by means of its ability to evoke emotion, possesses both epistemic and behavioural moral force, then it also possesses, for some people, some of the time, the power to build moral character. If music enlarges our capacities of emotional empathy (not for everyone, not for all music and not on all occasions), then it straightforwardly has a role to play in building moral character. It can affect the way we see and think about things. Here, then, is a connection not only between Wollheim’s account of the emotions and aesthetics but also between aesthetics and ethics.* *My thanks to Gary Kemp and Tamas Pataki for very helpful comments.

Notes 1 On the Emotions (1999) is based on Wollheim’s 1991 Ernst Cassirer lectures at Yale. 2 See, for example, Wollheim (1987a).

264 Michael Levine 3 The scarce attention paid to connections between his aesthetics and theory of emotion may be part of the broader problem Wollheim himself boldly identified. He says that Virtually all those who are not either ignorant of Freud or totally sceptical of his findings believe that he altered, radically altered, our conception of the mind. He effected a change in what we think we are like, and it was a big change. Astonishingly enough, it is philosophers who have been of all people the slowest to recognise this fact. They have been slowest to recognise that this fact has anything to do with them (1993: 91) But even if philosophers generally don’t recognise, or flatly reject, the philosophical significance of psychoanalytic theory, whether generally or for aesthetics in particular, why didn’t Wollheim make the connection as clearly in his account of emotion (1999) as he did in Painting as an Art (1987a)? 4 Levinson (2001:33): ‘It is hard to overestimate the keen interest that viewers of paintings naturally take to bringing simultaneously into relation, or alternating systematically between the recognitional and configurational, or the picture and the picturing, in different styles of depiction’. 5 Essays in this book may convince me otherwise. But additional treatments on Wollheim’s account of seeing-in remind me of the joke about asparagus: ‘My mother cooks asparagus to perfection, and then an hour longer’. Seeing-in has been done-to-death. Still the essays keep on coming (Newall 2015). 6 Wollheim says: The experience of seeing-in has then a recognitional and a configurational aspect. But in one important respect these aspects of one single experience behave as though they were what it is crucial to remember they are not: that is, separate experiences. For both aspects are conditioned by the cognitive stock that the spectator holds and brings to bear upon the representation that he confronts. Both how he sees the marked surface and how he sees what he sees in the marked surface are determined partly, of course, by how the surface is, but partly they are determined by the concepts and beliefs that the spectator has and mobilizes. (Savile and Wollheim 1986: 47–8) 7 Levinson prefers to put the matter this way: much remains of what Wollheim has urged in what I am willing to affirm on this vexed topic: Pictorial representation involves the intentional marking of a flat surface so as to elicit a distinctive sort of visual experience in appropriate spectators. An experience we may continue to call seeing-in as long as we understand that this sometimes amounts only to what may be more transparently labelled seeing-from, where such experience is indeed elicited from those spectators in virtue of their attending to the surface as marked. (2001: 36) 8 Compare Levinson’s account of the phenomenology of pictorial experience, and critique of seeing-in theory with that of Lopes (1996), whom Levinson (2001: 37 (note 7)) notes has influenced him. 9 And more: Levinson (2001: 29) says ‘imagining seeing X in viewing Y implies, as a default, imagining you are face-to-face with X . . . ’ It may in some cases, but why is this the default? Indeed, I would think that imagining seeing X in viewing Y generally implies not imagining one is face-to face with X. The first imagining (imagining seeing X in viewing Y) rules the second imagining (imagining you are face-to-face with X) out.

Wollheim’s ekphrastic aesthetics


10 Tamas Pataki suggests (in correspondence) that I may be misreading Wollheim here. He says ‘Wollheim says that what one sees is “partly determined by . . . ”; which doesn’t appear to imply that one must be aware of the surface’. But if Pataki is right, this would not be a case of seeing-in. 11 Wollheim may mean something different by acentrally imagining. Pataki (in correspondence) suggests that in acentral imagining one is not imagining from the point of view of a participant in the scene or drama. Instead one would just be the external observer without projecting oneself into a character or internal perspective. Perhaps. But I am unsure as to whether acentral imagining is possible in this account either. 12 See Gardner (1992: 42–4); Levine (2000: 244–5). 13 Pataki says: I’ve never accepted this account of Gardner’s and it’s clear now in the light of recent work in affective neuroscience that it’s wrong. Emotion comes first. So I would not regard the Gardner type view as THE psychoanalytic view, but perhaps as one. (in correspondence) I am unclear as to how neuroscience can show this, but quite willing to regard Gardner’s view as one psychoanalytic view among others. 14 See note 3. For a discussion of Freud’s aesthetics see Levine (2015). 15 Pataki cautions: I would probably restrict the scope of the wish-fulfilment account more than you do. I doubt that all, or perhaps even most of the things, commonly classified as art (lots of decorative stuff, pottery etc., even high art) would be wish-fulfilling in any reasonably straightforward sense. There is the expression of states of mind, it seems to me, that are not associated with wishes. And it’s strange that Wollheim doesn’t appear to use the notion of reparation which is central to Klein’s contribution to the theory of art. (in correspondence) 16 See Cox and Levine (2015) for an account of whether music can make you a better person. 17 My thanks to Gary Kemp for this point. 18 See Sachs (1982).

Bibliography Carroll, Noël (2011). ‘Art Interpretation’, 2010 Richard Wollheim Memorial Lecture. British Journal of Aesthetics 51:2, 117–35. Cox, Damian and Levine, Michael (2015). ‘Music and Ethics: The very mildly interesting view’. Music: Oxford Handbooks Online. Oxford: Oxford University Press. Eugenides, Jeffrey (2002). Middlesex. New York: Farrar, Straus and Giroux. https:// msantologia.files.wordpress.com/2013/11/middlesex-jeffrey-eugenides.pdf Freeman, Damien (2012). Art’s Emotions: Ethics, Expression and Aesthetic Experience. Montreal and Kingston: McGill-Queen’s University Press. Gardner, Sebastian (1992). ‘The Nature and Source of Emotion’. In J. Hopkins and A. Savile (eds), Psychoanalysis, Mind and Art: Perspectives on Richard Wollheim. Oxford: Blackwell, pp. 35–54. Grünbaum, A. (1988). ‘Précis of The Foundations of Psychoanalysis: A philosophical critique’. In P. Clark and C. Wright (eds), Mind, Psychoanalysis and Science. Oxford: Blackwell, pp. 3–32.

266 Michael Levine Hopkins, J. (1988). ‘Epistemology and Depth Psychology: Critical notes on The Foundations of Psychoanalysis’. In: P. Clark and C. Wright (eds), Mind, Psychoanalysis and Science. Oxford: Blackwell, pp. 33–60. Hopkins, J. (1982). ‘Introduction: philosophy and psychoanalysis’. In: R. Wollheim and J. Hopkins (eds), Philosophical Essays on Freud. Cambridge: Cambridge University Press, pp. vii–xlv. Kivy, Peter (2009). Antithetical Arts: On the Ancient Quarrel between Literature and Music. Oxford: Oxford University Press. Levine, Michael: (2015). ‘Freud’s Aesthetics: Artists, art and psychoanalysis’. In Simon Boag (ed.) Psychoanalysis and Philosophy of Mind. London: Karnac, pp. 137–62. Levine, Michael: (2000) ‘Lucky in Love: Love and emotion’. In M. Levine (ed.), The Analytic Freud: Philosophy and Psychoanalysis. London and New York: Routledge, pp. 231–58. Levinson, Jerrold (2001). ‘Wollheim on Pictorial Representation’. In Rob Van Gerwan (ed.), Richard Wollheim on the Art of Painting: Art as Representation and Expression. Cambridge: Cambridge University Press, pp. 28–38. Lopes, Dominic (1996). Understanding Pictures. Oxford: Oxford University Press. Lovibond, Sabina (2002). ‘Review of Richard Wollheim’s On the Emotions.’ The Journal of Philosophy 99: 153–6. Newall, Michael (2015). ‘Is Seeing-in a Transparency Effect?’ British Journal of Aesthetics 55: 131–56. Pataki, Tamas (2014). Wish-Fulfilment in Philosophy in Psychoanalysis: The Tyranny of Desire. London and New York: Routledge. Pataki, T. (1997). ‘Some Aspects of Writing’. Quadrant 41: 46–52. Prinz, Jesse (2002). ‘Review of Richard Wollheim’s On the Emotions’. Ethics 113: 188–90. Robinson, Jenefer (2005). Deeper than Reason: Emotion and Its Role in Literature, Music and Art. Oxford: Oxford University Press. Rorty, Amélie, O. (1986). ‘The Historicity of Psychological Attitudes: Love is not love which alters not when it alteration finds’. In Midwest Studies in Philosophy X: 399–412. Reprinted in Mind in Action: Essays in the Philosophy of Mind. Boston: Beacon Press (1988), pp. 121–34. Sachs, David (1982). ‘On Freud’s Doctrine of Emotions’. In R. Wollheim and J. Hopkins (eds), Philosophical Essays on Freud. Cambridge: Cambridge University Press. Savile, Anthony and Wollheim, Richard (1986). ‘Imagination and Pictorial Understanding’. Proceedings of the Aristotelian Society, supplementary volumes 60: 19–60. Sayers, J. (2007). ‘Freud’. In Freud’s Art: Psychoanalysis Retold. New York: Routledge, pp. 9–28. Sterba, R. (1940). ‘The Problem of Art in Freud’s Writings’. The Psychoanalytic Quarterly 9: 256–68. Wittgenstein, Ludwig (1922). Tractatus Logico-Philosophicus. London: Routledge and Kegan Paul. Wollheim, Richard (2001). ‘On Pictorial Representation’. In Rob van Gerwan (ed.), Richard Wollheim on the Art of Painting: Art as Representation and Expression. Cambridge: Cambridge University Press, pp. 13–27. Wollheim, Richard (1999). On the Emotions. The Ernst Cassirer Lectures, 1991. New Haven: Yale University Press.

Wollheim’s ekphrastic aesthetics 267 Wollheim, Richard (1993). ‘Desire, Belief, and Professor Grünbaum’s Freud’. In The Mind and its Depths. Cambridge, MA: Harvard University Press, pp. 91–111. Wollheim, Richard (1987a). Painting as an Art. London: Thames and Hudson. Wollheim, Richard (1987b). ‘Painting, Omnipotence, and the Gaze: Ingres, The Wolf Man, Picasso’. In Painting as an Art. London: Thames and Hudson, pp. 249–304. Wollheim, Richard (1984). The Thread of Life. Cambridge: Cambridge University Press. Wollheim, Richard (1980). Art and Its Objects: With Six Supplementary Essays. 2nd edition. Cambridge: Cambridge University Press. Wollheim, Richard (1973). On Art and the Mind: Essay and Lectures. London: Allen Lane. Wollheim, Richard (1970). ‘Freud and the Understanding of Art.’ British Journal of Aesthetics 10: 211–24. doi:10.1093/bjaesthetics/10.3.211 Young, James O. (2001). ‘Review of Richard Wollheim’s On the Emotions.’ Journal of Aesthetics and Art Criticism 59: 336–7.

12 Visions Wollheim and Walton on the nature of pictures David Hills

1 I evolved a way of looking at paintings which was massively timeconsuming and deeply rewarding. For I came to recognize that it often took the first hour or so in front of a painting for stray associations or motivated misperceptions to settle down, and it was only then, with the same amount of time or more spent looking at it, that the picture could be relied upon to disclose itself as it was . . . I noticed that I became an object of suspicion to passers-by, and so did the picture I was looking at . . . There seemed to me no point in setting painting in a new light, unless I could satisfy myself that it was a true light. (Wollheim 1987: 8–9) Two people taught me most of what I know in aesthetics, Kendall Walton and Richard Wollheim, and I’ve been lucky enough to have both of them as friends. Each has argued that to better understand art’s hold on us—the values we accord it, the satisfactions it affords us—we must better understand our hold on art: the thoughts, experiences, and actions in which art makes its value felt. Together they deserve much of the credit for making aesthetics what it is today: a thriving, empirically informed, self-correcting branch of contemporary philosophy of mind. So it was disappointing and baffling when communication between the two of them broke down over Ken’s make-believe-based account of depiction, an account intended to provide substantiating detail for Richard’s own account based on seeing-in. Those of us who’d followed Ken’s work closely found it hard to make connected sense of Richard’s objections, while Richard, for his part, seemed sure the new light in which Ken set depiction was a false light, sharply at odds with the spirit of his own suggestions.1 In “What Makes Representational Painting Truly Visual?” (Wollheim 2003), he divides his own position into two distinct claims: a broad structural one, to the effect that pictorial representation needs to be understood by reference to the nature and content of the experience to be had in front of a picture; and a

Visions: on the nature of pictures 269 narrow substantive one, to the effect that the appropriate experience involves seeing the subject in the picture. He continues: For most philosophers of art, perhaps because they are more concerned with philosophy than they are with art, my substantive claim, or the account that I have offered of the mode of perception involved in looking at representational paintings, has proved to be the more interesting part of my claim about the visuality of pictorial representation. Furthermore my account has achieved widespread acceptance. But widespread acceptance of the account has not equalled general agreement on the topic . . . Philosophers have reinterpreted [seeing-in] so as to get the theory of representation that they favor. They have smuggled into the appropriate experience modes of perception, hence into aesthetics theories of representation, that I had hoped talking about seeing-in would block. Specifically I want to consider two ways in which seeing-in has become for philosophers of art a Trojan Horse to secure victory for their favored theory of representation. . . . My argument will be that neither adventure succeeds. (Wollheim 2003: 133) That stings. Who wants to be caught smuggling? Who wants to care more about philosophy than about art? Who wants to be a geek? (Beware of geeks bearing gifts!) But set that aside. Wollheim thinks make-believe-based theories of depiction (in the manner of Walton) and experienced resemblance theories (in the manner of Peacocke, Budd, and Hopkins) spell out his narrow substantive claim—that representation is to be understood in terms of seeing-in—in ways deeply at odds with his broader structural claim—that the content of a picture is that of the visual experience we are meant to have in front of it. What does this second claim amount to, and is it true? Can there be such a thing as the way a picture is meant to be experienced? I’ll return to this question when we’re better prepared to address it.

2 A vast renewal of philosophical interest in depiction got under way in the fifties with the work of the art historian E.H. Gombrich. Gombrich followed Helmholtz, Popper, R.L. Gregory and others in holding that appearances, the contents of human visual experience, are products of an unconscious inference to the best explanation of available retinal stimuli on the part of the human visual system. Perceptual inferences can deploy whatever concepts the perceiver happens to possess, so the appearance an object takes on for any particular person depends on the terms in which she is prepared to

270 David Hills think about it. There is no such thing as the appearance a thing possesses when accurately seen from a particular physical viewpoint. A naturalistic image, a depiction, a picture in the strong sense, is a marked surface designed to reproduce or replicate more or less fully, for appropriately equipped and appropriately prepared spectators, an appearance its subject can take on for such spectators. In that respect, it aspires to stand in or substitute for its subject in the visual experience of such spectators. Other images lack such appearance-capturing aspirations. They stand in or substitute for their originals in other connections and for other purposes, sharing with their originals only the specific properties needed for the relevant kind of standing in. (Think of how a hobby horse stands in for a real horse or a doll for a real baby.) Consider a naturalistic image maker, out to capture an appearance some particular object is able to take on. There is no way for her to set aside effects of past encounters with other objects in her efforts to see this one accurately, or past attempts to depict other objects in her efforts to respond appropriately to how she now sees this one. She must therefore exploit what she can’t set aside. She relies on habits, routines, and formulas inherited from earlier image-making practice, naturalistic or otherwise, to produce a rough generic image of thing of the right general sort, which she revises and elaborates in a trial-and-error manner until she achieves an adequate likeness to this particular object as seen by her here and now. Naturalistic image making is a process of schema and correction, making and matching. Small-scale explorations by individuals in the course of making particular pictures contribute to large-scale explorations by communities of picture makers as they invent, refine, and promulgate redeployable appearance-capturing routines, informed by hard-won empirical insights into how human vision works. Foreshortening, tonal modeling, and the various perspective systems are major inventions of this sort; there are countless minor ones. Platonic replicationism to the contrary, one needn’t and can’t make one’s depiction of an object visually indistinguishable from the object itself. It suffices to contrive a stimulus that combines (a) a selection of the cues by means of which we visually detect the object and its properties in real life; with (b) decisive counter-indications that what we are in fact face to face with is not the object at all but a marked flat surface instead. The result is that we recognize the depicted object (thanks to reproduced perceptual cues) in its acknowledged absence (thanks to the counterindications). The history of art . . . may be described as the forging of master keys for opening the mysterious locks of our senses to which only nature herself originally held the key . . . Of course, once the door springs open, once

Visions: on the nature of pictures


the key is shaped, it is easy to repeat the performance. The next person needs no special insight—no more, that is, than is needed to copy his predecessor’s master key. (Gombrich 1960: 359–60) For Gombrich this went along with a definite conception of the phenomenology of picture perception, a kind of mitigated illusionism. As depiction takes hold, we lose sight of the surface markings in favor of the object they depict; as attention returns to the markings, the object disappears. We have two “readings” of what’s before our eyes that can’t be sustained together, alternate over time, and are separated in time by a Wittgensteinian aspect shift. Once depiction takes hold and for as long as its hold lasts, we see the marked surface as the depicted object; it stands in for that object in ongoing perceptual activities of ours; we take the marked surface for the object in a manner that stops just short of confounding the two.

3 Wollheim admired Gombrich’s key-forging metaphor and the associated understanding of pictures as stimuli that mix otherwise reliable cues with decisive counter-indications, stimuli that trigger the recognition of objects in their manifest absence. But he regretted Gombrich’s illusionistic phenomenology. Seeing-as should yield its place to seeing-in, a form of seeing double that predates its controlled deployment in depictions. Referring to an Aaron Siskind photograph, Chicago 10, 1948 (Figure 12.1), he writes: Seeing-in is a distinct kind of perception, and is triggered off by the presence within the field of vision of a differentiated surface . . . When the surface is right, then an experience with a certain phenomenology will occur . . . The distinctive phenomenological feature I call “twofoldness”, because, when seeing-in occurs, two things happen: I am visually aware of the surface I look at, and I discern something standing out in front of, or (in certain cases) receding behind, something else. So, for instance, I follow the famous advice of Leonardo da Vinci to an aspirant painter and look at a stained wall, . . . , and at one and the same time I am aware of wall, and I recognize a naked boy . . . In virtue of this experience I can be said to see the boy in the wall . . . The two things that happen when I look at, for instance, the stained wall are, it must be stressed, two aspects of a single experience that I have, and the two aspects are distinguishable but also inseparable . . . They are neither two separate simultaneous experiences, which I somehow hold in the mind at once, nor two separate alternating experiences, between which I oscillate . . . None of this is to deny that there is an important causal traffic between seeing-in and seeing face-to-face. Children learn to

272 David Hills

Figure 12.1 Aaron Siskind, Chicago 10, 1948 (photograph). Source: © Aaron Siskind Foundation, courtesy of Bruce Silverstein Gallery, New York.

recognize many familiar and unfamiliar objects through first seeing them in the pages of books. The twofoldness of seeing-in does not, of course, preclude one aspect of the complex experience being emphasized at the expense of the other. In seeing a boy in a stained wall I may very well concentrate on the stains, and how they are formed, and the materials and colours they consist of, and how they encrust or obscure the original texture of the wall, and I might in consequence lose all but a shadowy awareness of the boy. Alternatively, I might concentrate on the boy, and on the long ears he seems to be sprouting and the box he is carrying—is it a bomb, or a present for someone, and thus have only the vaguest sense of how the wall is marked. One aspect of this experience comes to the fore, the other recedes . . . In a community where seeing-in is firmly established, some member of the community—let us call him (prematurely) an artist—sets about marking a surface with intention of getting others around him to see some definite thing in it: say, a bison. If the artist’s intention is successful

Visions: on the nature of pictures 273 to the extent that a bison can be seen in the surface as he has marked it, then the community closes ranks in that someone who does indeed see a bison in it is now held to see the surface correctly, and anyone is held to see it incorrectly if he sees, as he might, something else in it, or nothing at all. Now the marked surface represents a bison . . . —(Wollheim, 1987), 46–8. Sometimes our experience of a differentiated flat surface involves two distinct aspects: (a) a configurational aspect, thanks to which we are visually aware of the surface itself and its variations in color and texture; and (b) a recognitional aspect, thanks to which we are visually aware of various robustly three-dimensional objects or states of affairs that aren’t and aren’t believed to be before our eyes at the time. The “objects” figuring in the representational aspect may no longer exist (Napoleon), may never have existed in the first place (Icarus), or may be merely generic in character (a brown-haired woman holding violets who isn’t any particular woman holding any particular violets). The configurational aspect can be described as analogous to a veridical simple seeing of a differentiated surface, which it resembles both intrinsically and in a functional role. The recognitional aspect can be described as analogous to a face-to-face seeing of the things we in fact merely see in the surface (or, alternatively, as analogous to an optically unaided visualization of those same things). Yet I can be aware of a differentiated surface in the particular way exhibited here only by employing it to discern absent three-dimensional things; and I can be aware of discerned absent things in the particular way exhibited here only by being aware of a differentiated surface whose features enable me to discern them in it. In at least this sense, (a) and (b) are inseparable aspects of a single experience rather than independent experiences which happen to occur simultaneously. And though it can be described an analogous to the simpler experiences just mentioned, there is a sense in which a detailed point-for-point comparison between it and such simpler experiences is out of the question: seeing-in and the simpler experiences to which it is in various ways analogous are “phenomenologically incommensurate.” Such is the twofoldness involved in seeing-in. A painting represents a given subject matter when we are manifestly intended to see that subject matter in its surface and can in fact do so. We aren’t just moved think in visual terms about the absent or generic object we see in a picture’s marked surface; we have visual experience of it; we see it. But this is “seeing” of a sort not countenanced by causal theories of perception along familiar Gricean or Strawsonian lines. We aren’t face to face with what we see in a picture’s surface; we don’t causally interact with it by means of our eyes; it needn’t figure among the causes of our visual experience in any way whatsoever.

274 David Hills

4 Wollheim complains that if Gombrich’s illusionism were true, we’d never be in a position to experience the marked surface as such and the picture’s subject matter as such at the same time, hence never be in a position to visually explore, visually experience, visually enjoy and appreciate the way in which the first gives rise to the second—why the second is as it is, given that the first is as it is. When we admire the great achievements of naturalistic art we do so because we think of them as very lifelike representations of objects in the real world: but to think of them in that way is quite incompatible with taking them to be or seeing them as . . . the objects themselves. Indeed, if we took the picture of an object to be that object, what would there be left for us to admire? (Wollheim 1974a: 277) By insisting that we are simultaneously aware of the marked surface and the represented object, twofoldness allows us also to be aware of the fineness of the match between the two that the great painters effected. (Wollheim 2003: 147) Up to a point, at any rate, we can experience, visually experience, how it is that a picture’s configurational aspect gives rise to its recognitional aspect. Up to a point, we can watch the configurational give rise to the recognitional before our very eyes. Watching this happen is an important source of the pleasure we take in paintings and drawings. Especially poignant or suggestive relationships between the what and how of depiction are a fertile source of visual interest and pictorial meaning. The exploration of these relationships was a chief interest of Wollheim’s comrade-in-arms Michael Podro (Podro 1987, 2001, 2004). If it weren’t for twofoldness, we couldn’t visually experience the “because” of depiction. This is a perfectly sound objection to Gombrich’s mitigated illusionism. Yet twofoldness is by no means sufficient for experiencing the “because.” For one thing, an experience can exhibit the dual-aspect structure Wollheim ascribes to seeing-in even when there is no “because” to experience, since the surface-oriented aspect of the experience doesn’t give rise to the depthinvolving aspect at all. It is a summer day without a breath of wind. You sit in the grass by a small pond surrounded with meadow, gazing up at the sky. The sky is cloudless except for a few contrails left by passing jets, contrails which present themselves as blurred white streaks adhering to a pale matte blue ground. You switch from gazing up at the sky above to gazing down at the pond’s surface. Your visual experience now has two interdependent aspects, either of which can be emphasized at the other’s expense. To the extent you continue to attend to the sky above via its reflection, your experience is still of (or as of) a flat marked surface, hence analogous to the more simply

Visions: on the nature of pictures 275 structured experience you had when gazing up at the sky directly; to the extent you instead peer through these surface markings, attending to the fish and weeds in the pond and the pebbles on its bottom, your experience is of (or as) of fish and weeds and pebbles seen in depth, analogous to the more simply structured experience you would have if you donned Polaroids to extinguish the sky’s reflection. Precisely because it presents both a flat marked surface and things arrayed in depth by means of a single visual stimulus, your experience is phenomenologically incommensurate with either of these simpler ones in precisely the manner called for by Wollheim’s characterization of twofoldness. Neither aspect of the experience could have the character it does in the other’s absence, yet neither gives rise to the other in the special way a configurational aspect gives rise to a recognitional aspect. The marked surface aspect isn’t configurational and the depth-involving aspect isn’t recognitional, in the senses Wollheim has in mind anyway. What’s seen face to face, somebody might want to say, are the fish and weeds and pebbles, not the sky’s marked surface. But of course, neither sky nor fish really are where they are visually presented as being. The sky has been relocated visually thanks to reflection, but the fish have likewise been relocated thanks to refraction. There’s no deep and principled difference between these two ways of bending light rays. In a more principled and pertinent sense, a Grice–Strawson causal sense, we see both the sky and the fish face to face, since both contribute to causing our experience of them in the manner characteristic of face-to-face seeing. Since both the sky (with its contrails) and the fish (with their accompanying reeds and pebbles) are present and causally active, neither aspect of the experience arises from the other via a Gombrichian process of recognition in absence, since neither is absent. So neither aspect of the experience is a recognitional aspect in Wollheim’s sense. There are three subtler points to notice here as well. Each bears why one might want to incorporate imagination into a fuller account of seeing-in and its workings. 1


