About CA
Contact CA

Search Journal

Editorial Board

Permission to Reprint


Site Map


One Song, Many Works: A Pluralist Ontology of Rock

  Dan Burkett

A number of attempts have been made to construct a plausible ontology of rock music.  Each of these ontologies identifies a single type of ontological entity as the “work” in rock music.  Yet, all the suggestions advanced to date fail to capture some important considerations about how we engage with music of this tradition. This prompted Lee Brown to advocate a healthy skepticism of higher-order musical ontologies.  I argue here that we should instead embrace a pluralist ontology of rock, an ontology that recognizes more than one kind of entity as “the work” in rock music. I contend that this approach has a number of advantages over other ontologies of rock, including that of allowing us to make some comparisons across ontological kinds.

Key Words
Lee Brown, Stephen Davies, Theodore Gracyk, Andrew Kania ontology, performance, rock music, song, track, work

1.  Introduction

Painters create paintings, sculptors create sculptures, and photographers create photographic prints.  What, then, are the art works of rock musicians?[1]  The Beatles created songs, tracks, concerts, and albums, but which of these should we consider to be their “works”?  Stephen Davies[2] argues that a work of rock music is a song-for-studio-performance, while Theodore Gracyk[3] and Andrew Kania[4] both identify the recorded track as the work.  Each of these positions is found wanting, however, as none of them successfully accounts for the way in which we engage with music in the rock tradition.

This difficulty in identifying the work of music is specific to the rock genre and stems from the unique importance placed upon the recorded track in the rock tradition.  This is especially evident when it is compared with the contrasting genres of classical music and jazz. In classical music, the work of the classical musician (the composer) that forms the central focus of critical attention tends to be the composition[5] (more on this in section 1).  

Conversely, the work of the jazz musician which forms the central focus of critical attention tends to be the performance.  While classical and jazz music may both be recorded, the role of this activity is far less important than the activities of composition and performance in each of these respective genres.  Classical and jazz recordings are not usually intended to be works in their own right but are a way in which performances of music from these traditions can be more easily disseminated.  This contrasts strongly with the rock tradition, where a track is often intended to be a work in and of itself and commonly contains features not found in the song or performance (more on this in section 3).

Lee Brown denies the plausibility of an ontology of rock and advocates for a healthy skepticism of ontologies devoted to specific forms of music (what he calls “higher-order music ontologies”).  I think it is too soon to reach this conclusion.  Because current ontologies of rock focus on one kind of entity as “the work” and not on all the entities that rock music produces, it does not thereby mean that we ought to abandon higher-order ontologies altogether. My solution is to embrace a pluralist ontology of rock, an ontology that recognizes more than one kind of activity or entity as a work of rock music.

Why should we be concerned with establishing what the musical works are in the rock tradition?  The motivation behind this inquiry is two-fold: first, we judge artists by the works they create; thus to judge rock musicians, we need to identify what their works are.  Do we, for example, judge rock musicians by assessing the songs they write, the tracks they record, the performances they give, or something else entirely?  

The question is far from trivial.  While some rock musicians may engage in and excel at all of these activities, many others have their specializations.  Who is the better rock musician, the artist who creates exquisite tracks yet delivers atrocious live performances, or the artist whose recordings are mediocre but whose live shows are like no other?  A question like this cannot be answered until we establish which, if any, of these entities is considered a work in rock music.  Second, once we have established what we should assess, we need to establish how we should assess it.  Kania has suggested that one of the reasons that rock music “is held in lower esteem by some” may be because its art works “have been misunderstood to be of the same kind as classical musical works.”[6]  Are rock works possibly entirely different from musical works of other genres?  In that case our traditional methods of aesthetic assessment (i.e., those that we bring to bear on works of classical music or jazz) may be wholly inadequate.

