THE STYLE THEORY OF ART BY JAMES D. CARNEY A historically oriented, relational theory of art is one of the more promising current approaches to the search for a definition of art. As an attempt to improve or give a more detailed version of the historical approach, this paper combines parts of the Kripke-Putnam story of what determines the extension of a term with a historical approach to defining art.^ The idea that art can be identified historically is not new. The idea is contained in Arthur Danto’s famous remark that “to see something as art requires something the eye cannot decry—an atmosphere of artistic theory, a knowledge of the history of art: an artworld.”^ Both Jerrold Levinson and Noël Carroll have recently offered versions of a historical theory of art which I consider in section V.^ I understand attempts to define artworks historically as attempts to provide accounts of the nature of artworks. Common to all historical theories of art is the notion that we identify a new work as art in virtue of a relation in which it stands to things already acknowledged as art. What is essential to being an artwork is having the right sort of relation to previous artworks. Objects are works of art in virtue of standing in some designated relation to past artworks. It is part of the Kripke-Putnam story that what determines the exten sion of a term is the “nature,” broadly conceived, of the kind of thing to which the term refers.^ ‘Artwork’ is not a natural-kind term if a naturalkind term must designate a kind that is part of nature, as opposed to culture. ‘Artwork’ is a social-cultural-historical kind term, not a kindfrom-nature term. Still ‘artwork,’ it seems, is a term for a real kind, and I wish to propose that its extension is fixed historically according to something like the Kripke-Putnam pattern. Specifically, W is an artwork if it is of the same kind as indexically picked out samples. Being of the same kind consists in having the right sort of relation to past artworks. What I propose the relation to be is explained next. Pacific Philosophical Quarterly 12 (1991) 272-289 0279-0750/91/0400-0272 $1.80 Copyright © 1991 University of Southern California 111 THE STYLE THEORY OF ART /. 273 The Style Theory o f A rt What I will call the historical/style theory of art is anchored to an extension of the Kripke-Putnam theory of reference for natural-kind terms, though ‘artwork’ is not construed as a natural-kind term. What is borrowed from this well-known view is that W is an artwork if and only if it is of the same kind as the W’s to which ‘artwork’ has been applied, where same kind is construed as standing in a specified relation to past artworks, and where the “originating samples,” Ur Art, are indexically designated. Most individuals in a community that use the word ‘artwork’ need not have the identifying knowledge. The word ‘artwork,’ on the view, exhibits what Hilary Putnam calls a “socio-linguistic division of labor.” It is enough for users to rely on knowledge and expertise at large in the community. Three elements are borrowed from the KripkePutnam account: (a) indexical component for ‘artwork’; (b) linkage via same kind, where kind need not necessarily be natural-kind; and (c) sociolinguistic division of labor by way of relying on knowledge of experts, those who are suitably informed about art. ‘Same kind’ here is vague, and ‘knowledge of experts’ is obscure. A style theory approach provides explication of ‘same kind’ and ‘expert’ that, I hope, is both promising and interesting. Five features make up the theory. They are: (1) At any time in history there are a finite number of objects regarded as artwork, allowing for borderline cases, and most of these are classified by art historians in terms of general styles. (2) Though there are no uniquely correct style classifications, no in variant general styles, there are limit constraints that allow for a degree of objectivity in style classifications, and thus for knowledge and expertise. (3) An object is artwork if and only if it can be linked by those suitably informed, along one or more of various specific dimensions, to a past or present general style or styles exhibited by prior artworks. (4) The ways to make the linkage are open-ended. Some of the common ways are: show that the object in question is a repetition, amplification, repudiation, synthesis or reinterpretation of a general style or styles.^ Other ways include: preserving, distorting, extending, restructuring, reanimating, or radical reinterpreting. ' (5) Following Danto, anything whatever can in some possible arthistorical context be an artwork, though certain works could not be inserted as artworks into certain periods of art history, due to the un availability of linkage to general styles.^ Of the five features, (3) is the most important since it is the one that gives conditions for an object’s being an artwork. If an object stands in the designated relation to past artworks, then it is an artwork. But are there other sufficient conditions for arthood that are independent of any 274 PACIFIC PHILOSOPHICAL QUARTERLY historical connection of this sort? If so, the conditions in (3) cannot be necessary conditions. Before giving my reason for believing that there may be no sufficient conditions for arthood independent of the causal history of an object, I shall provide some details of the theory and some illustrations. This discussion which follows focuses on visual artworks, primarily paintings. I believe there are no theoretical or practical obstacles that prevent the theory from applying to all forms of artworks. II. Determining General Styles Crucial to the view is an account of what are correct art history classi fications, and of what gets hooked up to historical styles when an object is a repetition, amplification, repudiation, and so on of a general style. Three preliminary points need to be made. First, there are two ways to understand ‘the style of x.’ This can be taken as referring to the general style of Xor the individual style of x. Picasso, Braque and scores of other artists painted in the Cubist style, yet each had his or her individual way to do Cubism. For example, Diego Rivera during his Cubist period incorporated a use of color after Matisse and incorporated a formal structure influenced by Cézanne. Second, it is important to distinguish the style of an artwork, either general or individual, from interpretations of the artwork. Style explains why a work looks the way it does. Style, in the visual arts, results from the kind of approach used to solve pictorial problems of expression, representation, and form. It is a schema for conveying significant content. Style can also include characteristic choices of subject matter. I take the style of an artwork to consist of those features of the form and content of a work that are characteristic of artist, period, place or school. Characteristic style features cut across the form/content distinction.