The folk concept of art

contributioncontent_tagsWhat is the folk concept of art? Does it track any of the major definitions of art philosophers have proposed? In two preregistered experiments (N = 888) focusing on two types of artworks (paintings and musical works), we manipulate three potential features of artworks: intentional creation, the possession of aesthetic value, and institutional recognition. This allows us to investigate whether the folk concept of art fits an essentialist definition drawing on one or more of the manipulated factors or whether it might be a disjunctive or cluster concept. The results suggest that none of the three manipulated properties alone suffices for an object to be considered art. The folk concept of art might thus well be a cluster concept instead of an essentialist concept.

1 Introduction

concepts↗concepts↗sensesThe definition of art has been one of the central topics in philosophical aesthetics in the last century. However, finding a satisfactory definition of art turned out to be very difficult: traditional definitions of art often include objects that are not artworks or exclude objects that are, in fact, artworks. Thus, a second question has gained more and more attention in the last few decades: Is it possible to define art at all? If so, what kind of definition would it be? Are there any individually necessary and jointly sufficient conditions for an object to be art? Is it rather a disjunction of minimally sufficient conditions that makes something art? Are these conditions (or set of conditions) determinate or indeterminate, i.e., is the concept of art closed, or is it open?

In response to these questions, three kinds of accounts of art have emerged: essentialist accounts propose individually necessary and jointly sufficient conditions. Disjunctive accounts claim there are individually sufficient and jointly necessary properties. According to cluster theories of art, there is a set of properties that make an object art, but the set is not determinate and might change: the concept of art is open if its conditions of use are “amendable and corrigible” (Weitz, 1956: p. 31).

object_of_studyresearch_questionconcepts↗concepts↗concepts↗concepts↗sensescontent_tags↗Although most philosophers see the definitional project as (at least partially) descriptive – that is, they regard it as dependent on how we do think about art rather than on how we should think about art – there is very little empirical research on what ordinary people identify as art. In this paper, we distinguish between two kinds of projects in the philosophy of art: essentialist definitions of art on the one hand, and non-essentialist definitions – cluster or disjunctive accounts – on the other. Our goal is to explore which of these approaches correspond best to the folk concept of art, or the concept of art shared by non-experts. We present an overview of different types of accounts of art (Sects. 2 and 3) and discuss the role empirical research can play in elucidating the concept of art (Sects. 4 and 5). In Sects. 6 and 7 we report two experiments that aim to answer the questions of whether (i) the folk concept of art is an essentialist, cluster, or disjunctive concept, as well as (ii) the relative importance of three properties of artworks – intentional creation, aesthetic value, and institutional recognition– that are often considered the most important ones by philosophers.

2 Essentialist and non-essentialist accounts of art

concepts↗content_tags↗content_tags↗As already sketched above, there are three categories of theories of art: essentialist definitions that specify individually necessary and jointly sufficient conditions, disjunctive definitions that present a list of properties at least one of which must be satisfied, and cluster accounts of art that hold that none of the properties are individually sufficient.

In the following, we will briefly present a few essentialist definitions.Footnote 1 Historical accounts emphasize the relation between a work of art and some earlier works, such as the intended resemblance of an artwork to some acknowledged earlier pieces (Carroll, 1993). One prominent historical account– the intentional-historical view– requires artistic intentions for an artwork to be regarded in the same way as other objects belonging to this category. Whether something falls under the concept of art thus depends both on what was seen as art in the past, and on whether the creator’s intention tracks previous art in relevant ways (Levinson, 2002). For example, whether or not a painting counts as art would depend, as Levinson outlines, on factors such as the painter’s attention to colour and painterly detail, awareness of stylistic features, the artwork’s historical background, and so forth (1989: p. 24).

concepts↗concepts↗Consider a different kind of essentialist definition. According to functionalist accounts, art is defined by its purpose: there is some specific function that all works of art must embody, to satisfy some specific need that art can fulfil. The most common functionalist account is the aesthetic view, according to which the function of art is to satisfy the need for aesthetic experiences (Beardsley, 1982; Zangwill, 1995). Other functionalist definitions (e.g. Stecker, 2005) do not specify a precise function. There are many possible functions of art, for example, according to Pignocchi, “bearing formal or expressive properties, imitating nature, inducing an aesthetic experience, bearing embodied meaning, inducing awe, being key elements of ceremonies, faithfully representing a certain reality, drawing attention to a social or ecological problem, stimulating thinking about art, criticizing its futility, ostensibly violating the rules tacitly employed by an artistic genre in order to revive its language” (2014: pp. 428–429).

A hybrid functionalist account– the procedural-institutional view– claims that the status of being art is conferred upon an object by someone with authority (Danto, 1981; Davies, 2004a, 2004b; Dickie, 1974). Advocates of this account mention procedures such as the object’s being created by an artist, held by someone with authority, presented to the artworld, and discussed by art critics and art historians.

concepts↗unhoused_relationssensesclaim_relationscontent_tags↗All of the discussed definitions can be interpreted as essentialist: they invoke necessary and sufficient conditions for an object to be considered an artwork, or at least emphasize one factor as the most important, while others are treated as more peripheral. This view is challenged by anti-essentialists, who object to essentialist definitions with counterexamples that are intended to demonstrate that the respective views are over- or underinclusive. We can distinguish between cluster and disjunctive anti-essentialist theories, both of which draw inspiration from Wittgenstein’s idea of ‘family resemblance’ (Wittgenstein PI, 2009: §67, Weitz, 1956; Ziff, 1953; Dean, 2003). To illustrate this concept, take Wittgenstein’s example of a ‘game’. There are board games, card games, Olympic games, ball games etc., but, Wittgenstein argues, there is no single property (or set of properties) they all have in common. However, and this is what it means for two concepts to bear a family resemblance, if something shares similarities with several things that have been called games before, the concept of game is extended to include a new example. The concept of game thus has no rigid boundaries.

