I. General 3. The role of method in musical analysis

3. The role of method in musical analysis.
Many of the classifications that have been formulated for musical analysis have distinguished between types of analytic practice according to the methods used, which can then be grouped together into broader categories. For example, there is the widely accepted division into ‘stylistic analysis’ and ‘analysis of the individual work’ which was described above as pragmatic but theoretically unnecessary. There is the threefold classification into ‘constructional analysis’, ‘psychological analysis’ and ‘analysis of expression’ put forward by Erpf in MGG1 (1949–51). This classification does not correspond exactly with, but is roughly equivalent to, Meyer’s distinction (1967, pp.42ff) between ‘formal’, ‘kinetic-syntactic’ and ‘referential’ views of musical signification. Dahlhaus (RiemannL12, 1967) made a fourfold distinction: ‘formal analysis’, which explains the structure of a work ‘in terms of functions and relationships between sections and elements’; ‘“energetic” interpretation’, which deals in phases of movement or tension spans; and Gestalt analysis, which treats works as wholes; these three make up among them the field of analysis proper, which he distinguished from his fourth category, ‘hermeneutics’, the interpretation of music in terms of emotional states or external meanings. The first, second and fourth of these correspond broadly with the three categories of Erpf and Meyer, while the third deals with analyses based on the idea of organism.

The principal difficulty with these classifications is that their categories are not mutually exclusive. Thus, for example, Riemann is generally cited as the prime example of a formal and constructional analyst, and yet his work rests on a fundamental idea of ‘life force’ (Lebenskraft, lebendige Kraft, energisches Anstreben) that flows through music in phases and is actualized in phrase contours, dynamic gradings, fluctuations of tempo and agogic stress. This idea is closer to the kinetic view of music; it suggests that Riemann’s work belongs to two of Meyer’s three categories.

A different way of identifying analytical methods is partly historical in nature. For example, Schenkerian analysis, so called, has its origins in the work of Heinrich Schenker (1868–1935); but, as a label that identifies a type of analytical practice today, ‘Schenkerian analysis’ includes a number of developments that have accrued since Schenker’s death and are due to his pupils, his pupils’ pupils and others (see §II, 5–6). This ‘method’ is circumscribable because, seen as a tradition communicated orally by teaching and also through a modest number of written sources, it remains reasonably concise. The status of motivic analysis as an identifiable complex of methods is less easy to describe: naturally, it has a history (see §II, 3–5), but the accumulation of developments has been such that some filtering out has also taken place; significantly different synoptic descriptions of motivic analysis as practised today are thus possible. In the case of harmonic analysis, the range of meanings is such that no contemporary synoptic description of it as a ‘method’ can properly be offered: only its history retains a degree of integrity.

A good example of the emergence of a method by accumulation and selective filtering is seen in the analysis of form, which Cook (Guide, 1987) explicitly included in a category of ‘traditional’ methods. Broadly speaking, one may describe formal analysis historically, identifying principles and refinements as they were newly introduced; and one may present an overview of what is meant by formal analysis today. Neither of these approaches alone, however, can fully reflect the fact that formal analysis has a two-dimensional history of changing accumulations: that the difference between formal analysis around 2000 and formal analysis around 1900, for example, cannot be measured solely in terms of the new ideas that have been added in the intervening century. But it can be traced in the differences between attempts by responsible authors to provide synoptic definitions at various times and places.

The history of formal analysis tells us that during the late 18th century and the 19th, music theorists defined certain structural patterns – not genres or species such as concerto or minuet, but more widely applicable processes of formal construction common to many genres and species – that were reducible to two fundamental patterns: AB and ABA. These were subsumed in German terminology under the single term Liedform (first proposed by A.B. Marx, 1837–47) in its ‘two-part’ (zweiteiliges) and ‘three-part’ (dreiteiliges) form, and distinguished in English terminology as Binary form and Ternary form. Broadly speaking, these terms referred to small-scale forms; they applied most directly to instrumental dance movements of the 17th and 18th centuries, and relied on the concept of regular phrase structure with the eight-bar period as the principal unit of construction. Later in the history of formal analysis, large-scale formal models came to be regarded as extensions to one or other of the two fundamental patterns: thus Sonata form was the extension of the binary pattern, and Rondo of the ternary.

An overview of what is meant by formal analysis might begin with the three basic form-building processes proposed in §2 above: ‘recurrence’, ‘contrast’ and ‘variation’, expressible as AA, AB and AA′. It might further identify a distinction between two basic processes of extension: that of a succession of formal units, and that of development. The former (in German, Reihungsform or plastische Form) relies on proportion and symmetry, and is architectural in nature; the latter (Entwicklungsform or logische Form) relies on continuity and growth. The rondo, ABACADA, extends ternary form by succession; sonata form extends binary form by development. And the two processes are both brought into operation in the so-called sonata rondo: ABACAB′A. There is a further process by which larger forms may be created out of one of the two basic patterns: by the operation of one or both patterns at more than one level of structure (Potenzierung, ‘exponentiating’). By this means, such structures as A (aba)B (cdc)A (aba) are produced. Related to this is the concept of Cyclic form, whereby movements in recognizable forms are grouped together to form larger units such as the suite and the sonata.

