NOT MUSIC, BUT MUSICS: A CASE FOR CONCEPTUAL PLURALISM IN AESTHETICS

We argue for conceptual pluralism about music . In our view, there is no right answer to the question ‘What is music?’ divorced from some context or interest. Instead, there are several, non-equivalent music concepts suited to different interests – from within some tradition or practice, or by way of some research question or field of inquiry. We argue (1) that unitary definitions of music are problematic, (2) that the role music concepts play in various research questions should motivate conceptual pluralism about music, and (3) that taking music pluralism seriously grounds a fruitful research programme in aesthetics. We suspect that pluralism about music is agood test case for the utility of pluralism in aesthetics more generally, and we present it as such.


I. INTRODUCTION
It is uncontroversial that music is multi-faceted: from medieval monophonic plainchant to the heterophony of Javanese gamelan, from rigorous military marches to passionate ritual drumming, from a simple, subtle lullaby to the complexities and cacophonies of the avant-garde, music manifests itself in extraordinary heterogeneity. Music's theoretic features are also highly diverse. Take Western art music. Keys, in the context of a composition, produce a variety of hues: from sombre minors and triumphant majors, to grieving blue pentatonics and enigmatic aeolians. Whole chromatic systems, families of keys, themselves diverge both in expressive range and formal structure. Moreover, music plays many roles: it provides catharsis, ceremonial backdrops, pedagogical props, social glue, and individual inspiration. Unsurprisingly then, the investigation of music proceeds from a bewildering array of angles. The music theorist might revel in music's stark mathematical structures, the composer in its expressive power, the anthropologist in its ubiquity and variability, the neuroscientist in its multi-modal complexity, and the metalhead in virtuoso technique and fine-grained genre distinctions.
It is our contention that music's multi-faceted nature -at least in terms of music concepts -runs deep. That is, in our view, there is no fact of the matter about 'What music is' -no unitary answer to the question 'What is music?' -divorced from some angle, some interest. In this article, we argue for conceptual pluralism about music. 1 A conceptual pluralist about some category believes that there are at least two, non-equivalent, equally legitimate concepts pertaining to that category. We understand a'category' to be an ontological notion: a category is a way the world is carved. Given that we are primarily interested in concepts rather than categories here, we leave questions about realism and natural kinds -questions about whether there are mind independent categories, or whether the world comes to us 'pre-carved' -aside. A concept, among other things, decides category membership. A music concept, then, decides which events, entities, or processes count as music, and which do not. 2 For a monist about music, there is one true music concept; there is a tout court right answer to the question 'What is music?' 3 The music pluralist denies this, arguing that music ought to be carved up in different ways -by different music concepts -in different circumstances. Our argument for conceptual pluralism in discussions about music could be taken as a case study for aesthetic concepts more broadly. Indeed, we shall suggest that pluralism has the potential to provide a framework for fruitful research directions in aesthetics. Our approach diverges from traditional philosophical approaches to music. Conceptual analyses of music typically take one of two forms: reductive analyses encompassing individually necessary and jointly sufficient identity conditions, or multidimensional, though unitary, analyses comprising 'salient features' . We argue that these approaches fail, and this failure, in part, motivates our pluralism.
Pluralism, as we discuss it, makes available a methodological programme for aesthetics. This programme involves identifying legitimate research interests as well as characterizing, critiquing, and perhaps designing concepts of music appropriate for those interests. Recently, Jonathan McKeown-Green has argued that the intuition-driven methodology prevalent in the philosophy of music (and aesthetics more generally) is problematic because what counts as music is not narrowly determined by how we conceive of it. As he says, 'Exclusive appeals to intuitions and classificatory practices only work if the nature of the thing being defined is determined by our conception of it, that is, by the way we construe it. ' 4 For McKeown-Green, what music is cannot be defined by our current conception Not Music, but Musics: A Case for Conceptual Pluralism in Aesthetics 2 There are many ways of considering concepts; we focus on a functional notion. In other contexts, we might consider concepts as psychological objects, and pluralism can matter here as well. See Daniel A. Weiskopf, 'The Plurality of Concepts' , Synthese 169 (2009): 145-73. 3 That said, a monist might argue that 'music' is a vague term (of which a paradigm example is 'heap') and deny that there is an answer to 'Is this music?' questions in vague cases. As far as we are aware, this position has yet to be defended about music, and will be set aside here. because conceptions change through time (we draw apart the notions of 'conception' and 'concept' below). 5 Moreover, conceptions of music can be indexed to particular socio-cultural contexts. This is well accepted in the ethnomusicological literature. 6 As Curt Sachs writes, 'our generic term music is misleading. For it implies an all-embracing concept of the various elements that reach our ears and therewith a common rule for each of them.' 7 Our pluralism claims that 'music' is polysemous, and that what counts as a legitimate music concept depends upon context. 8 On our view, then, music is disunified diachronically, as McKeown-Green allows, but also synchronically, as Sachs indicates. Our pluralism further grounds a methodology that uses traditional philosophical tools (conceptual analysis, for instance), but is not vulnerable to concerns such as McKeown-Green's.
