Editing Cantometrics (section)

==Criticisms and responses==
The reaction of musicologists to Cantometrics was complex, as some critics questioned whether one could ever have enough statistics to prove anything about music and culture. One, Richard Middleton, called it an example of sociological [[homology (sociology)|homology]].

The musical examples for Cantometrics had not been chosen randomly, however, but for their representativeness, following the scholarly guidance of specialists who had studied the regions and /or supplied the audio or film clip samples. This was done because it was found that style tends to be very repetitive, and in most instances relatively few examples captured the stable performance norms, so that in most cases, coding many examples per culture yielded little new information. Lomax and Grauer settled on about ten examples per culture for Cantometrics, which had the largest dataset. Ten examples per culture usually sufficed, although for some cultures one or more sub-styles needed to be sampled as well.

Alan Lomax himself stressed repeatedly that completeness of sampling was not the point:
<blockquote>The adequacy of any of these samples is . . . subject to this test: Will another sample of a similar kind taken in the same culture produce a similar performance profile? From this point of view, I believe that the majority of our samples will hold up. Even "secret" songs generally tend to be stylistically close to the more familiar music of a culture. The truth is that, with any one culture or subculture, singing is a rather standardized kind of behavior. It must be so since a main function of song is to . . . permit groups of performers to vocalize together and their listeners to share in a common experience. Cantometrics is a study of these standardized models, which describe singing rather than songs. Therefore, it is not primarily concerned with "complete" collections and descriptions, as are most scholarly endeavors, but with locating verifiable regularities and patterns. (Lomax [1968] 2000).</blockquote>

Victor Grauer noted that: “Cantometrics is a statistical method.  Its patterns and correlations are based on multiple instances, not just one.  There will always be exceptions, which is why statistical methods are generally preferable for comparative studies than the close examination of individual examples.”<ref>[http://www.mustrad.org.uk/articles/cantome2.htm Victor Grauer, "Song and Social Culture, A Response",  ''Musical Traditions'' web magazine, July 27, 2005.]</ref>