2.  The ontology of music generally

I begin this discussion with a consideration of the ontology of music generally.  Kania notes that in classical music the production of the sound event for an audience is the result of two “quite distinct groups of actions.”[7]  The first is the composer creating the musical work by writing a score, and is followed by the performing artist/s performing the work.  This claim runs counter to Platonic realism, according to which musical works are universals.[8]  On the Platonic view, a work such as the 1812 Overture has always existed, and thus was not created but merely discovered by Tchaikovsky.  Kania’s claim instead seems to echo a nominalist position, such as that put forward by Nelson Goodman.[9]  On this view, “the distinguishing feature of a work lies not in there being some abstract entity existing apart from its performances and score-copies, but in there being for each work a special kind of score.”[10]  This approach holds that a classical composer creates the musical work by writing the score.  For Goodman, “the constitutive properties demanded of a performance of [the musical work] are those prescribed in the score,”[11] and whether or not a performance is a performance of a particular musical work will depend upon whether it “has or has not all the constitutive properties of [that] work.”[12]

This traditional ontology of music fails to describe rock music for two reasons:  first, many rock songs have no score.  They are fluid creations born of improvisation and collaboration and are rarely transcribed in any detail.  If this is the case, how are we to establish when a performance is a performance of a particular work of rock music?  Where are we to find the constitutive properties of the work? Second, even where a score is created, it is not necessary for a performance of a musical work of rock to have all (or even many) of the constitutive properties outlined in that score.  Rage Against the Machine's cover of Bruce Springsteen's "The Ghost of Tom Joad" utilizes a completely different melody from the original.  Likewise, Hilary Duff’s cover of Depeche Mode’s “Personal Jesus” features entirely new lyrics and even a new title, “Reach Out.”[13]  Despite these differences, in both cases fans clearly identify the artists as performing the same rock songs.

Kania registers strong disagreement with the “unintuitively stringent” requirements of Goodman’s wider view.[14]  Jonathan Neufeld shares this concern, noting that Goodman’s account of performance is “conceptually consistent but overly rigid.[15]  For Neufeld, it is not clear “that judgments concerning the identity of the performance…do not bleed straight away into aesthetic judgments”—that is, that “the supposedly non-aesthetic, non-normative commitment to certain identity criteria of performance…is in itself aesthetic.[16]  It is this observation which leads Neufeld to argue that “ontological constraints on musical performance should be embraced, if at all, as normative claims from within the musical public sphere rather than as metaphysical constraints imposed from without.”[17]  In this way, claims about musical ontology—about what counts as a musical “work”—should be treated as claims that are critical, not metaphysical.

Neufeld, then, wants to defend the possibility of allowing “the diverse members of music practice” to determine musical ontology.[18]  It is this very sentiment that underpins my current inquiry.  The reason that the traditional ontology of music fails to describe rock music is that it does not fit with the way that members of the practice (that is, rock musicians, rock critics, and rock fans) ordinarily engage with works from that tradition.  In what follows, I will survey and assess several competing ontologies of rock (including my own) against this standard.

3.  Competing ontologies of rock

Gracyk holds that in rock “the musical work is less typically a song than an arrangement of recorded sounds.[19]  In this view, works of rock music are not ontologically “thin” songs to be instantiated in different performances but rather ontologically “thick” structures recorded as audio tracks and properly instanced through the playback of that recording.[20]  Gracyk’s ontology of rock can be presented diagrammatically as follows:


Songs are “thin” in that they carry all of the necessary elements for the literal instantiation of a particular musical work (such as who plays what, when, in what key, and at what tempo).  Tracks, on the other hand, are maximally “thick” in that they contain many more nuances than a score could ever provide.  While a correct instantiation of Beethoven’s “Für Elise” would  require adhering to the constitutive properties outlined in the score (as in the nominalist view of Goodman), a correct instantiation of the Rolling Stones’ track “Gimme Shelter” would require an exact replication of all elements of the original recording—right down to Merry Clayton’s voice breaking at three-minutes-four-seconds into the song.

In the track-centered view, rock music “is essentially dependent on recording technology for its inception and dissemination.”[21]  Davies disagrees with this interpretation, noting that “more groups play rock music than ever are recorded.”[22]  According to the track-centered ontology, such bands have not created any works of rock music.  Let us call this the “no-works problem.”  A parallel problem concerns the status of rock songs that have not yet been recorded as tracks.  In the 2011 documentary Back and Forth, Dave Grohl, lead singer of the Foo Fighters, describes how the song “Enough Space” was written and introduced into the band’s repertoire mid-tour to cater to the European mosh pits.  The song became the Foo Fighter’s concert opener and was performed live for many months before being recorded.  Much energy and experimentation were spent on the song’s conception, and it subsequently became a focus of critical attention by fans and commentators alike.  Despite this, a track-centered ontology would hold that until the song was recorded it failed to qualify as a work of rock music by the Foo Fighters.  This seems wrong.