In § II, we provide a general and schematic discussion of conceptual pluralism illustrated with a discussion of pluralism about gene concepts. We then move onto our argument. In § III, we discuss problems with the traditional approach to music-concept analysis. We begin by critiquing three conceptual analyses of music as unitary. We then argue that any particular reductive analysis of music will either admit dubious cases, be hopelessly ad hoc, or unwieldy. These considerations motivate shifting to an alternative approach that attempts to capture music across a set of dimensions. We argue that although this approach has its attractions, determining which dimensions are admissible seems to depend on the investigative programme or explanatory interests of the theorist.
This leads us to pluralism and § IV. Here, we articulate our position and distinguish between some possible legitimate, non-equivalent music concepts. But we don't take our article to have the last word on the actual/legitimate music concepts out there: we argue that pluralism provides a fruitful and useful avenue for philosophical research. Pluralism, then, is justified by the futility of unitary approaches, the plausibility of pluralism, and its fruitfulness. In § V we tentatively consider pluralism's prospects in aesthetics more widely. Before beginning, we should briefly explain why we focus on concepts, and clarify our claim. Philosophical analysis can have different targets: linguistic, conceptual, or ontological, for instance. In this article we focus on concepts, largely putting linguistic and ontological questions aside. When we say 'We are music pluralists' , then, we mean pluralists about music concepts (we consider below whether this has ontological upshots). By making this choice, we do not imply that fruitful philosophical work cannot be carried out in these other domains. We do think, however, that establishing conceptual pluralism matters.
First, linguistic pluralism is surely obvious -different languages have different musical terms. But this apparently shallow claim belies interesting questions that relate the linguistic to the conceptual. Do, for instance, our linguistic capacities influence our concepts, or vice versa? An understanding of the nature of the concept of music (or, as we argue, concepts of music) would help answer such questions. Second, understanding the kind of category that music is -that is, the aim of asking 'What is the ontological status of music?' -is influenced by understanding music conceptually. Pluralism about music concepts could feed into both of these approaches. Again, this doesn't deny that entering into the discussion from these other points would not also be useful and interesting, but we think that starting with concepts clearly is. 9 There is a rich, widespread discussion in philosophy about the purpose and nature of conceptual analysis, but here is not the place to defend the practice (or, for that matter, decide whether we are in the business of conceptual analysis 'properly understood'). 10 Let's clarify our position. We are not merely arguing that music may be investigated from many viewpoints. Surely anything worth understanding is worth understanding from different perspectives. Nor are we merely arguing that there are different conceptions of 'the' music concept, since people may indeed have different conceptions of a unified concept. Often, particularly in moral and political philosophy, a distinction is drawn between 'concepts ' and 'conceptions' . 11 Theorists may agree on some concept -moral value, say -but disagree on what it takes to be morally valuable -you might think moral value lies in human flourishing, while we might think it lies in maximizing pleasure and minimizing pain. In such contexts, theorists are often not talking past each another; they have different conceptions of the same concept. In principle at least, disagreements that arise pertaining to the concept at hand should be resolved, perhaps by giving up one of the conceptions. The analogue in discussions of music would be to argue that music has a single concept but there are contested conceptions. This strikes us as the status quo: theorists proffer different definitions of music yet take themselves to be explicating the same concept as each other.
We argue instead that there are multiple music concepts, and thus some debates about music are in fact due to conceptual confusion. In some contexts, debates about the nature of music are merely semantic in the pejorative sense: both sides target different concepts, and thus talk past one another. 12 It is also worth distinguishing conceptual pluralism from eliminativism.
Eliminativism, in this context, has two senses: ontological -music doesn't existand discursive -we should eliminate 'music' from our vocabulary. As we'll expand on below, we are not here concerned about the appropriate metaphysical stance we should take to the music category in light of concept-pluralism. Regarding eliminativism about music discourse, although this is not the place for a long discussion, we should mention that we prefer pluralism for two related reasons.
Firstly, although music concepts are non-equivalent, they are nonetheless clearly related. We suspect that there are more or less deep, and illuminating, connections between music concepts, and this is reason enough to maintain aspects of our typical discourse. Secondly, non-eliminativism is more amenable to adopting the integrative, multi-disciplinary work that music research requires. 13 So we claim that there is no one privileged music concept, where concepts are (amongst other things) necessary for resolving ontological questions. That is, to determine whether or not some music-candidate is music, some concept is required to delineate music from non-music. It is with regard to this notion that we are pluralists: there is no unique answer to such questions, just as there is no unique concept. 14 II. CONCEPTUAL PLURALISM: A SKETCH A pluralist believes that some apparently unified category is in fact captured by at least two, legitimate, non-equivalent concepts. It is useful to think of concepts 12 For scientific examples, see Adrian Currie and Anton Killin, 'Musical Pluralism and the Science of Music' , European Journal for Philosophy of Science 6 (2016): 9-30. 13 Ibid.
14 Perhaps one could generate uniquely correct answers by supervaluationist means: that is, only accept as music that which counts as music on each legitimate concept. But why one should prefer 'supertruth' over our usual concept of truth is not obvious to us. as functions from events or states of affairs, or from objects, to categories. A music concept, then, distinguishes music from non-music. Multiple non-equivalent music concepts will result in multiple non-equivalent ways to make that distinction. In this section we provide an example of conceptual pluralism, and draw on that to provide a recipe for establishing such positions. Conceptual pluralism has been most discussed in scientific contexts: it has been suggested about species, 15 homology, 16 and biodiversity, 17 sparking both renewed interest in analyses of these categories and a fruitful methodology for individuating and integrating equivocal scientific concepts. We will use a simplified account of gene pluralism in order to illustrate the idea. Needless to say, this illustration falls short of a full-fledged defence of pluralism about gene concepts, and is not intended as such.
