8
PERFORMANCES AND RECORDINGS

Andrew Kania and Theodore Gracyk

The most common musical experience today, across most of the globe, is that of listening to a recording. For many centuries, however, music was only experienced live, since recording technology did not exist. As a result, much of the philosophy of music is rooted in the idea that music is a performance art, and recordings have been met with some skepticism (when they have been discussed at all). In this chapter, we investigate the nature of musical performances and recordings, and compare views about their respective values.


Performances

General features of musical performances

Not just any musical event is a musical performance. Consider a CD playing in an empty house. While there is music going on in the house, there is no performance within its walls. For there to be a performance going on, there must be people performing. Performance is thus a kind of action – something only people (not machines, such as CD players) can do. But not every musical action is a musical performance. We standardly distinguish between just messing around on an instrument, practicing, rehearsing, and performing. What distinguishes performance from the other musical activities in this list seems to have something to do with the presence of an audience. When you mess around, practice, or, rehearse, you play your instrument or sing, but you do not do so for an audience.

Is the requirement that a performance be for an audience merely intentional, or is it a success condition, in the sense that if there is no audience, there cannot be a performance? Both Stan Godlovitch (1998: 41–9) and Paul Thom (1993: 190–3) argue for the stronger claim: an actual audience is a necessary condition on there being a performance. They do so on the grounds that performance is essentially communicative, and thus requires two parties – performer and audience. Godlovitch describes two situations which, he suggests, we would only describe as “performances” in some secondary or derivative sense. In the first, a performer decides to go ahead with the evening’s performance even though no one has turned up to hear him. In the second, the performer plays a politically incendiary work in defiance of the government officials who have locked the audience out of the hall (Godlovitch 1998: 43). These thought experiments do not quite show what they are intended to, however. For in both these cases there is neither an actual audience nor an intended audience in the relevant sense. Of course these performers “intended” to perform for people, but the past tense of the verb is telling. Performances are intentionally for an audience in the sense that one’s actions (playing, singing, etc.) are guided by the belief that one is playing for people who are capable of listening to the sounds one is making. The performers in Godlovitch’s cases do not have this belief, since they know there is no one else present. One can have the relevant belief mistakenly, though. Imagine a case where the performer comes onstage and plays for the audience in the hall, only realizing after the performance, when the blinding stage lights are dimmed and house lights come up, that there is no such audience. Such a performer has the relevant intention despite the absence of an audience, and thus might be said to have performed. This view does not undermine the analogy with ordinary communication. If one is convinced there is a burglar in the house, one might utter a warning, such as “Who goes there?” Such a speech act is intentionally directed at whoever is in the house, even if it turns out that one is mistaken, and there is no such person. (This kind of thing may happen in cultures where musical works are performed for the gods. If there are two such cultures, with beliefs in incompatible deities, then if Godlovitch and Thom are right, at most one is actually engaged in musical performance. This seems wrong.)

Paul Thom gives a different argument for the necessity of an actual audience, arguing that the address of a performer to an audience is different in kind from that of non-performance artists, such as painters or novelists. The latter make a “hypothetical” address, according to Thom, “to whoever happens to be the addressee,” while as a performer, “I make a categorical address to the audience, whom I assume to exist. In performing I believe myself to be referring to present persons, to whom I am in effect saying, ‘You, attend to me.’” (1993: 192). To the extent that Thom refers here only to a belief or assumption that the audience exists, it does not establish the need for an actual as opposed to an intended audience. What remains is the idea that the audience for a performance must be (at least believed to be) present. But this condition is also too strong. For musicians can perform a live broadcast for “the folks at home” without any audience present where they play. It seems, then, that the attitudes of performing artists are not at base so different from those of other artists. They present their efforts to whomever is in a position to appreciate them. This argument could also be extended to the production of some musical recordings.

In sum, a performance requires the intention to play music for an audience, but there need be no actual audience. You might think this point is usually moot, since the performers themselves may count as the audience, in the absence of any other listeners. But it is not clear that musicians are in the right position to be the audience for their own music-making (Godlovitch 1998: 42–3; Gracyk 1997: 149 n. 6; cf. Thom 1993: 172).


