PRINT October 1979

Recent Esthetics and the Criticism of Art

PERHAPS THE ALMOST TOTAL LACK of contact between art critics and philosophers writing about art is justified. Estheticians characteristically treat historical issues—the relation of Kant’s esthetics to his epistemology, Hume on taste, Dewey on emotion—in purely historical terms; or they discuss issues that critics might find interesting in ways that critics will certainly think dull. For example, perhaps the most widely discussed recent work in esthetics is the institutional theory of art. Much traditional esthetics offers an answer to the question: “What properties are necessary to make something an artwork?” Thus, for Plato artworks are representations, and for Hegel art is a kind of expression. After Duchamp’s readymades and much art influenced by him, no such definitions of art are convincing. The institutional theory suggests that something is an artwork if people in the “artworld” call it an artwork. Thus described, the theory seems true but vacuous: art is what people who know call “art.” What would make this theory illuminating is an explanation of why art critics call certain things art. But to follow the philosopher who developed this theory in saying that Duchamp’s urinal called Fountain has “qualities similar to those works by Brancusi and Moore which many do not balk at saying they appreciate,”1 evades the question of why Duchamp’s anti-art gesture ended with his work’s becoming art. So, at best art critics are likely to find this theory irrelevant to their interests.

If critics are likely to find esthetics boring, estheticians tend to find critics’ approach to problems of theory uncritical. When it comes to the theory of art, critics adopt all too often the attitude of a bricoleur. Just as Harold Rosenberg assembled the notion of “action painting” from bits and pieces of then-popular existentialism, so now the fashionable French work of Lacan, Derrida and others influences criticism. What is of value here can be found, I believe, only by a person of enormous erudition, someone familiar with Freud, phenomenology, Kojeve on Hegel, and much more. As the philosopher Richard Wollheim has written, and as the work of the Art-Language group illustrates, “Philosophy has virtually nothing to offer those who would rifle it.”2 In their pursuit of the fashionable, critics seem a bit like people who, becoming interested in religion, turn to the exotic traditions of Zen Buddhism or Hinduism. Just as such would-be religious people might do better to see what is available to them in their own Western traditions, so art critics interested in theory might find profit in looking at some native philosophic traditions.

But is there anything for such critics to find? Within the English-speaking world, the dominant tradition is that of analytic philosophy. For these philosophers, the German tradition of Hegel, Schopenhauer and Nietzsche, together with phenomenology and more recent French philosophy, remains an exotic tradition. Despite the ingenuity and intelligence of some individuals drawing on that tradition,work in it has something of the status of the study of Plotinus: it is a subject for specialists, not a matter of general interest. This situation profoundly affects the status of esthetics. Whatever the differences between Hegel, Schopenhauer and the more recent continental philosophers, all give art great importance. When Nietzsche calls making art “the truly metaphysical activity of man,”3 he could be speaking for that entire tradition. But the focus of analytic philosophy on problems of language, and, often, its emphasis upon scientific knowledge, goes together with a lack of interest in art. One of the best-known articles of a couple of decades ago described the esthetics developed within this tradition as dreary, yet 10 years ago an analytic philosopher produced a treatise on esthetics which is of great importance.

Prior to that date, Nelson Goodman was already known as a philosopher of extraordinary brilliance. Since then, his Languages of Art (1968) has been widely written about by estheticians, although it seems to remain almost entirely unknown to nonphilosophers. The book is technical and closely argued, and it is not obvious why his arguments should interest nonestheticians. Goodman himself has said: “When I wrote Languages of Art, I had hopes that practitioners and theorists of literature, painting, music and other arts might find something relevant in it . . . This happens rarely, since few people are competent enough in both philosophy and one of the arts to make the connection.”4 That it takes time for such work to become accessible to art critics should not be surprising. It took a century before Kant’s very abstract account of art was translated by Wölfflin and Fry into terms useful for art historians and critics.5 Languages of Art, like Kant’s work, will probably not have a major effect on critics and historians until popularized versions of it are published.6 My more modest aim is to suggest why Goodman raises issues of interest to art critics. I will offer neither an explication nor a critique of his book. I will argue that Goodman offers both a philosophical justification for some methods employed by critics and, if he is correct, shows that some other widely used critical approaches are not acceptable.7

Goodman divides arts, or artworks, into two kinds: autographic and allographic.8 As the term suggests, autographic arts are those whose identity is defined by their history. Painting is autographic because only that one object made by Giorgione is his Tempest. Written literature is allographic because any text with the right words in the right order is a true “copy” of the book Lolita. The word “copy” can cause confusion here: my postcard of The Feast of the Gods is a copy of the unique artwork in the National Gallery; my copy of Lolita is an instance of that artwork, of which there can be an indefinite number of instances. Assuming Nabokov’s manuscript has been printed correctly, a literary scholar can use any copy of the book. But an art historian who could see only copies of the Bellini painting would never be sure if some subtle but significant feature of the original were not missing from those copies.

