Genres of and Approaches to Classification

Scholars engaged in classification have devoted primary attention to the products of folk music, relegating processes to a position of secondary importance. Most systems have therefore arranged and ordered the individual pieces constituting a large, presumably coherent repertory. Because the earliest classifiers in the nineteenth and early twentieth centuries worked from collections and were themselves often not collectors, they had little evidence at their disposal concerning processes of change or social function. Through classification they hoped to achieve two ends. First, they hoped to devise a way of gaining easy access to and identifying the individual pieces in large collections. Second, they sought ways to define the interrelation of items in large collections, usually motivated by historical and nationalistic considerations. The nature of these goals necessitated an approach that concentrated initially on specific items and then expanded to broader levels of comparability. Accordingly, the early classifiers concerned themselves more with the establishment of repertory boundaries than with analytical approaches that would allow for exploration beyond such boundaries.

The first attempts to classify relied on song texts; melody received relatively little attention until the twentieth century (Herzog 1937:49, Mosely 1964:9, and Sorce Keller 1984:100). As a criterion of classification, song texts permitted a fair degree of specific identification: dialect could locate the text geographically and subject matter presumably established function. Many of the early classifiers were linguists whose primary concerns lay with the explication of textual matters. Many early collections of English folk song reflected this textual impetus. The English and Scottish Popular Ballads by Francis James Child (1882-98) contained texts only. Not only were melodies lacking, but the classification system that later became formalized as the “Child ballads” was arrived at completely on the basis of textual subject matter. So entrenched was this early method of classification according to text that subsequent attempts to classify Anglo-American ballads have simply accepted it, in essence working around it when ascribing tune types to the narrative structure of the ballads (cf. Seeger 1966 and Kolinski 1968 and 1969). German anthologies employed a somewhat different approach to classification by text, incorporating aspects of social function together with more standard literary typologies. Thus, the genres identified in the three-volume Deutscher Liederhort (Erk 1893-94) included such categories as “songs of departure and wandering” and “historical-political songs” as well as “ballads” and the “songs of legends.” This mixture of literary, functional, and contemporary genres has persisted in German classification systems and is basic to the broad conceptualization of genre that characterizes German folk music scholarship today (cf. Brednich, et al. 1973 and 1975 and Bausinger 1980:263-94).

Melodic classification began in earnest somewhat later than textual approaches for both practical and intellectual reasons. Even in collections that included musical aspects, notation was indefinite or was modified by antholo-gizers to suit practical ends, as in arrangements in harmonized versions for choral performance. Many collectors, moreover, possessed only a modicum of musical skills, which prevented them from transcribing musical aspects during fieldwork—a difficult enough task for anyone before the advent of sound-recording devices, when transcriptions often required repeated performances or reliance on the collector’s memory.

As collections grew larger, the need for musical classification became more pressing. The early architects of classification responded to two primary motivations. The first resulted from the need for easy access to large collections. The second sought to interpret the ways in which the melodies of large collections were related to each other. The first comprehensive attempt to incorporate these two methods appeared in articles by Oswald Koller (1902-3) and Ilmari Krohn (1902-3; cf. also Nettl 1983:120). Koller proposed a means of locating a tune among many others when—and only when—one knows the exact pitch complement. Krohn devised a means of abstracting tunes so that many tunes might be compared to discover patterns of relationship. Koller’s approach was primarily lexicographic: one could easily find the tune in a melodic dictionary if one had an idea of the correct spelling; the relatedness of tunes was a question of objective components. The tunes examined by Krohn, in contrast, exhibited a relatedness based on structural similarities, which suggested similar patterns of development and change. Such patterns were further interpreted in the light of genetic relatedness, and scholars soon labeled them with the appropriate term tune families. Krohn himself and his many successors refined and modified the genetic-historic approach to classification, and its underlying principles remain a part of many contemporary classificatory methodologies.

A third realm of classification has also been important to folk music study during the past century. This one has as its subject musical instruments. Unlike other areas of folk music classification, studies of instruments have often emphasized the comparative and universal more than the specific and the regional. Not surprisingly, the resulting methods reflect the disciplinary concerns of anthropology and systematic musicology (cf. Sachs 1929 and Schaeffner 1936). Even though evidence from specific cultures illustrates many organological theories, it also serves to justify claims for the universality of many instruments. Organologists have often aimed to conquer rather expansive theoretical turf, such as the evolution of instrument types throughout history (Sachs 1940). In general, approaches to the classification of instruments have been more lexicographic, although family-like groupings result also from the underlying evolutionary impetus.