Art, or so anti-essentialists argue, is similar in this regard. Although artworks do not all share a single property (or set of properties), each artwork has at least one property in common with other objects that are considered art. In Berys Gaut’s words,

Suppose we can construct some set of properties […]. And suppose it can be shown that if various subsets of them obtain, then an object is art, that none of these properties has to be possessed by all artworks, but that all artworks must possess some of them. Then we cannot define “art” in the sense of giving individually necessary and jointly sufficient conditions for it, but we can offer a characterization of it– and account of what it is in terms of criteria or characteristics. Note that this account allows a great deal of indeterminacy in whether the obtaining of a particular subset of properties is sufficient for something to be art: there will be many cases where it is not clear whether this is so; what is important is that there are some subsets the obtaining of which is sufficient for something to be art. (Gaut, 2000: p. 27)

There might be cases that would invite us to reconsider the use of the concept of art in a way that would allow it to cover new objects (Weitz, 1956: p. 31). According to Roy Cook, the openness of the concept of art should be seen as an epistemological thesis, in the sense that we are unable to specify which objects are or could be art (either at present or in the future) (Cook, 2013: p. 69).

One defender of the first type of anti-essentialism, i.e. a cluster account, is Gaut. On his view, properties that are ordinarily regarded as counting towards an object’s status as a work of art include (1) possessing positive aesthetic qualities; (2) being expressive of emotion; (3) being intellectually challenging; (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; (8) being an artifact or performance which is the product of a high degree of skill; (9) belonging to an established artistic form; and (10) being the product of an intention to make a work of art (Gaut, 2000: p. 28). Dutton (2009) has argued for an account with a similar cluster of properties.

concepts↗content_tags↗Other anti-essentialists have put forward disjunctive accounts (Davies, 2015; Hazelwood, 2021; Longworth & Scarantino, 2010; Stecker, 1994). Such accounts specify a disjunction of sufficient conditions for an object to be considered art, with no condition’s being individually necessary: properties such as P or Q are individually sufficient and disjunctively necessary. For example, Stephen Davies offers such a definition: he holds that an object is art

(a) if it shows excellence of skill and achievement in realizing significant aesthetic goals, and either doing so is its primary, identifying function or doing so makes a vital contribution to the realization of its primary, identifying function, or (b) if it falls under an art genre or art form established and publicly recognized within an art tradition, or (c) if it is intended by its maker/presenter to be art and its maker/presenter does what is necessary and appropriate to realizing that intention (2015: p. 375).

Properties included in disjunctive definitions are seen as constitutive of what it is to be an instance of art. Nevertheless, such disjunctive accounts are incompatible with classical essentialist definitions, because possessing any property from the list suffices to be art, and there must be more than one property on the list. To illustrate, on Longworth and Scarantino’s disjunctive theory, the following conditions must be met for an object to be considered art:

(DTA): ∃Z ∃Y (Art ↔ (Z ∨ Y)), where (i) Z and Y are either non-empty conjunctions (e.g. P & Q & R) or non-empty disjunctions of conjunctions (e.g. (Q & R & S & T) ∨ (P & Q & W) v...); (ii) there is some indeterminacy over exactly which disjuncts are sufficient; (iii) Z does not entail Y and Y does not entail Z; (iv) Z does not entail Art and Y does not entail Art. (2010: 163)

This definition specifies that none of the disjuncts entail each other, and none of them entail an object’s being art. It ensures that disjuncts are not individually necessary, and that art cannot be defined merely by listing types of art.

In brief, although cluster and disjunctive theories standardly share some of the very properties that are individually emphasized by other (essentialist) definitions of art, the main difference is that cluster and disjunctive theories deny that any of these properties are individually necessary. In contrast to disjunctive theories, cluster accounts of art hold that none of the properties are individually sufficient.

3 The most influential essentialist accounts

concepts↗concepts↗concepts↗content_tags↗content_tags↗From among the considerable variety of classical views briefly surveyed in the previous section, three accounts have proven particularly influential: The institutional account (Dickie, 1974), the aesthetic account (Beardsley, 1982), and the intentional-historical account (Levinson, 1979). Dickie’s institutional definition claims that a work of art is “(1) an artifact, (2) a set of the aspects of which has had conferred upon it the status of candidate for appreciation by some person or persons acting on behalf of a certain social institution (the art-world)” (Dickie, 1974: p. 34).Footnote 2 In other words, an artifact is considered an artwork if it is recognized as such by people belonging to and acting in the name of relevant institutions. According to Beardsley’s aesthetic definition, an artwork is “either an arrangement of conditions intended to be capable of affording an experience with marked aesthetic character or (incidentally) an arrangement belonging to a class or type of arrangements that is typically intended to have this capacity” (Beardsley, 1982: p. 299). Put differently, artworks are, roughly, objects with dispositions to trigger aesthetic experiences.Footnote 3

According to Levinson’s historical-intentional definition of art,

X is an art work at t = df X is an object of which it is true at t that some person or persons, having the appropriate proprietary right over X, non-passingly intends (or intended) X for regard-as-a-work-of-art, i.e., regard in any way (or ways) in which objects in the extension of ‘art work’ prior to t are or were correctly (or standardly) regarded (1979: 240).Footnote 4

Differently put, on Levinson's view, an artwork is something that has been intended by someone to be treated in a way that some earlier artworks were treated (Levinson, 2007: 74).

These three definitions are the most obvious candidates for an empirical study for several reasons. First, according to Stephen Davies’ seminal book “Definitions of Art” (1991), there are two basic approaches to the question of how to define art: procedural and functional, and all (essentialist) theories of art can be reduced to one or the other of these two types of definitions. Dickie’s institutional and Beardsley’s aesthetic definitions are typical examples of procedural and functional approaches, respectively. Art, just like other social kinds, is often defined in terms of its function, as people create objects in order to satisfy certain needs. The function of art can be defined in different ways but is usually expected to provide a rewarding aesthetic experience.

concepts↗unhoused_relationsThe requirement for intentionality is so common in the literature that almost all definitions of art require an artwork to be the product of an intentional action. For instance, Christy Mag Uidhir claims that this position is almost universally accepted in the literature. He argues that any artwork must be the product of a successful attempt to make art (Uidhir, 2013). Moreover, the requirement of intentional creation is also highly influential in the psychological literature. Paul Bloom has extended Levinson’s theory of art to all kinds of artifacts, claiming that an object becomes a member of an artifact kind if it is created with the intention for it to belong to that kind (Bloom, 1996).

