This is a variation of the social-cohesion hypothesis
. The social-cohesion hypothesis, succinctly stated, is that music is a characteristic that builds bonds within a community. The bond between parent and child is one of the strongest, and some researchers have studied the role of lullabies in facilitating this bond
. Apparently, lullabies across the world share similar features that include simple structure and falling contours. So far, it seems that scientists have only studied mother-child bonding and lullabies, but the results
indicate that maternal singing is an adaptive function that allows female humans to soothe their babies without having to touch them. Apparently, this could have been adaptive way back in the day when humans were hunter-gatherers--women could put down the infants and soothe them by singing to them while they gathered water or berries or twigs or whatever. Again, though, the problem with this theory is that it's pretty gendered and relies on outdated understandings of early hunter-gatherer human behavior. While some tribes relied on a strict division of labor, with men hunting and women gathering, there is a lot of evidence that some tribes didn't
do that. The women of the Agta people, for example
, hunt with bows and arrows and are accompanied by dogs. When they do that, maybe the menfolk watch the babies and sing lullabies. Also, hunting is the primary form of subsistence only at certain latitudes. In some societies, it is both men and women who engage in gathering, hunting and all of the activities
that go toward feeding the tribe.