|Babies exposed to sign language babble with their hands, even if they are not deaf. The finding supports the idea that human infants have an innate sensitivity to the rhythm of language and engage it however they can.
Everyone accepts that babies babble as a way to acquire language, but researchers are polarised about its role. One camp says that children learn to adjust the opening and closing of their mouths to make vowels and consonants by mimicking adults, but the sounds are initially without meaning.
The other side argues that babbling is more than just random noise-making. Much of it, they contend, consists of phonetic-syllabic units - the rudimentary forms of language.
Laura-Ann Petitto at Dartmouth College in Hanover, New Hampshire, a leader in this camp, has argued that deaf babies who are exposed to sign language learn to babble using their hands the way hearing babies do with their mouths.
Petitto believes that the hand-babbling is functionally identical to verbal babbling - only the input is different. But critics counter that deaf children cannot be directly compared with their hearing counterparts.
Now Petitto and her colleagues have tested three hearing babies who, because their parents are deaf, were exposed only to sign. Three control infants had hearing, speaking parents.
To analyse the hand movements of the six children, the researchers placed infrared-emitting diodes on the babies' hands, forearms and feet. Sensors tracked the movements of the babies' limbs as they engaged in a variety of tasks, including grasping for toys and watching two people communicate.
Petitto reasoned that if her opponents were right, then what the babies did with their hands would be irrelevant - and indistinguishable. Instead the team found that the two groups had different hand movements.
Sign-exposed babies produced two distinct types of rhythmic hand activity, a low-frequency type at 1 hertz and a high-frequency one at 2.5 hertz. The speech-exposed babies had only high-frequency moves.
There was a "unique rhythmic signature of natural language" to the low-frequency movements. "What is really genetically passed on," Petitto says, "is a sensitivity to patterns."
But Peter MacNeilage, of the University of Texas at Austin, is not persuaded. "She makes a blanket statement that there is an exact correspondence between the structures of speech and sign," he says. "But there is no accepted evidence for this view at the level of phonological structure or in the form of a rhythm common to speech and sign."