This well-known and often debated 13th Century "rota" is one focus of a fascinating essay on medieval British polyphony by musicologist Shai Burstyn: "Gerald of Wales and the Sumer Canon" (The Journal of Musicology, Vol. 2, No. 2. Spring, 1983). Gerald of Wales, or Giraldus Cambrensis, was a medieval (c. 1146-c. 1223) archdeacon, famous for his writings on Welsh history and culture. According to Burstyn, "Gerald's musical comments and Sumer are, I believe, related and may illuminate one another. I shall endeavor to interpret Sumer, and the musical tradition I take it to represent, in terms of Gerald's frequently quoted description of Welsh Polyphonic singing:"
[N.B. (added on March 28): to my chagrin I've recently discovered that Shai Burstyn is male, not female. This just goes to show how much I have to learn not only about names -- and making assumptions based on them -- but also about the current state of music history and the historians who inhabit it, a field that I must admit I have not kept up with for several years. I'll now proceed to correct my mistake in this and the other posts where it appears. Sorry.]
He quotes his comments at some length, in both the original Latin and an English translation. Here are some highlights from the latter:
As to their musical euphony, they [the Welsh] do not sing uniformly as is done elsewhere, but diversely with many rhythms and tunes, so that in a crowd of singers such as is the custom among these people, you will hear as many different songs and differentiations of the voices as you see heads, and hear the organic melody coming together in one consonance with the smooth sweetness of B flat. . .Gerald's comments on the musical habits of the Welsh closely resemble Jordania's descriptions of Georgian singing:
[Both the Welsh and the northern British] have acquired this peculiarity [i.e., singing in parts] not by art but by long usage which has made it, as it were, natural. Moreover, it prevails in both countries and is now so deeply rooted there that nothing musical is performed simply but only diversely among the former people [i.e., the Welsh] and in two parts among the latter [i.e., the north British]. And what is more remarkable, children scarcely beyond infancy, when their wails have barely turned into songs observe the same musical performance [pp. 135-136].
I remember, during my fieldworks in Georgia (mostly during recording sessions at a traditional “supra” feast) I would often realise that all present were contributing to the choral singing (sometimes including myself). So, although this may sound bizarre to some readers, in the societies with a tradition of polyphonic singing there often are no listeners at all, as all the members of the event are actively involved in the music making . . . (p. 10).(Could it be a coincidence that Wales, like Georgia, and so many of the other places in "Old Europe" where polyphonic traditions survive, is a mountainous "refuge" area?)
Gerald's description of the Welsh singing "diversely with many rhythms and tunes, . . . [with] as many different songs and differentiations of the voices as you see heads . . . ," which, for Burstyn, "seems to describe singing in which the active parts are heard as independent from, though presumably coordinated with, one another" could easily be a description of African Pygmy or Bushmen group vocalizing, where everyone present typically joins in with his or her own independent part.
Burstyn considers and rightly rejects various interpretations that associate Gerald's comments with heterophony, singing in parallel parts, etc.:
More plausible is the idea that what Gerald describes is the singing ofBurstyn also considers the possibility that the Sumer canon could be a "learned" (two syllables) composition. "Even if it were possible to prove that Sumer is a composed piece (in the conventional meaning of this term) it would still not necessarily follow that it differed stylistically from contemporary music of oral tradition [p. 139]." Which brings us to the matter of the "corrected" and original versions, as mentioned in my previous post. To judge from the corrected version, which is the one usually reproduced, one might conclude that this is indeed the work of a trained composer, as the upper parts, at least, seem free of the sort of "errors" that can so easily creep into "folk" polyphony. The original version, however, contains exactly this sort of thing -- note for example the parallel unisons between the upper two parts in measure 5. Similarly "awkward" unisons can be found between the 1st and 5th parts in measure 3, even in the "corrected" version. It's useful to note, by the way, that Sumer, while usually described as a four part canon over a two part ground, can, according to Burstyn, "be sung by as many as twelve singers" [p. 149]. In other words, it can be performed as a true round, or perpetual canon, with no real ending, very much in the tradition of well known rounds such as "Three Blind Mice" -- and also in the tradition of Pygmy/Bushmen polyphony, also performed as a continuous flow of interlocking parts, with no end in sight. When performed in 12 parts, by the way, Sumer would indeed contain many "awkward" parallels and duplications, at the unison and octave, of the sort that could be expected in a true "folk" performance, but would be forbidden in an "art" work.
a rondellus, rota, or at any rate some polyphonic technique based on voice
exchange. Assuming this voice-exchange texture to rest on a pes will bring
us still closer to Gerald's description. As Sumer is in fact a four-part rota
over a two-part rondellus pes, this interpretation clearly draws the work
into the sphere of the type of polyphonic singing possibly described by Gerald [p. 138].