If the “because” of depiction really is experienced and really is experienced visually, the configurational aspect and the recognitional aspect can’t jointly exhaust the phenomenology of seeing-in, since the visual experiencing we do as we “watch” the configurational give rise to the recognitional belongs to neither the configurational nor the recognitional. We must enlarge the structure of seeing-in to accommodate a third aspect and relate it to the two already recognized. What if the watching, thanks to which we experience the “because,” were a form of reflective monitoring, achieved in part by distinctively visual means, of our own imaginative activity? If Gombrich’s key-forging idea is right as far as it goes, depiction and seeing-in more generally depend on the simultaneous presence in the stimulus of cues and counter-indications: cues that readily and reliably indicate the presence of particular objects and kinds of object, or particular properties of and relations among objects, when we are face to face with them

276 David Hills


in real life; and counter-indications making it clear that this is not real life and we are in fact face to face with nothing of the sort. Such mixtures of cues and counter-indications give rise to recognitions in the acknowledged absence of the things recognized. It is thanks to recognition in absence that there is such a thing as the recognitional component. But what is cued for doesn’t exhaust what gets seen in the picture’s surface. Nothing in the stains and scribbles on Siskind’s wall is a cue for the presence of a box or an ear, for example, yet we certainly see these things in it. (Talk of a recognitional component may disguise this, since we can plausibly take “what’s recognized” either narrowly, for features cues for which are actually available in the visual stimulus and actually responded to, or broadly, for whatever we end up seeing in the wall that serves as our stimulus.) Where does the part of the seen-in that isn’t separately or directly cued for come from in the first place? What if it is in some sense supplied by the imagination? The cues we can experience to be at work in recognition in absence form a small and special subset of those at work in veridical face-to-face recognitions of actual things and kinds, actual properties and relations in real life. These can be surprising (and moving) in precisely what they turn out to be and precisely how they turn out to operate. Yet a rough sense of what they are and a rough sense of how they operate is available to us without controlled psychological experiment or sketchbook trial-and-error, thanks to our ability to respond to them as cues in the first place. Recognize a famous face in a caricature or a famous voice in the performance of an actor doing impressions and you’ll generally have a rough idea what you’re going by, a rough idea how you’re going by it, and a capacity to refine and test these rough ideas by reflecting on your pictorial experience as it stands. Cues active in veridical real-life perception needn’t be at all like this. Recognize a familiar face on the street or a familiar voice on the telephone and unless you’ve made a special study of such matters you may be completely clueless about what you’re going by and how you’re going by it. With cues like those, discovering the “because” is hard work and experiencing it is out of the question. If this is right, it needs explaining. What if the cues active in picture perception are circumstances we notice, heed, and take into account in efforts to work out what we should be “seeing” in some imagination-involving recognitional sense, given what’s there to be seen in a Grice–Strawson configurational sense? We pre-reflectively participate in the operation of these special cues; that’s what gives us relatively unimpeded reflective access to what they are and how they work.

5 Walton’s account of depiction draws on a broader conception of makebelieve. Having made a twiddling gesture at the end of a shoebox, we seat a teddy bear in the box. We ask, “Where’s the soap?”, pick up a wooden brick, and

Visions: on the nature of pictures 277 rub Teddy’s back with it. [A] two year old joins in by lifting Teddy out of the box, [announces] “He’s all wet,” and wraps him in a piece of paper. (Harris 2000: 9) Once after supper, Wittgenstein, my wife and I went for a walk on Midsummer Common . . . It occurred to Wittgenstein that the three of us should represent the movements of the sun, earth, and moon, relative to one another. My wife was the sun and maintained a steady pace across the meadow; I was the earth and circled her at a trot. Wittgenstein took the most strenuous part of all, the moon, and ran around me while I circled my wife. (Malcolm 1965: 51–2) With an ease and swiftness that would be astonishing if it weren’t so familiar, two games of make-believe are negotiated into being, games in which what participants are called on to imagine being the case at a given stage in the proceedings, what’s fictional at that stage, and how they are called on to imagine what’s fictional—for example, from what points of view each of them are to do so—depend in an intricate but mutually understood manner on what participants say and what’s actually perceivably the case at that stage concerning various of the game’s props. The sayings and doings of the participants and their interactions with one another and with inanimate props spin out a sort of yarn for their mutual entertainment and edification. Some strands of the spun-out yarn get put in place by the sheer stipulative say-so of one of the parties, others by various kinds of enactment, still others by borrowing for present imaginative purposes assumptions the participants are used to making for other purposes in other contexts. Publicly discernible states of affairs (at a given stage in the proceedings) serve to determine what’s fictional (in the corresponding episode of the spun-out yarn) in accord with rules of make-believe—principles of generation—players of the game manage to put in place as they go. How players manage this, how they develop and maintain a single shared set of rules to play by, rules they’ve never played by before and will never play by again, rules they may be unable to state even as they play by them, is a good question—one of the most urgent and under-studied questions in the cognitive psychology of human culture. What each player has to go on in interpreting the proceedings and deciding which rules should govern his own imaginings, consists for the most part of: (a) what he can see or hear or otherwise discern about the proceedings themselves, his experience in the here and now of the actual states and doings of available potential determiners of fictionality, witting and unwitting, animate and inanimate, things Walton calls props; (b) a stock of precedents, customs, conventions, and expectations deriving from his past experience as a player of similar games with similar partners;

278 David Hills (c) his personal sense of the relative imaginative interest or promise or appeal of different available ways of interpreting what’s happened to date and different available ways of proceeding from here on out; and (d) last but not least, a thin stream of verbal and gestural signals from the other players, calculated to convey something about their working understanding of the rules, the states and doings of relevant props, and what the two in combination call on all concerned to imagine, according to them. In the Harris game, two such signals are the adult’s “Where’s the soap?” and the two-year-old’s “He’s all wet.” In them the parties commune with one another in their capacity as co-directors of a game of make-believe and co-authors of what they are subsequently called on to imagine. The adult refers to one of his props, a block, by means of a role he expects it to assume in the game he hereby initiates, that of a cake of soap, all in the course of talking himself through a serious effort to locate the block. The child for his part expresses what he takes to be fictional at a certain stage in the proceedings, thereby helping the adult to discern principles of generation required to make proper imaginative sense of what he does next. Whether the parties simultaneously speak “in character” in their capacity as role-playing actors—whether we are to imagine a real bather seriously seeking to locate a real cake of soap by seriously asking himself where it is, whether we are to imagine a real dryer-off seriously trying to point out a real bear’s real wetness to a real bear-bather—may be unclear or indeterminate in this case. Signals reflect the speaker’s working impression of the states and doings of relevant props, his working understanding of the rules linking such states and doings with what participants are called on to imagine, and his working sense of what all of us are in fact called on to imagine, so that the clearer a given listener already is about one of these matters, the more informative the signal will be about the others. The generation of what’s fictional comes in four pervasive major modes (there are many minor modes as well). We Spell Things Out when we propose to our fellow players that something be fictional (either at the current stage of play or until further notice) by going through the motions of asserting it—speaking in what we might call a let’s say spirit. If nobody shows any sign of resisting our proposal, it is often accepted and takes effect. We Carry Things Over when a bit of common knowledge or familiar folklore becomes sufficiently salient to all concerned, manifestly coheres with what is already fictional, and is such that its fictionality would be of real imaginative interest in the game at hand. If nobody resists, the relevant bit of lore often becomes fictional without ever getting put into words. We Work Things Out when a good inference, deductive, inductive, explanatory, or whatnot, running from things already taken as fictional to some further conclusion, becomes sufficiently salient to all concerned, the conclusion manifestly coheres with what is already fictional, and its fictionality would be of real imaginative interest

Visions: on the nature of pictures 279 in the game at hand. If nobody resists, the conclusion often becomes fictional, despite the fact that neither it nor the inference leading up to it are ever put into words. But it is Making Things Up by Acting Them Out, enactment with suitable actors and inanimate props, that makes pickup make-believe the distinctive and striking form of human activity that it is. In Acting Out, the determination of what’s fictional proceeds via a kind of as-if showing (rather than the as-if telling that occurs in Spelling Things Out). The very state of things we are to imagine is as if set before our eyes, as if shown to us. But it is hard to get the nature of this as-if showing in clear focus. Plato conceived it as a matter of replicating the perceivable outsides of perceivable things or actions, to the extent permitted by one or another medium of portrayal. That won’t do at all, yet there are undigested residues of Platonic replicationism in the accounts of make-believe offered by developmental psychologists and philosophers even now. A twiddle manages to suggest the actual deliberate opening of a tap without resembling such an opening at all in any independently specifiable respect, let alone replicating it. If you doubt this, first play Harris’s game yourself, then go to the bathroom and actually open an actual tap.

6 So far, I’ve followed Walton’s own account of make-believe pretty faithfully, but at this point I need two emendations that will matter later on. The first concerns the involvement of an imaginer in the content of her own imaginings, the second concerns how best to understand the form of generation I call enactment, or Acting Things Out. The Shakespeare scholar A.D. Nuttall wrote his thesis under Gilbert Ryle and locked horns with his supervisor more than once. Gunning for Cheshire Cats in The Concept of Mind, Ryle writes: The pictured smile is not, then, a physical phenomenon, i.e. a real contortion of the doll’s face; nor yet is it a non-physical phenomenon observed by the child taking place in a field quite detached from her perambulator and her nursery. There is not a smile at all, and there is not an effigy of a smile either. There is only a child fancying that she sees her doll smiling. (Ryle 1949: 248) To which Nuttall responds: I can imagine (or fancy) that I am looking at a rose, and I can also imagine the rose, quite simply. In the language which admits talk about objects, in the first I am a personage in my own imagined scene, in the second only one object is present, the rose. (Nutall 2007: 69)

280 David Hills Walton sides with Ryle on this; I find I must side with Nutall instead. Sometimes we imagine the obtaining of various states of affairs, the truth of various propositions, and that’s that. Other times we do our imagining from some particular perspective or point of view on the states of affairs we imagine to obtain, the subject matter of the propositions we imagine true. I’ll call such perspective-involving imagining envisioning. Perspectives from which we envision things seem always to be ones from which we or others could perceive them, at least in principle. The perspectives from which a given person can perceive things, the perspectives she occupies as a perceiver, depend on various epistemically momentous relations she stands in to those things, relations that enable her to perceive part of the truth about them and preclude her from perceiving the rest. There are the various sense perceptual perspectives on our physical surroundings we can occupy as perceivers by standing in particular places, facing in particular directions, and making free and effective use of various bodily senses. There is a special “central” or “inside” perspective on a person’s thoughts and feelings and mental activities that she and she alone can occupy as a perceiver by cultivating an alert, reflective state of mind and simply being who she is. There may be still further kinds of perspective as well, associated with less obvious forms of perceptual awareness. Now envisioning and perceiving from a given perspective seem linked in a striking manner. When and insofar as we envision things from a given perspective, the properties we imagine things to possess and the relations we imagine them to stand in will be ones we imagine to be perceptually detectable from it by those (if any) who are able to occupy it as perceivers. We could make tidy and connected sense of all this by supposing the following: (a) Imagining, even the perspective-free kind, has a built-in reflexive aspect that differentiates it from mere supposition. In imagining that p, S ipso facto also imagines being aware that p, imagines true a proposition to the effect that she herself is aware that p in some way or other. Imaginings are always about the imaginer, at least in part. (b) What “perspective” adds to purely propositional imagining is simply further propositional imagining on S’s part, to the effect that she herself is aware that p because she herself perceives that p from some particular perspective. Such is the account of self-imaginings Walton offers (1990: 28–35).2 Some would question the coherence of these suggestions. When it comes to (a), can’t I imagine that unbeknownst to us all, q? And in that case, won’t I be imagining both that none of us know that q and that I do? Walton would reply that although do I imagine both those things, I needn’t imagine their conjunction or needn’t do so at all vividly. Imaginative projects tolerate this sort of incoherence remarkably well. The Adventures of Huckleberry Finn represents itself as a carefully crafted, deeply reflective 300-page memoir by a barely

