Ethnomusicology, as it is now known, was first manifested as an organized academic field of study under the rubric "comparative musicology" in the late-nineteenth century. A common view is that one of its founders was the tone-deaf^[13] British physicist A. J. Ellis, whose best-known work was his 1885 article "On the musical scales of various nations". As a physicist and musical enthusiast, Ellis’ interest was in using the methods of acoustics to measure the characteristic pitches and intervals of musical scales of different musical cultures with the aim of comparing them. Ellis’ starting point, however, was his understanding (as a physicist) that the interval between any two notes:

As Jaap Kunst explains in a discussion of Ellis’ paper, an octave is represented by a 2:1 ratio, a perfect fifth by a 3:2 ratio, and a perfect fourth by a 4:3 ratio. When the two pitches have no lowest common denominator, the ratios become very unwieldy and a logarithmic table is used.^[15]

His long paper was originally a talk given to the Society of Arts with many accompanying illustrations on various instruments, including a dichord, a number of English concertinas specially tuned to different scales, a sitar, a koto and a set of Chinese bells. The research on which the talk was based was done with instruments obtained privately or through museum collections, some of which had fixed tuning (and therefore which could be investigated independently), and others which had to be played by native musicians; these included the Arabic oud, the Scots highland bagpipe, the Indian sitar and vina, Chinese flutes, mouth organs, gongs and stringed instruments, the Japanese koto, and both slendro and pelog gamelan orchestras. His treatise is full of tables and charts showing the exact frequency measurements and intervals of all of these instruments, as well as scales derived from them. Ellis reveals an ethnomusicological sensibility that is somewhat ahead of his time when he writes that it is necessary to be a native musician is to hear the real pitch of a musical scale, that there might be considerable variation of such a scale from musician to musician, and that, at any rate, there is a significant difference between knowing the notes used in a piece of music and the musical theory which underpins it.^[16]

In Ellis’ conclusions, he compares the principal intervals used in these different scales from all over the world, proposing how particular intervals were altered to produce distinctive scales, and even suggesting how particular musical features from one region of the world influenced another. For his final conclusion, however, Ellis writes: "the Musical Scale is not one, not 'natural,' nor even founded necessarily on the laws of the constitution of musical sound, so beautifully worked out by Helmholtz, but very diverse, very artificial, and very capricious."^[17]

So, even though Ellis’ starting point and methodology were based on the Western conception of the physics of sound, there is also an explicit recognition that the musical practices of different cultures might be based on quite different principles. Most importantly, it has been observed that his work provided the empirical foundations for comparative musicology.^[18]

Perhaps the most important aspect of Ellis’ work was his development of a system of "cents" for the study of intervals between pitches. To measure non-tempered intervals in terms of ratios of the two pitches in question is cumbersome in the extreme, and to use a logarithmic scale is only somewhat less so. Ellis’ most enduring contribution to ethnomusicology was to propose the division of each semitone in the equal-tempered scale into one hundred equal increments which he called "cents"; although a difference of one cent is impossible to discriminate, Ellis felt that most sensitive ears could register a difference of five cents.^[19] By freeing comparative musicology from the need to negotiate complex logarithms, Ellis provided a methodological tool that not only facilitated the comparison of different tonal systems, but also maintained a strong orientation around tonality as a primary concern of the discipline.

A prominent influence on early scholars of non-Western music like Ellis was psychological theory — the idea that the study of various aspects of music, like rhythm or tonality, could reveal principles of the functioning of the human mind. Indeed, comparative musicology and psychology were very closely integrated in the late-nineteenth and early-twentieth centuries,^[20] and many early research problems were oriented around the investigation of the origins and development of music and universal musical principles.^[21] Carl Stumpf’s Tonpsychologie (1883 and 1890) was particularly influential, developing a theory of tone sensation based on individual cognition: for instance, examining the perception of similarity and difference in tonal stimuli, which led to a major focus on tonal distance and scale formation advanced by Ellis and many others.^[22]

