NON-WESTERN ART AND THE CONCEPT OF ART: CAN CLUSTER THEORIES OF ART ACCOUNT FOR THE UNIVERSALITY OF ART?

artworks. Yet many artworks are not intended for these kinds of regards. There is no feature in Gaut’s theory that can prevent this kind of appropriation. His cluster account could be supplemented by a historical constraint: the featured criteria can only count towards arthood if the criteria in question are historically ‘correct regards’.61 Artefacts are created, dealt with, and appreciated in concrete historical settings. The way in which they were dealt with and appreciated are correct regards for these artefacts. An artwork may be ‘formally complex and 60 Note that I write ‘conceptions’ and not ‘conception’, since there is no one Western conception of art, but there are multiple conceptions. 61 See Jerrold Levinson, ‘Refining Art Historically’, Journal of Aesthetics and Art Criticism 47 (1989): 26. Annelies Monseré Estetika: The Central European Journal of Aesthetics, XLIX/V, 2012, No. 2, 00–00 161 Zlom2_2012_Sestava 1 14.11.12 8:26 Stránka 161


I. INTRODUCTION
The question of whether or not much of what we call 'non-Western art' is in fact art continues to trouble philosophers, art historians, and anthropologists alike.
On the one hand, it seems obvious to see the sculptures, paintings, and songs of other cultures as art. Moreover, it appears to be an empirical fact that we can recognize and appreciate art cross-culturally. 1 On the other hand, it is equally obvious that most non-Western cultures do not fully hold 'our concept of art' .
Though there is considerable overlap in the ways they talk and see their artefacts, often there are also important differences: their artefacts do not necessarily have Non-Westerm Art and Concept of Art: Can Cluster theories of Art account for the universality of Art?
This article is based on a paper entitled 'Gaut's Cluster Account and the Identification of Non-Western Art' , which I presented at the 2011 British Society of Aesthetics Conference. I thank the conference participants for their insightful comments, especially Berys Gaut, David Davies, and Simon Fokt. I am also grateful to Bart Vandenabeele and Violi Sahaj for their useful remarks. This work was supported by the Research Foundation Flanders (FWO). the same functions and meanings as our art. 2 I will argue that, in order to maintain the idea that art has transcultural and transhistorical significance, a theory of art needs to distinguish between the broad concept of art and its many different historical and cultural conceptions. Thus, on the one hand, a theory of art must show that 'art' is a unified concept; the reasons a Greek vase, a Baule sculpture, Chinese calligraphy, and a Jane Austin novel are art must be coherent. On the other hand, a theory must acknowledge that there are many different conceptions of art. Art-making features are culturally and historically conditioned; what these features entail and when they apply depend on the cultural and historical context in which these features appear. 3 Many philosophers of art recognize the universality of art. Yet their definitions cannot adequately account for non-Western art. Formalist definitions are rightly criticized because they include non-Western art on account of one narrow Western concern -namely, form: they start from one conception of art, the formalist conception, in order to define and clarify the concept of art. 4 Institutional and historical definitions define art in terms of the artefact's relationship to certain institutional structures 5 or to earlier uncontested artworks, 6 to which much non-Western art is unrelated. In order to account for non-Western art, they have to stretch their definitions almost beyond recognition. 7 Though they can easily acknowledge that there are many conceptions of art, they cannot form a unified concept of art, that is, they cannot show how and why art is a human universal. 8 Cluster theories try to identify or define art in terms of a list of recognition or cluster criteria that are disjunctively sufficient, but not necessary, for 'arthood ' . 9 As such, cluster theories do not take any of these conditions -namely, form, institutional and historical embeddedness -to be necessary conditions for arthood. Hence, they seem to be in a better position to elucidate the arthood of non-Western artworks in a non-appropriating way. Moreover, cluster theories can explain why an artmaker, or even a culture, does not need a conscious concept of art. The intention to make art is a non-necessary feature of art; hence the fact that much non-Western art lacks this feature does not imply that these artworks are not art. These artworks can be art due to other cluster or recognition criteria they do possess. 