The Classical Recording as a Work of Phonography

Tore Simonsen

Norwegian Academy of Music


An unlikely candidate

I would like to take as my points of departure two sentences, one from a paper read at the Twentieth World Congress of Philosophy in Boston 1998, the other from the first CHARM symposium in London 2005. In London, Philip Auslander in his discussion of visual media said that

[...] regardless of the ontological status of recorded music, its phenomenological status for listeners is that of a performance unfolding at the time and in the place of listening. (Auslander 2005:1)

Because of this position he made the conscious decision to discuss audio recordings of music in the present tense and video recordings of performances of the same music in the past tense.
Auslander's research field is popular music of a special kind using strong theatrical elements, like glam rock. He described his position as a dilemma. He had argued that audio and audiovisual recordings of music should count equally as performances, at least for the purposes of performance analysis. But, as he said, in a spirit of phenomenological honesty, he had to admit that a playback of a video recording did not constitute the same kind of performance that the playback of an audio recording did. The video recording seemed to be bound to its origin while the audio recording seemed timeless. He did not feel comfortable with the situation at all.
In Boston, Lee Brown read a paper about his concept Work of Phonography [WP], defined as Soundconstructs created by the use of recording machinery. Let me return to his definition later, the main point here is that he tried to establish some limits for this kind of recording. Electronic music and rock music are obviously well inside this definition; in the paper he would argue that jazz recordings also were artifacts of the recording studio and as such acquired at least some features of WP. But he was not quite certain where to break off the list:
Consider an unlikely candidate-a recording by Enrico Caruso. Is his "La Donna Mobile" [sic] not an artifact of the recording studio? (Brown 1998:VI 2).
Let me at this point make clear my own background. Having had my own classical label for more than thirty years I would have no problems of describing all recorded music-classical or otherwise, from Caruso and up-as artifacts of the recording studio. In this paper I will argue for this position.

Classical music

One main thread in discussions about the ontology of music is about music as a process versus music as an object. There is no debate among musicologists (or music philosophers) that Western art music has been transformed from a partly improvised process to an exactly notated object, and that this transformation happened as modern society raised its head around 1800. This was also the background for the rise of musicology as an academic research area; music-and specifically German music-was accessible for objective measurements. Music could be analyzed and described as permanent musical objects.
This was the position for many art philosophers. Nelson Goodman would argue that dynamics, timbre, phrasing, even tempo, were unimportant for the definition of the musical object as such (Gracyk 1997:140). Edward Lippman would regard improvised music as having a social function only, the art definition being limited to notated music (Davies 2001:18). The music was an abstract entity we could experience only indirectly, by listening to the sound of it. Other philosophers tried to incorporate the performance in the definition of music: Stephen Davies suggests that "a musical work is a prescribed, performed sound structure"; music has its original structure, but must also be contextualized through its performance (Davies 2001:97).
Within the New Musicology the original text has lost even this position. It does not define the work at all, Stanley Boorman says, "it is just a version of the work, carrying elements believed to be essential to that composition" (Boorman 1999:422). And José Bowen goes even further, saying
If we want to discover what lies at the core of a work, we cannot simply go to the score, for our own perception of the score is guided by current period style. [...] If there is such a thing as the 'work itself ', it is fully embodied and inseparable from its performance and reception (Bowen 1999:444).
But if we cannot separate the work from its performances, the consequence must be that music does not exist until it is heard in performance (Dusman 1994:131). Thus we may establish the position that there exists no past music at all, all music is "reconstructions shaped in the present" as Stephen Davies formulates this most extreme position within the ontology of classical music (Davies 2001:95).

