Why did humans evolve to be musical? This question has vexed scientists since Darwin first posed it more than a century ago.
Several theories have emerged over these years – that music evolved like the peacock’s tail, as a sexy signal to potential mates; that music evolved to help groups of people coordinate their actions or to be a more cohesive group; that music is simply a byproduct of other human faculties like language — but all of these theories have had their problems. In their place, Mehr & Krasnow (2017) proposed a new theory of music evolution: that music evolved in humans out of the conflict between parents and their infants over attention.
Unlike the young of most other species, human babies are born pretty helpless and they need the attention of a parent or caregiver to stay safe. That parent has other uses for her attention, however, including paying attention to her other children. Genetically, parents and infants are not expected to “agree” on the optimal amount of attention that parents should give. Sexual reproduction means that we pass on a particular gene to a child only 50% of the time. For a parent, that means that all of her children have the same chance of having that gene. Thus, to maximize their spread, her genes should “prefer” a relatively equal split of attention between her children. But for an infant with that gene, siblings are only 50% likely to have it as well. Thus, that same gene in the infant should “prefer” a selfish share of the parent’s attention. This asymmetry is the root of parent-offspring conflict generally, and shows up in lots of places, from conflicts over sleeping and nursing to sibling rivalry.
But what does this have to do with music? Attention is a covert property of the parent’s mind; a parent can be looking in the direction of their child but be distracted, leaving the child’s safety needs unmet. A good design in the child, therefore, would be an appetite for an honest signal of the parent’s attention. Just like our peahens above, on the prowl for peacocks with good genes, but unable to see those genes, who instead rely on a signal of those good genes (a big flashy tail that only a peacock with good genes could afford to build), an honest signal of attention from the parent should be conspicuous and hard to fake.
An elaborate vocal signal like song ticks these boxes, as a parent singing cannot fake how far away from the infant she is, and couldn’t adjust the song to infant’s ongoing state without actually attending to her state. Moreover, these two mechanisms, the appetite for honest signals of attention in infants and the capacity to produce them in parents, should be under what’s known as arms-race coevolution. Infants with stronger appetites do better and put pressure on parents to provide better signals. Over time, this arms race could start with simple infant-directed vocalizations and end up with something close to modern-day lullabies.
Of course, a theory is only as good as the predictions it makes. Interestingly, this theory seems to have some legs. Follow up work by Mehr, Krasnow, and their colleagues testing the theory has found that some of our genes (specifically on the long arm of chromosome 15) that regulate early life demands on parents (in feeding and sleep) also regulate our appetite for music.
Lullabies also seem to be a human universal — appearing in nearly all of the world’s cultures — and also universally perceivable: taking short clips of music recorded in 86 mostly small-scale societies and playing them to people all over the world, these researchers found that these listeners could identify the lullabies regardless of what culture they came from.
In ongoing work, the team is finding that these lullabies also work regardless of where they come from: US infants are soothed more by lullabies than other kinds of music, even when this music comes from totally unfamiliar cultures.
These findings are described in the article entitled Parent-offspring conflict and the evolution of infant-directed song, recently published in the journal Evolution and Human Behavior. This work was conducted by Samuel A. Mehr and Max M. Krasnow from Harvard University.