Performance as the Bermuda Triangle of Musical Ontology: Identity versus Variety, and the Persistence of the ‘Text’

Anthony Pryer, Goldsmiths’ College, University of London

Introduction

One of the central targets of modern performance theory has been the traditional hierarchical distinction between the original work and its performances. Indeed, the organizers of the Helsinki conference specifically asked its participants to consider: ‘how and why and in whose interests does that hierarchy have to be there?’ Moreover, the implied impatience behind that question is understandable. First, performance is now commonly seen less as an act of reproduction than as an event, and, moreover, as an event with its own independent revelations, values and social meanings. [1] Second, there have always been certain kinds of work that seem to arise as a result of performance (rather than the other way around) through acts of what we call ‘improvisation’. And third, many developments in modern theory – the death of the author, the blurring of textual boundaries, the interest in performative meanings, the questioning of the ‘metaphysics of presence’, the engagement with the Body as a central signifier of meaning – have sought to undermine the primacy of the ‘work-concept’ [2] and instead attempted to explain the import and interest of music as a by-product of the ‘aura’ and ‘authority’ of performers and their activities. [3] Hence the very clear orientation of the title of the Helsinki conference: ‘The Embodiment of Authority: Perspectives on Performances’.

However, in the background to this rich array of new perspectives lie three questions that we must face if we are to dislodge the traditional hierarchical model from its pedestal. First, can works simply ‘arise’ from the sum of their performances? Second, if the work does not have primacy, what are musical performances performances of? Third, if the identity of the work is forever in doubt, how are we going to claim that at least some of the aesthetic qualities that are displayed through performance are indeed those of the artwork and not merely those of the (perhaps uniquely whimsical) performance? These questions sum up in a general way the central problems of the ontology of music; [4] they represent, if you like, the Bermuda Triangle - the plughole - into which our theories and explanations have tended to disappear, forever lost in the watery darkness. 

Not surprisingly, the temptation has been to gloss over these difficulties with some hastily constructed substitute notions – for example, that the work is not a exactly a ‘text’ but rather something akin to a dramatic ‘script’; [5] or that the performance itself can be treated as a work in its own right because it has aesthetic qualities; [6] or that a work is simply the sum of its performances. This last formulation has gained something of a foothold in writings about performance, [7] and it serves as an interesting example of a notion that may look promising but which does very little to dispel the ontological difficulties of music. In the first place, if we reverse the traditional hierachy and simply say that what we call ‘the work’ emerges from the sum of its performances, we then have to explain how we know exactly which performances are relevant to the construction of a particular work without already having some shadowy notion of the work against which to check them. Next, some performances might contradict others, or the last performance might be cancelled - and in the latter case we would need to give some account of how we knew that it was indeed the ‘last’ performance, the one needed for final completion to take place. Finally, what could those performances possibly be ‘interpretations’ of, since the work – we are told - does not exist until they are all over? Clearly, in attempting to dismantle the hierarchy between works and performances the Bermuda-Triangle effect has a disconcerting tendency to reassert itself. We need to proceed with caution.

If we are to loosen up the hierarchy between works and performances we need to understand in some detail the nature of the links between a work and its performances; and we need to locate exactly where the potential for greater freedoms between them might lie. We also need to understand a little about ‘hierarchy theory’ since not all hierarchies are arbitrary, or formed only by obsolete habits and elitist ideas. Some hierarchies are rational or causal in origin – a fact that may emerge more clearly as this paper progresses. [8] The following investigation of these issues will fall into four sections. First, the special functions of performance in the performing arts will be analysed, particularly in relation to those performances that claim to be of a pre-existent text (or score). Next there will be some consideration of the work-performance relationship in apparently ‘non-texted’ performances such as improvisations. The third section will look at the potential for flexibility, first in relation to the notion of the work, and then in relation to performance. And finally the fourth section will make some observations about the location of ‘musicness’ in all this, and suggests some ways in which the relationship between performance and the work might be reconfigured. 

1.  Performance Functions in Relation to Pre-existent Works

One way to get at the relationship of performance to text in the performing arts is to ask a simple question: since we perform out the meaning in our heads of all the artworks that we come across – whether novels, paintings or beautiful buildings – why are not all of the arts performing arts?  The usual answer to this centres on the fact that in interpreting pictures, for example, the seriation, the exact sequence of information through time, does not matter, whereas in musical performance it does. We might call this way of discovering the information we receive the seriation function of performance. 

