Nim Chimpsky. If language adapted to our brains, why didn't it adapt so well to his brain?
The previous two posts (here and here) have reported a theory that says:
- Syntax is not arbitrary.
- Syntax evolved by adapting to the human brain.
- Because syntactic rules are adapted to the brain, children can start speaking easily without requiring specially evolved language modules in their brain.
Those two posts focused on explaining the theory and its arguments, so in this final post I want to look at some of its problems. My own opinion is that the article (“Language as Shaped by the Brain” by Morton H. Christiansen and Nick Chater in Behavioral and Brain Sciences; final draft here) marks a great stride that ties together many trends in the study of speech origins, but there are still some questions, starting with the FOXP2 gene. We know that speaking proper language depends on this gene, so how can Christiansen/Chater dismiss the claim that syntax is the result of biological evolution?
FOXP2
The Christiansen/Chater paper dismisses the FOXP2 issue in a footnote that says the gene is important to sequential learning generally, not just speech. The answer is too cavalier, suggesting the authors were in a hurry and unwilling to really grapple with the issue. As I understand their argument, their primary quarrel is with the notion of the evolution of an arbitrary, universal grammar. They are willing to posit the evolution of functional changes, such as changes to the vocal tract that promote articulation. Last January, this blog reported work with FOXP2 in songbirds that indicates the gene is critical in helping an individual learn to imitate the sounds of its neighbors. (See: Birds Also Use FOXP2) This is exactly the sort of task that humans must master, no matter what language they are destined to speak, and that chimpanzees need never trouble themselves over. It is well known that speaking clearly is one of the most difficult elements children face in mastering speech, much more difficult than syntax. So we would expect genetic support in this task.
As for the grammatical problems of speakers with mutated FOXP2 genes, this supplies a great test for the Christiansen/Chater theory. Generative grammarians propose that some genes result in the formation of modules that impose syntax on the speaker. Does FOXP2 do that? Probably not, but then generative grammarians have a fall back position. They can just say the gene is not that one. And if no gene is ever found, they can say the module is a spandrel. It is difficult to falsify the generative claim. The rival position is more testable. Genes that impede speech should do their damage by altering the mind’s geometry so that language and brain do not fit so easily. If FOXP2 works that way, well and good. If not, the theory will need some adjustment.
At this point we do not even have a good description of the syntactic problems faced by mutant speakers, let alone a strong enough account of the mind’s linguistic geometry to make predictions. But in principle the Christiansen/Chater idea can be falsified or buttressed once we do have an understanding of the gene’s results.
Ape Sign Language
By this time a number of chimpanzees, bonobos, and gorillas have learned to use sign language to a modest extent. They have also put two signs together in a way that sometimes seems coherent, but they have not been able to fashion sentences. Why not, if all it takes is for words to come into contact with a brain? And we do have evidence suggesting primates can think in what generative grammarians would call noun phrases. (See: Just How Old Are Noun Phrases?)
The question may sound a bit naïve, given the many obvious differences between the brains of chimps and humans, but the authors do specifically say, “we suggest that language does not involve any qualitatively different mechanisms compared to extant apes” so some kind of grammar might be expected to emerge from ape constructions. Yet it did not. Why not?
Lets remember the process by which language is supposed to have adapted to the human brain. Speakers say things and, as the words are passed from speaker to speaker, word usage adapts to the geometry of the users’ minds, introducing structure as a matter of course. The process has been demonstrated in the lab using both computers and humans. But apes do not pass the signs on. Having learned signs, they did not begin to converse among themselves. Thus, there really is no opportunity for ape language to evolve and adapt itself to ape brains.
The Christiansen/Chater theory concerns syntax; linguistic modules for organizing sentences may not have evolved in the brain, but we still had to evolve the capacity to learn from one another, to imitate one another, to be willing to pay joint attention to one another. Pinker’s Language Instinct may be at risk with the new theory, but not the idea that the origins of speech are biological and evolutionary.
The Critical Period for Language Learning
Although children have an easy time learning an unknown language, adults do not. What is easily learned at four becomes difficult at sixteen, and in those rare cases of abused children who do not come in contact with language until they are twelve or so, learning to speak seems impossible. How can those facts be reconciled with the idea that language is adapted to the brain?
I don’t suppose they can be reconciled if you take a naïve view of the brain as a static organ like the liver or pancreas. But the brain is extremely dynamic. Anything written on its blackboard cuts a groove. Gerard Edelman’s theory of “neural Darwinism” proposes a brain in which connections are preserved through a selection process. A critical period is a natural result of his theory in which the brain adapts itself to its own experience. A process of neural selection in early adolescence changes the brain from an organ on which many things can be written to one in which there is a very strong bias toward what has already been marked on its board. So language can adapt to a youthful brain and become fixed, so to speak, in the adolescent head. (see: Darwinian Linguistics). A dynamic brain is in keeping with Christiansen and Chater, the observations of a critical period, and the further observation that adults forced to forgo their mother tongues will speak a pidgin, but their children will transform the pidgin into a full language.
Now, what we really need is a good navigator’s map of the mind, one that shows where the chalk will go if you start drawing on the blackboard.
Comments