1/3/2002

Narrowing the Gap: Composition and Music Theory

Beginning primarily in the 19th century, there has been a tendency in aesthetics to isolate the world of the artist from the rest of us. From that rarified space, said the Romantics, a genius creates his masterworks in a way we ? and often the artist himself ? can?t hope to comprehend. Although countless philosophical tides have swept over the musical landscape since their heyday, the image of Beethoven and Bach on an inaccessible pedestal still pervades our culture in many ways. I don?t believe this view has been given the requisite critical treatment to justify its continued role, however softened, in much of our common speech and thought. I hope to investigate in what ways and for what reasons it affects our other notions about music for good or ill, and at the same time explore other ways we might analyze the creative process.

I don?t mean to imply that the model of creativity in which inspiration dominates more familiar forms of labor is itself something new. The list of platitudes about music is a long one but it is not a very satisfying one, to borrow a quip from Alperson?s introduction; this notion of composition is not the least among them. For example, the Greeks had the Nine Muses to inspire their writers, poets, and musicians; to be prolific and/or brilliant was to be in their favor, no less fortuitous than fair weather from Poseidon was to a sailor. Nor were the early philosophers ignorant of this view?s implications; Plato, for instance, accepted that artistic invention was different from the usual skills, but concluded that the former was the inferior pursuit. As Wolterstorff explains in ?The Work of Making a Work of Music,? because inspiration of this sort required no familiarity with Form comparable with ordinary craft, it did not incorporate ?true knowledge.?

But Plato?s epistemology was hardly in vogue in the secular West when the idealization of the creative artist began again in earnest to seize the popular imagination. The Muses might have been accepted in a symbolic role, but criticism based on ?true knowledge? would have been dismissed out of hand by 19th-century relativism. In music, for example, they postulated a celestial ?Urklang? which the composer vainly attempted to capture in worldly notation; to the modern humanist, though, there was no doubt that the Urklang was itself a product of the composer?s genius (and as such, no more or less ?true? than any other human pursuit). De-emphasized, then, was the model of passive reception, though in its place grew a reverence for the artist?s cosmic struggle. In this way, the metaphysics being offered were changed, but the essential gap between artist and audience was widened if anything.

The most evident consequence of this gap is seen in our attempts to answer the question ?what is music?? I believe Sparshott offers many useful answers in his essay ?Aesthetics of Music: Limits and Grounds,? centered on his analogy between music and conversation. For example, the conversational model allows each participant ? composer, performer, listener ? to perceive a piece of music in different ways without violating its identity as a single work. Each participates subject to some guidelines suggested by the others? work, but never bounded in a strict way. Being a model, this view is itself in no way necessary, but as we will see, where competing models might conflict Sparshott offers some means for criticism.

When a gap exists of the kind suggested by Romantic views of composition, our hypothetical conversation begins to break down, and with it the harmony between the theoretical elements it allowed. If a composer?s role is reduced to communicating an Urklang, no matter how excellent, it is as if his contribution to the conversation is scripted. At the same time, idealizing the Urklang urges us toward treating the work as an end in itself ? as a command instead of a recipe, in Sparshott?s terms. Thus the inevitable variations in performance once again threaten its identity in a way to which the new theory seems ill equipped to respond; the art of listening is given a renewed normative bent that might enhance appreciation, but proves detrimental to the individual?s ability to perceive expressiveness on his own terms (as, I feel, we must).
If we are inclined to merely recognize the value of music in this fashion, whatever we gain is surely outweighed by the loss of music?s communicative role. Music?s adoption of some functions of language has been explored by many philosophers, but Sparshott is especially compelling: although [as noted by Langer] music is cannot be referential as an ordinary message?s words can be, to communicate meaning it must be neither wholly predictable nor unpredictable since in either case it would carry no information(1). Yet music as Urklang invites criticism from both extremes: if a ?masterwork? embodies perfection a priori, its internal features must perfectly follow its internal order (i.e. predictably); at the same time, these patterns exist in the otherworldly realm of the imagination, giving even the composer difficulty.

