This paper deals with the issue of how a language family's range (how geographically dispersed it is) provides information about its age (how old it is).
I believe that a similar methodology should be applied to the spread of gene variants as well (e.g., Y-chromosome haplogroups). At present, haplogroup ages are estimated only by internal (genetic) information, e.g., the diversity of linked STR loci. However, as I have argued before, there are additional pieces of information: (i) how big the haplogroup is, and (ii) how geographically dispersed it is. Both (i) and (ii) depend on a haplogroup's age.
CURRENT ANTHROPOLOGY Volume 49, Number 6, December 2008 DOI: 10.1086/592436
Language Spread Rates and Prehistoric American Migration Rates
Spread rates for language families can be calculated from the family's range (which is generally known rather precisely) and age (which is only rarely known precisely but can often be estimated with useful accuracy). Spread rates are calculated here for a number of different language families and subfamilies in different cultural and economic contexts. Deliberate long-distance migrations, imperial conquest, and transport using wheels or sails make for very rapid spread rates. The rate of prestate, pretransport spreads depends primarily on ecology (latitude, coast vs. interior, mountains, vegetation, climate); presence versus absence of food production and movement into inhabited versus abandoned land have little impact on spread rates. This fact makes rates of recent language spreads applicable to early prehistory, where they can be used to model prehistoric colonization rates. Average rates for various ecologies are calculated for a spread from a North American entry point to the archaeological site at Monte Verde, Chile (14,500 calendar years ago). The time required gives a latest possible date for the first settlement of the Americas. Entry dates postdating the end of glaciation all require implausibly fast rates of spread.