Mainstream models of early language acquisition assume a progressive construction of linguistic units from small (phonemes) to large (words) ones in explaining how infants acquire a lexicon (i.e. Kuhl, 2004). However, there is now ample evidence showing that the process of establishing a phonetic space runs in parallel with the construction of the lexicon, in a fully interactive way. In the present project we want to test the hypothesis that former atypical results in bilingual infant phoneme discrimination tasks can be accounted by a model where phoneme acquisition is seen not as the results of the computation of rather local distributional regularities, but as also reflecting the distribution of information at the lexical level. To this end, two separate series of experiments will explore monolingual and bilingual 8-month-old infants' phoneme discrimination capacities. A third series will explore the perceptual capacities of adult monolinguals and bilinguals. The experiments will vary the linguistic context testing the contribution of general domain mechanisms, in particular the  effect of "Acquired Distinctiveness".