Visions: on the nature of pictures


literate young man preparing to light out for the territories because the prospect of any work that forces him to sit still scares him half to death. Is imagining knowing something that never gets known any harder than imagining reading something that could never have got written? When it comes to (b), can’t I visualize myself giving a speech from outside—for example, from somewhere in the second row of my own audience? Can’t I envision from inside Napoleon’s thoughts and feelings at a moment of triumph on some smoking battlefield? Walton could respond that I can imagine myself occupying the relevant perspectives by imagining myself to be someone else, imagining true a de se proposition to the effect that I myself am that someone else.3 I would resist it on more descriptive grounds instead. When I imagine things from a sense-perceptual perspective, I needn’t imagine the perspective to be perceptually occupied by anyone at all. And when I imagine things from someone else’s central or inside perspective, I needn’t imagine that perspective to be perceptually occupied by me; the only perceptual occupant I must imagine it to have is the only one it does or can have as a matter of fact: that particular other person. Here are some considerations I find compelling; Walton doesn’t find them compelling at all. In so-called observer memory (Freud, Nigro and Neisser)4 one recalls an event or scene from one’s own personal past from a viewpoint one couldn’t have occupied at the time, with the result, sometimes, that one’s memory includes a vague but unmistakable view from without of one’s own past self. An observer memory of mine doesn’t represent me as having seen myself from without at the relevant time in the past; it needn’t represent its viewpoint as having been available to anyone at all at the time. If memories can involve viewpoints remembered to be devoid of perceivers, why can’t imaginings involve viewpoints imagined to be devoid of them? To follow an informal proof in solid geometry, I may have to envisage a particular configuration of lines and planes from the external point of view adopted by a diagram in the text. But I don’t thereby represent myself (or anyone else) as occupying and enjoying that point of view as a perceiver. The called for imagining isn’t about me or about otherwise unidentified perceivers—it’s just about the lines and planes. Some Spellings Out call on us to visualize their propositional content from a more or less definite point of view. Here’s an example from the psycholinguist Herb Clark. The opening of Hemingway’s “The Killers” reads: The door of Henry’s luncheon opened and two men came in. This calls on us to envision the entry of the two men from a point of view inside the luncheon. They turn out to be the killers of the title. The point of view from which we envision them intially lacks an owner in the story but gradually attaches itself to a character already seated in the luncheon, Nick Adams. Had Hemingway written The door of Henry’s luncheon opened and two men went in.

282 David Hills we would have to envision their entry from outside the luncheon, quite possibly from a viewpoint in the middle of the street. Would we thereby need to imagine that we or anyone else witness their entry from the middle of the street? No. Neither the actual story nor our imagined modification has it that there is anyone in the middle of the street outside the luncheon when the killers enter. The Haida of the Pacific Northwest have a trickster tale. Once upon a time the world was in total darkness. This bothered Raven, who got tired of groping and bumping into things. He heard rumors about an old man, living alone with his daughter, who kept all the light in the universe in a tiny box, the innermost box in a nest of bigger and bigger boxes, which he stored on a shelf in his lodge. Raven decided to steal the light, and came up with a plan. He waited for the old man’s daughter to come to the river for water, turned into a hemlock needle, and fell in the river just as she dipped her water basket into it. When she drank from the basket, she swallowed the needle. Raven grew and grew inside the daughter until eventually she gave birth. The old man quickly came to dote on his strangely birdlike grandson, who begged and begged to be allowed to play with the biggest box, then the next biggest, and so on, until he finally held the light itself in his chubby little hands in the form of a tight glowing ball. At that point he turned back into Raven, gave a mighty caw, caught the ball in his beak, flew out the smokehole of the old man’s lodge, and threw the ball into the sky where it became the sun and moon and stars. Hearing this tale, we envision Raven’s frustrated stumbling about, the affair with the hemlock needle and the water basket, the beaklike nose of the strange child Raven becomes—despite the fact that none of this was seen or could have been seen by anyone at the time, since the relevant episodes occur in total darkness. We don’t confine the terms in which we envision them to relations and properties accessible to touch and hearing; we envision them in visual terms, just as Raven must have done when plotting out how to steal the light in the first place. Nor are we confronted here with the mixture of logical incoherence and fictional coherence we encountered with Huck Finn. It isn’t as if to respond to this tale as it demands we had to imagine both that the lights are out (so Raven can’t see) and that they are on (so that we can.) According to the tale, the lights are out until Raven contrives to turn them on and that’s that. Whatever structural affinities there may be between sensory perceivings and envisionings, I can envision things without thereby imagining that I myself perceive them—indeed, I can envision things without thereby imagining that they are or could be perceived by anyone. Wollheim himself sides with Nutall and me against Ryle and Walton on this point (Wollheim 1974b). He regards “seeing” as a genus with two species: face-to-face seeing of the sort Grice and Strawson had in mind, a kind that uses the eyes; and visualization, a kind that uses only the eye of the mind, a form of seeing that is at the same time a form of imagining. In visualization

Visions: on the nature of pictures


I start with a set of conceptual specifications for an object or scene and proceed to conjure up something that answers to those specifications, presenting it to myself in visual terms from some particular visual point of view. To be sure, Wollheim treats “visualize” and “imagine seeing” as more or less synonymous. Yet he doesn’t view the latter as short for “imagine that I myself see”; he appears to view it as an unbreakable idiom.5

7 The second emendation, concerning enactment, is a little more intricate. We’re playing cowboys; I point and yell “Bang! You’re dead!” The “You’re dead!” part is a signal. The word Bang! participates in the enactment of a killing or, at least, the enactment of a gunshot, participates or tries to participate in rendering it fictional that someone is shot dead—this despite the fact that nothing remotely resembling that word could contribute in any significant way to rendering it actual that someone was shot dead. (The noise of a gun’s discharge is in no way lethal; it’s a nonlethal side effect of a lethal causal process. It doesn’t sound much like the word Bang! anyway.) The uttered word Bang! provides the soundtrack to a moving picture of a shooting whose visuals are provided by my pointing finger, a non-static tableau vivant whose content depends on conventions of portrayal vaguely akin to the conventions of depiction in the visual arts. Those conventions may be familiar and stereotyped (as they are here) or put in place mysteriously on the fly (as they are in more typical cases). What makes it fictional that someone is shot dead, if anything does, is that someone yells the word Bang! while pointing with sufficient accuracy at a sufficiently critical portion of a would-be victim’s anatomy. This is what it takes (nearly enough, in a game of this excruciatingly familiar sort) to make it fictional that one has just now shot one’s would-be victim dead. In pickup games of make-believe we assign roles or parts to concrete perceivable actors and props. In playing their assigned roles, the actors come to portray various, often distinct, often very different, people or kinds of people; the props come to portray various, often distinct, often very different, objects or kinds of object. We have: (a) definite role assignment, where one assigns X the role of Y; and (b) indefinite role assignment, where one assigns X the role of a G (some G or other, no particular G in particular); (c) generic role assignment, where one assigns Fs in general the role of Gs (without assigning any particular F the role of any particular G). What am I called on to do with my imagination while such a role assignment is in force? For starters: on sighting or otherwise spotting X [an F], I am to envision a G and do so from a perspective suitably modeled on that from which I experience X [the F] here and now. But there’s more to it

284 David Hills than that. As our reflections on Bang! just showed, concrete perceivable actors and props get to play the roles of other things—get to portray other things to determinate fictionally propositional effect—only because various of their perceivable properties and relations and doings fictively count as generally distinct, generally very different properties and relations and doings, perceivable or otherwise, appropriate to the things portrayed. And vice versa: perceivable properties and relations and doings fictively count as other properties and relations and doings, perceivable or otherwise, only because appropriate roles are assigned to concrete actors and props, actual or possible, that might bear the properties, stand in the relations, engage in the doings. In the determination of what’s fictional, concrete actors and props on the one hand and abstract bearers of fictional identities on the other go together like subjects and predicates. One way for it to become fictional that a particular person or thing Y possesses property Q is for some actor or prop X that has what it takes to portray Y, what it takes to assume the definite role of Y, to possess some property P that for its part has what it takes to fictively count as a form of property Q: X






counts as

| Y

| has


For instance, in Wittgenstein’s game it is fictional that the sun moves with uniform velocity because Mrs Malcolm assumes the definite role of the sun—striding along at a steady pace in a roughly straight line fictively counts as moving with uniform velocity and, as a matter of perceivable fact, Mrs Malcolm does stride along at a steady pace in a roughly straight line. (If she misbehaved, it will become fictional that the sun itself misbehaves, so a role this important in a game such as this is a sacred trust.) One way for it to become fictional that some G or other (no particular G in particular) possesses property Q is for it to become fictional that some G has Q, for an actor or prop X that has what it takes to portray a G, someone or something that has what it takes to assume the indefinite role of a G, to possess a property P that has what it takes to fictively count as a form of Q. X





counts as

| aG


| has


Visions: on the nature of pictures 285 For instance, in Harris’s game it is fictional when a certain stage in the game is reached that a towel (no particular actual towel) serves to dry Teddy off because a certain piece of paper then has what it takes to assume the indefinite role of a towel and (the property of) being loosely covered in a thing like that (the piece of paper) has what it takes to fictively count as (the property of) being snugly wrapped in and dried off by a towel (some towel or other). And so forth. (Whew!) Roles often accrue to things and kinds and fictive identities to properties and relations via special forms of Carrying Over peculiar to role-play: a salient knowledge or folklore to the effect that X is Y (or an F) will often suffice to give X the role of Y (or an F); salient knowledge or folklore to the effect that to have P is ipso facto to have Q as well will often suffice to make P fictively count as a form of Q, etc. The relation between rendering fresh things fictional and the assigning of fresh definite, indefinite, and generic roles is a two-way street, thanks to special forms of Working Out peculiar to role-play: If X already has the role of a G and it becomes fictional by whatever means that Gs are Hs, X thereby assumes various further roles: that of a G, that of a thing that is at once a G and an H, etc. If X already has the role of a G and it becomes fictional by whatever means that there is one and only one G, namely Y, X thereby assumes various further roles, that of Y, that of a thing that is at once a G and Y, etc.

8 What I’ve said so far about role playing and enactment is perfectly compatible with Walton’s views as they stand. But we’re about to part company again. Someone might contend that I’ve so far left out the simplest and most fundamental thing a role assignment calls on me to do. For it is natural to suppose that: (a) When X is assigned the (definite) role of Y, I’m called on to imagine that X is Y, the is in question being that of identity. (b) When X is assigned the (indefinite) role of a G, I’m called on to imagine that X is a G, the is in question being that of predication. (c) When Fs are assigned the (generic) role of Gs, I’m called on to imagine that Fs in general are Gs, the are in question being the plural form of an is of predication. In other words, when a role is assigned, I’m called on to imagine true a statement of identity, membership, or inclusion that imaginatively assimilates the player (or players) of the role in question to the thing or kind that it (or they) are being called on to portray. Now the forms of imaginative assimilation

286 David Hills described in (a)–(c) are perfectly possible imaginative feats. But one may question whether role assignment involves or calls for such feats. It’s widely and understandably assumed that it does. Consider this from Northrop Frye: As for metaphor, where you’re really saying “this is that,” you’re turning your back on logic and reason completely, because logically two things can never be the same thing and still remain two things. (Frye 1964: 32–3) Or consider this, from Gombrich himself: When we make a snowman we do not feel, I submit, that we are constructing a phantom of a man, we are simply making a man of snow. We do not say, “Shall we represent a Man who is smoking?” but “Shall we give him a pipe?” . . . We feel tempted to work the snow and balance the shapes till we recognize a man . . . a member of the species man, subspecies snowman. (Gombrich 1960: 99–100) One gets drier formulations to a similar effect in contemporary developmental psychologists: Harris, Leslie, and others. Call the view laid out in (a)–(c) assimilationism. If assimilationism is correct, role players invariably figure in the content of the fictions they help to generate just as much as the things and kinds they serve to portray do, thanks to a standing imaginative assimilation of the former to the latter. Consider this expression of assimilationist sentiment from Walton: Spectators imagine of Sir Laurence Olivier, when he plays Hamlet, that he is Prince of Denmark; there is a real person before them who they imagine to be faced with the task of avenging his father’s murder, to hesitate in carrying it out, and so forth. (Walton 1990: 26) According to Walton, fictional truths about fictional characters like Hamlet are generated from corresponding fictional truths about the actors playing them by modes of generation already at work in our response to stipulative narration, in some such manner as the following: 1 2


We see Olivier say and do various eminently perceivable things (on stage). It is therefore fictional about Olivier that he says or does these eminently perceivable things (in some circumstances or other). (By a form of Carrying Over.) Thanks to prior events on stage, it is already fictional about Olivier that he is in circumstances where saying or doing such things would be good evidence that one is called on to avenge one’s father’s death and trying to act accordingly.