Kinds of musical performances

Many musical performances are performances of independent musical works, that is, works that would exist whether or not these particular, or indeed any, performances of them existed. Philosophers have disputed what is required for a performance to be of a given work. One appealing first pass at an answer is that one must play all the right notes. But most work-performances include wrong notes, and we do not discount them as performances for that reason. On the other hand, it seems clear that if you play none of the right notes, you have failed to perform the work in question. The kicker is that it seems an impossible task to decide how many, or what proportion of notes must be correct for a performance to count as of a given work. All this suggests that some other connection between performance and work is at least necessary. One popular suggestion is that the performers must intend to play the work in question; others have suggested that there must be a particular kind of causal chain running from the work (or its composition) to each performance. (For an excellent overview of the literature on these questions, and a consideration of how to spell out these proposals, see Davies 2001: 152–84).

Another important part of this debate has been the discussion of “authentic performance practice,” which is usually centered around the question of whether a (proper) performance of a musical work ought to involve the use of the kinds of instruments contemporary with the work. The literature on this question dwarfs that on any of the others considered in this chapter; it is thus treated separately in this volume. (See Chapter 9, “Authentic performance practice.”) Many performances, on the other hand, are not performances of works. The most obvious examples are free improvisations. Such performances need not emerge ex nihilo; rather, they are cases where any materials they are based on are treated as jumping-off points for the performer’s creative activity, instead of something the performer centrally intends to present to the audience through performing it. (See Chapter 6, “Improvisation,” this volume.) Are such performances musical works in their own right? The answer turns, unsurprisingly, on the nature of the concept of a musical work. On the one hand, such performances are the primary focus of appreciation in traditions such as jazz, suggesting that if there are works of art in jazz they include such performances. On the other hand, work-performances are a primary focus of appreciation in classical music, yet we do not typically think of these as works. We could, of course, simply stipulate that performances that are of works cannot be works in their own right. It may be, though, that central to our (or one of our) concept(s) of a work of art is the idea that they are enduring entities. If that is right, then we might deny that there are works of art in jazz (and other similar traditions). This may sound like an insult to the tradition, but it should not if the sense of ‘work of art’ being employed here is not an evaluative one.


Evaluating performances

Some evaluative criteria seem applicable to any kind of performance. The ability to play one’s instrument or sing well is valued in any performance, for instance, and it may be exalted in virtuosic performances (Mark 1980). We also evaluate the musical properties of the performance, for example, its melodies or harmonies, and how they are developed over the course of the performance. The way in which such features are evaluated depends upon the kind of performance we are listening to. The virtuosity and musical features of a work-performance will be attributed to the work or its composer, while those of an improvisation might be attributed to the performer. (It is worth remembering that in attending to a work performance we attend to at least two things – the work and its performance.) Other evaluative criteria depend on the kind of performance evaluated. In evaluating an improvisation, we value the spontaneous risks the performer takes in attempting to fashion a worthwhile musical event in the moment. In evaluating a work-performance, on the other hand, we value a faithful adherence to the work. There are other things we value in work-performances, such as a performer’s ability to interpret the work, and thereby show us something new and interesting about it. Moreover, as Jerrold Levinson (1990a) has argued, there are many legitimate yet irreconcilable perspectives from which to evaluate a work-performance. A good performance for a first-time listener, for instance, may emphasize broad structural and expressive elements of the piece, while a good performance for a seasoned listener may emphasize the role of a particular motif that should not be foregrounded for a first-time listener. There are, of course, illegitimate perspectives, such as that of the monomaniacal percussionist who values the loudness of the cymbals over all else. And there may be some difficult cases. Levinson judges the perspective of a jaded listener, who values idiosyncratic performances, legitimate (1990a: 380). But there will doubtless be cases that fall in a hazy border between the legitimate and illegitimate. The variety of legitimate perspectives arises precisely because the kinds of musical works we have been considering are intended for multiple performances. This suggests that it is pointless to ask what the ideal performance of a given work would be like.


Live non-performance music-making

Musical performances are “art” in the loose sense that they are produced for an audience that is supposed to appreciate the performance in some way. But there is much live music-making that does not fit this description. Two broad types are, or have been, common. The first is communal music-making, such as the singing of hymns in church or folk songs around a campfire. In these cases there is something like a performance of a work – the singers attempt to get the notes, words, chords, etc., right – yet they are not singing for an audience (not even for each other) – in the sense in which the concert performer does. Rather, they are singing with each other. The two types of music-making may occur simultaneously, as when the audience joins the band in singing along with a hit song at a concert. The band is performing, in the sense of the term we have been using; the audience is not.

