Binding of Reflexives in Polish as Agree, Move, and Late Spell-Out

This paper considers components necessary for a successful account of A-binding relations in Polish, a language with subject-oriented reflexives and a binding domain delimited by the Tensed Sentence Condition. Following the presentation of major relevant data points in Polish, two comprehensive theories of binding—the Agree-based theory, presented in Reuland 2011, and the Move-based theory, presented in Boeckx et.al. 2008—are briefly outlined and applied to said data. It turns out that the two theories, in their most orthodox forms, fall short of achieving empirical adequacy. Subsequently, a positive theory of A-binding is proposed which combines upward Agree, movement (and copy pronunciation) of the bound element, similar to movement of clitic/weak pronoun in Polish, and a lexicalization algorithm modeled upon the proposals in Safir 2014 and Nikolaeva 2014. It is shown in a number of derivations with possessives how both the subject and the object engage in binding relations as antecedents and how their dependents become lexicalized as either reflexive or pronominal.