^ A style feature of Fauvism is the use of bright, primary colors for expressive purposes. A style feature of Suprematism is the use of colored geometric form for representing gravitation-free objects. Interpretation involves fixing the particular meaning(s) of an artwork executed in a style. Third, aesthetic features of an artwork are style constrained features. Since in another place I present some arguments for this last view, I will ignore them here.^ The idea is that representational, expressive, or exemplified features of an artwork depend, in part, on the style supposed for the artwork. Linkage to past or present general art styles requires that there be some non-arbitrary, largely empirical way to identify general art styles. I suggest that there are at least three levels of constraints that are implicit in the practice of art history and are strong enough to provide something of an empirical base for classifying artworks in terms of their style. THE STYLE THEORY OF ART 275 Level 1: Any serious art historian is a revisionist to some degree. Art historians typically reevaluate established views. Any degree of innovation is made possible by changing the existing order. Thus new classifications or reinterpretations of old classifications start with the consensus classification. Level 2: Single out the artists in target classification whose quality and originality make them prototypes of artists working in variations of the target classification. With respect to these prototype artists, determine their individual styles.^ Here, I suggest, a set of limiting conditions comes into play which provides constraints on what is taken as the correct S for an artwork. I suggest that the following limiting conditions, none of which alone is sufficient, count toward a style S being the individual style of the works of an artist: - Limit 1: The fact that art-historical consensus has it that S is correct (unless there is a good reason not to follow consensus in every respect). - Limit 2: The fact that a relatively large number of features are standard with respect to S. (This condition is taken from Kendall Walton’s well-known set of criteria for determining the correct cat egory for W.)^® - Limit 3: The fact that S accords with the artist’s aims and goals. - Limit 4: The fact that S counts toward explaining the relationship between W and past general styles. - Limit 5: The fact that S helps explain the relationship between W and culture and history of the time. Level 3: Relate individual styles to the general style by providing an interpretation of the general style that allows for the individual styles’ being variations of it. To illustrate, consider the general style Symbolism. One starts with the artists being regarded as Symbolists and with the consensus interpretation of Symbolism, Level 1. On Level 2 select those artists which best exemplify the style, for example, Odilon Redon, Gustave Moreau, Ferdinand Hodler, and George Minne. Locate these artists in their social and historical context. For example, identify the ideologies, life-styles, and interchanges among painters and other artists and critics that character ized the climate in the Symbolist decade (late nineteenth-century). Pro vide an account of individual styles that explains the recurrent themes such as Salome, the Holy Grail, Orpheus, and so on; and the occurrence of decorative motifs, and so on. Relate the individual styles to the general styles, such as those of the Pre-Raphaelites and Impressionists, and the Expressionists who succeeded them. Consider how these artists understood their own works. Taking the works of the selected artists as variations of Symbolism, provide an interpretation of Symbolism, Level 3; for example. Symbolism rejected objectivity in favor of the 276 PACIFIC PHILOSOPHICAL QUARTERLY subjective, and turned away from the direct representation of reality in favor of a synthesis of aspects of it, aiming to suggest ideas by means of themes such as Salome, and so on. Though the aims and goals of prototype artists are one of the Level 2 limiting conditions, neither this condition nor the others are taken as crucial or basic; none has priority over the others. They lean on one another. A prototype historical account of a general style is a holistic account that best blends the limiting conditions. When the limiting conditions compete or conflict, trade-offs are made, different balances are struck. Thus it is possible that in some cases little weight is given to a prototype artist’s aims and goals. Cases like this occur when the expressions of intentions are murky and/or contradictory. Kasimir Malevich (1878-1935), the Soviet artist who pioneered the general art style Suprematism, is an example of an artist whose remarks about the aims and goals of his art are notoriously obscure and inconsistent. Some times an artist is inexplicably aloof with respect to how he or she thinks of his or her works, for example, the neo-Dadaist Jasper J o h n s . ( I n limiting condition (3) ‘aims and goals’ is narrowly interpreted in terms of conscious psychological states. If ‘aims and goals’ is broadly defined as a pattern or patterns in a whole series of actions, then there seems to be little difference between an artist’s individual style and an artist’s aims and goals.) Most artworks have a correct style. In cases where the limiting con ditions do not signal an S, they constrain what goes into the set of possible correct S’s. For example, neither Johns nor Malevich would be taken as a prototype Symbolist, and one could hardly interpret Symbol ism as art that places stress on the importance of color and structure, as in itself enough to provide the content of a work of art, as is true with Fauvism. Such an interpretation of Symbolism would not explain how Symbolism was an outgrowth of the style of the Pre-Raphaelites, and how it influenced Expressionism; it could not relate the artworks to the social history of the time; it would conflict with how the artists under stood their art; and so on. Style classification unavoidably reflects both the values and attitudes of the historians doing the classification and the functional theories of art favored by the historians. Currently art historians give greater relative weight to social