Live music-making can also be functional. Examples here include work songs and lullabies. The musicians in these cases produce music primarily for some purpose other than the appreciation of an audience, whether it is to coordinate their actions, make the time pass quickly, lull a baby to sleep, or express one’s love. (See Chapter 40, “Song,” this volume.)


Musical recordings

“These modern gramophones are a remarkable invention,” remarks Sherlock Holmes in “The Adventure of the Mazarin Stone” (Doyle 1921: 296). Holmes has just used a phonograph recording of a solo violin performance of Offen-bach’s barcarolle from The Tales of Hoffman to fool jewel thieves into thinking that he was playing his violin in a neighboring room. Heard through a wall, it is plausible that they might confuse the playing of a primitive recording with a very different thing, a performance. In any case, the phenomenon of recorded music was sufficiently familiar to the general public in 1921 to serve as a plot device in a popular detective story. Fifteen years later, Walter Benjamin and Theodor Adorno staked out opposite positions on the effects and desirability of this “mechanical reproduction” on listeners (Benjamin 1968; Adorno 2002) – after which there is a long silence on this topic in the philosophical literature. As late as 1990, philosophers simply took it for granted that listening to recorded music constitutes listening to music, without pausing to discuss whether audience response differs when listening to recordings (e.g. Levinson 1990b: 306). However, in the ensuing decades a number of philosophers took up the topic of recorded music and its role in musical experience (e.g. Gracyk 1996, 1997; Fisher 1998; Brown 2000; Kania 2009). Two general topics have emerged concerning musical recordings. First, what is the nature of recorded sound and what is its relationship to the music it records? Second, should we be concerned that so much of our musical culture now takes the form of listening to recordings? It is best to take up the two questions in that order, for it is doubtful that we can achieve an evaluative consensus when we do not yet agree on the nature of the phenomenon being evaluated (Kania 2008: 69–73).


Kinds of recordings

Consider the simple case of Sherlock Holmes “playing” Offenbach’s barcarolle on his gramophone. In 1921, it would have been a mechanical recording of an uninterrupted performance of that piece. While Offenbach’s barcarolle allows for multiple instances through multiple performances, each performance is a singular event. Yet the multiple playbacks of a single gramophone recording (and the multiple playings of multiple copies of the recording) present us with the ontological peculiarity that a single musical performance can be heard by a temporally and spatially dispersed audience. Because one cannot listen to a musical performance years after the performance ends, it seems relatively obvious that audiences for musical recordings do not actually hear the performance. They hear an imitation or representation of the sonic dimension of that performance. (However, see the following discussion of transparency.) This intuition about representation poses three problems. First, does this relationship hold for all recorded music? As will become apparent, this is unlikely. Second, where it does hold, does the recording provide an instance of the music? Third, where it does hold, can the recording faithfully capture the sonic dimension of the performance?

With the advent of electronic music (both synthesized and musique concrète), it became apparent that some musical works depend essentially on recording technology and playback. Subverting the ontological priority presupposed by Holmes’s use of the gramophone recording of the Offenbach barcarolle, these recordings directly instantiate music that cannot otherwise exist. There are no performances of such works, for their only instances are playbacks (e.g. Pierre Schaeffer’s Étude Pathétique and Milton Babbitt’s Composition for Synthesizer). Stephen Davies calls these works “for playback, not for performance” (Davies 2001: 7–8). Following Aron Edidin’s alternative terminology, these “recording artifacts” should be distinguished from two other kinds of recordings: recordings of performances and recordings of compositions (Edidin 1999). Whereas recordings of performances provide access to musical works by documenting performances of some work (e.g. Holmes’s recording of the barcarolle), recordings of compositions employ studio editing and manipulation to construct sonic manifestations of musical works that can also be instantiated in real-time performance. The intended aesthetic appeal of such recordings is not confined to their documentary function of capturing the sonic dimensions of musical performance. Thus, two different recordings of Glenn Gould’s interpretation of Bach’s Goldberg Variations possess distinct functional relationships to Bach’s music and thus have different ontological status: Gould’s 1981 studio sessions and his 1959 Salzburg live performance furnished a recording of a composition and recording of a performance, respectively. Mere listening does not necessarily reveal the appropriate category. The functional relationship to performance practice, rather than the kind of musical work that is presented, determines which kind of recording presents the music.