10 Furthermore, some prominent advocates of cluster theories defend their theories in terms of their ability to account for non-Western art. Denis Dutton, for one, claims that the recognition criteria in his list can be found in the arts cross-culturally: 'The list could be described as inclusive in its manner of referring to the arts across cultures and historical epochs ' . 11 Although these cluster theories seem to show promise when accounting for the universality of art, it will be argued they are not fully satisfactory. The aim of this essay is twofold. First, I will develop an argument in favour of the universality of art and clarify the requirements that theories of art must meet in order to account for this universality. Second, I will contend that cluster accounts fail to meet these requirements and conclude that they only provide some starting points for a truly cross-cultural theory of art.  51 (1993): 425-35. Though the term 'cluster theory' is most frequently used by Gaut, the other theories mentioned are very similar in structure and content and can therefore also be seen as cluster theories of art. The main difference between Gaut's account and Dutton's is that Gaut presents his theory as an anti-essentialist identification theory, whereas Dutton presents his theory as a definition. Moravcsik, however, presents his theory as a middle way between essentialism and antiessentialism. Stephen Davies, amongst others, has convincingly argued that Gaut's cluster theory is best seen as a highly disjunctive definition of art. See Stephen Davies, 'The Cluster Theory of Art' , British Journal of Aesthetics 44 (2004)

II. ART AS A HUMAN UNIVERSAL
This essay starts from the highly debated idea that a theory of art must be able to account for non-Western art. Some authors argue that only fine art, as it developed in post-Enlightenment Europe, truly qualifies as art. However, like Stephen Davies, I prefer 'to acknowledge the music, dance, drama, literature and picturing found in all cultures as art, and to regard Western fine art as one species within a wider genus that also includes religious art, domestic art, and so on' . 12 Similar views can be found in the philosophies of art of, among others, Arthur Danto,Noël Carroll,and Denis Dutton. 13 Yet no consensus is in sight. Initially, the view that only the West has art implied cultural hegemony: only 'we' could have a valuable cultural practice like art.
However, contemporary art historians, anthropologists, and philosophers who argue that art is a uniquely Western phenomenon do not aim to argue for Western superiority; rather they maintain that we cannot understand other cultures through 'our' Western categories, such as art. In this view, the inclusion of non-Western artefacts in the domain of art is tantamount to cultural appropriation. The West grants arthood to non-Western artefacts that bear obvious similarities to Western canonical art, regardless of the intentions of the makers and the meanings and functions of the artefacts under consideration. This mechanism is indeed obvious, for example, in formalist definitions of art: artefacts of non-Western traditions are granted arthood purely based on formal similarities with Western modern art. 14 Nonetheless, these criticisms do not necessarily suggest that non-Western cultures do not have art. The crucial question is whether the cultural artefacts we call non-Western art are art 'in themselves' or are art only because Westerners have conferred art-status on them. Many anthropologists take the latter view. Concerning African artefacts we call art, the anthropologist Sidney Littlefield Kasfir, for example, simply states: the view 'that from an African perspective, these objects are not art in the current Western sense is too well known to discuss' . 15 Accordingly, non-Western Arthur Danto (Beyond the Brillo Box, 110) maintains that: 'My sense is that the philosophical structure of African artworks is the same as the philosophical structure of artworks in any culture. ' Noël Carroll ('Art and Human Nature' , 98) links art to human nature and claims that 'once we cease to allow ourselves to be misled by the eighteenthcentury theory that art is exclusively an affair of disinterested contemplation, we find that our operative conception of art coincides approximately to what we find elsewhere in other cultures' . Dutton has argued that the differences between world cultures have been much overemphasized at the expense of recognizing similarities and pancultural universals, such as art. See Denis Dutton, '"But They Don't Have Our Concept of Art"' , in Carroll, Theories of Art Today, 217-40. art is only art when it is placed outside its indigenous milieu and recontextualized in the Western artworld. 16 In other words, it is argued that (most) non-Western cultures do not have art unless the identity of their artefacts is illegitimately altered by their introduction into a Western artworld. From the fact that many, if not most, non-Western cultures have no concept for art, the advocates of this view claim that virtually only the post-Enlightenment West has art.