Of course, other types of artworks (such as novels) also require a fixed seriation of the information for understanding to take place, but if we can read (and so pick up the connected concepts), the process is relatively straightforward. On the other hand, exactly how we might read meaning into the seriations, the sequences of sound, in an abstract art such as music is a complex issue. 

One such way is provided by what we might call the human-agency function of performance. The fact that performances are usually rendered by human beings makes it easier for us to hear the intensities of the material sounds as somehow appropriate to, or as an accompaniment for, human feelings, gestures and responses. In other words the human-agency aspect of performance assists us to hear the sounds figuratively or symbolically – it provides not only a focus of human identification but, in so doing, it fulfils a metaphor function. It becomes one of the chief means by which a mere set of notational events, a mere sequence of sonic intensities, can be heard as if it is a meaningful narrative, or a dialogue, or a developing argument. 

That said, any particular human performer will select from the many potential ways of displaying the notes, particular ingredients and relationships he or she wishes to foreground – in other words they will impose individualized patterns of intensity and suggestion upon the music and in so doing then fulfil the function of aspectival emphasis, a function which we usually call ‘interpretation’.

But showing that a set of sounds can be emphasized in a particular way, or might be taken as a metaphor, is, at heart, only a neutral piece of information. What happens next is that, by their manner of display, certain performances win us over to the view that such musical interpretations – perhaps of question and answer, coherence or struggle, closure or incompleteness – are convincing, or disturbing, or partly obscure, or are worthy of our reactive imagination and sympathy. Thus musical performance has yet another function which we might call the persuasive function: it persuades as well as informs the audience, and through its apt intensifications, convinces us to commit to a certain aspectival way of listening, and to enjoy it. 

Indeed, so entrenched is the assumption of persuasive human agency in performance, that even when there is no real-time, live performer – as in certain types of computer-generated composition, or studio-based mixes in popular music – we tend to hear the display of the music as if it is a performance, and not merely an effect carefully engineered behind the scenes by someone making calm judgments about which knobs to turn. And because the human-agency element of performance provides persuasive metaphors of coherence, we can also hear non-texted performances such as improvisations as if they are revelations of a text, even if that ‘text’ is unfolding for the first time before the ears of the performer as well as the listener. Moreover, the gestural surface of the display makes us feel as though the performance must be about something, and we take that ‘something’ to be a ‘text’ of some kind, even if the ‘text’ is little more than a concatenation of the playing techniques acquired by the performer. In these circumstances the performance is ‘about’ the skill and variety of the performer’s technical facilities.

Musical performance, of course, takes place in sounds. This means that a performance – unlike a written score, or the composer’s idea, or a critic’s description of a performance – can be directly sensed as displaying what we might call ‘musicness’, a  soundscape that we perceive as having meanings than cannot be completely replaced by any other medium, whether words, analyses, expressive splashes of paint, or anything else. This is what we might describe as the musically specific, non-substitutable aesthetic function of musical performance. This means that we do not just glean information about details of the work from the performance. Rather, our knowledge and understanding, gained in that particular way, are infused with the special manner of their begetting, and it is through this manner of begetting that we move beyond the realm of mere meaningful remarks and into the realm of remarkable meanings. And those remarkable meanings are taken to be aesthetic.

But the relation of aesthetic qualities to the status of performance needs to be treated with some caution because even if we could clearly establish that performances had aesthetic qualities themselves, that would not be enough to show that performance was an independent art form in and of itself. After all, a sunset might be aesthetically pleasing, but it cannot be well-crafted or ironic or a compelling commentary upon the human condition, and we would need to marshal some very tortuous arguments indeed to plead that sunsets were artworks in virtue of merely having aesthetic qualities. (Even the notion that they were ‘created by God to sustain our sense of wonder’ would not do, since not all ‘creations’ are artworks, and neither are all ‘message-mediums’). 