Visions: on the nature of pictures 4 5 6


So: it is fictional about Olivier that he is called on to avenge his father’s death and trying to act accordingly. (By a form of Working Out.) It is fictional about Olivier that he is (identical to) one Hamlet, Prince of Denmark. (Role Assignment.) So: it is likewise fictional that one Hamlet, Prince of Denmark, is called on to avenge his father’s death and trying to act accordingly. (By a second form of Carrying Over.)

Now not only is the final claim about Hamlet fictional: we are to imagine its truth from a perspective on Hamlet and his doings modeled on that we enjoy as perceivers on Olivier and his doings. Where might this additional requirement come from, if enactment’s characteristic effects are to be explained in terms of Carrying Over, Working Out, and imaginative assimilations of portrayers and portrayeds in the manner illustrated by the argument I just gave? If to visually envision were to imagine seeing, one could see how this might go: I see Olivier from such and such a distance, at such and such an angle. Often, where I do see Olivier from will be where I should imagine seeing him from, by a form of Carrying Over. If to envision is to imagine seeing, where I should imagine seeing Olivier from is ipso facto where I am to envision him from. But it is fictional that Olivier is Hamlet (Role Assignment). So by a form of Working Out, where I should envision Hamlet from, where I should envision Olivier from, and where I do see Olivier from are one and the same perspective. Assimilationism needs the mistake about envisioning I criticized in Section 6. But the difficulty I want to focus on is a different one, embodied in the italicized part of the Walton passage and step 4 of the argument above. In “Imagination and the Self,” Bernard Williams discusses a second performance by Olivier—in Othello, as it happens—and strenuously balks at counterparts to step 4 (Williams 1973). He’s right to balk, I think. Even if Olivier’s performance in Hamlet calls on us to imagine various things to be the case about Olivier himself, it’s going too far to suppose we’re called on to imagine particular things about Olivier’s relations to his actual family. Leave his dad out of it! Things might be different if Olivier were a method actor and we were a method audience. But he isn’t and we aren’t. The difficulty generalizes: assimilationism makes the generation of fictionality by enactment depend on the prior fictionality of things that intuitively aren’t fictional at all. The trouble is, assimilationism seems not just available but inescapable. It is hard to deny, given what we’ve already said, that When X is assigned the role of Y, I allow X to appear in the guise of Y, imagine X to be Y. When X is assigned the role of a G, I allow X to appear in the guise of a G, imagine X to be a G.

288 David Hills When Fs are assigned the role of Gs, I allow Fs in general to appear in the guise of Gs, imagine Fs to be Gs. When P takes on the fictive identity of Q, I allow P to appear in the guise of Q, imagine P to be Q. And imagining this to be that, isn’t it just a matter of imagining that this is that? Arguably the two formulations are an innocent syntactic transformation apart, the second obtained from the first by an innocent form of subject raising. And isn’t imagining that this is that simply a matter of imagining true a certain proposition, the proposition that this is that (in the sense of is appropriate to the case at hand)? So it may seem—yet when one reflects at all candidly on any actual game, the following seem undeniable: When X is assigned the role of Y, I’m not thereby called on to imagine that X and Y are one and the same, that Y and X are one and the same, that any one thing is (identical to) both Y and X. When X is assigned the role of a G, I’m not thereby called on to imagine that X belongs among the Gs, that among the Gs is to be found X, that any one thing is both X and a G. And when Fs are assigned the role of Gs, I’m not thereby called on to imagine that the Gs include the Fs, or (even) that some Gs are also Fs. Once a role is assigned, am I or am I not called on to imagine that the thing (or kind of thing) playing the role is the thing (or kind of thing) whose role it plays? Didn’t we just now answer this important question both ways? What gives? I’m an Austin-style romantic, wedded to the notion that ordinary language never did betray the heart that loved it, but surely it betrays us here. Some of the imaginings involved in role-play concern portrayed things and events exclusively. Others concern portrayers and portrayeds at once, since in them we imagine suitable portrayers to be suitable portrayeds. The content of an imagining of either sort can be specified by means of a that-clause indicating what we are said to imagine. But imaginings of the second sort aren’t a matter of imagining true the propositions we employ in specifying their content. They aren’t a matter of imagining true any propositions whatsoever. They only appear to relate imaginers to propositions. They are pseudopropositional imaginings. Something similar happens with belief locutions. Suppose Jones is abstractly aware of slender four-footed unfroglike amphibians known as salamanders. But whenever he encounters a salamander he takes it for a lizard—he takes his own casual observations of salamanders as confirming or disconfirming his own casual generalizations about lizards, etc. It will be natural to say (in some contexts and for some purposes) that Jones believes

Visions: on the nature of pictures


salamanders to be lizards, or even that he believes that salamanders are lizards, despite his not believing true the proposition that salamanders are lizards. Jones just takes, mistakes, salamanders for lizards. After all, he may know full well (as a matter of general conceptual bookkeeping) that salamanders are amphibians, lizards are reptiles, and never the twain shall overlap. (The same general sort of thing happens with other propositional attitude verbs as well.) Not all believing that Fs are Gs is a matter of believing true the proposition that Fs are Gs; some of it is taking Fs for Gs. By the same token, not all imagining that X is a G is a matter of imagining true the proposition that X is a G; some of it is fancying as—where to fancy a stick as a horse is to give that stick the role of a horse in some suitable game, and to fancy a twiddle as the opening of a tap is to treat that twiddle as an enactment of the opening of a tap in some some suitable game. Assimilationism gets its grip from our failure to recognize that verbs like “imagine” and “believe” (and “want” for that matter) can pair with forms of the verb “to be” to report attitudes that are only pseudopropositional in their content. Once we reject assimilationism, it becomes natural to contend that role players—portrayers—figure in the propositional content of the fictions they help generate only when they double as portrayeds—self-portrayal being an interesting and common special case. To be sure, we imagine things about unportrayed portrayers, in that we imagine them to be the things they portray. They are among the objects of our imaginings, hence not mere props, to that extent. Yet they don’t figure in the content of what the relevant game renders fictional—don’t figure in the content of the spun-out yarn—since the imaginings in which they figure aren’t propositional. If anti-assimilationism is right, there are at least three things we can mean by “imagining X to be a G.” First, one can imagine it true (de re) of or about X that it is a G. That option has been on the table all along. Second, if X is a concrete thing capable of serving as a prop, a suitably shaped lump of plastic perhaps, one can imagine it to be, for example, a gun by giving it the role of one. Third, if X is an action, event, or state of affairs capable of enacting some other action, event, or state of affairs—for example, a suitable combination of pointing and yelling—one can imagine it to be a killing by treating it as an enactment of one.

9 I turn at long last from make-believe in general to depiction in particular, and Walton’s claim to offer Wollheim’s seeing-in a much needed fleshing out. Early in Philosophical Investigations II. xi, “pages which cast alternating bands of light and darkness on the topic of [pictorial representation]”:6 Here it is useful to introduce the concept of a picture-object. For instance, the figure would be a “picture-face”.

290 David Hills

In some respects, I engage with it as with a human face. I can study its expression, can react to it as to the expression of the human face. A child can talk to a picture-man or picture-animal, can treat them as it treats dolls. (§119) To which Walton would add: I can engage with it in imagination as I engage with a human face for real. Images substitute or fill in for their originals only in the sense that they enable us to do and think and feel in imagination things their originals enable us to do and think and feel for real. And they enable us to do and think and feel these things in imagination by calling on us to do so, by helping to make it fictional that we do and think and feel these things (for real)—helping make it fictional, since what it is we fictionally do and think and feel depends jointly on what the images are like and how we proceed to interact with them. Glossed in this somewhat fussy manner, Wittgenstein’s talk of “picture-objects” and Gombrich’s of “species horse, subspecies hobby horse” are genuinely suggestive. An image is a prop in a game of make-believe, an engaging educational toy. This account of images in general applies to appearance-capturing images, depictions, in a highly specific manner. First, the games we play with depictions as props are visual games, in a twofold sense. We play them by interacting visually with depictions, inspecting them in various ways, and experiencing them visually in various ways as a result. And the yarns we spin out by visually interacting with depictions as we do in such games have as their core subject matter fictitious visual interactions between us and the depicted objects. Both points are relevant to the sense in which, for Walton, the experiences depictions are meant to afford us are visual experiences. Second, depictions contrast with effigies (dolls, hobby horses, etc.) in that they don’t portray the things they represent, don’t assume the roles of those things in the enactment of fictional truths: [Does one imagine] the brown mass of color that one perceives on the canvas to be a horse? Savile says that the viewer does not do this, and surely he is right. (Walton 2008: 137) Depictions are in no sense imagined to be, hence in no sense seen as or taken for, the things they depict. They aren’t objects of our imaginings at all. A depiction doesn’t stand in for the thing(s) it depicts: the most that happens is that our view of it stands in for a view of them. Walton retains the overall structure of Wollheim’s account of representation, with its contrast between configurational and recognitional aspects. But he holds that imagining figures helps to explain the recognitional aspect

Visions: on the nature of pictures


Figure 12.2 Meindert Hobbema, Mill with the Great Red Roof 1662–4 (Art Institute of Chicago, oil on canvas, 81cm × 110cm).

of seeing-in—something Wollheim himself consistently denied. Here are the details, as stated in Mimesis as Make-Believe:7 The viewer of Meindert Hobbema’s Mill with the Great Red Roof [Figure 12.2] plays a game in which it is fictional that he sees a redroofed mill. As a participant in the game, he imagines that this is so. And this self-imagining is done in a first person manner: he imagines seeing a mill, not just that he sees one, and he imagines this from the inside. Moreover, his actual act of looking at the painting is what makes it fictional that he looks at a mill. And this act is such that fictionally it itself is his looking at a mill; he imagines of his looking that its object is a mill. (Walton 1990: 293) Our imaginings and our perceptual experiences when we look at pictures are even more intimately connected than this. One does not first perceive Hobbema’s picture and then, in a separate act, imagine that perception to be of a mill . . . The imaginings called for when one looks at a picture inform the experience of looking at it. The seeing and the imagining are inseparably bound together, integrated into a single complex phenomenological whole. (ibid: 295)

292 David Hills We should add that the games must be sufficiently rich and vivid visually. They are rich to the extent that they allow for the performance of large variety of visual actions, by virtue of actually performing visual actions vis-à-vis the work. It will be fictional of the viewer of Hobbema’s painting, depending on how he actually looks at it, that he notices a woman in the doorway . . . or that he searches the trees for squirrels or examines the wood for worm holes, that he gazes casually toward the fields in the distance or stares intently at them, and so on. (ibid: 296) Walton proposes that a depiction D of object O is a prop in a game of make-believe, a prop whose function in that game has the following features: 1

Seeing. The spectator is to look at and hence come to see the depiction D. How he looks at D determines how he is to imagine he looks at O. What he then sees about D and how he then sees it determines what he is to imagine seeing about O and how he is to imagine seeing it. In particular:


Reconstrual. He is to imagine his act of looking at D to be an act of looking at O from a perspective D contrives to specify in some way or other. And he is to imagine the resulting experience of seeing D to be instead an experience of seeing O resulting from the aforesaid (fictional) looking at O. This condition is designed to capture an important difference between the imagining (in visual terms from a visual perspective) that we do in response to a visually evocative written description and the imagining (in visual terms from a visual perspective) that we do in response to a picture. In each case, actual observations that concern one thing settle what we are to imagine observing concerning something else. In the description case, we aren’t called on to imagine anything about our observations of the printed words. But in the depiction case, we’re to imagine our observations of the marked surface to be something they actually aren’t: observations of its subject.


Perspectives. He is to do all this so that he ends up imagining looking at and thereby coming to see O from the picture’s own viewpoint—ends up imagining it true that he himself looks at and thereby comes to see O from that viewpoint—where he does this both (a) vividly and (b) from the inside. This condition is designed to capture the thought that a picture has to land us as imaginers in the very perspectives we imagine ourselves occupying as perceivers. On the one hand, we need to end up imagining O itself from the very (outer sense, spatial) perspective from which we imagine ourselves seeing it. On the other, we need to end up imagining

Visions: on the nature of pictures 293 our seeing of O from the perspective we regularly have as perceivers of our own seeings: the “from the inside” perspective afforded us by being who we are and cultivating a reflective turn of mind. Reconstrual is expected to make the fulfillment of this further condition possible, and we must endeavor to see how this works. 4

Richness. The game leaves the spectator free to look at D in a wide variety of different ways, where how he ends up seeing D and what he gets to see about it depend in a rich and detailed manner on which way he looks at D in the first place. Yet its rules are such that how it is he is to imagine seeing O and what he is to imagine getting to see about it depend in turn in a rich and detailed manner on how he actually sees D and what he actually gets to see about it.