Work of Phonography

Musicologists and music philosophers had no interest in the recorded sound whatsoever before the post-war advent of electronic music. Early writers on electronic music, like Linda Ferguson, arrived at the conclusion that it could not be called music at all; it had to be another, still nameless, art form, since no performance exists (Ferguson 1983:17). Stephen Davies calls this work for playback (stored as machine-specific instructions and only realized as such; examples are Elektronische Musik and musique concrète from the mid fifties) (Davies 2001:25).
But this nameless art is just a small part of the recorded musical universe. Early attempts to categorize the rest of this universe have resulted in constructions like "produced music", "electronic works not for performance" etc. for studio productions of music not intended for live performances, and "recorded performances" for recordings of live music. A dichotomy like this, with "produced music" up against "recorded performances",  has many merits; most jazz recordings may be regarded as a way (and the only way) of experiencing a specific performance, and most studio recordings of rock and pop (what Theodore Gracyk calls rock music) are clearly produced music, not for performance (Gracyk 1996).
Gracyk's work on the aesthetics of rock music has been hailed as one of the first attempts to view studio productions in a greater musicological and ontological context, but one problem has been that he limits his definition of studio recordings to popular music. Lee Brown discusses this and reminds us of many classical productions that can clearly only be realized as studio productions (and not as recordings of performances); examples are Stokowski's recordings from between the wars and Georg Solti's opera recordings from the sixties with John Culshaw as producer (Brown 2000:367). It is in this context that he builds his definition of recorded music: Works of Phonography (Brown 1998:I 1). His definition is wider than Gracyk's Rock music. He points out that there is an element of sound manipulation in all recordings-intended or not-regardless of the record industry's focus on "transparent sound", "life-like recording", and so on. As already noted, his main point was to argue that jazz recordings were more of a WP and less of a performance.
To argue for the classical recording as a WP is not an easy task. Aron Edidin summarizes the ontology of the classical recording like this:
The function of soundings of recordings of classical performances is more complicated. Such soundings allow us to hear the performances in question, but in the composition-centered practice of classical music, the principal point of performances is to realize compositions. So the ultimate point of soundings of recordings of classical performances is that they enable us to hear compositions as performed on the recorded occasion (Edidin 1999:38).
The real problem here is in the construction "compositions as performed". A classical recording is a representation of two distinct and separate artistic creations, one composition and one interpretation. The composition is an example of what Nelson Goodman calls allographic art: it may be represented in an infinite number of different ways, all of equal value. He calls the opposite for autographic art: art objects where we have just one original; where even an exact duplication of this original does not count as genuine (Goodman 1976:113).
There is a paradox here: the main reason why the new musicology has been able to do research on performances-and so concentrate on music as a process-is the ability to record them. But at the same time this very ability transforms the sounding of the performances from one of many ways of interpreting the allographic composition to an autographic work of art, a sound structure impossible to duplicate in a new recording. The instantiation of the process has made it into a new object-easily accessible for research.
The connection between the original allographic composition and this new autographic sound object is therefore open for discussion. Glenn Gould commented this already in the sixties when he discussed-and hailed-the new possibilities for the listener:
At the center of the technological debate, then, is a new kind of listener-a listener more participant in the musical experience. The emergence of this mid-twentieth-century phenomenon is the greatest achievement of the record industry. [...] He is [...] a threat, a potential usurper of power, an uninvited guest at the banquet of the arts, one whose presence threatens the familiar hierarchical setting of the musical establishment (Goodman 1976:113).
He called this a necessity; the coupling to the history destroys our ability to evaluate a work of art, he said, "we have never really become equipped to adjudicate music per se" (Gould 1966:342). But not only is the link to the composer weakened, the same happens with the link to the performance. The American composer Elliot Schwartz wrote as early as in 1973:
The recording of Mozart's Don Giovanni has more in common with Gesang der Jünglinge or the Jimi Hendrix Experience-in terms of our experiences, perceptions, and responses- than it has with a live performance of [this] or of any other opera (Schwartz 1973:172).
What they do have in common is that they all are recordings, and I find Lee Brown's Work of Phonography as fitting a definition as any other. He mentions some distinctive features of WPs (Brown 1998: II):
* All characteristics of the sounding are constitutive of the WP. Brown uses electronic music as his example; however I would contend that we may expect that any recording, when regarded as an autographic work of art, must be replicated in all details for each presentation of the same work.
* The WP is the recording itself, not the various physical media used in transmission of the recording. Authentic playbacks of a given WP are those generated, on appropriate equipment, by appropriate means, from authentic playback artifacts.
* WPs are only accessible through audio playback.
"Is not Caruso's 'La Donna Mobile' [sic] a WP?" Brown is not very precise in his question; Enrico Caruso recorded in total three versions of this aria for Victor, two with piano, one with orchestra. All are representations of the same allographic work; all may also be regarded as quite true time-pictures of the recording sessions. But at the same time they are three separate, autographic works, permanently fixed in their sounding, and impossible to recreate exactly, even for the artist(s) themselves, should they have had the inclination and the possibility to do so.

