Monroe C. Beardsley, Aesthetics: Problems in the Philosophy of Criticism, 1st ed., 1958; 2d ed., 1981.

5. Representation

The discussion of pictorial representation in the first edition (Chapter 6) was a first attempt at answering what seemed to me an important and puzzling question that had hardly been recognized, much less seriously discussed: what is it for a picture to represent something? -- or what are the truth-conditions of such a statement as "Picasso's Girl Before a Mirror represents a girl"? Although not carried through in rigorous detail, the discussion led to a number of distinctions that are still necessary and useful. But it became evident that the proposed analysis on the most important and difficult concept, that of depiction, was in need of considerable emendation, to take adequate account of a number or points: for example, that we want to distinguish between depicting a picture and merely copying it; that two prints drawn successively from the same etched plate do not depict each other, no matter how similar; that objects and state of affairs that never existed can be depicted, even quite realistically; that caricatures and cubist fantasies can represent; and so on. An ingenious suggestion made by Paul Ziff a few years afterward, but independently ("On What a Painting Represents"), made essential use of the concept of "visual aspects," found both in a picture and in the object it depicts, so there can be a match; and the difficulties in applying this suggestion (it is hard to think of Picasso's painting as containing anything that could be called a visual aspect of a girl before a mirror!) pointed up still more clearly the deficiencies of my original proposal. About the same time (1960), E. H. Gombrich's Art and Illusion burst upon the scene and began to make us understand far better than before the element of convention in pictorial representation -- the fact that we have to learn how to read pictures, to master the rules of the system of representation employed (that, for example, in a line drawing the enclosed area is to be taken as depicting a solid, that cross-hatching is shadow, and so forth). (Cf. David Novitz, "Conventions and the Growth of Pictorial Style.")

The problems about representation came to a head (in 1968) with the publication of Nelson Goodman's Languages of Art, which sharply challenged (in Chapter 1 and in Section 1 of Chapter 6) the prevailing, indeed I should say universal, assumption that pictorial representation necessarily involves a significant similarity between the picture and what it depicts, even between some of the shapes (gestalts) in Girl Before a Mirror and some shapes that might characteristically be found in a young woman before a mirror. Goodman's arguments to show that similarity is not necessary for representation opened up a great deal of fresh discussion of the problems.

Goodman's radical claim is that everything we need to say about pictorial representation (or representation in music, etc.) can be said in terms of two fundamental concepts. There is, first, representation tout court, which he holds to be simply denotation, in the usual sense: a picture of a car denotes cars in general, just as the term "car" (or more precisely the predicate "is a car") denotes them. The difference is that the picture and the verbal predicate belong to different types of symbol system (in his rigorously worked out scheme, pictorial symbol systems -- unlike languages -- are both semantically and syntactically dense). But second, we can also classify a picture along with other pictures in terms of its pictorial properties -- that is, in terms of predicates that denote it -- distinguishing, say, car-pictures from perpetual-motion-machine-pictures. Of course perpetual-motion-machine-pictures do not denote anything, since there are no such machines; and even a car-picture need not denote actual cars (say if it is a thirteen-wheel-car-picture) -- nor need a picture be a car-picture to denote cars (for example, if it is a tiny speck toward the top of a photograph or drawing of an enormous parking lot in which the distant cars appear only as specks).

Even Goodman's critics have often admired the skill with which these two basic concepts, that of denotation in a dense symbol system and that of being a picture, are developed and applied, and the proposal has often been found convincing. I am still troubled by certain basic features of it, especially by the treatment of representation as denotation (see "Languages of Art and Art Criticism"). Although Goodman has said that he does not object to the concept of depiction as a form of indefinite reference -- the photograph depicts a car -- he considers representation of existing things (as in paintings and sculptures) to have the same basic semantic character as verbal labels, even though he acknowledges that there are striking differences between them (for example, that predicates can be logically conjoined and disjoined, while pictures cannot -- but see Robert Howell, "Ordinary Pictures"). I believe that depicting is an important semantic function of many visual works of art, and that it involves selective similarity: essentially, my notion is that the depicting surface has some visual features that match distinguishing, characteristic, telling features of objects of a certain kind -- a car-shape, however distorted, for example (although in other pictures, in different stylistic systems, the selecting feature might be color or mass or location in picture space or pattern of light-and-dark, and so forth).