Discussing studio recordings of works such as the Goldberg Variations (i.e. Edidin’s category of recordings of compositions), Davies observes that they normally aim at a simulated performance that emphasizes accuracy, consistency, and finish (Davies 2001: 313–17). However, many such recordings are not limited to the function of simulating a performance. Since the 1960s it has been common for popular music recordings to employ studio techniques that create sonic events with electronic effects that cannot be reproduced in real-time performances. Such effects include movement within the stereophonic soundscape, singers who sing multiple harmonies with themselves and drum kits with cavernous echo that sounds distinctively different from the echo effect on the vocal performances on the same recording. Here, the studio manipulations furnish musical effects that can exist only in playback and which are intended to be appreciated as such. Furthermore, the recording process often serves as a non-documentary compositional tool, allowing new compositions to emerge through trial and error as additions are recorded at different times and by multiple contributing musicians (Gracyk 1996: 46–50).

Like Davies’s works for playback and Edidin’s recording artifacts, these composite, studio-enhanced “tracks” are distinct musical works, intended to be appreciated for composed musical effects that go beyond real-time performance effects (Gracyk 1996; Zak 2001). We might consider, for example, Pink Floyd’s Dark Side of the Moon, which employs sound manipulation and montage techniques to create musical patterns from “found” sounds. Such manipulation is particularly conspicuous at the beginning of “Money.” Because it is also possible to perform the song live, in real time, Davies contends that the recording is a simulated performance of a musical work of a special type: a work for studio performance (Davies 2001: 34–35). Davies further contends that this composition is the only musical work to be appreciated when listening to the fifth track of Dark Side of the Moon. Gracyk (1996) and Kania (2006) contend that non-documentary studio tracks engage listeners with two distinct kinds of musical works. There is a representational display of the basic properties of an ordinary musical composition and there is also the studio-constructed track for playback (i.e. Edidin’s categories of recordings of compositions and recording artifacts, respectively).

Against Davies, there is no reason to fabricate a special type of composition for the songs on Dark Side of the Moon, nor two types of performance – live and studio. If works for playback by Schaeffer and Babbitt are independent musical works, then so is Dark Side of the Moon. We do not require a special ontological category of musical composition for such music. We need only distinguish between three distinct modes of providing access to performable compositions: (1) real-time performance instantiations, (2) recordings of such performances, and (3) studio-constructed representations. Thus Pink Floyd’s song cycle can be heard – and differently appreciated – in its performances, in documentary recordings of its real-time performances, and in the recording of the composition that is Dark Side of the Moon.


Repeatability and transparency

Christy Mag Uidhir (2007) notes that sound recordings do not necessarily provide repeatable playbacks, for they might play back from a source that can only be used once. However, the recording technologies that interest us here were developed in such a way that complex sound sequences can be preserved and then repeated. Sound recordings are templates for generating multiple aural instances, and they can function as representations of other sound events, just as photography developed multiple-instance representations (Davies 2001: 318–19).

In grounding categories of musical recordings in distinct representational relationships to musical compositions and their performances, we have suggested that the mere activity of listening can be insufficient for determining which sort of recording one is hearing. Listening to Glenn Gould or Pink Floyd, a listener might confuse a recording of a composition with a recording of a performance, and so might admire Gould’s precision and Pink Floyd’s ensemble interaction on false grounds, the way that a naïve film viewer might attribute the feats of the stunt double to the leading man. Therefore recordings of compositions are sometimes viewed with suspicion as detrimental to musical culture (Gracyk 1997). Lee B. Brown (1996) and Davies (2001) worry that a musical culture centered on recordings will desensitize listeners to music’s interactive and performative aspects. Recordings undermine the social practice of performing music, because their repeatability counterbalances their documentary function: “the music stands in an adverse relationship with the calcifying medium with which we document it” (Brown 2000: 122). Furthermore, Brown worries that the technology has a destructive effect on improvisational music, particularly jazz, because it encourages audiences to treat non-repeatable performances as repeatable, re-identifiable compositions (Brown 1996).