The classical recording

Still, to regard the classical recording as an autographic work of art is a controversial position. As we weaken the link to the performance, the production and the postproduction gain more artistic credit; the finished product is therefore not necessarily about a performance, but a separate artistic statement. On one hand this is quite contrary to normal classical studio practice where High Fidelity has been a leading star for the last generation or so and where true representation of the performance-edited or not-has been the main goal. On the other hand we all know that the finished classical recording today is not a sound picture taken from the "best seat in the hall", but a highly manufactured product, with instrumental balance and dynamics far removed from the concert hall experience.
This is where I am reminded of Philip Auslander's phenomenological honesty. The sounding of a recording-even a classical one-is here and now, even if I play a "live recording", with all its elements in place of being a memento of something that happened somewhere, sometime. As a listener I have a certain power in this situation; even if I would not go as far as to try Glenn Gould's well known dial twiddling, actively shaping or editing the listening experience, I may start and stop the music whenever I like, play it loud or soft, listen actively, passively, or not at all.
Auslander's-and my-understanding of the presence of the autographic sounding leads to more questions, not least about the interpretations in a historical context. Listening to old recordings as WPs removes the historical dimension somewhat, we may read the age of the recording out of the technological limitations we are hearing-mono, low bandwidth, noise, distortion etc.-but we may accept the various interpretations as being in the present. The historical dimension becomes compressed, as listeners we are presented with a number of concurrent but different autographic art objects.
In the world of classical recordings, the postproduction stage has never been regarded as an artistic process, but instead mainly as the bringing together of all the bits and pieces from the sessions to a performance-like product. As such, it has certain well-defined limits to what is acceptable. Still, even in this context, artistic conflicts may arise. From one of my own releases I remember one production where the artist in question would not accept the producer's edits. Maybe a conflict of this kind would not be as clear-cut in Gracyk's rock music as in our classical environment, but we sided with the artist and asked for a re-edit of the session tapes. The result was that the producer no longer wanted to be connected with the end product; he would not have his name printed in the CD booklet. The result did not any longer have the artistic qualities he had envisioned during the production and post-production process.
The post-production stage may obviously make possible artistic manipulations of a much more serious character than in this example. If starting from a master with a great number of highly separated tracks the producer may take on the role of a re-creator, using his own artistic talent to highlight or suppress qualities inherent in the session files. If this might be regarded as reducing the performing dimension of the phonographic end product, I would contend that this is exactly what may be expected when considering even the classical recording a Work of Phonography.
In 2007 Glenn Gould's legendary recording of Bach's Goldberg variations was "re-performed" on a Yamaha Disklavier, controlled by MIDI signals derived from the same recording. The intention was to recreate the 1955 production, less its alleged failures like tape hiss, chair squeaking and pianist humming. But the artistic decision to limit this production to a simple recreation was solely in the hands of the producers. They could have-and in my opinion with the rights given them by the new artistic and technical possibilities-just as well decided to use this data from Glenn Gould's performance as control parameters for other sound events. This could be as small as playing the music in an authentic baroque tuning to a completely re-orchestration of the work, but still with the pianist's timing and phrasing.
Where should one break off the list? As I see it one should not do it at all-the Glenn Gould Goldberg re-performed on Yamaha is a true Work of Phonography-and so is the Caruso recording of "La Donna e Mobile" made one hundred years earlier.


Auslander, Philip (2005) Sound and vision: Record of the past or performance in the present?, CHARM Symposium 1 [ s1Auslander.pdf (2005.10.10)].
Boorman, Stanley (1999) The musical text. I Nicholas Cook og Mark Everist (red.) Rethinking Music, Oxford University Press.
Bowen, José A. (1999) Finding the music in musicology: Performance history and musical works. I Nicholas Cook og Mark Everist (red.) Rethinking Music, Oxford University Press.
Brown, Lee B. (2000) Phonography, rock Records, and the ontology of recorded music, Journal of Aestetics and Art Criticism, 58, s. 361 - 372.
- (2005) Sound and vision: Record of the past or performance in the present?, CHARM Symposium 1 [ s1Auslander.pdf (2005.10.10)].
Davies, Steven (2001) Musical Works and Performances: A Philosophical Exploration, Oxford: Clarendon Press.
Dusman, Linda (1994) Unheard-of: Music as performance and the reception of the new. Perspectives of New Music, 32, s. 130 - 146
Edidin, Aron (1999) Three kinds of recordings and the metaphysics of music, British Journal of Aestetics, 39, s. 24 - 39.
Ferguson, Linda (1983) Tape composition: An art form in search of its metaphysics, Journal of Aesthetics and Art Critisism, 42, s. 17 - 27.
Goodman, Nelson (1976) Languages of art, Indianapolis: Hackett.
Gould, Glenn (1966) The prospects of recording. I Page (red.) s. 331 - 353.
Gracyck, Theodore (1996) Rhythm and noise: An aestetics of rock, London: I.B.Tauris & Co.
- (1997) Listening to music: Performances and recordings, The Journal of Aestetics and Art Critisism, 55(2), s. 139 - 140.
Page, Tim (red.)(1987) The Glenn Gould Reader, London: Faber & Faber.
Schwartz, Elliot (1973) Electronic music: A listener's guide, London: Secker & Warburg