Although it is rightly argued that many non-Western artefacts are illegitimately recontextualized into our artworld and thus misappreciated and misunderstood, this does not exclude the possibility of a broader, more universal concept of art. 17 From the fact that most non-Western cultures do not have a concept of 'art' , it does not follow that they have no art. Anthropologists and philosophers have argued that no one needs a conscious understanding of the concept 'art' in order to make art. Moreover, it is fair to ask whether a culture even needs a concept of art at all, conscious or unconscious, in order to have art. 18 The anthropologist Philip Ravenhill agrees that Wan men do not call their Goli masks 'art' , but they do recognize that the nature of these artefacts is radically different from that of other manmade objects. Ravenhill explains: It is because of the place that it occupies in their life and thought that I have no problem in considering it art as I understand that term, since for me it has the power that I attribute to art: the power to transform, to create new realities, to offer new possibilities, to create responses, to confound, and to do things that otherwise cannot be done. 19 Ravenhill points out that although the formalist notion 'It's art because it's like our art' is rightly rejected, it is replaced by an equally problematic notion that says: 'It's not art because they have no word for art' . 20 Likewise, the philosopher Arthur Danto has shown that the distinction between artefact and art is not lexically marked in the vocabularies of African languages generally, but claims that the absence of lexical markers can hardly be taken as evidence that the distinction cannot be made or that it is not made in the linguistic community in question. 21 Non-Western Art and the Concept of Art: Can Cluster Theories of Art Account for the universality of Art? McIver Lopes ('Art without "Art"' , 2) has persuasively pointed this out: 'We must be wary if, as the reception history suggests, there is some possibility that non-Western art is a figment of the Western imagination -if what we see in primitive art is in this way a projection of our own conception of art. The sceptic simply elevates this wariness to a methodological principle. ' Dominic McIver Lopes goes to the heart of the matter when he questions the so-called 'concept dependence thesis' underlying the view that the West is almost alone in having art. This thesis states that a culture has art only if the members of a culture have a concept of art. 22 He compellingly points out that a theory of art that supports the concept dependence thesis implies that it is impossible for a culture to have art that is made incidentally. Cluster theories, or so Lopes argues, can show that artefacts not made with a conscious concept of art can be art because they share many features with other artworks. Many artefacts made within 'primitive' or 'traditional' societies are used in religious or social rituals, and 'they are suited for use in ritual contexts because they afford aesthetic experiences, have aesthetic properties, express emotion, challenge the intellect, communicate complex meanings, express personal view points, are original, or exhibit skill. ' 23 Hence, the makers of these artefacts intended them to have these (art-making) features: these features are not secondary; we do not appropriate these artefacts by recontextualizing them. The non-Westerners who made them were aiming for appreciable effects parallel to those that mark our art as such. 24 There is consequently no reason to maintain that these artefacts are not art and to accept the concept dependence thesis. Accepting the thesis leads to cultural exclusion, but rejecting it does not necessarily lead to cultural appropriation. 25 Moreover, holding on to the concept dependence thesis and rejecting the possibility of 'incidental' art implies that most of the acknowledged masterpieces of art in any culture will not count as art. 26 The rejection of the concept dependence thesis has, as Lopes points out, quite serious implications: If art is not just what Westerners have a concept of, then theorists must look beyond Western art to ascertain the nature of art. They cannot assume that the nature of art is determined by the Western concept of art. Thus anthropological studies of art in different cultures are essential resources in developing theories of art. 27 Yet looking at anthropological studies of art in different cultures is only a starting point. Subsequently, a theory of art must elucidate the broader concept of art.
Which criteria must a theory of art meet in order to do this? It is to this question I will now turn. 22 McIver Lopes, 'Art without "Art"' , 3. 23 Ibid., 10-11. 24 I am indebted to an anonymous referee for this suggestion. 25 Consequently, these artefacts possess art-making features in light of their function; they do not have these features because 'we' project these features into these artefacts. 26 McIver Lopes, 'Art without "Art" ' , 11. 27 Ibid., 14.