No wonder then that the incomprehensibility of the Urklang became a direct (though somehow admired) corollary. If we abandon our model in this way, however, separating brilliant music from noise reasserts itself as a philosophical burden, with possibilities like expressiveness seeming a very distant goal. Our conversants have become not just stilted, but tongue-tied. Contrast this picture with Berleant?s view of composition as growth: not only do the conflicts with our conversation-model disappear, but we see at the outset the potential for linking the composer?s dynamic process with one of our own ? experiencing emotion. The constraints on performance also seem relaxed in a positive way. He quotes in ?Musical De-composition? an observation of Elliott Carter that the best performers generate an improvisatory flair; the illusion that music is being created at that very moment (vs. a timeless Urklang) makes its communication that much more effortless.

Another problem with the ?Urklang? hypothesis is its overemphasis on the composer as an individual and his work as necessarily unique. Trivially, we might point to composers who have worked together ? or much more commonly, influenced each other and incorporated material from their predecessors. These criticisms could be easily explained away, though, without rejecting the underlying notions of holism, transcendentalism, and so on. If a work remains more than the sum of its parts, of what consequence is it that some of those parts are adapted from the work of others? As long as all the parts originated in the unfathomable realms of some composer?s imagination, wouldn?t they remain inaccessible to concrete analysis?

More critical is the way this radical individualism defines a work: it implies that it must be in some sense complete, its internal structure self-sufficient. The Urklang is not just ontologically prior to a work?s performance and perception; it is perfect in a way no realization in our world can be(2), though composer and performer must mitigate the difference as best as possible. Therefore we saw with the dawn of the Romantic period a trend for composers to specify more and more nuances within the score. When compositional details began to be considered necessary in some sense, there was less need seen for external references; a score could be elevated to the position held by a ?text? in literary criticism or perhaps even a ?theory? in the concurrently evolving world of science.

Thus, with the rise of such ?absolute music,? the semantic role that in my view defines music grew less important. The Urklang need not be ?listened for? in the Sparshott sense ? it simply is. Profound, no doubt, but this philosophy stretches our model of music as conversation to the breaking point and beyond. The composer is no longer just reading from a script; he is silencing the counter-dialogue. Rose Subotnik offers especially devastating criticism of this development: while 19th century music maintained its grasp on Western culture by offering alternative categories with which listeners might still identify (the masterwork, the concert repertoire), once even these were shunned as unnecessary(3), 20th century music lost ?the illusion of itself as a self-evidently necessary structure? ? in effect, its audience.

Moreover, by viewing a composer?s work ? genius though it may be ? as fundamentally transcending our understanding, we cripple not only his critics but also his potential heirs. The sort of ?structural listening? encouraged by Hanslick has drawn more than its fair share of criticism for his claims that it alone is valid, but certainly the other extreme is no less restrictive. Yet such is precisely the result of a large gap between composer and listener: if we cannot comprehend the purely musical devices employed in a work, we cannot put our analytical tools to work in their full capacity. Imagine if the only possible reaction of a composition student to a masterwork were pure awe ? Schopenhauer?s ilk might be pleased, but he would have learned nothing. This example is extreme, to be sure, but hopefully it illustrates ad absurdum some consequences of maintaining such a gap.

I doubt I could call myself a composer-in-training, but I certainly listen as a student of music and musical theories. The first part of this essay contrasted the supposed gap separating us from musical works with their ability to converse with us on an emotional level; in addition to being moved in this fashion, though, I?m also constantly aware of the mere ?measurable tones? at work in a way I believe equally incompatible with this gap. For me, the relationships between tones on the smaller levels of melody and especially harmony are most significant; barring something like a rondo, I don?t notice the overall organization unless there?s a score in my hands. Nevertheless, there?s no doubt other listeners may find attention to structure and form fruitful, relegating the study of individual tones to the necessarily more detached mode of analysis employed by the eyes.