The most widely accepted theory of instrumental ordering, published by Curt Sachs and Erich von Hornbostel in 1914, established four large groups of instruments on the basis of construction material and sounding mechanism: idiophones are “self-sounding” instruments (e.g., bells or xylophones); sound emanates from a stretched skin or other membrane in membranophones; strings are plucked, bowed, or struck to produce the sound from chordophones; and setting a column of air in vibration distinguishes aerophones (Hornbostel and Sachs 1961). The Sachs-Hornbostel method has encountered some criticism because of its imposition of hierarchical order (Lysloff and Matson 1985) and because nonacoustical (largely electronic) instruments do not lend themselves to classification in this manner, but the method remains the standard reference for most discussions of folk music instruments (e.g., the Diagram Group 1978).

While classification systems have focused primarily on the products of folk music, they have not completely ignored the processes, regarding them as links among the various products. Processes appear under three general categories: tune histories, tune families, and melodic change. Tune histories constitute the broadest category, but specific theories to describe the history of melodies are relatively undeveloped. The distinguishing feature of tune history as a category is what one might call the objective behavior of musical materials. Acceptance of such behavior is evident in studies of tunes that are diffused over vast cultural areas (e.g., Wiora 1953), sometimes with the implication that melodies or parts of tunes travel about more or less randomly and on their own volition (Tappert 1890). Such objectivity is relatively absent and unimportant in concepts of the tune family, and instead one finds explanations of how tune evolution takes place and why some versions appear and others do not. The genetic basis of the metaphorical family predetermines the extent and shape of representative melodic repertories. Melodic change, often a greater concern to ethnomusi-cologists studying folk music, may be less the product of an internal genetic cohesion than of contact with external musical and social forces. In general, however, many types of externally instigated change have not been primary interests in the development of classification systems. Some classifiers see radical change of this sort as a component of urban and modern influences, and therefore not related to the formation of folk music repertories; others regard external influences as disruptive of the order necessary for classification and therefore tantamount to a disintegration of the discursive boundaries of folk music (cf. Bartók 1931:53-80).

Classification: Concepts and Vocabulary

The initial step in most attempts to devise a classification system is determining which elements lend themselves to meaningful comparability and therefore identify the ways in which relatedness exists in the collections or repertories under examination. These elements range from very small units, some form of melodic morpheme, to large and abstract units, perhaps some type of tune gestalt (Suppan 1973 and Kuckertz 1963). The elements may be intrinsic to the structure of each piece or extrinsic to the piece and broadly characteristic of the larger repertory; an example of this sort of extrinsic element is mode or social function. Even these extrinsic factors tend to be measured by their influences on smaller elements, as in the complement and weighting of pitches in a particular piece that establish a relation with the modal possibilities of an entire repertory.

Smaller elements can be ordered in very precise ways, and for this reason lexicographic systems often concentrate initially on the smallest elements. The opening and final pitches of a piece, the initial interval, the relation between ambitus (the distance separating the highest and lowest notes) and central pitch all can be quantified, numerically or otherwise. Quantification allows one to establish an order, to spell out each melody and place it in a melodic lexicon near those melodies that have a similar spelling. The greater the number of small units that one quantifies, the more extensively the lexicographic system determines a place specific to individual melodies.

The next level often assumed by classification comprises the grammar and syntax of melody. This level moves from the strictly quantifiable and objective to the empirical and subjective. Thus, one not only identifies the elemental units but also formulates the basic patterns in which these units regularly occur. It is theoretically possible to interpret the patterns formed at this level for lexicographic purposes, but most classifiers concerned with aspects of grammatical and syntactical patterns also investigate them to determine the bases for tune families or histories. Genre is first evident at this level, in some cases in only general characteristics but in others through fairly specific structures and processes of change. Whereas ballad and epic require more extensive narrative structures, lyric folk songs and the blues may allow essential flexibility at this level. In studies of lyric folk song, Judith McCulloh has suggested that examining the function of the “smallest rearrangeable unit of text” provides the key to establishing the level at which understanding and exploration of the genre can take place (1970:7). These units appear throughout repertories of lyric song and in essence define the repertories by the particular clusters that they form (McCulloh 1983:42).

