摘要

An investigation into the evolution of protolanguages is described, exploring the tension between faithful reproduction of memes to ensure consistency and hence comprehensibility in a language and the need to generate new memes to encourage the diversity observed in evolution of new languages and dialects. Agents were given different genetic strategies for interpreting, memorising and imitating memes. The process of memetic reproduction over 50 generations was simulated with competing fitness criteria to optimise consistency of reproduction on the one hand, and diversity in the meme set on the other, while simulating variations in the complexity of seed meme sets. The results showed that selection among interpreting and memorising strategies is more critical for faithful reproduction of memes; in contrast, there was little difference between imitation strategies that selected memes either randomly, or based on frequency and similarity of the memes received in previous exchanges. More complex meme sets produced more variation favouring diversity, while simpler meme sets favoured consistency. Shorter memes were more stable than longer memes over time, but increasing the number of seed memes had less effect. The implications for the evolution of memetic reproduction and protolanguages are discussed.

  • 出版日期2012-3