Loading field map
Music analysis is the scholarly practice of examining a piece of music in detail to understand how it is put together and how it produces its effects. Where music theory typically develops general systems for describing musical structure, analysis applies, tests, and refines those systems on individual works or repertoires. The analyst asks not only "what happens in this music?" but "why does it happen, and what does it make possible?" The field's central questions concern the relationship between a score or performance's observable features—pitch, rhythm, harmony, timbre, form—and the larger patterns, meanings, or experiences those features generate. Because music is a temporal art, analysis also grapples with the problem of representing a process that unfolds in time as a static object of study.
A foundational issue is deciding what exactly is being analyzed. For much of the field's history, the object was the notated score, treated as the authoritative repository of the composer's intentions. This orientation suited the Western classical canon, where detailed notation preserves pitch and rhythm with relative precision. But even within that canon, notation leaves many dimensions underspecified: tempo, dynamics, articulation, and phrasing are approximate, and their realization varies among performers. Analyzing a score therefore risks mistaking a blueprint for the building.
This problem intensifies when analysis turns to music without a score, or where notation is secondary. Oral traditions, popular music, jazz, and much contemporary electronic music exist primarily as sound. Analyzing them requires transcription, which is itself an interpretive act, or the use of spectrograms and other audio visualization tools. The field has responded by expanding its methods to address sound directly, but the tension between score-based and sound-based analysis remains unresolved. A related difficulty is the problem of segmentation: deciding where one musical unit ends and another begins. This decision, seemingly technical, is deeply interpretive, since different segmentations yield different structural accounts.
Systematic reflection on musical structure is ancient—medieval theorists codified modal systems, and Renaissance and Baroque writers described counterpoint and harmony. But music analysis as a distinct, self-conscious practice emerged in the nineteenth century, alongside the rise of the concert repertoire and the canonization of past masters. Early analysts sought to explain how the great works of Beethoven, Bach, and others achieved their coherence and expressive power. Their methods were often descriptive: identifying themes, charting key relationships, and outlining forms such as sonata form.
A decisive development came in the early twentieth century with the work of Heinrich Schenker, whose approach transformed analysis from description into a rigorous explanatory system. Schenker argued that the surface of a tonal composition—its actual notes—is a prolongation, or elaboration, of a deep underlying structure. He proposed that all well-formed tonal music reduces to a fundamental melodic descent over a harmonic bass, and that the analyst's task is to show how the surface arises from this background through successive levels of diminution. Schenker's graphs, with their characteristic slurs and stemmed notes, became a distinctive visual language for showing hierarchical relationships among pitches.
Schenker's system was revolutionary but also controversial. It was developed almost exclusively on the German-Austrian canon, particularly Bach, Beethoven, and Brahms, and it assumed that tonal music is a unified, organic whole. Music that does not fit this model—much French, Russian, or Italian music, let alone non-Western traditions—was either ignored or judged deficient. Moreover, Schenker's theoretical claims were entangled with nationalist and organicist ideologies that later scholars have had to disentangle from his analytical insights. Despite these problems, Schenkerian analysis remains influential, particularly in North American universities, where it became the dominant method for teaching tonal analysis in the mid-twentieth century.
The mid-twentieth century saw the proliferation of competing analytical systems, each addressing perceived limitations in the others. Set theory, developed by Milton Babbitt and refined by Allen Forte, provided a method for analyzing atonal and twelve-tone music that lacked the tonal hierarchies Schenker took for granted. Set theory treats pitch classes as unordered collections and classifies them by their interval content, allowing the analyst to identify motivic and harmonic relationships in music that does not conform to traditional tonality. Its strength is its generality and precision; its weakness is that it can produce analyses that feel arbitrary, since the relationships it identifies are not always perceptually salient.
Neo-Riemannian theory, developed in the 1980s and 1990s by David Lewin and others, addressed a different gap: chromatic music that is neither strictly tonal nor fully atonal, such as the music of Wagner, Liszt, and early Schoenberg. Neo-Riemannian analysis models transformations between triads—for example, changing a major chord to a minor chord by lowering one note—and studies the group structure of these transformations. This approach captures the fluid, non-hierarchical quality of chromatic harmony better than Schenkerian analysis, which struggles with music that does not prolong a single tonic.
