In lieu of an abstract, here is a brief excerpt of the content:

Notes 59.1 (2002) 97-99



[Access article in PDF]

Book Review

The Cognition of Basic Musical Structures


The Cognition of Basic Musical Structures. By David Temperley. Cambridge, Mass.: MIT Press, 2001. [xvi, 404 p. ISBN 0-262-20134-8. $45.] Music examples, illustrations, bibliography, index.

David Temperley's The Cognition of Basic Musical Structures is an important contribution to the field of tonal music theory. His preference rule approach appears to reflect very well the cognition of meter, phrase, counterpoint, harmony, and key, and with further testing this approach may help answer some basic questions about the foundations of music conceptualization. A preference rule theory such as this will bring to mind the work of Lerdahl and Jackendoff (Fred Lerdahl and Ray Jackendoff, A Generative Theory of Tonal Music [Cambridge, Mass.: MIT Press, 1983]), and the book is indeed a useful extension of this earlier work. (Temperley was a student of Lerdahl.)

The preference rules are used to design computer models whose outputs are then judged for accuracy against the author's intuitions and largely self-evident human [End Page 97] analyses of music scores. For example, Metric Preference Rule 6 reads: "Prefer to align strong beats with changes in harmony" (p. 51), which matches normal tonal practice as well as the rules one would find in most textbooks; the four Harmonic Preference Rules are tested on examples from Kostka and Payne (Stefan Kostka and Dorothy Payne, Tonal Harmony, 3d ed. [New York: McGraw-Hill, 1995]) and judged against these authors' analyses (which are presumably the result of traditional ear-and-eye analysis). Some of the rules reflect simple common sense, such as Metric Preference Rule 9: "Prefer to assign parallel metric structures to parallel segments . . ." (p. 51). For example, if a motive is understood as beginning on a weak beat, and that motive is repeated, then the repetition should likewise be understood as beginning on a weak beat. Although most of the rules might seem to go without saying, this is just the point (as I take it): many of these rules have gone without being said, and some have been assumed without explicit understanding. Temperley's preference rule system offers a reasonable hypothesis about how we conceptualize the basic structures upon which musical meaning is founded, and it promises to offer an explicit understanding of these basic structures.

Phrase Structure Preference Rule 2 is another sensible rule (if one is not too picky about the definition of "phrase")—"Prefer phrases to have roughly eight notes" (p. 69) —but while the author notes how well this rule applies, he then adds parenthetically that the reasons for this are unclear. He cautiously suggests a correspondence between vocal breaths and phrase length, as well as the limits imposed by the psychological principle of "chunking." These are both good suggestions, but the rule is nonetheless cast as mysterious, having an unclear basis. One reason for this is the premise of the entire project, stated at the outset: "If we assume that harmony, metrical structure, and the like are real and important factors in musical listening, then listening must involve extracting this information from the incoming notes. How, then, is this done; by what processes are these structures inferred?" (p. 1). That is, music has certain properties which we then extract and infer. Given this perspective, it would seem somewhat mysterious why so many phrases should correspond to the eight-note rule—and indeed most if not all of the preference rules are similarly mysterious: they model cognition well, but why they do so remains an unanswered question. This is not a criticism of the preference rules, but is rather an observation regarding the starting point of this project: the embodied experience of music is largely ignored by these models, leaving a lacuna between sound and cognition. For example, Contrapuntal Preference Rule 1 reads: "Prefer to avoid large leaps within [melodic] streams." This seems a very sensible rule, but why so? If the voice were the paradigmatic reference, we could begin to approach an answer: it feels different to sing steps versus leaps. (Why the voice should be the paradigmatic...

pdf

Share