III. NON-WESTERN ART AND THEORIES OF ART
If a theory of art is intended to clarify the broader concept of art, it needs to clarify not only the arthood, for example, of Baule art, Western art, and Chinese art, but also the transhistorical and transcultural concept of art. Thus, a theory should not only provide criteria to include non-Western art, but also needs to show that art is a coherent concept. Moreover, the concept of art cannot be illuminated by pointing at culture-and time-specific criteria for arthood. It must therefore take into account the historicity of art: every criterion put forward for arthood is historically conditioned. If it is claimed that something is art because of the aesthetic properties it exhibits, then it has to be acknowledged that what can be seen as an aesthetic property now may not have been seen as an aesthetic property in the past or in another culture. While it is generally agreed that Matisse's paintings exhibit aesthetic properties like beauty and expressiveness, these paintings would not have been perceived as beautiful or expressive in the Renaissance. Moreover, they could not even have been made in the Renaissance. This view is very much in line with Danto's claim that the philosophical structure of all artworks is the same, but not all artworks can be art at the same time. History makes new conceptions of art possible, but that does not mean we lack a coherent concept of art. 28 In short, the fact that the concept of art has transcultural and transhistorical significance does not mean that there are no multiple conceptions of art and vice versa. 29 Both facts have to be acknowledged by a theory of art. Advocates of cluster theories of art have criticized other theories because these illegitimately take standards and traditions that exist in the philosopher's contemporary place and culture as essential to arthood. 30 Denis Dutton has compellingly pointed out that although aesthetic theories might claim universality, they are usually conditioned by the aesthetic questions and debates of their times. 31 If the features of one particular conception of art are seen as art's essential defining features, then the arthood of culturally and historically remote art cannot be illuminated.
In this case, these objects are only granted arthood in the light of one particular Non-Western Art and the Concept of Art: Can Cluster Theories of Art Account for the universality of Art? Furthermore, as we have seen, cluster theories of art can show us why the concept dependence thesis can be discarded, without this leading to the false formalist notion that says: 'it is art because it looks like our art. ' The question is, however, if cluster theories of art are, in content and structure, better suited for clarifying the arthood of non-Western art. Can they show why dissimilar artefacts from different cultures belong to the same category of art without ignoring their substantial difference? In the next part of this essay, I will examine what these accounts entail.

IV. CLUSTER THEORIES OF ART
The cluster theories that will be addressed here are the ones developed by Berys Gaut, Denis Dutton, and Julius Moravcsik. Gaut's formulation of the cluster theory of art is the most advanced and straightforward to date. His aim is to show that art is better understood as a cluster concept, since the concept 'art' does not exhibit necessary and sufficient conditions, but only conditions that are jointly sufficient, but not necessary. Gaut does not seem very preoccupied with the problem of non-Western art. Almost all the examples he uses to defend his theory are taken from the Western history of art. Only in order to point out that not all artworks are the products of an intention to make art, and that this criterion, accordingly, is not an essential or necessary feature of arthood, does he turn to non-Western art. Much primitive art is made without a conscious intention to make art and, as such, such an intention is not a necessary condition for arthood. He defends the claim as follows: '"primitive" societies tend not to have anything like our concept of art, but we accept some of their products as art, and probably much that we now accept as "folk art" was never intended by its makers as art' . 32 This suggests that Gaut has a quite different outlook from the one defended in this essay. He seems to agree with Kasfir's statement that these societies only have art in virtue of Western recontextualization. Nevertheless, analyzing his cluster theory is relevant since it will show that other cluster theories, despite seeing art as a transhistorical and transcultural phenomenon, exhibit the same problems as Gaut's cluster theory with regard to non-Western art. Both Dutton and Moravcsik argue that art is a cross-cultural phenomenon. Non-Western art is art not because of Western intervention, but because it possesses a sufficient amount of art-making features. 33 32 Gaut, '"Art" as a Cluster Concept' , 32. In order to defend their lists of art-making features, they refer to Western as well as non-Western examples. As such, they take seriously Lopes's assertion that theorists must look beyond Western art to ascertain the nature of art and must use anthropological studies of art in different cultures as essential resources for developing theories of art. 34 The first incarnations of cluster theories of art occur in Morris Weitz's and Paul Ziff's anti-essentialist philosophy of art. They, amongst others, introduced the idea that art cannot be defined, bearing in mind centuries of failed attempts at defining art and the diversity of artistic practice. 35 They argued that definitions take only one feature of art as its defining feature. This is unwarranted since many features are indicative of arthood. Applying Wittgenstein's analogy of family resemblance to art, Weitz argued that the arthood of an artefact can be identified, but not defined, by its similarities to other artworks. 36 Roughly two methods followed from this approach: the resemblance-to-paradigm account and the cluster account. The former suggests that to identify arthood, one has to decide whether the artefact under consideration is similar to paradigmatic artworks. 37 The latter starts from the idea that there are bundles of properties, none of which needs be present but most of which are present when we describe things as works of art. 38 Both accounts are hinted at in Weitz's writings.
