How the mouth took speech away from busy hands

Did our language evolve from hand gestures, and did we talk our fellow hominids to death? Michael Corballis looks at the evidence.

It seems natural to suppose, as Charles Darwin did, that human language evolved from animal cries. Primates can be noisy, and even parrots can give a fair imitation of human speech. Yet linguists generally agree that animal cries bear little if any resemblance to true human language. They are tied largely to emotional events - such as danger, territorial threat, the discovery of food or mating rituals - and have little of the variety and unpredictability of human language. We humans have the unique capacity to combine sounds to create an infinity of possible meanings.

Perhaps, though, we might understand the origins of language better if we suppose that it evolved not from vocal cries but from manual gestures. Given that we are such a vocal species, the idea may seem preposterous. But if you watch people as they talk, you will see that spontaneous speech is nearly always accompanied by expressive gestures of the hands and face. Moreover, deaf people all over the world have spontaneously developed signed languages that have all the grammatical and lexical sophistication of spoken languages. Language is not tied to the vocal medium.

Careful observations of great apes in the wild reveals that they gesture freely to one another, and their gestures have more language-like properties than their vocalisations. Attempts to teach our closest relatives, the chimpanzees and bonobos, to speak have proved fruitless. Much better progress has been made toward teaching them to communicate with a simplified form of sign language or by pointing to visual symbols on a keyboard. The communication developed in these ways still has nothing like the grammatical sophistication of human language, but the animals are at least able to build up a vocabulary of several hundred "words".

Chimps and bonobos have sophisticated visual systems and the ability to move their limbs flexibly and voluntarily, whereas their vocalisations are relatively inflexible and emotionally bound. We now know, too, that the area in the frontal lobe of the monkey brain corresponding to Broca's area in the human brain is involved with making and perceiving hand and arm movements, whereas Broca's area in humans is one of the primary speech areas. At some point in primate evolution, this area shifted its allegiance from hand to mouth.

This idea that language may have sprung from manual gestures is not new. It was proposed in the 18th century by the philosopher Condillac and has been restated intermittently ever since. But it has not found favour, perhaps because of a big stumbling block. If manual gesture provides so obvious a medium for the evolution of language, how and why did vocal speech take over? Is it not more parsimonious to suppose that language evolved from animal cries?

One way to address these questions derives from the view, developed by the late Alvin M. Liberman and his colleagues at the Haskins Laboratories in the United States, that speech is better conceived as a series of articulatory movements than as a sequence of sounds. This can explain why parts of words that look similar are quite different acoustically, but result from similar vocal gestures. Moreover, what we hear is quite strongly influenced by what we see, as a good ventriloquist can demonstrate.

The evolutionary scenario may have run like this. The early hominids, originating some 6 million years ago, were distinguished from their ape cousins by their bipedal walking. This would have freed the hands, perhaps leading to the development of more sophisticated manual communication. But manual communication would have competed with other manual activities, such as throwing, carrying and the subsequent manufacture of tools. This led to a gradual shift from manual to facial gesture, which may have accelerated with the emergence of stone-tool manufacture about 2.5 million years ago.

Facial gesture could then have been augmented by the addition of vocal sounds. Adding voicing can create the distinction between voiced and unvoiced consonants, thereby increasing the repertoire of possible gestures. Voicing can also provide access to gestures produced at the back of the mouth or tongue, which are inaccessible visually. Speech is facial gesture half swallowed, and with added sound effects. Once sound was incorporated into gestures, other advantages of vocalisation would emerge, such as the ability to communicate in the dark. Beginning perhaps with Homo erectus some 2 million years ago, natural selection may have led to the gradual alterations to the vocal tract, and to the more sophisticated control of breathing and vocalisation, that resulted eventually in vocal speech becoming the dominant mode.

Autonomous speech may even have been ultimately a cultural invention - although we have not completely lost the manual component since people gesture even when talking on the phone. Until some time in the past 100,000 years, though, language may have continued to depend partially on manual and facial gestures. There is now evidence that it was people migrating out of Africa from some 50,000 years ago who eventually replaced all other hominids. Could it be that their principal advantage lay in the ability to communicate purely vocally, finally freeing the hands for the explosion of technology and manufacture that followed? Did we in effect talk our fellow hominids out of existence?

Michael C. Corballis is professor of psychology at the University of Auckland. From Hand to Mouth: The Origins of Language is published this week by Princeton University Press, £19.95.