Dialects as Genetic Barriers.

Yakov Pichkar and Nicole Creanza have a preprint “Subtle cultural boundaries reinforce genetic structure in England” whose abstract reads:

Genes and languages both maintain signatures of human history. The evolution of genetics and of culture both have features that can track population movements and demographic history. Further, cultural traits may themselves impact these movements and demography. In particular, while speaking a different language appears to act as a barrier to gene flow, it is not clear whether more subtle dialect-level linguistic differences within a language can influence mating preferences and thus affect genetic population structure. We examine the strength of cultural barriers and of association within England using the spatial similarities between rates of linguistic and genetic change. We find that genes and dialect markers have similar spatial distributions at all geographic scales, though these similarities are more pronounced at larger scales. This covariation, in the absence of geographic barriers to coordinate linguistic and genetic differentiation, suggests that some cultural boundaries have maintained genetic population structure in England.

Interesting stuff, and I’ll be interested to see what Hatters have to say. Thanks, Dmitry!


  1. … suggests that some cultural boundaries have maintained genetic population structure in England.

    It would interesting to add economic class to the variables studied, and to test for correlation between economic status and dialect. Then, to get fancy, distinguish between dialect only speakers and dialect speakers who code switch with RP.

  2. David Eddyshaw says

    As far as I can make out, what the authors feel that they have demonstrated is that dialect diversity (among elderly rural people in the 1950s) correlated to some degree, in some places particularly, with genetic diversity among rural people all of whose grandparents were born within 80 miles of each other, around 1895.*

    The rest is all unconstrained speculation based on this somewhat unstartling news, with occasional gems like

    These patterns of genetic clusters reflect waves of migration to Great Britain and of later movements of people

    Well, yes … that does seem to cover all the options …

    * For some reason I am powerfully reminded of Cold Comfort Farm. I mun cletter the dishes now …

  3. What a load of old cobblers (as we in the apparently new genetically-unmodified England say)*. I wonder who will get a PhD from testing these “theories”. People have migrated in and out of our islands for eternity and left bits in our gene pool, and bits in our language, too. As David Eddyshaw says, we English are much more diverse and are no longer living in the society of the nineteen fifties. The authors, I am afraid, are showing their ignorance of Britain today.

    *Maybe I should contact the lab that gave me my DNA analysis to say that I’m not the hybrid they make me out to be.

  4. clettering (an impractical method used by Adam for washing dishes, which involves scraping them with a dry twig or clettering stick).

    “Fake but authentic-sounding local vocabulary (in particular Sussex and West Country accents)”, sez the WiPe. Not the Adam, that would be anachronistic. But genetically covariant, I suppose.

  5. David Eddyshaw says

    When Adam cletterid and Eve span, who was thanne a gentilman?

  6. David Eddyshaw says

    The authors, I am afraid, are showing their ignorance of Britain today

    To be fair (when am I not?) I think the primary objective is not really to shed light on contemporary Britain: their statement of what they are trying to achieve goes

    However, these studies that observed evidence of gene-language covariation have not been able to analyze its causes. They tended to compare people speaking different languages, and few used quantitative measures of language variation (22, 28, 29). These limits were due, in part, to the limited availability of quantitative data describing cultural variation and the spatial resolution of available genetic data. These limits only allow the comparison of genes and language at the scale of ethnic groups and large regions, where the effects of population movement and cultural homophily cannot be disentangled. Therefore, the impact and spatial extent of these forces has not been thoroughly studied at smaller scales, such as within a geographic region where individuals generally speak the same language.

    I don’t think their methodology is actually able to adduce causes as such; at most it can only find correlations, which is why they then disappear into a morass of speculations, some of which do seem to show a surprisingly poor grasp of the actual history of Britain.

    I would be moderately interested if they tried this out in northern Ghana, actually; their assumption that on smaller geographical scales people all speak the “same language” is based on their preconception that modern Europe is typical of how these things work. And it would be much easier to exhibit robust differences between northern Ghanaian languages than to go fishing for relatively trivial English dialect differences which seem to be based on selecting from what information was conveniently to hand rather than deciding before you conducted the study what your linguistic criteria should be.

    This may possibly be an important point. As far as I can see, there is no information on whether all the linguistic data were used in the statistics, or if only a selection, and if the latter, how this selection was made, and (crucially) at what stage of the study this selection was made. Unless such a selection was made beforehand, on criteria quite independent of what the study is attempting to discover, I would have thought that the statistics are frankly meaningless. But this may all be in the supporting material, I suppose.

    I would be interested if it turned out that speakers of Western Oti-Volta languages showed significant genetic resemblances to one another, compared with speakers of other Oti-Volta languages. It’s not at all clear whether the undeniable fact that the relatively uniform WOV subgroup covers half the land area where Oti-Volta languages are spoken is to do with the expansion of the historical Mossi-Dagomba states. (Though if I had to guess, I would say, yes it is, but the actual speakers are probably not particularly closely related genetically, but are descended from locals who adopted “imperial” WOV languages in place of whatever they spoke previously.)

    Getting all the necessary genetic data together would be a major task though. I don’t suppose it’ll ever happen …

  7. Lars Mathiesen says

    It’s easy to be fair when you have backup coming from A Centauri innit.

  8. David Eddyshaw says

    As one of your own sages has remarked: With great power comes great responsibility.

  9. I haven’t read the article yet, but I’d expect dialect barriers to correlate with (subtle) genetic differences for no other reason than that they both reflect patterns of historical contact.

  10. David Eddyshaw says

    Yes, their implication that this has all come about because nice girls don’t marry chaps what don’t talk proper (and vice versa) seems – unjustified by their evidence. Their actual findings (assuming that they haven’t vitiated their statistics by the way they set up the study, which I would hope population geneticists would be too savvy to do) seem in themselves to be totally unsurprising.

    “Reinforce” implies causation. That’s pure speculation.

  11. January First-of-May says

    nice girls don’t marry chaps what don’t talk proper (and vice versa)

    Reportedly many different far-flung parts of the world have developed a tradition where people aren’t supposed to marry within their language/dialect (presumably as a proxy for inbreeding avoidance). But I guess that would be much less relevant in modern(ish) Europe.

  12. David Eddyshaw says

    Famously, the Vaupés in Amazonia, where taking a spouse with the same mother-tongue is (or was) regarded as incest. (Though in fact you tend to end up marrying a cousin.)

    Although at most only trivial language differences are involved, the Kusaasi do not marry within their clans, which are locality-based. I was once told “the first thing a young man looking for a wife needs to do is to buy a bicycle.”

  13. Trond Engen says

    There’s a new paper out on Early Medieval migrations from the continent, I posted that here. It would have been fascinating if the two works were connected, but it doesn’t look like it.

Speak Your Mind