Skip to main content

Art, Ontology of


Ontology is concerned with what exists. So one may think the ontology of art is concerned with whether artworks exist. However, most people take the existence of artworks for granted. (See Dilworth 2004 for someone who does not.) The main issue for the ontology of art is what kind or kinds of objects artworks are. A second important issue is about the identity and individuation of works. Concerning both of these issues there is wide disagreement along a variety of parameters.

Objects that Are Artworks

one kind or many

One parameter along which there is disagreement is whether all artworks belong to a single kind or whether they belong to irreducibly different kinds. The second view seems more plausible, at least initially. A painting, such as one made with oils or watercolors, is an entity that has physical properties, such as spatial dimensions, that exists in a single place at a single time, and, for these reasons, may be plausibly taken for a physical object. A novel could be said to exist in many placeswherever there is a copyor in no place, because no copy or even the original manuscript is the novel. For this reason, novels could not be physical objects. One of a kind sculptures are more like paintings in the respects mentioned above, whereas many musical works are more like novels.

Nevertheless, there are a variety of attempts to argue that artworks belong to a single kind. One strategy for doing this is to argue that all works are types or kinds of some sort, thereby assimilating those, such as paintings, that appear to be physical objects to the category to which novels and musical works more obviously belong. One proposal is that all artworks are structural types. Musical works are sound-structures and literary works are linguistic structures (or possibly, in some instances, plot structures). Paintings also have a structure that could be defined in terms of patterns of colors and shapes, or defined in some other way. This structure is duplicated in a copy of the painting, perfectly duplicated in a perfect copy. Prints and sculpture produced from a model seem to fit this proposal better than paintings, because people currently recognize that prints and sculpture that share a common structure belong to a single work. Current practice does not do this for paintings, no matter how perfect the copy. One can imagine a future time when painters produce a work in two stages. First they paint something. Second, they authorize a certain number of mechanically produced copies to be housed in several different museums or galleries as instances of the work, just as there are now several authorized instances of Henry Moore's sculpture King and Queen on different sites in different parts of the world. However, the possibility of imagining this new practice does not show that paintings are really abstract structures. If anything it shows the opposite, because the imaginary practice stands in stark contrast with the actual one. This actual practice does not recognize copies as instances of the work, asserts that the work is deteriorating when the paint applied by artist to canvas deteriorates despite the existence of good copies, and so on. Because painting and some sculptures are not structural types, there are other works that also do not fit the proposal, even though they are not physical objects. Improvisations and happenings are nonrepeatable events. So the strategy of arguing that all artworks are abstract structural types fails.

These considerations also speak against a second proposal: that artworks are action-types (Currie 1988). The type is the discovering of an abstract structure in a specific way (a "heuristic path"). The proposal recognizes a consideration that is discussed at length below: a work's pattern or structure and the context of its making are distinct sources of important artistic properties. However, if the reasoning of the last paragraph is correct, the present proposal has a defect similar to the first proposal in misidentifying the sort of objects that paintings and uncast sculptures are. These are not types of achievement; rather, they are specific concrete objects that are appreciated only in part for what they achieve. Even genuinely abstract works seem to be objects brought about by a type of activity rather than that action-type itself.

A different strategy for arguing that all artworks belong to a single ontological category is to argue that they are all concrete objects of some sort rather than abstract objects. One proposal on the table that fits this approach claims that all artworks are action tokens, in particular, the creative activity of artists that bring into existence those objects normally thought of as works of art (D. Davies 2003). On this proposal, the actual work is uniformly the creative activity, the product of that activity being dubbed the work-vehicle and distinguished from the work itself. One may wonder what this accomplishes other than a renaming. Both the creative act and the object are recognized by everyone, and as such, no novel entity is involved in the act-token conception of artworks. So why depart from normal practice and assert that the artwork is not the object produced by the artist's activity, but is the activity itself? Simplifying a complicated argument, the main reason is the importance for art appreciation of reconstructing the artist's creative activity. The claim is that the only way to acknowledge this importance fully is to identify the work with the activity. This claim is unjustified. An object has many relational properties in virtue of its origin and recognizing these properties may be crucial to fully appreciating the object as an artwork. People can accord recognition to the artist's creative activity by understanding that it is in virtue of the creative act, and of the project that gives rise to itthat the work (object) has the relational properties crucial to appreciating it. There is no need to identify the work with the creative activity itself. Hence the renaming that the act-token view proposes is neither necessary nor desirable.