Phenomenal Unity. Imaginings can relate to things I do or undergo in order to imagine in two different ways. I can do or undergo something with an eye to inspiring, prompting, or otherwise inducing subsequent imaginings in myself. In that case my imaginative goal and my imaginative means are distinct existences standing in a external relation of effect to cause. Instead I can do or undergo something so as to thereby already imagine the thing in question. I can do or undergo something in such a manner that imagining the thing in question is part and parcel of doing or undergoing the thing in question in that manner. The qualitative nature of my doing or undergoing would be different, unless I were imagining the thing in question. In that case my imaginative goal is internally related to my imaginative means, built into the manner in which I do or undergo something else. Depiction-involving games require the relation between fiction-generating doings and undergoings and the imaginings they call for to be of this second internal kind. They don’t mandate that we manage our imaginings as called for by doings and undergoing already in place. Instead they mandate that we keep the details of our ongoing doings and undergoings and the imaginings inherent therein as called for by the broader outlines of those doings and undergoings. Called for imaginings and the perceivings that call for them must come into existence together via a process of mutual accommodation. This last condition is designed as a gloss on twofoldness: the two aspects, one experience portion of Wollheim’s claims about seeing-in. Perception (looking and seeing) of the kind appropriate to depictionbased games can only exist in the presence of the imaginings appropriate to such games—and, of course, vice versa.

10 Reconstrual does seem important to capturing the difference between depictions and evocative descriptions like Hemingway’s. But it is problematic if

294 David Hills we read it in the manner Walton himself appears to insist on, a de re propositional manner: [The viewer] imagines her actual perceiving of [a painting of a horse] to be an act of perceiving a horse. (She does not imagine her perceiving to be both a perceiving of a canvas and a perceiving of a horse, of course; she imagines of the perceiving which is in fact a perceiving of the canvas that it is a perceiving not of the canvas but of a horse.) (Walton 2008: 137) More formally: 2.* The spectator is to imagine it true of or about his act of looking at D that it is instead an act of looking at O (from a particular perspective that D contrives to indicate somehow). He is to imagine true of or about his resulting experience of seeing D that it is instead an experience of seeing O resulting from the aforesaid (fictional) looking at O. We may sometimes imagine of some particular experience of ours that its nature, object, and content are other than they actually are in various dramatic respects, where this is sometimes only subtly different from imagining that instead of having this experience we have some other experience with a different nature, object, and content in its place. Plausible cases of imagining an experience to be otherwise in this or that respect are diverse, since the terms on which we individuate experiences in the first place are diverse and subtle. Yet I find it hard to believe this sort of de re propositional “imagining otherwise” about particular actual experiences plays any decisive role in depiction. We’ll imagine as a depiction calls on us to imagine, it seems to me, if the sorts of experience of depicted objects we imagine ourselves to have relate in called-for manners to the sorts of experience we actually have of the depiction itself. The numerical identities of our actual experiences don’t figure here on either side—in either the content of what we’re ultimately called on to imagine true or in the specification of the facts on the ground to which our imaginings true are ultimately responsible. Perfectly parallel things can be said about evocative descriptions, of course. “On observing Hobbema’s canvas, one imagines one’s observation to be of a mill.”— I imagine my observation of a portion of the canvas to be an observation of the mill on some understanding of the “imagine to be” locution, but it needn’t be one on which I am to imagine exotic things to be true of my actual observation of the canvas. Offhand, what one’s perusal of the canvas renders fictional in no way concerns that very perusal.— “One imagines of one’s observation that it is of a mill.”— In some sense, but we have seen that such locutions admit of pseudopropositional readings. Am I also to imagine that among the observations I make of the mill is this very observation of mine (which is really of part of the canvas)? We already know that not all

Visions: on the nature of pictures 295 imagining about is a matter of imagining true propositions about.—“But in the depiction case, my visual experiences aren’t just props, they’re objects of my imaginings as well. That differentiates depiction from depiction.”—I don’t deny that for an instant; I want to understand it. At which point it may occur to a defender of 2* to contend that certain de re propositions about actual experiences are and must be fictional after all, that their fictionality is important to understanding how Reconstrual explains Perspectives, how Reconstrual serves to place us as imaginers in relation to O and our own imagined seeings of O. The thought is that calledfor imaginings derive their perspectives from the perceptions that call for them along lines already explored with Olivier: I see D and am appropriately reflective. I therefore perceive from inside this (my current seeing of D). I imagine it true (de re) of this (my current seeing of D) that it is a seeing of O from P, the perspective on O that D contrives to signal. (Reconstrual.) So I am in a position to imagine from inside that I see O from P. (Working Out.) But to do that is to imagine seeing O from P (from my special inside perspective), and to imagine seeing O from P at all just is to imagine O from P. We thus get both the appropriate imaginative perspective on O itself and the appropriate, “from inside” imaginative perspective on our imagined seeing of O. But this proposal is as suspect as the Olivier one on which it is modeled. It relies on both the hermeneutic strategies of assimilationism and the doctrine that to visualize is to imagine seeing, things we’ve found reason to reject. Both criticisms of step 2* point in the same direction. Perhaps we can preserve the felt pertinence of step 2 to understanding both the depiction/ description contrast and the perspectives from which we imagine in response to depictions if we can reinterpret its imagine to be locution along more promising pseudopropositional lines. The natural thing to try is: 2.** The spectator is to treat his act of looking at D as an enactment of an act of looking at O (from a particular perspective D contrives to specify in some way or other). And he is to treat his resulting experience of seeing D as an enactment of an experience of seeing O that results from the aforesaid (fictional) looking at O. We imagine things about enactments: we imagine them to be the states of affairs, events, or actions they serve to enact. But that kind of imagining-

296 David Hills to-be is only pseudopropositional. Enactments don’t figure as such in propositional content of the fictions they make possible. As for the depiction/description difference, that falls into place as a special case of the deep difference between the as-if showing I call Acting Things Out and the as-if telling I call Spelling Things Out. And Phenomenal Unity is on hand to indicate why seeings of depictions can’t elicit called for imaginings in quite the way seeings of descriptions can. Depiction would place us as imaginers where Perspectives calls on us to be, thanks to the more general principles that govern role playing and enactment across the board. Step 2** leaves us with unfinished business. How can visual activities involving depictions enact anything if depictions aren’t role players in their own right? Can we avoid the idea that pictures are scripts without falling back on the idea that they are effigies? How can our interactions with pictures serve to enact anything unless both ourselves and the pictures with which we interact are role players? The answer is implicit in my remarks about Gombrich a section back. A depiction doesn’t stand in for the objects(s) it depicts: the most that happens is that our view of it (the depiction) stands in for a view of them (the depicted objects). Associated with any perspective on an object, with its characteristic opportunities and limitations for those who occupy it as perceivers, is a corresponding view of that object. In looking at X one explores one’s view of X. Indeed, our looking at it just is our exploration of our view of it In seeing it, we take in that view. Indeed, our seeing of it just is our taking in our view of it. The role players who collaborate in the determination of what’s fictional in a depiction-based make-believe game are the spectator (who plays herself, in the absence of special instructions to the contrary) and her view of the depiction (which takes the part of a suitably related view of the depicted objects). Lookings get enacted as explorations of views, seeings as takings-in of views.

11 On the make-believe account as we have it so far, what actually happens when I play a depiction-based game is that I visually explore the depiction’s marked surface with my eyes, visually experience the depiction’s marked surface in the manner afforded me by my visual explorations of it, and imagine accordingly. What fictionally happens is that I visually explore the subject in suitably related ways (from the picture’s own point of view) to suitably related experiential effect. The result is a spun-out yarn, a stretch of fictional perceptual autobiography—“The Subject of the Picture As I Found It”—whose twists and turns correspond episode for episode with my actual exploration of the picture’s marked surface. The relation of the content of the generated fiction to what we might think of as the built-in content of the picture itself—what the picture depicts and how it depicts what it depicts as being and looking—is rather delicate. The more thoroughly I explore the

Visions: on the nature of pictures 297 picture’s surface, the more I am called on to imagine having seen about the picture’s subject. And things I am to imagine seeing about the subject are things I am to imagine true of it. So a picture’s depictive content is a matter of what eventually comes out fictional about the subject in each and every playing of the picture’s game that eventually gives the picture’s surface a sufficiently thorough looking over. Walton’s conditions suffice to make the imaginative experience of the subject any given player of the depiction’s game is called on to have a visual experience as well. But different players are to have different imaginative experiences depending on how they go about perusing the picture, something the picture itself leaves entirely up to them, given the Richness condition. The make-believe account makes depiction a narrative art, in which the spun-out yarn centrally features visual explorations and experiences on the player’s part unfolding over time; and a collaborative art, in which the shape and interest of the spun-out fiction depends on how the spectator chooses to explore the painted surface. Is this a bug or a feature? In “On Pictorial Representation,” Wollheim restates the seeing-in account and offers cogent and powerful objections to various competing approaches to the nature of depiction, among them semiotic approaches (in which a picture is a special sort of conventional sign of its subject) and experienced resemblance approaches (in which a picture is experienced as resembling its subject in certain special visual or visually detectable respects). He contends that neither of these other approaches does justice to the sense in which painting is a visual art, the understanding and appreciation of which takes the form of some appropriate visual experience. He insists: (One) if a picture represents something, there will be an experience of it, call it the appropriate experience, that determines that it does so; (two) if a suitable spectator looks at the picture, he will, other things being equal, have this experience; and (three) this experience will be or, or include, a visual awareness of the thing represented. (Wollheim 1998: 219) We come to the question we left long ago: can there be any such thing as the way a picture is meant to be experienced? No doubt the question needs sharpening before it can be confidently answered yes or no. We inspect a representational painting bit by bit, experiencing corresponding portions of its depicted subject matter as we do so, lingering over one item, hurrying past another, returning repeatedly to items of special interest. Precisely what will catch and hold the eye of a given spectator is hard to anticipate and even harder to control. Old-fashioned art-appreciation folklore to the contrary, pictures are seldom if ever constructed so as to encourage us, let alone require us, to attend to particular details in a particular order or accord some details closer and more sustained attention than others. No doubt fleeting and forgettable glimpses of this or that detail contribute to more lasting and

298 David Hills substantial impressions of the detail’s nature, interest, and place in a picture’s overall organization, impressions which can acquire the standing character of memories without losing the characteristic organization of perceptions. Yet such standing impressions can themselves mutate repeatedly and spectacularly, even for highly knowledgeable spectators, as concerns and interests and viewing conditions change—a point beautifully confirmed by T.J. Clark’s heroically self-revealing journal of viewings and re-viewings over successive days and weeks and months of a pair of Poussins hanging on adjoining walls in the Getty (Clark, 2006). Contrary to what Wollheim’s description of his evolved way of looking might suggest, the impression that forms once things settle down needn’t be a permanently stable impression. Consider ekphrasis, the traditional rhetorical exercise in which one describes a depiction by describing the things depicted in it, as if one were face to face with those very things. Since he is to do his describing by means of words and their arrangement into sentences and sequences of sentences, the composer of an ekphrasis must choose which of a scene’s features to include in his description, the order in which to describe the ones he includes, and the amount of detail to accord each. Since he has to make these choices anyway, he had better make them in ways that reflect the interest and connection he accords the features he is called on to describe. The result of these choices is a description of the scene at a time that’s readily reheard as the narrative of an exploration of it over time, where the order in which features are described corresponds to the order in which they were explored and the detail with which they are described corresponds to the closeness with which they are attended to. In his volume on ekphrastic poetry, John Hollander discusses three different tours of Bruegel’s Hunters in the Snow (Figure 12.3), offered by three different ekphrastic poets: Walter de la Mere, John Berryman, and William Carlos Williams. The poems are startlingly different, but each does justice to the picture in its own way. Each makes its way through the depicted scene on its own trajectory, at its own pace, set by its own distinctive interests and priorities.8 The hobby-horse account should feel surprising, even after one decides to understand depictive content in make-believe terms, for in the simplest cases of narration-generated and enactment-generated fiction, what a physically passive audience member is called on to imagine at each stage seems settled once and for all, without her participation, by the pertinent text or performance and principles of generation governing texts or performances of that general kind. If the make-believe account is correct, this is never the case with depiction-generated fiction. With depiction, the spectator is always a character in a fiction that concerns the depicted subject only because it concerns the spectator’s visual experience of that subject. And the spectator is always an actor in and co-author of that fiction, helping to generate that fiction by inspecting the picture in the particular contingent way he does. Why think of things this way? Doesn’t doing so make a picture too much dolls and other effigies, too little like a text? Doesn’t it carry

