One of the persistent questions about language origins asks when did it begin. This blog has always favored a very ancient date. Truth is, I suspect the earliest Homo genus had language, in the sense of using joint attention to consider a topic. At that stage speech might have been no more complicated than a modern 24-month-old toddler today. In other words, they could speak phrases, but not full sentences with two nouns and a verb. The evidence for that is two-fold: (1) all healthy, modern humans go through such a developmental phase today, and (2) chimpanzees and gorillas are already smart enough to use language at this level. I’m fully in tune with Michael Tomasello’s argument that apes lack the motivation to use language, not the smarts.
Meanwhile, the elite consensus seems to be that language is no more than 100 thousand years old, which is to say that it appeared a hundred millennia after the appearance of Homo sapiens. Their evidence is that symbols only become a regular feature of archaeological remains at about that time. The many biological adaptations for speech must therefore have evolved for some other purpose. What other purpose might that be? Can’t say, but it must be so.
Bart de Boer recently stepped into this fray. He has been mentioned several times before on this blog because of his interest in the evolution of the vocal tract. He has published a short overview of his work in a paper titled “Evolution of Speech and Evolution of language” (abstract). In this paper de Boer investigates anatomical evidence, particularly the hyoid bone, the position of the larynx, and the thoracic vertebral canal. The hyoid bones in human throats differs from that of great apes because apes have air sacs that enable impressive bellowing but limit sound differentiation, while humans do not have such sacs. Neanderthal hyoids are like humans, suggesting that they too lacked the air sacs. The position of the larynx helps make clear vocalization. Whether Neanderthals had such a larynx is debated, but de Boer’s own work persuades him that Neanderthals too had a lowered larynx. Finally, the thoracic vertebral canal (also known as the spinal canal) refers to the interior of the spine that carries nerves along the chest. Humans have larger canals than apes, probably because they feed more neurons into the chest, allowing for more precise control of the outbreaths used to support speech. It turns out that Neanderthals have these enlarged thoracic vertebral canals as well.
None of this evidence is conclusive, but taken as a whole they give a consistent pattern supporting ancient vocalizations, and de Boer concludes that at least by the time of the last common ancestor of Neanderthals and humans (i.e., at least half a million years ago) our lineage had evolved the ability to produce and control the sounds needed for modern speech.
Does that mean our ancestors so far back actually talked? De Boer offers three choices: (1) language co-evolved with our ability to control vocalizations; (2) first we developed the ability to vocalize and then introduced language; or (3) first we developed language and then got the ability to vocalize.
If choice 2 is correct, we evolved the ability to vocalize for some reason other than speaking. A popular notion is that we developed the ability to sing tra-la-la type songs first. If choice 3 is correct, we developed the ability to think linguistically first and then externalized it. Chomsky and company are associated with the idea that language began internally, but they do not like the old date this gives for externalization. De Boer rejects choice 2 on the grounds that if we were smart enough to sing, we should have been smart enough to talk. “It seems more parsimonious to propose that language and music evolved together and are more likely two sides of the same coin.” De Boer dismisses the Chomskyan idea out of hand, but gives more attention to the idea that we might have first used a gestural language and then switched to vocalization. He admits it is not impossible, but notes that “evolution tends to stick with solutions that work well.” In other words, if we had first developed a sign language, we probably would have stayed with it even if speech might be a better way.
This reasoning leaves us with choice number one: language co-evolved with our ability to control our vocalizations. Added to this view is that singing and gesturing came right along with words and grammar. So we can imagine a Homo community of 500 thousand years ago, talking, singing and gesturing together.
What was the language of that common ancestor like? We do not need to suppose it was like modern language. The brain continued to evolve, offering the interesting possibility that Neanderthal and Homo sapiens languages might have had some critical differences and were not fully translatable.
De Boer argues that some anatomical and cognitive co-evolution is almost inevitable. For example, the difference between vowels tends to be maximized automatically. This difference is the result of anatomy, but then there will be some pressure to reliably articulate these differences, leading to adjustments in the nervous system furthering control of the sounds produced. This further control leads to still more self-organization of sounds, leading to more nervous system adjustments.
It sounds like a classic virtuous cycle, but de Boer does say he thinks the initial trigger was cognitive. Skills like vocal flexibility and imitation, even at a very crude level, had to be there first. In fact, we see these things today in babbling babies who make different and imitative sounds before they start speaking words.
The kind of dramatic mutation favored by Chomsky and melodramatic playwrights everywhere looks to be a naïve fantasy. It may be true that modern language of the metaphorical, myth-making, blah-blah we know today is only a hundred thousand years old, or even younger, but the cooperative contemplation of topics has been co-evolving for a much longer period of time.