From the cursory examination of data presented in this chapter, several general tendencies emerge concerning the phonological knowledge and performance of heritage speakers. First, heritage speakers tend to amplify properties that separate their two languages; for example, German-dominant heritage speakers of Italian exaggerate Italian gemination, while English-dominant heritage speakers of Polish exaggerate final obstruent devoicing. This supports the notion that heritage bilinguals maintain separate representations of each sound system and do not identify L1 sounds with L2 sounds. As more work is done on categorization in phonology, more compromise (“good enough” or “best-case scenarios”) may emerge. This result supports the differentiation between underlying representations and online performance by heritage speakers. It is often tempting to attribute differences or deficits in the heritage system to the latter, assuming that the former is intact and native-like. While it is true that production reveals heritage speakers’ deficits in a more apparent manner, the phonological discrimination data presented here also indicate that differences between heritage speakers and baseline speakers run deeper and cannot be reduced to performance limitations alone.