This distinction is not a distinction between arts where the artwork is one thing and those where it “is” many things. For engraving is autographic—all genuine prints are made from the same original plate—and there may be only one copy of an unpublished novel. In general, arts tend to move from being autographic to being allographic. We have the transition from oral to written poetry; from improvised to scored music; and, today, the development of a notation for dance. We attempt to make the artwork something that can exist apart from, and after the death of, its original creator. A poem recited by Homer or a ballet choreographed by Ballanchine depends upon the presence of these persons. But if they can be handed down, and either a written text of the poem or a notated description of the dance given, then later generations can have access to the poem or the dance by these artists (allowing a higher degree of variation in dance, where nuances of performance not subject to notation are vital).

The question “Could the visual arts become allographic?” might be phrased in two related ways. First, can those artworks already in the museum become allographic? Second, can a new allographic visual art be developed? In the first case we might imagine making full-size Polaroids of Bellini’s Feast of the Gods so good as to be indistinguishable from the original. Such an example raises problems that can only be noted here.9 Noting these problems is worthwhile because, as we will see shortly, they raise questions for the art critic. Some puzzling evidence of Titian’s hand in this painting can be seen only on x-ray photographs of the original: so, even if the Polaroid copy were indistinguishable from the original, seen in normal light that original would still possess properties of esthetic significance not found in the copy. Then, suppose the original, but not the copy, faded. Would we say, “The Feast of the Gods has faded?” Finally, as long as what counts is that such copies are made from Bellini’s original, painting could remain a largely allographic art in which the original would only have the special status of defining the nature of the artwork, like the manuscript of a published book. All the esthetically significant literary features of Nabokov’s handwritten copy of Lolita are duplicated in my paperback edition. The identity of my copy as a copy of that novel can be checked by checking the order of the words, even if the original has been destroyed.

Much art of this century challenges the autographic status of visual art. Duchamp’s readymades are distinguished from identical-looking urinals, bottleracks and snowshovels because they are the objects chosen by him (more exactly: after the originals he chose were lost, he chose some others). Just as Bellini’s painting is that object made by him, Duchamp’s readymades are those objects chosen by him. But put this way, the parallel seems misleading. Close visual scrutiny, or in this case the use of x-rays, reveals esthetically significant features of The Feast of the Gods. But it would be absurd to imagine a would-be Berenson studying Duchamp’s Fountain with a magnifying glass. Given Duchamp’s intentions, would it matter if we discovered that the “right” Fountain were not the one displayed in his shows? (If so, how much would it matter? How different would one of Duchamp’s own copies be in this regard?)

LeWitt has given instructions for making a work and has left the execution partly or entirely to others. To think of him as being like Rubens with his studio assistants would be misleading. Connoisseurs seek to distinguish those works painted entirely by Rubens from works in which he painted only the faces, or provided only the overall design. Wondering which if any of the colored lines were drawn by LeWitt himself would he pointless. Suppose I follow his instructions. Can what I draw be sold as a LeWitt? If not, what is the relevant difference between my drawing and a drawing made by assistants hired by him who work from the same instructions?

I would love to possess, but will never be able to afford, a Stella “Protractor.” Suppose I have one of my art students make a copy. This copy might seem quite different from a forgery of a Vermeer or Rembrandt. We think of those Dutch masters as providing not only the design for a painting, but also as executing every detail of their works. A forged work fails to duplicate the “touch” of these painters. But in some contemporary painting—Stella and Lichtenstein, but not Olitski, Frankenthaler or Diebenkorn—we might think of the painter as producing a “design” for the finished work. In viewing such painting, our concern with the particular artist’s touch is negligible. Our evaluation of a particular Stella, but not a Diebenkorn, might not change were we to discover that the artist provided only a drawing, leaving the execution to someone else.

Wollheim describes this difference by indicating why in some cases we are right to be concerned with the artist’s touch: “An artist could come to regard the work of art . . . as the summation of all (un-erased) states of that work. Of course some of those states, buried under others, would no longer be . . . discriminable. Nevertheless, the artist’s preparedness to look upon those states as belonging to, indeed as constitutive of, the work of art would naturally affect his attitude to the work.”10 X-rays show us an earlier state of The Tempest which is constitutive of that work in the way that ruled lines overpainted later in a hard-edge painting are not a “state” of that painting.