Here are two ways of defining 'gene' . 18 First, a functional definition: genes are the unit of inheritance responsible for the transmission of phenotypes across generations. Second, a physical definition: genes are a sequence of nucleotides in either DNA or RNA. By the first definition, to be a gene is to play a certain kind of role; by the second, we pick out a class of objects by their physical properties.
There may be other useful definitions of 'gene' too; we will restrict ourselves to these two for present purposes. 19 These definitions are non-equivalent, that is, they identify different sets of objects as genes. The physical definition will include sequences of DNA that have nothing to do with the inheritance of phenotype -DNA sequences that, for instance, play other roles in the developmental system or (potentially) no role at all. The functional definition admits objects that are not sequences of DNA and RNA. Perhaps epigenetic factors, sets of proteins, or other streams of inheritance play the specified role. This non-equivalence could be illusory, of course: it could be that there is some other concept that does all the work we want, or near enough.
In any case, these two definitions of 'gene' are more or less appropriate given the interests of different scientific fields. Classical genetics, for instance, investigates patterns of heredity using Mendelian theory and breeding experiments using model organisms (traditionally, fruit flies). It is plausible that a functional definition of 'gene' suits their purposes. After all, Mendelian theories, of allelic dominance and so on, say nothing overtly about the physical realizers of heritable units, and non-molecular studies of fruit flies do not grant direct epistemic access to physical realizers. Molecular genetics is a different story.
The use of microscopy, direct interventions on DNA, and mechanistic developmental theories, make a physical definition of 'gene' appropriate. Where for the Mendelian geneticist genes are unobservable functional postulates for explaining patterns of heredity, in molecular genetics genes are the target of direct investigation and are conceptualized in physical terms. If something like that is right, 20 then insisting that one or the other definition takes precedence, and, worse still, demanding that one unitary definition be adopted, is likely unjustified. 21 Conversely, it may turn out that either classical or molecular genetics is not truly legitimate. Perhaps classical genetics has been superseded by molecular methods, or perhaps molecular genetics is somehow wrong-headed.
If so, one or both gene concepts might not be legitimate.
Conceptual pluralism, then, is grounded in two claims. First, non-equivalence: there is more than one concept of some category on the table, these concepts divide up the world in different ways, and they cannot be successfully unified. Second, legitimacy: these different concepts play ineliminable roles in distinct, and legitimate, investigations. If the case for gene pluralism satisfies both nonequivalence and legitimacy conditions, then gene pluralism is justified. Moreover, pluralism occasions a research methodology for approaching scientific concepts.
The different concepts are divided among the relevant researchers (for example, classical versus molecular), and findings are integrated into more inclusive models or systems. 20 Which it may not be. C. Kenneth Waters, 'What Was Classical Genetics?' , Studies in History and Philosophy of Science Part A 35 (2004): 783-809, argues that the institutional split between 'classical' and 'molecular' genetics is artificial. 21 As hinted at earlier, this scenario contrasts with cases in which two conceptions of some unified concept give rise to substantial dispute. The analogue for genes: perhaps there are contrasting conceptions of what it takes for a (functional) gene to be a unit of phenotypic inheritance, or for a sequence of DNA or RNA nucleotides to be a (physical) gene.
As we suggested earlier, one might be tempted to think that pluralism about concepts pertaining to a category entails claims about ontology, specifically eliminativism about that category, but this is not so. A possible conclusion from our simplified gene pluralism is that there are no genes, but rather functional genes and physical genes. But eliminativism is not entailed by pluralism. Some philosophers, for instance, have argued that although we ought to be pluralists about concepts such as species and homology, such general categories survive as required for historical exposition, and as explananda: the targets of explanation. 22 Conceptual pluralism and eliminativism have a tricky relationship, and here in this article we will be mostly silent about whether pluralism about music concepts leads to eliminativism about the music category (although it is briefly considered in § IV.2). 23 For now, let us simply note that if the lesson of concept pluralism is ontological eliminativism, it is at least an atypical route to eliminativism. For instance, moral error theorists are not eliminativists about the category of moral facts because they think that there are multiple, non-equivalent moral concepts. Rather, they are eliminativists because they think there is nothing like moral facts. 24 In what follows, we advance conceptual pluralism about music. To support non-equivalence we will consider attempts to provide unitary analyses, and we'll then shift to a discussion of legitimacy.

III. UNITARY ANALYSES OF MUSIC
Unitary conceptual analyses of music have been proffered by various theorists.
In our view, reason to think that these accounts fail lends support to our claim that the non-equivalence condition is satisfied.
We do not pretend that we have the final word on unitary conceptual analyses of music: needless to say, we can only consider so many of them, and we can only follow so many avenues. However, we provide strong inductive grounds for thinking that any such analysis will ultimately fail. Moreover, as we discuss in § IV, we also provide a positive reason to endorse pluralism: it provides the framework for a fruitful research programme.

