Talking our way out of Africa

19 April 2011

Interview with

Quentin Atkinson, University of Auckland

Chris -  Also this week, how our early human ancestors talked their way out of Africa. There's a study marrying up the diversity of the sounds that are used in languages around the world with what genetics tells us about how people migrated out of Africa and across the globe. And it shows that it was probably the art of conversation that got people moving. Quentin Atkinson, the author of this new study works at the University of Auckland in New Zealand.

Quentin -   Ever since Charles Darwin, it's been recognised that the evolution of languages parallels, in many ways, the evolution of species.  So you get this process of descent with modification. Like you get genetic mutations arising, you get what linguists called innovations in languages.  And so, after a period of time when two populations have been separated, they can no longer understand one another, and you get new languages forming, just like you get new species forming when they can no longer breed with one another.  So it was with that background I started looking at global patterns of phoneme diversity. Phonemes are sounds in languages. They're used to Human_originsdifferentiate words with different meanings.  So "cat" has a different meaning to "bat" because there's a different phoneme at the front of it.  One of the key pieces of evidence supporting an African origin for our genetic ancestry is that genetic diversity is highest in Africa, and it decreases with distance from Africa.

Chris -   So you wondered whether the same thing would apply to language.

Quentin -   That's right.  That pattern fits with an idea called the "serial founder effect."  We expect an ancestral population in the homeland to have been there for quite a while and generated a lot of genetic diversity. And during an expansion, small groups will break off from that ancestral population and carry a subset of the diversity with them.  And then, as the expansion continues, small groups will break off from those groups and carry a further subset of diversity.  So the further out you go from the origin, the less diversity you see.

Chris -   How do you counteract the effect of population size? Because, obviously, with a small population, you have a small gene pool.  You're also going to get a small phoneme pool - small number of ranges of different word sounds. That could be an effect of a bottle-neck in a population. It could also be just the fact that a population happens to be small anyway; so how do you get around that?

Quentin -   So, part of the motivation for looking at a serial founder effect in phonemes was that population size has been shown to be correlated with the number of sounds that a language uses.  So languages with more speakers use more sounds.  So that fits with this idea of a founder effect.  So I was looking for a global decrease in the number of phonemes used around the world from some origin point. And you're right, one of the things you need to consider if you're looking at that is that this is not just a result of differences in modern population sizes. So if some parts of the world happen to have, on average, smaller populations, maybe any global patterns we see is just the direct result of that.  So one of the things I did was control for modern population size and looked to see whether we still find a global decrease in phoneme diversity with distance from an origin, and the origin turned out to be Africa.

Chris -   How many different languages did you consider around the world?

Quentin -   In total, 504 languages.  So they're part of a data set called the world atlas of language structures.

africaChris -   Presumably you've got the whole world represented on a map of where people went in evolutionary time, and you've got the languages from those sorts of territories. When you marry the two together, do they agree?

Quentin -   Yeah.  The pattern that we see in the phoneme diversity, the number of phonemes, matches quite nicely with the patterns we see in genetic diversity.  So just like the genetic diversity points to an African origin, so too it seems does the phoneme diversity.

Chris -   So this is another strand of evidence that supports the genetic diversity. People have also done this with Helicobacter pylori - people in Cambridge have done that actually - and also the facial phenotypic appearances. So this is another strand of evidence, but does it actually show anything that those other investigations didn't, or does it fill any gaps that they were incapable of addressing?

Quentin -   I really think it does, because this is not some genetically inherited thing we're tracking.  This is language.  It's culturally inherited to the extent that our cultural ancestry, like our genetic ancestry, can be traced all the way back to Africa, which I think is quite remarkable. And that means that just as we're one big genetic family, we're also one big cultural family.

Chris -   But an even more fundamental question must be, are you suggesting that the thing which catalysed the 'out of Africa' migration, the thing that led to humans taking over the whole world, was that we actually got modern language? Because if the root of all languages is Africa and then they slowly diversify from there, that would suggest that people had to have that core language function first before they went anywhere?

Quentin -   That's right.  It looks like, from these results I've presented, that humans carried with them in their toolkit from Africa language and all the advantages that it confers in terms of cooperation and coordination. And I think language therefore could've been key in giving us the competitive advantage that we obviously had over a lot of the other hominid species which we ultimately obliterated.

Chris -   Quentin Atkinson - he's based at the University of Auckland in New Zealand and he published the work you've just been hearing about this week in the journal Science.

Add a comment

This question is for testing whether or not you are a human visitor and to prevent automated spam submissions.