So, there is considerable evidence to suggest that the reasons for thinking of old-master visual art as autographic are not such strong reasons to think of some contemporary visual art as autographic. Two opposed reactions to this trend can be found in contemporary art. There is the attempt to produce works which, though not actually allographic, are thought of as being available fully, or available only in reproductions. Some earthworks, which most of us will see only in photographs,and much Conceptual art move in this direction. Art in which whatever physical changes occur in the works made by the artist are to be thought of as part of his work, is similar in one respect. “The control of the artist over his work is relaxed.”11 Imagine this trend taken a step further. The artist might provide just a design, and leave to different individuals the task of executing it. Children’s coloring books are a primitive model here. Such an artist would be like a composer who leaves the interpretation of his work to the performer. So, just as we compare the performances by Pollini and Rosen of late Beethoven sonatas, we might compare different interpretations of the same design. Such art would seem a natural outgrowth of that trend noted by Walter Benjamin, Andre Malraux and Edgar Wind: much of our appreciation of visual art tends to depend upon reproductions, not the original works.

The opposed trend in contemporary art is the attempt to produce works whose esthetic qualities are unreproducible. In photographs the sensuous effect of Kelly’s vibrating large areas of color is lost; Reinhardt’s late works appear to be solid fields of black; the subtlety of much color field painting disappears. Finally, Duchamp’s Etant donnés is not to be photographed. Though making a photograph would pose no technical difficulties, photographs would destroy that condition of erotic suspense Duchamp desires in the spectator.

Generalizing a point made by some Conceptual artists, we may see the move toward allographic visual art, and the opposed trend to make works whose effect gets lost in reproduction, in relation to the tendency for such precious objects as artworks to become commodities.12 Although the political reasoning of these Conceptual artists was confused—are we to believe that allographic arts like literature or music are any less a part of capitalism than painting?—their identification of visual art as moving toward becoming an allographic art was insightful.

I turn now from Goodman’s autographic/allographic distinction to a more general account of his esthetics. For him there are three basic ways of describing art: representation, expression, and exemplification. All three are defined in terms of one fundamental notion, denotation, in the sense that a symbol denotes what it refers to or stands for. A descriptive label—“is gray” or “is square,” for example—denotes a picture if that picture possesses the property of being gray or being square.

The simplest form of representation is just denotation. Constable’s Wivenhoe Park denotes that English estate. Goodman rejects any definition of representation in terms of resemblance between the picture and what it represents. Unlike Gombrich, whose Art and Illusion is the most sophisticated recent reworking of that notion, Goodman denies that perspective is an objectively valid way of depicting the world. For him, perspective is just one conventional method of representing. The relation between a symbol and what it denotes is entirely arbitrary (readers of French philosophy will recognize the affinities with the sign/signifier distinction). Of course, a familiar system of denotation seems natural to those who use it; just as for English speakers it seems natural that “dog” should denote what in German is denoted by “Hund.”13

A fuller account of representation becomes more complex. Pictures of unicorns do not denote animals that exist. There Goodman speaks of the label “unicorn-picture” as denoting the picture. Other complications occur in dealing with caricatures or symbolic pictures. Churchill can be represented as a lion, and St. John the Evangelist as an eagle. The labels “lion-picture” and “eagle-picture” apply to these pictures which denote Churchill and St. John. (An open question is that of the “iconographic” denotation of persons; i.e. the way a picture may denote both a real person and abstract ideal, such as a virtue.)

Exemplification is a more complex form of symbolization. A tailor’s booklet contains swatches of cloth exemplifying the color and texture of the clothes he makes. The sample does not exemplify its properties of being square, or being bound in a book. So, to exemplify a property a symbol must both possess that property and refer to it. Knowing which of its properties a symbol refers to requires knowing the relevant symbol system. A tailor’s samples exemplify the pattern, but not the size of the material; a baker’s samples the size of the cake, but not the date of baking.

Expression is a special form of exemplification. A painting is somber if the label “somber” metaphorically denotes it, and the painting refers to that label. People are literally somber, joyous and lively: such labels can only be applied metaphorically to art. Two sorts of properties are not expressive: first, some, like “being red,” are literally possessed by the painting; second, others, such as “being a gold mine,” are metaphorically possessed but are not expressive qualities.

What are the implications of this analysis for criticism? An artwork is a symbol functioning in relation to a particular symbol system. Just as a pawn functions as such only in chess,14 and not when used as a paperweight or admired as a sculpture, so a painting denotes and is denoted by particular labels only within a given symbol system. In another symbol system, Wivenhoe Park could represent a beached whale,15 and late Rothkos might be light and cheerful. Goodman’s radical relativism does not make him a subjectivist. Relative to a particular symbol system, questions about the properties of an artwork are not subjective: “ ‘Sad’ may apply to a picture even though no one ever happens to use the term in describing the picture; and calling a picture ‘sad’ by no means makes it so . . . by practice or precept, the use of ‘sad’ . . . is not arbitrary.”16

Goodman’s account rules out three sorts of criticism. First, a teleological view like Gombrich’s, that sees the history of art as the perfection of the means of representation, is wrong. For different systems of representation are just different conventions.