The establishment of criteria for using melodic contour for classification has been one of the more persistent ways of de-emphasizing the role of smaller elements while bringing light to bear on the overall form of melody. The earliest advocate of classifying primarily on the basis of contour was Sirvart Poladian, who aimed to find some means of extending methods that worked well for specific repertories to a larger, more comparative framework. The salient features of contour were those that had widespread distribution in a large number of very different repertories—for example, descent to the final note, which Poladian held to be “one musical feature [that] seems common to most folk songs” (1942:207). Perhaps most important to Poladian was his belief that contour was one of the few melodic factors that song variants did not alter, thus making it an extremely stable representation of any repertory or style (ibid. :210-11). More recently, George List has made similar claims for the salience of melodic aspects abstracted by contour, again noting that melodic contour may provide one of the most effective keys for determining the existence and limits of universality in folk music (List 1978 and 1985; see also Mosely 1964:11-12).

Few areas of classification have weathered the storms of criticism as indomitably as modes. The application of modal theories often takes the form of esoteric studies directed to a small fraternity of cognoscenti. That may be the stimulus of or response to the polemical debate that modal theories sometimes spur, but rarely do modal theories seem designed for broader appeal or the methodological corollary that lexicographic ordering permits. In several cases, debates over the role of modes in folk music take on a life of their own, abandoning empirical studies of standard collections and the new evidence that fieldwork would impart (cf. Cazden 1971 and Bronson 1972). In Anglo-American ballad scholarship, polemic was a part of modal research from virtually the moment it appeared. In the first edition of his English Folk Song: Some Conclusions in 1907, Cecil Sharp included a chapter on modes (pages 36-53), in which he claimed that the melodic material of English folk song was essentially related to the diatonic modes of the medieval Latin church. His aim was twofold. On one hand, this connection with an earlier stratum of European musical life overlaid folk song with the patina of age, concomitantly distancing it from popular, urban music employing major and minor modality. On the other hand, Sharp hoped that folk song would be a new font for the English composer, hardly an unusual desire in an age of nationalism but one that would benefit markedly from a demonstration of the theoretically sound structure of folk music. Sharp’s rather prescribed goals notwithstanding, many successors took up the modal banner, insisting that Anglo-American folk songs were rooted in intricate and occasionally fantastic designs for the interrelation of the church modes. Bertrand H. Bronson conceived of this pattern as a seven-pointed star, with each mode forming one point and mixed modes located elsewhere about the star (1969:85).

The persistence of modal debates has done little to clarify problems of classification and has instead occluded the potential for establishing more accessible criteria for relatedness. That this persistence continues to petrify some of the most sacred of canonic formulae, including those encasing the Child ballads, cannot be denied. But mode has not always worked to the detriment of classification. Its role in many non-Western musical systems is undeniable, and we base our systematic understanding of such musical cultures as those of Iran, the Arabic Middle East, and the Indian subcontinent on elaborate modal frameworks. Several European scholars were more judicious in their claims for modes; Bartók, for example, used mode as one criterion for the age of a folk song, thus employing mode to corroborate, rather than set, the discursive boundaries of Hungarian folk music.

For a system of classification to offer significant insight into the organization of folk music repertories and style it must ideally combine the various levels and foci that this section has surveyed. One of the most masterful syntheses of diverse approaches is that which Béla Bartók, in collaboration with Zoltán Kodály, applied to Eastern European folk music, with especially valuable results in the examination of Hungarian collections (Bartók 1931, Kodály 1960, Erdely 1965:43-73, and Herzog 1937:54). Bartók combined lexicographic methods developed by Finnish scholars with classificatory criteria that would identify both genetic and historical patterns. He began by simply counting the number of lines in a melody. Next, the intervallic distance of the final note of each line from the finalis of the song was measured. The number of syllables in each line was then represented, and finally the ambitus of the song was stated as two numerals, one the lowest note of the melody and the other the highest. A string of numbers, then, appeared as an abstract lexicographic representation of the melody. Bartók’s method may be illustrated with the folk song “Mikor a nagy erdön kimész” (example 3). The figures Bartók used to designate this song are , 8, 1-8. These figures mean that the song is a four-line tune whose first and second lines end on and whose third line ends on . All lines consist of eight syllables, the lowest note being , and the highest (Bartók 1931:7).