Second, properties emphasized by these three definitions of art also usually figure in the cluster and disjunctive theories, and they are usually among the first to be discussed. For example, they are part of both Gaut’s (2000) and Dutton’s (2009) sets of properties that “count towards being art”. The same can be said about Longworth and Scarantino’s Disjunctive Theory of Art (2010) because it borrows properties from Gaut’s list. Our three key properties also correspond to the three disjuncts in Stephen Davies’ definition of art (2015) and are similarly emphasized by Stecker (2000) and Hazelwood (2021).

The third reason for choosing intentional creation, beauty and institutional recognition as the key factors of our research regards extant empirical work, in which they also figure prominently. We will discuss this point in further detail in the following sections, in particular Sect. 5.

4 The role of empirical research

Many philosophers of art claim that the project of defining art is descriptive, i.e. that any plausible definition must not stand in tension with how art is commonly thought of (e.g. Kania, 2011: p. 5, Stecker, 2000: p. 60) or that it must “cohere with a wide spread of intuitions about the terms in which art is discussed and interpreted” (Davies, 1991: p. 46).

An empirical inquiry into the folk concept of art can fulfil two tasks: (i) It can reveal the structure of the concept of art (e.g. whether it is essentialist or not), And (ii) it can settle the question of descriptivists, namely which– if any– of the many proposals from the philosophical literature on offer is consistent with the folk concept of art. Both strands of the inquiry are of key importance for philosophical debates about art (Cova et al., 2015). According to Pignocchi, the work of cognitive scientists and the work of philosophers who seek descriptive definitions are strikingly similar, except that philosophers rely on their own intuitions while cognitive scientists test them with a wider set of participants (2014: p. 427). Denis Dutton, for example, claims that it is the folk concept of art that is being analysed, and the folk concept guides further expert investigation:

The cluster criteria tell us what we already know about the arts. They may be adjusted at the edges, with items subtracted or added to it, but they can be expected to remain largely intact into the foreseeable future, governing what counts as investigation into the arts by neurophysiologists, philosophers, anthropologists, critics, or historians. (Dutton, 2009: p. 60)

concepts↗concepts↗sensesconcept_edgesIn contrast to descriptive accounts, there are views that provide a normative, or partially normative, definition of art (e.g. Gaut, 2000: p. 39). Definitions of this sort aim to describe not how art is actually thought of, but how it should be thought of, or what the concept of art really consists of (whether or not the folk agree). In this case, empirical findings may also be important for philosophers seeking to create normative definitions. Empirical findings can provide philosophers with a background against which the normative theories might be created and refined; they uncover problems and inconsistencies found within actual usage (Monseré, 2016: p. 15). Any normative project must start with a descriptive step, as it is impossible to revise a concept without knowing how it actually functions (Meskin et al., 2018). A philosopher offering a normative definition must explain, in a first step, what people’s intuitions are, and only then can they offer an explanation of why said intuitions are mistaken (see also Pignocchi, 2014).

To date, empirical work in aesthetics has mostly focused on topics such as folk objectivism and subjectivism about aesthetic judgements (i.e. whether the folk see aesthetic judgments as objective or subjective; Cova & Pain, 2012; Rabb et al., 2020; Cova et al., 2019; Bonard et al., forthcoming), imaginative resistance (i.e. psychological difficulties in engaging in certain imaginative activities; Black & Barnes, 2017, 2020; Kim et al., 2019; Liao et al., 2014), aesthetic adjectives (terms that are used to communicate aesthetic judgments; Liao & Meskin, 2017; Liao et al., 2016), the ontology of musical works (for instance, the question of what it means to play the same musical work twice; Bartel, 2018; Mikalonytė, 2022; Mikalonytė & Dranseika, 2020, 2022; Mikalonytė & Canonne, 2023; Puy, 2022), the interaction between aesthetic and moral qualities (Doran, 2021), and other topics (for a review, see Cova (forthcoming)). However, to date, there is very little empirical research on the folk concept of art.

Mikalonytė and Kneer (2022) have investigated intuitions concerning art created by artificial intelligence. When presented with stories about either an AI system or a human creating a painting and asked to what extent they agree with the claim that the painting is art, people expressed similar levels of agreement regardless of whether the painting was made by AI or a human creator. However, despite this, participants were not very willing to call AI-creators artists, and they refused to attribute artistic intentions to them. These results suggest that an object does not have to be created intentionally to be considered art. However, many other factors that figure in the most influential definitions of art have not been tested in that study.

concepts↗concepts↗concepts↗concept_edgesconcept_edgescontent_tags↗Liao et al. (2020) claim that the ordinary concept of art is a dual character concept: a concept that has both descriptive and evaluative dimensions. In their interesting paper, they present evidence that the concept art behaves distinctly from clear-cut evaluative concepts such as masterpiece or stroke of genius on the one hand, and also from clear-cut descriptive concepts such as comic or sonnet.

There exist two studies that were designed specifically to put various definitions of art to the test (Kamber, 2011; Kamber & Enoch, 2018). Kamber (2011) ran an online survey with a large number of objects (photographs, paintings, poems), asking participants– the vast majority of whom were art professionals– whether they considered them art. Most objects were “hard cases” discussed by philosophers of art. Kamber found many discrepancies between the intuitions of art professionals and philosophers’ claims about hard cases: for instance, almost half of the art professionals in the study categorised a Bugatti car as art. He concluded that none of the existing philosophical art theories succeed in tracking lay intuitions.