history, ^^ por example, we find historians giving great weight to the relation between Humanism and the Italian Renais sance, stressing the social issues at stake in the art of Die Brücke, focusing on the utopian aspirations and the art of the Soviet Constructiv ists, noting Baroque art. as a response to the aims of the CounterReformation, connecting American Western art of the last century to Manifest Destiny, and seeing American art between the 40’s and 80’s as uncritical of culture and taken up into the market system because of the THE STYLE THEORY OF ART 211 economic shift from production to consumption. Not too long ago, historians who extolled the sanctity of the flat surface under formalistic influences gave greater relative weight to how a general style is to be understood as a development out of earlier styles—art understood as development within a flat medium that is contentless and politically neutral, and is evolving to pure abstract art. Before formalism there was a time when scholars thought that the artist’s biography should be given greater relative weight as part of style fixing. A scholar may believe that there is some kind of linear or progressive model for art history. For example, E.H. Gombrich seems to believe that representation, by “making and matching,” evolves towards some kind of verisimilitude, and that most of twentieth-century abstract art is lamentable due to the rejection of realism. An unsympathetic attitude toward twentiethcentury abstract art could motivate a historian to ignore such art in relating new twentieth-century realistic styles to previous abstract art styles. Because of this irreducible value component in style fixing, there are no uniquely correct classifications; no invariant classifications. Terms for general styles are also social-cultural-historical terms and not kindfrom-nature terms. But the dismissal of a uniquely correct classification does not block discrimination between correct and incorrect classification. The limiting conditions provide enough constraints for a degree of ob jectivity and for knowledge and expertise within an overall relativistic framework. General art style classifications have changed. Some classifications fall out of use. More commonly the extension of a general style or the interpretation of the style or both undergo change. For example, the general style ‘Classical’ for a long time referred to Greco-Roman antiquity, and to periods that drew their inspiration from the ancient world, especially the Renaissance and seventeenth-century France. Today, the art of Greece and of Rome are seen as totally disparate. Historians tend to restrict the Classical art in Greece to the late fifth and early fourth centuries, and the Classical art in Rome to the Augustan period. Thus there has been a narrowing of the extension of ‘Classicism.’ The mòre recent general style. Abstract Expressionism, has undergone a change in interpretation rather than in extension due, 'in part, to the values of historians and critics who have examined it.^^ In the 50’s it was interpreted in two ways: as Action Painting, according to which the artists express their personal emotions with little attention to form, style, and subject matter; and as representing traditional American concerns with emotional honesty. In the 60’s the formalist Clement Greenberg and most critics and historians located Abstract Expressionism as a stylistic evolution of European modernism to pure art. Since the 70’s critical attention has been turned to subject and meaning as well as to form. 278 PACIFIC PHILOSOPHICAL QUARTERLY Abstract Expressionism is now seen as an art with a subject matter, and it is examined within a social and historical context. IIL Linkage to General Styles On the theory, there is an open-ended set of strategies for linking works of the present to past general styles—repetition, amplification, repudi ation, and so on. On the theory, what is repeated, amplified, and so on, are style characteristics of past and present general styles. As an illustration consider Academic Art. Jean Léon Gérôme, French, 1824-1904, is amost everyone’s prototype of the academic painter. His style was founded on intense admiration of ancient art, the primacy of accurate draftsmanship, high finish, archaeological realism, anecdote, and exotic themes from historical episodes. Using Gérôme and others as prototypes of academic painters, one can obtain an interpretation of the general style of Academic Art. Such an interpretation is made of a set of characteristics. For example: representation by accurate likeness through skillful use of techniques; expression achieved by narrative content; representational and expressive content having primacy over any other considerations; use of biblical, classic, and exotic themes from history; and morally uplifting themes which serve as an instrument of national propaganda. Linking a painting to a previous general style or styles by repetition, reinterpretation, repudiation, and so on involves explaining the way a painting looks. Consider, again, the characteristics of Academic Art. Amplification is explaining how a painting looks in terms of following the practices of Academic Art to a higher degree than usual. In this regard compare the works of Lebrun with those of Ingres. (Though Ingres was not following the practices when, for example, he gave his Odalisque an extra vertebra in her back.) Repudiation is an explanation of a painting in terms of conflicts with the stylistic practices of Academic Art. Here one could cite Gustav Courbet and his realism, the view that art should be about the everyday world, and the idea that art should serve the aspirations of the working class. Synthesis is blending some of the practices of Academic Art with the practices of one or more other historical general styles. Delacroix, for example, followed many of the academic practices but freed color from the restrictions Academic Art had put on it. He was inspired by Constable’s The Hay Wain, laying down areas of color rather than blending them. The result was a more lifelike color. Radical reinterpretation is delineated as a new way to, for example, represent or achieve moral narrative content. It is arguable that Anselm Kieffer’s art is, in part, a radical reinterpretation of Academic Art. THE STYLE THEORY OF ART 279 These examples are to illustrate, not to suggest that any of the works of Courbet, Delacroix, Ingres, Gerome or Kieffer were publicly ques tioned as art, or that they had to be identified as art. Objects can be linked by quasi decision. Most artworks are artworks because of de