One may conclude that the heterodox view that artworks belong to irreducibly different kinds is not only more plausible initially, but more plausible after reflection as well. Taking this for granted, the next question is how to more sharply define these kinds.

the role of intentions and context

According to the heterodox view, some artworks are abstract types or kinds, others are concrete objects with physical properties, and there are still others that are particular events or processes. One individuating feature of abstract artworks, such as novels, plays, and pieces of music, is that instances of each work share a common structure. Is this sharing of a structure sufficient to individuate a single work? It is clear that this is not always so. Consider the case of a sculpture that has multiple instances. The structural element here consists of a material such as bronze being shaped in a certain way. Wherever there is a piece of bronze so shaped there is an object that has a structure in common with the sculpture. But this is clearly not sufficient for the object to be an instance of the sculpture. Someone who produced pieces of bronze with shapes identical to those belonging to King and Queen would not thereby produce an instance of that sculpture. For the pieces to be such instances, they would have to be produced from the cast Moore supplied to a certain foundry chosen by the artist and be one of a specific number of instances as indicated by Moore. This much is obvious. Controversy arises when one asks why this is so and whether a common structure is equally insufficient to individuate musical and literary works.

One explanation of the insufficiency of structure to individuate works such as cast sculptures and prints appeals to a purported distinction between autographic and allographic works (Goodman 1968). The latter are those that, because they are made in a notational symbol system, are in fact individuated by a shared sequence or structure of symbols. A thought experiment suggesting that musical and literary works are allographic relies on the impossibility of forging a musical or literary work by copying the score of one or the sequence of words of the other. This simply would produce the score of the musical work or a copy of the novel rather than something to be passed off for one of these. If, however, someone copied a cast sculpture by creating a new cast that produced a piece of bronze identical in shape to the sculpture, that would be a forgery. To be the sculpture, even one that has multiple instances, each instance must derive in the right way from the hand of the artist. This is what makes sculpture an autographic art form.

One can accept a version of the autographic/allographic distinction that simply says that some works are made in notational symbol systems and others are not. What this version of the distinction does not imply is that if a work is made in a notational system, it is individuated entirely by notational structure. A different thought experiment suggests that even for works in notational systems, instances sharing a common structure are not necessarily the same work. The experiment revolves around structurally identical items from different periods or cultures. Because of the different historical contexts, the items will have different artistic properties despite sharing, say, the same sequence of words. A well-known version of this thought experiment is the often-cited story by Jorge Luis Borges, "Pierre Menard, Author of the Quixote" (1970). In this story, Borges imagines a late-nineteenth-century writer, Menard, who produces a manuscript word-for-word identical to some chapters of Cervantes' great novel. Borges plausibly proceeds to note the huge differences in style and meaning between the two works. Cervantes' style is colloquial, whereas Menard's is self-consciously archaic. The latter contains allusions to contemporary philosophic thought that the former could not possibly have. Hence even ignoring that Menard's text is identical to only a small part of Cervantes', the two are different works in virtue of different authorial intentions and contexts of creation.

Once one recognizes that intentions and context play roles in individuating works that have the ontological status of types or kindswhether or not they are also "allographic" in the weak sense noted in the preceding paragraphone can also recognize that intention and context play similar roles in the case of concrete works such as paintings and uncast sculptures. To recall another famous example from Arthur Danto's The Transfiguration of the Commonplace (1981) consider three pieces of canvas covered with red paint by the hand of three different, independent artists. Three "structurally identical" red canvases could form parts of a single work of art, a triptych, if produced with that intention by a single artist or a group working together. That three distinct physical objects are produced in isolation from each other, the product of three different "hands," implies that, if each red canvas is or constitutes a work of art, there are three distinct works. However, when does a red-paint-covered canvas constitute a work of art, and what sort of entity is the art object so constituted? The answer to the first question once again appeals to intentions and context. For a canvas uniformly covered with red paint to be a work of arta paintingan art-historical context must be in place that permits certain intentions to count as art making ones. In eighteenth-century France such institutions were not present, whereas in twentieth-century America (or France) they were. Second, the art-making intentions must actually exist. If one canvas became red because the artist needed an empty spray paint can and got it by discharging the paint onto this canvas, there is no artmaking intention and no artwork. If the canvas became red as the result of an intention to produce a work in the color-field genre, then there is an art-making intention, and thus an artwork.

