Musical Ontology

Musical ontology is the study of the kinds of musical things there are and the relations that hold between them. The most discussed issues within this field have been the metaphysical nature of works of classical music, and what it is to give an ‘authentic performance’ of such works. Recently there has been growing interest in the ontologies of other musical traditions, such as rock and jazz, and discussion of the methodology and value of musical ontology. (For a more detailed overview of these debates, see Kania 2008a and forthcoming b.)

2.1 The Fundamentalist Debate
Musical works in the Western classical tradition are multiple entities, in the sense that they admit of multiple instances (performances). Much of the debate over the nature of such works thus reads like a recapitulation of the debate over the ‘problem of universals’; the range of proposed candidates covers the spectrum of fundamental ontological theories. We might divide musical ontologists into the realists, who posit the existence of musical works, and the anti-realists, who deny their existence. Realism has been more popular than anti-realism, but there have been many conflicting realist views. I begin with three unorthodox views before moving on to more orthodox Platonist and nominalist theories.

Idealists hold that musical works are mental entities. Collingwood (1938) and Sartre (1940) respectively take musical (and other) works to be imaginary objects and experiences. The most serious objections to this kind of view are that (i) it fails to make works intersubjectively accessible, since the number of works going under the name The Rite of Spring will be as multifarious as the imaginative experiences people have at performances with that name, and (ii) it makes the medium of the work irrelevant to an understanding of it. One might have the same imaginative experience in response to both a live performance and a recording of The Rite of Spring, yet it seems an open question whether the two media are aesthetically equivalent.

David Davies argues that musical works, like all works of art, are actions, in particular the compositional actions of their composers (2004). Thus he revives what we might call an ‘action theory’ of the ontology of art. (An earlier defender of such a view is Gregory Currie (1989), who argues that artworks are types of action, rather than the particular actions with which Davies identifies them.) Although deciding between theories of musical ontology is always to some extent a matter of finding a balance between the benefits of a theory and its cost in terms of our pre-theoretic intuitions, action theories have a particularly hard row to hoe since they imply that an instance of a work is some action performed by a composer, rather than a performance. In order to make up for such damage to our intuitions the theoretical benefits of an action theory will have to be quite extensive.

Guy Rohrbaugh has proposed a new ontological category for musical, and other multiple works of art (2003). He argues that since the kinds of things we attribute to musical and other art works, such as modal and temporal flexibility, cannot be accounted for by any extant ontological theory and, moreover, that these problems arise not just for artworks, but such things as ‘species..., clubs, sorts of artifact, and words’ (199), we are justified in positing a new kind of entity: a historical individual that is ‘embodied in’, but not constituted by, physical things, such as scores and performances. (For criticism of this view, see Dodd 2007, 143–66.)

Most theorists seem to think that some kind of Platonist or nominalist theory of musical works is more plausible than those so far considered. Nominalists hold that musical works are collections of concrete particulars, such as scores and performances (Goodman 1968, Predelli 1995, 1999a, 1999b, 2001, Caplan and Matheson 2006). While this view is attractive because it appeals only to the least problematic kinds of entities, it faces serious challenges. Though many of our claims about musical works may be paraphrasable into claims about sets of possible performances, some seem to make intractable reference to works. For instance, most performances of The Rite of Spring – even including the possible ones – include several wrong notes. Thus it is difficult to imagine how the paraphrase schema will avoid the nonsensical conclusion that The Rite of Spring contains several wrong notes. The solution to this problem seems to lie in an appeal to the work as independent of its various performances, but such an appeal seems unavailable to the nominalist. (For a recent defense of nominalist theories against some standard objections, see Tillman 2011.)

Platonism, the view that musical works are abstract objects, is currently the most popular view, since it respects more of our pre-theoretic intuitions about musical works than any of the other theories. On the other hand, it is the most ontologically puzzling, since abstract objects are not well understood. Nonetheless, Platonism has been tenacious, with much of the debate centering around what variety of abstract object musical works are. What we might call ‘simple Platonism’ (known simply as ‘Platonism’ in the literature), is that works are eternal existents, existing in neither space nor time (Kivy 1983a, 1983b, Dodd 2000, 2002, 2007). According to ‘complex Platonism’, musical works come to exist in time as the result of human action. The view is motivated by a number of features of musical practice, including the intuition that musical works are creatable, the attribution of various aesthetic and artistic properties to works, and the fine-grained individuation of works and performances (e.g. in terms of who composed them, or what instruments they are properly performed upon) (Ingarden 1986; Thomasson 2004b; Wolterstorff 1980; Wollheim 1968, 1–10, 74–84; Levinson 1980, 1990c; S. Davies 2001, 37–43; Howell 2002; Stecker 2003, 84–92).