In another study, Kamber and Enoch (2018) presented their participants with the same set of objects and asked them to categorize the objects as art or not art. In addition, they asked the participants to consider a set of fourteen possible reasons for their judgments. The latter included that the object was made by a conscious agent, intended by its makers to be an object of aesthetic interest or appreciation, intended by its makers to evoke or communicate intention, or that the object invites interpretation, provides an imaginative experience, or is beautiful (Kamber & Enoch, 2018: p. 21). Although intentional justifications were the most popular, the results suggest that none of the tested art definitions are fully in accordance with intuitions of art professionals. Kamber and Enoch note that Beardsley’s aesthetic definition of art is more successful than the others. Based on Kamber’s findings, Winner (2018) has argued that the concept of art is open, as proposed in a Wittgensteinian manner by Weitz (1956): there are no necessary properties, just family resemblances, and the conditions for the application of the term might even change over time.

In both studies conducted by Kamber, the majority of the participants were art professionals (2018) or half art professionals, half “art buffs” (both groups consisted mainly of university faculty members) (2011). As such, these studies do not speak to the folk concept of art. Moreover, the studies have been described as employing an “informal approach” of testing a large number of definitions using only one or two instances of artworks (Winner, 2019: p. 21). Thus, although these results are interesting and important, the actual folk concept of art remains largely unexplored.Footnote 5

5 Empirical work relating to the three core factors

Let us take a deeper look at three factors– intentional creation, aesthetic value, and institutional recognition– and at the cognitive mechanisms that could lie behind the folk concept of art. Each of these three factors enjoys some support from the psychological literature and from psychological studies that were conducted with a view to exploring only one or two aspects of the folk concept of art. We will survey the literature relevant to each of these factors: aesthetic value (5.1.), institutional recognition (5.2.), and intentional creation (5.3.).

5.1 Aesthetic value

mechanisms↗Alessandro Pignocchi hypothesized that the concept of art seems intuitively appropriate in cases where people consider an artifact as intended to fulfil a function that other artworks typically fulfil (2014: p. 439). Two factors, he claims, are key: the functions that the judging person (e.g. the study participant) accepts as being normally fulfilled by art, and the intentions that they ascribe to the creator of that artifact. Although Pignocchi mentions several possible functions of art, we will limit ourselves to what Davies claims to be the most typical– that of inducing aesthetic experiences.

mechanisms↗Empirical research has shown that artifacts are often categorized in terms of functions and intentions (Boyer & Barrett, 2004). Function-based explanations of artifacts, and even natural objects, are held to be a cognitive default (Kelemen, 1999; Kelemen et al., 2013) inherent in the reasoning of children (DiYanni & Kelemen, 2005) and adults (Kelemen & Rosset, 2009), also in non-Western cultures (Schachner et al., 2017).

An additional motivation to test the aesthetic definition derives from Kamber and Enoch’s (2018) results, which suggest that Beardsley’s aesthetic definition of art is more successful than the others. If the folk concept of art is not radically different from that used by art professionals, we might expect their intuitions to favour the aesthetic definition. Moreover, research by Pelowski et al. (2017) shows a strong correlation between the classification of objects as art and ratings of liking, which suggests that appraisal and classification are interrelated. These results also offer some prima facie support for the aesthetic definition.

5.2 Institutional recognition

mechanisms↗McCallum et al. (2019) offer an explanation of how the institutional theory of art might work from a cognitive point of view. They claim that the institutional theory matches the main claims of relevance theory (Sperber & Wilson, 1986). Presentation of an object as a work of art (such as putting it into a gallery) is an ostensive act. It suggests for the viewer two kinds of authorial intentions: not only informative (to communicate some information to the audience), but also communicative (an intention to point out the informative intention to the audience). According to McCallum and colleagues, institutional recognition and presentation “as a candidate for appreciation” is a kind of ostensive act which not only intentionally draws attention to an object, but also gives a reason to think that interpreting the object is worthy and will provide worthwhile cognitive effects (2019: 25). This theory, then, points to both institutional and intentional factors in art status ascriptions.

Kamber’s studies suggest that both institutional recognition and aesthetic value might be important in judgments on what is art. However, in the 2011 study, more than half of the people were willing to categorize an aesthetically bad painting as art, while an even bigger number of participants categorized works as art that were institutionally recognized but unlikely to provoke aesthetic experiences. These results suggest that neither the aesthetic nor institutional theory fully tracks folk intuitions on what is art on their own.

5.3 Intentional creation

mechanisms↗Inferences about the mental states of creators are as important in categorizing works of art as they are for other artifacts. According to Bloom (1996), authorial intent is relevant to determining the kind of an artifact even if it is not directly connected to the object’s appearance or function. Thus, intuitions on whether an object was created with the intention for it to belong to the category of artworks must play an important part in judgments on whether an object should be categorized as art.

Several empirical studies have suggested that intention is essential for the way we think about art. Intentions have been found to be crucial for what people categorize as art (Jucker et al., 2014; Newman & Bloom, 2012)– even more important than the appearance of the object (Newman & Smith, 2018). Moreover, in Kamber’s studies, when asked to provide reasons for categorizing an object as art, participants mentioned “being made by a conscious agent”, “being intended by its makers to be an object of aesthetic interest”, and “inviting interpretation” more often than other kinds of justifications for their judgments (Kamber, 2018: p. 81). If ordinary folk deploy the same concept of art as art professionals, intentions must be an important factor. However, in another study on folk intuitions as to whether AI can create art, Mikalonytė and Kneer (2022) found that intentional creation is not seen by non-experts as a necessary condition for an object to be considered art. These results reveal that the exact role of intentional creation in judgments of what counts as art is still not fully understood.

6 Experiment 1: painting

concepts↗content_tags↗Our first experiment explores the extent to which the three most common factors in definitions of art– intentional, aesthetic, and institutional– are consistent with the folk concept of art. We thus manipulated all three features in a between-subjects design, i.e. (i) whether or not the object was created intentionally, (ii) whether or not it is beautiful,Footnote 6 and (iii) whether or not it received institutional recognition. As a second core dependent variable, we explored to what extent people were willing to deem the creator of the different types of objects an artist. Furthermore, we wanted to explore the mental-state requirements as well as their possible interaction with beauty and institutional recognition and thus asked participants to what extent the creator desired, believed, and intended to make a work of art.

claim_relationsAt a more general level, our study, as described, explores (H1) whether any of the three features constitute necessary and/or sufficient properties for something being art, and (H2) whether the folk concept of art is essentialist, or whether it is a cluster or disjunctive concept. In contrast to Kamber’s studies, our participants are not art professionals or amateur-experts, but, rather, ordinary people. The online appendix, as well as materials, data, and preregistrations of this and the second experiment can be found on the project’s OSF page, https://osf.io/96f57/.