The second, strictly ontological, question asked above is: What sort of entity is the art object? Is it the painting that results from covering the canvas with red paint? Is it identical to the paint-covered canvas or not? To grasp the ontological puzzle here, it is easier to turn to a different example: a piece of clay shaped into a human figure. Is the sculpture identical to the lump of clay? Obviously not, because the lump existed before it was shaped to create the sculpture, but the sculpture itself did not exist. An alternative answer to this question is that the sculpture is the human-shaped lump. This entity came into existence when the sculpture did and could be regarded as a phase of the existence of the lump itself. However, even such a "phase" or "time-slice" could be understood as having its shape contingently. That is, if it is possible for it to continue existing as one and the same phase of the lump while radically changing in shape, the phase is not identical to the sculpture because the sculpture would not survive such a radical change in shape. Also, if it is possible for this entity to come into existence independently of any human intention, it could not be identical to the sculpture. A sculpture cannot come into existence exclusively through natural processes. These considerations imply that the entity identical to the sculpture is not simply a lump of material structured in a certain way but such a lump structured to fulfill an artistic function or intention typically made possible by certain institutions or practices. Exactly the same is true of the red painting. It is a canvas covered by red paint to in order fulfill an artistic function or intention made possible by certain institutions or practices (Levinson 1996).

are all artworks created?

Concrete artworks such as the red paintings and the clay sculpture just discussed are obviously created. Are the abstract worksthe novels, musical pieces, and so onalso created? Some scholars, such as Jerrold Levinson (1980), have argued that it is a condition on a satisfactory ontology of art that the ontology accounts for the createdness of abstract artworks. Others, such as Peter Kivy (1993) and Julian Dodd (2000, 2002), have disputed this. Underlying these conflicting views are conflicting intuitions. One intuition is that novels, plays, and musical works are just as much the products of creative activity as are paintings. The other intuition is that abstract objects cannot be created because of the sort of objects that they are.

It may seem that the argument of the preceding section supports the claim that abstract artworks are created. It was argued that these works are not identical to abstract structures per se, but to structures tied to certain intentions and contexts. What could "tied to" mean but created with certain intentions in a certain context? But this raises an important question: What are these entities that are purportedly created? They are not pure abstract structures, because these are really uncreated and eternal, and it has already been denied that they are the artworks. The best known proposal on this score is Levinson's. He claims that they are indicated structures "a structure-as-indicated-by-P-at-T-in-[art]-historical-context-C" (1996, p. 146). The dashes are intended to indicate that this is not a set-like ordered quadruple but something more "unified," a type that comes into existence with the act of musical or literary composition.

There are a variety of objections to Levinson's view. Stefano Predelli asserts that indicating does not in general create new entities (1980). If I point out my favorite house in the neighborhood, I haven't created a new entity: the-house-as-indicated-by-me. So it is implausible that indicating creates one when authors or composers indicate abstract structures. There are two ways of replying to this objection. One reply would be to claim that new entities are always created by indicatings, but people pay no attention to most of them because the indicatings are of no interest them. The house-as-indicated-by-me is an entity that has about as much interest as a scattered object such as a nose-tie consisting of Bill Clinton's nose and a tie he left in a hotel during a visit to Australia. Both nose-ties and indicated-buildings nevertheless exist. The other reply claims that some indicatings are special because they occur within institutions or practices that endow them with special properties and give them special recognition. Sentences can be regarded as abstract syntactic structures, which, when used (when indicated by a speaker or hearer) creates a new entity, which has semantic or pragmatic properties not possessed by the abstract sentence type. The ability to convey something distinct from the semantic meaning of the sentence type results from linguistic conventions combined with the intentions of language users and the context of use. Writers are just special cases of people who use (indicate) strings of sentence types to convey something through the creation of a complex literary object. Composers do something similar with abstract sound structures. The two replies are consistent with each other, though the second is available to those who would resist the first.

A second objection is that abstract entities, such as structural types, cannot be created because they cannot enter into causal relations. Being created means being caused to exist and, if an entity cannot enter into a causal relation, it cannot be caused to exist. This claim is said to apply to any abstract type whether it be of the pure unindicated variety or an indicated structural type. A related third objection should also be mentioned at this point. It could be said that even if there are indicated types, they are just as eternal and uncreated as any other abstract thing. Such types exist just in case a property corresponding to the type does, and all properties exist eternally. Hence the property of being a structure indicated by P at T in C exists eternally. Therefore the indicated structural type does too. Hence it is not created (Dodd 2000, 2002).