The underlying issues involve the evaluative appreciation of music. There is concern that an audience for recordings will form improper expectations for performances, and so will improperly evaluate both performances and undoctored recordings of performances. (These worries are distinct from concerns about auditory degradation, which will be taken up in relation to the issue of transparency.) Such concerns are partially mitigated by noting that audiences bear some burden of responsibility for understanding that different recordings “promote different values” depending on the functional intentions behind their production (Davies 2001: 317). Furthermore, even if recordings do mislead some listeners, they provide many compensatory advantages, such as ease of access to multiple interpretations of the same composition (Gracyk 1997).

While there are important gains in being able to compare Gould’s 1955, 1959, and 1981 recordings of the Goldberg Variations, and to compare these in turn to Murray Perahia’s more recent interpretation, we may remain concerned that all sound recordings lack documentary transparency. Recordings are sonically inadequate to provide the timbral musical nuances that can be heard in a good performance venue. Furthermore, recordings are never stylistically neutral; all recordings of performances introduce some degree of sonic departure from the sound of the documented performances (Gracyk 1997; Hamilton 2003). Against this view, Joshua Glasgow argues that “transparent” recordings are possible. At least some parts of some recordings are qualitatively identical with their sources (Glasgow 2007). While such transparency is not always desirable, Glasgow defends its possibility.

Glasgow’s emphasis on sonic accuracy appears to miss the point, developed by Gracyk (1996) and Kania (2009), that transparency is fundamentally an ontological issue. Even allowing for the possibility of recordings that sound just like their sonic sources, does a documentary recording actually permit someone to hear the music? Albrecht Dürer’s self-portrait of 1500 may look very much like him, yet one does not literally see Dürer by looking at it. Paintings are not transparent. Glass windows, in contrast, are transparent. In 1500, someone could look through a window and see Dürer on the other side, and the viewer would see him even if the glass was uneven and thus produced distortions in how he looked. Kendall Walton (1984) has argued that photographs are similarly transparent, for they allow us to see (albeit indirectly, and with certain distortions) the actual things that are photographed. Can recordings of performances do the same with music? Do we literally see and hear Judy Garland sing “Over the Rainbow” when we watch The Wizard of Oz (1939)?

If sound recordings are transparent in this sense, then recordings of compositions are worrisome entities. Listening to Gould’s 1981 Goldberg recording, we cannot hear how many recording “takes” were needed, how many partial performances were spliced together, and how many days of performing were involved to produce the thirty-two musical segments. Therefore it is not possible to evaluate Gould’s playing, for we cannot determine his capacity to produce those sounds in the manner Bach intended, that is, by playing them consecutively at one sitting. The “distortion” here is not a matter of sonic fidelity. The distortion comes in a listener’s inability to keep track of what performance activity is transparently heard as the music moves forward, instant to instant. Combined with the fact that sonic fidelity is more an ideal than a practice, the merits of transparency are frequently at odds with the effects of studio manipulation and sonic infidelity (though see Kania (2009: 32) for an attempt at resolving this tension).


Conclusion

Musical performances and recordings are all alike in being essentially aimed at providing listeners with musical experiences. But this broad commonality masks a host of differences both between and within each category. Musical performances differ in their nature and aims. Some musical recordings are aimed at replicating the experience of one or another kind of performance. But other recordings are works of art in their own right, to which, in fact, some performances may bear a derivative relation. Philosophers and other theorists of music, particularly those interested in the listener’s musical experience, ought not to ignore such matters.

See also Authentic performance practice (Chapter 9), Improvisation (Chapter 6), Jazz (Chapter 39), Ontology (Chapter 4), Popular music (Chapter 37), Rock (Chapter 38), and Song (Chapter 40).


References

Adorno, T. (2002 [1938]) “On the Fetish-Character in Music and the Regression of Listening,” in Essays on Music, ed. R. Leppert, Berkeley: University of California Press, pp. 288–317.