Mikor a nagy erdön kimész

Translation

EXAMPLE 3
(cf. Bartók 1931:7, 107, and Section Al, P. 6)

Bartók’s classification system offers other descriptive insights, most importantly those linking style to age (cf. Kodály 1960:23-68). When a piece has a restricted ambitus or lines with a small number of syllables, it comes from an older stratum of the repertory (Bartók 1931:8-9). Rhythmic and metric considerations are also significant determinants of age and style. The Old Style Hungarian melodies had a free, speech-like rhythm that Bartók calls parlando rubato. In contrast, a tempo giusto style, characterized by a more regularized, dance-derived rhythm, predominates in the New Style (ibid.:9). Bartók’s method is remarkably thorough, and it elucidates many intricacies of style and structure in Hungarian folk music. It is less successful in its explanation of non-Hungarian repertories, where, for example, syllable count in each line plays a less significant role than in Hungarian folk music. The system founders when applied to non-Western music. The criteria of his system required that Bartók claim Arabic folk music to be at an incipient evolutionary stage because of the prevalence of short lines and single melodic motifs (Bartók 1920:489), whereas variable line lengths and the appearance of simple and complex melodic motifs together are widespread in all genres of Arabic music. Bartók’s approach to classification was, therefore, most effective when applied to the repertories that he knew best, whose idiosyncracies he understood most intimately; that the approach did not succeed in becoming universal is hardly exceptional in the history of folk music classification.

A wide range of regional and ethnic criteria has produced many approaches to classification. The boundaries that such criteria demarcate are, of course, geographic and social, and therefore one recognizes the considerable potential for determining some aspects of universal ordering. A frequent qualification of indigenous classification is the association of certain pieces with the place in which one lives and the understanding that certain other pieces come from elsewhere. The venue for the pieces from elsewhere may vary; it may be another region or another social setting—an urban center, for example, in contradistinction to a rural area. Folk taxonomies in northeastern Iran and Afghanistan almost always stress this contrast between rural and urban, even when the real sources of the folk music repertories are more complex and mixed (cf. Blum 1974:86 and Sakata 1983:53-63). Folk musicians who perform throughout an ethnically pluralistic area often rely on a complex indigenous classification system to indicate which repertories are appropriate for the audiences in different parts of the area (cf. Blum 1972 and Bohlman 1980). Regional and ethnic criteria often classify repertories in a fairly loose fashion. The criteria with which Vance Randolph ordered his large collections of Ozark folk songs do not transfer as a cohesive group to other repertories, for despite his use of some general categories (e.g., “British ballads and songs,” in Randolph 1980, vol. 1) he also uses more specific regional categories (e.g., “Songs of the South and West,” in ibid., vol. 2). The larger classification system itself only applies to the Ozarks, whereas the individual parts might describe other areas of the United States.

The most ambitious attempt to forge a classification system by combining regional and ethnic criteria with social and functional aspects of music has been Alan Lomax’s cantometrics project (Lomax 1968 and 1976). Cantometrics posits that social structure is the model for musical performance and sound and that repertories form on the basis of unified expression of social structure. In a sense, this idea is an extension of the culture-area (Kulturkreis) concept so important to comparative musicology. But in this concept, history and diffusion of cultural traits determine large areas, whereas in cantometrics cultural boundaries form from within the culture—they are buttressed by the uses and functions of culture. Most important to the classificatory potential of cantometrics is the assertion by its advocates that the relation between social structure and musical expression is quantifiable: thirty-seven criteria of musical production can be measured easily and compared accurately using computers (figure 2). In addition, a detailed map reveals classification according to six world regions, fifty-six culture areas, and 233 specific cultures. The difficulties of precise comparability notwithstanding, cantometrics is not unlike many other classification systems in that it begins by attempting to identify the appropriate levels of comparison and then examines them in diverse settings to establish criteria that universalize the structure of music’s discursive boundaries. And, like other systems with an agenda of universality, cantometrics has encountered the persistence of differentiation within those boundaries, thus limiting comparability to the specific and requiring modification of the classificatory criteria when cantometrics approaches the universal.

