People all over the world sing songs, and the different songs they sing have certain things in common. It has long been known that songs tend to be constructed from musical scales; that is, from a limited number of pitches or intervals. While there is great variety in song structures the world over, certain patterns of melodic shape (that is, how pitch is patterned over time) tend to turn up again and again. Such similarities are even more evident when we consider folk songs and traditional songs. (That is, the songs meant to be sung by anyone in a group, rather than by musical elites.) Cross-cultural research into folksongs has shown a bias for arch-shaped and descending musical contours within phrases and for small pitch movements between adjacent pitches in a melody, and a tendency for the final note in a phrase to be relatively long. For an easy example exhibiting each of these features, consider the traditional song "Home on the Range."
If songs across the globe and from a variety of cultures tend to have certain things in common, why might this be so? One possibility is that these similarities are "hard wired." That is, they spring from innate predispositions within the human mind. Noam Chomsky is well-known for defending such a view about the similarities that seem to underlie all human languages. Recently a group of researchers has put forth an alternative hypothesis about the origins of song structure, and they have come up with an ingenious way to test it.
In a recent article, Adam T. Tierney, Frank A. Russo and Aniruddh D. Patel suggest that widespread similarities in the melodic shape of songs can be traced to the workings of the body rather than of the mind. The actions that human beings must use to produce different melodic shapes vary in the amount of energy they require. (For example, compare the ease with which most of us can sing "Home on the Range" with the effort required to sing "The Star Spangled Banner" - a song whose musical structure exhibits fewer of the common patterns.) The authors suggest that song structures that require less energy will tend to be favored over those that require more energy. If they are correct, and song similarities originate in motor constraints, then we would expect that non-human animals with similar motor constraints would sing songs with similar melodic features, and that animals with different motor constraints would favor different melodic shapes. Birds make an ideal comparison class, because their mechanisms of song production are both similar to and different from those of humans.
To test their hypothesis, the authors needed to find and compare a large number of human and avian songs. For the human part of the study, they relied on the Essen Musical database, which contains notation for over 9000 songs, including more than 2000 songs from China. Avian songs were a little more difficult to find, but the authors came up with recordings of 80 different songs from as many species. Once the data were assembled, they used specially written software to analyze, first the human songs, and then the birdsongs, examining melodic shape, distance between adjacent pitches, and length of the final notes in phrases. They found, as they had predicted, that both human and birdsongs shared common features, with arch-shaped or descending contours predominating, and a lengthening of final notes relative to the other notes in a phrase. In addition, they found more of a bias for small pitch intervals in human than in birdsong. They attribute this difference to the fact that songbirds have two sets of sound-producing labial folds (as compared to the one set in humans), allowing for pitch leaps without an acrobatic swing in labial tension.
Vocal music is an important part of human culture, with application in ritual and religion, the care of children, as well as entertainment and leisure. Despite this ubiquity (or perhaps because of it?) singing has tended to be taken for granted. The research by Tierney, Russo, and Patel is a valuable corrective. I hope that they (or others) will continue to test the "motor constraints" hypothesis with other animals. Do the cries and duets of gibbons (our primate cousins) exhibit similar melodic structures? What about the songs of dolphins and whales? Learning more about why our songs tend to have the structures that they do should contribute to our understanding of the origins of music and language in human prehistory.
Reference:
Tierney, Adam T. ; Russo, Frank A.; and Patel, Aniruddh D. The motor origins of human and avian song structure. PNAS 2011 108: 15510-15515.
