Conflicts may arise when listeners encounter a language with a different number of vowels. For instance, Spanish has 5 vowels, English at least 9 and Spanish listeners have many difficulties perceiving the /i/–/I/ contrast in some (though not all) varieties of English. Spanish learners of English solve this few-to-many problem by incorporating a new auditory dimension (vowel duration) into their categorization system. Perhaps surprisingly, the reverse scenario, namely encountering a language with fewer vowels, is also problematic. For instance, Dutch listeners of Spanish (Dutch is comparable to English in its vowel inventory size) have a boundary mismatch when listening to the Spanish /a/–/o/ contrast, and they hear the extra category /I/ when listening to Spanish /i/–/e/. Dutch learners of Spanish shift their boundaries toward what would be appropriate for the target language and they gradually stop perceiving the extra category. It turns out that both in few-to-many and many-to-few situations learners are capable of changing their perception to become like or almost like the native speakers of the target language. This can be explained within a formal model of speech perception and its development.