Fotb Maggie Jochild has reminded us of a study that was published in Science in April and publicized in The New York Times. Biologist Quentin D. Atkinson applied mass comparison methods familiar in genetic research to the analysis of phonemes, the sounds that languages use to distinguish one word from another. If a geneticist found the same pattens in a set of single-nucleotide polymorphisms that Atkinson found in the phonemes of the world’s languages, that geneticist would likely conclude that the set represented descent from a single common ancestor. So, Atkinson suggests that all languages known to us, both those currently spoken as mother tongues and ancient languages known to us only through writings, descend from a language spoken in Africa tens of thousands of years ago.
When Atkinson’s paper appeared, it attracted strong criticism from some of our favorite writers. To mention only the two posts about it that appeared on the incomparable Language Log, Mark Liberman analyzed Atkinson’s methodology, and Sarah Thomason reproduced a letter that anthropologists Ives Goddard and Bruno Frohlich sent to Science, but which that journal refused to publish. Professor Liberman is lightly skeptical of Atkinson’s methods, but optimistic that his conclusions might prove true, while Professors Goddard, Frohlich, and Thomason express more severe reservations.
For my part, I’ve often wondered how many languages were spoken when language was first spoken. For various reasons, it would be interesting if the idea of “a language” were familiar to the first generations of language speakers. The New York Times article says that Atkinson’s work “implies, though does not prove, that modern language originated only once, an issue of considerable controversy among linguists.”
I don’t think it does imply that, actually. Even if it were proven that all languages available to us for study shared a common ancestor, it would not thereby be proven that “language originated only once.” That language may well have been one of many languages spoken at the time, and may well have been descended from other languages spoken many millennia earlier. Consider the languages of ancient Italy. Latin is the ancestor of many languages spoken around the world today; the dozens of other languages that flourished on the peninsula before the rise of Rome have left no descendants. If it weren’t for the written evidence that has come down to us from antiquity, we might be tempted to conclude that language originated in Italy only once, and in Latium. Extending the processes that singled Latin out as the only Italian language of the mid-first millennium BC to leave descendants back another 50,000 years to a time when all of our ancestors may still have lived in Africa, we can see that they may well have left only one language spoken at that period with descendants, even if at the time it had thousands of contemporaries.
Indeed, that might have happened even if that hypothetical mother tongue had not stood out among its contemporaries. To resume for a moment the parallel with Latin, before the late fifth century BC it would have been a rare observer who would have guessed that varieties of that language would grow into languages that would continue to be spoken for thousands of years after the language of the far wealthier and more powerful Etruscans had died without issue. Going back further, the people who spoke the language that was the common ancestor of the Anatolian and Indo-European language families probably lived sometime around the year 4000 BC. They were likely a rather scruffy group of nomads who lived in some of the less desirable corners of the Black Sea coast, what Moe the Bartender would no doubt call “one of them loser countries.” * Even the early Romans would have seemed impressive set against them. Yet something like half the people in the world now speak one or another of the hundreds of languages that descend from theirs. If such an undistinguished group can launch a language that crowds out so many of its contemporaries over a period of 6,000 years, surely there is little we can say about the career in the world of a language that is likely 1o times that old.
*For an excellent introduction to what we know of the early Indo-Europeans and their predecessors, see David W. Anthony’s The Horse, the Wheel, and Language (Princeton UP, 2007.) The author of the New York Times piece would have benefited greatly from reading Anthony’s book; he cites a 2003 study by Atkinson which argued that “Indo-European was much older than historical linguists had estimated and hence favored the theory that the language family had diversified with the spread of agriculture some 10,000 years ago, not with a military invasion by steppe people some 6,000 years ago, the idea favored by most historical linguists. ” Anthony treats of the theory that Indo-European came out of southwest Asia with agriculture in pages 75-82 of his book, and his analysis is devastating. While the languages we have available for study in historical time change ceaselessly and are as a rule considerably more diverse than the material cultures of the peoples who speak them, this theory requires proto-Indo-European to have entered Europe sometime shortly after 7000 BC, to have spread over that continent, and to have remained unchanged for 3000 years while an extremely wide variety of material cultures flourished there. This scenario is so remote from any attested elsewhere that it would qualify as a miracle. Moreover, all branches of the Indo-European family of languages include vocabulary descended from a rather large group of words related to wagons. The archaeological record makes it clear that wagons cannot have existed in Indo-European speaking areas much before 3500BC; the early entry hypothesis therefore requires that the Indo-Europeans, at a time when they had been separated from each other for three and a half millennia, simultaneously adopted the same vocabulary for this new invention. Again, if Atkinson’s method compels this conclusion, Atkinson’s method is reduced to absurdity.
As for the “military invasion” hypothesis, it is not “the idea favored by most historical linguists,” as Anthony makes clear at several points. On pages 117-120, he introduces the first of a series of models of language diffusion developed by anthropologists working in contemporary African and Asian societies that have once and for all displaced the nineteenth century vision of the warlike Aryans, the “blond beasts” whose image has caused so much more harm than good. These models are a major theme of Anthony’s book, and in fact allow far less significance to military invasions than does the early entry hypothesis.