Diffusion models


Ethnos and language are intimately associated, and are known to be much more resistant to change than culture and social stratification, and thus changes in material culture are not to be equated to changes in language, even if ethnicity may take on new meanings[Kristiansen 2000].

Demic diffusion refers to a migratory model of population diffusion into an area that had been previously uninhabited by that group, possibly displacing, replacing, or intermixing with a pre-existing population[Cavalli-Sforza and Minch 1997]. It is theoretically the simplest (and thus strongest) link with ethnic and linguistic change, since it shows the predominance of a new people that displaces or absorbs the original one. This is usually accompanied by a decline in Y-DNA variation, since certain chiefdoms and clans usually predominate in the expansion of a population.

Founder effect refers to a loss of genetic variation caused by a colonization and genetic separation of a subset of the diversity present within the original population, and is different from a bottleneck, where the original population loses its prior diversity by a similar mechanism[Jobling et al. 2014]. It seems theoretically second to demic diffusion, in explaining the replacement of genetic make-up without replacement of language. To resort to a founder effect to explain population changes when enough ancient DNA samples are lacking to suggest them is dangerous: the scarcity of ancient DNA samples makes the interpretation of their meaning – in relation to actual ancient areal occupation – a matter of subjective evaluation, in conjunction with archaeological finds[Campbell 2015].

Cultural diffusion in a strict sense (opposed to demic diffusion) refers to the spread of cultural traits – including ideas, technology, and language – between individuals, without a need of a migration. Multiple models have been proposed, but all offer a weaker potential explanation for linguistic change than demic diffusion or founder effect, since it implies that language spreads by way of economic or cultural (e.g. religious) domination. Given the strong ethnic connection of language, examples of such an event were probably exceptional before the creation of the first states. On the other hand, population expansion into certain territories and decline of the original population are followed in some cases by a rising of the original paternal lineages and admixture component[Brandt et al. 2015]. A static genetic situation observed after that process could be quite simply interpreted as cultural diffusion, supposing that no (or almost no) population exchange has taken place. Admixture analyses are quite useful to investigate these cases.

Other potential models can only be weaker than these main three. It seems logical that weaker models should not be used lightly, and clear proof of their applicability (and non-applicability of the stronger models) should be given in each case.


  • [Brandt et al. 2015] ^ Brandt, G., A. Szecsenyi-Nagy, C. Roth, K. W. Alt, and W. Haak. 2015. Human paleogenetics of Europe--the known knowns and the known unknowns. J Hum Evol 79:73-92.
  • [Campbell 2015] ^ Campbell, Lyle. 2015. Do Languages and Genes Correlate? Language Dynamics and Change 5 (2):202-226.
  • [Cavalli-Sforza and Minch 1997] ^ Cavalli-Sforza, L. L., and E. Minch. 1997. Paleolithic and Neolithic lineages in the European mitochondrial gene pool. Am J Hum Genet 61 (1):247-54.
  • [Jobling et al. 2014] ^ Jobling, M. A., E. Hollox, M. MHurles, T. Kivisild, and C. Tyler-Smith. 2014. Human Evolutionary Genetics. Second edition ed. New York and Abingdon: Garland Science, Taylor & Francis Group.
  • [Kristiansen 2000] ^ Kristiansen, Kristian. 2000. Europe Before History. Cambridge: Cambridge University Press.