The Workings of Co-Evolution

Yoruba and Italian peoples can learn each other's languages, but slightly different facial features bias  the evolution of phonemes in slightly different ways.

Here’s a hypothesis that I expect will become known as a law: Language develops in the individual as the interaction between the individual’s genetic bias, the population’s genetic bias, and cultural history. The surprising part of that statement is the reference to population genetics. That idea is the point missed by Chomskyan linguists, cognitive psychologists, and associationists. This statement brings together several themes that have appeared on this blog since the end of last summer, and it explains what happens when language and people co-evolve.

Last September, there were several posts (see: Was the First Language Tonal? and Do Genes Bring Up the Rear?) that reported a monumental study by Dan Dediu and Robert Ladd of the University of Edinburgh (author’s summary available on-line here, complete with links to the full article) which showed that linguistic behavior can be shaped by the genetics of a population rather than by the genetics of the behaving individual. Populations where a particular gene is carried by a majority of members tend to speak tonal languages; populations with a different genetic structure tend to speak non-tonal languages. The genes of an individual are not decisive in shaping that individual’s speech; the genes of the population are critical.

I wrote at the time “I have the feeling that this discovery is one of those things, like Robinson Crusoe observing a footprint, that changes many expectations. The Dediu-Ladd paper on the relation between tones and genes calls for a new way of thinking about thinking. The nature-nurture discussion just took a new turn.” So, of course, I’m glad to see Dediu and Ladd (with Anna R. Kinsella) are back discussing the implications of their finding and that they put their ideas in the context of the new paradigm that I saw emerging in Barcelona.

That other idea is the importance of co-evolution of language and people, “Language can adapt to the abilities of the speakers by changing so that it is easier to learn.” (see: Co-Evolution Idea Won Big in Barcelona). In the current issue of Biolinguistics Dediu, Ladd, and Kinsella participate in a “Forum” on “Language and Genes” (available here, registration required) and they describe a co-evolutionary relationship between genes and language (although they never use the word “co-evolutionary”).

Here’s how the authors present the hypothesis I cited at the start of this post:

… we are not proposing any sort of deterministic relation between genes and language, only a very indirect and probabilistic one; we certainly are not suggesting that there are “genes for Chinese”. But we believe that the broad outlines of an explanation based on the interaction of bias [both individual and population] and cultural transmission are very plausible indeed. [p. 120]

How does the hypothesis work?

It is now well established that genes affect speech and language in individuals. By this we mean that there are demonstrable associations between inter-individual differences in genetic makeup and inter-individual differences in speech and language abilities. The best known case to date is that of the FOXP2 gene. [115]

The interesting thing about this level is that while it determines something about an individual’s linguistic abilities, it has no impact on language beyond the level of that speaker and contributes little to nothing to co-evolution.

It is also well established that genetic and linguistic diversity are correlated at the level of populations. That is, geographical inter-population differences in allele frequencies tend to match the distribution of language varieties [115]

These correlations tend to be spurious; that is they are not causally related. Although language and ethnicity are distinct, they can be correlated, especially in the case of languages with a small number of speakers. The group’s linguistic and genome distinctions will be correlated, but neither causes the other. At the same time, it is important to remember that language is evolving. This is the cultural history mentioned in the opening hypothesis. Not many people will be surprised that it matters, but the point is a bit more radical than it sounds. Chomskyan linguists lean toward the argument that the decisive element is an individual’s genetics and the cultural history, while important to the details of which language one acquires, is much less important than the innate knowledge.

A third possible type of relationship between genetic and linguistic diversity … [is theoretically possible] between population genetics and language typology. [116]

The point is that genetics matters, but population genetics shape the language, not the individual’s. Besides the example of tonal language that Dediu and Ladd reported last year, the authors give an interesting case from an old, 1984, paper by Peter Ladefoged, late of UCLA’s linguistics department, which showed that the very slight differences between the vowel system of Yoruba speakers (of Nigeria) and Italian speakers could be explained by the very slight differences between the typical face of a Yoruba speaker and that of a typical Italian speaker. We assume, of course, that a Yoruba child raised in Italy will speak Italian like a native, and that the same is true for an Italian child raised as a native speaker of Yoruba.

But suppose a group of Yoruba immigrants settled in Italy and raised their children as Italian speakers. Then, after a few generations in which the Yoruba become fully assimilated as Italians, suppose their descendants are suddenly isolated from the rest of Italian speakers. They would continue speaking Italian, but as the language was passed on from generation to generation we might expect the language spoken to take on the characteristics of Yoruba vowels.

Why would we expect that? Because of the co-evolution of language. Language changes to fit existing genetic biases within a population. Individual speakers must keep up, whether it is easy or not. No wonder that languages typically have one or two features that native speakers find troublesome. The authors note that in the Yoruba/Italian distinction,

The linguistic bias in this case is unrelated to any biologically selective pressures … linguistic differences can merely be indirect by-products of characteristics that have independently evolved [119]

A critical implication of this notion about linguistic differences is that, even though the authors [and this blog] take evolutionary arguments very seriously, we cannot assume that language evolution must be explained in adaptive terms, or at least not as an adaptation to the speaker’s environment. Language may adapt to the peculiarities of a population’s genome and be quite irrelevant to the environment around the speaker. So there may be reasons for a particular language's form, but that reason has nothing to do with language iself. As the authors put it, the observation “sits uneasily with the idea of language as a perfect and economical system.” [121] Here they are making a point already stressed in Gary Marcus’s new book Kluge (see: Just How Sane Are We?)

The environment in which humans act is shaped by traditions that reflect both nature and nurture over time. The human “capacity for language … is not fixed and uniform across the species, but diverse and dynamic.” [121] Some linguistic universals may indeed be the result of some deep, genetic cause, but we cannot assume that all variation in language is mere some surface output of a deeper universal. The difference between Yoruba and Italian phonology, or tonal and non-tonal languages turn out to reflect population differences that are real and yet unrelated to linguistic selection.

Population-based co-evolution goes a long way toward explaining one of the deep mysteries of language. Why does language require a community for it to form in the individual? Chomsky has proposed that language acquisition is as natural and automatic as the onset of puberty. It comes, no matter what. But children who reach adolescence will go through puberty, community or not. So-called wolf children raised apart from any community will not develop a language.

This hypothesis, if it endures, also puts the end to theories that language evolved at the individual rather than the community level, or that it evolved for thought first and communication second.

Comments

[watercat]

Children do not develop language unless they participate in a community, as famously in Nicaragua and the Sinai, but sometimes not even then. William Washabaugh studied a community of deaf persons who did NOT create a language. The crucial difference was nothing to do with alleles, it was the opportunity to negotiate consensus with one's peers.