In contrast to all these realists views stand those of the anti-realists, who deny there are any such things as musical works. An early proponent of such a view is Richard Rudner (1950), though it is difficult to say whether he is best interpreted as an eliminativist or a fictionalist, the two anti-realist views currently on the table. According to eliminativists, there are no such things as musical works, and thus we ought to stop trying to refer to them. Ross Cameron (2008) defends such a view, but only with respect to ‘Ontologese’ – the language we speak when we do ontology. He argues ordinary English locutions such as ‘there are many musical works’ can be true without there being any musical works. (For critical discussion, see Predelli 2009 and Stecker 2009.) According to fictionalists, the value of discourse about musical works is not truth, and thus we ought not to abandon the discourse despite the non-existence of its subject matter, but rather adopt a different, make-believe attitude towards it (or perhaps we already do so). (See Kania 2008c, forthcoming b.)

Much of this debate over the fundamental ontological category to which musical works belong has turned on ‘technical’ issues, that is, controversial general metaphysical claims about the nature of properties, causation, embodiment, and so on (e.g. Howell 2002, Trivedi 2002, Caplan and Matheson 2004, 2006, Dodd 2007, Cameron 2008). In the face of this, some theorists have pointed out that musical works are cultural entities, and thus the methodology appropriate to uncovering their ontological status might be quite different from that of general metaphysics (Goehr 1992, S. Davies 2003c, D. Davies 2004, Thomasson 2006, Kania 2008c). There currently seems to be as much interest in the methodological questions as in first-order theorizing. (For recent examples, see Kania 2008c, D. Davies 2009, Predelli 2009, Stecker 2009, Dodd 2010.)

2.2 Higher-level Ontological Issues
It might seem that, since musical works are multiple entities, once we have figured out their true nature, we will know what relation holds between the work and its instances. However, since the fundamentalist debate is about the basic ontological category to which works belong, resolving that debate will leave open many questions about the instantiation relation. For instance, is the use of a harpsichord required to instance Bach's Brandenburg Concerto No. 5 in performance? Would producing harpsichord-like sounds on a synthesizer do just as well? What about using another keyboard instrument from Bach's time, or a modern piano? Learning that musical works are, say, eternal types will not necessarily help settle this issue of ‘authentic performance’, which is perhaps the most discussed ontological issue, of interest to philosophers, musicologists, musicians, and audiences alike.

There have been two sources of widespread confusion in the debate over authenticity in performance. One is a failure to recognize that authenticity is not simply a property, but a relation that comes in degrees and along different ‘vectors’. Something may be more authentic in one regard and less authentic in another (S. Davies 2001, 203–5). Another is the assumption that authenticity is an evaluative concept, in the sense that ‘authentic’ implies ‘good’. That this is not the case is clear from the fact that an authentic murderer is not a good thing (S. Davies 2001, 204). Thus, our value judgments will be complex functions of the extent to which we judge performances authentic in various regards, and the values we assign to those various kinds of authenticity.

The central kind of authenticity that has been discussed is authenticity with respect to the instantiation of the work. Most agree that the fullest such authenticity requires the production of the right pitches in the right order. Pure sonicists argue this is sufficient (e.g., Kivy 1988a). Timbral sonicists argue that these pitches must also have timbres reflecting the composer's instrumentation (e.g., Dodd 2007, 201–39). Instrumentalists argue that such sounds must be produced on the kinds of instruments specified in the score (e.g., Levinson 1990d). Much of the debate is over what kinds of aesthetic or artistic properties are essential to musical works. If the limpid textures of Bach's Brandenburg Concerto No. 5 are essential to it, then one cannot authentically instance the work using a grand piano instead of a harpsichord. As such, the debate reflects a wider one in aesthetics, musical and otherwise, between formalists (or empiricists, or structuralists), who believe that the most important properties of a work are intrinsic ones, accessible to listeners unaware of the historical and artistic context in which it was created, and contextualists, who believe that a work is strongly tied to its context of creation. Stephen Davies has argued for a strong contextualism, claiming that one cannot give a single answer to the question of whether particular instrumentation is required for the fully authentic instantiation of a work. Works can be ontologically ‘thicker’ or ‘thinner’ as a result of the specifications of a composer working within certain conventions (1991, 2001). The more properties of an authentic performance a particular work specifies, the thicker it is. Thus for some works (typically earlier in the history of Western music) instrumentation is flexible, while for others (for example, Romantic symphonies) quite specific instrumentation is required for fully authentic performances.