As Dieter Christensen has pointed out, Stumpf "believed in the unity of the human mind",^[23] and required musical data from all cultures to substantiate his theories. Toward this end, and under his direction, a large collection of phonograph recordings — which eventually became the famous Berlin Phonogramm-Archiv — were amassed that would be analyzed for the purpose of psychological research.^[24] The Phonogramm-Archiv also became the basis for the development of a unique form of evolutionary theory focusing on music and, in particular, on pitch and intervallic organization.^[25] As Eric Ames notes, Stumpf and his colleagues used recordings of non-Western music to emphasize musical evolution. However, evolutionary thought at the turn of the twentieth century took a variety of forms beyond a strict linear model, and Stumpf was able to construct an evolutionary discourse that accommodated notions of both development and geographical diffusion.^[26]

Stumpf’s colleagues, notably Erich von Hornbostel and Otto Abraham, continued to investigate psychological aspects of music, based primarily upon the examination of tonal systems, scales, intervals, and the like. Regardless of their commitment to the intricacies of psychological theory, other comparative musicologists certainly maintained a dominant interest in scales and melodies well into the twentieth century. As Stephen Blum notes, for many comparative musicologists "'musical system' was in effect synonymous with 'tone system'".^[27] Blum quotes Robert Lach as an example:

This emphasis on scales, melody, pitch, heterophony and polyphony — all aspects of music associated with pitch or tone — was prominent in much comparative musicology, particularly in the work of Hornbostel.

A well-known example is his theory of the cycle of blown fifths, discussed in detail by Jaap Kunst. Hornbostel’s theory has it that an ancient Chinese tone-sequence or scale was developed as follows: a tone of 366 Hz (the so-called huang chong, or "yellow bell") was produced on a length of stopped bamboo 230 mm in length and 8.12 mm in diameter. When this tube is overblown, it produces a tone a twelfth above the fundamental, which is then transposed down an octave to give the second tone of the sequence, a fifth above the fundamental; an overblown twelfth above this second tone, transposed an octave down, produces the third tone, a fifth above the second tone; and so on. A cycle produced in this way using pure fifths (702 Hz, a "Pythagorean" interval based on string-measurement) would arrive back at the same note from which it started after 12 jumps. However, Hornbostel noted, since the blown fifth is roughly one-tenth of a tone flat, the cycle is not completed until 23 jumps. Hornbostel postulated that before the Chinese began to construct scales based on Pythagorean principles, they must have developed their scale based on this cycle of blown fifths. Hornbostel and others discovered scales from around the world which could be derived from this principle.^[29] Hornbostel’s complex theory has been criticized since its inception, but it reveals two key concerns of the comparative musicology of his day: first, a concern with the origins of musical features; and second, a belief that such origins may be revealed through the study of tones, scales and intervals.

Another example is Hornbostel’s ideas about the concept of pure melody. Unlike European music, which since roughly 1600 has been based on harmony, all other music is based on pure melody. Hornbostel believed that harmony was superimposed on "natural" traits rooted in the "psycho-physical constitution of man": a natural melodic movement downward ("like breathing or striking, from tension to rest"); small melodic steps of at most a major third; a "melodic unity" with a disregard for fixed scale steps; and a concern with the "distance" between notes — "the size of steps" —rather than their consonance.^[30]

As with Stumpf, there is an undercurrent of evolutionism in Hornbostel’s speculation on melody.^[31] He states that short, repetitive melodies of a few notes less than a fourth apart are indicative of "an early stage of development" and are due to "a narrow range of consciousness",^[32] and that the natural evolution of melody is from this level to a larger number of notes, a greater range, and longer phrasing.^[33] Furthermore, in his study of African music, Hornbostel believed that polyphony was a natural development from antiphony (the alternation of solo and chorus singing), probably by accident when the two parts unintentionally overlapped. However, he is clear that this polyphony is the result of a melodic, rather than harmonic, principle.^[34]

Hornbostel’s theoretical orientation is well-revealed in a 1933 article called "The Ethnology of African Sound-Instruments". Hornbostel was an advocate of a form of evolutionism, but one which was combined with the geographical diffusion of cultural traits.^[35] In fact, at certain points he seems to deny the validity of evolutionary theory for the purpose of understanding cultural phenomena. Of the idea that we might reconstruct the history of cultural phenomena, such as musical instruments, by arranging them according to differential features which would reveal their relative ages, Hornbostel writes: "Plausible as this reasoning sounds, its utility as a guide to method is doubtful, and theories of evolution, however ingenious, can contribute little to the classification of cultural phenomena in chronological order, which has always been accepted as one of the most important problems of ethnology."^[36]

In support of this critique of evolutionism, Hornbostel cites two instruments familiar to the Australian ethnographic literature: "the bull-roarer and the boomerang impress us by the subtlety of their technique, yet they belong to the Australian aborigines"^[37] - which certainly qualifies as a case of damning with faint praise.