Davies attempts to provide an ontology of rock that better addresses these issues.  He argues that, as in the case of classical music, works of rock music are songs created for performance.  There is a distinction however. While classical works are created for live performance, rock works are created for studio performance.[23] Davies’s ontology of rock can be represented as follows:

In Davies’s ontology, unrecorded songs like “Enough Space” will be considered works so long as they are created with the intention of eventually being performed in the recording studio.  (This intention is implicit in the “for” of “for-studio-performance.”)  This ontology, too, still suffers from the “no-works problem.”  Davies merely replaces Gracyk’s reliance on recording with a reliance on the intention to perform a song in the studio at some point.  In highlighting this problem, Kania notes that works-for-studio-performance seem to necessarily require a sound-engineer.  But, he continues:

[a]lthough many garage and pub bands may hope to be recorded one day, it is not clear that they write their songs with a part for a sound engineer even implicitly in mind....[T]hese bands seem to think they are providing audiences with fully authentic performances of their songs, not with performances missing a part. [24]

Indeed, many bands write rock songs without any intention of one day performing them in a studio.  In the works-for-studio-performance ontology, however, these songs would not be considered works of rock music.

Kania’s own ontology of rock attempts to join Gracyk’s idea “that the primary work in rock music is the ontologically thick recording”[25] with Davies’s notion that “rock is importantly a performance art.”[26] Kania argues that “rock musicians primarily construct tracks.  These are ontologically thick works... and are at the center of rock as an art form.”[27]  According to Kania, these tracks manifest songs without being performances of them.[28]  Kania’s ontology holds that a song is an ontologically thin structure of “melody, harmony and lyrics,”[29] and despite the central importance of tracks, such songs may also be manifested in live performances.  Put simply, songs are a sort of basic framework that may be instantiated later as either an audio track or a live performance.  It is only tracks however, not performances, that are legitimate musical works.  Kania’s more complicated ontology of rock can be represented as follows:

As Kania notes, the term “manifestation” is used in order to provide an intermediate between something authentically instantiating a work (or nonwork object, such as a rock song) and bearing no relation to that work.  A manifestation “represents the work [or nonwork object], displaying many of its properties, without necessarily being an instance of it,”[30] something which both the recorded track and the live performance do.  Note that Kania’s privileging of tracks as the musical “work ” of rock does not necessarily entail that songs and performances receive no critical attention.  Instead, the track-centered ontology is best understood as implying that while songs and performances may receive some critical attention, they are not the primary focus of critical attention in the rock tradition.  

Unfortunately, the revised track-centered ontology falls prey to the very same criticisms as Gracyk’s original track-centered ontology.  The “no-works problem” rears its head again, as do the concerns about the status of rock songs that have not been codified as tracks.

4.  Brown’s skepticism and a pluralist ontology of rock

The challenge of identifying the work in specific musical traditions has led to Lee Brown’s skepticism about higher-order musical ontologies altogether.[31]  He observes that the ontologies of rock outlined above all attempt to provide an answer to the question, “Is the work of rock music a track?  Or, is it a song?”[32]  Brown seems sympathetic to Gracyk’s and Kania’s views that the track is the primary focus of  critical attention when discussing rock music,[33] but he notes that “we can and do bring as much critical discourse to bear on rock songs as on rock tracks.”[34]  For this reason he argues that “it is unconvincing to privilege tracks ontologically over songs. But it would be equally unconvincing to privilege rock songs over rock tracks….”[35]

The central problem is that regarding neither the song nor the track as the work in rock music provides a full and satisfactory explanation for the way in which we engage with music in the rock tradition.  As Brown points out, once “we begin to choose one position over others, we have already taken the bait.”[36]  He believes that answering the question, “What is the work in rock music?” already presupposes “that there exists an entrenched concept of the work of rock music.”[37]  Brown argues that the most plausible position is therefore to deny the existence of any such concept as the work

I believe that such skepticism about the existence of a single work in rock music should not lead to a more general skepticism about works of rock music.  Instead, I contend, it should give us reason to embrace an ontology of rock that allows for a plurality of ontologically diverse works.  Why doesn’t Brown’s line of reasoning lead him to consider such a pluralism?  The most likely explanation is that it is precluded by the second basis of his skepticism—namely, that the very concept of a “work” of rock music is an “artefact of philosophy.”[38]  He argues that works of rock music “play no role in musical discourse” and can only be known as “inventions of the philosophers who conjure with them.”[39]  But this does not seem quite right.  “Works” are simply those things that the artist creates and upon which we focus some amount of critical attention.  (Just how much critical attention will be discussed later.)  To talk of works of rock music is merely to talk of those things in the rock tradition that meet this description.