Both of these objections are too tendentious to be decisive. The issue of whether types can be caused to exist is not settled by their being abstract; they are abstract because they have instances or tokens. Someone could claim that a type does not exist until at least one token of it does, or instructions for creating a token are present. In either case causing the token (or the instructions for making tokens) to exist in effect causes the type to as well. There are many types that it is plausible to conceive in these terms. Consider artifact types, one example of which is an automobile model. It is plausible that automobile designers bring this type into existence when they create the design for a car model. This plausible claim is deniable. It could be consistently maintained that the type Volkswagen Beetle would exist even if intelligent life had never evolved anywhere in the universe. Though consistently maintainable, the claim is implausible. Saying it is tendentious is perhaps an understatement. If this is true for car models, it would be equally true for literary and musical works. So one may perhaps set aside the second and third objections to the idea that indicated structural types are a kind of entity that can be brought into existence.

Individuating Artworks

What has been demonstrated thus far is that indicated structures are distinguishable from unindicated ones, and that the idea that they come into existenceindeed, are brought into existenceis, at least, plausible. However, there is a final set of objections to them that question whether they individuate musical and literary works correctly. Are such works always essentially tied to the precise times they are created, to their creators, and to their context of creation? This is what is denied by the final set of objections.

Look first at authorship. It may be true that Cervantes and Menard (had there been such a person) could not possibly create the same work. But imagine two contemporary writers, composers, or even painters who belong to the same school working at the same time. There are two different scenarios to consider. One occurs when both produce identical works. Suppose Mozart and Haydn had produced, independently of each other, identical scores for a string quartet in the year 1787. Would they both have independently composed the same work? An alternative scenario can be created by supposing a possible world in which Haydn instead of Mozart composed a score identical to the score for Mozart's G major quartet K.387 and in which Mozart produced no such score. Would Haydn have composed in this possible world the same quartet that Mozart composed in actuality? Some people would answer yes to both of these questions; but, if that answer is right for either one, the identity of the artist may not be essential to the identity of the work. The first scenario does not raise a problem when it comes to painting because two numerically distinct painted canvases from the hand of different artists are different paintings even if they are indistinguishable. The second scenario, however, raises the same question for painting as it does for music or literature. In a possible world in which Braque rather than Picasso had painted a portrait of Gertrude Stein exactly similar to Picasso's actual painting, would Braque be the artist responsible for the work Picasso actually made (Currie 1988, S. Davies 2001)?

Something similar can be said about the time at which the work is indicated or brought into existence. Is this always an essential property of artworks? Some works seem to be tightly tied to their time of production. Hemingway's fiction is closely tied to the World War I generation. Picasso's Les demoiselles d'Avignon seems even more tightly tied to its moment of production. But consider traditional African sculpture from a particular region, some forms of traditional Chinese painting, or the naive work of an amateur artist, all of which may remain unchanged in style over many years. In these cases, it may seem plausible that the same work could be produced many years apart in different possible worlds. However, it seems possible that even those works that seem most closely tied to a moment in time might have been produced at slightly different times or, in special circumstances, very different times. Consider a possible world that duplicates the history of European art, but in which that whole history begins two hundred years earlier than it in fact did. In that world, Picasso paints Les demoiselles in the early eighteenth century (D. Davies 2003).

The contextual variable is perhaps immune to considerations such as those just raised about artist and time of production. Works from different eras, traditions, styles, or works made with different intentions will not be the same no matter how superficially similar they appear. This is an important point of the Menard example. Neverthetheless, a case may be made for the possibility of the same work in different contexts by appealing to the idea that slight differences in context in different possible worlds may still result in the same work. This is especially plausible if the specific difference in context would not make a difference to the creator of the work in question (D. Davies 2003).

This set of objections raises two broad concerns for the idea that musical and literary works are indicated structural types and that paintings are contextually identified physical objects. It raises objections to Levinson's specific proposal regarding the individuation of indicated structures, but it also questions whether any general formula appealing to any of the variables under discussion can individuate artworks correctly.

Before concluding that these concerns are correct, one needs to evaluate the objections on which they are based. Do the objections show what they set out to show? One problem with them is that they rely on uncertainties in human modal intuitions about artworks, which point more directly to epistemic rather than metaphysical possibilities. That is, in the face of the sorts of examples considered thus far, many individuals will be uncertain what to think, and so it will be epistemically possible, relative to their beliefs, that a certain principle of individuation is wrong. That, however, falls short of showing it is wrong.