Two very noteworthy alternatives to this view have recently been offered. Kendall Walton (see "Pictures and Make-Believe" and later essays) has developed a subtle and impressive theory of "fictional truths," according to which, for example, the proposition "Dracula was not reflected in a mirror" is "fictionally true" in Bram Stoker's novel because it is made true in the world of the novel by the text but is not true of the real world (this is a rough paraphrase of a precise and complicated definiens). One way of being fictionally true is being "make-believedly true" (as contrasted with being "imaginarily true"): that is, being fictionally true in virtue of some fact other than someone's imagining, pretending, and so forth. Representational pictures are then said to be "props in games of make-believe": truths about objects, events, persons in a painting are fictionally true in virtue of the shapes, lines, colors on the canvas, which make it possible for us to perform the (make-believe) actions of seeing, noticing, recognizing, and so forth, (say) Socrates and his companions of the Phaedo in David's famous painting. And what a painting depicts in a particular society is determined by the socially-accepted rules specifying what propositions are made make-believedly true in the painting by the painting's visual properties. According to Walton, although resemblance to an object is not necessary for depicting it, since perceiving an object-depiction counts as a make-believe perception of that object, the game will generally be facilitated by resemblance.

Walton's account is worked out in detail, and takes care of a number of puzzles about representing. It also seems to me to arouse other puzzles: for example, how can it be make-believedly true"in David's painting that Socrates drank the hemlock (since it is historically true)? And I haven't shaken off my sense that perceiving Picasso's depiction of a girl before a mirror in no way involves a making-believe that I am perceiving a girl before a mirror. (For other comments, see Eddy Zemach, "Description and Depiction.") An alternative account developed with care and clarity by Goran Hermeren (see "Depiction" and Representation and Meaning in the Visual Arts, Chapter 2) builds on Wittgenstein's concept of "seeing as." Hermeren allows for a number of possible senses of "depict," but in the central (and recommended) one, a painting depicts an apple when its visual properties are so arranged that they are or can be seen as an apple. He and Walton both note that this proposal has affinities with Walton's, although Hermeren develops it in a different direction. Robert Howell's technical analysis of "The Logical Structure of Pictorial Representation," which makes use of possible-worlds semantics to explain statements about pictured objects and events, takes "seeing-as" to be an essential element in the concept of representation. Jenefer Robinson ("Some Remarks on Goodman's Language Theory of Pictures") argues for a "seeing-as" account, if the term is taken broadly to include "cognitive" as well as visual apprehension. This goes well beyond Hermeren's position, I believe, and perhaps makes the position weak enough to fit with any adequate account. What troubles me is that the term by itself explains so little. If we define depiction in terms of "is seen as," it becomes uselessly relativized; if in terms of "can be seen as," it is uselessly general; if we put the emphasis (where it should be) on "is to be seen as, given such-and-such a configuration of visual properties," then the real question becomes that of articulating the rules or conventions that enable just those properties to single out and refer to things of just that kind.


Beardsley, M. C.
"Languages of Art and Art Criticism" and reply by Nelson Goodman. Erkenntnis 12 (1978).
Black, Max.
"How Do Pictures Represent?" In Maurice Mandelbaum, ed., Art, Perception, and Reality. Baltimore: The Johns Hopkins University Press, 1970.
Gombrich, E. H.
Art and Illusion: A Study in the Psychology of Pictorial Representation. Princeton: Princeton University Press, 2d ed., 1961.
Goodman, Nelson.
Languages of Art. Indianapolis: Hackett, 2d ed., 1976.
Hermere'n, Goran.
"Depiction: Some Formal and Conceptual Problems." Journal of Aesthetics and Art Criticism 37 (1978).
Representation and Meaning in the Visual Arts: A Study in the Methodology of Iconography and Iconology. Lund: Berlingska, 1969.
Howell, Robert.
"The Logical Structure of Pictorial Representation." Theoria 40 (1974).
"Ordinary Pictures, Mental Representations, and Logical Forms." Syntheses (1976).
Novitz, David.
"Conventions and the Growth of Pictorial Style." British Journal of Aesthetics 16 (1976).
Robinson, Jenefer.
"Some Remarks on Goodman's Language Theory of Pictures." British Journal of Aesthetics 19 (1979).
"Two Theories of Representation." Erkenntnis 12 (1978).
Savile, Anthony.
"Nelson Goodman's 'Languages of Art': A Study." British Journal of Aesthetics 11 (1971).
Sparshott, Francis E.
"Goodman on Expression." The Monist 58 (1974).
Walton, Kendall L.
"Are Representations Symbols?" The Monist 58 (1974).
"Points of View in Narrative and Depictive Representation." Nous 10 (1976).
"Pictures and Make-Believe." Philosophical Review 82 (1973).
Zemach, Eddy M.
"Description and Depiction." Mind 84 (1975).