In Udi, person markers occur either within the verb, in positions determined by internal morphological structure, or else at the right edge of a syntactic constituent in focus. Harris reasons that since describing the distribution of these person markers requires reference to both morphological and syntactic entities, this phenomenon challenges the lexicalist separation of morphology and syntax advocated by the Lexical Integrity Hypothesis. I argue that one can maintain a lexicalist approach by adopting lexical sharing, which allows words to instantiate multiple terminals. Udi person markers are positioned within the word by purely morphological alignment constraints; words containing a person marker instantiate an additional terminal, which is positioned by purely syntactic alignment constraints. This analysis preserves the separation of morphology and syntax, showing that the data surrounding Udi person markers are not, in fact, damaging to lexicalist theory.