Many of the earliest definitions of art were probably intended to emphasize salient or important features for an audience already familiar with the concept, rather than to analyse the essence possessed by all art works and only by them. Indeed, it has been argued that art could not be defined any more rigorously, since no immutable essence is observable in its instances. But, on the one hand, this view faces difficulties in explaining the unity of the concept – similarities between them, for example, are insufficient to distinguish works of art from other things. And, on the other, it overlooks the attractive possibility that art is to be defined in terms of a relation between the activities of artists, the products that result and the audiences that receive them.
Two types of definition have come to prominence since the 1970s: the functional and procedural. The former regards something as art only if it serves the function for which we have art, usually said to be that of providing aesthetic experience. The latter regards something as art only if it has been baptized as such through an agent’s application of the appropriate procedures. In the version where the agent takes their authority from their location within an informal institution, the ‘artworld’, proceduralism is known as the institutional theory. These definitional strategies are opposed in practice, if not in theory, because the relevant procedures are sometimes used apart from, or to oppose, the alleged function of art; obviously these theories disagree then about whether the outcome is art.
To take account of art’s historically changing character a definition might take a recursive form, holding that something is art if it stands in an appropriate relation to previous art works: it is the location of an item within accepted art-making traditions that makes it a work of art. Theories developed in the 1980s have often taken this form. They variously see the crucial relation between the piece and the corpus of accepted works as, for example, a matter of the manner in which it is intended to be regarded, or of a shared style, or of its being forged by a particular kind of narrative.
If the purpose of a definition of art is to facilitate the unequivocal identification of items as art works, then it should characterize a property, or some combination of properties, displayed by each and every art work and belonging exclusively to art works, that is, a feature or set of features marking all art works and only them. Such a definition is called ‘real’ or ‘essential’ (see Definition ); it specifies one or more necessary conditions which in the combination indicated are together sufficient for anything to be of the kind in question.
Definitions can serve goals other than that of unequivocal identification and they need be no more rigorous than is required by the chosen purpose. We might look to a definition simply for the sake of knowledge; for instance, in seeking a precise and systematic catalogue of things. We might aim, alternatively, to teach the meaning of a term and will use such definitional methods as are adequate to achieving that end – ostension, enumeration, dictionary meanings, reference to paradigms. We could wish to prescribe a new meaning or use for a term by an act of stipulation, either for a special purpose (in which case the definition is sometimes called ‘operational’) or in order to change its meaning altogether (as in revisionist definitions). We might be concerned, instead, to characterize a thing’s typical features, or the properties that are significant in our use of that thing; these emphases might result in partial definitions drawing attention to non-essential features. As is apparent from this list, sometimes the task of definition is purely descriptive and at others it is regulative; sometimes it is concerned with the way the world is and at others with linguistic practices; sometimes it is controlled by our interests and at others is largely independent of them.
2 Early definitions of art
Many of the famous theories of art offered in the past – Plato’s conception of art as mimesis (imitation or representation), Tolstoy’s view of art as the communication of feeling, Clive Bell’s account of art as significant form – fail very obviously when treated as real definitions. If the key notions are construed so broadly that all art works cannot help falling under them, these notions are also bound to cover many things that are not art works. If the central terms are read narrowly, then, they still seem certain to apply to some things that are not art, as well as not applying to some pieces that generally are agreed to be art. It is best to treat these views as recommending fruitful approaches to art’s interpretation, or indicating art’s more salient or valuable features, rather than as real definitions. Indeed, this is the spirit in which most were offered. These theories are addressed to an audience already skilled in the identification of art, and take that common understanding for granted.
Is a more rigorous approach to a definition of art possible? Morris Weitz (1956) has famously and influentially argued that art has no fixed essence and, hence, that no real definition of art can be successful. He notes that when we look, we find no property common to all works of art. Art-making is creative and, hence, inevitably defeats the definer’s attempt to congeal what is a fluid process. Weitz explains the unity of the concept of art with the idea of a network of ‘family resemblances’, a notion he adopts from Wittgenstein’s Philosophical Investigations (see Wittgenstein, L. §§10-12 ). Works of art are appropriately grouped together in terms of similarities that link them, though there is no single feature or set of qualities shared by all.