facto linkage, without the need for those who have expert knowledge about the artwork to provide an explanation. The question whether something is an artwork has often been raised, for example, in connection with such objects as Readymades, Found art. Minimal art. Marginal art (quilts, folk art, Keith Haring cartoons, etc.) and recently with photographs such as those by Robert Maplethorpe and the works of Ruby, an elephant in the Phoenix Zoo. Some of the painted billboards in Los Angeles today are regarded as artworks. However when Gino Raffaelli executed his classic 1939 billboard for the Forest Lawn cemetery, it was not seen as art by any art community. Raffaelli’s 1939 classic billboard appears to be a spotlighted marble sculpture of a mother and child that came so close to being successful trompe-Voeil art that many spectators thought they were looking at a sculpture. It was only with later works such as Kent TwitchelFs nine-story-tall painting of Edward Ruscha completed in 1987 that L.A. billboard art was taken as art.^^ This was brought about in part by linkage to Photorealism, and by works such as Chuck Close’s 1969 Phil, The artists of these handpainted billboards used techniques pioneered by Michelangelo and made use of bold, knock-out colors reminiscent of the Fauvists. Interestingly Raffaelli’s earlier work got linked to past art styles via a style that followed his own works, i.e.. Photorealism. IV. Advantage o f Historical Approach to A historical Approaches In Anglo-American philosophy the attempts to define artwork can be sorted out in terms of four approaches: functional approach, anti definition approach, institutional approach, and historical approach. In this section I will very briefly consider the three ahistorical approaches. (1) Functional approach. Though artworks have 'many useful functions, some theorists find that artworks have a special function in culture and that this function is the primary reason we value artworks. Definitions for artworks are then constructed around this function. On this approach art is a functional concept, and whether or not something is art is seen as the question whether it meets what is regarded as the reason art is valued. The traditional theories of art proposed by such figures as Kant, Tolstoy, Collingwood, Fry, Langer and Beardsley take this approach. 280 PACIFIC PHILOSOPHICAL QUARTERLY Some examples: An artwork is an artifact that is intended to have the capacity to provide an aesthetic experience (Beardsley). The nature of art is to give conscious expression to feelings by giving them definite form (Collingwood). Art is iconic symbols of the forms of feelings (Langer). Each of these approaches is ahistorical in two ways. First, artwork is not explicitly defined in terms of any relation to past artworks. Second, whether an object has the relevant property for arthood can be deter mined from a point of view external to the history of art. For example, anyone who is sensitive to form can be aware of the unity and intensity of complex regional qualities of an artwork responsible for sustaining an aesthetic experience, independently of any knowledge about the relation of the artwork to past artworks. There are two common objections to defining art simply in functional terms. One is that it is too restrictive; there are too many counter examples. The other is that there are a variety of reasons why artworks are valued, and there is no compelling argument to make one of these paramount. Neither of these objections is insurmountable. It is not inconceivable that a functional definition can take a form that involves all the functions that art has at some given time.^^ Such a definition may not present functionalism in its most plausible or convincing form. But I wish to consider another issue. A piece serves the function(s) of art by virtue of aesthetic features. It is having aesthetic features that allows it to meet the function(s) of art. Aesthetic features, including exemplified features, are style dependent. They are not discoverable independent of some understanding of the history of the artwork. The identification of such features presupposes correct style assignment. The colors of Fauvist landscapes do not apply to what is depicted because in Fauvism colors are used for expressive purposes. The abstract colored shapes in a Malevich Suprematist paint ing depict gravitation-free objects in space because of the painting’s historical style. The abstract colored shapes and colors in an Ad Reinhardt painting do not depict anything because they are Minimalist artworks. Hudson River Western landscapes do not depict any actual places because the painters used a composite style in order to represent the ideal of what the West ought to look like. A Whistler painting lacks narrative or symbolic content not because of what is to be seen in the painting, simpliciter, but because of what is to be seen relative to the correct historical style. A historically minded functionalist must concede that only an artist working within a historical tradition can bring about an object with aesthetic features. A historically minded functionalist cannot maintain that objects are art simply in virtue of a connection to previous artworks. W is a work of art, for the historical functionalist, if and only if W THE STYLE THEORY OF ART 281 exhibits some style and the maker of W intended it to fulfill a function. But if W is regarded by the cognitive art community as exhibiting a style, not just showing one, why is this by itself not sufficient for its being an artwork? (2) Anti-definition approach. The anti-theory movement of the 60’s tried to explain why traditional functional theories failed by arguing that artwork cannot be defined. The reasons given, however, are compatible with the possibility of an essential definition. Where they may have had some success is in arguing that if a definition characterizes the nature of art in terms of properties that can be discovered (if discoverable at all) by examining the artworks themselves, then there are no such features common to all artworks, and if one insists that some such feature is necessary for arthood, then this could legislate out style innovations. If no essential definition is possible, then it is not unreasonable to infer that one needs to look to the concepts that people associate with ‘artwork’ to account for how things get classified as artworks. The anti theorists hypothesized that individuals rely on various criteria for recog nition; thus the concept of art is open and is made up of open disjunctive sets of criteria of recognition that have a family resemblance.