Tuesday, August 7, 2007

80. The Power of Cantometrics: 5

Let's now examine the table presented at the end of the last post, on the distribution of interlock and yodel in Africa. Again, I'll remind you to right-click and select "Open in New Window." Those of you who peeked at this yesterday will notice that it's been expanded, as I decided to take as close a look at those groups who do not fit the pattern as those who do.

You'll notice that the Pygmy groups have been divided in two, with the great majority on the first row and only a small 4 song sample from the so-called "Twa" Pygmies in the second. Actually the word "Twa" is misleading as various Pygmy groups have been given that name by various people at various times. What I'm referring to are the "Twa" of Rwanda, who, in the relatively small sample I've been able to find for them, show no signs of either interlock or yodel. They are the only Pygmy group whose music I've heard who (apparently) do not vocalize in this manner.

Not only do all the other Pygmy groups in the sample employ interlock, but, as should be evident from row 1, this mode of group vocalization would seem to predominate (68%). The Ju'hoansi (aka !Kung) Bushmen are also coded with interlock in the great majority of cases (71%). As with the Pygmies, however, not all Bushmen groups vocalize in this manner. I've already discussed the Khwe group (||Anikhwe) included in the sample, who seem to have assimilated to at least some extent with neighboring Bantus, and sing in typically Bantu "call and response" style.

The Mikea and Wayto (or Weyto) each have a row of their own, as it's not clear to me where they belong. Both are hunter-gatherers, but neither is classed with either Pygmies or Bushmen. As you can see, both groups have been coded with interlock (50%), though the Wayto sample is too small to properly assess. Interlock has not been coded for any of the other hunter-gatherer groups in our sample (El Molo, Hadza and Sandawe), though only the Hadza sample is really adequate to date.

The first seven rows represent hunter-gatherers exclusively. The last two enable us to assess the degree to which interlocked vocalizing (and yodel) is found among all other Sub-Saharan groups sampled. As can be seen in row eight, from a total of 873 songs, only 88, or 10%, employ interlock. Row nine represents a subset of the above, all songs from all such groups with at least one instance of interlock. These are the non-hunter/gatherer groups included, in either red or blue, on the map labeled "Interlock & Yodel in Africa" (see previous post). Here we see an important difference between these groups and (nonassimilated) Pygmies and (Ju'hoansi) Bushmen, because interlock is found only 34% of the time among the former, but 68% and 71% respectively among the latter.

Now let's turn to a consideration of yodel, another highly distinctive and also quite rare, mode of vocalizing. It's found among (mainstream) Pygmies 57% and Ju'hoansi 71% of the time, just about as common as interlock in both groups, and 50% among the Mikea -- but not at all among the Wayto, Twa or Khwe groups, nor any of the other hunter-gatherers. (I must add that yodelling has not been found among the Bedzan, the only Pygmy group in our sample, aside from the "Twa," which lacks this very distinctive and characteristic vocal device, suggesting something unusual in their history that warrants further examination.) From rows eight and nine we see that yodel is found in only 5% of our non-hunter/gatherer groups and not much more, 9%, among all such groups using interlock. Clearly, the use of both interlock and yodel is characteristic of most Pygmy and Bushmen vocalizing, yet rare in either Africa or anywhere else.

In post 76, I wrote:
it's not enough to simply identify similarities between one group and another. To establish that such similarities are meaningful, one must also identify significant differences between these groups and all others, then look for patterns based on both. This is where the method Lomax called "Cantometrics" comes in and why it is so important. Using Cantometrics we can search not only for similarities but also differences, seeking out patterns of all sorts, from the strictly local to the most wide ranging, encompassing vocal styles from all over the world.
This is exactly what I've been doing in these last two posts. We can learn only so much from informal listening, as we have no way of knowing about all the many examples we haven't heard, and whether they would be similar to or different from the few we have. Working ones way through the literature isn't much better, because it's all too easy to get lost in all the details and too difficult to keep track of what goes with what and what doesn't. It may seem all too obvious to suggest that some sort of methodology, either Cantometrics or something similar, is needed for meaningful comparisons to be drawn, but the sad truth is that Cantometrics has been rejected and no other methodology has emerged to replace it. I won't comment further on the current state of ethnomusicology in this post.

By using various types of relatively simple, easily understandable statistics, drawn from the Cantometrics database, to assess both similarities and differences, using both the worldwide and African samples, we have been more objectively able, thanks to the power of Cantometrics, to support the claim that there is indeed something very special and unusual about the musical traditions of the Pygmies and Bushmen of Africa.

No comments: