Online research in philosophy

Certain simple thoughts about pictures suggest that the contents of pictures are closely bound to vision. But how far can the striking features of depiction be accounted for merely in terms of the especially visual contents which belong to pictures, without considering, for example, any issues concerning the nature of the visual experiences with which pictures provide us? This article addresses that question by providing an account of the distinctively visual contents belonging to pictures, and by using that account to explain many notable general facts about depiction. Some implications of the resulting framework for the main stream of current theorizing about pictorial representation are also discussed.

Congenitally blind people can make and understand ‘tactile pictures’ – representations form of raised ridges on flat surfaces. If made visible, these representations can serve as pictures for the sighted. Does it follow that we should take at face value the idea that they are pictures made for touch? I explore this question, and the related issue of the aesthetics of ‘tactile pictures’ by considering the role in both depiction and pictorial aesthetics of experience, and by asking how far the experience of those engaging with representations through touch can approximate to that of those engaging with them through sight.

Depiction is the form of representation distinctive of figurative paintings, drawings, and photographs. Accounts of depiction attempt to specify the relation something must bear to an object in order to depict it. Resemblance accounts hold that the notion of resemblance is necessary to the specification of this relation. Several difficulties with such analyses have led many philosophers to reject the possibility of an adequate resemblance account of depiction. This essay outlines these difficulties and argues that current resemblance accounts succumb to them. It then develops an alternative resemblance account, drawing on Grice's account of nonnatural meaning and its role in determining sentence meaning to argue that something depicts an object if it bears intention-based resemblances to the object that jointly capture its overall appearance. In addition to solving the metaphysical problem of what it is for something to depict an object, this account also sheds significant light on the epistemological issue of how we are able to work out that something depicts an object. This essay argues that our ability to work out that something depicts an object results from both our more general ability to identify intentions from the products of communicative behavior and our knowledge of stylistic conventions. This account avoids the difficulties that face rival attempts to analyze depiction in terms of resemblance. It also clarifies and explains the features that distinguish depictive from nondepictive representation.

1. A new wave of subjectivism in the theory of pictorial art began around forty years ago; and since then it has gathered pace in tandem with changing fashions in the philosophy of mind. The initial impetus was provided by the publication of Ernst Gombrich’s 1956 Mellon Lectures, Art and Illusion.1 In this book, and in many subsequent articles and lectures which elaborate its theme, Gombrich argues that the development of Western art – essentially the art of ancient Greece and the art of Western Europe from Giotto to Cezanne – consists in a series of discoveries about the nature of visual perception, and the means by which the effect of visible objects on our senses can be simulated. ‘What may make a painting like a distant view through a window’ he writes, ‘is not the fact that the two can be as indistinguishable as is a facsimile from the original: it is the similarity between the mental activities both can arouse.’ And in another place: ‘The goal which the artist seeks with such self-critical persistence is … a psychological effect.’2 These remarks are concerned with a specific artistic tradition, and with specific pictorial devices, such as foreshortening and shading, which I shall not write about here. But Gombrich’s work launched a search for a general theory of depiction based on the same approach. The search intensified in the wake of Nelson Goodman’s Languages of Art, which persuaded most of its readers that the resemblance theory of depiction could not be made to work, but did not provide a plausible alternative to it. Today there is broad agreement among philosophers that the nature of pictorial art cannot be explained by analysing the relationship between the marks on the surface of a picture and the kinds of objects that they represent. The consensus is that it can only be explained by defining the psychological effect that these marks produce.3 Today, the two most influential theories of depiction that are guided by this general idea are Richard Wollheim’s and Christopher Peacocke’s..

I offer two, complementary, accounts of the visual nature of representational picturing. One, in terms of six features of depiction, sets an explanatory task. The other, in terms of the experience to which depiction gives rise, promises to meet that need. Elsewhere I have offered an account of this experience that allows this promise to be fulfilled. I sketch that view, and defend it against Wollheim's claim that it cannot meet certain demands on a satisfactory account. I then turn to Wollheim's own view, arguing that it suffers from crucial obscurities. These prevent it from meeting the explanatory commitments I describe, and are only exacerbated by the demands Wollheim himself imposes.

Philosophical discussions of depiction sometimes suffer from a lack of differentiation between several questions concerning the nature of pictorial representation. To provide a suitable framework I distinguish six such questions and several levels on which one might want to proceed in order to answer some of them. With this background, I reconstruct Goodman's and Elgin's answer to the specific question: What distinguishes the pictorial from the verbal or linguistic? I try to reveal some major motivations behind their system-oriented approach and to indicate some reasons why a strategy of this kind is to a certain extent mandatory to grasp the nature of the pictorial. The system-relative and functional character of depiction has to be captured by every adequate theory.

I propose a number of criteria for the adequacy of an account of pictorial realism. Such an account must: explain the epistemic significance of realistic pictures; explain why accuracy and detail are salient to realism; be consistent with an accurate account of depiction; and explain the features of pictorial realism. I identify six features of pictorial realism. I then propose an account of realism as a measure of the information pictures provide about how their objects would look, were one to see them. This account meets the criteria I have identified and is superior to alternative accounts of realism.

Pictorial representation is a subject of interest to both cognitive science and aesthetics. Standard theories of depiction often draw on vision science, and vision science must give an account of picture perception. I offer a critical overview of standard theories of depiction and argue that none of them is adequate. I then describe ways in which new theories of perception blend elements of representationalism with an emphasis on attention and motor control. Such theories, in effect, limit the reliance on mental representation in perceptual tasks. This work provides the basis for a theory of depiction in which pictorial representation is explained in terms of both mental representations and perceptual strategies. I argue that, in the case, the mental representations are most plausibly individuated by the functional and conceptual roles, rather than by causal links to the external world.

This book presents an original theory of the nature of pictorial representation. The most influential recent theory of depiction, put forward by Nelson Goodman, holds that the relation between depictions and what they represent is entirely conventional. Flint Schier argues to the contrary that depiction involves resemblance to the things depicted, providing a sophisticated defence of our basic intuitions on the subject. Canvassing an attractive theory of 'generativity' rather than resemblance, Dr Schier provides a detailed account of depiction, showing how it illuminates and resolves many of the enigmas of pictorial representation while remaining true to our basic intuitions on the subject.

By defining both depictive and linguistic representation as kinds of symbol system, Nelson Goodman attempts to undermine the platitude that, whereas linguistic representation is mediated by convention, depiction is mediated by resemblance. I argue that Goodman is right to draw a strong analogy between the two kinds of representation, but wrong to draw the counterintuitive conclusion that depiction is not mediated by resemblance.