The possibility of a pluralist ontology of rock seems to be a position with which Kania tacitly agrees.  He states that if our critical attention was split amongst several kinds of ontologically varied objects in a musical tradition, “then the correct conclusion to draw would...be that there are two primary foci of critical attention in the given tradition.”[40]  I take this proposal further.  A correct ontology of rock will include three primary foci of attention:  rock songs, rock tracks, and rock performances.  Consider how rock works are created.  The initial work created by a rock musician is the ontologically thin song.  This song includes a number of basic features, the presence of some number of which will be sufficient to provide the bases for a manifestation of that song.  A lack of a sufficient number of these features will not indicate a bad manifestation of a song but rather a failure to manifest that song at all.

A rock song can be manifested as either a track or a performance, and can be manifested in the same way any number of times (so there can be many different tracks and many different performances of the same song).  Both tracks and performances are also works of rock music.  The song “The Ghost of Tom Joad” is a work of rock music and so too are the tracks of the same name by Bruce Springsteen and Rage Against the Machine, as well as any live performances of that same song.  This ontology can be represented as follows:

Whereas Gracyk, Davies, and Kania all select one kind of thing as the work, I contend that an ontology that embraces songs as works and tracks as works and performances as works best reflects the way we engage with rock music.  Pluralist ontologies are a common feature of newer art forms.  Consider, for example, works of computer art. Dominic Lopes defends an ontologically pluralist account of computer art in which works of computer art create displays that can then be considered works in their own right.[41]

The argument for the track as a work of rock music is well established.  This is the thesis underpinning both Gracyk’s and Kania’s positions, and has been touted by Franklin Bruno as the “emerging consensus about the ontology of rock.”[42]  What is more controversial is the assertion that songs and performances should also be afforded the status of works.  Of these two types of entities, songs in particular receive short shrift.  Davies states that “very thin works, such as songs…are usually not of much interest in themselves…[and that] as pieces become thicker, they become more worthy of interest.”[43]  Kania concurs, holding that rock songs are not “the, or even a, primary focus of critical attention in rock” and that for this reason they are not musical works.[44]  This move from a lack-of-critical-attention to a lack-of-work-status is based in Kania’s own definition of a work of art, according to which “a work of art is an art object that (1) is of a kind that is a primary focus of critical attention in a given art form or tradition, and (2) is a persisting object.”[45]

5.  Songs as works

Is it really the case that rock songs fail to be a primary focus of critical attention?  Brown, as noted above, believes that songs clearly receive at least some critical attention.  He argues that in the rock tradition:

…we can critically compare tracks considered as tracks—Rolling Stones tracks with those of the Beatles, say. But we can also compare songs—by John Lennon and Paul McCartney, say—with those of Mick Jagger and Keith Richards.[46]

Furthermore, we can coherently discuss and evaluate rock songs without relying on features of their best-known manifestations.[47] I can coherently discuss and evaluate the song “Hallelujah” without referring to the tracks recorded by either Leonard Cohen or Jeff Buckley, or to any particular live performance.  I might refer to the song’s skillful chord progression or to the surprising lyrics of the second verse.

Bruno agrees that songs are of central importance to the rock tradition.  He observes “the relevance of songs and songwriting to the making, appreciation, and evaluation of rock music”[48] and notes that a number of rock musicians “are frequently credited with excellence or merit as songwriters by listeners and practitioners.”[49] He argues that the existence of such judgments “indicates that songs are, on some occasions, a focus of attention for those who appreciate and evaluate rock music.”[50]  Bruno further supports the argument for the importance of rocks songs by pointing to the way in which “evaluations of rock music often appeal to the relationship between a song and its realization in a particular rendition.”[51]  We might, for example, make the claim that a particular track is “over-produced,” or “under-produced.”  As Bruno notes, “there is nothing incoherent about preferring, on balance, an indifferently produced recording of what one judges to be a good song to a well-produced recording of a bad or pedestrian one.”[52]  Such a view would, however, be incoherent if it were only recorded tracks that were worthy of critical attention.

