Cantometrics Revisited

Unknown-1The late American musicologist Alan Lomax’s attempt in the 1960s to find patterns by comparing the ways in which people sing in different folk traditions around the world — “Cantometrics” — was widely discredited by ethnomusicologists in the ensuing decades.

I’ve been learning about the Dutch geneticist Armand Leroi’s more recent work — which he calls “Neo-Cantometrics” — on applying the rules of evolutionary biology to Lomax’s theories. Leroi uses an algorithm known as “Grouper” and large amounts of computing power to hunt for patterns in the data sets containing just over 5000 folk songs from all over the globe collected by Lomax.

It’s fascinating stuff. In this 2007 video presentation, Leroi uses maps and song samples to illustrate similarities between the genetic makeup of geographically disparate groups of people and the vocal traits inherent in their singing.

I do feel somewhat skeptical of  Leroi’s ideas. The findings in his paper on the subject are inconclusive. And he is quick to dismiss musicologists’ criticisms of Lomax’s work on the grounds that humanists are “frightened of numbers,” which is ridiculous.

Still, I’m curious about how I might apply some of Leroi’s research to work that I am doing on the evolution of singing culture. If there are patterns in the ways people use their voices to sing, then are there also patterns in the participation and consumption habits around vocal music? And how have these evolved over time in different parts of the world?

P.S. Leroi has conducted another interesting project around vocal music, in which he created an experiment which draws on Darwinian principles of natural selection to “compose” the “perfect song.” The DarwinTunes project received quite a bit of media attention last year. An interesting 30-minute radio segment about this work, which aired last year on the UK’s Radio 4, can be heard here.

Share on FacebookTweet about this on TwitterShare on RedditEmail this to someone


  1. says

    Hi Chloe. I found this website via a Google search and was pleased to learn of your interest in Cantometrics. I too have been curious to learn more about LeRoi’s Cantometric research. The paper to which you refer is actually a response to a paper of mine, published in the previous issue of World of Music, and you are right, it is inconclusive. It lays out a research project that, as far as I can tell, was never actually carried out, or at least none of the results were ever published. The video to which you refer draws heavily on the work of both Lomax and myself, yet he never mentions me at all. So forgive me if I don’t trust this guy.

    As Lomax’s collaborator on Cantometrics I consider myself in an excellent position to evaluate his work and I must say I have for many years had mixed feelings. I see Cantometrics itself as a valid method, and have in fact used it extensively in recent years. However, the evolutionary scheme Lomax came up with, based on Cantometrics, is riddled with problems and it is really this, and not the methodology itself, that was imo the principal reason for its rejection.

    I neglected both Cantometrics and ethnomusicology for many years, but have once again gotten involved, and have in fact written a book based on my experiences in this field. It might interest you and is freely available on the Internet:

    I’m pleased to learn of your interest in the voice and I’m curious to learn more about your research.

    • Chloe Veltman says

      Thanks Victor for being in touch. Your work sounds fascinating and I will check out the link you sent me as soon as I have a moment!

  2. says

    Like some kinds of abstract physics, there are theory’s so profoundly aesthetic you hope they will be proven out of a desire to see a correlation between truth and beauty. That’s why I really like Lomax’s concept of cantometrics. Even if it is reductive or overly general, there is something very beautiful about its ideals.

    Another reason I like cantometrics is that I believe that we sing our worlds into being. Through art we formulate our human identity. Cantometrics attempts to turn that concept into science, even if it seems a hopeless task.

    In the early 90s I was visiting a musicologist friend, Andy Kaye, in New York. He took me down to a warehouse where he and Roswell Rudd were working on Lomax’s Global Jukebox. It was a set of CD-ROMs based around a map of the world. One could click anywhere on the map and listen to the region’s folk music. The geographic evolution of musical styles and their correlations to other cultural attributes seemed so clear. Musical styles moved across the map like the gradually shifting colors of a prism. One could easily hear, for example, that the music of Siberian hunters was closely related to Native American music. Profound clues to human history and identity seemed to be revealed through song.

    Cantometrics can only be substantiated through the collection and classification of enormous amounts of data that define song. Lomax used 37 attributes of singing which have to be applied to tens of thousands of field recordings. I think cantometrics was ahead of its time, because it is only through computers that we are learning how to manage that amount of data. And we are only beginning to learn how much such mega-data can reveal. I hope we haven’t heard the last of Lomax’s concepts and ideals. Surely theories so beautiful have to have truth in them somewhere.

  3. says

    Nice to see this interest in Lomax’s work! We just published an article related to these questions today looking at correlations between music and genes among the indigenous peoples of Taiwan using Cantometrics and related techniques (it’s available free online for the rest of the month here:

    There was also some press coverage here: