… thou hast clapped thine hands, and stamped with the feet.
It is clear that the source of music is in the human body. Likewise, dance, though dance is a single thing, while music can be either vocal or instrumental. Song, like language, is produced by the vocal apparatus, while instrumental music emerges, like dance, from the body’s movement. Language and song, dance and instruments are two symmetrical pairs, evident beyond the murky realm of origin and anteriority. Born from the human body two by two, they are simply recombined in operatic drama; Balinese theater, among other traditions, brings this principle of reunion to the extreme, by mingling speech with song and dance, and by keeping the orchestra onstage and visible, truly incorporating the instrumentalists’ gestures into the action performed.129 Music and dance are intimately bound together, and while song could perhaps have existed without language, instrumental music in its most primitive forms always presupposes dance: it is dance. Humans strike the ground with their feet or hands, beat time on their own bodies, or move all or parts of their bodies, shaking the jingles and ornaments they wear. These, no doubt, were the first forms of instrumental music to come into existence: barely dissociable from dance, while essentially different from singing. The widespread idea that instrumental music arose from imitating song is untenable. Indeed, there is no proof that any attempt to imitate the human voice with the help of instruments was ever made. Lucretius cautiously suggests that even before singing, we imitated “the liquid song of birds” with our mouths, and that the sound of the wind in the reeds inspired us to blow into their hollows:
At liquidas avium voces imitarier ore
ante fuit multo quam levia carmina cantu
concelebrare homines possent aurisque juvare.
Et zephyri, cava per calamorum, sibila primum
agrestis docuere cavas inflare cicutas.130
Does Lucretius mean that we might have whistled before we sang, and proceeded from the natural whistling of the lips to whistles or reed flutes? If so, then instrumental music modeled itself on something other than the human voice. Even supposing, somewhat problematically, that human bodies were originally aphasic, the same bodies might well have been able to grasp the rudiments of music if they were guided by their first dancing or laboring gestures. In the present day, the most primitive of musicians might beat the ground or strike a piece of wood, and we would not assume that the gesture was motivated by the desire to imitate the sounds of the voice, nor even to mimic sounds made by beating out rhythm on the body. A whole misguided perspective on the art of music stems from misusing the concept of imitation; the ill effects of this perspective could be detected far and wide. Human instrumental music, despite what it receives from vocal music—and what it teaches vocal music in return—has its own trajectory, which fails to coincide with that of song. Even if singing preceded the use of instruments, the prior existence of one does not guarantee that it presided over the invention of the other. The theoretical authority of song over instruments comes from the importance that humans ascribe to their speech, and nothing more. In rejecting this excessively vocal conception of music, however, we need not adopt the opposite thesis, asserting—by false analogy with tools, which were created to extend the action of the hand—that all instruments have a manual origin. The hand, like the voice, should not be cast as prima donna.
Let us not yet set aside the idea of song as a possible model for instruments. In all likelihood, it is the abstract unity bequeathed upon song that strengthens the appeal of this apparent relationship of model to copy. But just as the beginnings of instrumental music are not restricted to the mere action of the hands, or feet, or any other part of the human body, there are also many vocal techniques. It is fallacious to speak, as we often do, of song as if the word designated a single and homogeneous art. We have but to think of the infinity of timbres produced by the varied resonance of vowels within a single language, the friction or vibration of semivowels, the percussive effects of certain consonants: the flexibility of the vocal apparatus far outstrips what the current state of mechanical reproduction can match. Each of these phonetic systems belonging to a specific language131 has a corresponding, often unique mode of singing, and this diversity is intensified by the variety of ethnic particularities of vocal physiology—deep bass tones of Slavic or Negro peoples, high piercing tones of Far Easterners—and the ways in which the music of these cultures draws on these specificities. By a reversal of the natural sequence of voice types according to sex and age, castration produces a striking and distinctive soprano voice. In other contexts, singing may be entirely nasal or remain entirely glottal; make abundant use of yodeling (as with the Tyrolean joddle), and include in its natural purview (and often in that of —choral singing) diverse sounds made with a closed mouth, hiccups, pants, clucks, hisses, whistles, strange cries; in brief, it may tend toward a certain form of the “grotesque,” reminding us of singing’s origin in magic.132 Considering all these factors, including the phonetic resources of the languages in which we sing, national and regional pronunciations, diversity of registers, variety of