Weitz’s positive view faces serious difficulties. Similarity could not provide a basis for recognizing the first works of art, since these had no artistic forebears that they might resemble. Nor does the appeal to similarity explain the status of more recent art works. Some art works, such as ready-mades or representational works, more closely resemble things that are not art than they resemble art works; for instance, art films are more like TV ‘soaps’ and home videos than sculptures. A counter to this objection might insist that only relevant numbers, kinds, or degrees of similarity are significant in establishing the classification of things as art. To enumerate and clarify the types of resemblance that count towards something’s being a work of art is to return to art’s definition, however, for one would have to specify the set of similarities that are necessary and sufficient for something to count as art. Weitz’s reliance on the notion of resemblance does not replace the need for definitions of the type he declared to be impossible.
On the face of it there is no significant property perceptible in all art works. If so, this counts against the kinds of theories Weitz was keen to attack, namely, those proposing that art might be defined in terms of shared aesthetic properties, these being conceived as qualities revealed directly to the senses. But it is not clear that Weitz has demonstrated the impossibility of defining art, for the relevant properties might be imperceptible. (One cannot distinguish uncles from other males merely by examining their appearances, but this does not show that the idea of an uncle is indefinable.) It is plausible to expect that some complex, imperceptible relation between creators, the things they make and the audience that receives them will lie behind a definition of art. Hence, even if Weitz is correct in claiming that we do not see a property common to all art works, this does not show art to be indefinable.
What of Weitz’s further claim – that a real definition of art will be refuted and repudiated by artists’ creativity? Again, the claim appears plausible only when directed against definitions holding that art works must possess aesthetic qualities (given a limited set of these). A definition relating artists, their products and audiences might easily accommodate innovative kinds of art, because it emphasizes the context of creation and reception rather than the constitution of the piece involved in this transaction.
I have suggested that neither Weitz’s arguments nor the fact that most adults have a secure grasp of the concept shows the irrelevance or impossibility of defining art. Moreover, there is an obvious need for such a definition, since the claim to art-status of many pieces created in the twentieth century is hotly debated. Some artists have deliberately produced works that challenge the border between art and non-art, provoking the question ‘But is it art?’ If we could define art we would have a means of resolving disputes about ‘hard cases’ of this sort. And even if the attempt to formulate a correct definition is likely to remain controversial, we might come to a deeper understanding of art and its context through the pursuit of such a definition. While Weitz’s arguments have been influential and the impossibility of defining art is still asserted, the number of publications presenting new definitions indicates that reports of the death of the enterprise have been greatly exaggerated.
It might be said that it is not so much for us to discover whether things are works of art as a result of applying to them an independent standard captured in a definition but, rather, to decide whether they are art. I regard this response as misguided. As an aspect of culture, the nature of art is socially constructed and historically malleable, depending on human interests and judgments. If the nature of art is relative to, and affected by, human concerns and practices, this will be mentioned in an adequate definition; such a definition could play a role in settling the appropriateness of our deciding a particular hard case in one way or another. Even if it is for us to decide whether something is art, it does not follow that that decision can be entirely arbitrary, for there must be a difference between our coining an additional meaning for an old term and our resolving that some controversial case is to be properly grouped with undisputed paradigms under the same conceptual umbrella.