Lewin's broader contribution was the concept of the "transformational attitude," which shifts the analyst's question from "what is this music?" to "what does this music do?" Transformational analysis focuses on the operations that connect musical events, rather than the events themselves. This perspective has been influential beyond pitch analysis, extending to rhythm, timbre, and form.
Rhythmic and metric analysis developed along a separate track. The work of Maury Yeston, Fred Lerdahl, and Ray Jackendoff, and later Christopher Hasty, established rhythm as a domain with its own hierarchical structures, not merely a secondary parameter. Lerdahl and Jackendoff's A Generative Theory of Tonal Music (1983) attempted to unify pitch and rhythm under a single generative framework, borrowing concepts from linguistics. Hasty's Meter as Rhythm (1997) challenged the idea that meter is a fixed grid, arguing instead that it emerges from the ongoing process of musical becoming.
A parallel tradition has focused not on pitch or rhythm in isolation but on the large-scale organization of musical time. Formenlehre, the study of musical form, has roots in the nineteenth-century conservatory tradition, where students learned to recognize sonata, rondo, and theme-and-variations forms. In the late twentieth century, scholars such as James Hepokoski and Warren Darcy revitalized this tradition with "Sonata Theory," which treats sonata form not as a fixed mold but as a set of default expectations against which individual works make choices. Their concept of "deformation"—a work's deliberate deviation from generic norms—has become a standard analytical tool.
The question of musical meaning has generated its own approaches. Some analysts, following the hermeneutic tradition, read music as expressing or representing extra-musical content: emotions, narratives, philosophical ideas. This approach was marginalized during the mid-twentieth century, when formalism dominated, but returned forcefully with the "new musicology" of the 1980s and 1990s. Scholars such as Susan McClary and Lawrence Kramer argued that analysis cannot be value-neutral, and that musical structure is entangled with social, political, and gendered meanings. This turn did not replace structural analysis but insisted that structure and meaning be studied together.
A related development is the analysis of music as performance. Rather than treating the score as the work and the performance as its imperfect realization, performance analysts study recordings and live performances as texts in their own right. This approach examines how tempo, dynamics, and articulation shape musical meaning, and it has been aided by computational tools that can measure these parameters with precision.
The late twentieth and early twenty-first centuries brought a new kind of analysis grounded in empirical methods. Music cognition researchers study how listeners perceive and remember musical structures, using experiments, behavioral measures, and neuroimaging. This work has sometimes confirmed and sometimes challenged the assumptions of score-based analysis. For example, listeners do not always hear the large-scale structures that Schenkerian analysis identifies, raising questions about whether those structures are real or merely notational artifacts.
Computational music analysis uses algorithms to analyze large corpora of scores or audio. Corpus studies can reveal statistical regularities—common chord progressions, typical phrase lengths, stylistic fingerprints—that are invisible to the analyst working on a single piece. Machine learning techniques can classify music by genre or composer, and can generate analyses that are consistent across thousands of works. These methods have been criticized for reducing music to features that are easy to compute, but they have also opened new questions about style, influence, and historical change.
The relationship between computational and traditional analysis is not always comfortable. Computational methods excel at pattern discovery but struggle with interpretation; traditional analysis excels at interpretation but is difficult to scale or verify. Some scholars advocate for a "computational hermeneutics" that combines both, using algorithms to generate hypotheses that human analysts then interpret.
Contemporary music analysis is pluralistic. No single method dominates, and most practitioners are conversant in several. The field has also expanded its repertoire beyond the Western classical canon. Popular music analysis, jazz analysis, and the analysis of world musics have developed their own methods and questions, often borrowing from ethnomusicology. The analysis of electronic and experimental music has required new concepts for timbre, texture, and spatialization.
This pluralism reflects a broader shift in the discipline's self-understanding. Analysis is no longer seen as the neutral discovery of structures that are "in" the music, waiting to be found. It is understood as an interpretive practice, shaped by the analyst's choices, assumptions, and cultural position. This does not mean that anything goes—analyses must still be accountable to the musical evidence—but it does mean that the field acknowledges multiple legitimate ways of hearing and explaining the same music.
The enduring value of music analysis lies in its capacity to deepen attention. A good analysis does not merely label what is already audible; it reveals relationships that transform the listener's experience, making the music seem more coherent, more surprising, or more meaningful than it did before. This is true whether the analysis is Schenkerian, set-theoretic, narratological, or computational. The methods differ, but the goal is shared: to understand how music works, and in understanding, to hear it better.