The Interplay of Language and Human Evolutionary Histories 

Dan Dediu 

Language Evolution and Computation Research Unit, University of Edinburgh, UK

ddediu@hotmail.com

The evolution of the capacity for language, including linguistic
universals, cannot be meaningfully separated from the history of
languages, but also the reverse holds true.  The evolution of language
and languages have taken place in the broader context of human
evolution, the feedback cycle being mediated, on one side, by cultural
selection on learnability and usability of languages, conducing to
glossogenesis, and on the other side, by a form of linguistic Baldwin
effect. 

This realization implies that different models of human evolution should
be correlated with different models of language (and languages)
evolutionary history.  At the two extremes there are the
Recent-Out-of-Africa and Multiregionalism, but there is also a class
of intermediate models.  The nowadays almost universally accepted
Recent-Out-of-Africa should imply a recent common ancestor of (almost)
all modern languages as well as a small individual and regional
variation in linguistic capacity.  Multiregionalism should imply a
deeper common ancestor of modern languages as well as a more pronounced
individual and regional variation in linguistic capacity. 

Empirically testing this hypothesis is feasible, even if not a simple
task, using a series of tests like: adopted children in cultures
speaking languages of different linguistic family than their own,
comparative performance on equivalent linguistic tasks of adults
speaking different languages, etc. 

The present study employs a computer modeling approach.  It involves
regional populations of linguistic agents, able to split, migrate,
replace or interbreed with pre-existing populations as well as
linguistically interacting with them.  The language is simulated using a
series of linguistic properties which are genetically encoded but
flexible enough as to be modified through learning during the critical
period (there are learning costs associated in order for the Baldwin
effect to take place).  The preliminary study suggests that in this case
the presence of the Baldwin effect can have detectable consequences on
the linguistic map and it thus indicates that a more complex study
targeted at testing the relationship between human evolutionary models
and language evolution is relevant and worth pursuing.