Philosophy Looks at the Arts
ed. Joseph Margolis
(Third Edition, 1987)
[137]
Part Two
The Definition of Art
Aestheticians are perennially trying to define what a work of art is. The variety of answers advanced is itself worth noting, because it suggests that the question may be irregular in some way. Still, there is perhaps no quicker introduction to classical discussions of art than to summarize the master definitions that have been provided, say, from Plato to Clive Bell. It should be noticed that "work of art" is used in two entirely different ways, at the very least. For one, it is a value-laden term applied to things in virtue of certain alleged excellences. And for a second, it is one of the most basic category-terms in aesthetics, designating the principal objects that are to be examined from a certain point of view-objects possibly, though not necessarily, entitled to be characterized as "works of art" on the first use of the phrase. This distinction is often overlooked, though it would appear to affect decisively the nature of any effort to define "work of art."
The trouble with any effort to fix the basic category-term, "work of art," is that it will depend on what counts as an aesthetic point of view. But what counts as an aesthetic point of view cannot itself be decided by some simple inspection of actual usage. Philosophers seem to decide, more than to find, what the boundaries of aesthetic interest are (see Part One). And the definition of "work of art" may vary according to the varying boun- dares assigned to the aesthetic. This means that at least some apparent incompatibilities in definition may reflect the shifting decisions of philosophers regarding the definition of the aesthetic. Furthermore, as already remarked, the "aesthetic" tends to ignore the distinction of art as such- whatever we may choose to feature in speaking of art: even its history as a specialized term has only an uncertain linkage, systematically, with the nature and appreciation of art. To make the definition of "art" depend on how we specify the meaning of "aesthetic," then, is inevitably to invite a measure of conceptual disorder.
[138]
Comparatively recently, the question had been raised whether (following the lead of Wittgenstein) it is at all possible to isolate the essential and distinctive properties of works of art. It has been argued, for instance, that works of art exhibit only "family resemblances" or "strands of similarities" but not essential and distinctive properties common to all admitted in- stances. The question is a vexed one, because it is not entirely clear what sort of initial restrictions may properly be placed on the collection of things for which this claim, or the counterclaim, could be confirmed Hence, if no restriction is allowed, it would seem trivially true that no definition of the required sort could be put forward, since an expression like "work of art" is probably used in a great variety of somewhat unrelated ways. (Is a sunset a work of art, for instance? Is life a work of art? Is driftwood art?) And if an initial restriction is allowed, will it be a logical or an empirical matter that a definition of the required sort cannot be found? The question at stake is not, specifically, the definition of art but the eligibility of the effort to define art. Though if "art" be definable, we surely would want to know what the best formulation is.
Still, if "work of art" is a basic category-term, its importance probably is to be located elsewhere than in the presumed effort to discover its essential conditions (which may, nevertheless, remain legitimate). Because to identify such essential conditions is to summarize the findings of other primary investigations--for instance, the nature and orientation of criticism upon the fine arts. That is, the definition probably serves to indicate the focus of a systematic account of other questions of aesthetics more directly concerned with the professional and amateur examination of works of art themselves.
Morris Weitz's Matchette Prize essay (1955) has undoubtedly been responsible for a lively reconsideration of what are perhaps the best-known theories in aesthetics, which have regularly been advanced in the form of definitions of art. It represents a turning point in aesthetic theory precisely by raising a "'meta-aesthetic" question. But as with so many parallel efforts in other fields, it proves to be extremely difficult to separate sharply "object"-level questions and "meta"-level questions, that is, questions about the status of "object"-level questions- in particular, the question, What is art? In effect an entire industry has developed canvassing the force of Weitz's challenge.
Intuitively, it seems preposterous to deny that it is possible to say, and worth saying, what fine art is. The nature of Weitz's demurrer is rather less obvious than may at first appear. Weitz seems to press the point that if it is to be provided, a definition of art should identify what is really essential to it, what its necessary and sufficient conditions are. Apparently, a definition can be provided "for a special purpose, "but this, Weitz believes, is not the same thing.