Visions: on the nature of pictures


Figure 12.3 Pieter Bruegel the Elder, The Hunters in the Snow 1565 (Kunsthistorisches Museum, Vienna, oil on panel, 117cm × 162cm).

consumer sovereignty in the visual arts entirely too far? This natural question has a persuasive answer, I think—I owe it to conversations with Patrick Maynard. Consider the alternative. How else might we try to understand depiction in make-believe terms? In light of our earlier reflections, we might be inclined to hold out for an account that builds a particular standing fictional content into the picture itself, a content in which the subject figures but the spectator doesn’t. How might such an account go? It would need to say: A picture depicting subject S from viewpoint V calls on us to imagine that—is such as to render it fictional that—S would look thus and so, or would reveal itself to be configured thus and so, were it to be viewed from V. Further fictional truths about the subject may be generated from this core by Carrying Over, Working Out, etc. The thus and so signals some sort of Quinean deferred ostension: we indicate or point to a pertinent way of being or looking by (first) indicating or pointing to the picture’s marked surface. But how does the marked surface serve to specify a particular way of being or looking? There are only two salient natural possibilities. Perhaps the surface specifies it by exhibiting it, nearly enough, in which case we are committed to some sort of

300 David Hills resemblance-based or experienced resemblance-based account of depiction. Otherwise the surface specifies a way of being or looking by serving as some sort of conventional notation for it, in the manner of semiotic accounts, in which case the question of how we became privy to the relevant notational conventions will be pressing. In neither case will our general capacity to play games of make-believe on a pickup basis play any essential role in our account of how pictures in particular get understood. And we’ll fail to do justice to how pictorial understanding is pursued on the occasions when it is eventually achieved. It is by letting ourselves play with them as best we can (under culturally favorable circumstances) that we understand dolls as dolls (when we can manage this at all). Similarly, it is by letting our eyes play with them and over them as best we can (under culturally favorable circumstances) that we understand pictures as pictures (when we can manage this at all). Our capacity to understand pictures feeds on and feeds back into our more general capacity to play games of make-believe on a pickup basis. The hobby-horse account can accommodate this fact; its supposed alternative can’t. Pictures just are more like dolls than they seem.

12 Two further issues call for a final adjustment. First, the account copes resourcefully with static depictions of things that are thereby depicted as moving or changing, until the point is reached where distinct successive stages in our actual perusal of a picture serve to enact distinct successive episodes in our fictional perusal of its subject. The world seldom sits still. At any rate, it seldom sits nearly as still as pictures of it typically do. It seems to follow that once the aforesaid point is reached, I’ll be called on to imagine that the corner of the world I’m fictionally exploring remains eerily unchanged throughout my fictionally prolonged fictional explorations—as if it had been unnaturally frozen or slowed down. Call this the Grecian Urn Problem. Second, the account says that when I use a picture as it is designed to be used I observe it; having observed it, I am to imagine about the observing by me of it that it is an observing by me of the subject; hence I am to imagine there is an observing by me of the subject; hence I am to imagine that I see the subject face to face here and now. But aren’t there cases where it is fictional that I don’t see a depicted scene from the relevant point of view— cases where it is fictional that nobody at all sees it or where somebody else sees it in my place? Indeed, aren’t the cases where I do imagine myself face to face with the picture’s subject rare and special in all sorts of ways? One final adjustment, then: I really am called on to imagine that I myself experience the subject of the picture in a manner that employs and adopts the picture’s own viewpoint on its subject. But not all viewpoint-involving experiencing is seeing in the strict literal Grice–Strawson sense of “see” that requires me to occupy the point of view from which I see, requires the content of my

Visions: on the nature of pictures


seeing to be caused and constrained in familiar ways by the nature of what I see and the viewpoint from which I see it, and so forth. The bit of fictional inner autobiography my inspection of a picture calls on me to conjure up might instead be a fictional bout of recollection, involving fictional experiential memories of the subject from points of view I don’t (fictionally) occupy now and perhaps (fictionally) never did occupy. It might be a fictional bout of visualization, fictionally figuring in a fictional daydream of mine. It might be a fictional effort to envision what some unheard and merely fictional voice is telling me about events I fictionally never witnessed firsthand. And there are further possibilities. In short, the rules of a depiction-centered game of make-believe might count my actual seeings (of the picture) as enactments of fictional visual experiences (of the subject) that aren’t seeings in the strict literal Grice–Strawson sense. There are pictures where part of what’s depicted is to be understood as a vision on the part of someone, often a saint, depicted alongside it somewhere. Such visions were a Mannerist specialty. I’m suggesting that the whole depicted content of many and perhaps most pictures is best understood as some sort of vision on the part of the spectator herself. I am to (introspectively) experience my own seeings of the picture so as to thereby imagine myself visually experiencing the picture’s subject from the picture’s own viewpoint in a manner appropriate to face-to-face seeing, a manner appropriate to recollection, a manner appropriate to daydream, a manner appropriate to following another’s words—or what you will. But that’s an argument for another day.

Notes 1 For the objections see Wollheim (1998); for replies see Walton (2008); Hills (2002). 2 For a searching and resourceful application of this conception to puzzles in film theory see Wilson (2011). 3 For pertinent considerations see Walton (1990: 29, 33). 4 See Schachter (1996: 18–22). 5 For a nonstandard account of the syntax of “imagine seeing” that accords well with these views of Wollheim’s see Vendler (1979). 6 Wollheim (1998: 221). 7 For parallel formulations see Walton (2008: 136–8); Maynard (2005: 88–92). 8 Hollander (1995: 243–8).

References Clark, T.J. (2006) The Sight of Death: An Experiment in Art Writing. Yale University Press, New Haven and London. Frye, N. (1964) The Educated Imagination. Indiana University Press, Bloomington and London. Gombrich, E.H. (1960) Art and Illusion: A Study in the Psychology of Pictorial Representation. Princeton University Press, Princeton.

302 David Hills Harris, P.L. (2000) The Work of the Imagination. Blackwell Publishing, Malden and Oxford. Hills, D.J. (2002) “Review of Rob van Gerwen, ed., Richard Wollheim on the Art of Painting: Art as Representation and Expression”. Notre Dame Philosophical Reviews, August 9: https://ndpr.nd.edu/news/23141-richard-wollheim-on-theart-of-painting-art-as-representation-and-expression/ Hollander, J. (1995) The Gazer’s Spirit: Poems Speaking to Silent Works of Art. University Of Chicago Press, Chicago and London. Malcolm, N. (1965) Ludwig Wittgenstein: A Memoir. Oxford University Press, Oxford and New York. Mayard, P. (2005) Drawing Distinctions: The Varieties of Graphic Expression. Cornell University Press, Ithaca and London. Nutall, A.D. (2007) Two Concepts of Allegory. Yale University Press, New Haven and London. Podro, M. (1987) “Depiction and the Golden Calf”. In Philosophy and the Visual Arts: Seeing and Abstracting (ed., Harrison, A.), Dordrecht and Boston: Reidel Publishing Company, pp. 3–22. Podro, M. (2001) “The Artistry of Depiction”. In Richard Wollheim on the Art of Painting: Art as Representation and Expression (ed., van Gerwen, R.), Cambridge University Press, Cambridge and New York, pp. 112–34. Podro, M. (2004) “On Richard Wollheim”. British Journal of Aesthetics, 44, 213–55. Ryle, G. (1949) The Concept of Mind. Hutchinson’s University Library, London and New York. Schachter, D.L. (1996) Searching for Memory: The Brain, the Mind, and the Past. BasicBooks, New York. Vendler, Z. (1979) “Vicarious Experience”. Revue de Métaphysique et de Morale, 84, 162–73. Walton, K.L. (1990) Mimesis as Make-Believe: On the Foundations of the Representational Arts. Harvard University Press, Cambridge, Massachusetts and London. Walton, K.L. (2008) “Seeing In and Seeing Fictionally”. In Marvelous Images: On Values and the Arts, Oxford University Press, Oxford and New York, pp. 133–42. Williams, B. (1973) “Imagination and the Self”. In Problems of the Self: Philosophical Papers 1956–1972, Cambridge University Press, Cambridge, pp. 26–45. Wilson, G.M. (2011) Seeing Fictions in Film: The Epistemology of Movies. Oxford University Press, Oxford and New York. Wollheim, R. (1974a) “Reflections on Art and Illusion”. In On Art and the Mind, Harvard University Press, Cambridge, pp. 261–89. Wollheim, R. (1974b) “Imagination and Identification”. In On Art and the Mind (ed., Wollheim, R.), Harvard University Press, Cambridge, Massachusetts, pp. 54–83. Wollheim, R. (1987) Painting as an Art. Princeton University Press, Princeton. Wollheim, R. (1998) “On Pictorial Representation”. Journal of Aesthetics and Art Criticism, 56, 217–33. Wollheim, R. (2003) “What Makes Representational Painting Truly Visual?” Aristotelian Society Supplementary Volume, 77, 131–47.


Acting Things Out see Enactment Adoration of the Shepherds, The 126–9 Adventures of Huckleberry Finn, The 280–2 An Artist Drawing a Seated Man 220 Anna Karenina (Tolstoy) 249 Art and Illusion 169 ‘Art Interpretation’ (Richard Wollheim Memorial Lecture) 242 Art and Its Objects xiv, 38, 49, 92, 171, 172–3, 174–7, 261 aspect-blindness 81, 194 ‘aspect change’ 83–4, 93, 101–6 aspect-perception: appreciating as 96–7; description 94; historical context 78–9; indeterminancy of the mental 80–1; introduction 77–8; Kohler, Wittgenstein and objects of perception 82–5; paradox 85–7; reductionism 92–4; secondary meaning 81–2; ubiquity thesis – exegetical perspective 89–90; ubiquity thesis – substantive perspective 90–2; univocality of perception 94–6; Wittgenstein – preoccupation with aspects 79–80; Wittgenstein – solution to puzzle 87–9 Aspect View: perceptual experience 228–30; pictorial experience 205; seeing-in 233–4 Atget, Eugene 155–6 Auerbach, Frank 143–4, 145–7, 149 Bacchus and Ariadne 146–7 Bacon, Francis 149 Bang! term 283–4 Bathers, The 152–4 Baxandall, Michael 141

Beethoven, Ludwig van 46 Bellini, Giovanni 164, 186 Berryman, John 298 Beschreibung der unmittelbaren Enfahrung, des seherlebnisses 41–2 Bildgegenstand (picture-object) 174 Black Circle 209–10 Botticelli, Sandro 198 Braque, Georges 121–2, 136 Bruegel, Pieter (the Elder) 299 Budd, Malcolm 175–6, 177, 201, 215, 230 Carroll, Noël 242 Carrying Over 278, 286–7 Cavell, Stanley 151, 154, 156 Cézanne, Paul 121, 136, 152–4 ‘change of aspect’ (Aspektwechsel) 44–5, 70, 79, 86, 93, 104, 107, 146, 170 Chardin, Jean-Baptiste-Simeon 154–6 Chateau Noir 121 Chicago 10 271–2 Clarke, Kenneth 101 Clark, Herb 281 Clark, T. J. 298 Collier, Evert 194 Collingwood, R. G. xiv, 160–1, 163, 164 Conan Doyle, Arthur 21 Concept of Mind, The 279 continuous aspect-perception 50, 69–70, 72–3, 89, 170, 194 Count Johann 219 Critique of Judgement, The 72 Critique of Pure Reason 65 Dalmatian Dog 228 de Kooning, Willem 199

304 Index de la Mere, Walter 298 depiction: aspect-perception 77; description 207, 270; ekphrasis 298; Gombrich Ernst 269–70; imagination 294; imagined to be 290; makebelieve (game) 276–7, 292–3, 296–7; philosophy 208, 269–70; surface 271; visions 297; Walton 276–9; Wollheim, Richard 289 Drury, John 126–8 duck–rabbit figure: ambiguity 169, 221–2, 231; Clarke, Kenneth 101; communication 42–3; Gombrich, Ernst 102; perception 46; perceptual indeterminancy 70–1; picture puzzle 79, 87; seeing 25, 66, 108; ubiquity thesis 91; Wittgenstein, Ludwig 37, 40, 68, 79, 87, 93, 106, 142, 165; Wollheim, Richard 169, 171–3 Dürer, Albrecht 220 Ehrenstein illusion 232–3 ekphrasis 298 Enactment (Acting Things Out) 279, 283, 296 Equinox 132 Erkenntisse (Kant) 64, 69, 72 Eugenides, Jeffrey 257–8 Experienced Resemblance View 212, 216, 218, 220, 221–3, 224, 227, 232 Fact and Fable in Psychology 169 Factory Horta de Ebro 121–2 Flag 135 Floyd, Juliet 71 foundational account (Wollheim) 188, 200 Freud, Lucian 149, 164 Freud, Sigmund 242–3, 249, 257–8, 260–1, 262, 281 Gardner, Sebastian 258 Gestalt: faces 68; perception 67; psychology 66, 69, 79, 82; theories 78 Gombrich, Ernst: duck–rabbit figure 102; ‘illusion’ 107, 232; perception 169–71, 187; seeing as 101–3; visions 269–71, 274–5, 286, 290, 296; Wittgenstein, Ludwig 173, 177 Gorgonzola cheese 29 Grecian Urn Problem 300 Grice, Paul 193, 196, 273, 275–6, 282, 300–1