The Dialectical Concerns of Classification

Accessibility to the general and differentiation of the specific: ideally, all classification systems must achieve both. The classifier must discern how each piece is like the others in a given repertory while maintaining the integrity of the piece by recognizing its differences. Classification is, at its best, a dynamic process that allows for the changeability of folk music and its repertories and for influences from diverse sources. On the surface, classification may seem to seek ways of closing a repertory, but underneath it must be flexible enough to respond to the dynamic forces that constantly reshape folk music aesthetically and functionally. Classification must confront and explain both the products and the processes of folk music in a dialectical way if it is to be as broadly effective and meaningful as possible.

FIGURE 2. Cantometrics profile for Europe. The solid line traces the course of the main type of European song—solo, unaccompanied performance of foursquare strophes in simple meter. The dashed line shows a second type—simple group organization in polyphony with good tonal blend. This profile most strongly resembles Old High Culture. (Source: Lomax 1968:100. Reprinted by permission of the author.)

Description and prescription. Descriptive approaches seek to define repertories in relation to elements that are stable and structurally fundamental over time. Description is inductive in the sense that it observes the specific first and then devises the appropriate classification system on that basis. Theoretically, the potential for true universality is greater with descriptive approaches, but practically, the census of characteristics consistently observed over sufficiently long periods can never be quite complete enough. Prescription, in contrast, allows the making of choices on the basis of limited examples that permit the extrapolation of an idealized repertory and the classification thereof. The classifier may prescribe those levels that he or she believes salient to comparability and then define the canon accordingly. The expansion of field collections may or may not justify the canon, but this may not be important to the relevance of classification, for prescription accepts that certain patterns of relatedness are ideals, or abstractions of a census too unwieldy to represent in all its details. The combination of description and prescription depends on the repertory or collection that the scholar aims to classify and on the role fieldwork will play in continuing to expand the collection; it also depends on the relative needs for lexicographic access or stylistic analysis.

Emic and etic. By referring to emic and etic approaches as dialectics I do not mean to pair folk with outsider classification. Rather, I wish to suggest two philosophical goals that underlie many approaches to classification. One extreme argues that true classification of a repertory lies in the collective unconscious of the community sharing that repertory; the other extreme believes that there are objective rules of structure and form over which the folk exert no control. Those approaches that attempt to reflect emic classification demonstrate great variety. Some aim to elucidate an extensive epistemological understanding of the way music is expressive in different cultures; Steven Feld’s studies of the Kaluli of Papua New Guinea, for example, have revealed a complex representation of the culture’s avian-derived mythological concepts in its attitudes toward and performance of music (1982). Other emic approaches rely on a deeper historical understanding of a society’s classification of folk music; Stephen Blum, for example, has shown that the attitudes toward folk music in northeastern Iran (Khorasan) are related in complex ways to the long interaction of oral and written genres of music and poetry (1974). To some extent, all classification systems reflect the etic motivations of their designers. When etic motivations bear little or no resemblance to emic concepts, a classification system may result that has little historical validity. The efforts of Cecil Sharp, Olive Dame Campbell, and others to impose on many Appalachian communities folk music repertories replete with Child ballads and morris dances is one of the most glaring examples of the disjunction between emic and etic concepts of classification (cf. Whisnant 1983, esp. 105-79).

Aesthetic-formal and cultural-functional. Aesthetic and formal criteria have been fundamental to most genetic-historical classificatory schemes. Form and structure provide the basic vocabulary for lexicographic ordering and for the relatedness of pieces and genres one to another. In some cases, such approaches only prescribe the aesthetic realm of a single repertory; they may also elucidate the structural elements that make the formation of repertories possible, thereby extending their application to a more general level. Reliance on the role of social function in shaping repertories often produces a somewhat different way of ordering. In extreme cases, functional criteria completely replace formal relatedness. Many of the most effective approaches, including the approach developed by Bartók and Kodály, combine aesthetic and formal considerations with the cultural and functional, thus explaining a vast array of differentiated elements with a single classification system.

