Universals in possessive morphology

In this paper, we test the hypothesis that possessive pronouns have the same basic structure containing the genitive pronoun, plus, in some languages, some extra structure, as suggested by Caha (2009). In order to unravel the structure of these pronouns, we use the same logic applied by Caha (2009) and Bobaljik (2012) that excludes so-called ABA-patterns. If possessive pronouns are built on top of the genitive, we derive several predictions. First, we predict that there are languages in which the possessive pronoun comprises the genitive pronoun plus an extra affix (complex morphology). Furthermore, we predict that there are no possessive pronouns that have the same form as the accusative, or the nominative pronoun, to the exclusion of the genitive (*ABA). And thirdly, we expect that any syncretisms between possessives and other pronominal forms respect the proposed hierarchy in the sense that only structurally adjacent forms may be syncretic. Our data provide ample evidence for the claim that possessive pronouns are “bigger” structures than the accusative or ergative pronouns, suggesting that the possessives are indeed constructed from these structures. However, the data in our sample do not give crucial evidence for the claim that the possessives are more complex than the genitive. The data leave open the possibility that the genitive is in fact “bigger” than the possessive. Only in a few languages do we find ABA-patterns. We argue that these ABA-patterns are only apparent counterexamples to the proposed structure. Therefore, we conclude that there is broad typological evidence for the hypothesis that possessives are built from pronouns expressing a dependent (accusative/ergative) case