Infant-directed song might later have evolved into the more complex forms of music we hear in our modern world, researchers said.
Music is a tricky topic for evolutionary science: it turns up in many cultures around the world in many different contexts, but no one knows why humans are the only musical species.
"There has been a lot of attention paid to the question of where music came from, but none of the theories have been very successful in predicting the features of music or musical behaviour," said Max Krasnow, from Harvard.
At the core of their theory, Krasnow said, is the notion that parents and infants are engaged in an "arms race" over an invaluable resource - attention.
More From This Section
"Particularly in an ancestral world, where there are predators and other people that pose a risk, and infants don't know which foods are poisonous and what activities are hazardous, an infant can be kept safe by an attentive parent," he said.
Researchers predict that children should 'want' a greater share of their parents' attention than their parents 'want' to give them.
But how does the child know it is has her parent's attention? The solution is that parents were forced to develop some method of signalling to their offspring that their desire for attention was being met, said Krasnow.
"But I could be looking at them and thinking of something else, or looking at them and focusing on my cell phone, and not really attending to them at all. They should want a better signal than that," he said.
Signals in form of a song is more honest due to the cost associated with them - meaning that by sending a signal to an infant, a parent cannot be sending it to someone else, sending it but lying about it, researchers said.
The research was published in the journal Evolution and Human Behaviour.