Inquiring into language’s origins can seem like a quixotic adventure, but it does bump you into the heart of the classical humanist question of what makes a human human. Two issues in particular stand out:
- How much of language is cultural and how much inborn? This is a variant on the widespread dispute about nature or nurture, but it provides a specific focus. It does not seem unreasonable to think the question might have an answer.
- Why is language so different from other animal communications? Animal signals are nothing like sentences, either in semantics, syntax, or vocabulary. Other animals do not discuss topics together. They do not construct a body of lore that can be useful at some later date. They cannot call the dead to mind by the simple expedient of speaking a name. Yet, how can we square the differences being so great with the Darwinian premise that species do not evolve powers they do not need to survive? A prey animal needs to evolve the ability to outrun its predators by only a little bit. No need to run at 100 miles per hour when the fastest predator can only go 50.
This blog’s answers to these questions run: (1) language is cultural, but it must work with an inborn perceptual apparatus; and (2) ‘twas continuous competition with other groups of speakers that kept pushing us to greater and greater powers.
The dominant answer in America’s linguistic departments is: (1) all essential elements of a language are inborn; and (2) human language benefits from an unlikely mutation that gave us language’s full power in a single bound.
I’m used to being in the minority, but I don’t necessarily like it. So I am glad to see a brief paper—Bart de Boer, Modeling Co-evolution of Speech and Biology (abstract)—that addresses both these issues. In this paper de Boer asks whether it was even possible to evolve a sensori-motor system that supports the needs of language production and perception. Not to keep readers in suspense, he answers yes. The pressures of language usage can lead to biological changes, and the biological changes can expand usage.
This sort of circular evolution is called co-evolution and not everybody has been convinced that the co-evolution of language and body is possible. Previous work on co-evolution has found that the biological changes must be the result of functional pressures, not ones produced by arbitrary drift in usage. That's because a language's arbitrary features change too fast to become biologically fixed. However, even with functional selection the effect is only small. Since language is very different from animal communications, the small effect discovery seems a problem.
De Boer hypothesizes that, as in my answer #2, small effects can evolve into big differences if the change is along a continuum and the pressure to change persists. He then did the math by conducting a series of computer simulations to see what kind of changes to language production, perception, and usage might co-evolve. Not surprisingly, he found that everything moves slowly up a scale if the pressure persists. His results “show that it is in principle possible for strong [observable] cognitive traits to evolve under selection for use in language.” It is, therefore, probably not just a happy accident that human hearing is most sensitive in exactly the ranges produced by the human voice and not the ranges produced by chimpanzees.
This finding is right in keeping with this blog’s position that apes could start using a primitive language today if their motivation to keep secrets gave way to a motivation to share knowledge. That is to say apes are already at the rudimentary start of language, able to associate arbitrary symbols with percepts, able to direct one another’s attention, and able to keep something in mind for a small amount of time.
One of the systems de Boer examined was purely anatomical: the ability to produce vowels sounds. The other was cognitive: the ability to recognize vowels. Both systems are assumed to be functional, since without the relevant anatomy we would be unable to speak and without the relevant cognitive power we would be unable to understand what was said.
The actual vowels produced and recognized, however, are arbitrary. One culture may say toe-mah-toe and another tuh-may-tuh even though members of each group are all born with the ability to use either vowel set.
This finding argues that cultural differences are built into the linguistic machinery. We can evolve the functional universals of a system, but the arbitrary cultural elements are left to their own devices. I like this finding because, if you look back up at my answer to question #1, you will see that this difference is old news on this blog. Some linguists wrestle with the question of why languages differ and many proposals have been offered, but there may be a mechanical explanation. Languages are bound to drift over time because we cannot evolve the ability to stabilize a language’s arbitrary features.
Chomsky and disciples have been saying something similar for decades, but they have never placed their universals on a continuum; so they cannot have co-evolved. Chomsky is right when he says that either language appeared in a single leap or his system is off base.
De Boer concludes, “Researchers in language evolution may … have been searching for the wrong kinds of adaptations to language. Rather than the discrete mechanisms [favored by Chomsky and others] … perhaps the mechanisms that have undergone selection related to language are the continuous traits that have [long] been considered of secondary importance…”
On this blog, besides vocalization and perceptual range, those continuous traits would include powers of attention and working memory.