Defining the musical work in the age of recording
Music history, as a discipline, relies heavily on recorded sound. After all, we live in the age of streaming music—more recorded music is available than it ever was for any previous generation—and it might seem that there is no better way to understand music history than through historic recordings. But even though recordings are important for music historians, there are still problems that emerge in their use.
First of all, we have to deal with the concept of the musical work and how this intersects with recorded sound. The work concept proposes that there is an essential set of core elements that define a work of art, so that one can talk about “Beethoven’s Fifth Symphony,” “Miles Davis’s Kind of Blue,” or the “The Banana Boat Song” as objects that exist in reality, with clear defining features that make them what they are and allow them to be easily identified.
The work concept has been subject of a great deal of academic debate, particularly among literary scholars; much of this debate is readily translatable to music. The core question remains: who defines the work? The author (composer) who first imagined it, or the reader (listener) who interacts with it? Or is the text itself (the music itself) an independent object that has its own voice? One can pair these types of questions easily with aesthetic issues; an author-focused work concept gives a canon of Great Composers and their Great Works, with all the cultural baggage that entails, for example. For our purposes, it is enough to say that the work concept remains useful despite efforts to contest it, because it allows students to conceptualize a piece of music in coherent way.
The work concept was chiefly developed for written literature and the visual arts—art forms which can be held in one’s hand or seen in a single glance. Music, on the other hand, exists in multiple formats such as recordings, sheet music, and live performances, and it can only be understood in time. Music is defined by its relationship to time in ways that many other art forms are not—poets and novelists do not concern themselves with how quickly or slowly the reader can read, and painters and photographers rarely design their works with the explicit intention that they must be studied for a certain length of time.
This relationship to time complicates any concept of the musical work. If you look at a painting for five seconds, you can grasp tangible features very quickly—it’s circular, there is a lot of red and green, there is a figure in the center that looks like a horse. Listening to a song for five seconds tells you only about the five seconds that you heard—there is a clearly marked beat, a saxophone riff, and vocals that sound like a sigh—but it cannot tell you what to expect in the rest of the song. Perhaps the saxophone will fall silent, perhaps the beat will grow more complex. Perhaps the song will continue exactly as it started. Who can say?
But let’s go back to our list of musical works, each of which I chose to illustrate a different aspect of our subject. “Beethoven’s Fifth Symphony” shows the problems of trying to establish a musical work in the period before recorded sound—it’s not as if Beethoven made a definitive recording to which we can turn. What does the work consist of? Does it lie in Beethoven’s manuscript of the symphony or in the first published edition? These questions lead us toward textual scholarship—establishing the written text exactly. With the notated text as our only authority, it then follows that musical performances are representations of this ideal text, and not necessarily the work itself. After all, the work is altered in performance—so that “Beethoven’s Fifth Symphony” is subtly different if it is conducted by Leonard Bernstein or Daniel Barenboim or Marin Alsop.
Which is the text? Ludwig van Beethoven, Symphony no. 5, manuscript in Beethoven's handwriting, c. 1807 - 1808, posted on imslp.org by the Staatsbibliothek zu Berlin, and edition published in Leipzig by Breitkopf and Haertel in the mid-nineteenth century, posted on imslp.org by the Bayerische Staatsbibliothek.
Now here is an interesting conundrum. All serious performers of Western classical music have been inoculated against taking excessive liberties with the scores, and yet the differences in the recordings persist. But when we study music via recordings in a music history class, we often study these recorded performances as a way of accessing the work. In other words, one does not listen to Claudio Abbado conducting Beethoven, but to Beethoven as conducted by Abbado. And often, the first recording one hears of a historic work becomes the work itself; for example, I first heard Beethoven’s “Appassionata” sonata in a recording by Rudolph Serkin, and that recording shaped my idea of the “Appassionata” as a discrete work. This is an unavoidable problem, one that is not solved by being named, but one that needs to be named nonetheless.
Now for something completely different: “Miles Davis’s Kind of Blue.” This jazz album differs from the Beethoven example, in that the recording itself is the musical work and the object of study. This recording-focused approach to the work came to define jazz, particularly as jazz fans developed a culture of obsessively studying and replaying iconic recordings as a substitute for and as preparation for attending live shows. There is no room for debate about Kind of Blue’s definitive form, and there is no need to try to realize Davis’s creative vision through modern performances—Davis and his band realized that vision, committed it to recording, and moved on to other projects. (Modern editions of the album sometimes contain outtakes that do complicate this idea somewhat, but it would be unthinkable for a major jazz group to play a note-for-note cover of the album).
This example captures the difference in work concepts between jazz and classical music. Because the recording of Kind of Blue is the work, there was no need to release a definitive notated score for the album as one would for a classical work. Notated lead sheets for its tracks are designed to provide a loose template for creative improvisation rather than a strict set of guidelines to be followed. In this case, as with many other genres of music in which the recording is the musical work, having a fixed sonic text frees musicians to take creative approaches when they deal with historic music.[1]
Finally, let’s turn to the “The Banana Boat Song.” Most Americans who know this song probably know it through the recording Harry Belafonte released on his Calypso album in 1956. But this recording, although it has been sampled many times, is not the first recording of this Jamaican folk song—for example, both the Jamaican folk musician Louise Bennett and the Trinidadian singer Edric Connor recorded the tune as “Day Dah Light” in the early 1950s. Neither Bennett, Connor, nor Belafonte claimed to have composed the song. They are all copies of a missing original.
So where does the musical work lie in a folk song? Unlike Western classical musics, there is no notated score to fall back on, and these recordings cannot all represent the same musical tradition. The recordings are vastly different from each other—Bennett’s only uses voices and a drum and uses Jamaican patois extensively; Connor’s vocal timbre and piano accompaniment resemble a Paul Robeson recording; Belafonte clearly had a large group of highly-trained studio musicians available and employs a complex arrangement. It’s tempting to say that Bennett’s version is the closest to folk practice, but of course folk songs often have many local variants. Bennett’s drumming might sound more rural, and Connor’s piano might sound more urban; but are folk songs less authentic if they are sung in a city? Just as with classical music, it is very easy to confuse recordings with the musical work when it comes to folk songs.
The problems posed by folk songs capture the biggest problem with relying on recordings to understand the music of the past: simply put, that not everything gets recorded. There is a similar problem with relying on notated scores alone—historically, notation was linked to educated people who only notated the music that mattered to them or that could be sold as sheet music to a literate audience. While recordings might seem more democratic, they often function as products to be sold. So only the profitable bits of an artist’s repertoire get recorded.
Today, recordings are often marketed as if they represent an artist’s full creative work—a statement that is often true in our recording-dominated age, but becomes less accurate the further back we go in time. For example, no recording exists of Louis Armstrong playing excerpts from Pietro Mascagni’s opera Cavalleria rusticana, although he played them in live shows.[2] Presumably, record executives felt that there was little potential profit in selling recordings of Armstrong playing this music.
In conclusion, the very nature of music and of musical recordings complicates the work concept. In many genres, it is reasonable to think of the recording as the definitive version of the musical work, but in others, this attitude will lead us astray. Finally, an overreliance on recorded sound can lead students to imagine that recordings represent all musical activity in a given period.
[1] For explanation of how similar concepts work in rock music, see Andrew Kania, “Making Tracks: The Ontology of Rock,” Journal of Aesthetics and Art Criticism 64 no. 6 (2006): 401–414.
[2] Elijah Wald, How the Beatles Destroyed Rock’n’Roll (Oxford University Press, 2011), 20.