Second, any approach like formalism, relying on a distinction between properties belonging to the artwork itself and properties the artwork possesses in relation to something else, is mistaken. For properties of a symbol belong to it by virtue of its function in a symbol system. Roger Fry’s distinction between the formal properties of the painting and those literary qualities which require reference outside the work, and Michael Fried’s distinction between modernist art which contains internal relations and theatrical art which possesses its qualities only in relation to the spectator, are both wrong. A conservative critic like Edgar Wind, who criticizes modern art for lacking the symbolic meanings of a Botticelli, is just as mistaken as the formalist who believes such meanings are nonesthetic. And the teleological view of art history as the progressive discarding of those properties of painting not “internal” to its definition as a medium is unacceptable.

Third, the historicists’ suggestion that all interpretations of art of the past are subjective is to be rejected. Learning the symbol system needed to interpret such art may be more difficult than learning the symbol system for familiar art, but the difference in difficulty is only a matter of degree.

Nothing here implies that some symbol systems may not allow finer discriminations, or a broader range of discriminations than others. But those differences do not necessarily make one system superior to another. That Pontormo’s Deposition depicts bodies in a variety of complex poses unavailable to earlier artists does not make it a more powerful religious statement than Giotto’s work.

The function of art criticism is to determine what an artwork represents, expresses and exemplifies by determining what symbol system it is to be fitted into. Art of the past has already been placed with reference to symbol systems. What makes criticism a difficult and interesting activity is that the critic must decide which of the old conventions of symbolization apply to a novel-looking work. The critic is like a linguist trying to learn a new language. Although the Europeans who concluded that the native word for an Australian animal was “kangaroo” were mistaken—in the native language, the word apparently means “what’s he saying?”—the principle of investigation which they used was a good one. A preliminary mistaken judgment can always be revised. The critics who found Pollock merely decorative, Pop art reflecting a love of popular culture, and Minimal art essentially anonymous, made fruitful errors because they gave later critics some starting points. It is unlikely that a new work will upset all of our expectations, for then it would be hard to place. To speak of the importance of continuity in art is less to note a fact about art than to underline a necessary heuristic principle for criticism.

Because Leo Steinberg believes “all the haste and waste involved in his self-education should be preserved in a book,” his description of his reactions to Jasper Johns’ early work is helpful here.17 We watch him using his knowledge of earlier art to search for the right description, testing his account at each stage to see if it matches his experience of the art before him. A selection of criticism from the March 1979 Artforum further illustrates these procedures.18 Robert Ryman’s works are compared to Rothko’s and Newman’s. Why, the critic asks, do I find their use of a fixed motif satisfying, while Ryman’s use of such a motif bores me? William Bailey’s still-lifes, like Matisse’s works, rely on an “idealistic conservatism”; but he lacks the greatness of that artist, his safe color range and “smooth blankness” reminding one of Andrew Wyeth, and his bad cropping suggesting a deceptive connection with Minimal art. Another critic finds Stella’s new works mixing a whole series of categories: sculpture/painting, color-field painting/Punk; though the result is “startling and cathartic,” whether a lasting direction is established is unclear.

These critics place these new works by comparing them to known ones. Bailey’s paintings do not look much like Matisse’s or Wyeth’s, but those comparisons help establish the expressive qualities of his work. Comparing Ryman’s paintings to those of other artists, or trying to find the right categories in which to place Stella, serves a similar function. Such examples suggest why writing criticism is like developing a scientific theory. A scientist who just noted “facts” would be as inept as a critic who mechanically made a list of the properties possessed by an artwork. Determining which properties an artwork exemplifies requires determining which of those many properties it possesses it refers to. That is as creative a task as finding what theory would give an elegant and convincing account of the date.

Of course, science is cumulative in a way that art is not. Einstein’s general theory of relativity goes beyond Newton’s theories in a way that Matisse cannot be said to go beyond Raphael.19 But in one sense, art is cumulative. One defining quality of the esthetic use of symbols is “multiple and complex reference, where a symbol performs several integrated and interacting referential functions, some direct and some mediated through other symbols.” And a relatively simple example of this multiple reference occurs in contemporary artworks that “quote” or otherwise refer to older art.20 Lichtenstein’s The Artist’s Studio with Dancers quotes Matisse’s Dance, and Larry Rivers’ The World’s Greatest Homosexual refers to David’s Napoleon. In more complex cases of cross-reference we understand one work by comparing it to another. The texture of Brice Marden’s surfaces may be contrasted to that of a Johns or Kelly, and his abstract reference to nature contrasted to the representational depictions of Cézanne.