Hamlet 286–7 Harnett, William 138 Harper’s Weekly 169 Harris’s game 278, 279, 285 ‘Heavenly Sight and the Nature of Seeing-In’ 197 Hemingway, Ernest 281–2, 293 Hinton, J. M. 29 Hobbema, Meindert 291, 294 Hoffman, Hans 131–2, 136, 138 Hopkins, Robert 183, 188–91, 199, 212, 215 Hunters in the Snow, The 298–9 Icarus 186–7 imagination: xiv–xv; Kant 88–9; seeing-in 275–6; Wittgenstein 107–8; Wollheim’s concept of 63, 175–6, 177, 179, 243, 249–53 Imagination View 212, 223, 224, 225–7 intention xv, 7, 13, 83, 84, 86, 88, 89, 94, 117–18, 133, 137–40, 142–3, 161–3, 172, 241, 243, 249–59 Ishiguro, Hide 179 James, Ronald C. 228 Johns, Jasper 135–6, 207 Kanizsa’s triangle 232 Kant, Immanuel: Critique of Pure Reason 65; Erkenntisse 64, 69, 72; imagination 88; intuitions without concepts 196; ‘nature’ 64; thoughts without intuitions 196; transcendental psychology 89; unity of experience 67 Kant, Immanuel and Wollheim: aspects and schemata 191–5; conclusions 200–1; further ideas 195–7; Hopkins, Robert 183, 188–91; integration and inflection 197–200; introduction 183–4; preliminaries 184–5 ‘Killers, The’ 281 Klein, Melanie 242, 258 Koffka, Kurt 66 Köhler, Wolfgang 78–9, 82–5, 103, 113 Kulvicki, John 197–8 La Nappe (The Tablecloth) 155 Large Bathers, see Bathers, The LBW (cricket) 27 Lectures on Philosophical Psychology 103

Index Leonardo: seeing scenes in cracked walls 195; see also Leonardo’s challenge Leonardo’s challenge: introduction 117–18; perception/projection dichotomy 118–19; Rembrandt and Wittgenstein 126–31; Wittgenstein and ‘see’ term 120–6; Wollheim and aesthetic perception 117, 131–6, 149–56; Wollheim and representational seeing 136–49 Les Demoiselles d’Avignon 200 Letter Rack 194 Levinson, Jerrold 191, 242–3, 245–50 Liberty Leading the People 136 localisation requirement (Wollheim) 187 ‘look’ (verb) 5 Lovibond, Sabina 254 Madonna of the Meadow 186 make-believe: depiction (game) 276–7, 292–3, 296–7; see also Imagination View Malevich, Kasamir 209–10, 213 Manet, Édouard: single figure paintings 251 meaning-blindness 81 ‘mental image’ (inneres bild) 176–7, 179 Merleau-Ponty, M. 67–70, 72 Mill with the Great Red Roof 291 Mimesis as Make Believe 291 Monet, Claude 156 Moore, Henry 199 Napoleon 191, 273, 281 Necker cube: A-cube 5; expert authority 25; Gombrich, Ernst 102; MerleauPonty 69; Tractatus 4; visual ambiguity 15–18, 111, 221–2; Wittgenstein, Ludwig 79 ‘networks of factive meaning’ 63 Nevelson, Louise 198 Newman, Arnold 123–4, 134 Night Café 189 Nuttall, A. D. 279–80, 282 Olivier, Laurence 286–7 On Drawing an Object 172, 177 On the Emotions 242, 256 ‘On Pictorial Representation’ 102, 257, 297 outline shape, occlusion shape 189, 191, 192, 217–22, 227–8


Painting as an Art 102, 136, 154–5, 177–9, 242, 254, 256 perception: aspects/concepts 61–5; aspects as perceived internal relations 65–8; conclusions 72; grammar and phenomenology of Wittgensteinian aspects 51–7; introduction 49–51; perceptual indeterminancy 68–72; term 118–19; Wollheim on seeing-as 57–61 Phenomenology of Perception 69–70, 72 Philosophical Investigations 38, 41, 51, 106, 117, 120, 150, 161, 169, 176, 289 Philosophical Psychology (lectures) 112–13 Philosophy of Wittgenstein, The 111 Picasso, Pablo 121–2, 123–4, 134, 136 pictorial experience 206, 207–8, 211–12 ‘picture-objects’ 290 picture puzzles (Kippfiguren) 78 Plato 279 Podro, Michael 177, 198–9, 273, 274 Primavera 46 Principles of Art, The 163 Prinz, Jesse 254–5 Proceedings of the Aristolean Society 189 ‘projection’ term 119 Proust, Marcel 154–6 Pythagoras 190 Question of Content 205–6, 211, 221, 227, 231 Question of Depth 205–6, 211, 222, 227, 231 Raphael 190 Raven trickster tale (Haida of the Pacific Northwest) 282 reading in 118, 120, 130, 131, 137, 138 Reconstrual 293–5 reductionism 77, 92–4 Remarks on the Philosophy of Psychology 22–3 Rembrandt, van Rijn: Leonardo’s challenge 126–31; religious thinker 136; Sylvius, Jan Cornelisz 228–9; Wollheim, Richard 141, 251 representation (Darstellung) concept 176–7 Robinson, Jenefer 262

306 Index room in a view, the: two forms of authority 18–22; collapse 11–13; disjunctivism 29–32; Frege’s challenge 13–15; generality 8–9; introduction 3; invisibility 9–10; structurability 10–11; things not in view 22–9; Tractatus speedy collapse 4–8, 11–12; visual ambiguity 15–18 Rorty, Amélie O. 254–5 Rubin vase figure 46 Ryle, Gilbert 279, 280, 282 St Francis Giving his Cloak to a Poor Knight 141 Savile, Anthony 249 School of Athens 190 seeing-as 4, 13, 15–16, 22–3, 28, 38–9, 43, 49–50, 54–5, 57–61, 68, 79, 89–90, 92–3, 94, 101–13, 125–6, 170–2, 172–3, 271 ‘seeing aspects’: seeing-in 3; telling stories about it 37–47 seeing-as/seeing-in xiv, 38–9, 40–1, 47, 92, 135, 183, 184–9, 193, 195–7, 200–1, 271–2 seeing-in: aesthetics 242–53; Gombrich, Ernst 271; meaning-content 136; Wittgenstein, Ludwig 3, 183; Wollheim, Richard 38, 117, 131, 134, 136–9, 154–5, 161–3, 166–7, 169–79, 242–8, 256, 289, 297 seeing-in as aspect perception: Aspect View 233–4; depiction philosophy (Wollheim) 208–11; Experienced Resemblance View 212, 216, 218, 220, 221–3, 224, 227, 232; Imagination View 212, 223, 224, 225–7; introduction 205–6; Kanizsa’s triangle 232; outline shapes 217; pictorial experience 206, 207–8, 211–12; Question of Content 205–6, 211, 221, 227, 231; Question of Depth 205–6, 211, 222, 227, 231; surface shapes 217; visual awareness 206–7 ‘see’ (verb) 5 Sein und Zeit 78 Self-Portrait (Picasso) 123 Shakespeare, William 140 Siskind, Aaron 271–2, 276 Sketch from Titian’s ‘Bacchus and Ariadne’ 146–7 Spellings Out 278–9, 281, 296 Sterba, R. 259–60

Strawson, P. F. 89–91, 273, 275–6, 282, 300–1 Study after Titian I 143 Study after Titian II 143–4 Study Table, The 138 ‘surface’ as expression of intention: figure and ground 67, 70, 87, 112, 151, 274; introduction 160–1; ‘psycho-methodological account’ 163–6; seeing-in 161–3, 166–7 surface shapes 217 Sylvius, Jan Cornelisz 228 Target Paintings 207 Tarquin and Lucretia 145–6 Tempest, The 140 Thread of Life, The 242 Three Views of Rouen Cathedral 7 Titian 143–4, 146–7 Tolstoy, Leo 249 Tractatus 4–8, 11–12, 110–11 Travis, Charles 62, 63 trompe de l’oeil 194–5, 232 ‘twofold thesis’, twofoldness 47, 101, 102, 134, 137–8, 145, 174, 188, 209, 212, 215, 230, 244–6, 241, 242, 248, 271–2, 274–5, 293 ubiquity thesis 77, 89–90, 90–2 van Gogh, Vincent 189, 213–14 Venus 198 Viaduct at L’Estaque 121 visions (Wollheim and Walton on nature of pictures): enactment 283–5; Gombrich, Ernst 269–71, 274–5, 286, 290, 296; Haida of the Pacific Northwest 282; imagination 275–6; introduction 268–9; make-believe 276–7, 296–7; Reconstrual 293–5; Siskind, Aaron 271–2 Vorstellungen 14–15, 22, 65 Walton, Kendall xv, 223–4, 225–6, 268, 276–7, 279, 280–1, 285–7, 290–1, 292–4, 297 Wheat Field with Cypresses 213–14 White Tablecloth, The 155 Wildschwein 5 Williams, William Carlos 298 Window Grid 219 Wittgenstein, Ludwig: aesthetic problems 37, 44; aspect-seeing 37; aspects as perceived internal relations 65–8, 72; Bildgegenstand

Index (picture-object) 174; duck–rabbit figure 37, 40, 68, 79, 87, 93, 106, 142, 165; Erfahrung/Erlebnis term 42; ‘experience’ term 43; faces 68; Gombrich, Ernst 173, 177; ‘leave judgement external to perception’ 65; linguistic ‘intuitions’ 38; locutions 47; ‘mental image’ (inneres bild) 179; Necker cube 79; noticing an aspect 43; paradoxical elements 44–5; perceptual indeterminancy 68–71; Philosophical Investigations 38, 41, 51, 106, 117, 120, 150, 161, 169, 176, 289; Rembrandt 126–31; ‘see-as’ 28; seeing-as 38, 68, 101–13, 125–6; ‘seeing aspects’ 3; seeingas/seeing-in 183; ‘see’ term 120–6; Tractatus 4–8, 11–12, 110–11; triangle 150–1; see also aspect-perception; Leonardo’s challenge Wittgenstein, Ludwig and seeing as: aspect changes and paradox of sameness and change 103–4; perceived picture (internal relations/ organisation) 110–13; perceiving subject (inner picture, primitive reactions, attitude, will and interpretation) 104–10; Wollheim, Gombrich and Wittgenstein 101–3 Wollheim, Richard: aesthetic perception 117, 131–6, 149–6; aesthetic problems 37; Art and Its Objects 37, 92, 171, 172–3, 174–7, 185, 261; aspects as perceived internal relations 65; depiction 77; depiction philosophy 208; depth awareness 210–11; duck–rabbit figure 169, 171–3; Foundational Account 188, 200; localisation requirement 187;


On Drawing an Object 172, 177; On the Emotions 242, 256; ‘On Pictorial Representation’ 102, 297; Painting as an Art 136, 242, 254, 256; perception 41; Rembrandt, van Rijn 141, 251; representational seeing 136–49; resemblance 165; seeing-as 57–61, 102; seeing aspects 69; seeing-in/seeing-as 38, 47; seeing representations 39; ‘straightforward’ term 43; ‘The Art Lesson’ 161; ‘twofoldness’ 248; twofoldness 272, 291; ‘twofold thesis’ 134; ‘visualize/ imagine seeing’ 283; Wollheim, Gombrich and Wittgenstein 101–2; see also Kant, Immanuel and Wollheim; Leonardo’s challenge; seeing-in as aspect perception; ‘surface’ as expression of intention; visions Wollheim, Richard – ekphrastic aesthetics (emotion and its relation to art): description 253–8; introduction 241–2; music cannot convey emotion? 261–3; psychoanalysis 258–61; seeing-in and aesthetics 242–53 Wollheim, Richard – seeing-in: art as a form of life 161–3; Art and Its Objects 171, 172–3, 174–7; depiction 289; development 172–3; introduction 169–70; On Drawing an Object 172, 177; Painting as an Art 136, 177–9; perception 167; seeing-as 170–2; ‘twofold thesis’ 134, 174; Wittgenstein, Ludwig 173 Woman series (de Kooning) 199 Young, James 241 young lady–old witch figure 83