One routine question in various forums is "What is polyphony?" and one example which is always raised by discussion is the drone. Simply put, drones repeat a few notes in the background, and provide something of a framework over which primary melodic lines are spun. Drones are ubiquitous in the classical music of India which, together with the body of scholarly writing on the subject, makes it the natural choice around which to center a discussion of drones. In the terms of Indian music, a drone is not a voice in polyphony, because the music is not polyphonic. Some have suggested that this should be stated otherwise for Westerners, as a means to pay proper respect to Indian music. In other words, we don't want people saying that the West invented polyphony. What with East Asian gamelan, the various African ensembles and all the rest, it seems rather blind to say that the West "invented polyphony" and implicate all of these examples as somehow unworthy. So maybe we should say that the drone makes polyphony. On the other hand, I am not too keen on the idea of taking various unconnected developments around the world and trying to decide which do or don't fit into some arbitrary definition. It is more compelling to look simply at what each has to offer and to appreciate the diversity. It is not as though we owe royalties for musical inventions... it is a point of trivia.
The issues the drone raises can be interesting, however. Although it is not considered a "voice" and in fact the performer of the drone instrument traditionally is not named, it is still highly significant for Indian classical music. The tuning of the drone is very precise, and does vary based on the individual raga, although it does not necessarily use only notes of the raga. It remains fixed during a raga development, and of course the usual lack of modulation in Indian classical music means that the drone emphasizes the fundamental note very very clearly. The simplest motivation, especially for a poor singer such as myself, is that the key note is always present and so maintaining intonation becomes easier. One can certainly speculate that this situation has facilitated the complexity of intonation in Indian music. The drone is articulated as an arpeggiated chord, and should be performed in a disinterested manner, without reacting to either the rhythmic or melodic motion of the lead performer. In that sense, it is not what would usually be called accompaniment in the West. The drone is traditionally performed on the plucked string tampura (or variant spellings), by someone sitting in the back of the ensemble. It is increasingly common, especially in Carnatic music, to have an electronic drone box, which is consequently extremely precise and quintessentially disinterested and so provides some compensation for the loss of real string sonority. Since the three or four notes of the drone are tuned precisely in overtones of the fundamental, they also produce clear combination & difference tones, making the spectrum quite rich. In a good acoustic environment, the notes of a raga can almost be heard as ornamentation to the drone.
Indian classical vocal music actually has two other forms of accompaniment. One is the percussion accompaniment, which is also tuned, and which at its best directly extends the rhythmic patterns of the soloist into filigree units too difficult for the voice to negotiate. Casually speaking, it is the opposite of the drone, and so I will leave the topic for another day. The other accompaniment is melodic, and serves both to allow the voice some rest as well as to echo the phrases of the main melody. Especially given the improvisatory nature of this music, the latter can serve as an essential springboard for the next level of elaboration. This accompaniment is strictly secondary to the main line, but reacts to it in a way that the drone does not. To return briefly to the opening remarks, it is not viewed as a second polyphonic melody either. Rather, it gives an alternation of the main line from one instrument to the other, with the melodic accompanist frequently playing a few long basic notes while the singer performs passagework, just as the drone does. Something in this style is quite typical of emerging performance traditions for Western medieval secular music, such as the songs of the troubadours or the post-trouvère style in the monophonic songs of Machaut. It is also tempting to look for origins of Western polyphony in the drone. Although the very slow tenors of Perotin could almost be viewed as drones backing the upper parts, especially in view of my earlier remark on hearing the raga as ornamenting the drone, earlier generations of liturgical polyphony actually feature more note against note writing. A contrapuntal basis of this sort is more uniquely Western, although I hesitate to make a stronger remark. Two-part writing in some virelais by Machaut presents a more intriguing case, and in fact reconstructions such as those by Emmanuel Bonnardot are strongly suggestive of the idea that the second part derived from improvised drone- & alternation-based accompaniment.
If one posits a partial origin for the notated second and third lines of the polyphonic chansons of Machaut in a parallel accompaniment tradition, rather than in liturgical polyphony alone, the implications certainly stretch forward into the artistic dominance of the three-voice ballade and in turn the songs of the early Renaissance. In some respects, this suggestion would make the ascendance of the true part-song short-lived, as the basso continuo technique of the Baroque is a partly improvised accompaniment and in some ways related to the idea of drone. Although in this case the basic unit becomes the chord, contemporary sources increasingly suggest that continuo playing should adopt something of the "disinterested" quality of the drone and not react to voice leading or contrapuntal issues in the melodic parts. When played on an arpeggio-based instrument, the analogy becomes that much more direct. I would also like to suggest that the idea of drone takes on a new meaning in our contemporary society, which is fundamentally noisy. Certainly "new age" and electronic music have adopted various drone effects, and indeed if we return to the example of the electronic drone box in Carnatic music, the use of electricity has another simple implication. Although we are so accustomed to it that we might not notice, the alternating frequency of ordinary electricity is within the audible range. Incorporating inescapable drones into the idea of drone accompaniment may be extremely practical, even if I remain an advocate of acoustic instruments. What it can provide is a means of framing acoustic space, especially as that space is increasingly distorted by the environment in which it is embedded.