Does music carry universal meaning?

30 January 2018

Interview with

Dr Samuel Mehr, Manvir Singh - Harvard University

Music from different cultures can sound extremely different from one other: a Scottish ballad may not have too much in common with an Australian healing song by the sound of it. But is there something intrinsically in music that crosses cultural divides to tell us what the music was originally intended for? Georgia Mills spoke to Samuel Mehr and Manvir Singh from Harvard University...

Samuel - We started by working with a project called the Natural History of Song project. And that’s a project where we’re basically taking modern techniques from the cognitive sciences, from data science, from a few other fields to try and systematically characterise different features of music and musical behaviour from around the world.

So, basically, we took this dataset that we had built for this bigger project which was a systematically built collection of four different kinds of songs from 86 small scale societies around the world. They’re dance songs, lullabies, healing songs and love songs.

We took this sample; we’ve got 118 different songs split across those genres. We randomly selected a very short snippet of each song - 14 seconds long, and then we took all those snippets and we played them to people all over the world. People who lived in 60 different countries and who were recruited on the internet.

All we did is that for each little excerpt that they listened to we asked them a series of questions. We said tell us if you think that the singers definitely do not use the song for such and such or definitely use the song for such and such? In these questions such and such was something like use the song for dancing, or use the song to soothe an infant, or to heal illness? So the question is: can these naive listeners on the internet who’ve never heard these songs, they don’t know anything about these cultures, can they tell what a song is for, what it's function is on on the basis of its musical forms - what’s actually there in the recording?

Manvir - My name is Manvir Singh. I’m a PhD student in human evolutionary biology, also at Harvard University. What we found was that people seemed to be very very good at identifying dance songs. The effect for people rating dance songs as being used for dancing compared to other songs was huge. People are also really good at identifying lullabies. People are okay at identifying healing songs, so there was an effect but that was smaller than the effects that we saw for dance songs and lullabies. Then we found that people are actually not able to identify love songs. Kind of relatedly we also asked two extra questions on the side, so is it used for mourning the dead and is it used for telling a story? Interestingly, people rated healing songs as especially high on the dimension mourning the dead. Although they couldn’t recognise love songs as being used to communicate love, they did rate love songs as especially high on the dimension being used to tell a story, which is kind of interesting finding that maybe has implications for what love songs are and how they do communicate love.

Sam - There was pretty much strong agreement among experts in the field that we were not going to find the things that we did actually find in the study. It was the standard view was music is culturally produced, it doesn't have things in common with one another across cultures. Each culture has their own interesting, unique musical idiom on the world and our study is not saying that that’s wrong. Our study is saying that even though it’s really clear that from culture to culture there’s incredible variation and insane differences in how dance songs are used, and what they sound like, and there’s all sorts of differences. Underlying all of that really interesting variance is the simple fact that these songs are all produced by human minds, and human minds have features in common with on another. Just because there’s a lot of really interesting variance doesn’t necessarily mean that cultures don’t have things in common with one another.

Manvir - We definitely don’t know from our study why there are dance songs and why there are lullabies. But one thing we do know is that people around the world respond to a certain string of musical stimuli in a similar way. People around the world have minds that when they hear some certain kind of stimulus they want to dance, they want to move their body. And, similarly, people around the world have minds that respond to a very very different kind of sound by being calmed and falling asleep. Why humans have those minds is an open question, but our study at least kind of pushes us forward in showing us that we do seem to have those minds. Kind of relatedly, and much more weird, people seem to have minds that in response to certain kinds of musical stimuli seem to think they’re being healed spiritually, which is more more bizarre and fun question.

Song information - in order of appearance:

- Love song, Scottish Highlands, 1953
- Love song, Q'ero Quichua - Central Andes, 1964
- Dance song, Chuuk - Micronesia, 1984
- Love song, Rwanda, 1952
- Healing song, Otavalo Quichua - Central Andes, <1960
- Lullaby, Nyangatom - Eastern Africa, 2012


Add a comment