Furthermore, as the philosopher Arthur Danto notes in his elegant article “The Artworld,”21 and as some critics have remarked, appreciating contemporary art may change our perception of art of the past. Ron Davis’ use of perspective may change our view of perspective in 15th-century art, and seeing some unstretched or unframed works may make us more conscious of the significance of the stretched and framed canvas in older art. Rejecting both the teleological accounts of Gombrich or Greenberg and Wind’s conservative critique of modern art, we are free to compare art of many different periods in a way that enriches our perception of all art.

Using such comparisons is no novelty, but a staple of much traditional criticism. Both Vasari and Reynolds, for example, define the qualities of the two men they believe to be the greatest masters, Raphael and Michelangelo, by contrasting the former’s sweetness and versatility with the latter’s feel for the sublime.

Discussing WöIfflin’s formalism, Gombrich argues that there is a fallacy inherent in such comparative studies.22 Comparing a Rubens to a Leonardo, it makes sense to think of Rubens choosing to paint differently from Leonardo; but the converse comparison attributes, absurdly, to Leonardo the choice not to paint in a manner he could not have known. Gombrich makes a good point here about how we can think about artists, but we need not accept his suggestion that there is something wrong about critics making such comparisons. Seeing that Leonardo da Vinci painted differently from Rubens is a useful way of drawing attention to some aspects of his work, even if it is not an observation that Leonardo himself could have made.

This discussion of Languages of Art suggests that the description “institutional theory of art” might better be applied to Goodman’s account than to the theory of that name described earlier. Goodman quickly rejects that theory: “What distinguishes what is from what is not a work of art? . . . That it is exhibited in a museum or gallery? No such answer carries any conviction.”23 Goodman’s account of the symbol systems suggests that the real task of criticism is to give a detailed and explicit analysis of the working institutions of the art world.

What is the status of these symbols that Goodman describes? Some theorists believe representation or expression are nonconventional, their status being defined by the nature of the world or the characteristics of our human nature. Some symbol systems depend on rules everyone knows. “Drive on the right, stop at red lights” are conventions we agree to when getting an American driver’s license. But Goodman’s symbol systems are neither nonconventional nor conventions based upon such explicit agreements. Why does he think judgments about those systems are objective?

Earlier theorists who, like Goodman, thought of criticism as a creative activity, often produced very subjective criticism. Today Pater’s famous account of the Mona Lisa is likely to seem more a projection of his fantasies about “fatal women” than an objective analysis of Leonardo’s painting.24 How do we know that in 2079 our art criticism won’t appear equally subjective? Joseph Masheck raised the relevant question here: “One might suspect that stylistic comparison in art depends too much on random, very possibly inappropriate, comparisons . . . Juxtapositions . . . have a way of opening lines of thought as well as closing them. . . . Stylistic propositions are more demonstrable than matters of opinion (perhaps they are instead like matters of belief).25

Here a parallel with linguistic conventions is helpful. I know that “The cat is on the mat” is grammatical, and that “cat mat is on the” is ungrammatical. Offhand, I do not know the theory of English grammar that supports that claim. But my judgment is not subjective. I am sure all English speakers would agree with me. What more objectivity could there be here?26

In the past century judgments about the importance of many artists have changed drastically. Guercino has become an important master, and Bouguereau a nonentity (and then, again, an entity of sorts). These changes might make a skeptic feel that criticism cannot be objective. Such a skeptic might be consoled by noting that often radical disagreements about an artwork’s value are compatible with agreement about its symbolic properties. Roger Fry’s condemnation of Caravaggio as a movie director before his time shows he understands the properties of Caravaggio’s works, and finds them disgusting just because he does understand them.

Here is another approach to this question about objectivity. The critic faced with a new work, I suggested earlier, tries to see what rules apply to that work. This critic is in a position like that of the artist making the new work. As Duchamp said: “The artist may shout from all the rooftops that he is a genius; he will have to wait for the verdict of the spectator in order that his declarations take a social value and that, finally, posterity includes him in the primers of art history.”27 When the critic’s judgments are accepted by posterity, those judgments become agreed-upon conventions. Consider again the parallel between criticism and science. Some metaphysicians claim that the world has a certain structure. Were they correct, then the aim of science would be to find that way the world is. Goodman rejects any such metaphysics. The discovery of new data, inconsistencies in the theory itself, or the discovery of a theory that covers better the existing data, may lead to rejection of a scientific theory. Analogous considerations can lead to the rejection of a critical theory. We can reject formalism because it fails to account for much art of the 1960s, or because a certain concept of modernism is incoherent, or because other accounts better describe the relation between “old master” and “modernist” art. That scientific theories are open to rejection is reason to believe science gives us objective knowledge. The trouble with pseudosciences like astrology is that they are not open to change in this way. If such change of theories in science should be a reason to call scientific theories objective, why should not the same be true of critical theories?