[139]
We are bound to raise several questions here. Are there any definitions that do not serve a special purpose? Do definitions, if they serve a special purpose- for instance, facilitating reference and focus for a set of compet- ing theories about art-_fail for that reason to capture the essence of art? And is it the case that serviceable definitions must, and must be supposed to, capture the real essence of what they characterize? It is in fact a feature of the more recent, changed reception of definitional requests (for instance, in the account provided by Hilary Putnam) that we may characterize the distinctive traits or conditions of art without pretending to fix its essence or without insisting that empirical definitions be addressed to real essences.
One of the most notable recent discussions of the peculiar nature of works of art is Arthur Danto's paper "The Artworld" (1964). In effect, Danto does not disqualify definitions but rather shows why it is that, particularly with regard to art, the inventive possibilities oblige us to keep adjusting our antecedently stabilized characterizations. In doing so, Danto clearly links the treatment of the definition of art with the deeper question of the metaphysical status of a work of art. He hints at a difference between physical nature and human culture and he introduces, without development, what he calls the "'is' of artistic identification." What we may suggest, here, without pursuing that issue (see Part Three) is that definitions may serve only to fix the properties of what, in accord with the prevailing current of competing theories, are thought to be the normal or central instances. In this sense, as we say, definitions fix only the nominal, not the real, essences of things; that is, definitions are practical and alterable instruments servicing developing theories that cover at least certain undeniable specimens. It is entirely possible, therefore, that works of art that are somewhat deviant relative to the standard cases can be admitted as works of art and can be admitted to have properties quite different from those focused by our definition, without contradiction at all. It is only when such discrepancies begin to take on a systematic importance that earlier serviceable definitions will have to give way: then, the meaning of "art" will have to change, whereas it need not have changed in accommodating our changing beliefs about what a work of art could be like. This explains in a sense the tolerance that is possible in speaking of driftwood, readymades, l'art trouvé, machine art, the art of chim- panzees, and so-called conceptual or idea art. It also suggests the sense in which we may proceed by genus and difference, by necessary and sufficient conditions, or by characteristic conditions, without in the least violating any logical constraints on the formulation of empirical definitions. The definitions of empirical terms in the sciences are intended to facilitate discourse that is primarily explanatory (in the causal sense) and predictive. Where art is concerned, no such constraints obtain, though there is a clear sense in which definitions must be brought into fair agreement with the Pai
[140]
general sorts of theories and activities that characterize our aesthetic concern with the arts themselves. The concession, it may be admitted, points as well to the possible vacuity of definitional disputes. But, more important, it signifies a refinement in the theory of definition and a sense of the function and validity of particular definitions of art.
Danto has, more recently, pursued (1981) some of the intentional complexities of art, characteristically ignored in aesthetic theory particularly of the Anglo-American sort. It would not be amiss to say that, as with the "Artworld" paper, he is benignly influenced, here, by a reading of Jean-Paul Sartre' phenomenologically oriented account of imagination (1984). Nevertheless, he does not return to clarity the "is" of artistic identification.
Sartre himself, of course, has--in a deliberately paradoxical but also conceptually troublesome way-raised art to a higher status than the "merely" real by dubbing it "unreal." (This use of "mere" is itself not merely accidental in Danto's papers on art.) Sartre's intention, here, is to dramatize the intentional, imaginative complexity of what is said to be represented in painting and the other arts. But his account obliges us, admitting his generic thesis about art, to consider what we may usefully say about the definition of art. And Danto, in a lively and even more recent paper, "The End of Art" (1984), has usefully collected the histori- cally unfolding record of the art tradition in a way that confirms the conceptual puzzle of what we take ourselves to be doing in defining art-_both with respect to his claim that art has "become transmuted into philosophy" and that the boundaries between the various arts have "become radically unstable." Still, what we should understand by the definitional question remains for others to examine (for instance, Robert Matthews, 1979), particularly if there is no final or essential theoretical discovery about the nature of art (see for instance Francis Sparshott, 1982).