Benjamin, W. (1968 [1936]) “The Work of Art in the Age of Mechanical Reproduction,” in Illuminations: Essays and Reflections, ed. H. Arendt, trans. H. Zohn, New York: Harcourt, Brace and World, pp. 219–53.

Brown, L.B. (1996) “Phonography,” in D. Goldblatt and L.B. Brown (eds) Aesthetics: A Reader in Philosophy of the Arts, Upper Saddle River: Prentice Hall, pp. 252–7.

—— (2000) “‘Feeling My Way’: Jazz Improvisation and its Vicissitudes – A Plea for Imperfection,” Journal of Aesthetics and Art Criticism 58: 113–23.

Davies, S. (2001) Musical Works and Performances: A Philosophical Exploration, Oxford: Clarendon Press.

Doyle, A.C. (1921) “The Adventure of the Mazarin Stone,” Strand Magazine 62 (October): 288–98.

Edidin, A. (1999) “Three Kinds of Recording and the Metaphysics of Music,” British Journal of Aesthetics 39: 24–39.

Fisher, J.A. (1998) “Rock ’n’ Recording: The Ontological Complexity of Rock Music,” in P. Alperson (ed.) Musical Worlds: New Directions in the Philosophy of Music, University Park: The Pennsylvania State University Press, pp. 109–23.

Glasgow, J. (2007) “Hi-Fi Aesthetics,” Journal of Aesthetics and Art Criticism 65: 163–74. Godlovitch, S. (1998) Musical Performance: A Philosophical Study, New York: Routledge.

Gracyk, T. (1996) Rhythm and Noise: An Aesthetics of Rock, Durham: Duke University Press.

—— (1997) “Listening to Music: Performances and Recordings,” Journal of Aesthetics and Art Criticism 55: 139–50.

Hamilton, A. (2003) “The Art of Recording and the Aesthetics of Perfection,” British Journal of Aesthetics 43: 345–62.

Kania, A. (2006) “Making Tracks: The Ontology of Rock Music,” Journal of Aesthetics and Art Criticism 64: 401–14.

—— (2008) “Piece for the End of Time: In Defense of Musical Ontology,” British Journal of Aesthetics 48: 65–79.

—— (2009) “Musical Recordings,” Philosophy Compass 4: 22–38.

Levinson, J. (1990a) “Evaluating Musical Performance,” in Music, Art, and Metaphysics, Ithaca: Cornell University Press, pp. 376–92.

—— (1990b) “Music and Negative Emotion,” in Music, Art, and Metaphysics, Ithaca: Cornell University Press, pp. 306–35.

Mag Uidhir, C. (2007) “Recordings as Performances,” British Journal of Aesthetics 47: 298–314.

Mark, T.C. (1980) “On Works of Virtuosity,” Journal of Philosophy 77: 28–45.

Thom, P. (1993) For an Audience: A Philosophy of the Performing Arts, Philadelphia: Temple University Press.

Walton, K.L. (1984) “Transparent Pictures: On the Nature of Photographic Realism,” Critical Inquiry 11: 246–77.

Zak, A. (2001) The Poetics of Rock: Cutting Tracks, Making Records, Berkeley: University of California Press.


Further reading

Davies, D. (forthcoming) The Performing Arts, Malden: Blackwell. (An excellent introduction to a range of philosophical issues raised by the performing arts.)

Day, T. (2000) A Century of Recorded Music: Listening to Musical History, New Haven: Yale University Press. (An extended exploration of the impact of recording technology on performance practices, particularly in classical music.)

Doğantan-Dack, M. (ed.) (2008) Recorded Music: Philosophical and Critical Reflections, London: Middlesex University Press. (A strong collection of essays on recorded music.)

Eisenberg, E. (1987) The Recording Angel: Explorations in Phonography, New York: McGraw-Hill. (Essays on recorded music, whose themes were taken up by Lee B. Brown and Theodore Gracyk, among others.)

Katz, M. (2004) Capturing Sound: How Technology Has Changed Music, Berkeley: University of California Press. (Trained in both philosophy and musicology, Katz covers some of the same material as Day but extends the discussion by speculating on the impact of digital technology.)

Philip, R. (2004) Performing Music in the Age of Recording, New Haven: Yale University Press. (Focusing on classical music, includes an interesting essay on “authenticity” and the Early Music Movement of the twentieth century.)