^^ (Cur rently, asking individuals about their concept of art is likely to elicit “Art is what an artist says is art,” “Art is what the artworld regards as art” or some functional theory.) The historical/style theory is compatible with a variety of different ways for people to recognize artworks. The theory is an attempt to explain the nature of artworks. The theory is not an analysis of the various prototypes of art that individuals rely on in using the word ‘artwork.’ The extension of ‘artwork’ is not wholly explained in terms of individual subjects’ dispositions. Something is artwork, on my proposal, if it stands in the posited relation to objects to which the term ‘artwork’ has been applied. Consequently, ‘artwork’ can be an open concept, if by this one means that a nominal definition of ‘artwork’ is likely to result in some kind of open-ended disjunction. But this is compatible with there being an essential definition for artwork. (3) Institutional approach. Works are artworks as the result of the place they occupy within the established practices of the artworld. On the new version of George Dickie’s theory, “a work of art is an artifact of a kind created to be presented to an artworld public. No historical condition is explicitly included in the influential view. Many and various objections have been raised to the th e o r y .I believe that the most telling objection to the view is related to its circularity. In particular, in Dickie’s new theory the definition for artwork is one of four definitions. Each art term is defined in terms of another, resulting in a circular set of definitions. (Dickie, of course, is aware of this but argues that circularity is unavoidable.) Dickie’s definition allows for more than one model. If 282 PACIFIC PHILOSOPHICAL QUARTERLY ‘work of art’ is replaced with, for example, ‘baseball game’ or ‘fashion show’ in the definition of work of art, one obtains: a baseball game is an artifact of a kind created to be presented to a baseball public. The definition shows us that works of art, along with baseball games and fashion shows, are x’s that are presented to an x-public. But this defini tion is useless in determining the extension of ‘art’ or ‘baseball game’ other than showing that x’s are members of a class that includes art, baseball games and fashion shows, and that x differs from things that are not artifacts and that are not in some way present to an x-public. As Robert Stecker makes this point, “unless artworld systems can be dis tinguished from non-art artifact presentation systems, the framework Dickie gives us tells us no more than that art resembles many other kinds of a r t i f a c t s . Be l o w I argue that a historical theory can avoid such circularity. Advantage o f Historical/Style Theory to Historical/Intention Theory Both Noël Carroll and Jerrold Levinson have proposed that a historical approach to defining an artwork is the best approach. For Carroll W is an artwork if there is a narration that relates W to historical practices by showing that the new object is a repetition, amplification or repudiation of the historical practices. Though Carroll’s view is one of the chief sources of inspiration for my view, I, like Stecker, find the view “overly vague. 1 find Carroll’s view in need of more details and improvements which I propose can come in part from the Kripke-Putnam story and in part from historical art styles. Without further details, Carroll’s view faces the very difficulty he and others have pointed out in the openconcept approach. If “narration” is simply explained as describing a resemblance between W and historical practices, then without some kind of account of resemblance in what respect, the explanation is without much use. However, a more serious concern for this paper is that Carroll sees his view as “compatible” with Levinson’s view and “perhaps mutually informing. ”23 Levinson’s view, if I understand it correctly, is contrary to the historical/style theory. His version of the historical approach is a historical/intentional theory. An artwork, on Levinson’s theory, is a thing intended in a specified way. As he most recently expressed his theory: A work of art is a thing intended for regard-as-a-work-of-art: regard in any of the ways works of art existing prior to it have been correctly regarded. 24 THE STYLE THEORY OF ART 283 The regarder is a person with proprietary rights to the artifact. This person is usually the artist, though Levinson allows for some member of the art community (“subsequent proxy”) getting proprietary rights, and appropriating the would-be artist’s work at a later date with the right intention, and so bringing it into the sphere of art.^^ Someone can make art without having in mind any particular past artworks; the agent can have a transparent or intrinsical intention; or someone can be making art in virtue of intending his object for regard as some particular past artworks are correctly regarded, without having in mind any specific regards; the agent can have an opaque or relational intention. The most usual case is “mixed,” involving a relational regard that is a regard in some way. On the historical/intention approach the required relation between now objects and past artworks is that of making something which is in tended for regard as previous artworks have been correctly regarded. On a historical/style approach, the relation is a linkage between the object created by the artist to some past or present general styles exhibited by some prior artworks. The relation is not achieved exclusively through intentions of the would-be artmaker. There needs to be a work with certain features, and there also needs to be the disposition on the part of the cognitive art community to regard the object as linked to prior artworks. On Levinson’s view, an individual’s intentions play a para mount role; on the style approach, intentions comprise only one of a set of considerations for correctly figuring out the style of an artwork. The style theory also provides a reason why would-be art objects are regarded as artworks: linkage to past styles. What troubles me most about the historical/intention approach is that it makes an individual’s psychological state, if it answers a certain description, sufficient for something to be art. It seems to me that there can be cases where: (a) an object has certain features; (b) it has them as a result of an agent’s action; (c) these features are put there intentionally, with the anticipation they will be regarded in one of the ways some pre existing artworks were correctly regarded; and (d) the object