3 Functionalism and proceduralism
Many definitions offered in recent decades can be classed as functional or procedural. Functional definitions give centrality to the necessary condition that works of art serve a purpose or purposes distinctive to art, whereas procedural definitions stress that they are created according to certain conventions and social practices. A composite definition mentioning both of these necessary conditions, as well as others, is possible. In practice, though, these two kinds of definition oppose each other, because the procedures by which the status of art is usually conferred have been used to create pieces that fail to serve functions traditionally met by art. Indeed, items may be presented as art, though they have as their point the goal of opposing the attempt to appreciate them in the orthodox fashion. Some functionalists offer their definitions with the goal of excluding such pieces from the realm of art, whereas proceduralists aim to include them. These approaches also differ concerning the connection between something counting as art and its having artistic value. Functionalists see the possession of a degree of aesthetic value, measured in terms of an item’s success in fulfilling one or more of the functions of art, as essential to its qualifying as art, while proceduralists regard the artistic evaluation of a thing as separable from the determination of its status as art. The proceduralist’s definition is purely descriptive, having little to say about the significance of art or about the reasons that might lead someone so authorized to confer art-status on one thing rather than another. By contrast, the functionalist’s definition is normative.
Monroe C. Beardsley (1982), a functionalist, characterizes an art work as either an arrangement of conditions intended to be capable of affording an aesthetic experience valuable for its marked aesthetic character, or (incidentally) an arrangement belonging to a class or type of arrangement that is typically intended to have this capacity. A more recent version of functionalism is given by Robert Stecker (1994 ), according to which an item is a work of art at time t if and only if either (a) it is in one of the central art forms at t and is intended to fulfil a standard or correctly recognized function within the set of central art forms at t or (b) it is an artefact that achieves excellence in fulfilling a function belonging to the set of functions for central art forms (whether or not it is in a central art form and whether or not it was intended to fulfil such a function).
Among the tasks and difficulties faced by functionalist accounts are as follows. (1) Specifying the functions of art. For Beardsley, the main purpose is that of providing an aesthetic experience. (2) Acknowledging both that the point of art might alter through time and that the art-historical context of creation affects the aesthetic character of the work and, thereby, its functionality. The historicism introduced by Stecker’s time-indexing is designed to cover such considerations. (3) Explaining the dysfunctionality of very poor works of art. Both Beardsley and Stecker do so by allowing that something intended to serve the point or points of art might become an art work even if that intention is unsuccessful. (4) Resolving the status of the hard cases mentioned previously. Beardsley denies that Duchamp’s pieces are works of art, whereas Stecker argues that, within their art-historical setting, they serve accepted functions of art (reference to and rebellion against former artistic types and practices). He notes that they could not have served equivalent functions in earlier times.
The most detailed version of a procedural account is the institutional theory developed by George Dickie. His most recent definition ( 1984 ) runs: (a) an artist is a person who participates with understanding in the making of an art work; (b) a work of art is an artefact of a kind created to be presented to an artworld public; (c) a public is a set of persons the members of which are prepared in some degree to understand an object which is presented to them; (d) the artworld is the totality of all artworld systems; (e) an artworld system is a framework for the presentation of a work of art by an artist to an artworld public. The ‘artworld’ is the historical and social setting constituted by the changing practices and conventions of art, the heritage of works, the intentions of artists, the writings of critics, and so forth.
Among the difficulties faced by the proceduralist are as follows. (1) Showing that the relevant procedures are established (and, in its institutional version, demonstrating that they mark an informal institution distinguishable from similar institutions with different goals). (2) Accounting for the art-status of works never presented to, or intended for, a public, including the products of isolated artists, of the earliest artists in history and of those working outside the officially recognized boundaries of the artworld, such as embroiderers. Dickie’s definition requires not that the piece be presented, but that it be of a kind suitable for presentation; also, he could allow that some pieces are enfranchised as art from within the institution after their creation. (3) Avoiding a vicious circularity in characterizing the procedures, or the institution in which they are applied, without assuming their products to be art works. Dickie claims that the circularity in his own account is benign. (4) Resolving the status of the hard cases mentioned previously. Dickie sees it as an advantage of his theory that it accommodates Duchamp’s ready-mades, but one might wonder if the procedural account is able to explain what makes such cases hard.
4 Recursive definitions
Weitz’s suggestion that something is a work of art in virtue of its resemblance to other (prior) works of art indirectly acknowledges the historicist character of art-making. Artists frequently draw on, refer to or react against their predecessors. Moreover, what constitutes art and what can be done within art depends on what has been art and what has been done within art in the past; the art of the distant past of a culture might differ in many respects from the art of its present, despite the continuity of the process that links one to the other. The historicist character of art has received growing recognition within philosophical aesthetics since the 1950s; more recent attempts at a definition reflect this.