Like Brown, Bruno is convinced that “songs are legitimate objects of artistic interest and attention within rock practice.”[53]  In discussing the allegedly “uninteresting” nature of rock songs, Bruno considers that:

it is unclear why a philosopher’s judgment about these matters should override paradigmatic evidence that audiences and musicians treat songs as objects of critical and evaluative interest, especially in the context of an avowedly descriptive metaphysical project.[54]

As the track-centered ontologist alleges, the rock track may be the primary focus of critical attention and thus deserving of work-status, but shouldn’t the substantial attention that songs also receive qualify them, too, as worthy of work-status?  The primary focus of critical attention for an architect may be the building she designs, but we also treat her conceptual designs and sketches as works in their own right.  I contend that the same criteria should apply in the rock tradition.  Brown appears sympathetic to this position:

Suppose it could be shown that rock tracks get roughly seventy per cent of the critical attention devoted to rock music, with only thirty per cent devoted to songs.  Not even this unbalanced mix would warrant the strange conclusion that the rock work is a track, not a song.[55]

Both rock songs and tracks receive critical attention, and both should be considered works of rock music on their own merits.  As a result a pluralist ontology of rock is both logical and inevitable.

6.  Performances as works

What about live performances?  I earlier expressed my belief that they are works of rock music, as well.  I structure my argument to accept live performances in much the same way as my earlier argument for songs, namely by examining the manner in which audiences and musicians treat performances.  As with songs, we give much critical attention to live performances, distinguishing, for example, between “beautiful renditions” and “lackluster recitals.” Comparisons are also commonly made between live performances.  If the argument for pluralism succeeds and work status is given to any musical rock entity that is a focus of critical attention, I believe that live performances are equally deserving of such status as songs.

That rock performances receive critical attention is not the issue here; the real hurdle for live performances being afforded work status is that if they are not recorded, they are not enduring entities, a quality that Kania’s definition of a work requires.  A valid pluralism of rock music, however, does not depend on live performances being afforded the status of works.  For the present we will simply have a limited pluralism in which only rock tracks and rock songs will be identified as works of rock music.

A pluralist ontology of rock has a number of advantages over other ontologies of rock.  First, a pluralist ontology uniquely recognizes that our critical attention is focused on rock songs and rock tracks and rock performances, as opposed to only one of these.  This approach avoids the major pitfall of track-centered ontologies, which require that a band must record a song before it has created a work of rock music.  A pluralist ontology acknowledges that a rock band has created a work upon completing the writing of a song.[56]  The band then creates additional works of rock music when it performs this song live.  This ontological change addresses a similar problem with Davies’s works-for-studio-performance ontology that a band need not intend to record songs to create works of rock music.

7.  Comparisons among different works of rock music

As with any other art form, we compare works of rock music of the same ontological kind.  We compare songs with songs, tracks with tracks, and performances with performances.  The pluralist ontology may also allow us to evaluate works of rock music of different ontological kinds, for example tracks with performances, so long as those works instance the same song.  Such comparisons are common in our everyday dealings with rock music.  I might praise Pearl Jam’s track “Jeremy,” but in my criticism write that “the live performance from their concert last week was so much better.”

Kania is resistant to such a possibility, arguing that there is an important metaphysical difference between tracks and performances that should prevent us from making comparisons between the two.[57]  This difference arises from the fact that with studio recordings, one can “always in principle go back and change something until one is happy with the result,”[58] something that cannot be done when performing live.  In addition, while bands often employ similar instrumentation in live performances, they don’t usually attempt to create a “sonic doppelgänger”[59] of the original recording.  Instead, bands make use of techniques such as “extended introduction[s]... alternative lyrics... interpolated verses; improvisatory instrumental breaks; and...extended coda[s]”[60] to provide different instantiations of the same song.