This means that even if we could show that performance on its own had aesthetic features, this would still not confirm performance as an artform in its own right, and nor could it show that performance was anything more than a craft, because craftworks have aesthetic features as well. [9] For musical performance (as opposed to performance art) to be an independent artform, we would need to show that it operated in a genuinely independent way – that there could be brilliant, valuable performances of banal works; that we had clear notions of what constituted performing genius (aside from being able to understand someone else’s works well); that performance had its own art-historical traditions of discovery and innovation, independent of changing compositional challenges and styles; and so on. Clearly the claims of performance to some kind of ‘artness’ would need to be made separately from the status of the things they perform, and this may not be easy. After all, the performance attributes described so far of seriation, human agency, aspectival emphasis, persuasiveness and aesthetic import seem to arise inevitably from the requirements of the artwork itself – in that sense they come second to the work in the ‘causal hierarchy’. The issue of whether a performance might, in some circumstances, be synonymous with a new artwork, and so free itself from this hierarchy, will be explored in the next section.

2.  ‘Un-notated’ Performances and the Persistence of the ‘Text’

It could be argued that although we have explained why a musical text might need performance, we have not explained why performance might need a musical text.  Indeed it is often assumed that certain kinds of performance, such as improvisation, do not need a text. However, the situation here may be slightly more complicated than it seems, and to uncover those complications we need eventually to divide non-texted performances into four rather different types (which may all, none-the-less, imply the shadowy presence of some kind of text even if in different ways). The four categories are: (1) those that are composing centred (such as group improvisations); (2) those that are performer centred (such as ‘featured artists’ concerts where ‘works’ are ‘spontaneously’ created, or used merely as ‘pretexts’ for performer display); (3) those that are performance centred (such as certain kinds of performance art, where the performance, rather than any separate, extractable content, is the ‘work’); and those that are idea centred (as in certain types of conceptual art where any  performance employed is purely for the purpose of ‘idea-exemplification’, a purpose which, it seems, could frequently be fulfilled in another way entirely). [10]

Since time is short today I will talk only about improvisations. At one level the spontaneous aspects of improvisation would seem to suggest that any musical ‘text’ emerges from, rather than is prior to, the performance. But if this were simplistically true it would be a rather odd kind of text. After all, no matter how skilful, free improvisation tends to lead to isolated splashes of sound, unbalanced forms, arbitrary voice-leadings, and a tendency to privilege only a small number of parameters – such as fragmentary melodic exchange, or rhythmic gesture, or the methods of call-and-response, and so on. And the somewhat locally generated nature of these elements can make it difficult for us to read meaning into the emergent seriation of sounds. [11]

Against these random tendencies there is another factor at work, however, and that is preparation. There is much evidence to show, for example, that in some oral cultures, the performance traditions and the particular works presented (including their apparently free embellishments) are at least as rigidly controlled as in written cultures - if not more so – either by tradition, or through the hallowed, controlling influence of established masters of the art, or through the repertory of techniques, or formal structures, or favourite melodic fragments and harmonic progressions already assimilated by the performer through training and practice. In other words, in both oral and written cultures, a great deal of the performance is prepared, even within the so-called improvising practices of both cultures.  With this in mind we can draw a kind of map – a notation-culture matrix – where both oral and written cultures have the same type of division at their heart, the division between prepared elements and unprepared elements, and where they are posited as producing similar categories of work (see Figure 1). And if in oral cultures the remembered, hallowed instructions of the master act as a kind of virtual notation, they still constitute a sort of controlling ‘text’ – and one which is prior to the performance in terms of the causal hierarchy. 

Of course, whatever kind of notation the performers may use – whether real or virtual – there are bound to be at least slight variations in the performance, and so the next question to ask is what level or kind of variation could still allow us to claim that the various performances involved are of the same work? And this brings us to the next sections on attempts to theorize the limits of allowable flexibility, both within the text itself, and within the performances that may emanate from that text. I should just mention here that some of you will have noticed that, in my list of performance functions – seriation, human agency, aspectival emphasis, the persuasive, the aesthetic - I have not included straightforwardly the function of displaying the work. To be sure, all of these functions are related to understanding the work, but I have reasons, which I shall come to shortly, for wanting to avoid implying that the performance simply overlaps with the work in some way. 

