In this paper, we present an agent-based model to describe the emergence of bilingualism in an artificial language learning communication task as a result of the competition of social conformism (Asch, 1956) and bias towards regularization (Smith, Skarebela, & Tamariz, 2010;Kirby, Cornish, & Smith, 2008). We compare our numerical results with preliminary results from an experiment performed at the LEC in Edinburgh (Tinits, 2012(Tinits, , 2013 and with results of Lupyan and Dale (2010) and obtain good qualitative agreement.