is not an artwork because the object does not support being regarded in the intended way. For example, imagine a young artist entranced with Marcel Duchamp and his Readymades. He also is enthralled with tall buildings and one day designates the Sears Tower in Chicago as a Readymade. He uses a spray can and writes ‘Mr. Mutt’ on the building. He regards the building in the various ways that the art community have regarded Fountain—as, for example, expressive of the idea that there is no reason to use an antiquated medium such as painting or sculpture to create visual art, or what is important in visual artworks is as much the conceptual as anything else. If necessary we can add to the story that he is also a very 284 PACIFIC PHILOSOPHICAL QUARTERLY rich Japanese investor who just bought the building. That the person lacks authority in the art community cannot explain the failure, for Duchamp himself actually tried to transfigure the Woolworth Building in New York into a Readymade. Whether something is art is not reducible to the exercise of an individual’s intentions. It seems that an artist must succeed in some way in giving features to an object, either by modifying by work or by bringing about some cultural significance for the object (as Duchamp did in Fountain) that supports the art community’s regarding it in the way some previous artworks have been regarded. (The indi viduals doing the regarding that makes the link, on a style approach, are members of the relevant cognitive community, in this context, knowledge able members of the art community.) There also needs to be a good reason for intending something for regard-as-a-work-of-art. Levinson is sensitive to the condition and sees this intuition explicitly reflected by adding another necessary condition to his definition: “so that an experience of some value be thereby obtained.” For W to be an artwork, is it necessary that its would-be artmaker intend W for regard in some way pre-existing artworks are regarded, so that an experience of some value be thereby obtained? Some art critics have not liked or admired some paintings. This has been true for Pop art and Minimalist art. That a Warhol Cornflakes Box can be important in twentieth-century art strikes some as the saddest thing one could say about the art of our time. A would-be artmaker could regard his or her output in the same way, and nevertheless succeed in producing art. VI. Some Objections to the Historical/Style Theory I will end this paper by exploring certain objections that can be raised to the historical/style theory. (1) New works are artworks because of a relation to past artworks; artworks of the recent past are art because of a relation to not-so-recent art; and so on. This string of relations ends with Ur art, objects that are artworks but cannot stand in the relation to past artworks. How can the theory account for Ur art? Carroll tries to solve this problem by maintaining that Ur Art will be identified as artworks in virtue of their function, so historical narration is not the only means of identifying artworks—though, he maintains, it is the primary means of identifying objects as artworks.^® The result is that there are sufficient conditions for arthood that are independent of the posited relation in the historical theory. And if Ur art can be functionally located, then why can’t later objects? Levinson’s general strategy to accommodate a historical definition to THE STYLE THEORY OF ART 285 Ur art is, I believe, more promising. A historical definition is put forward as an account of the nature of art. The aim is to display an extension of those things that are artworks. The definition does provide a procedure for ferreting out all examples of Ur art from, as Levinson puts it, their historical hiding places—trace them back to the initiating samples, the roots of Western art. Perhaps the tracing of a hand on a cave wall in the Paleolithic Age would turn out to be Ur art. The defini tion can generate the extension of those things that are artworks—the set of Ur art plus all the objects that stand in the posited relation. Ur art gets tagged as artworks indexically (the “originating samples”) and not because of providing aesthetic experience or some other function, family resemblance, or because of being proposed as candidates for appreciation. (2) An object can have aesthetic features independent of exhibiting a style, and this can be the reason for arthood; thus the conditions in the style definition cannot be necessary conditions. For example, an indi vidual isolated from any art community who grows up alone can, say, mark a stone in such a way that it is a representation of a recognizable tiger. This object could initiate a process that results in an art culture, and thus relative to this culture the object could be Ur art. But does the object, given its causal history, have aesthetic features? In particular, does it depict something? We can see figures and other things in clouds and rock formations, so we could see a tiger in the marked stone. Does the marked stone depictively represent a tiger? On some accounts of depiction it seems that it does. Richard Wollheim takes seeing-in as the key to pictorial depiction. From seeing figures in things such as stained walls there is, on his theory, a “smooth transition to seeing a water-mill in Hobbema’s c a n v a s . F o r Wollheim “a picture represents that which it has the power to cause a suitably sensitive, suitably informed spectator to see in its surface and which the artist marked the surface so that the spectator should see in it.” ^^ The problem I have with Wollheim’s account is that it is incom plete.For Wollheim the experience of seeing-in has two aspects: when one looks at a tiger-picture one recognizes a tiger and one is aware of the marked surface. In what sense does -a spectator recognize a tiger, since the spectator does not recognize an actual tiger? The explanation that what one recognizes are marks that resemble a tiger is not an option for Wollheim. If, however, one selected this option, an account of the resemblance involved would be necessary. Here it seems to me the only plausible account is that the style supposed for the artwork provides the standard for resemblance; consequently, the standard for resemblance varies with different styles. (3) Circularity enters the theory in the same way it does in the insti tutional theory. Reviewing the definition reveals that the word ‘art’ 286 PACIFIC PHILOSOPHICAL QUARTERLY appears on the right side of both versions of a historical definition. The historical definition is no more useful than the