In addition to the question of what constitutes authenticity, there has been debate over its attainability and value. Those who question its attainability point to our historical distance from the creation of some works (Young 1988). We may no longer be able to read the notation in which the work is recorded, or construct or play the instruments for which it was written. If so, full authenticity is not attainable. But we rarely have no idea about these matters, and thus we might achieve partial authenticity (S. Davies 2001, 228–34). Those who question the value of authenticity often target kinds other than work-instantiation. For instance, one might question the value of producing a performance that authentically captures the sound of performances as they took place in the context of a work's composition, on the basis that musicians were not as highly skilled then as now, for instance (Young 1988, 229–31). Such arguments, though, have no consequences for the value of work-instantiation. Some argue that although we might attain an authentic instance of a work, the idea that we might thereby hear the work as its contemporaries heard it is wishful thinking, since the musical culture in which we are immersed enforces ways of listening upon us that we cannot escape (Young 1988, 232–7). Thus the point of such authenticity is questioned. In response, we may consider not only the possibility that we are in a better position to appreciate historical works than contemporary ones, but also the remarkable flexibility people seem to show in enjoying many different kinds of music from throughout history and the world (S. Davies 2001, 234–7).

(For an excellent overview of the authentic performance debate, see S. Davies 2001, 201–53. For an investigation of authenticity with respect to things other than instantiation of the work, see Kivy 1995, Gracyk 2001, Bicknell 2005, 2007a. Some recent work has, like the fundamentalist debate, taken a methodological turn; see S. Davies 2008, Dodd 2010.)

A second area that is independent of the fundamentalist debate is that of comparative ontology. Just as classical works from different historical periods may be ontologically diverse, so may works from different contemporary traditions. Theodore Gracyk has argued that instances of works of rock music are not performances. Rather, the work is instanced by playing a copy of a recording on an appropriate device (1996). Stephen Davies has argued that rock is more like classical music than Gracyk acknowledges, with works for performance at the heart of the tradition, albeit works for a different kind of performance (2001, 30–6). One can defend Gracyk by trying to find a place for live performance and performance skill within his basic framework (Kania 2006).

Work on the ontology of jazz has centered around the nature of improvisation, particularly the relation between improvisation and composition (Alperson 1984, 1998, Valone 1985, Brown 1996, 2000, Hagberg 1998, Gould and Keaton 2000, Sterritt 2000, and Young and Matheson 2000). Even apart from widening the scope of the philosophical discussion of music, this has been a useful reminder that not all music is the performance of pre-composed works (Wolterstorff 1987, 115–29). However, it must be noted that improvisation can occur within the context of such a work, as in the performance of an improvised cadenza in a classical concerto. Some have argued that there is not as significant a distinction between improvisation and composition as is usually thought (Alperson 1984). Others have argued that all performance requires improvisation (Gould and Keaton 2000). Yet others restrict the possibility of improvisation to certain kinds of musical properties, such as ‘structural’ rather than ‘expressive’ ones (Young and Matheson 2000). However, the arguments are not compelling. Usually they turn on equivocal use of terms such as ‘composition’ and ‘performance’, or beg the question by defining improvisation in terms of deviation from a score or variation of a limited set of ‘expressive’ properties.

