Evolution of language diversity: Why fitness counts

Many recent studies exploring the evolution of language in a population of communicating agents assume that the possession of a common language increases the fitness of its users (e.g. Cangelosi 2001; Reggia et al. 2001; Nowak et al. 1999; K. Smith 2001). Language-conferred fitness is defined as the increase in individual survival probability that stems from successful communication: a group whose members communicate well is supposed to leave, on average, more offspring. Although the assumption that language increases human evolutionary fitness is intuitively appealing, its validity needs to be examined in the light of indications that a coherent language can emerge in the absence of language-related fitness (Kirby 2000; Oliphant 1999; Oliphant and Batali 1997; Briscoe, Chapter 14)-for example, through a combination of genetic drift and random population diffusion.