institutional definition. One has to know what artworks are to use it. Here Levinson has also provided a solution. The definition can be reformulated without presupposing the concept of art, but only pre supposing the past extension of ‘artwork’ at some time: W is a work of art if and only if W is an object of which it is true at t that it is related in the required way to objects in the extension of ‘artwork’ prior to t. (4) There is the intuition that there is some connection between arthood and being a good or valuable artwork. In functional theories this is a deep-rooted intuition. In the historical/intentional theory of Levinson “so that an experience of some value be thereby obtained” gets added to the definition. Where is good art connected to the historical/style theory? The style theory accommodates a weak connection between art and good art through the indispensability of aesthetic values in art classifica tion, and in the linking-regard of now objects to past artworks. (5) The theory suggests that in order to create artworks, in particular some new style or revolutionary style of art, the artists must limit them selves to variations of prior general styles. Readymades are a prototype of a revolutionary style that is different from revolutionary styles, such as the first systematic school of nonrepresentational art, Suprematism. Suprematism was just a step away from prior styles, such as Cubism and Futurism, and thus exhibited similarities with past art styles. (The style theory does not require that all aesthetic features of an object be linked to prior art.) Levinson suggests two strategies for reconciling a historical theory to radical revolutionary art, such as Readymades.These can be “translated” into the historical/ style approach. The first is that the Readymade gets regarded in the way that past artworks have, thus becoming linked to past styles; and when this proves to be frustrating and unrewarding, the public is prodded into regarding the work in a new way. Many have, in fact, regarded Fountain in the way that traditional sculpture is regarded; take Dickie’s reply to the contention that it cannot be appreciated: ''Fountain has many qualities which can be appreciated—its gleaming white surface, for example. In fact, it has several qualities which resemble works of Brancusi and Moore. The second strategy—and the one I favor—does not require that the Readymades be regarded in the way past sculptures have character istically been regarded, but that they be linked by way of contrast to or against previous art styles. Readymades get linked by being regarded partly as repudiations of past ways to do art, but also partly as retaining some features of past artworks: Readymades show that the artist can THE STYLE THEORY OF ART 287 make art without messing around with a medium as long as there are concepts and ideas that are expressed. (6) In the earlier discussion some examples of successful linkage are given. However something can be, for example, the repetition of a style and not be a work of art. For example a reproduction of a painting—say, a forgery—reproduces the painting’s style, but may not be a work of art. What distinguishes real links from such cases? What is the repetition relation such that objects that bear it to a past or present style are themselves works of art? Analogous problems can be raised for the various other dimensions for linkage. Arthur Danto argues that replications are logically styleless. A replica tion can show a style but not have a s t y l e . O n e can distinguish the relation of reproduction by copy—copying, for example, a Fauvist landscape—from the relation of repeating a style—painting a landscape in the way the Fauvists did. Usually when an artist copies a painting, say, a Fauvist landscape, this is the relation of repetition by copy; it shows a style, but it does not have a style—the copier does not paint a landscape in the way the Fauvists did. To do this involves, among other things, making use of vibrant colors for expressive purposes. (7) Sometimes artists paint copies of their own paintings. The eighteenth-century French artist Jean-Simeon Chardin from time to time painted copies of his own paintings—for example. Soap Bubbles, 1733—and may have tried to pass them off as originals. In such cases the copies not only show a style but have a style. I am Inclined to say that the Chardin case, and self-copying in general, are cases where an object both shows and has a style and is, consequently, on the theory, an artwork. Artist self-copies tend to be treated as if they were artworks. Each of the permanent collections of the Los Angeles County Museum of Art, the National Gallery, and the Metropolitan Museum has a version of Soap Bubbles. The practice of honoring and appreciating creativity from the hands of the artist is evident in connection with all the versions. Arizona State University Tempe, Arizona NOTES * In earlier papers I have proposed that the concept o f art can be defined historically. “Defining A rt,” British Journal o f Aesthetics, vol. 15, 1975, pp. 191-206 and “What is Art?” Journal o f Aesthetic Education, vol. 17, no. 3, Fall, 1982, pp. 85-92. There the attempt made was to combine the Kripke-Putnam account with traditional art theorizing. My approach suggested that ‘artwork’ is a natural-kind term. The shortcomings o f this approach have been well highlighted by Thomas Leddy, “Rigid Designation in Defining 288 PACIFIC PHILOSOPHICAL QUARTERLY A rt,” Journal o f Aesthetics and A rt Criticism, vol. 45, no. 3, Spring, 1987, pp. 263-272. In the earlier papers, being the same kind as the indexically identified samples, past artworks, is identified with the help o f art theories. On my new theory, being of the same kind is identified with the help o f art history. New objects are identified as art through the history of art rather than through art theories. 2 Arthur Danto, “The Artworld,” reprinted in Dickie, Sclafani, and Roblin (eds.). Aesthetics: A Critical Anthology, 2d ed. (New York: St. Martin’s Press, 1989), p. 177. ^ Jerrold Levinson, “Defining Art Historically,” British Journal o f Aesthetics, vol. 19, no. 3, Summer, 1979, pp. 232-250 and “Refining Art Historically,” Journal o f Aesthetics and A rt Criticism, vol. 47, no. 1, Winter, 1989, pp. 21-33. Noël Carroll, “Art, Practice, and Narrative,” Monist, 71, 1988, pp. 140-156. ^ Saul A. Kripke, Naming and Necessity (Cambridge: Harvard University Press, 1980). Hilary Putnam, “The Meaning o f ‘Meaning,’ ” in Putnam, Mind, Language, and Reality: Philosophical Papers, vol. 2 (New York: Cambridge University Press, 1975), pp. 215-271. 