Is there a way of sharpening intuitions to arrive at principles of individuation? Perhaps this can be done by getting clearer about what the Menard example and other similar examples reveal about structurally identical works. Cervantes and Menard had different artistic projects and, in pursuing these, each achieved (did) different things in their respective works. This pair of differences, concerning artistic project and artistic achievement, is crucial in individuating works and in identifying important artistic properties of them. In highlighting these differences, the analogy mentioned earlier between abstract sentence types and utterancesor, more broadly, sentences-in-useis a helpful one to remember. The language user in question, along with the user's intentions, the time of utterance, and the context of use, all commonly contribute to fixing what the utterance conveys beyond or in distinction from the semantic content of the sentence. However, the precise role each of these items plays may vary in different uses of language. Further, it is possible for different utterances to convey precisely the same thing. Regarding artworks, something similar is true (Stecker 2003): They are individuated by being a specific abstract or concrete structure used by an artist in pursuing such and such a project and achieving so and so. Usually the three variablesidentity of artist, time of creation, and artistic contextare crucial in constituting projects and what they achieve, yet their exact role can vary in different art forms and different traditions, as well as for many other reasons. The emphasis on the artistic project and the artist's achievement recalls the idea that artworks are action types or tokens. However, those views identified the work with the wrong entity. As the indicated-structure view emphasizes, the artwork is the product that results from the project and embodies the achievement.

See also Aesthetics, History of; Aesthetics, Problems of; Art, Definitions of; Danto, Arthur; Existence; Ontology; Ontology, History of.


Borges, J. L. "Pierre Menard, Author of the Quixote." In Labyrinthes, translated by J. E. Irby, 6271. London: Penguin, 1970.

Caplan, B., and C. Matheson. "Can Musical Works be Created?" British Journal of Aesthetics 44 (2004): 113134.

Currie, G. An Ontology of Art. London: Macmillan, 1988.

Danto, A. The Transfiguration of the Commonplace. Cambridge, MA: Harvard University Press, 1981.

Davies, D. Art as Performance. Oxford: Blackwell, 2003.

Davies, S. Musical Works and Performances. Oxford: Oxford University Press, 2001.

Dilworth, J. The Double Content of Art. Amherst, NY: Prometheus Books, 2004.

Dodd, J. "Musical Works as Eternal Types." British Journal of Aesthetics 40 (2000): 424440.

Dodd, J. "Defending Musical Platonism." British Journal of Aesthetics 42 (2002): 380402.

Goodman, N. Languages of Art. New York: Bobbs-Merrill, 1968.

Howell, R. "Ontology and the Nature of Literary Works." Journal of Aesthetics and Art Criticism 60 (2002a): 6779.

Howell, R. "Types, Indicated and Initiated." British Journal of Aesthetics 42 (2002a): 105127.

Kivy, P. The Fine Art of Repetition. Cambridge, U.K.: Cambridge University Press, 1993.

Levinson, J. "What a Musical Work Is." Journal of Philosophy 77 (1980): 528.

Levinson, J. The Pleasures of Aesthetics. Ithaca, NY: Cornell University Press, 1996.

Predelli, S. "Musical Ontology and the Argument from Creation." British Journal of Aesthetics 41 (1980): 279292.

Rohrbaugh, R. "Artworks as Historical Individuals." European Journal of Philosophy 11 (2003): 177205.

Stecker, R. Interpretation and Construction: Art, Speech, and the Law. Oxford: Blackwell, 2003.

Wollheim, R. Art and its Objects. 2nd ed. Cambridge, U.K.: Cambridge University Press, 1980.

Wolterstorff, N. Worlds and Works of Art. Oxford: Clarendon Press, 1980.

Robert Stecker (2005)

Cite this article
Pick a style below, and copy the text for your bibliography.

  • MLA
  • Chicago
  • APA

"Art, Ontology of." Encyclopedia of Philosophy. . 21 Sep. 2018 <>.

"Art, Ontology of." Encyclopedia of Philosophy. . (September 21, 2018).

"Art, Ontology of." Encyclopedia of Philosophy. . Retrieved September 21, 2018 from

Learn more about citation styles

Citation styles gives you the ability to cite reference entries and articles according to common styles from the Modern Language Association (MLA), The Chicago Manual of Style, and the American Psychological Association (APA).

Within the “Cite this article” tool, pick a style to see how all available information looks when formatted according to that style. Then, copy and paste the text into your bibliography or works cited list.

Because each style has its own formatting nuances that evolve over time and not all information is available for every reference entry or article, cannot guarantee each citation it generates. Therefore, it’s best to use citations as a starting point before checking the style against your school or publication’s requirements and the most-recent information available at these sites:

Modern Language Association

The Chicago Manual of Style

American Psychological Association

  • Most online reference entries and articles do not have page numbers. Therefore, that information is unavailable for most content. However, the date of retrieval is often important. Refer to each style’s convention regarding the best way to format page numbers and retrieval dates.
  • In addition to the MLA, Chicago, and APA styles, your school, university, publication, or institution may have its own requirements for citations. Therefore, be sure to refer to those guidelines when editing your bibliography or works cited list.