Modelling of factors underlying the evolution of human language

Abstract

A central question in the evolution of human language is how it emerged. Based on recent research across disciplines, we identified three processes proposed as potential driving factors behind the evolution of ‘modern’ language phenotype: i) a reduction in reactive aggression entailing a boost in prosociality and cooperation, ii) a change in early brain growth trajectory that impacted structures like the cerebellum and striatum, and thus likely impacted the (procedural) memory circuits these regions support, and iii) a demographic expansion of H. sapiens during the Middle Pleistocene. While extensively researched on their own, the interaction between these three processes has yet to be investigated systematically. We develop an abstract agent-based model to interrogate the relationship between these three factors and how they influence transmission of information within a population, which we take to be the essence of language. The model abstracts linguistic capacity to an ‘array of skills’ and investigates under what conditions the number of skills increases. The results demonstrate that there is an optimal degree of cooperation and memory capacity at which the amount of transmitted information is the highest. Our model also shows that separate linguistic communities arise under circumstances where individuals have high levels of memory capacity and there is at least a certain degree of non-cooperation. In contrast, we find no significant direct effects for population size in the process of linguistic community formation. Taken together, these results highlight the explanatory benefits of combining insights from cognitive science, archaeology, and computational modelling.