III.1. REDUCTIVE ANALYSES
There are two main contexts in which theorists offer strict biconditional accounts of music. First, reductive conceptual analyses that attempt to define music in nonmusical terms. Second, operational definitions which appear in some scientific contexts. Operational definitions are not intended to be unitary analyses, so we will not be concerned with them here. 25 A third possibility, recently defended about art, is that of a cluster concept: a decidedly messy set of desiderata, subsets of which, when satisfied, are sufficient for the concept's application. 26 Stephen Davies has shown that cluster concept accounts reduce to standard, albeit disjunctive, conceptual analyses. 27 For our purposes, it suffices to say that, qua reductive analyses, such attempts to find cluster conditions for music will fail, due either to counterexample, being ad hoc, or to bloated complexity. This motivates the idea that there are non-equivalent concepts of music.
We consider and critique three leading reductive analyses of music. Our critique takes the form of counterexample; that is, we shall consider whether what counts as music (simpliciter) by these conceptual analyses coheres with other legitimate concepts of music. Some are 'folk' views, others ethnomusicological, and so forth. Although this involves some appeal to intuitions, recall that our ultimate aim is to show that a single musical definition cannot meet the requirements of a range of different interests. On the assumption that appeals to intuition track one or more of these different interests, they serve our aim. Irving Godt has defended what he calls a'practical' definition of music: '(2) Music is humanly organised sound, (3) organised with intent (4) into a recognisable aesthetic entity (5) as a musical communication (6) directed from a maker (7) to a known or unforeseen listener, (8) publicly through the medium of a performer, (9) or privately by a performer as listener. ' 28 Godt's first condition, 'Unwanted sound is noise' , will be set aside. His tenth, and final -that all human societies make 'music' -is discussed below. According to Godt, then, music must be an organized act of communication, self-reflexive or otherwise, through sound. Godt's definition is targeted towards evaluative, rather than descriptive, purposes: it aims to set a bench-mark for something's counting-as-'music' . He is ostensibly interested in a definition which encourages a critique of music. Such a definition might prove to be legitimate for Godt's normative purpose -although we doubt this, skewed as it is towards Western art music. Importantly for us, it fails if construed as a unitary analysis of music simpliciter, for the following reasons.
First, Godt's definition is too narrow. Godt does not consider aleatoric music, for instance, to be music, and is rather damning of John Cage's chance-based music (for example, Variations IV and Imaginary Landscape No. 4; he does not consider performances of these to be music). 29 Dismissing (disliked) music from being music is incoherent at worst, ad hoc at best. 30 Godt's definition admittedly 'embodies a deliberate bias toward the music of Western European culture'; yet Godt claims that 'There are no human societies without music' -an error, certainly, if Godt's own definition is supposed to apply to the term 'music' . 31 Much of the world's music is not, for instance, 'intentionally organised' into a clear, recognizable 'aesthetic entity'; that is, it is not intended to be an object of aesthetic value, or an object to be perceived or evaluated aesthetically. Lullabies sung by parents or caregivers with the intent to soothe infants' arousal and produce changes in affect, investigated by developmental psychologists and anthropologists, might be one example of this found in many cultures. 32 Ethnographic investigations of music often find that music is entangled with both dance and ritual, or indeed other aspects of social life (such as hunting), and that it need not be primarily communicative (in his narrow sense, by 'musical communication' Godt means 'a controlled sequence of sounds with a definable and ultimately perceivable grammar'), 33 but rather be affective. Several cultures lack a synonym for the English word 'music' , but possess 'music-dance'; for these peoples, movement is indispensable for the concept's application. 34 For some, such as the Blackfoot and Sioux Native American tribes, music is often a communal, whole-group affair, and it contributes emotionally to various Not Music, but Musics: A Case for Conceptual Pluralism in Aesthetics 29 Ibid., 86. 30 An anonymous reviewer points out that music and musical works are conceptually distinct, and that many events will count as music just because they are performances of musical works. contexts (for example, ritual). 35 Although the 'makers' of these instances of music may also be the only 'listeners' , it is not a'private' experience' , undermining Godt's distinction. An imitative 'bleating calf' song was long used by Blackfoot hunters to round up a bison herd at the top of a cliff face so the beasts could then be spooked over the cliff. 36 If performances of this song were not intended to give rise to an aesthetic entity qua musical communication, it would be hard to see how they could be accounted for in Godt's definition.
We turn now to Levinson's analysis of music: 'Sounds temporally organised by a person for the purpose of enriching or intensifying experience through active engagement (e.g. listening, dancing, performing) with the sounds regarded primarily, or in significant measure, as sounds. ' 37 Like Godt's definition, Levinson's definition is both too broad and too narrow, construed as a unitary analysis.
Spoken nonsense poetry, for example, seems to satisfy this definition and thus count as music. Moreover, it excludes potential music-candidates. Background music at a function might not be attended to, nor might it enrich or intensify experience, but it is nonetheless music (for instance, see Davies on Tafelmusik). 38 Warm-up scales and muzak may also constitute music, albeit poor or base music.