In crude outline, a historicist definition of art has two parts. The first explains how the first works in history came to be art – perhaps by stipulation, or because they served an appropriate function. The second, recursive part states that ‘Something is an art work if it stands in an appropriate relation to art that predates it.’ The ‘appropriate relation’ is characterized in various ways. A suitably historicized functionalist definition, for example, would construe the relation as holding between the (intended, central, significant) function of the present candidate and the (intended, central, significant) functions of past works. A suitably historicized proceduralist definition would construe the relation as holding between the procedures applied to the present candidate and the procedures used successfully in conferring art-status on prior works. (I have already noted the historicist aspect given to functionalism by Stecker. The institutional theory is ripe for and would be improved by a similar treatment.)
Some recent historicist definitions conceive the defining relation neither in terms of function nor procedure. Jerrold Levinson (1979) sees the defining relation in the intended treatment of the candidate – a work of art is a thing that has been ‘seriously intended for regard-as-a-work- of-art’; that is, regard (meaning treatment, taking, engagement with or approach) in any way pre-existing works of art are or were correctly regarded. James Carney (1991) characterizes the defining relation as a shared style: an object is a work of art if and only if it can be linked by those suitably informed, along one or more various specific dimensions, to a past or present general style or styles exhibited by prior works of art. Noël Carroll (1993 ) takes the unifying relation to be that of narrative continuity, though he denies offering this as a definition. In his view, something is an art work if it can be linked to preceding art-making practices and contexts by a narrative committed to historical accuracy that reveals the piece as an intelligible outcome of recognizable modes of thinking and making of a sort already commonly adjudged to be artistic. If there is dispute about the artistic nature of the context from which the candidate work arose, then this is to be settled by appeal to a meta-narrative that links that context with acknowledged artworld practices, procedures and processes.
The detail of each of these theories might be examined critically. For instance, one might ask if Levinson can distinguish the art-making intention from other intentions that similarly invite a regard of something as if it were art without aiming, directly or indirectly, at making that thing art; and one might consider whether Carney could analyse the notion of artistic style, or Carroll could develop the relevant notion of continuity in narrative, without begging the definitional question. (Of course, a theorist might avoid such queries by further generalizing the recursive part of the definition – something is a work of art if and only if it stands in the appropriate art-creating relation to previous works. This approach meets these objections, though, only by emptying the definition of content.)
Instead of pursuing such matters here I will mention one concern about the general strategy. It seems that there is more than one tradition of art-making and appreciation; also, what is possible at a given time within one tradition might not be possible at the same time, or at any time, in others. Recursive definitions explain how something is art by relating it in the appropriate way to a given tradition. Such definitions will be at best incomplete, because so much of the explanatory burden is carried by the implicit, undefined notion of an artistic tradition. If something is a work of art within only one of many possible traditions, then the notion of art is not fully explicated until a basis is provided for distinguishing traditions of art from other historically continuous, cultural processes or practices and, also, for individuating one artistic tradition from another.
Two ways of attempting to dismiss this point fail, I think. First, it would be both false and offensive to confine art to a single cultural tradition, such as that arising from western Europe, and to dismiss other traditions merely as generating non-art that serves functions similar to those of art. And even if we allow for the many human artistic traditions, it might be implausible to reject the possibility of non-human, non-terrestrial art. Second, it would be an error to suggest that the proposed definition allows that something is a work of art if it relates appropriately to any pieces in any tradition of art-making, for the work then becomes decontextualized. This is unconvincing because it implies that, if something could become art within one tradition, it could become art in any; if Duchamp could make a work of art of a urinal in USA, a Chinese artist might have done the same in China. Rather than emphasizing that the art status of a piece depends on the piece’s historico-cultural location, this approach treats the place of the piece in its given tradition as irrelevant to its status as art.