6.1 Participants

content_tagsWe recruited 797 participants on Amazon Mechanical Turk. The IP address was restricted to the United States. In line with the preregistered criteria,Footnote 7 336 participants who were not native English speakers, failed an attention check or took less than twenty seconds to answer the main questions (including reading the prompt) were excluded, leaving a sample of 461 participants (female: 48%; age M = 44 years, SD = 14 years, range: 19–90 years).

6.2 Methods and materials

method↗concepts↗content_tags↗content_tags↗content_tags↗We explored peoples’ concept of art in a 2 (intention: intentional v. accidental creation) × 2 (beauty: beautiful v. ordinary-looking) × 2 (institutional recognition: present v. absent) between-subjects study. There were thus 8 conditions in total, to one of which participants were randomly assigned. The scenarios were composed of the following elements (labels in bold excluded, stating either the presence ( +) or absence (−) of the key features):

Intention

[+] A person decides to create a painting. She takes an empty canvas and applies paint onto it.

[−] A person accidentally brushes against some jars of paint that spill onto an empty canvas.

Beauty

[+] The resulting object looks beautiful, featuring an elegant interplay of different lines of paint. It captures the viewers’ attention and evokes awe and wonder.

[−] The resulting object looks ordinary and uninteresting. It leaves the viewers bored and unimpressed.

Recognition

[+] Soon this object gets recognized by art critics, finds its way into a museum and some years later it appears in art history books.

[−] This object never gets exhibited in art galleries or museums, and it never receives any attention from art critics.

Having read the scenario, participants had to rate to what extent they agreed with the following claims on a Likert scale anchored at 1 with “completely disagree” and 7 with “completely agree” (labels in brackets excluded):

(1) “The object is art.” [Art]

(2) “The object was made by an artist” [Artist]

(3) “The person wanted to make a painting” [Desire]

(4) “The person believed they were making a painting” [Belief]

(5) “The person intentionally made a painting” [Intent]

6.3 Results

content_tagscontent_tagscontent_tagscontent_tagsA between-subjects ANOVA for art determined that participants were more inclined to judge the painting to be art if it resulted from intentional action (F(1,453) = 59.02, p < 0.001, ηp2 = 0.12, a medium-large effect), if it was beautiful (F(1,453) = 43.52, p < 0.001, ηp2 = 0.09, a medium-sized effect), and if it received institutional recognition (F(1,453) = 7.67, p = 0.006, ηp2 = 0.02, a small effect). No significant interactions were observed (all ps > 0.255). As suggested by the ANOVA main effects and illustrated in Fig. 1, the impact of institutional recognition was marginal. What is interesting, however, is that– even in the absence of recognition– either beauty or intentional action suffice for judgments to significantly exceed the midpoint (one sample t-tests, all ps < 0.009, see Online Appendix, Table 6).Footnote 8 Differently put, on the folk view, one can accidentally create genuine art if the creation is beautiful. One can also intentionally create genuine art if the creation is not beautiful. But if intentionality and beauty are lacking, institutional recognition does not, by itself, suffice to make the work a piece of art.

content_tagsAs regards the creator’s status as an artist (see Fig. 2), people were more willing to judge her as an artist if the painting was made intentionally (F(1,453) = 49.04, p < 0.001, ηp2 = 0.10, a medium-sized effect), whereas the painting’s being beautiful or receiving institutional recognition did not influence the judgment of the creator’s status as an artist (ps > 0.480). The results revealed small but significant interactions between intention and the other two factors (ps < 0.035, ηp2s < 0.03). All other interactions were nonsignificant (all ps > 0.074 see Online Appendix, Table 3). In contrast to the object’s status of art, being deemed an artist requires acting intentionally: The creators of accidental artworks, even if beautiful and enjoying recognition, are not deemed artists (significantly below the midpoint 4 for all four accidental conditions, one sample t-tests, all corrected and uncorrected ps < 0.001, see Online Appendix, Table 6).

content_tagsThe manipulation checks suggest that participants understood the scenario as intended. Aggregating across beauty and institutional recognition, attributions of belief, desire, and intention to make a painting in the intentional condition significantly exceeded those in the accidental condition (all ps < 0.001, all ηp2 > 0.35, indicating large effects, see Online Appendix, Tables 7–9 for detail). All three mental states correlated strongly with the judgment that the resulting object was art (r = 0.57-0.59; all ps < 0.001; see Online Appendix, Table 4) and judgments that its creator is an artist (r = 0.76-0.81, all ps < 0.001).

6.4 Discussion

sensesclaim_relationsclaim_relationsOur experiment produced several findings. First, among the three manipulated factors, intentionality and beauty manifested a significant and pronounced impact on art judgments, whereas the effect of institutional recognition was significant yet small. Importantly, the creation was judged art (significantly above the midpoint of the scale) if it was specified either as intentionally created or beautiful. This suggests that, on the folk view, neither intentionality nor beauty are necessary features of an artwork (as suggested by Denis Dutton and Berys Gaut). However, each of them is sufficient (at least for the scenario at hand). Our results are consistent with Liao et al. (2020): the ordinary concept of art seems to be associated with both descriptive (institutional) and evaluative (aesthetic) criteria. The results are also consistent with Mikalonytė and Kneer (2022), who found that intentionality is not a necessary condition for being art.

By contrast, the status of being viewed as an artist depended only on whether the object was created intentionally. Whether or not the object was beautiful or recognized institutionally did not matter. Our finding that beautiful accidental creations are art, even if they have no artist, is very surprising, bearing in mind that there is no artist to claim credit for aesthetically valuable artworks created in this way.

Paintings are not the only objects of art. In order to explore whether our findings generalize to another domain of artistic creation, we ran a similar, second experiment for music.