Recent manifestations of the cluster theory of art are clearly inspired by these writings. A cluster theory of art starts from the idea that there are criteria of recognition of works of art. These criteria can be found in classical definitions of art. There, however, they are presented as necessary and sufficient conditions; they constitute the essence of art, at the expense of other criteria. 39 Yet, in Weitz's view, 'None of the criteria of recognition is a defining one, either necessary or sufficient, because we can sometimes assert of something that it is a work of art and go on to deny any one of these conditions ' . 40 Non-Western Art and the Concept of Art: Can Cluster Theories of Art Account for the universality of Art? 34 McIver Lopes, 'Art without "Art"' , 14. 35 Weitz, 'Role of Theory'; Paul Ziff, 'The Task of Defining a Work of Art' , Philosophical Review 62 (1953): 58-78. 36 Weitz, 'Role of Theory' , 31. It is a matter of debate whether Wittgenstein's notion of 'family resemblance' can be used as a technical notion for classifying concepts or for application to new cases like art. See, for example, Ben Tilghman, 'Kaufman on Art, Family Resemblances, and Wittgenstein' , British Journal of Aesthetics 48 (2008): 86-88. This debate will not concern us here. 37 Weitz, 'Role of Theory' , 31-32. These theories suffer from many problems since they do not specify which artworks are paradigmatic and what kind of resemblances are sufficient. For more or less 'friendly' criticisms, see Carroll, 'Historical Narratives' , and Gaut, '"Art" as a Cluster Concept' . 38 Weitz, 'Role of Theory' , 33. 39 Ibid. 40 Ibid., 34.
These main ingredients appear in recent formulations of the cluster theory.

Gaut developed the insights of Weitz and argues:
In holding that 'art' is a cluster concept, I mean that there are multiple criteria for the application of the concept, none of which is a necessary condition for something's being art. A criterion is a property, possession of which conceptually counts towards an object's falling under a concept; there are several criteria for a concept. 41 It follows that, firstly, if all criteria are fulfilled, the object is art and if some are fulfilled, then the object might be art. Secondly, not one of the proposed criteria is a necessary criterion for being art. 42 Thirdly, the criteria are disjunctively necessary: at least some of them must be instantiated if the object is to fall under the concept. 43 Though Gaut mainly wants to defend the form of the cluster account, rather than the concrete criteria put forward, he formulates ten criteria and claims: 'these criteria are good prima facie candidates for those which should appear in a cluster account, and I will defend the form of the account in terms of these specific features' . 44 The properties he puts forward are derived from the way in which we talk about art and, as such, they largely coincide with conditions proposed by classical definitions of art. 45 Dutton also formulated a highly disjunctive theory of art. He calls it a naturalist definition of art. In structure and content it is very much in line with Gaut's cluster account, since he also proposes a list of jointly sufficient, but non-essential conditions that are reminiscent of conditions proposed by definitions of art. Unlike Gaut, he starts from a broad range of cross-cultural examples of art. He does not focus on the Western history of art, and even claims that we should not worry too much about the inclusion of avant-garde artworks since they are 'an infinitesimally small class of cases' , but tries to look for characteristic cases of art worldwide, in order to understand arthood. 46 He defends this method as follows: We must first try to demarcate an uncontroversial center that gives the outliers whatever interest they have. I regard this approach as 'naturalistic' , not in the sense that it is biologically driven […], but because it depends on persistent cross-culturally identified patterns of behavior and discourse: the making, experiencing, and assessing of works of 41 Gaut,'Cluster Account' , Both Dutton and Gaut claim that there are necessary criteria for arthood: Gaut thinks that being the product of a human action is a necessary condition and Dutton argues that being an artefact is a necessary condition. Neither one, however, includes these conditions in his list of proposed criteria. I thank an anonymous referee for this observation. 43 Ibid. 44 Gaut, '"Art" as a Cluster Concept ' , 29. 45 Ibid., 28. 46 Dutton, 'Naturalist Definition' , 368. art.