[Figure 1]

3. Performance Variety and the Identity of ‘The Work’

Just how rigid or flexible is the identity of a musical work? Only by exploring this question can we begin to get some idea of how we might dismantle the causal hierarchy between the work and the performance. A typical way in which philosophers deal with this issue is to say that we should imagine the work as a kind of abstract template, a so-called ‘type’, and that any particular performance is therefore a concrete token of that type. [12] There are, for example, many different kinds of editions and copies of Jane Austen’s novel Pride and Prejudice. They are in different formats, on different coloured paper, with different print sizes and paginations – but all of these matters are merely contingent. What makes them tokens of the type ‘Pride and Prejudice’ is that they share what philosophers call ‘the sameness of spelling’. But ‘sameness of spelling’ here does not refer to the fact that all of the words have the right letters in the right order (it is not a matter of simple orthography), rather it refers to the fact that the novel has all of the right words in the right order – no more and no less – and only by that means does the copy reflect the abstract type “Pride and Prejudice”. 

One problem of the type-token approach is that a token of a type (a particular performance, say, of Beethoven’s Ninth Symphony) must share exactly and completely the right number of notes in the right order, otherwise – philosophically – it is a token of a different type, a different abstract work. This might be called the ‘one-wrong-note problem’, and it weakens the theory considerably, because it is counterintuitive, and because it requires us to postulate a new abstract type every time an adventurous or mistake-prone musician takes the stage. In other words, it does not allow us to account for variations in performance in relation to the identity of the work – the very issue we are interested in.

One way around this would be to look at the notational aspects to see if any of them actually invite variation by their very nature as it were. Some theorists have tried to capture a view of music in this mold by characterizing scores not as texts, but as scripts. [13] After all, scripts in plays, films and other types of drama do seem to invite what we might call a kind of looseness of realization. 

But there are some problems with this analogy. For example, a play by Shakespeare might be open in certain productions to a looseness of setting in terms of what we might call the contextual semiotic. If we place Othello in sixteenth century Venice we get one kind of implication, if we place it in present day Zimbabwe we get another – particularly if we make all the main characters black apart from Othello who is now represented as white. [14] But what we need to remember is that the identity of the play can survive this kind of transference precisely because the plot and the characters have been drawn in carefully controlled detail by Shakespeare, and they are sharply drawn in virtue of the actual words designated for that purpose by the author. Moreover, the words – unlike the contextual setting – are not normally subject to transference, deletion or embellishment. It seems clear that appeals to the so-called looseness of the script generally refer to the looseness of the contextualising of the script (rather than the looseness of its detailed content), an operation that in itself need not lead to much freedom or deviation from the text by the individual performers. 

But there may be another way of considering the various types of information contained in scripts and also in musical scores. We could begin by dividing that content into what we might call strategies and archives. In music, for example, the archival elements would consist of specific rhythms, pitches, forms and the like, whereas the strategies would involve invitations and intentions to produce certain kinds of effect in the most appropriate manner on the occasion of the performance. Some of these strategic elements are already quite familiar to us in terms such as ‘cadenza’, or ‘rubato’, or ‘espressivo’ – after all, we know what effect is required by the term ‘espressivo’ but it would be very difficult to define what archived feeling, if any, was supposed to be represented in an ‘espressivo’ moment. To go alongside these familiar kinds of strategic terms, there are also, one imagines, strategies implied by the score that inform even our use of the archival elements such as particular tempo markings, rhythms and dynamics. Such implied strategies might include  ‘please use all the archival elements to create in some way a sense of dénouement at bar 200’; or ‘please create a sense of balance in the texture and formal structure – even if your ways of achieving this are not ones envisaged by me, the composer’. 

I mention this last condition because we have evidence from several composers, but especially from Brahms, [15] that on occasions they were delighted by interpretation-solutions that they had never foreseen or were at odds with what they had originally intended. Normally, of course, the archival elements in a score are far more in evidence than the strategic ones, but there are two further points to make here. One is that in any case the strategic elements can certainly on occasions take precedence over the local archival details (think how a cadenza might be notated as compared with what any particular performer might actually do), and that therefore performance variety is sometimes subsumed in the notation usage. The other is that some works – particularly avant-garde compositions of one kind or another – are almost entirely written in strategic notation rather than archival notation. A famous case in point is Cage’s 4’33”, the notation of which provides a strategy for creating a certain kind of event, but it can make no attempt to archive the specific details of the ambient sound content on any particular performance occasion. [16] All of this is to say that the identities of works can be defined as much by strategies that leave their exact realizations to the performer, as by archives which do not. 

