Artificial language learning paradigms can be used to examine learning biases and their relation with linguistic typology. Focusing on phonology I will present two case studies, one on vowel harmony and the other on syllable reduplication.