Mechanisms for cross-situational learning of word-referent mappings: Empirical and modeling evidence