So much for potential freedoms already clearly encoded within the text, as it were, whether by archives of information, or carefully defined strategies for producing particular effects. But what of freedoms that may spur on the performer apparently to neglect the exact text or break away from it – freedoms, for example, of expression, display, and characterisation? Part of the difficulty here is that we have traditionally assumed that the identities of works are completely encoded in their notational and formal features, and we need to consider for a moment whether the identities of artworks have more to them than this. We can perhaps approach this question by using an analogy. 

Suppose we were looking for representations of a  lion. Normally we would expect four legs, two eyes, a long tail, sharp teeth, and all the other usual attributes. But supposed our physically complete lion was timid and sickly, and a rather poor hunter, whereas another specimen had been in many fights, had lost an eye and half its tail, but was brave, fearless and dangerous. It would be physically incomplete at a technical level, but would vividly exemplify at another exactly those attributes that made a lion a distinctive, admired and attractive animal. What this tells us is that identity does not have to be represented by the complete archive of attributes, so long as the most vivid and distinctive ones are exemplified. In other words, depending on your purpose, there is always going to be a hierarchy of identity markers, and the trick is to get the right ones for the right occasion – a vet or a zoologist will have a different set of criteria from someone who wishes to make a lion the central character in a children’s story or who wants to choose a mascot for a football team. 

When we choose a work for performance what we are doing is choosing a particular context for the reception of that work, with a generally known and anticipated set of expectations attached. And just as sickly timid lions will not do for mascots, mechanical, undifferentiated but notationally accurate performances of Rachmaninov’s piano concertos will not do for concert receptions either, even if they do have all of the notes in the right order. Of course, the context-relevant identity-markers of a Rachmaninov piano concerto performance must have enough notational features for it to resemble other instances of the work, but not so many that the resemblance collapses into mere replication, or suppresses other characteristics essential to the performance context.

Just how we might negotiate this space between resemblance and replication will be addressed in the final section, but there is one further point to make about the lion analogy. The person who chooses to exemplify certain attributes of the lion is not revealing those attributes for the first time to a world ignorant of lions or their characteristics. They are rather selecting attributes already known, and which they have some confidence will be widely recognized as attractive and desirable. In other words, they are not reproducing an identity, they are using an already known identity (‘lion’) to exemplify attributes likely to be taken to have wider reference and application. Similarly, the performer is not introducing an unknown type of object to the public – artworks and piano concertos and expressive pieces of music are already familiar categories – rather he or she is attempting to demonstrate how a particular notational scheme may exemplify or be shown to have a relationship to those wider categories. The task of the performer is to take a set of neutral marks or instructions from the composer (which of themselves hold little aesthetic or emotional interest) and assimilate them through performance into an already-known wider referential world of ‘style’ and ‘musicness’ and ‘expressive expectation’. And to do this he or she must select from the many distinctive, identifying archived features of the score those aspects most likely to make that assimilation vivid and compelling. During this process, some archive features not absolutely essential to the impact of the assimilation – that is, some features rather low down on the hierarchy demanded by that particular setting – may fall by the wayside, even though different operations (such as music editing or analysis) would clearly have different hierarchies and selections.

4. Performance and ‘Musicness’

We now need to suggest some specific ways in which we might theorize the links between performance variety and the fixed identity of the work. Part of the problem seems to be that when we refer to what we call a musical work, we are really referring to something that has two rather separate components – there is what we might call the identity bit (the things that distinguish this creative design from other creative designs), and there is the ‘musicness’ bit (that is those aspects that result from the performer transforming the created instructions into audible music). Note that even a visitor from Mars could probably spot the distinct identity of a notational scheme without even knowing what music was. True, the Martian might mistake the marked page for a unique kind of wallpaper, but he or she (or it) would not need to know that the marks represented sounds. Of course we, not being Martians, know that notations and scores and the ideas of composers however signalled - even though they usually have distinct identities - are not yet music, they are only potentially music and need to be realized through performance. For that reason I would not call scores ‘music’, but would rather, for the sake of clarity, call them ‘performables’. [17] What then happens is that the performer comes along and performs the notation, and out of that transaction the work is revealed to have musical import as well as identity import. 