Yet these are not attended to for experiential enrichment either. Moreover, as Kania has noted, one might 'sneak into a friend's bedroom and play the opening of the first-violin part of Strauss's Don Juan to startle him awake, with no intention that either of us attend to or engage with these sounds at all, let alone for the purpose of enriching or intensifying our experiences ' . 39 Although the occurrence of it may be unlikely, a sleepwalker picking up a guitar and filling her bedroom with the sounds of an unintentional rendering of, say, Anthony Ritchie's Melancholia surely counts as producing music, yet does not have the requisite intent. Comparative psychology raises more problematic examples. As we've discussed elsewhere, an important inroad into the nature of music -particularly in an evolutionary context -involves examining musicality in animals. 40 Such work involves identifying commonalities and contrasts between humans and non-human animals in terms of the presence and purpose of organized sound. In such contexts, requiring that true music involves certain kinds of intention -and thus asking whether various animal cases are truly 'music' or instead some kind of 'music-like' behaviour -strikes us as a red herring (although of course examining the function of musical behaviour in music is productive).
Levinson's account fails in part due to not meeting common intuitions about music. Where Godt's definition could not cope with some ethnographic and scientific contexts, Levinson's does not 'capture the phenomena' , that is, our common ideas about what counts as music. On our view, this is a different project than, say, anthropology, but may nonetheless be a legitimate philosophical investigation. On the pluralist view, a concept's suitability depends crucially on context: specifically, it depends on which project we are engaged in.
The bottom line is not that these counterexamples are knock-out objections.
Sometimes, classifying spoken nonsense poetry as music might be a legitimate move. Rather, these examples emphasize the futility of the monist approach to music. Theorists might attempt to massage our intuitions so that we agree that music as used in ritual isn't really music simpliciter, or that spoken nonsense poetry just is, but if this is where the focus is then this begins to take the appearance of a degenerate research programme.
Finally, we consider Kania, who argues that 'music (1) is any event intentionally produced or organized (2) to be heard, and (3) either (a) to have some basic musical feature, such as pitch or rhythm, or (b) to be listened to for such features' . 41 Kania's analysis is promising. The third, disjunctive condition allows for the inclusion of much avant-garde experimental music -music that lacks standard features, like pitch or discernible rhythm. It also distinguishes between music and non-musical sound art (Foley tracks, perhaps), since these are not intended to be listened to for pitch, rhythm, and so on. However, his analysis is too broad. Poetry audiences do listen for metre, rhythm, pitch contour, and so on. As spoken poetry is intentionally produced/organized, to be heard, and has basic musical features such as rhythm, it counts as music according to this definition. Moreover, Davies argues that Morse code also counts as music according to Kania's account, given that it is listened to (and systematically organized) with regard to rhythm. 42 Acts of rhetoric, such as Churchill's speeches, are surely events intentionally organized to be heard, which are certainly sometimes listened to (and performed) for their musical features (cadence, repetition, assonance, and so on), and it seems a mistake to categorize them as music simpliciter.
Kania's definition is also too narrow. His first condition discounts the unintentional performance of Melancholia, and our intuitions at least are offended by this.
The second condition discounts examples that are not intended to be heard, and we think this is problematic. A deaf pianist (say) rehearsing the Allegro moderato from Bartók's Piano Sonata, alone, and unheard, is producing music, despite the fact that the sounds are not, on this occasion, produced to be heard by anyone. Moreover, by focusing on intention to produce something that is to be heard, Kania's definition, on our view, misses something important about a range of examples, such as ritual music that is not produced to be heard, but to have a certain effect on the participants. For instance, military marches are primarily for coordinating action and stirring comradery. Moreover, consider trance-inducing ritual music in voodoo or similar practices -the purported spiritual influence and significance of the music may be an indispensable dimension. 43 The music might need to be heard to have the desired effect, yet it is not primarily produced or organized to be heard. Further, the appeal to intention is again potentially problematic in the context of comparative psychology. Finally, it strikes us that imagining and appreciating music written on a score is a legitimate aesthetic experience of music, even if it is not, nor is intended to be, heard.
What can we say about these definitions? It looks like there are several, perhaps irreconcilable, tensions in trying to arrive at a unitary account of music. Here are some: (1) Intention. It is undoubtedly true that intention plays an important role in music. It is tempting to think that music is non-linguistic intentionally organized sound. But it is also true that not all intentionally organized sound is music and not all music is intentionally organized.
(2) Formal features. The vast majority of music can be understood in terms of metre, pitch, rhythm, and so forth. But, particularly in a Western art-music context, contemporary composers will actively resist this. Many contemporary examples of music are, if anything, music by virtue of subverting those features.
(3) Social role. Music plays a variety of, sometimes conflicting, roles in human society: from specialists playing for elites in various contexts, to social cohesion, to encouraging discord, to simple individual pleasure. 43 Music has widely been used by peoples the world over, sometimes in conjunction with drugs, 'in order to communicate with the spirits, and then return to their senses in order to heal the sick, control animals or change the weather' , and so on; David Hendy, Noise: A Human History of Sound and Listening (London: Profile, 2013), 37. It strikes us as plausible that these different features matter more or less for different explanatory projects. If we give up on monism and allow music to admit of a plurality of concepts, such tensions would likely dissolve. Three possible conclusions strike us. First, if tensions such as these prove irreconcilable, then non-equivalence is true. That is, there is more than one music concept. Second, even if some unitary account is available, it will be so bloated with epicycles, caveats, and so on, that the capacity for it to do any work for us will be undermined. Third, that unitary account will surely involve some ad hoc decisions about what will and will not count as 'music' . Moreover, as McKeown-Green emphasizes, unitary accounts are, at best, definitions of what we consider to be music now: a definition of music is temporally constrained. As musicians continue to flout conventions, and as music continues to take surprising turns in its Others might have differing intuitions about the cases above: you might insist that music, to be music, must be fundamentally auditory, or, more persuasively, you might adopt a broader conception of aesthetic value which could undermine our claim that lullabies are not primarily aesthetic. 45 And indeed, we might disagree. But ultimately this is not the point: although wrangling about intuitions can be an important and useful exercise, we hope this discussion shows that in the context of monism -of insisting that there is a privileged music concept -it is at best unhelpful.