Though jazz is not necessarily improvisational, and very few jazz performances lack any sort of prior compositional process, the centrality of improvisation to jazz presents a challenge to the musical ontologist. One might argue that jazz works are ontologically like classical works – composed for multiple, different performances – but that they tend to be thinner, leaving more room for improvisation (Gould and Keaton 2000, Young and Matheson 2000). The difficulty is to specify the work without conflating one work with another, since tokening the melody is not required, and many works share the same harmonic structure. As a result, some argue that the performance is itself the work (Alperson 1984; Hagberg 2002; S. Davies 2001, 16–19; 2003, 156). One problem here is parity with classical music. If jazz performances are musical works in their own right, it is difficult to deny that status to classical performances of works. This seems to multiply works beyond what we usually think is necessary. A third possibility is that in jazz there are no works, only performances (Brown 1996; 2000, 115; Kania 2011b). This is counterintuitive if ‘work’ is an evaluative term, but it is not obvious that this is the case.

A third topic of ontological discussion at the ‘higher’ level is the nature of the elements of musical works, such as melodies, harmonies, and rhythms, and how they come together to form complex wholes. At present only Roger Scruton (1997, 19–79; forthcoming) and Stephen Davies (2001, 47–71) have addressed these questions in any depth, but they will be important to understanding the very nature of music, if an appeal is made to such ‘basic musical features’ as pitch and rhythm in the definition of music (see section 1.2, above).

2.3 Scepticism about Musical Ontology
Several philosophers have expressed scepticism about the possibility and value of studying musical ontology. Aaron Ridley (2003a; 2004, 105–31) argues that such study is a waste of time. He argues (i) that there aren't any genuine ontological puzzles about music, (ii) that musical ontology depends on musical value judgments, and thus (iii) that musical ontology has no implications for musical aesthetics. The first claim is itself puzzling, since it is hard to understand why there would be such disagreement among philosophers about the ontology of music if there were easy answers available. Moreover, Ridley does not set out the supposedly obvious ontology of music, but rather makes substantive and contentious ontological assumptions throughout his argument. His second claim is questionable, since it seems that, like many other things, one can only correctly judge the value of something in terms of the kind of thing it really is (Walton 1970, 1988), and the kinds of things musical works, performances, and so on, are is precisely the central question of musical ontology. If Ridley is wrong about his second claim, then he is wrong about his third. However, it might be argued that some debates in musical ontology are more distantly connected to questions of musical value than others. For instance, while the debate over authentic performance is closely related to the question of the relative value of different performances, the debate over the fundamental ontological category of musical works seems more neutral with respect to that same question. (For further discussion of Ridley's scepticism, see Kania 2008b, Bartel 2011.) Lee B. Brown (2011) argues for similar conclusions with respect to ontological theories of particular musical traditions, such as rock or jazz. The discussion continues in Kania 2012 and Brown 2012.

Amie Thomasson has expressed a more measured scepticism about certain art-ontological debates (2004a, 2005, 2006). She points out that in grounding (and re-grounding) an artistic kind-term, such as ‘symphony’, there is a problem of identifying the kind of thing one intends to pick out with the term. (This is an instance of a more general ‘qua-problem’. For instance, when naming your car, there is a question about how you can successfully name the whole thing, rather than just the windshield, or a temporal part of the car, and so on.) Thus, one must have some basic ontological sortal in mind which determines the identity and persistence conditions of the thing being picked out. Since art-kinds are determined through widespread social practices, the way to discover the basic sortal is to investigate artistic practice. Thomasson's scepticism arises from her view that our artistic practices may be vague or incomplete with respect to some questions, such as how many notes one can get wrong before simply failing to perform a certain work. On the other hand, the answers to some questions are unequivocally revealed by our practice, according to Thomasson, such as that musical works are created. Thus, while her theory is sceptical about some issues, it is quite the opposite about others, ruling certain theories, such as Platonism, out of court.

One worry for Thomasson's view is that artistic practice might be even messier than she supposes. If artistic practice offers not just vague or incomplete, but contradictory answers to some questions, then it is presumably the ontologist's job to offer a theory that is the best overall reflective reconstruction of the practice (Dodd 2005, 2010). The view might also be criticized from a different direction. Even if artistic practice implies unequivocally that musical works are created abstracta, if it can convincingly be shown that there are no such things as created abstracta, then practice must be wrong. This strategy raises once more the question of the nature of musical works, and whether traditional forms of metaphysical argument are appropriate in the cultural realm. (See the references at the end of section 2.1, above.)