7 Experiment 2: music

In Experiment 2, we want to replicate the findings of Experiment 1 for another aesthetic domain. We thus adapted the vignette to the context of music, again manipulating intentionality, beauty, and institutional recognition.

7.1 Participants

content_tagsWe recruited 714 participants on Amazon Mechanical Turk. The IP address was restricted to the United States. In line with the preregistered criteria,Footnote 9 287 participants who were not native English speakers, failed an attention check, or took less than twenty seconds to answer the main questions (including reading the prompt) were excluded, leaving a sample of 427 participants (female: 48%; age M = 43 years, SD = 13 years, range: 21–77 years).

7.2 Methods and materials

The design of the experiment was near-identical to the one employed in Study 1. The only difference concerned the mode: This time, the agent does not create a painting but a musical work. We again manipulated intention, beauty, and recognition, making for a total of eight individual conditions to which participants were assigned randomly. The scenarios were composed of the following elements (labels in bold excluded, stating either the presence (+) or absence (−) of the key features):

Intention

[+] A person decides to compose a piece of music. She opens a blank stave sheet on a music notation software, writes notes on it, and carefully chooses instrumentation.

[−] A person uses a new music notation software for the first time. She opens a blank stave sheet, writes random notes on it, and chooses random instrumentation.

Beauty

[+] The result sounds beautiful, featuring an elegant interplay of parts. It captures the listeners’ attention and evokes awe and wonder.

[−] The result sounds ordinary and uninteresting. It leaves the listeners bored and unimpressed.

Recognition

[+] Soon the piece gets recognized by music critics, is performed in concert halls and some years later it appears in music history books.

[−] The piece never gets played in concert halls, and it never receives any attention from music critics.

content_tags↗As in Experiment 1, participants were asked to rate on a 7-point Likert scale whether the resulting work is art, its creator an artist, and whether the latter had a belief, desire, and intention to produce the piece.

7.3 Results

content_tagscontent_tagscontent_tagsFor art (see Fig. 3), an ANOVA revealed a small main effect of intention (F(1,419) = 14.76, p < 0.001, ηp2 = 0.03), a medium main effect of beauty (F(1,419) = 59.10, p < 0.001, ηp2 = 0.12) and a small main effect of institutional recognition (F(1,419) = 14.77, p < 0.001, ηp2 = 0.03). There was a significant yet small beauty*recognition interaction (F(1,419) = 9.47, p = 0.002, ηp2 = 0.02). As in Experiment 1, even in the absence of recognition, either beauty or intentional action (by and large) suffices for a creation to be deemed art, as none of the ratings are significantly below the midpoint (for corrected and uncorrected one-sample t-tests, see Online Appendix, Table 15). In contrast to visual art, in the context of music, an accidentally produced, non-beautiful piece is viewed as art if it receives institutional recognition (M = 4.47, SD = 1.68, uncorrected p = 0.041, corrected p > 0.05, see Online Appendix, Table 15).

As regards the creator’s status as an artist (see Fig. 4), an ANOVA revealed that people were more willing to judge the creator as an artist if she acted intentionally (F(1,419) = 23.41, p < 0.001, ηp2 = 0.05, a small to medium effect), if the piece was beautiful (F(1,419) = 24.07, p < 0.001, ηp2 = 0.05, a small to medium effect), and if it was institutionally recognized (F(1,419) = 5.83, p = 0.016, ηp2 = 0.01, a small effect). There was also a small yet (just about) significant beauty*recognition interaction (F(1,419) = 3.90, p = 0.049, ηp2 = 0.01). All other interactions were nonsignificant (all ps > 0.169, see Online Appendix Table 12). In contrast to the domain of painting (Experiment 1), the requirements regarding the status of an artist are less stringent for the domain of music. Independently of recognition, those who produce musical works are deemed artists as long as the resulting works are either made intentionally or beautiful (significantly above the midpoint 4, one-sample t-tests, all uncorrected ps < 0.006, see Online Appendix, Table 15).Footnote 10

As in Experiment 1, the manipulation checks suggest that participants understood the scenario as intended. Aggregating across beauty and institutional recognition, attributions of belief, desire, and intention to produce a musical work in the intentional condition significantly exceeded those in the accidental condition (all ps < 0.001, all ηp2 > 0.04, see Online Appendix Tables 16–18 for detail).

All three mental states correlated strongly with the judgment that the resulting object was art (r = 0.43–0.48, all ps < 0.001) and its creator an artist (r = 0.61–0.62, all ps < 0.001, see Online Appendix Table 13 for detail).

7.4 Joint analysis

content_tagscontent_tags↗We conducted joint analyses for the entire sample of the two experiments. This allowed us to make artistic domain (visual art v. music) a fourth factor beyond intentionality (intentional v. accidental), beauty (yes v. no), and institutional recognition (yes v. no). For art judgments, we found a small though significant main effect for artistic domain (p = 0.029, ηp2 = 0.01). The beauty*recognition and the intention*domain interactions revealed significant though small effects (ps < 0.009, ηp2s = 0.001), for detailed results, see Online Appendix, Table 20). For artist judgments, we found a small though significant main effect for artistic domain (p = 0.008, ηp2 = 0.01). The interactions for beauty*recognition, beauty*domain, recognition*domain and intention*beauty*domain revealed significant though small effects (ps < 0.046, ηp2s = 0.001), for detailed results, see Online Appendix, Table 21. Overall, we found a significant yet rather small impact of domain on both art and artist judgments.Footnote 11

7.5 Discussion

phenomenonConsistent with Experiment 1, beauty had a considerable effect on art judgments, whereas institutional recognition, again, had but a small effect. As concerns intentionality, the findings differed somewhat across experiments: This factor had a strong effect on judgements of art in the domain of visual art, whereas its effect was considerably smaller in the realm of music. Interestingly, each of the three factors sufficed for an object to be judged as art. Only in the absence of all three factors was mean art attribution significantly below the midpoint.