Kania appears to be claiming that unless two works are of a metaphysically identical type, they cannot be fairly compared.  This seems an overly strict requirement for the practice of comparison.[61]  While there are important ontological and metaphysical differences between recorded tracks and live performances, this should not preclude us from making at least some comparisons across those ontological divides.  Henry Pratt, for one, argues that “we can in principle compare artworks that derive their value from different sets of properties.”[62]  This should come as no surprise.  We often compare all sorts of ontologically different kinds of works.  We prefer a Royal Shakespeare Company’s performance of Hamlet to Kenneth Branagh’s film adaptation of the same play, or the BBC miniseries Pride and Prejudice to the 2005 film of the same name. Such cross-ontological-kind comparisons are possible because the different works share a causal ancestry:  They are performances or interpretations of the same text.  Tracks and performances which are instantiations of the same song are no different.  In the case of rock music, our ability to compare tracks and performances also stems from the works’ shared causal ancestry:  Different instantiations of the same song all possess a sufficient number of features in common.

8.  Some potential objections to a pluralist ontology of rock

One important objection to ontological pluralism about rock music goes as follows:  When  using the expression “the musical work ‘Eleanor Rigby,’ ” we are understood to be referring to a single, unique entity (that is, the musical work “Eleanor Rigby”).  However, if ontological pluralism is adopted, there is no single entity that is the musical work “Eleanor Rigby.”  Instead, there are many different entities,[63] each of which is a musical work appropriately referred to as “Eleanor Rigby.”  Thus the sentence “the musical work ‘Eleanor Rigby’ was created by The Beatles” is not true.[64]

If the expression “the musical work ‘Eleanor Rigby’” can only apply to a single work, it is not the fault of ontological pluralism, but rather because the phrase itself is ambiguous.  When I say that my mother enjoys Pride and Prejudice, I could be referring to the novel, the film, or the televised miniseries.  Ontological pluralism allows for a multiplicity of works to use this name; therefore, when we use the expression “the musical work ‘Eleanor Rigby,’” we should not be understood to be referring to a single, unique entity.

A further objection is that if ontological pluralism is accepted, both tracks and live performances will be recognized as musical works, and instance other musical works (namely, songs).  Therefore each time we listen to a track or live performance, we actually listen to two musical works—the track or performance and the song that it manifests.  Davies considers the possibility that “a rock recording presents more than one work of art:  an electronic piece that is replete with constitutive properties and…a realization of a much thinner song.”[65]  Nonetheless, he rejects this on the possibility that “very thin works, such as songs…are usually not of much interest in themselves.”[66]  My intention has been to successfully disprove this allegation of the “uninterestingness” of rock songs.  I have argued above that we, at least sometimes, find rock songs interesting and bring critical attention to bear on them.  There seems to be implicit support from Davies that if songs are interesting, there is nothing very problematic about rock tracks and live performances simultaneously presenting us with an experience of two different musical works.  Bruno agrees, stating that:

Some rock recordings, including many too canonical to dismiss, afford listeners access to two objects of potential artistic interest:  that is, two works.  One is the recording itself (or, perhaps, the sound-event-type instanced on playback).  The other is the song of which it is a rendition.  Listeners make and articulate appreciative, critical, and evaluative judgments about works of both kinds.[67]

Given ontological pluralism, it is possible to experience more than one work simultaneously; this occurs in many art forms.  When I watch a film adaptation of Macbeth I also experience two works, the director’s film and the bard’s original play, and there does not appear to be anything particularly controversial about such a claim.

If a compelling case has been put forward for a pluralist ontology of rock, an ontology of rock that allows for more than one kind of entity to be a work of rock music, one objection remains.  I have here argued here for three kinds of entities as musical  works of rock on the basis that each receives critical attention in the rock tradition. But what stops us at these three?  If the criterion for something being a work of rock music is that it receives critical attention, then surely there are other entities that might qualify as works of rock.

A pluralist ontology is supposed to provide a third alternative to having either a single work of rock music (as Gracyk’s, Davies’s, and Kania’s ontologies do) or no work of rock music (as Brown’s ontology does).  Any ontology must provide an acceptable explanation for which entities or things we consider as works of rock music.  This is not complicated to decide if we accept that there is only one kind of thing that is the work in rock music and that it is that entity which is the primary focus of critical attention.  We simply pick that thing upon which we focus our critical attention.  A pluralist ontology is more challenging and, if we are not discerning about the lines we draw, we may find ourselves descending a slippery slope towards “megapluralism.”[68]

There are two ways in which this descent might occur.  First, one could begin to treat components of works as works in themselves. We might, for example, consider a particular verse of a song to be a work of rock music.  We might then go further, insisting that a particular line should receive this same status, and so on towards a reductio.  The simplest way to avoid this is to restrict ourselves to being pluralist about works of different kinds, but not pluralist about different parts of the same work.  We cannot, simply put, divide an already identified work into additional separate works.