5 Carroll, “Art, Practice, and Narrative,” suggests that linkage is by a narration that connects the object in question to previous art practices by way o f repetition, amplification, repudiation, synthesis, or reinterpretation. 6 Arthur Danto, The Transfiguration o f the Commonplace (Cambridge: Harvard University Press, 1981), pp. 44-45. Nelson Goodman writes “style consists o f those features o f the symbolic functioning of a work that are characteristic o f author, period, place, or school.” “Style and Signature,” Ways o f Worldmaking (Indianapolis: Hackett Publishing Company, 1984), p. 35. Except for “symbolic functioning” my account is similar to his. Both accounts take style as cutting across the form/content distinction. ® James D. Carney, “Style and Formal Features,” Southern Journal o f Philosophy, vol. 29, no. 4, Winter, 1991, pp. 431-444. 9 My suggestion for what are the limiting conditions for individual style determination is found in “Individual Style” Journal o f Aesthetics and A rt Criticism, vol. 49, no. 1, Winter, 1991, pp. 15-22. There I argue that the only difference between individual style and general style is that the latter is true o f more than one individual. Kendall L. Walton, “Categories of Kri,” Aesthetics: A Critical Anthology, pp. 394-414. Walton suggests four rules to determine correct category, C, for a work o f art, W: (i) the presence in W o f a relatively large number o f standard features with respect to C; (ii) the fact that W is more aesthetically pleasing when perceived in C; (iii) the fact that the artist who produced W intended it to be perceived in C; and (iv) the fact that C is well-established in the society in which W was produced. He writes that “. . .in almost all cases at least one o f the historical conditions, (iii) and (iv), is o f crucial importance” (p. 411). All my conditions except Condition 2 are historical conditions. Only (i) is borrowed from Walton’s set. “Intentions” can be understood in several ways, so I choose “aims and goals o f the artist”—the conscious way the artist understood his or her work. Also I do not take any of the conditions as crucially important. Even those who know him are sometimes struck by how oblique Johns can be. “You ask Jasper what time it is and the kind o f answer you get might be something like, ‘I didn’t know you were interested in time.’ ” Helen Dudar, “Enigmatic, Distant, Jasper Johns Is at the Top of His Form,” Smithsonian, 21, 1990, pp. 56-68. ^2 Bradford R. Collins, in his recent reviews o f popular and new art history textbooks, is critical o f the older approach and is enthusiastic for the historical approach. A rt Journal, 48, 1989, and 49, 1990. For E. H. Gombrich’s view on the evolution o f representation see A rt and Illusion (New Jersey: Princeton University Press, 1962). For his view on 20th century art see his The Story o f Art, 14th ed. (New Jersey: Prentice-Hall, 1985), pp. 451-474. These different ways to interpret Abstract Expressionism (and others) are discussed in THE STYLE THEORY OF ART 289 Stephen Polcari’s “Abstract Expressionism: New and Improved,” A rt Journal^ 47, 1988, pp. 174-180. Here I am following the discussion o f Doug Stewart in “In Los Angeles, Billboards Are Seen as High A rt,” Smithsonian, 21, 1990, pp. 99-111. *6 This possibility is suggested by Stecker in “The Boundaries o f A rt,” British Journal o f Aesthetics, vol. 30, no. 3, July, 1990, pp. 269-271. *”7 Stephen Davies provides one o f the best accounts o f the objections that have been raised to the anti-definition approach in his Definitions o f A rt (Ithaca: Cornell University Press, 1991). He also provides illuminating and helpful accounts o f the problems for other approaches to defining artwork. What is especially relevant for this paper is that he criticizes Levinson along similar lines as I do and raises against the functional definition some o f the worries I raise (while allowing for the possibility of historical functionalism). Davies favors the institutional theory and attempts to formulate it in a way that avoids circularity. Morris Weitz in “Art as an Open Concept,” Aesthetics: A Critical Anthology, p. 159, writes that we do not need a theory to answer what is art because we rely on reasons that relate to “disjunctive sets o f nonnecessary, nonsufficient criteria.” 1^ George Dickie, The A rt Circle (New York: Haven Publications, 1985), p. 80. Also see “The New Institutional Theory o f A rt,” Aesthetics: A Critical Anthology, pp. 196-205. 20 I find that Robert Stecker in “The End o f an Institutional Definition o f A rt,” Aesthetics, pp. 206-213 has raised the key criticism o f the theory. 21 Stecker, “The End o f an Institutional Definition o f A rt,” p. 208. 22 Stecker, “The End o f an Institutional Theory o f A rt,” p. 269. 23 Carroll, “Art, Practice, and Narrative,” p. 155. 2^ Levinson, “Refining Art Historically,” p. 21. 25 Levinson, “Defining Art Historically,” p. 242. 26 Levinson, “Refining Art Historically,” p. 25. 22 Levinson, “Refining Art Historically,” p. 29. 28 Carroll, “Art, Practice, and Narrative,” p. 155. 29 Levinson writes that objects that are Ur art are “simply stipulated to be art works.” “Defining Art Historically,” p. 249. It seems that he wants no part o f indexicality except in the case where the artist’s regard is relational. He writes that “it is no part o f my view that art is anything like a natural kind.” “Refining Art Historically,” p. 31. 30 Richard Wollheim, “A Note on Mimesis as Make-Believe,'' Philosophy and Phenom enological Research, vol. 51, no. 2, June, 1991, p. 408. 31 Wollheim, “A N ote,” p. 403. 32 The incompleteness problem for Wollheim’s seeing-in account o f depiction is raised by Kendall Walton in several places, most recently in “Reply to Reviewers,” Philosophy and Phenomenological Research, vol. 51, no. 2, June, 1991, p. 425. Not having resem blance available to complete the account, Walton naturally suggests that only make believe is left. 33 Levinson, “Defining Art Historically,” pp. 241-242. 34 George Dickie, “A Response to Cohen: The Actuality o f A rt,” Dickie and Sclafani (eds.). Aesthetics: A Critical Anthology (New York: St. Martin’s Press, 1977), pp. 199-200. 35 Danto, Transfiguration, p. 204. 36 I wish to express my appreciation to the anonymous referee o f this journal who had the patience to review several stages of this paper. The referee rooted out a set o f fallacies and is responsible for the formulation of several of the objections in the last section. I also wish to express my gratitude to Thomas Leddy and my colleague Bernard W. Kobes for helpful suggestions.