How did early man learn to talk? James Hurford discovers that psychologists, anthropologists, linguists and neurologists all have something to say on the subject.
What was the origin of human language? Nobody now believes that any language that we can point to, ancient or modern, was "the first language". The remaining hunter-gatherer societies are "primitive", in the sense that they resemble the societies from which our own must have evolved. As far as we can see, hunter-gatherer languages are just as grammatically complicated as the languages of modern industrial societies. No languages that we know of today (discounting Pidgins, which are cobbled-together codes, not the first language of any society) display any characteristics that can be identified as "primitive". Similarly, no group of humans shows any "primitive" mental traits.
Most historical linguists agree that they can only reconstruct past languages as far back as 10,000 years ago. Languages change so much that the trail fades cumulatively with every century. By the time you try to reconstruct languages beyond 10,000 years ago, there are so many speculative assumptions that any reasonable claim to certainty disappears. The trail in time peters out at a stage where linguists are able to discern about 20 major language families, between which there is not enough similarity to justify any further claims to relatedness.
The question of the origin of language has now become a cluster of different questions that researchers from a bewildering range of disciplines try to answer. They all involve crucial changes that mankind underwent to become fully modern, possessing a full linguistic capacity. To anthropologists, the changes of interest were in social organisation; to neurologists, the changes were in the structure and function of the brain; to psychologists, the changes were in the ability to represent "other minds" and to conceive of meanings relating not just to the here-and-now, but to distant or even hypothetical situations; to formally inclined linguists, the changes were in the ability to build "recursive" mental structures, mirroring the complex hierarchical organisation of sentences; to phoneticians, the changes were physiological, with a two-chamber vocal tract emerging to make modern vowel sounds possible. What exactly were these changes, and when did they happen?
We cannot insist that there is only one answer, or even a single question. Only if one insists on a very narrow view of what language "is" does the question of "its" origin remain, perhaps comfortably, narrow. The fossil record gives us a clear outline of the time-frame in which to locate the crucial changes. Our first bipedal ancestors, the Australopithecines, appeared about three to four million years ago, in Africa, and never moved out of that continent; the next creature to attract much discussion is Homo erectus, who flourished between two million and half a million years ago and got from Africa to Europe and East Asia. And finally, somewhere between 200,000 and 100,000 years ago, we, anatomically modern humans, or Homo sapiens sapiens, appeared. Over all this time, brain size increased dramatically, perhaps in spurts. (Neanderthal man was a close relative of ours, whom we probably ousted from Europe about 35-40,000 years ago.) This outline is blurred by a controversy between palaeontologists and geneticists, known as the quarrel between "recent out of Africa" theorists and "multiregional evolution" theorists. Nobody denies that our ultimate ancestors came from Africa. What is disputed is whether we owe anything genetically to the Homo erectus populations of Europe and Asia of a million years ago. For now, the academic wind is blowing in favour of the theory that the emergence of Homo sapiens sapiens is a relatively recent (150-100,000 years) event, after which our species swept the planet, not mingling with pre-existing Homo erectus. Intermediate views, which hold that there was limited intermingling of genes, and perhaps even of rudiments of culture and language, are possible, but in the absence of data, debate is polarised.
The "recent emergence" theory appeals to many linguistic theorists, who believe that the human capacity for language indicates a single highly significant difference between us and our nearest relatives. The formation of modern humans as a new and distinct species in the course of evolution is consistent with a view of language being the single, genetically determined, trait that distinguishes us most markedly. Clearly possession of language would have given us massive adaptive advantage over rival related species.
Some linguists and neurologists are sympathetic to the idea of a very specific genetic change setting us apart from rival species. Asymmetry in basic body design is rare; most things that we have pairs of, such as lungs, eyes, ears, arms and legs, are symmetrically distributed. And parts that we have just one of, such as genitalia, anus, backbone, mouth and nose, tend to be located symmetrically in the middle of our bodies. So our human asymmetric hand (and foot) preferences are anomalous. Our brains are also physically symmetrical, but their organisation for language, with a strong tendency for the language areas to be in the left hemisphere, is asymmetrical.
Most people are right-handed and control the actions of their dominant hand with the left hemisphere, the side of the brain that also controls language in most people. Half of lefthanders (who control their dominant hand with their right hemisphere) control language from their left hemisphere, too. There is one gene responsible for handedness, although environmental, non- genetic, conditions play a role in determining the handedness phenotype. Conceivably, then, there could be one gene for the side of the brain in which the language areas grow. But this would not imply that there is just a single gene for language, any more than the existence of a handedness gene implies that there is a single gene for all the characteristics of the upper limbs. "What to grow" and "where to put" its controlling site(s) in the brain are separate embryological problems.
Timothy Crow, professor of psychiatry at Oxford University, has reported that people with exactly balanced handedness, for whom it is impossible to say whether they are left or right-handed, are also significantly less able at certain language tasks, again suggesting the intriguing link between handedness and language. Chris McManus, a psychologist at University College London, has theorised that the gene determining manual preference, and eventually lateralisation of language, mutated millions of years ago from the gene that determines the location of another conspicuously asymmetric organ, the heart.
Another linguist who is attracted to a simple genetic explanation for language ability is Myrna Gopnik, who has identified a linguistic disability that runs in families; she set out her views in The THES two weeks ago. Derek Bickerton, also a linguist at the University of Hawaii, claims that there is a single step from Pidgin languages and the kind of communication that language-trained apes, children of 18 months, and language-deprived humans are capable of, which he calls "protolanguage", to full modern language. Bickerton's protolanguage is simply words strung together, with none of the organising grammatical markers (articles, pronouns, conjunctions, auxiliary verbs, etc) characteristic of proper language.
Did Homo erectus have this protolanguage? And was there a change in the brain of the first Homo sapiens sapiens that led to languages as we know them? Even if there was a speciation event some 150-100,000 years ago that gave modern humans their linguistic capacity, it is clear that a great deal else happened without which language would not have been able to become what it is today. Before the supposed event, there must have been a number of preadaptations in the hominid brain, and in the physiology of hearing and noise-making. There must have been social-psychological developments involving altruism, the ability to attribute intentions to others, and to produce imitations of the actions of others. Probably there was also a richly developed, though non-linguistic, way of mentally representing complex situations in the external world.
After the envisaged pivotal genetic innovation, giving humans the capacity for language, it may have initially done us little good at all. The capacity for language enables us to learn language from our elders. Early man would have had no one to learn it from. We had to invent it slowly, an activity involving the hit-and-miss trial of new words and grammatical patterns within communities. About 40,000 years ago, archaeologists notice a sudden dramatic improvement in human toolmaking abilities and artistic prowess, the so-called "upper palaeolithic revolution". Some have been tempted to identify this with the emergence of fully-fledged language.
What were our ancestors doing between 100,000 and 40,000 years ago? They seem to have been genetically adequately equipped for language, yet their material culture gives no sign of the richness we would expect from a language-wielding species. They were painstakingly, and through a cultural process, learning to fulfil their genetic potential.
James Hurford is professor of linguistics, University of Edinburgh.