limitationsconcepts↗concepts↗sensesconcept_edgesOur findings suggest that the folk concept of art might be a cluster or a disjunctive concept, but it is not evident which of the two accounts is correct from an empirical point of view. In the literature, cluster and disjunctive definitions are usually quite similar (Davies, 2004a, 2004b; Longworth & Scarantino, 2010). What distinguishes them is that a cluster concept, in contrast to a disjunctive one, is not governed by a definitive set of criteria for application. The fact that we have only tested three criteria is an evident limitation of our study. What is more, if the criteria of application can change in the future, as hypothesized by some, empirical studies of the sort conducted here will always be partially incomplete (Cook, 2013; Weitz, 1956).

unhoused_relationscontent_tagsThat said, we can draw some very tentative lessons from the data as regards the two non-essentialist accounts under consideration. Consider a contextualist interpretation of cluster concepts, akin to cluster accounts of the concept of knowledge in (see e.g. DeRose, 1992; Cohen, 1999; Stanley, 2004). If the difference between cluster and disjunctive concepts is essentially a question of whether their conditions of application are context-sensitive, then our results show some support for the cluster theory. The application conditions for the concept of art change across domains (painting v. music). In our study concerning music, institutional recognition was a sufficient criterion for an object’s being considered art, whereas in the study concerning visual art, it was not.

As regards the status of being an artist: In the context of visual art, only intentionality mattered; in the context of music, all three manipulated factors proved significant. Intentionality or beauty of the object was sufficient by itself for people to deem the creator an artist, and only in the case where the object lacked all three tested properties did people clearly refrain from judging the creator as an artist. While it is not surprising that being an artist and creating intentionally is tightly related, we found that, in some cases, intentional action is not a necessary condition for the folk to consider someone an artist.

8 General discussion

8.1 Overview of findings

concepts↗concepts↗concept_edgesAt the beginning of the paper, we distinguished between essentialist accounts of art on the one hand, and two particular non-essentialist accounts – the cluster and disjunctive views – on the other. Essentialism requires a set of individually necessary and jointly sufficient conditions for an object to be art. According to cluster and disjunctive accounts, by contrast, an object can constitute art if it manifests one or several individually nonessential properties from a set of features.

Our empirical inquiry has explored (Q1) whether any of the three properties are seen by the folk as necessary and/or sufficient properties for an object’s status of art, and, more generally, (Q2) whether the folk concept of art is an essentialist concept.

content_tagsAs regards the first question (Q1), we found that none of the three properties proved necessary: Mean art ratings exceeded the midpoint of the scale in both the visual art and musical context if the object manifested only beauty or only intentionality. Correlatively, this suggests that either property was sufficient for an object to be deemed art.Footnote 12 Beauty exerted a substantial impact on art judgments in both artistic contexts, whereas the impact of intentionality was less pronounced in the musical context. Institutional recognition was clearly the least important feature of art: In the visual art context, it proved insufficient by itself (though in the musical context, mean art ratings for objects which only had the feature of recognition exceeded the midpoint and just about reached the significance threshold). Furthermore, we detected an interesting interaction between beauty and recognition: For nonbeautiful objects, the presence of recognition tended to substantially raise the level of art judgments.

unhoused_relationsGiven that none of the three properties that feature most prominently in the literature capture a necessary feature of the folk concept of art, the folk concept of art might not be essentialist in nature (as stated, of course one could test further potential properties). Instead, it seems to be either a cluster or a disjunctive concept, and all three features tested seem to be part of the cluster of properties that can, by themselves or in conjunction with others, confer the status of art. According to our findings, in each context, beauty suffices to confer the status of art. Intentionality does in the visual context and is trending in the musical context. Recognition is (just about) sufficient by itself in the musical context, though does not suffice when it comes to paintings. However, recognition exerts a significant influence on art judgments when the object lacks beauty. Both the context-sensitivity of the factor’s relative impact and the significant beauty*recognition interaction provide some support for the hypothesis that the folk concept of art is a cluster concept: Consistent with what cluster theories would predict, the impact of each factor can change from situation to situation. The concept of art is sensitive both to artistic domain and the presence or absence of other features. In different contexts, different criteria might be relevant for categorising objects as art or not art.

claim_relationscontent_tagsWe have also gathered data on the question under what conditions a creator is deemed an artist. Broadly speaking, whether or not an agent is considered an artist does – expectedly – track whether or not the object they created is deemed art. However, and this is an interesting and surprising finding, in the visual context, we found a divergence: Beautiful, accidentally produced creations (recognized or not) are considered art, though their creators are not considered artists. What this suggests, at least prima facie, is that people are willing to attribute the status of art to objects that were created without a creative intention and not created by artists. This is surprising for several reasons: First, most authors in the philosophy of art assume that whatever is the real nature of art, it must be the result of intentional action (Dutton, 2009; Mag Uidhir, 2013). Artworks are almost always described as artifacts, which entails that they must be the product of intentional creation. The role of intention for art status and art interpretation has been emphasized extensively not only by philosophers but also in the psychological and empirical literature (Bloom, 1996; Jucker et al., 2014; Kamber & Enoch, 2018; Newman & Bloom, 2012; Newman & Smith, 2018).

Second, quite obviously, being an artist and being the source of art-making intentions is closely related, because any object that is the product of intentional action must have a creator. For instance, according to Mag Uidhir, “if what it is to be an artwork is to be the product of a successful art-attempt, then presumably what it is to be the artist of a particular artwork is to be the source of the intentions directing the actions constitutive of the successful art-attempt of which that particular artwork is the product. From this, it follows that an artwork must have an artist.” (2013: 45). According to the standard view, then, art must be intentionally made, and it must have an intentionally acting creator. That people are willing to recognize artworks that do not have an artist might sound paradoxical at the outset. However, it is less so when we think of art created by non-human agents, such as artificial intelligence. Although AI-driven robots are standardly not deemed artists, their creations (in experiments similar to the ones here presented) frequently are considered art nonetheless (see Mikalonytė & Kneer, 2022). One could wonder if people simply ascribe the creative intention to the human creator of AI instead of the AI-driven machine itself and consider the human creator the source of artistic intentions. However, the results of the current study show that people are open to the idea of art being created without a creator in a more general sense.

mechanisms↗concepts↗Although our results suggest that creative intention at the moment of an action that results in a work of art is not necessary, it is hard to imagine intentionality being entirely absent from art. In similar cases, intentionality often becomes relevant during the artwork selection or curation process. One familiar example is the creative process of AI-generated art, where the human user of the system is often presented with multiple outputs and selects the best one. Another example is animal art or human creations made by accident, both of which can be seen as instances of art made without artistic intention. However, the pieces ultimately presented to the audience are not chosen randomly but are often intentionally selected by humans. In the context of art created without a creative intention, intentionality is unlikely to disappear completely, at least not from the process of conferring art status: artistic intention here is replaced by curatorial intention. We leave these interesting, more sophisticated aspects of intentionality to be explored in future research.