[…] Characteristic features found cross-culturally in the arts can be reduced to a list of core items, twelve in the version given below, which I term recognition criteria. 47 In a similar vein, Julius Moravcsik endeavours to formulate a cross-cultural philosophy of art. He defends an approach that seeks to steer between essentialists, attempting analytic definitions of art, and anti-essentialists, who doubt the possibility of objective detached philosophical thinking, 'by delineating the subject matter as artworks in a wide cross-cultural sense, and by characterizing these works in terms of a feature system that allow us to formulate theoretically interesting hypotheses about artforms and art itself' . 48 He proposes a feature system. This is a set of characteristics which are jointly sufficient but not necessary for arthood and are found in art cross-culturally. Like the features put forward by Gaut and Dutton, they are reminiscent of, and derived from, the essential conditions put forward in classical definitions of art. Moravcsik claims that the definitions of art on offer provide some starting points for such a list. He explains his method as follows: Dropping the controversial uniqueness clause from these proposals we obtain characterizations that might not be rival definitions, but rather features that artforms and objects in various periods and cultures might have. Seen in this way these features can be incorporated into a single framework. With this equipment we can scan the global horizon, noting how widespread each feature is, and what generalizations account for their occurrences and co-occurrences. 49 Although the aims of Dutton and Moravcsik differ considerably from the aims of Gaut, their cluster theories are, as has, I hope, been shown, much alike in structure and even in content. First, the content, that is, the criteria put forward by these theories, will be explored.

V. CLUSTER CRITERIA AND THE CONCEPT OF ART
Does the idea that art is a cluster concept show us in any way why art is a concept at all or does it just arbitrarily lump together different conceptions of art? First, Gaut's criteria will be explored. Although Gaut mainly wants to defend the form of the cluster account, rather than the concrete criteria put forward, he formulates and defends the following ten criteria: (1) possessing positive aesthetic properties, such as being beautiful, graceful, or elegant (properties which ground a capacity to give sensuous pleasure); (2) being expressive of emotion; (3) being intellectually challenging (i.e., questioning received views and modes Non-Western Art and the Concept of Art: Can Cluster Theories of Art Account for the universality of Art? 47 Ibid. 48 Moravcsik, 'Why Philosophy of Art ' , 425. 49 Ibid., 432. of thought); (4) being formally complex and coherent; (5) having a capacity to convey complex meanings; (6) exhibiting an individual point of view; (7) being an exercise of creative imagination (being original); (8) being an artefact or performance which is the product of a high degree of skill; (9) belonging to an established artistic form (music, painting, film, etc.); and (10) being the product of an intention to make a work of art. 50 These criteria are reminiscent of criteria proposed by formalist, expressionist, aesthetic, mimetic, institutional, and historical definitions of art. 51 In order to select these criteria, Gaut explores how the concept of art is used in language. 52 Yet the way in which we talk about art is conditioned by our background. Thus, when we try to establish the arthood of certain artefacts, we usually refer to 'our' paradigmatic or uncontested artworks, which largely coincide with the Western canon.
Consequently, Gaut seems illegitimately to reinforce the centrality of these artworks. Still, he claims that the criteria are derived not only from paradigmatic artworks, but also from more contested artworks, since we also address these artworks when we talk about art. 53 Indeed, we might talk about contested artworks, including non-Western artworks, but these artworks are contested within 'our' framework. Moravcsik rightly emphasizes that practices that are peripheral to 'us' might be central for general human artistic practices. 54 It follows that Gaut's theory only clarifies the conceptions that are central 'to us' , that is, Western conceptions. In this sense, Gaut strengthens the idea that non-Western art is art because it is like our art. The Western conceptions of art stay referential and artworks are only allowed to enter the domain of art to the extent that they are in accordance with these conceptions. Dutton, on the other hand, tries to counter the idea that non-Western art is art because it is like our art. He argues that we should start from 'a natural centre' in order to think about art and that this natural centre should be cross-cultural. His list, derived from this natural centre, includes (1) direct pleasure, (2) skill or virtuosity, (3) style, (4) novelty and creativity, (5) criticism, (6) representation, (7) 'special' focus, (8) expressive individuality, (9) emotional saturation, (10) intellectual challenge, (11) art traditions and institutions, and (12) imaginative experience. 55 At first sight these criteria bear many similarities with those proposed by Gaut. Nevertheless, Dutton argues that his list has plenty of features that are relevant to Western as well as non-Western art. Larry Shiner, however, doubts whether Dutton's list can lead to a more universal concept of art.