The status of rock albums provides an example for this division dilemma.  Rock fans often bring critical attention to bear on both albums (e.g., Sgt. Pepper’s Lonely Hearts Club Band) and the songs on them (e.g., “Lucy in the Sky with Diamonds”).   Applying the pluralism I have presented, it is sufficient for albums to be afforded work status.  The division restriction, however, would preclude us from counting tracks as works, because tracks are a division of albums.

The pluralist might respond to this problem by claiming that when evaluating albums, rock fans are not bringing critical attention to bear on the album as a work in itself but rather as a collection of other works.  Consider an analogy with paintings.  I may say that “the Monet exhibition at the museum is truly inspiring,” but this does not mean that I am bringing critical attention to bear on the exhibition as a work in itself.  Instead, my utterance is shorthand for a bundle of judgments I have made about each of the works of Monet contained within the exhibition and, most important, the relations between these works (e.g., their thematic similarities, their arrangement within the gallery space etc.).  So, too, is it with rock albums.  The claim that “Sgt. Pepper’s Lonely Hearts Club Band is a phenomenal album” should be taken as shorthand for a bundle of judgments I have made about each of the tracks contained within the album and the relations between them as they are arranged on the album.  In this way, we can account for the critical attention we bring to bear on an album without having to posit it as an ontological entity over-and-above the songs that comprise it.

Even if we can appropriately apply the division limitation, the descent to megapluralism might result from deciding what kinds of things can properly count as a work of rock music. Consider other things the rock musician might create.  Clearly, we cannot label everything a rock musician creates a “work of rock music.”   We might say that a work will be a work of rock music where it has a specific causal relationship with our appreciation of a work that is musical in nature (i.e. something like the song, track, or performance). 

9. Conclusion

Ontologies of rock that select only one kind of thing as the musical work generally fail to reflect the way in which we ordinarily deal with rock music.  The solution to this problem is not to be skeptical about higher-order ontologies but rather to adopt a pluralist ontology of rock that recognizes the importance of rock songs, rock tracks, and rock performances.  I have defended this pluralist position, arguing that these three kinds of things are equally deserving of work-status, given the substantial critical attention each receives in the rock tradition.  I have also outlined several advantages of this approach, including its ability to allow for cross-ontological comparisons between recorded tracks and live performances of the same song.  In adopting a pluralist position, we embrace a unique ontology that is distinct from ontologies of other musical traditions and that more accurately reflects the way in which we engage with rock music.[69]


Dan Burkett

Dan Burkett is a PhD student at Rice University.  His areas of interest include social and political philosophy, the philosophy of time travel, and the philosophy of rock music.  He has contributed chapters to several books on popular culture and philosophy including Open Court Publishing’s Futurama and Philosophy and Homeland and Philosophy, and Blackwell’s The Ultimate Star Wars and Philosophy.

Published June 2, 2015.



[1] For the purposes of this paper, I adopt Theodore Gracyk’s definition of rock music as “popular music of the second half of the twentieth century which is essentially dependent on recording technology for its inception and dissemination.”  Theodore Gracyk, Rhythm and Noise:  An Aesthetics of Rock (Durham:  Duke University Press, 1996), p. 13.

[2] Stephen Davies, Musical Works and Performances:  A Philosophical Exploration (Oxford:  Clarendon Press, 2001).

[3] Gracyk.

[4] Andrew Kania, “Making Tracks:  The Ontology of Rock Music,” The Journal of Aesthetics and Art Criticism, 64,4 (Fall, 2006), 401-414.

[5] Contemporary composers may have some involvement in the creation of a live performance of their work, but this will be in a different capacity as a conductor or performer.

[6] Kania (2006), p. 401.

[7] Ibid.

[8] Lydia Goehr, The Imaginary Museum of Musical Works:  An Essay in the Philosophy of Music (Oxford:  Oxford University Press, 2007),
p. 14.

[9] See Nelson Goodman, Languages of Art:  An Approach to a Theory of Symbols (Indianapolis:  Hackett Publishing Company, Inc., 1976).

[10] Goehr, p. 21 [original emphasis].

[11] Goodman, p. 117 [original emphasis].