III.2. THE MULTIDIMENSIONAL APPROACH
A way of providing a unitary account, which nonetheless retains the flexibility required to accommodate music's multiple guises, is to conceptualize it along several dimensions. We think that it is worth pausing to give an example of this approach and articulate why we prefer pluralism.
Hamilton concedes that music is an elusive phenomenon but nonetheless believes that he can identify 'salient features' that 'can be elucidated by looking at three different directions of characterization: acoustic, aesthetic and acousmatic' . 46 These directions are not intended to be jointly sufficient for demarcating music, however. Much non-music will also be captured by the approach. We think it useful to represent these directions as a threedimensional conceptual space (see fig. 1).
The 'acoustic' dimension of music concerns features such as tone/pitch and rhythm. For Hamilton, instances of pure concert music that comprise a high preponderance of tones and discernible rhythms (jazz, Beethoven sonatas, and so forth) are placed towards one end of the 'acoustic' continuum, while the noisebased instrumental music of, say, Helmut Lachenmann will be towards the other end (that is, the less paradigmatically 'musical' end), and other music, such as the unpitched percussion music found in several non-Western cultures, will be somewhere in-between. 47 So the acoustic characterization picks out what counts as music according to certain criteria -rhythm, pitch, and so on -and orders instances of music according to the prevalence of such features. Just because something is lacking in pitch and discernible rhythm does not mean that it cannot count as music: however, it won't be a paradigmatic case of music.
The 'aesthetic' dimension picks out aesthetic properties that instances of music embody to greater or lesser extents -properties accessed through aesthetic attention (perhaps gracefulness, garishness, evocativeness, sluggishness, vividness, and so on) -and orders instances of music accordingly. Finally, the 'acousmatic' dimension picks out features of 'the listener's experience or response to sounds as abstracted from their worldly cause' , that is, 'detached from the circumstances of their production' . 48 In other words, features of sounds simply 47 Hamilton explicitly distinguishes music from sound art, citing the creative output of Bill Fontana and Alan Lamb as examples of sound art (Fontana's work includes recording and mixing the traffic sounds at Brooklyn Bridge; Lamb's work includes recording Australian telegraph wire humming). 48 Hamilton, Aesthetics and Music, 58. as sounds. These might include sound quality, timbre, and properties of the performance / acoustic space, considered intrinsically and decoupled from performance context. With reference to fig. 1, instances of music can be ordered along the axes of the three continua and plotted according to their acoustic, aesthetic, and acousmatic properties. The most paradigmatic instances of music will all be grouped together in one corner of the conceptual space: music rich in musical acoustic, aesthetic, and acousmatic features.
Hamilton's approach has been criticized before. For instance, devoting one dimension to aesthetics clearly skews the space towards 'art' music, yet not all music is art. Davies argues that instances of 'Happy Birthday' , for example, count as music, but are rarely, if ever, aesthetically appreciated. 49 Moreover, Hamilton's acousmatic angle might be too narrow to include properties of the musical experience of much music historically, prehistorically, and in some non-Western settings (recall, for example, the Blackfoot 'bleating calf' song; it remains to be seen whether taking it out of its traditional context and appreciating it as a song, for its aesthetic and acousmatic features, should make it more paradigmatically music).
Although we think that multidimensional approaches to music are undoubtedly useful, interesting, and an improvement on the traditional analyses, we reject the idea that there is such a thing as the 'best' unitary account of this nature. The basic idea is that (1) the conceptual space must include so many dimensions as to become unwieldy, or (2) the admissibility of dimensions will depend upon theorists' explanatory interests, in other words, pluralism.
The anthropologist, for instance, might demand that one ineliminable musical dimension is its integration with, and role in, human sociality, while the structure-focused musicologist will find such a dimension an unnecessary nuisance. Moreover, unless there are robust, systematic connections between dimensions like Hamilton's, it strikes us that a researcher who simply doesn't care about, say, the aesthetic aspect of music, has no reason to include that dimension. The jingle composer/producer might care about the psychological effect that her music could have, but deny outright that her music is art.
Consider also a caregiver singing a lullaby, or a certain rhythm played for a ritual purpose. It strikes us as mistaken to think that these are any less 'music' simply because they are non-aesthetic in purpose. A 'unified' conception of music along these lines would be, at best, gerrymandered and cobbled-together or, more likely, incoherent. We suspect a different approach is more fruitful. In the next section, we further advance our pluralism.

IV. A PLURALITY OF MUSIC CONCEPTS
As we have seen, music is elusive and difficult to define. This is explained by the realization that there is no one concept of music: rather, there are multiple music concepts, useful for different enterprises. An aesthetician might be interested in music as art; an anthropologist might be interested in music as an ethnographic phenomenon -and why should one researcher receive short shrift in favour of the other?
