Exocentric compounding and implicit heads

Abstract This chapter revisits the role of null morphemes in morphological theory, extending the discussion to exocentric compounding — an empirical domain where null morphemes and empty categories are often posited to resolve issues of headedness. Through a re-examination of exocentric compounds, I argue that the types of exocentricity identified by Scalise et al. (2009) arise from two distinct types of null elements: (i) a null category head that accounts for categorial exocentricity, and (ii) an empty noun that explains semantic and morphological exocentricity. The first type of null assigns a syntactic category to a phrase without necessarily aligning with the categories of its constituent members (e.g., n, v , or a ), thus creating a categorially exocentric structure. This category head, which takes a phrase as its complement, also reinforces lexical integrity, enabling unified movement and binding. The second type introduces an external (null) head in exocentric compounds, contributing additional morphological features and acting as an intervenor in agreement relations, thereby preventing any of the compound’s constituents from individually serving as agreement targets. The assumption of these two types of nulls is supported by evidence from argumental (deverbal and de-prepositional) and possessive compounds in Romance languages, which will serve as the empirical focus of this chapter.