This does lead to a slightly novel but, I think, useful conclusion, which is that there are no musical works as such, there are only musical performances. In fact I would be happy with this formulation since it splits the criteria for identity from the criteria for musicness. Thus we could then say that the so-called ‘works’ (performables) are not of music, they are for music. And this means that neither the score alone nor the performance alone encodes the musical work completely. Instead the musical work is something that emerges from a dialectic between the two. This suggests that there may be a kind of double ontology going on in what we call in a general way ‘music’ – the ontology of the individual performable, and the ontology of the generic type ‘musicness’. 

This might help us go at least a little way towards explaining the link between performance variety and the fixed identity of works. The fixed identity of a work comes from the performable, which already combines definite archival information (pitches, forms, and so on) with various types of strategic instruction (cadenza, ‘open the piano lid for 4’33” ‘, etc.).  These features in any case inherently leave open various pathways for their fulfillment, and go some way to explaining performance variety in the service of the same performable. But a second set of attributes (sonic aesthetic qualities) still needs to be contributed by the performance in the service of “musicness” – that is, in the service of the general type of artefactuality (‘music’) that the work will possess. And, as we know, performances can employ a varied range of techniques to evoke the generic type “music”, since in that facet of their activity they only have to be faithful to the type of artefactuality, not the particular work. 

Since the musicness of the work is created out of a combination of the performable and the performance, performers always have been able to share authority with the composer. But that shared authority operates (in causal terms) in an hierarchical chain – first the performable (whether a written score, or an imagined strategy, or a mentally prepared archive of techniques and notes), then the performance, then the revealed musicness of the work. And the proximity-in-action of the performance to the revelation of musical aspects of the work does not, on its own, invest the performer with more authority than the composer or the created performable. Moreover, the authority that the performer claims will only be deserved if he or she genuinely take on the responsibilities that go with such authority – the responsibility to understand and represent the performable by making best use of its archived attributes, and the responsibility to understand the characteristics of music well enough to create interesting, varied instances of it. This cannot be done by simply covering everything in a kind of emotive wash, or by playing a particular performing tradition rather than the work, or by using the performable as a pretext for performance narcissism. Performance authority, like any authority, implies certain kinds of responsibility. But once those special types of responsibility are grasped, then the medial position of performance in the causal hierarchy of the musical work can be transcended by another proper to the performer – a hierarchy that gives first place to qualities that arise from what the sociologist Max Weber called charismatic authority; [18] qualities of skill, insight, sensitivity and compelling and appropriate display. 

