Why we talk?: altruism and multilevel selection in the origin of language

Talking entails costs of production and time, although some of the information sent to hearers will be of value to them in general. We believe that the matter of why we talk at all is a key question for the origin of language, and the answer will shed some light on the mystery of human identity. This article focuses on altruism in communication, and aims to demonstrate evolutionary scenarios based on multilevel selection. We constructed a computational model to examine these scenarios. The evolutionary experiments showed that in the case of an unstructured population, a linguistic system hardly emerged due to the dynamics between interpretable utterance that imposes a penalty and correct interpretation that yields a reward, which is similar to prey-predator dynamics. However, in the case of a multigroup population, a linguistic system emerged owing to multilevel selection among the groups. In addition, the probability of success in conversation was higher in a group in more severe environmental conditions. This result supports Bickerton’s hypothesis based on the ecological gap between human ancestors and other ape species.

Authors and Affiliations

1. BS Division 2, Toyota Communication Systems Co., Nagoya, Japan

   Kana Sugiura

2. Graduate School of Information Science, Nagoya University, Furo-cho, Chikusa-ku, Nagoya, 464-8601, Japan

   Takaya Arita

Corresponding author

Correspondence to Takaya Arita.

This work was presented in part and was awarded the Young Author Award at the 15th International Symposium on Artificial Life and Robotics, Oita, Japan, February 4–6, 2010