Music really is a universal language, study suggests

There is a communicative power in music that runs deeper than words. In fact, as the German poet Heinrich Heine once wrote, “where words leave off, music begins.”

It has the power to stir up nostalgia; make the hairs on the back of your neck stand on end; and even make you cry or laugh. But just how universal is the language of music globally, to convey emotions and stir up specific feelings as it crosses cultures?

To answer this, researchers from Harvard University selected 86 songs from small-scale societies scattered throughout the globe. It asked 750 web users, in 60 countries to listen to 14-second excerpts from a random selection of these songs and the listeners had to evaluate whether they thought the songs were a lullaby; a love song; used for dancing; used to heal an illness; used for mourning; or used to tell a story.

After collectively listening to a total of 26,000 excerpts, the results – published in the journal Current Biology – showed accurate inferences about a given song’s function. Regardless of whether the music came from Highland Scots or Iroquois societies, the listeners were able to collectively decide what purpose a song has.

The clearest results came for lullabies and dance songs, which were the easiest for the majority of listeners to identify. Love songs and healing songs proved harder to categorise, but the collective conclusion still put these in their correct groupings.

There were actually no songs in the roster that were used for mourning or to tell a story, and these options were only added to stop listeners from assuming there were only four categories. Somewhat ironically, listeners tended to categorise healing songs as songs for mourning the dead.

Alphr asked the study’s lead author, Samuel Mehr, what the findings suggest about the commonality of human experience of music:

“I think they show quite clearly that despite the fact that the songs of the world are dazzlingly variable, with each culture shaping a great deal of the way its music sounds, at the same time, our shared human psychology leads to some very clear regularities underlying all that variance,” said Mehr.

“Nearly everybody seems to find dance songs ‘dancey’ and nearly everybody seems to find lullabies ‘soothing’ — but that doesn’t mean that all dance songs sound alike and all lullabies sound alike.”

Mehr continued that the song functions were chosen for two reasons. Firstly, work in anthropology and ethnomusicology shows these types of songs are found in many cultures. (“So they’re good candidate genres for looking at possible form-function links.”) Secondly, they are each related to theories of the biological and cultural evolution of music. Could there be underlying biological reasons for the common rhythms in music?

“Absolutely! We don’t yet know what they are, but there’s really cool work coming from evolutionary biology and psychology that develop testable theories of what those biological reasons could be,” said Mehr.

“For instance, in some of my other work with [Harvard Evolutionary Psychology Laboratory associate professor] Max Krasnow, I’ve proposed a formal explanation for the origins of infant-directed singing. Studies like ours can be used to test the predictions of those theories and help to rule them in or out.”

A follow-up experiment from the same researchers involved asking 1,000 people in the US and India to rate the excepts for contextual features such as the number of singers and instruments, and subjective features like melodic complexity, tempo and valence. While the results hinted at some relationship between those judgments and song function, they didn’t go far enough to explain exactly why people were able to correctly detect a song’s function.

To truly examine the universality of these responses, more work is need to test the reactions of listeners who haven’t had any experience of music from other cultures. Mehr said that the paper’s co-authors, Manvir Singh and Luke Glowacki, are “already leading field expeditions to isolated, small-scale societies, where they are running a new version of the same experiment to test the extent to which the current findings generalise to *all* humans, rather than just English-speaking internet users.”