On the syntax and semantics of appositive relative clauses

In this paper I analyze appositive relative clauses that modify quantified nominals. This new data supports a theory of appositives as a phenomenon at the interface between syntax/semantics and discourse. More precisely, I propose that the appositive relative pronoun is E-type (see Sells 1985a and Demirdache 1991), and that it needs to follow its antecedent. I also propose that appositives are propositions (type t), merged to the DP they modify in narrow syntax, and moved to the matrix CP after Spell-Out and at the discourse level. I claim that merge within discourse is linear order dependent, and that at the discourse level, both PF and LF representations are available to the computational component. The designed system accounts for differences between appositives and restrictives and makes the strong empirical prediction that prenominal relative clauses cannot be true appositives.