Abstract

The object of study of this thesis is control structures in Persian, more specifically the properties of the empty subject of embedded clauses occurring in Persian control structures. The null subject in these Persian contexts may be one of three different empty categories: pro, NP-trace/copy, or PRO. Obligatory Control contexts are defined by the following properties: they are made up of complex clauses contain a matrix clause and an embedded clause, an argument of the matrix clause must be coindexed with the subject of the embedded clause, and the subject of the embedded clause may be covert or overt, but must crucially be coindexed by the matrix subject. Current theories accounting for Obligatory Control contexts fall into two categories: those that seek to reduce the grammatical mechanisms necessary to account for Obligatory Control contexts and those that seek to account for these contexts by agreement without reducing the grammatical mechanisms.

The Persian data is examined to determine whether the empty subject has the properties of pro, NP- trace/copy, or PRO. The properties of Obligatory Control and the three empty categories are introduced and arguments are outlined for positing a structural subject in Obligatory Control contexts. The current theories accounting for Obligatory Control contexts are described and each is evaluated for its ability to account for the Persian Data. The analysis shows that the Persian Obligatory Control data is best accounted for by positing PRO as the empty subject in those contexts, licensed by agreement.