Some theorists analyse only a Western characterization of music. Yet other cultures have other concepts of music, which might exclude (say) an 'aesthetic' dimension, or might require dance or ritual, for instance. Some cultures do not have a word for music, yet they clearly create that which we would consider music.
Other cultures have different words for distinguishing different music concepts.
For example, Hiromi Lorraine Sakata, researching Afghanistan's music(s), distinguishes ghinā from musiqi; moreover, she says: 'It is clear that the concept of music that exists in Afghanistan does not parallel the Western notion of music. ' 50 Our pluralism encourages the analysis of cross-cultural divergence of music concepts without expecting them to adhere to a unitary one-size-fits-all concept.
Indeed, it is useless, given our pluralism, to ask whether something is music (simpliciter). For instance, there is debate in aesthetics about avant-garde music performances, which centres on the question 'Does this really count as music?' 51 If pluralism is correct, this is the wrong question to ask -rather, one should ask: 'Are they captured by some salient music concept?' or 'Which legitimate music concepts capture them?' Performances of Cage's 4′33″, for instance, might count as music for the purposes of one theorist's investigation (perhaps into the role of concert piano performance practice, or the history of twentieth-century American concert performances) and not of another's. 52 This is not to say that there are no substantial disputes about music concepts. If theorists employing the same music concept disagree about the status of (say) 4′33″, it is not merely verbal:  Levinson,Music,Art,and Metaphysics. 52 This situation results in a Chalmers-style deflationism about 'What is music?' questions: see David Chalmers, 'Verbal Disputes' , Philosophical Review 120 (2011): 538-43. different properties and will be shaped by the different roles they play in investigation.
Moreover, the concept of music as analysed by Godt, Levinson, and Kania above focuses on external sound-events heard. Yet is the score-reading fanatic, revelling in Mahler's Second Symphony (merely imagining a musical performance by reading a copy of the score) not appreciating music? Can she not aesthetically evaluate it too? Perhaps an aesthetician's concept will focus not on external sounds heard, but on the aesthetic properties of music accessible for evaluation from multiple modes of experience. 53 An anthropologist's or ethnomusicologist's concept might eschew aesthetics and the notion of 'art' in regard to their explananda. And an evolutionary theorist's concept might focus on human behaviours and capabilities, the traits that underlie these, and the interaction of genetics and environment in their regard. Different music concepts are useful for different investigations, and these are often complementary, because they ground legitimate concerns. Unlike disagreements about conceptions, then, different investigators with different concepts need not be at loggerheads. 54 In sum, we think that there are good grounds for conceptual pluralism about music. We have argued that (1) there is more than one, non-equivalent, music concept available and (2) these concepts ground equally legitimate investigations. It is worth noting that we have not provided an explicit, detailed account of a particular music concept, but have instead gestured to a large number of possibilities (ethnomusicology, developmental science, ordinary critique, philosophy, and so on). This is because our purpose in this article is to establish pluralism about music, not to argue for the legitimacy of some investigation of music. In § IV.2 we'll discuss how that philosophical project might go, but first we'll discuss an objection.

IV.1. BUT WE KNOW IT WHEN WE HEAR IT!
There is a straightforward objection to our view: there must be a unified music concept, we just don't know what it is yet. After all, we know it when we hear it! Davies writes, 'we can immediately and almost infallibly recognize [music], even when it comes from a culture that is foreign to us. Though I may be unable to predict how such music will continue or to recognize errors, and though it may sound strange to me, I can be in no doubt that it is music that I am hearing.' 55 Kania writes, 'most of us can tell whether, and which of, the sounds we are currently hearing are music. This is so whether or not what we are listening to is a familiar piece, a piece we have not heard before, or even music from a culture or tradition with which we are unfamiliar' . 56 Does our capacity to know music when we hear it undermine pluralism? Not at all. First, being a pluralist about music concepts does not disallow being a realist about music phenomena, or there being a music category. As discussed in § II, there is a debate to be had over whether conceptual pluralism entails ontological eliminativism. Second, given the amount of overlap in legitimate music concepts, it is unsurprising that, for easy cases, we both agree and immediately recognize something as 'music' . Third, in hard cases we do disagree: philosophers disagree over whether chance-based compositions, or other instances of the avant-garde, count as music; indeed, much contemporary Western art music is not immediately recognized by the untrained ear as music (and even theorists with a trained ear disagree -recall Godt on Cage's Variation IV and Imaginary Landscape No. 4).
Music is undoubtedly a 'phenomenon' -it is something which we recognize and want to explain, but it doesn't follow from this that music is unified. To take a well-worn example, we recognize jade when we see it, yet it is not unified for the mineralogist: jade reduces to jadeite and nephrite. And we distinguish between different species when we see them. Yet why think that scientists must constrain themselves to a unitary species concept? 57 Understanding music in its many guises, contexts, and functions requires a diverse array of contributions, from musicologists and philosophers, to anthropologists and ethnomusicologists, sociologists, archaeologists, evolutionary biologists, computer scientists, acousticians, neuroscientists, psychologists, and music therapists. That the study of music represents such an interdisciplinary research cluster itself might suggest pluralism; that given the range of explanatory interests, research questions, and lines of evidence relevant to the various fields, no unitary definition of music will do the job across the board. Recall that an art historian's or aesthetician's conception might take an aesthetic dimension to be ineliminable; an ethnomusicologist might reject this, and take social function as ineliminable. Both operate with different music concepts necessary for their research agenda.
