Linkers Are Not “Possession Markers” (but “Agreements”)

The empirical focus of this work is linkers in Albanian, Aromanian, and Iranian (Persian and Kurdish ezafes). We argue that linkers (at least in the languages considered) are neither copulas nor case assigners, but they are closer to what is usually called agreement. Specifically, the parallel is with the clitic pronouns/determiners of the Romance languages, which are also known to enter agreement (aka doubling) structures. We argue that so-called agreement morphology is interpretable as a partial saturation of argument slots. A cross-linguistic survey of determiners, linkers, and pronominal clitics based on this view is provided; we briefly extend our discussion to pre-genitival linkers, treating genitive case as a lexicalization of the part-whole relation, of which the linker saturates the external argument (the possessum).