DPs aren’t licensed, they’re selected (or not) MIT, February 21st, 2008 Thomas McFadden Universität Stuttgart tom@ifla.uni-stuttgart.de 1 Two approaches to the (non)-overtness of subjects A well-known set of facts that syntactic theory must explain is the distribution of various overt and non-overt DP types across syntactic positions. We can distinguish two main approaches: 1. DPs are inherently defective and thus can’t show up in a position unless some kind of licensing is available there. ☞ What we must explain under this approach are the places where lexical DPs are licit. 2. DPs have no special needs. They just have to be integrated into the interpretation like anything else. So they can show up anywhere this is possible – as long as independently motivated principles of grammar are satisfied. ☞ Under this approach, what we must explain are the places where lexical DPs are illicit. The central data for deciding between these approaches involve subject positions, in particular in non-finite and other embedded clauses. In very simple terms, certain kinds of infinitives don’t allow a lexical subject, but are fine if the subject position is left empty: (1) a. * Joe tried [Hank to buy the beer]. b. Joe tried [to buy the beer]. (2) a. * [Hank to buy the beer] would be difficult. b. [To buy the beer] would be difficult. Both approaches laid out above can handle these basic facts in a reasonable fashion: 1. There’s nothing to license the subjects of (these) non-finite clauses, so they’re bad by default. Finite clauses have subjects because they have something extra (Tense, Agreement. . . ) which can license them. –or– 2. There’s something special about (these) non-finite clauses that prevents them from having subjects. Finite clauses lack this property, so subjects appear there unproblematically. Of course, it turns out that lexical subjects are possible in non-finite clauses that are introduced by the preposition for (3a) or follow a transitive verb like expect (3b): 1