And yet, a chronological or developmental approach, whether or not we call it evolutionism, is present throughout Hornbostel’s work. He states explicitly that if one phenomenon belongs exclusively to a "primitive" culture and a second belongs exclusively to a "higher" culture, then the first phenomenon is chronologically earlier^[38] — which seems like a case of circular reasoning. Hornbostel also gives cautious approval to Tylor’s concept of "survivals", but does not agree with the full extent of Tylor’s reasoning. Hornbostel writes that Tylor felt that "inorganic and uncomprehended phenomena belong to an earlier stage of development",^[39] but he is more cautious: "[I]norganic and uncomprehended phenomena are therefore merely an indication that we have to do with older cultural phenomena, but in what connexion, and to what culture they must be assigned, can only be decided on its merits in each individual case."^[40]

On the subject of monogenesis vs. polygenesis, Hornbostel is unambiguously clear: it is highly unlikely that any specialized cultural phenomenon could develop independently in two different locations:^[41] "[T]here has never, so far as I know, been an historically verified instance of polygenesis in different cultures."^[42] Instead, Hornbostel advocates what he calls "the despised 'diffusion theory'",^[43] arguing that some cultural phenomena may survive for a long time and may travel great distances. Any specialized cultural phenomenon, according to Hornbostel, must only have developed in one place at one time — as he famously tried to demonstrate through his theory of the "cycle of fifths". Subsequent distance of the phenomenon from the central point of origin indicates relative age.^[44]

This leads us fairly directly to Kulturkreise theory, of which Hornbostel was an advocate. The idea is that cultures may be defined by a collection of individual characteristics (architectural, ritual, musical, social, etc.); complexes of characteristics shared by several cultures, called "culture-circles" (Kulturkreise), must have a common origin. These culture-circles have moved outwards from their points of origin, have mixed, overlapped and developed, resulting in the current situation of cultural diversity.^[45]

Another prominent figure in comparative musicology, Curt Sachs, betrayed a more obvious evolutionist stance than Hornbostel. In his The Rise of Music in the Ancient World: East and West, Sachs writes from the outset of "the plain truth that the singsong of Pygmies and Pygmoids stands infinitely closer to the beginnings of music than Beethoven’s symphonies and Schubert’s lieder".^[46] He continues by stating that "the only working hypothesis admissible is that the earliest music must be found among the most primitive peoples".^[47] Later, Sachs writes that:

Some of Sachs’ writing seems almost comical in its evolutionism by today’s standards, but he was by no means alone in his thinking among some scholars of his day. What is somewhat more remarkable is that this evolutionist line persisted into the 1940s, long after it had met a timely demise in mainstream anthropology. How is it that comparative musicology lagged theoretically behind cultural anthropology by at least two decades, when the two disciplines dealt with such similar subject matter?

Sachs’ interpretive framework reveals not only an underlying evolutionist conceptualisation of music and its development, but also how the evolution of music is revealed through particular musical structures. As with Hornbostel, melody features prominently as one key to understanding the contemporary music of so-called "primitive" people as representing the precursors to European music. Although, "[t]o the evolutionist, one-tone melodies as a first step before the rise of two- and three-tone melodies would almost be too good to be true",^[49] Sachs acknowledges that the best available information suggested that two-note melodies are the earliest that can be traced.^[50]

Sachs refers to these basic melodies which alternate between two notes a short distance apart as logogenic, or word-born. He considers them to be "a mere vehicle for words",^[51] and they evolved in an additive way; that is, certain other notes became added to the central core.^[52] Opposed to logogenic melodies were pathogenic melodies, which are due to "an irresistible stimulus that releases the singer’s utmost possibilities";^[53] these are characterized by great force and passion at the beginning of a phrase, only to diminish and weaken toward the end. "Descending melodies", Sachs writes, "recall savage shouts of joy or rage, and may have come from such unbridled outbursts".^[54] In pathogenic melodies, evolution proceeds in a divisive way; that is, the larger vocal range resulting from savage outbursts leads to the octave and larger intervals like fourths and fifths creating a skeleton for melodic development.^[55]