Shiner distinguishes between art in the older and broader sense that can be found 50 Gaut, '"Art" as a Cluster Concept' , 28. in nearly all cultures of the world and the narrow Western concept of 'art' . He maintains that Dutton's list can provide reasons to accord arthood to Western and non-Western art, but cannot show why both conceptions are part of the same concept. He states: 'Obviously, the ambiguity of the word "art" as a cluster concept permits the kind of mixture of properties Dutton proposes, but when the only properties on the list applicable to indigenous societies are those typical of the older sense of art, the claimed "universality" is dubious. ' 56 In other words, suppose an artefact possesses features 1, 2, 6, and 7, and another artefact possesses features 4, 5, 10, and 11, and, accordingly, both are called 'art' because of these features. Why is the term 'art' warranted in both cases?
These objects do not share any features. Nonetheless, they are put in the same category -'art' . This would not be a problem if the proposed criteria showed some kind of coherence, some kind of overarching function or meaning. Yet these criteria tell us, as Dutton admits, what we already know. They are used to claim that certain artefacts are art. But why these criteria are legitimately used and how they clarify the broader concept of art is left unexplained.
The same criticisms seem applicable to Moravcsik's cross-cultural philosophy of art. Like Gaut and Dutton, he proposes a list of features that are relevant for understanding the concept of art. He argues that different artforms at different times and throughout different cultures can be characterized, for example, as being representational, expressive of emotion, institutionalized, productoriented, conscious of art theory, related to functional utilitarian properties, and admitting the low/high culture distinction: 'Such a scheme will give us crossclassifications, and thus a"grid" system. Of course, other candidates might emerge as legitimate features for such grids, such as strong drive for novelty in form, or being influenced by political matters.' 57 Contrary to Dutton, who  grounds on which we can add features to the set and why the features in his list are valid. Consequently, the objectivity of the features in the grid system is not assured. Again, his cluster account cannot help us understand art; it just tells us what we already know: it clarifies our thinking about art, but does not show why and when our thinking, and thus the list of features, is correct.
I have, I hope, demonstrated that cluster theories cannot clarify the concept of art without further research into why the proposed criteria are criteria for arthood. Now the question whether cluster accounts can illuminate art's many conceptions will be addressed.

VI. CLUSTER THEORIES AND ART'S MANY CONCEPTIONS
At first sight, cluster theories of art easily make it clear that art has many conceptions. A twentieth-century modernist conception of art will showcase different cluster or recognition criteria from those in a Baule conception of art.
The former conception does not seem to be fallaciously extended to the latter; its criteria are not turned into art's only defining features.
As we have seen, Gaut seems mostly concerned about Western conceptions of art. 60 Nonetheless, even if we acknowledge that his aim is more limited than Dutton's and Moravcsik's, Gaut's theory cannot prevent the illegitimate application of certain cluster criteria within the domain of the Western conceptions of art. Gaut's cluster account does not specify which properties are applicable when, nor, consequently, does it take into account the fact that the reason something is an artwork in the twentieth century is not necessarily a reason why something is art in the Renaissance. Thus, an artefact can be accorded arthood for the historically wrong reasons. It follows that, for example, religious art can be bestowed arthood based on formal and other narrow aesthetic properties that are mainly incidental. These properties legitimately feature in his cluster account, since they are important for many modernist abstract artworks. Yet many artworks are not intended for these kinds of regards.
There is no feature in Gaut's theory that can prevent this kind of appropriation.
His cluster account could be supplemented by a historical constraint: the featured criteria can only count towards arthood if the criteria in question are historically 'correct regards' . 61 Artefacts are created, dealt with, and appreciated in concrete historical settings. The way in which they were dealt with and appreciated are correct regards for these artefacts. An artwork may be 'formally complex and 60 Note that I write 'conceptions' and not 'conception' , since there is no one Western conception of art, but there are multiple conceptions. coherent' and have 'positive aesthetic properties' , but if the maker did not intend the object to have these properties and the artefact was not appreciated because of these properties, they do not constitute correct regards for this artefact. Gaut's cluster theory, as it stands now, is an ahistorical bundling of characteristics or criteria. We should turn to history in order to see whether a criterion can be legitimately applied. History will also show us that not every criterion always amounts to the same thing. What is an aesthetic property in the twentieth century will not necessarily count as an aesthetic property in the fifteenth century. An artinstitutional context in the Middle Ages is something completely different from an art-institutional context now. It follows that every criterion will have many historical instantiations. 62 This will render the cluster theory even more complex, to the point of being quite uninformative. Moreover, it will even be more difficult to formulate a coherent concept of art.
Dutton's theory seems not to provide a way of escaping this kind of misapplication. He formulates twelve features, yet he nowhere indicates that we should first find out whether they have been justly applied to a specific artefact.
