What must be one of the earliest substantial references to music within anthropology appears in Franz Boas’ 1888 monograph The Central Eskimo,[68] which contained transcriptions and some analysis of two-dozen Eskimo songs, within the context of a comprehensive ethnography. Not only did this set the stage for an anthropological approach to the study of non-Western music, but it represents a scholarly effort almost as early as Ellis and Stumpf in comparative musicology. Indeed, Boas and Stumpf worked together in collecting and analyzing Northwest Coast indigenous music, subsequently published by Stumpf in 1886 in Vierteljahrsschrift für Musikwissenschaft, one of the earliest journals in comparative musicology.[69] So it would be a mistake to assume that comparative musicologists and anthropologists operated in isolation from one another. Nevertheless, their overall approaches to the study of non-Western music were very different in many ways. The comparative musicologists were motivated by theories about the origins and structure of music, and analyzed their materials accordingly. Boas’ work on music lacked a theory of this type,[70] and instead fitted into a framework of meticulous detail and ethnographic salvage work.
In addition to the music contained in The Central Eskimo, Boas published articles on Kwakiutl, Chinook, Eskimo and Sioux music between 1888 and 1925. Each of them, while focusing on music, contains the elements that have come to be associated with Boasian cultural anthropology. The article "On Certain Songs and Dances of the Kwakiutl of British Columbia"[71] contextualizes several musical transcriptions of songs (presumably done by Boas himself) by including accounts of ritual, song texts, and a very lengthy version of mythology relevant to the songs. His article "Chinook Songs"[72] contains three brief notations along with 38 song texts and a glossary of several dozen words. Two articles, both entitled "Eskimo Tales and Songs",[73] present song texts along with explanations of dozens of words used in them. His article "Teton Sioux Music"[74] addresses issues of musical form, including rhythm, phrasing, and structure, and includes 11 quite detailed musical transcriptions. Although he does not approach the technical and analytical detail of the work of some comparative musicologists, Boas was on par with many, and his work on music is under-recognized, probably due to the enormous volume of work on other subjects.
Boas’ impact on the cross-cultural study of music is felt most strongly not in his own research and publications, but in those of his students and their students, a legacy which has led to a number of the most significant ethnomusicologists working today. Edward Sapir, the ethnographer, linguist and poet, was also a musician and grew up in a musical family (his father was a professional cantor).[75] His first in-depth work on songs was in 1910, when he transcribed over 200 Paiute (northern Arizona/southern Utah) song texts[76] and recorded 120 songs on wax cylinders.[77] Along with this material, much of it unpublished, was copious contextual information on the performances, people and places associated with the songs.[78] Sapir’s recordings were transcribed by his father, Jacob Sapir.[79] This research led to Sapir’s article "Song Recitative in Paiute Mythology",[80] which included his own musical transcriptions and musical analysis focusing primarily on rhythm. This represents a notable difference, given the strong emphasis in comparative musicology on matters of melody, scales and intervals.
Some of the fundamental differences between comparative musicological and cultural anthropological approaches to the study of non-Western music can be found in a 1912 review Sapir wrote of a publication by Carl Stumpf (Die Anfänge der Musik). Sapir offers a précis of Stumpf’s belief that, in primitive music, notice was taken of "the unified effect of tones sung at consonant intervals", while other intervals "dissonant or relatively so, would in time arise by giving the voice free play within the fourth, fifth, and octave".[81] Sapir criticizes this approach as being difficult to prove or disprove, and as not being based on historical data.[82] As Sapir writes:
In the nature of things any such theory must be purely speculative, as the use of musical tones is far too ancient a heritage of humanity to yield its genesis to historical reconstruction. Failing historical evidence, a theory of origin can be fully convincing only when so well grounded in psychology as necessarily to exclude all other possible theories. This is hardly the case here.[83]
Sapir is also critical of Stumpf’s over-emphasis (which could be extended much more broadly in the comparative musicology of the day) on "the purely intervalic [sic] side of music":
Music is neither purely tone nor purely rhythm. Would it not be more suggestive to think of it in terms of an association of tone production, however it might arise, with the rhythmic impulse manifested in all of man’s artistic activities? Granted this impulse and the possession of vocal chords, adjustable for changes of pitch, various forms of musical expression might be expected to arise. Several paths seem possible, the actual course or courses traversed lie beyond our ken.[84]
And in a similar vein, critiquing Stumpf’s handling of melodic structure and musical form, Sapir writes:
I am inclined to doubt whether a purely musical study of this problem would be as fruitful as when taken in connection with song-texts, dance forms, and other features as musical execution is wont to be associated with in practice. The peculiarities of melodic forms are often due to factors that have no direct relation to musical problems as such, as witness our masses, lullabies, and bugle calls. These remarks are meant to indicate the necessity of studying the more complicated problems presented by primitive music in connection with associated cultural features. Stumpf’s relative neglect throughout the book of all features that are not strictly musical in character is naturally to a large extent unavoidable, but we must not fail to realize that such one-sidedness may lead us astray in our interpretations.[85]
Overall, Sapir’s review is quite favourable; these comments, however, indicate certain matters of interpretation and emphasis that differentiated comparative musicologists from anthropologists, matters that still, to some extent, characterize ethnomusicology.
Another of Boas’ students who was very influential in the development of ethnomusicology was George Herzog. Herzog is an interesting character in this narrative, as he really stands with one foot in each of the two camps that I am discussing. He first studied at the Royal Conservatory in Budapest, and was heavily influenced by the folk music research of Béla Bartók and Zoltán Kodály.[86] From 1923 until 1925, Herzog worked under Hornbostel at the Berlin Phonogramm-Archiv, the most important archival institution for comparative musicology until the Second World War, and thus was part of the so-called Berlin School of Comparative Musicology.[87] In 1925 he emigrated to the United States and studied anthropology under Boas.[88] Thus, in one individual we have the intellectual descendant of a variety of important approaches to the cross-cultural study of music. Bruno Nettl, Herzog’s most prominent student, wrote the following about Herzog’s letters, which give an interesting insight into his place in the development of ethnomusicology:
Throughout the correspondence one sees the hand of Hornbostel, the comparativist and the archive-builder, of Bartók the careful processor and analyst concerned with authenticity, of Boas the methodical fieldworker, of the confluence of folkloristics, linguistics, and ethnography…The multidisciplinary nature of American ethnomusicology is in part due to his many-sided reach.[89]
Another prominent student of Boas’ who was greatly influential in the development of the anthropological approach to non-Western music was Melville Herskovits, whose interest in the relationship between African and African-American culture was stimulated by the role of music.[90] Although Herskovits did not publish extensively on musical topics, he supervised the doctoral dissertations of both Richard Waterman and Alan Merriam (who would both be very prominent in the new field of ethnomusicology), and he made field recordings of music in Dutch Guiana, Haiti, Trinidad, Brazil, Dahomey, the Gold Coast and Nigeria.[91]