Consider Robert Rosenblum’s recent LeWitt catalogue in light of this discussion of criticism.28 Like Reinhardt, Kelly, Twombly and others, LeWitt strips art of “everything but elemental truths”; like Matisse, Newman, Chuck Close and others, he uses impersonal means to achieve a personal end; like Mondrian, Tobey and late Monet, his vocabulary is “rigorously restricted to fragmented geometric or quasi-geometric parts.” And so forth. There is hardly anyone in the history of art whom Rosenblum is not prepared to relate to LeWitt. Well! What right have Ito object? Did I not say that discovering the conventions applying to one artwork allows us to see that work in relation to the history of art? The answer is that criticism must aim at selectivity as well as breadth. Masheck’s relating of LeWitt to Neo-Classicism, Stella’s early works and the “Parsons Table” is helpful precisely because it is selective.29 Out of all the possible comparisons, he selects a few. Rosenblum’s account illustrates the esthetic theory I have discussed, but is a perplexing piece of criticism. The aim of criticism is not to exhibit every possible comparison, but to focus on a few relevant ones. As Goodman writes: “The search for accurate adjustment between symbol and symbolized calls for maximal sensitivity . . . we can decrease the risk of error by using more general terms; but safety is then gained by sacrifice of precision.”30

This discussion about the use of rules in criticism might usefully be placed in the context of recent debates about philosophy of language. Contemporary philosophers talk about linguistic rules in two ways.31 We may focus on the individual speaker, and talk about what he intends to communicate. Or, we may attend to the communication system, language, and talk about the conventions governing its use. It might seem as if these two approaches would come to the same thing. That is not necessarily so. Those starting with the language system find perplexing the talk of an individual speaker following rules he may not be aware of. In response, a theorist who focuses on the individual speaker writes: “The agent’s knowing how to do something may only be adequately explicable on the hypothesis that he knows (has acquired, internalized, learned) a rule . . . even though . . . he may not know that he knows the rule. . . .”32 The notion of such knowledge is complex. Can I know a rule I have no conscious awareness of? Why not say we know how to digest food, since we are able to digest it in our stomachs?33

Two different theories of the mind might help defend the notion of such knowledge. Chomsky’s linguistics requires that a speaker know a theory of his language, though only the linguist discovers that theory. And psychoanalysis leads to the conclusion that we have unconscious desires and thoughts. Of course, many find these two theories problematic because they do require us to have such knowledge.

This philosophical debate about rules in language is mirrored in the debate among estheticians about how artworks are to be described. Goodman’s definition of rules in terms of the intrapersonal symbol system—the equivalent to the focus on the language system—can be contrasted with Richard Wollheim’s account of rules as “something the individual painter internalizes.”34 Wollheim thinks of an artist painting in a given style because he has internalized the rules defining that style, rather as Chomsky thinks of English speakers as internalizing the transformational rules of that language. Thus, “In late Cézanne and much Matisse we encounter a kind of complexity which could be at least partly accounted for by the operation of an abstractly conceived figure-ground array of a certain rule—a rule that has the effect of successively retrieving the ground in such a way as to make it approximate to, or of equal weight with, the figure.”

Goodman finds such an approach to art as problematic as Chomsky’s linguistics. “Some artists . . . do have such overt theories . . . [But few] have such theories explicitly, and what it could mean to say they have the theories implicitly puzzles me as much as what it would mean to say the planets have the laws of motion implicitly.”35 For him, what counts is the relation defined by the symbol system, not the intentions of those “who perpetuated the system.”35

Wollheim writes: “The artist is active, but so also is the spectator, and the spectator’s activity consists in interpretation.”37 He and Goodman offer different ways of understanding that creativity. For Goodman, the artist thinks in terms of the media of art, and criticism must offer a verbal account of that thought. For him, Wollheim’s description of late Cézanne and much Matisse would be a creation of the critic. Wollheim believes we must attribute implicit knowledge of such a theory to the artist. He can appeal to a general conclusion in contemporary philosophy of mind. Classical philosophers from Descartes to Hume tended to assimilate thought as having pictures in the mind, while the contemporary approach is to emphasize the affinities between thought and language. So, even if the artist thinks in terms of the media of visual art, those thoughts are couched in language-like terms.