Between these two extremes is melogenic music, which represents a kind of middle ground characterized by elements of both logogenic and pathogenic. At this more developed melogenic level, where melodies tend to have a range greater than a third, particular intervals tend to crystallize, "determined by simple proportions of vibration numbers":^[56] the 2:1 ratio of the octave, the 3:2 ratio of the fifth, and the 4:3 ratio of the fourth. Sachs’ reasoning is sometimes far from convincing, as when he writes: "The strongest magnetic power emanates from the fourth — for physiological reasons it is here best to accept without attempting discretionary explanations."^[57]

This "magnetic attraction" results in the development of tetrachords and pentachords, and ultimately the complex melodic structures of "highly civilized peoples".^[58]

In a further extension of evolutionist principles, Sachs asserts that the earliest stages of music, represented by contemporary "primitive" people, also appears in the babbling songs of European toddlers. Sachs concludes: "Thus we cannot but accept their babbling as an ontogenetic reiteration of man’s earliest music and, inversely, conclude that the music of today’s most primitive peoples is indeed the first music that ever existed."^[59]

This was in 1943. It would be easier to dismiss such speculative thinking if Sachs was not so influential on others in comparative musicology — including, as will be discussed below, the early thought of Alice Moyle.

Ellis, Hornbostel and Sachs are only three, albeit quite influential, figures in comparative musicology, and it would be seriously misleading to suggest that their views are completely representative of an entire emerging discipline. Like any discipline, there was a great deal of internal diversity, which is suggested by the diverse backgrounds of the early practitioners: musicology, composition, physics, chemistry, phonetics. And yet, in their thought we can see some of the broad outlines which would come to characterize the "musicological" half of ethnomusicology. I have already commented on an underlying evolutionist stance, although this was not equally prominent among all scholars. Another feature was a unitary approach, viewing "the world’s musics as a group of separate units, stylistically distinct and internally homogenous",^[60] and therefore able to be characterized as wholes on the basis of very small samples which reflect a single set of principles.^[61]

The focus on melody, polyphony, scalar structures and other similar musical features has been interpreted by Bruno Nettl as reflecting some of the dominant values of Western music in the late-nineteenth century. In Germany and Central Europe, where many of the early comparative musicologists were based, standard musical training tended to stress intellectually difficult musical structures, such as melodic development and the simultaneous interaction of various parts. When the early scholars of non-Western music began their study, we see particular attention paid to these same features.^[62] Nettl continues:

Another feature of early comparative musicology, also related to the "musicological" half of ethnomusicology, was an emphasis on the transcription of non-Western music as an end in itself. Transcribing music onto the Western staff, even after the invention of recording technology, was and still is an important aspect of ethnomusicological training and practice,^[64] although the necessity and accuracy of transcription has always been subject to questioning and debate. Exact notation was not even a feature of Western compositional practice until the late-eighteenth century,^[65] and of course the compression of an entire musical system onto five lines and four spaces representing 12 semitones per octave tends to mould non-Western music to our own image. Any understanding of a non-Western music which is based on a transcription in Western notation will necessarily encode our own musical values.^[66] Nevertheless, if we think of the transcription of music as itself an interpretive orientation, it must be one of the most pervasive in comparative musicology through its first 50 years.

So, if early comparative musicology came substantially to inform the "musicological" half of ethnomusicology, what was the intellectual trajectory that led to the sometimes-opposed, but always intertwined, "anthropological" half?

Two prominent features of comparative musicology as it developed in late-nineteenth-century Europe are an overall lack of contextualization and a universalizing comparative perspective.^[67] This was implicit in Ellis’ large-scale comparison of scales, in Hornbostel’s meticulous analysis of recordings from around the world, and certainly in Sachs’ assumption of a single developmental framework for all the world’s musics. These two features also proved to be a key rift within the scholarly study of non-Western music, and led to the development of a distinctive perspective within American cultural anthropology.