Similarity and difference. The measurement of similarity and difference is central to many approaches to classification (Nettl 1983:118-19). Does one classify on the basis of similarity? Or are differences more relevant to the boundaries that a particular repertory exhibits? These questions raise even larger ones. Is a classification system determined by internal coherence? Or, rather, does order accrue as a result of contrast with external genres and repertories? Clearly, the answers to these questions differ from repertory to repertory. In some cases, similarity may be so widespread that it loses all meaning as a criterion for classification, and the classifier must turn to the ways in which differences appear. Extreme difference, in contrast, might belie the true sense of coherence that determines a repertory, perhaps because of even greater difference when compared to other repertories. Choices to emphasize the role of similarity or difference often reflect the balance between descriptive and prescriptive approaches that a classification system illustrates.

Diachronic and synchronic. When musical change serves as a measurement of the differentiation that folk music repertories exhibit, classification necessarily includes many diachronic considerations. The historical component of genetic-historical approaches is essentially diachronic. In the methods of Bartók and Kodály, diachronic evidence and results predominate; Nettl, too, has argued persuasively for the historical motive that pervades many approaches to understanding folk music (e.g., 1973:10-11). When stylistic and structural reasons seem to underlie the variation that appears in folk music repertories, synchronic analysis may prove more useful. We may find, for example, that there is a wide and relatively sparse distribution of tune or text types, regardless of the general patterns of diachronic change. Modal classification, too, relies relatively little on diachronic change, though the work of Harold Powers is an important exception (e.g., 1970). The more internal melodic elements are denominated as abstract entities, the more synchronic differences play a role in classification.

The specific and the universal. This dialectic is one of the most problematic in all classification endeavors. The classifier faces the paradox of refining the system for one repertory so precisely that it no longer has general applications or of accounting for the general characteristics of many repertories so broadly that no single repertory really fits very well. This paradox is seldom lost upon scholars who devote themselves to classification, for most demonstrate an awareness of the need to arrive at a synthesis of the two extremes (e.g., Poladian 1942 and Väisänen 1949). The most successful syntheses occur at a more moderate level, when one generally acceptable system is adapted to several other repertories. Bartók, for example, refined the Finnish genetic-historical method so that it would classify Hungarian folk music and other regional repertories of Eastern Europe. In most cases, his refinements offered new insight into the structure and history of Eastern European repertories. The same methods applied to such non-Western repertories as Turkish and Arabic rural musical styles produced spurious results. In short, Bartók significantly expanded the specificity of the Finnish method but was unable to render the method universal. The classificatory dialectic of particular versus universal underscores a fundamental question in folk music scholarship: is folk music really a universal genre of expressive behavior? If so, must its particular manifestations—its local and regional repertories and forms of musical behavior—be comparable on a universal scale? Even the most effective classification systems fail to provide firm answers to such questions, but they do serve to remind us that every repertory of folk music contains elements of both the particular and the universal.

Folk Music and Art Music:
Classification and the Composer

Although many of the classificatory concepts devised by folk music scholarship are primarily concerned with the interrelations of various folk music repertories, the urge to understand through classification has also guided much of the thinking about the ways in which folk music interacts with other genres, especially with “composed” art music. I emphasize the action of composing art music because this action identifies a significant aspect of this type of classification, whose end result is usually less a matter of inventory or schematic representation than of use; it is thus a restating of perceived meaning in folk music within a new musical context. Composers in many musical cultures turn to folk music in search of new meanings—or, more precisely, special meanings—and thereby advocate a personal understanding of the broader cultural role folk music stands to play: crystallizing nationalism, encapsulating the essence of musical style and structure, or serving as a font for expanding the musical vocabulary. The composer often bases his or her use of folk music in an art music context on classificatory decisions, and it is not uncommon to find the composer engaged in other aspects of classification in relation to or in preparation for the resituating of folk music in the art music context. Bartók and Vaughan Williams are but two of the twentieth-century composers whose collection and investigation of folk music were inseparable from their creative philosophy. In the history of Western art music, the motivations for using folk music have been many, but they have rarely been completely absent during the formation and change of a musical style (Wiora 1957).

Folk music appears in art music in two ways, each demonstrating many variations. In the first of these, the integrity of the piece of folk music remains. One can easily recognize the tune or another musical parameter that the composer attempts to represent with some degree of literalness. For Bartók the folk melody used in this way was “like a gem in its setting” (1972b: 169). In contrast, the composer may seek not to maintain the external integrity of a piece of folk music but rather to penetrate to the essence of folk song style and to appropriate this essence for the composition of art music. The composer may have very specific designs for folk music used in this way. Bartók, for example, wrote works in Bulgarian rhythms that did not contain direct settings of Bulgarian folk dances. Other composers may think of folk-like styles with more stereotypical goals in mind, as in writing in the style of a musette with a sustained drone and narrow melodic ambitus standing in for a more thorough consideration of the complexities of bagpipe playing.