Many manuals of form have separate descriptions of ‘the contrapuntal forms’ and allow a category of ‘free forms’. Nonetheless, the underlying idea of formal analysis is that of the ‘model’, against which all compositions are set and compared and measured in terms of their conformity to or ‘deviation’ from the norm. But if formal analysis may be distinguished from other kinds of analysis by its concern with the recognition of these processes and the description of works in terms of them, manuals of formal analysis vary in the ways in which they see the totality of musical formations, from the Middle Ages onwards and for all vocal and instrumental media, as governed by these fundamental patterns. Quite apart from the universality of the basic models, there are many difficulties in determining criteria for their recognition. For some analysts, identity or non-identity is determined by thematic character; for others, by key scheme; for others, by length of units. Thus, for Dahlhaus (RiemannL12, 1967), the prime conditions of the two-part Liedform |:A:||:B:| are, first, that the first part ends on a half-close in the tonic or a full-close in a related key, and, second, that the parts are melodically different (or related |:AX:||:AY:| or |:AX:||:BX:|). For Scholes (Oxford Companion to Music, ‘Form’) binary form rests on the same key scheme |:tonic–dominant (or relative major):||:dominant (or relative major)–tonic:|, and the absence of ‘strong contrast’ in thematic material. For Prout (1893–7, 1895), key scheme is not really a determinant at all for binary form, for he allowed |:tonic–tonic:||:remote key–tonic:|; nor is thematic relationship, for he allowed AA′BA″ as well as ABCB. The basic determinant for Prout was that the form should constitute ‘two complete sentences’. Thus the form |:A:||BA:|, which for Dahlhaus was three-part Liedform, was for Prout binary form unless the first part is itself a complete binary form, self-contained and rounded.

The question might be asked whether analysis as a whole can be described by listing and describing its methods – using this word in the sense explored above. Handbooks of analysis written largely for pedagogical purposes (e.g. Cook, 1987; Dunsby and Whittall, 1988) have adopted this approach virtually out of necessity. Such texts appeared at a time when analysis had emerged for the first time in the English-speaking world as a complex academic discipline it its own right, rather than as an adjunct, however valuable or essential, to other forms of musical activity or training. This moment in the history of analysis was brief, however – perhaps inevitably so, as the high profile of analysis encouraged the questioning of its assumptions and practices (see §II, 6). Arguably, those related disciplines from which this questioning emerged – notably criticism – have in the aftermath taken on many of the lasting priorities and occupations of analysis, themselves becoming significantly changed in the process. Conversely, analysis remains strong, but has revitalized its concerns through closer contact with disciplines that always left more room for debate about the nature and function of music than analysis had come to do.

It follows from all this that a thorough-going typology of musical analysis, widely applicable across times and places, would probably have to encompass several axes of classification. The analyst’s view of the nature and function of music would certainly be one of these. But his approach to the actual substance of music would be a second; his method of operating on the music would be a third; and the medium for presentation of his findings would be a fourth. Other axes might be concerned with, for example, the purpose for which the analysis was carried out, the context in which it was presented, the type of recipient for which it was designed.

Under approaches to the substance of music would be categories such as that a piece of music is (a) a ‘structure’, a closed network of relationships, more than the sum of its parts; (b) a concatenation of structural units; (c) a field of data in which patterns may be sought; (d) a linear process; and (e) a string of symbols or emotional values. These five categories embrace the approaches of formal analysts such as Leichtentritt and Tovey, structuralists and semiologists, Schenker, Kurth and Westphal, Riemann, hermeneutics, stylistic analysis and computational analysis, information theory analysis, proportion theory, Réti and functional analysis, and much else. The categories are still not exclusive. For example, (a) and (c) are not wholly incompatible in that approach (c) may lead to approach (a). Then again, two approaches may co-exist at two different levels of construction: perhaps (a) or (b) for large-scale form and (d) for small-scale thematic development.

Under methods of operating would be categories such as (a) reduction technique; (b) comparison, and recognition of identity, similarity, or common property; (c) segmentation into structural units; (d) search for rules of syntax; (e) counting of features; and (f) reading-off and interpretation of expressive elements, imagery, symbolism.

Under media of presentation would be categories such as (a) annotated score or reduction or continuity line (see fig.12); (b) ‘exploded’ score, bringing related elements together (fig.28); (c) list, or ‘lexicon’ of musical units, probably accompanied by some kind of ‘syntax’ describing their deployment (see figs.23); (d) reduction graph, showing up hidden structural relationships (figs.17–20); (e) verbal description, using strict formal terminology, imaginative poetic metaphor, suggested programme or symbolic interpretation; (f) formulaic restatement of structure in terms of letter- and number-symbols; (g) graphic display: contour shapes (fig.22), diagrams (fig.16), graphs (fig.15), visual symbols for specific musical elements (fig.14); (h) statistical tables or graphs; and (i) sounding score, on tape or disc, or for live performance. Such media can be used together within an analysis, and elements of two or more can be combined.