vocal noises, forcing of normal timbre and so on, a system for the instrumentation of vocal resources has been constructed, equal in richness and potential to that of “regular” instruments in terms of its complete repertoire of tools as well as of the immediate availability of the majority of those tools. This ready availability would allow the voice to surpass the instrument if instruments did not nearly always offer, in exchange, a pitch range and capacity for harmony inaccessible to the solo voice. Singing has the advantage of the inimitable diversity and mobility of its timbres; instruments have that of the range and simultaneity of their sounds. The organ, despite its vast musical capacities, may aspire to match nothing greater than an orchestra of wind instruments; it possesses the multiplicity of voices of a choir, but not the choir’s fluidity of timbre. Almost imperceptible by mechanical means, the continual passage from one timbre, and one sound to another is absent in both polyphonic and monophonic instruments: the most flexible and lavishly orchestrated ensemble is powerless against the halting sensation produced by the hiatus between timbres. Conversely, a single voice can glide from language to song, or within language from spoken to recited speech, or from a certain natural form of singing to a more violently stylized one. And when early harpsichord, flute or violin treatises invoke singing as a model for performance, drawing on the art of bel canto, this invocation is only effective insofar as the dematerialized voice, isolated from its sometimes monstrous physiology, retains its flexibility while remaining in all other ways an instrument. At this midpoint, the voice becomes a more accessible ideal and a symbol whose meaning is thenceforth clear. The song-like nature of the instrument—cantabile Art, as J. S. Bach terms it in the introduction to his Inventions 133—does not refer to the most characteristic features of the voice, but to that part which has essentially been converted into an instrument.
130 “Now, mimicking the liquid song of birds with the lips came long before people were able to make a practice of pleasing the ears by singing melodious songs. And it was the whistling of the west wind through hollow reeds that taught rustics to blow on hollow hemlock-stalks.” Lucretius 2009, 103, lines 1379–83.
131 Vendryes 1925, chs. 1 and 2.
132 Here we may mention an idea familiar to Professor Sachs: according to a certain theory, our idea of what is natural in singing does not in fact date back very far (no further than the Renaissance) and constitutes the sign, in a sense, of the secularization of our music. In this reading, modern-day “archaic” civilizations would be unaware of “natural” or profane forms of singing.
133 “Eine cantabile Art im Spielen zu erlangen” (“to arrive at a singing style in playing”), title page of J. S. Bach’s Inventions and Sinfonias, Cöthen, 1723. See also Clavierwerke, vol. 1, from the monumental Bach-Gesellschaft edition. [Hans, T. David, Arthur Mendel, and Christoph Wolff, eds., The New Bach Reader (New York: W. W. Norton, 1998), 98; J. S. Bach’s Johann Sebastian Bach’s Werke, vol. 3, Clavierwerke, vol. 1, ed. C. F. Becker (Leipzig: Bach-Gesellschaft, 1853).]
Lacking the degree of ductility of the vocal apparatus, instruments might yet have sought in singing the sort of general timbre that makes it possible to recognize the sound of a human voice, even from a distance. Yet only occasional effects such as vibrato, glissando, and unpredictable combinations of overtones in an orchestra lend themselves to (false and fugitive) imitations of the human voice. Once again, how implausible it is that instruments should have sought to imitate the voice.
To establish the moment at which singing itself ceased to cling to the model of speech in order to pursue an instrumental path is perhaps a more difficult problem. The Greek music theorists were the first to identify it (Aristoxenus of Tarentum in particular): since the voice is the organ of speech as well as of song, language and music, and since these are all distinct entities, it was needful that music establish some kind of fixed delimitation within the zone of perpetual transition we have described, one that posed no limitations on the freedom of timbre, and which perhaps drew on characteristics of instrumental music.
Let us then summarize and interpret the theory put forth by Aristoxenus of Tarentum in his Elementa Harmonica. The human voice moves in two different ways: if “it moves in such a way as to appear to the hearing to stand still nowhere, we call this movement continuous”—and thus we have language; but if “when it appears to stand still in one position and then seems to pass over some space, and having done this appears again to stand still” in another place, this intervallic movement characterizes singing.134 It is hardly worth mentioning here that the Greeks did not possess diastematic notation, our staff of multiple horizontal lines, which cannot help but introduce a notion of space into musical notation, and indeed, establishes the coordinates of such a space; for them, the height of pitches depended not on notes but on the strings of a lyre, and later on the position of playing holes, beginning at the mouthpiece of a chalumeau.135 The use of terms expressing stillness at a particular point, or intervallic movement, would have been clear in the minds of Greek readers who could draw connections only between, on the on...