No combination of these is an invalid approach to music as conversation, but music as Urklang would imply that no such attitude beyond mere appreciation could reveal anything of the true nature of the work. In this radically holistic view, any attempt to reduce it to its components cannot succeed. Yet what is ?success,? at the expense of the various communicative roles I?ve described? I would argue, in fact, that the ability of the mind to directly engage whatever pieces of a work it finds intelligible and follow them as they evolve in time is what enables a work to be expressive at all(4). (Even Peter Kivy?s simplistic model shows him finding beauty in the individual elements of a work.) Furthermore, if we never remove masterpieces from their pedestals, we can never learn from them. The goal of musical analysis is not deconstruction ? from which we might be justified in protecting our favorite geniuses with an aura of mysticism ? but edification.

The gap-theory?s reluctance to, say, clip short excerpts into a textbook is perhaps its most sympathetic role. Certainly no 8-measure passage can encapsulate Montiverdi or Messaien. How then can we claim to ever really understand them? I claimed a loftier ambition, but what do we really mean by learning? It is a truism in the pedagogical world that proficiency with a subject is easily observed by one?s ability to write about it in original and convincing fashion. How to extend this metaphor to the musical domain? Composing one?s own music certainly might demonstrate some influence from the works being studied, but would be analogous to criticizing a novel by writing fiction.

Moving past literature?s examples, consider instead a typical assignment in computer architecture: writing a program that simulates a chip. The student need not concern himself with every design consideration that went into the chip itself, but on the other hand must understand some of its functions so completely that he can codify in his own ?words? the rules that govern them. Musical comprehension develops in much the same fashion: a formal property is examined in context until it can be defined in general. I argue that this skill, too, coincides with the ability to write a computer program, since the thought process involves combining the flexibility of an original product (in the sense shared by a critical essay) with the logic mandated by the subject(5).

There is no need to claim holistic understanding of a work just yet, though, any more than someone simulating a computer chip necessarily knows how to build one. Even after writing a program that accurately identifies parallel fifths, awkward leaps, and so on, there is no contradiction in claiming less than complete mastery of counterpoint. However, once we?ve reduced the ?rules? governing the design of a CPU or a sonata to something a machine can understand, not only has the process enhanced our own understanding, but the path toward implementation is greatly clarified. If nothing else, we awaken the possibility of straightforward trial-and-error; more advanced methods rightfully belong to the domain of Artificial Intelligence, but even this branch of information science by its very existence removes much of the mystique that might intimidate a composer. Nor are the remaining difficulties deficits of the analogy; the reason I can?t write a program that completely simulates melancholy is simply that I don?t understand it well enough(6). To be sure, the computer scientist?s viewpoint narrows the gap without closing it, but it shifts our focus from a static object (a chip, a work) to its functionality ? how and why it affects the world as it does ? which is after all what we care about in the long run.


(1)Nevertheless, we must not strictly correlate information with value. There are far too many ways in which music is meaningful to us at various times & places.
(2)Goodman?s insistence on this sort of idealization (thus making any deviation ?wrong notes?) is the main reason I don?t bring up his theory of symbols in my discussions of music as language here or in the exam.
(3)Or in fact detrimental to the ?need? to be esoteric.
(4)This is the thesis of my response to question #2 and will be treated in depth there.
(5)Obviously, though, I don?t dismiss the abilities of the majority of music theorists who are simply not fluent in the necessary way, any more than I would literary criticism not written in Japanese.
(6)And even this is not to say I haven?t collected a lot of data about which musical elements induce melancholy (under various circumstances) and which do not. More generally, a machine that understands things like representation in music I don?t consider impossible, merely impractical. Only expressiveness, with its obligatory tie to experiencing a work sequentially and with concurrent mental reactions, seems unfit for simulation proper ? but even it can and should be subject to the same learning scheme.


Back to PhilosophyMusic