To portray the history of the interrelation of folk and art music as a commonplace activity of composers can easily oversimplify the more pervasive relationship between different genres of music that characterizes most musical cultures. One must exercise considerable caution when discussing the composer as the primary creative agent in the processes of exchange and cross-influence. The folk music specialist is often neither less adept nor less willing to exercise creativity by turning to numerous musical repertories for new sources. And unquestionably, religious and popular music genres draw from folk music, which in turn accepts influences from those genres. In short, the relationship between folk music and art music is not an isolated phenomenon based only on the decisions of a few musicians but is one expression of the complex patterns of change that continually shape different genres in all musical cultures (Ward 1986). It is an interrelation marked by dynamism and multidimensionality, which are among the most important markers of folk music in the modern world.

Folk music scholarship has not always accepted the multidimensionality of this interrelation. The theory of gesunkenes Kulturgut established primacy for art music, which yielded pieces to folk repertories through a process of “sinking” or deteriorating to meet the prerequisites of survival in oral tradition (Naumann 1922). The first articulation of theories accounting for individual creation of folk music also saw as fairly unidirectional the movement from folk music. John Meier’s motto for folk music resulting from this process was Kunstlieder im Volksmund (art songs in the voice of the folk). As far-reaching as Meier’s claims were in the first decades of the twentieth century, they recognized creativity in only one of the musical genres he sought to pair. Hungarian concepts of classification discern historical patterns within the movement among genres, portraying the interrelation of folk and art music as different potential lineages. One lineage yielded the folk-like art songs (volkstümliche Kunstlieder) of nineteenth-century Hungarian urban genres, which benefited from the influx to the city of rural Hungarians—the practitioners of folk music (Bartók 1972b: 164). A rather different and more direct history is possible when composers of art music penetrate folk music to determine its essence and then to use this essence in art music. Only certain results are possible from this process, giving it a somewhat limited dimensionality. Bartók, for example, could accept the possibility of molding only a tonal musical style on folk music. “Folk music [is] always tonal, and an atonal folk music [is] totally unimaginable. One could therefore not base an atonal twelve-tone music on tonal folk music” (ibid.: 174).

If we place the composer of art music in the role of classifier, it is hardly surprising that each has specific goals for his treatment of folk music. Again, the classifier has fewer universal than particular criteria for interpreting folk music. The ultimate result of using folk music for particular ends is stabilizing an individual style and sustaining its mediation of a canon. The abundance of literature devoted to Beethoven’s use of folk music, for example, largely shows how folk music came to shape his style. Folk music did not leave a deep impression of an especially Austrian folk melos, and in fact Beethoven ranged throughout the folk repertories of Europe in search of sources for the more than two hundred examples of folk melodies that scholars have identified in Beethoven’s oeuvre (Braun 1982-83:287). Bartók, too, did not limit himself to Hungarian folk music of the Old Style when composing. Despite the remarkable diversity in his personal style, it is very distinct from the many folk music styles that inspired him.

What are some of the qualities composers identify in folk music when drawing upon it as a source? One of the most common is the symbolism of the rural, as in the designation of a symphony movement as “pastorale.” Evoking this symbolism can involve explicit or implicit musical references; one may use folk music itself as a source or create a style that is folk-like. For some composers and scholars, the meaning of folk music for classical styles is much more profound than an evocative symbolism. Walter Wiora argues for a stylistic unity that pervades all German song but begins, nevertheless, with folk song. The nationalistic overtones and linguistic biases of such an argument are, of course, obvious, yet there can be no doubt that Wiora’s scheme for classifying the forms and history of German song are predicated on a sound understanding of the genre and a vast pool of examples (Wiora 1971b). The search for Ur-forms in folk music further identifies quite precisely the musical traits that distinguish the genre. Vaughan Williams, for example, transferred his belief in the modal moorings of English folk song to his own compositional style. Whereas many composers concerned themselves primarily with the melodic purity of some folk music, Bartók saw it as the basis for harmonic complexity, speculating that the simpler a melody was, the more individual its harmonization could be (Bartók 1972b: 170). As much as we might feel that Bartók is demonstrating homage to folk music, we must also wonder to what degree he is compelled to transpose the belief in vertical complexity so characteristic of Western art music to folk music. When he states that the simplicity of “Arabic peasant music” lends itself to many harmonic possibilities (ibid.), it seems likely that his concept of classification derives more from art music than from the folk music it seeks to interpret.