Notes

1. See, for example, Simon Frith, Performing Rites: Evaluating Popular Music (Oxford: Oxford University Press, 1996), especially Chapter 10.  On the opposition between ‘works’ and ‘events’ see: Carl Dahlhaus, Foundations of Music History, trans. J. B. Robinson, (Cambridge: Cambridge University Press, 1983), 132ff. The ontology of ‘events’ is at least as complex as that of ‘works’: see Lawrence Lombard, ‘Event Theory’, in Jaegwon Kim and Ernest Sosa, eds., A Companion to Metaphysics (Oxford: Blackwell, 1995), 140-144.
2. Lydia Goehr, The Imaginary Museum of Musical Works: An Essay in the Philosophy of Music (Oxford: Clarendon Press, 1992).
3. A classic account of the ‘aura’ of performers occurs in: Walter Benjamin, ‘The Work of Art in the Age of Mechanical Reproduction’, in his Illuminations, ed. Hannah Arendt, trans. Harry Zohn (London: Fontana, 1973), 211-244.
4. For a recent investigation into the ontological theories of music (albeit with a Platonist conclusion) see: Julian Dodd, Works of Music: An Essay in Ontology (Oxford: Oxford University Press, 2007).
5. For a sustained attack on this analogy see: Peter Kivy, The Performance of Reading: An Essay in the Philosophy of Literature (Oxford: Blackwell, 2006), 65ff. See also Nelson Goodman, ‘Script’ in his Languages of Art: An Approach to a Theory of Symbols (Indianapolis: Hackett Publishing, 1976), 199-201. 
6. For an introductory analysis of possible distinctions between ‘art’ and ‘the aesthetic’ see: Marcia Eaton, ‘Art and the Aesthetic’ in: Peter Kivy, ed., The Blackwell Guide to Aesthetics (Oxford: Blackwell, 2004), 63-77.
7. Contemporary composers of improvisatory or graphic compositions sometimes assert this, but since they rarely support the notion with a sustained argument it would be invidious to single out particular individuals. Philosophers tend to be a little more nuanced: for example, Nelson Goodman (Languages of Art, 186) defines the work as the class of performances compliant with the score, though his rules for ‘compliance’ are very strict and counter-intuitive; and Roman Ingarden in his The Work of Music and the Problem of Its Identity (London: Macmillan, 1986), at 154-6, defines the work as an intentional object but one which only reveals its full aesthetic properties through the history of its performances, though his criteria for the revealed aesthetic properties being of the work are nebulous.
8. See for example: Valerie Ahl and Timothy Allen, Hierarchy Theory (New York: Columbia University Press; 1996); and Albert Halsey, ‘Hierarchy’ in Adam Kuper amd Jessica Kuper, eds., The Social Science Encyclopedia (London: Routledge and Kegan Paul, 1985), 356-7.
9. The classic account of the distinctions between art and craft is in Robin Collingwood, The Principles of Art (Oxford: Clarendon Press, 1938), 15-41.
10. This is my own categorization, but the literature on these different approaches is vast. On the boundaries specifically between the performing arts and performance art see, for example: Stephen Foster, ed., ‘Event’ Arts and Art Events (Ann Arbor: UMI Research Press, 1988); Peggy Phelan, Unmarked: The Politics of Performance (London: Routledge, 1993); and Peter Goldie and Elisabeth Schellekens, eds., Philosophy and Conceptual Art (Oxford: Clarendon Press, 2007).
11. For further discussions of the ‘limitations’ of improvisations as compositions see: Carl Dahlhaus, ‘Composition and Improvisation’ in his Schoenberg and the New Music, trans. Derrick Puffett and Alfred Clayton (Cambridge: Cambridge University Press, 1987), 265-273; Edward Gioia, The Imperfect Art (Oxford: Oxford University Press, 1988); and Andy Hamilton, Aesthetics and Music (London: Continuum, 2007), Chapter 7.
12. The type-token distinction was first suggested by the philosopher Charles Sanders Peirce (1839-1914), but its application to aesthetics came later. For an early adoption of the terms in aesthetics see Richard Wollheim, Art and Its Objects (New York; Harper and Row, 1968), Section 35.
13. See, for example, Nicholas Cook, “Music as Performance” in Martin Clayton, Trevor Herbert and Richard Middleton, eds., The Cultural Study of Music: A Critical Introduction (New York and London: Routledge, 2003), 204-14, at 206ff. For two sustained analyses of the ‘script’ notion, see foonote 5 above.  For music-as-if-a-script in relation to opera arias (a special case) see: Martha Feldman, Opera and Sovereignty (Chicago: University of Chicago Press, 2007), 45ff. 
14. Someone has actually done this: in 1997 Patrick Stewart took the role of Othello with the Shakespeare Theatre Company, Washington D.C., in a "photo negative" production of a white Othello with an otherwise all-black cast.
15. See: Michael Musgrave and Bernard Sherman, eds., Performing Brahms: Early Evidence of Performing Style (Cambridge: Cambridge University Press, 2003), 5, 21, 23, 24, 31 and 32; and: Timothy Day, A Century of Recorded Music: Listening to Musical History (New Haven: Yale university Press, 2000), 187-8.
16. On the various notated versions of the score of 4’33” see William Fetterman, John Cage’s Theatre Pieces: Notations and Performances (Amsterdam: Harwood Academic Publishers, 1996), 69-84.
17. I have borrowed this term from the philosopher Nicholas Wolterstorff, but apply it in a rather different way. See: Nicholas Wolterstorff, ‘Ontology of Artworks’ in David Cooper, ed., A Companion to Aesthetics (Oxford: Blackwell, 1992), 310-314, at 310.
18. Max Weber, ‘The Nature of Charismatic Domination’, in: Walter Runciman, ed., Weber: Selections in Translation, trans. Eric Matthews (Cambridge: Cambridge University Press, 1978), 226-250. The original article dates from 1922.