[12] Ibid.

[13] Thanks to Lisa Seddon for providing me with this particular example.

[14] Kania (2006), p. 410.

[15] Jonathan A. Neufeld, “Musical Ontology: Critical, not Metaphysical”, Contemporary Aesthetics Volume 12 (2014), section 3 [emphasis added].

[16] Ibid. [emphasis added].

[17] Ibid., section 2.

[18] Ibid., section 3.

[19] Gracyk, p. 1 [emphasis added].  To be clear, Gracyk acknowledges both the rock song and the rock track as works of rock music, but concentrates his argument on a consideration of which of these is the primary focus of critical attention.

[20] Ibid., pp. 1-98.  These terms are first used by Stephen Davies in “The Ontology of Musical Works and the Authenticity of their Performances,” Noûs, 25,1 (March, 1991), 21-41.

[21] Ibid., p. 13.

[22] Davies, p. 32.

[23] One potential counter-example to this claim might be a film score, which clearly is not created for live performance.  However, while classical scores and film scores share many features (most noticeably, their tendency to use orchestral music), this does not necessarily entail that the latter is subsumed by the former.  The argument could be made that there are sufficient differences between the ways in which classical scores and films scores are composed and performed to justify classifying the latter as a different genre altogether.

[24] Kania (2006), p. 402 [emphasis added].

[25] Ibid, p. 404.

[26] Ibid.

[27] Ibid.

[28] Kania is extending Gracyk, who initially proposed that both tracks and performances are manifestations of songs (see Gracyk (1996), p. 18).

[29] Ibid.

[30] Kania (2006), p. 405.

[31] That is, ontologies of specific music traditions, such as rock and jazz.

[32] Lee B. Brown, “Do Higher-Order Music Ontologies Rest on a Mistake?” British Journal of Aesthetics, 51,2 (April 2011), 174.

[33] Ibid., p. 173.

[34] Ibid.

[35] Ibid., p. 174.

[36] Ibid.

[37] Ibid.

[38] Ibid., p. 181.

[39] Ibid.

[40] Andrew Kania, “In Defence of Higher-Order Musical Ontology:  A Reply to Lee B. Brown,” British Journal of Aesthetics, 52,1 (January, 2012), 100.

[41] Dominic McIver Lopes, A Philosophy of Computer Art (Abingdon:  Routledge, 2010).

[42] Franklin Bruno, “A Case for Song:  Against an (Exclusively) Recording-Centered Ontology of Rock,” The Journal of Aesthetics and Art Criticism, 71,1 (Winter, 2013), 65.

[43] Davies, p. 22.                             

[44] Kania (2006), p. 404.

[45] Ibid., p. 413 [emphasis added].

[46] Brown, p. 173 [emphasis added].

[47] Bruno, p. 67.

[48] Bruno, p. 65 [emphasis added].

[49] Ibid., p. 67 [emphasis added].

[50] Ibid.

[51] Ibid.

[52] Ibid.

[53] Ibid., p. 65.

[54] Ibid., p. 72.

[55] Brown, p. 173.

[56] The “completed” part is important here as, prior to completion, the song (or part thereof) is better categorized as a “work-in-progress” than a “work.”  Exactly when a song is complete will depend on the beliefs and intentions of the rock musician/s responsible for its creation.

[57] Kania (2006), p. 403.

[58] Ibid.

[59] Ibid, p. 407.

[60] Ibid.

[61] Thanks to an anonymous Contemporary Aesthetics reviewer for helping to clarify Kania’s argument here.

[62] Henry John Pratt, “Categories and Comparisons of Artwork,” British Journal of Aesthetics, 52,1 (January, 2012), 52.

[63] Including the song “Eleanor Rigby,” the track “Eleanor Rigby” that appears on the 1996 Beatles album Revolver, and the many different live performances of “Eleanor Rigby.”

[64] Thanks to Alexey Aliyev for bringing this potential objection to my attention.

[65] Davies, p. 33.

[66] Ibid., p. 22.

[67] Bruno, p. 72 [original emphasis].

[68] Thanks to Andrew Kania for bringing this concern to my attention.

[69] With thanks to Sondra Bacharach and Steven Crowell for their helpful feedback on earlier drafts of this paper, and to the Contemporary Aesthetics reviewers and copy editor for their incredibly useful comments.