The interrelation of folk and art music forms yet another classificatory dialectic, one that more often than not has slightly discomfiting ramifications because of the implicitly negative value judgments that too often appear in contrastive phrases like “high versus low culture.” But the dialectic also bears witness to some positive conclusions that contribute considerably to the study of folk music. First, the composer turns to folk music because of its creative potential; in the composer’s treatment the folk melody is not a frozen artifact whose survival in oral tradition depends on immutably simple structures. Second, the engagement of art musicians with folk music reveals even more clearly the fallacy of constructing any classification system based on stratified models of noninteracting genres; folk and art music are rarely not influencing each other in some way. Finally, the dialectical interrelation of folk and art music is remarkably dynamic and multidimensional, revealing complex processes of change that occur as much in folk music as they do in art music. It is hardly surprising, then, that this dialectic, too, has failed to produce a single, universal method of classification.

Classification and Defining (Defending) the Canon

Few successful classification schemes can avoid paring away some extraneous material from the repertories they order. If nothing else, classification must be manageable; it must extract from the amorphous whole some basic musical structures, whether these are abstract melodic types or simply common and persistent tunes. But the process of paring away is fraught with problems. The distinction between decisions that include or exclude is often difficult to make. A reductionist momentum may also build up, resulting in eventual claims that only a few tunes or texts truly represent a repertory; some of the otherwise fine attempts to classify the Anglo-American folk music repertory have suffered this fate (cf. Bayard 1950, esp. 3-4). Bartók, too, was far more comfortable with the exclusiveness of the Old Style and New Style, which were relatively purged of foreign, urban, and popular influences, than with the Mixed Style, which bore witness to the changes through which modern Hungarian folk music had passed.

The extreme form of the penetration of classification to the specific is the gradual replacement of observed and collected tradition with the system that has purported to define it. In other words, the classification system becomes the surrogate for the tradition itself. The emergence of a surrogate tradition is the most extreme and insidious product of canon-formation. The initial assertions that English ballads and dance genres were the archetypes of folk music in the Appalachians quickly became a surrogate tradition, spurring collectors to discover songs that would justify the appropriate classification systems, especially the system devised by Child. As a result, many of the same collectors ignored the more popular black-influenced styles that were flourishing in the same region. Under Nazism in Germany during the 1930s and early 1940s, folk music classification declared a sort of independence from ethnography, thus allowing the establishment of claims that Germanic and Nordic folk music was somehow central to European folk music in general (cf. Danckert 1987:21-26).

Classification is almost never the product of strictly objective goals. Whatever the claims for discerning a repertory’s “internal order” may be, value judgments inevitably attend the choice of criteria for comparability and, more pregnant as a force molding the ordering decisions, those pieces that actually constitute the repertory abstracted by classification. More often than not, those who most effectively use classification in this way are no longer aware of their redefinition of tradition. “The main danger in classification,” warned George Herzog, “is that it leads to assuming that objects in a storeroom were originally created and deposited in that order” (1937:51). The value judgments of classifiers may initially be very personal and idiosyncratic, but the discursive boundaries they set may assume a rigidity that renders them a surrogate of the tradition itself. Once a means of defining the canon, these discursive boundaries serve only to defend it.

Classification constantly concerns itself with the discursive boundaries—internal and external—of folk music. Whenever one talks about folk music in a systematic way, one engages in some form of classification. Our theoretical vocabulary is full of designations that accept or reject, acknowledge or ignore certain approaches to classification. Performers engage in a form of classification when they choose a particular means of interacting with different communities. The mass mediation of folk music, too, relies on widespread understanding of certain levels of classification. These various ways of discussing and portraying the interrelatedness of folk music may define it as a genre in restrictive or expansive terms. At some levels, the discourse of classification therefore serves only to perpetuate old canons; at others, it forges new canons.