8.2 Limitations

While we have delivered substantial evidence that the folk concept of art is a cluster concept and not an essentialist one, much work is needed to elucidate said concept in more depth.

limitationsFirst, our inquiry is limited to only two artistic contexts (painting and music). Given that even across these two contexts we detected some differences, it stands to reason to explore a wide variety of further artistic domains, and ideally employ several distinct scenarios for each to test for potential intra-domain variation. This would also help to ensure that differences found between artistic contexts have not been determined by differences between scenarios or perhaps differences in the plausibility of the stories.

limitationsSecond, as one reviewer has helpfully pointed out, a full factorial design with three distinct factors (and thus eight individual conditions), could give rise to some combinations of factors that might sound less plausible than others. It might thus be helpful to also test each factor in separate experiments, maximizing plausibility for both the absence and presence condition of the factor at issue.Footnote 13 It might also be helpful to rerun the full factorial design with extensive manipulation checks (ours are limited to intentionality).

limitationsThird, future experiments should randomise the order of the three (or more) factors. Since information about intention and beauty was always presented earlier than institutional recognition in our scenarios, order effects cannot be ruled out.

Fourth, although we have tested several properties that have prominently featured in both essentialist and disjunctive or cluster accounts of art, the net should be cast wider. Dutton, for instance, lists (though does not limit himself to) twelve distinct criteria in his cluster account; (1) direct pleasure, (2) skill and 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, (12) imaginative experience, and, in addition, intentional creation as a necessary condition (Dutton, 2009).

Fifth, in future research, it would be useful to make a clearer distinction between the intent to produce something that falls under the concept of art and other intentions of the appropriate kind that do not involve having a concept of art– for instance, an intention to create a rhythmical sound structure or to sketch something resembling a horse (see Lopes, 2007).

limitationsclaim_relationsSixth, empirical research in the social sciences tends to oversample populations from WEIRD (Western, Educated, Industrialized, Rich, and Democratic) countries in general, and US Americans in particular (Henrich et al., 2010, see also De Block & Kelly, 2022). Some have argued for a surprising stability of experimental-philosophical findings across populations and cultures (Knobe, 2019, 2021), citing studies that report cross-cultural convergence on topics ranging from the metaphysics of free will (Sarkissian et al., 2010), central intuitions in epistemology (Rose et al., 2019) and intentionality attribution (Lin et al., 2019), norms of assertion (Kneer, 2021), to key concepts in philosophy of law (Hannikainen et al., 2021, 2022). Others have countered this contention (Stich & Machery, 2022), pointing to studies where results vary strongly across cultures. In the experimental philosophy of aesthetics, there is but one study to date which reports some, yet limited, divergence concerning intuitions of intersubjective validity in aesthetic judgments across 19 different countries (Cova et al., 2019). A systematic inquiry into the folk concept of art thus requires cross-cultural experiments.

Seventh, given potentially diverging reasons for identical or similar responses, quantitative research of the sort reported in this paper might need to be supplemented with qualitative data, asking participants why they responded as they did. Given technological advances, corpus analysis constitutes another possibility (for examples see Baumgartner & Kneer, 2025; Sytsma et al., 2019; Sytsma & Fischer, 2023; Reuter & Baumgartner, 2024; for general discussion, see e.g. Andow, 2016; Chartrand, 2022; Fischer & Curtis, 2019; Plou, 2023).

Eighth, novel experiments are needed to better distinguish between cluster and disjunctive definitions. Future experiments can proceed in two directions: (1) further explorations of the context-sensitivity of the concept of art, and (2) the possibility of new cases: under what conditions might the current application conditions be reconsidered?

limitationsFinally, whereas the vignette-based approach has its strengths as it allows experimenters to tightly control the manipulated factors, one might argue that an essential part of aesthetic evaluation lies in a first-hand aesthetic experience, and the present studies do not involve such an experience. After all, textual vignettes leave out many details which must be filled in by a reader in her imagination. Readers tend to choose different details, which leads to different intuitive verdicts. However, this problem can– at least in the domain of aesthetics – be solved by using visual (or aural) stimuli (Weinberg, 2018: p. 237). It has been demonstrated in experimental philosophy of music, for instance, that using acoustic stimuli in addition to text vignettes (vs. text vignettes only) influences people’s intuitions (Puy, 2022). Hence, it might be helpful to include actual or potential works of art among the stimuli to explore the robustness of the findings here presented.

9 Conclusion

Whereas it is notoriously difficult to pick out the correct definition of art, the broadest question in the literature on art theories – and also the question we aimed to investigate in this study – concerns the type of concept the expression “art” most plausibly denotes. Can art be defined by individually necessary and jointly sufficient conditions or by a disjunction of minimally sufficient conditions? If the latter is true, is the set of conditions determinate or not?

claim_typecontent_tagsOur initial exploration of the three most widely discussed factors – intentional creation, aesthetic value and institutional recognition – showed that none of them is deemed individually necessary. Although further factors, such as emotional expressivity or formal complexity, could be explored in follow-up studies, our findings constitute tentative evidence that the folk concept of art is not an essentialist concept and that it is a cluster concept instead. If the correct account of art must be consistent with the folk concept of art – and there are good reasons to assume that this is the case– the cluster theory of art is the most promising one we have at our disposal.