Vico made the famous claim that products of human activity, like historical events or artworks, can be understood by us more directly than the external world of nature because they are our products. Wollheim could agree with Vico, and suggest that the critic’s account is a reconstruction in words of the artist’s thought processes. Wollheim has been influenced by the great, but as yet little discussed, critic Adrian Stokes, who, in turn, was much influenced by the analyst Melanie Klein. Just as Klein finds it appropriate to describe the thoughts of a very young child in language, so for Stokes it is natural to “assimilate a shape to a feeling, or to equate the use of a specific material with a fantasy.”38 I indicated earlier that there is reason to believe that accounts of rules in terms of the user’s intentions are not directly equivalent to accounts describing conventions. Stokes’ account of the artist’s intentions is an account of the attitude taken by the artist toward the media of art: I do not think his account can be translated into a description of how the viewer sees the work. Stokes suggests not just that we see the artwork as having certain qualities, but that we see it as made by an artist who had a certain attitude toward what he was making.

Goodman would disagree with Vico. Granted, artworks are made by people, and nature is not. But the critic facing the artwork, like the scientist analyzing nature, must describe in verbal or mathematical terms the objects of inquiry. This parallel is not meant to suggest that art is so mysterious that its significance can scarcely be captured in words. The problem is not just that criticism is in words and visual art is not. Making art is one kind of creativity; describing it, another. This creativity of the critic is not determined by the fact that the artwork is not in words. For the literary critic faces the same problems as the art critic. Finding the words that best describe a verbal text, like finding words to describe a visual artwork, is a matter of searching for that symbol system in relation to which the artwork can best be understood.

If we follow Goodman, we may believe Wollheim’s ideal of criticism founded on a psychology of the artist is mistaken. We will believe the critic should describe what the artist makes, not the mental processes that went into making that object. In response, Wollheim can suggest that all acceptable theories about the artwork are implicitly theories of what the artist does. Take one apparent paradigm of a theory not based on a psychology, WöIfflin’s formalism. Cannot even his analysis of the Baroque be read as a description of artists’ mental processes? Arranging objects in a recessive structure in the illusionistic picture space requires first thinking of the simpler plane structure, and then transforming that into a more complex structure.

Describing these disagreements between Wollheim and Goodman is meant to show, again, why philosophical disagreements are of interest to the critic. Choosing between their approaches here can lead to very different views about the best critical methods.

There is one approach to criticism that both Goodman and Wollheim give reason to reject. According to a view fashionable a decade or two ago, philosophers analyze the concepts of other disciplines. Critics talk about art, and estheticians talk about criticism. If criticism is a second-order activity—a discussion of artworks—then esthetics is a third-order discipline, talk about talk about art. The whole point of this essay has been to give reason to reject such a hierarchy. Art-making is one kind of creativity, art criticism another, and esthetics yet another. If estheticians could only talk about critics’ talk, then why would art critics have any reason to be interested in esthetics? The esthetician could only tell the critic what he or she does. Why would any critic be interested in that? I have argued that an interesting esthetician, like Goodman, develops a theory with implications about what kinds of criticism are acceptable. If I am correct, esthetics and art criticism are not, ultimately, separable activities. Which is why learning about esthetics can be worthwhile for the art critic, not to mention the artist reflecting on his own works.

David Carrier



1. George Dickie, Art and the Aesthetic, Ithaca and London, 1974, p. 42.

2. Richard Wollheim, preface to his On Art and the Mind, London, 1973, p. ix.

3. Friedrich Nietzsche, The Birth of Tragedy, trans. Walter Kaufmann, New York, 1967, p. 22.

4. Nelson Goodman, “Reply to Howard,” Erkenntnis, January 1978, p. 165.

5. For an account of the earlier stages of the popularization of Kant’s esthetics see Michael Podro, The Manifold in Perception, Oxford, 1972.

6. Unfortunately the one such article known to me dealing with Goodman and visual art. Epi Wiese’s “A Goodman Primer for Art Historians,” Art Journal, Fall 1976, pp. 39–44, both presupposes some knowledge of his technical vocabulary and offers a rather idiosyncratic application of his methods to art history.

7. The interpretation of Languages of Art given here is my own, and may not agree with what Goodman would say. My focus on his work should not overshadow my even greater intellectual debts to two other contemporary estheticians, Arthur Danto and Richard Wollheim, nor the significance of perhaps the best known American esthetician, Monroe Beardsley. I concentrate here on Goodman because Danto has not yet published his major work on art, while Wollheim’s important work is intimately linked with the visionary and elusive criticism of Adrian Stokes.