Instead, he emphasizes that the features on his list are 'easily graspable surface features': The features on this list are implicated, individually and more often jointly, in answers to the question of whether, confronted with an art-like object, performance, or activity, we are justified in calling it art. As recognition criteria, they therefore identify the most common and easily graspable 'surface features' of art, its traditional, customary, or pretheoretical characteristics. 63 'Easily graspable surface features' are precisely those features that tend to be easily misinterpreted. Danto has convincingly shown that objects with the same surface features can have a different object status. 64 Even though many surface features combined provide a convincing indication for arthood, it is still possible that the artefact in question is not art. Dutton's theory is therefore susceptible to the same criticisms as Gaut's theory: whether or not a criterion is applicable and what the criterion in question entails in the specific historical context it originated are not taken into account.
Moravcsik, on the other hand, does pay attention to the context in which the characteristics feature. He makes clear that every feature needs to be differentiated: no feature has an entirely fixed meaning. About the feature 'representation' , for example, he states: Representation is an achievement and we can succeed to a greater or lesser degree. Thus it has to be viewed in the context of different skill levels, available techniques, and technological means. What might count as a representational work of art in one context might count in a different context of available skill and technique as not meeting minimal standards. We also need to consider culture-dependent conventions of representation.
For example, what might seem to us as an unsuccessful representation of human arms in 'natural' positions may be in fact a successful representation -according to certain conventions -of humans in a pose of worship. 65 Accordingly, Moravcsik will not allow artefacts in the domain of art solely on the basis of surface features. Moreover, he calls for an investigation into what these features mean in certain historical and cultural contexts. But this does not render his theory much less problematic. On the one hand, the informative value of his feature or grid system can be called into question. If every feature has innumerable instantiations, then it is unclear how the employment of this grid system can 'help us in arriving at an insightful analysis of art across time and cultures' . 66 On the other hand, if Moravcsik cannot show even that the proposed features have a stable common core, how can he hope to show that there is unity in the concept of art?

VII. CONCLUSION
There are no compelling reasons, this essay hopes to demonstrate, to exclude non-Western artefacts from the domain of art. If we limit arthood to cultures that have a concept of art then we illegitimately restrict the domain of art, even excluding canonical works of the Western tradition. Yet if we want to include these canonical artworks, then there is no reason to exclude from the domain of art non-Western artefacts that showcase similar art-making features. Moreover, a theory of art needs to take this universality into account. It can neither simply start from 'our' art traditions and extend these conceptions of art to other cultures and times, since this would lead to cultural appropriation, nor can it resolve the matter by simply formulating separate criteria for non-Western art or 'art without a concept of art' , since this would mean that there is no unity in the concept of art. 65 Moravcsik, 'Why Philosophy of Art ' , 433. 66 Ibid., 434. At first sight, cluster theories of art seem capable of accounting for the universality of art. On the one hand, they can put forward a set of criteria which is derived from cross-cultural instances of art. Consequently, cluster theories need not start solely from Western art traditions. On the other hand, they are able to show that many criteria are relevant transculturally and transhistorically. Hence, these theories seem to suggest that there is unity in the concept of art. Yet cluster theories, as they stand now, remain unsatisfactory. Firstly, they simply systematize what is already known: they tell us only what kinds of criteria we use as indications for arthood. In this sense, they clarify our way of thinking about art more than they clarify anything about the concept of art itself. Moreover, they do not provide a method for distinguishing between correct and incorrect thinking about art, since they do not clarify why certain criteria feature in their clusters. Secondly, cluster theories cannot show us why two objects that share none of the features on the list can both be art. Consequently, they do not indicate why art is a unified concept. Thirdly, the majority of these theories do not take into account that not every criterion is valid in every given context and that all criteria have endless historical and cultural instantiations. It follows that either these theories, as formulated by Gaut and Dutton, are thoroughly ahistorical and include artefacts in the domain of art for the historically wrong reasons, or these theories, as developed by Moravcsik, take all these contextual factors into account and become highly complex and insufficiently informative.
Cluster theories can indeed provide us with starting points for a cross-cultural theory of art. This essay has sought to demonstrate, however, that the starting points they offer are still very undeveloped and many questions remain for further research. Advocates of cluster theories will need to point out what it is that grounds and binds the proposed criteria. Only then will they be able fully to address the universality of art.