Nelson Goodman (1977) has, for reasons rather different from Danto's, eschewed the definition of art. In speaking of the "symptoms" of art as opposed to its "defining properties, " however, he does actually consider a set of disjunctively necessary and conjunctively sufficient conditions that tempt us--against his demurrer--to view his thesis as definitional. These are linked to symbolic functioning; to features of symbols, to features sym- bolized, and to functioning at least by way of exemplification. One may perhaps say that, effectively, Goodman favors a functional essentialism with regard to art rather than a substantive essentialism--which is both subtler than the latter and also not entirely congruent with his own more recent commitment to a proliferation of "made" worlds. But if he were to concede that much, artworks might be shown not to exhibit the salient
[141]
symbolic functions he favors and yet remain significant artworks for all that, Goodman might well be obliged to curb considerably his semiotic orientation, both in general and in detail. The theory that works of art are symbols or symbolic forms is perennially of interest, of course. Goodman introduces a particularly subtle version of it, by means of which he is able to expose certain anomalies in formalist theories of art. But in doing so, he obliges us to consider once again how minimal the basis may be on which to construe an object as functioning symbolically —hence, the vulnerability of the theory that art exhibits in some strong sense a symbolic function.
My note says:
Perhaps both
of them have a point.
The key to Goodman's entire theory of art rests with the concept of exemplification. Granted that to exemplify is to symbolize; the question remains whether in possessing whatever properties it does possess (non-symbolic properties), a work of art must be said to refer to or symbolize any or all such properties. Goodman's account, therefore, suggests a possible asymmetry between validating and invalidating definitions of art or, more informally, characterizations of what serves as strong (or even decisive) evidence of the presence of art. It also suggests the variety of strategies by which we may recover some of the purposes of definition without conceding that the specification of "salient," "characteristic," "symptomatic,' » and similar traits are to count as actually definitional in nature.
Jack Glickman (1976) adopts another strategy. Instead of challenging what is normally thought to be a defining property of art, namely, being an artifact, Glickman attempts to show that being an artifact does not entail having been made by anyone. He explores, in passing, the notion of creativity and centers on the kind of art that is thought to be creative in nature. By this strategy, Glickman shows the possibility of dispute even about what is entailed by admittedly defining properties. And his claim that "particulars are made, types created" draws us on to the profound ontological problem of how a work of art can be a particular and combine concrete and abstract properties (see Part Three).
Perhaps the so-called "institutional" theory of art advanced by George Dickie (1969) is the most recently debated of standard attempts at defining art. Ted Cohen's criticism of it (1973) both introduces the thesis in a convenient and fair-minded way and musters the principal objections to it. There is no question that some sense of the institutional or societal complexities of art is essential to our theorizing. But Dickie's thesis has puzzled his readers primarily because, as Cohen very clearly shows, we cannot sort out satisfactorily whether the prior achievements of some putative artwork are marked as such by a knowledgeable clientele or whether some (somehow) authorized public body fixes the pertinent properties of a would-be artwork by selecting some artifact for artwork Perhaps bitt of them have a point
[142]
status. Questions both about how to legitimate those who legitimate the art status of artifacts and about the dubious (and parasitic) function of status-fixing itself have been raised against Dickie's view (for instance, by Beardsley, 1982; and by Wartofsky, 1980). Cohen's own emphasis rests chiefly with the performative feature of Dickie's defining conditions: both with respect to "conferring" (in Dickie's account) a status that must already obtain before the would-be enabling act itself and with respect to the difficulty of supposing that relevant conferring moves need be (or can be) directly linked with what Dickie has in mind in speaking of apprecia- tion. But all these challenges pale somewhat in the face of the fact that we generally lack a sustained account of the institutional, societal, and historical nature of art itself.