8. See Nelson Goodman, Languages of Art, Indianapolis, 1976, ch. 3.

9. My “Paintings, Conceptual Art and Persons,” Philosophical Studies (forthcoming), discusses this case in some detail.

10. Richard Wollheim, “Are the criteria of identity that hold for a work of art in the different arts aesthetically relevant,?” Ratio, 1978, pp. 29–48; here p. 41.

11. This phrase from Richard Wollheim, Art and Its Objects, New York, 1968, p. 73.

12. See Lucy R. Lippard, Six Years, The dematerialization of the art object, New York, 1973, preface, pp. 5–9, and “postface,” pp. 263–4.

13. This example raises a question I have not space to treat in this paper, and which provides a link between Goodman’s allographic/autographic distinction and his analysis of symbol systems. Given that both a picture of a dog and the word “dog” denote dogs, what is the difference between verbal and pictorial denotation? Goodman’s account of a notation (Languages of Art, ch. 4), discusses that problem in rather technical terms, and shows why autographic but not autographic arts possess a notation.

14. David Perkins and Barbara Leondar, eds., The Arts and Cognition, Baltimore and London, 1977, p.7. This volume is a collection of studies in experimental psychology by Goodman’s “Project Zero” at Harvard University.

15. This example from Goodman’s review of Art and Illusion, The Journal of Philosophy, 1960. pp. 595–99; here p. 599.

16. Goodman, Languages, p. 88.

17. Leo Steinberg, “Jasper Johns: The First Seven Years of His Art,” in his Other Criteria, New York, 1972, pp. 17–54.

18. Jeff Perrone and Deborah Perlberg, “Reviews,” Artforum, March 1979, pp. 61–4.

19. This important topic is discussed by Goodman in a tantalizingly brief footnote in his Ways of Worldmaking, Indianapolis and Cambridge, 1978, p. 140; next quotation from p. 68 of that book.

20. I use the word “quotation” in a different way from Goodman, his account is in Ways, pp. 41–56, mine in my “On the Depiction of Figurative Representational Pictures within Pictures,” Leonardo (forthcoming).

21. In The Journal of Philosophy, 1964, pp. 571–84.

22. E.H. Gombrich, “Norm and Form: The Stylistic Categories of Art History and their Origins in Renaissance Ideals,” in his Norm and Form, London, 1966, pp. 81–98; see esp. p. 90. Using the useful account developed by the literary critic E.D. Hirsch, Jr. in his Validity in Interpretation, New Haven and London, 1967, such categories describe not the artwork’s meaning, which is defined by the artist’s intentions, but the significance of the work has acquired for us. This reference was suggested to me by Gombrich; I will discuss his account of art-historical explanations at length in another article.

23. Goodman, Ways, p. 66.

24. Mario Praz, The Romantic Agony, trans. Angus Davidson, London and New York, 1970, pp. 25–37.

25. In his “Kuspit’s LeWitt: Has He Got Style?,” Art in America, Nov–Dec 1976, pp. 107–11, here p. 107, italics mine.

26. A complex and thorough philosophical discussion can be found in David K. Lewis, Convention: A Philosophical Study, Cambridge, Mass., 1969.

27. Marcel Duchamp, “The Creative Act,” in Gregory Battcock, ed., The New Art, New York, 1966, pp. 23–6, here p. 24.

28. Robert Rosenblum, “Notes on Sol LeWitt,” in the Museum of Modern Art catalogue Sol LeWitt, New York, 1978, pp. 15–21.

29. Masheck, “Kuspit’s LeWitt,” pp. 107–11.

30. Goodman, Languages, pp. 23–56.

31. A good brief summary of the complex philosophical issues is given in E Hirsch, Jr . The Aims of Interpretation, Chicago and London, 1976, pp. 68–71.

32. John R. Searle, Speech Acts, Cambridge, 1970, p. 42.

33. The example from Thomas Nagel, “Linguistics and Epistemology,” in Gilbert Harman, ed., On Noam Chomsky: Critical Essays, Garden City, N.Y., 1974, pp. 219–28, here p. 220.

34. Richard Wollheim, “Style now,” in Bernard Smith ed., Concerning Contemporary Art, Oxford, 1975, pp. 133–53, here p. 143; next quotation p. 151. Wollheim is preparing a book on style.

35. Nelson Goodman, “Comments on Wollheim’s Paper,” Ratio, 1978, pp. 49–51; here p. 50.

36. Nelson Goodman, “Reply to Kidrup,” Erkenrifnes, Januaray 1978, pp. 162–4, here p. 162.

37. Wollheim, Art, p. 75.

38. Wollheim, On Art, pp. 333–4.