Self-organisation of a learnable language: What happens when an alien language replicates through human brains

Through the use of an iterated learning experiment, we investigate the emergence of self-organised structure in a random initial alien language, as it is replicated through successive `generations´ of language learners in three distinct iteration chains. Experiments in the iterated learning paradigm have frequently shown that languages evolve to become more compositional and thus more easily learnable and replicable over the course of an iteration chain (see, e.g., Kirby et al. 2008). Here, we closely analyse the dynamics of this replication, identifying which features actually correlate with greater replicability, as well as suggesting which mechanisms may be driving the evolution towards a more replicable language. These findings ultimately show us that certain very broad generalisations about language structure can be explained purely by appeal to the self-organisational dynamics of replication through the cognitive environment of individual language learners, even prior to taking social and cultural interaction and population dynamics into account.