We take our discussion to have provided an inductive argument for thinking that there is not a unitary music concept. In what follows, we argue that there is also positive reason to be a pluralist about music: pluralism grounds a research programme in aesthetics. To the extent that this is plausible, it undermines the 'default status' of the view that music is a unified concept. 56 Kania, 'Definition' , 3. 57 Kitcher, 'Species'; Ereshefsky, 'Species' . Music may be what we think it is; it may not be […]. Music may be that to which some dance or pray or make love; but it's not necessarily the case. In some cultures there are complex categories for thinking about music; in others there seems to be no need whatsoever to contemplate music. What music is remains open to question. 58 Pluralism enables a methodology that is well equipped for tackling Bohlman's challenge. It grounds a research programme in aesthetics: the philosophy of musics. The project, as we see it, consists of (1) identifying what makes for a legitimate and an illegitimate music concept, (2) identifying different investigations of music, (3) identifying different music concepts and analysing them. This can help clarify crosstalk both in and outside philosophy.
The philosophy of musics also delivers a normative project: can we engineer suitable music concepts for different investigations? Although we are pluralists, we do not endorse a conceptual 'free-for-all' relativism. 59 It may be that there is a best, or true, analysis regarding some research interest -as is plausibly the case for species, genes, and so on. If it turns out that our brains process the basic musical features of sounds within an encapsulated module, then as McKeown-Green suggests, sounds with these features, which psychologists and neuroscientists could conceive of as music, might be worth investigating. 60 Anthropologists, on the other hand, might favour a concept that emphasizes music's societal/functional roles. In short, operational definitions, not usually discussed by philosophers, provide the way forward. (And, if we are wrong about music's plurality -if a unitary analysis of music is possible -our approach will help us achieve it.) For example, music cognition researchers Honing and Ploeger distinguish 'music' from 'musicality' as follows: 'We define musicality as a natural, spontaneously developing trait based on and constrained by our cognitive system, and music as a social and cultural construct based on that very musicality' . 61 And so, they take 'musicality' to be a human cognitive trait, and 'music' to be the varying anthropological and social expressions of that trait. In making this division, they hope to isolate the cognitive aspects of music in order Not Music, but Musics: A Case for Conceptual Pluralism in Aesthetics 58 to explain it evolutionarily. Assessing this move shouldn't proceed by way of counterexample, but rather by evaluating its suitability as a guide for focusing research on the phenomena investigated. Very briefly, Honing and Ploeger's conceptual division might be undermined if it can be demonstrated that splitting music from musicality along biological and cultural lines is unjustified due to the dynamic nature of bio-cultural co-evolution. That is, it may be that one cannot understand the evolution of music as a cognitive trait without taking its social expression into account. If so, the 'music/musicality' distinction may not do the work that Honing and Ploeger need it to do. If not, Honing and Ploeger might be justified in black-boxing the complexities of the influence of social and cultural forces in evolutionary psychological contexts. 62 To reiterate, the success of the concept turns on the work the concept is supposed to do.

V. CODA: PLURALISM IN AESTHETICS?
We are conceptual pluralists about music. A conceptual pluralist about some category must argue that (1) there is more than one non-equivalent concept applicable to that category, and (2) that those non-equivalent concepts are utilized by legitimate research streams. In diagnosing the failures of unified accounts of music, we motivated the former requirement. In our discussion in § IV, we have given examples of possible legitimate research interests (motivating the latter) and suggested that pluralism brings with it the advantage of fruitfulness.
According to our view, there is no interest-independent fact of the matter about what music is. However, our pluralism is not a 'free-for-all' relativism: the challenge for the aesthetician is to distinguish music concepts and understand which conceptions are admissible and which are not. For example, a pluralist about species might admit more than one concept of horse, yet deny that 'things which win the Derby' is an admissible horse-concept, even for the purposes of a race-goer. 63 We did not settle herein whether Godt's definition of music (or, for that matter, Levinson's or Kania's) was suitable given his specific purposes (although we noted that we are sceptical). There is debate to be had about the most promising, say, 'aesthetic' music concept.
Thus conceptual pluralism, we think, supports a fruitful research programme, and, moreover, one that could see philosophical methods put to good use outside philosophy. Defining music, however, is hardly the aesthetician's bread and butter. Does our discussion of such a fringe matter have upshots for aesthetics more broadly? We believe it does. We have shown how abandoning the search for 62 For further discussion, see Currie and Killin,'Musical Pluralism' . 63 Thanks to Greg Currie for this example.

Adrian Currie and Anton Killin
Estetika: The Central European Journal of Aesthetics, LIV/X, 2017, No. 2, 00-00 a unified music concept can be both motivated and fruitful. Similar claims may well be made about more central aesthetic concerns. Although we have not made the case herein, 'beauty' and 'art' , for instance, seem to us to be good candidates for pluralist treatments. Just as there is not music but musics, there may be multiple legitimate concepts of art, 64 and there may be multiple legitimate concepts of beauty or sublime.

Adrian Currie
Centre for Research in the Arts, Social Sciences and Humanities, University of Cambridge, Cambridge CB2 3QZ, United Kingdom ac2075@cam.ac.uk