Indulge me for moment, and have a listen to music clip below. Before you read on, take your best guess about what kind of song this is. Is it a love song, a song for dancing, or a lullaby?


If you guessed that this song from Australia’s indigenous Yalngu people is song for dancing, you’re right. In fact, a recent study from Harvard University’s new Music Labsuggests that we humans are quite good at picking out dance songs — no matter what part of the world we come from or what part of the world the song comes from.

Looking into The Music Lab at Harvard.
Edgar B. Herwick III/WGBH News

Why, exactly, are we good at it? That’s what the team at the Music Lab wants to find out.

"Not only do we have no idea why music exists ... we don’t have a good evolutionary explanation for where this behavior comes from," said Dr. Samuel Mehr, a cognitive scientist and musician who runs the Music Lab. "We don’t even have a lot of the basic facts about the behavior that we would need to evaluate any theory."

Using techniques ranging from old fashioned field work to 21st century data science, the Music Lab’s small, interdisciplinary team — which includes psychologists, musicologists and evolutionary biologists, to name a few fields represented — has undertaken the daunting task of starting to assemble those basic facts. 

"It’s all called The Natural History of Song," said Manvir Singh, an anthropologist and Ph.D candidate who works in the lab. "Which is, like, a sweet title in our opinion."

Their ambitious Natural History Of Song Project consists of two parts. No. 1: Old school. Collect as many field recorded songs from as many cultures in 30 different regions of world regions as possible, and assemble a huge discography of music to work with.

"And from every region we have a dance song, a love song, a lullaby, and a healing song," explained Singh. 

No. 2: New school. Assemble a massive database of every passage of text in every anthological paper ever written that deals with music. Code each one in a variety of categories, so they can search for patterns and commonalities.

"And we get just the raw text that they write, because just the text itself can be informative, whether you are a human reading it or a machine," said Mehr.

The hope is that the discography and database will be a wellspring for scientific studies that can inch us closer to a more fundamental understanding of why we make music.

The team’s first study, recently published in the journal Current Biology, starts small. They recruited 750 people online, from 60 different countries. The team played them clips of dance songs, lullabies, love songs and healing songs from their discography. The overarching question of the study? 

Dr. Samuel Mehr (left) and Manvir Singh.
Edgar B. Herwick III/WGBH News

"Will [they] be able to infer function from form? Will [they] know what this music is used for just from kind of naively and blindly listening to the tracks?" said Singh.

In other words, despite not knowing the song, people may be able to identify what each song was used for. What did the team learn? It turns out, people everywhere are pretty darn good at identifying both dance songs and lullabies.

"So, someone in Tennessee has a similar notion to how a dance song should sound or a lullaby should sound to some random woman in rural Punjab," said Singh. 

People were also reasonably good at picking out the healing songs, too — something that surprised the team. Love songs, on the other hand? Not so much.

"Maybe what makes a love song isn’t its musical features," said Singh. "Maybe all of it’s features make it just a song about it’s lyrics, a song that’s telling a story, and instead it’s the lyrics that are doing the work."

Or maybe it’s that while all cultures appear to have love songs, what constitutes love is unique across cultures, so the music itself varies.

It’s questions like these — and countless others — that the Music Lab team is aiming to put us in a position not just to ask, but to answer.

"People think of science, they think of a guy in a white lab coat and, like, test tubes, or whatever," said Mehr. "But this is also science. Things like the arts, music, visual arts, dance. These are topics that the sciences have something to say about. That’s one of the reasons we’re so excited about this work, is that this is kind of new."

Something new that might one day show whether that old adage — that music is a universal language — is scientifically sound.

If there is something that has recently piqued your curiosity, let me know about it. Email me at I have to get my story ideas from somewhere — why not from you?