The Category P Features, Projections, Interpretation

The Category P Features, Projections, Interpretation The Category P Features, Projections, Interpretation

12.09.2014 Views

223 b. dan hevi et ha-oto dan brough Acc the-car “Dan brought the car.” . (103) a. Dan wiped the table i [ AP clean i ] b. Dan wiped the table. Given the similarity of the le NP/PP in the OPC to the resultative secondary predicate, I will assume that the le NP/PP in the OPC is a secondary predicate. Following Rothstein (1983, 2001) this means that it is predicated of a thematic argument of the main clause predicate. 5.5.2.2 The syntax of secondary predicates: The syntactic issue concerning secondary predicates revolves around the question whether a secondary predicate is a daughter of a SC node (predicated of a PRO subject) (Chomsky 1981, Stowell 1983, among others), as in (104a), or rather an adjunct, generated without a subject as a daughter of some projection of the V, and predicated directly of the relevant argument of the main predicate (Williams 1980, 1983, 1994, Rothstein 1983, 1995, 2001, among others), as in (104b): (104) a. John [ VP wiped the table] [PRO clean] b. John [ VP [ V wiped] [ DP the table] [ AP clean]] As pointed out by Williams, the structure in (104a), at least for object-oriented secondary predicates, is problematic. PRO contained in the SC appears to be governed, as the SC has to be contained inside the VP. 66 Rothstein argues that SCs must be licensed by theta-marking, which is impossible in (104a). In what follows I will adopt the structure in (104b) for resultative and purpose secondary predicates ( le NP/PP). The syntactic representations of the Hebrew and English OPC are given in (105): 66 Chomsky (1981) raises the possibility that while theta-marked SCs (complements of ECM verbs) are transparent for government, non-theta-marked SCs (i.e. object-oriented depictives and resultatives) are opaque for government, thus allowing PRO.

224 (105) a. dan [ VP [ V hevi] [ DP et ha-oto] i [ leNP le-tikun] [Exti] ] b. Dan [ VP [ V brought] [ DP the car] i [ PP Op i to repair t i ] [Exti] ] . (105) correctly implies that the DP the car is the argument of the matrix verb, and that the le NP/PP is a VP-internal adjunct of some kind. Faraci (1974) presents five arguments to support the claim that the embedded constituent in the OPC (‘purpose clause’ in his terminology) is indeed a VP-internal adjunct, rather than a clausal one. I will mention here two of his arguments. Following Chomsky (1965), only phrases outside the VP could be preposed to the beginning of the sentence. Given this, the ungrammaticality of (106b) indicates that to practice on in (106a) is inside the VP: (106) a. John bought the piano to practice on b. *To practice on, John bought the piano (Faraci 1974, (7b), (8b)) The embedded constituent in the OPC can be part of a VP in focus position in pseudo-cleft sentences: (107) a. Marc bought Fido to play with b. What Marc did was to buy Fido to play with (Faraci 1974, (31a,b)) In sum, the le NP/PP in the OPC is a VP-internal adjunct. Its external slot (Ext) is discharged by assignment to the internal argument of the matrix verb (e.g. the car in (105)). 67 As already mentioned in 5.5.1.6, the saturated argument of the le NP/PP (x ARB ) is usually interpreted in the OPC as (quasi) existential (by default) (Dan brought the car to repair). Note, however, that following the discussion in 5.5.1.6, the saturated argument of the le NP/PP (x ARB ) in the OPC should, in principle, be able to be interpreted as arbitrary (universally-bound), if the appropriate licensing is provided. This is indeed the case, as shown in (108): 67 Williams (1983) argues that the correct way to capture the independence of a secondary predicate from the primary one is to revise the Theta-Criterion in such a way that it states that no argument can be theta-marked more than once by the same head.

223<br />

b. dan hevi et ha-oto<br />

dan brough Acc the-car<br />

“Dan brought the car.”<br />

.<br />

(103) a. Dan wiped the table i [ AP clean i ]<br />

b. Dan wiped the table.<br />

Given the similarity of the le NP/PP in the OPC to the resultative secondary<br />

predicate, I will assume that the le NP/PP in the OPC is a secondary predicate.<br />

Following Rothstein (1983, 2001) this means that it is predicated of a thematic<br />

argument of the main clause predicate.<br />

5.5.2.2 <strong>The</strong> syntax of secondary predicates: <strong>The</strong> syntactic issue concerning<br />

secondary predicates revolves around the question whether a secondary predicate is a<br />

daughter of a SC node (predicated of a PRO subject) (Chomsky 1981, Stowell 1983,<br />

among others), as in (104a), or rather an adjunct, generated without a subject as a<br />

daughter of some projection of the V, and predicated directly of the relevant argument<br />

of the main predicate (Williams 1980, 1983, 1994, Rothstein 1983, 1995, 2001,<br />

among others), as in (104b):<br />

(104) a. John [ VP wiped the table] [PRO clean]<br />

b. John [ VP [ V wiped] [ DP the table] [ AP clean]]<br />

As pointed out by Williams, the structure in (104a), at least for object-oriented<br />

secondary predicates, is problematic. PRO contained in the SC appears to be<br />

governed, as the SC has to be contained inside the VP. 66 Rothstein argues that SCs<br />

must be licensed by theta-marking, which is impossible in (104a).<br />

In what follows I will adopt the structure in (104b) for resultative and purpose<br />

secondary predicates ( le NP/PP). <strong>The</strong> syntactic representations of the Hebrew and<br />

English OPC are given in (105):<br />

66 Chomsky (1981) raises the possibility that while theta-marked SCs (complements of ECM verbs) are<br />

transparent for government, non-theta-marked SCs (i.e. object-oriented depictives and resultatives) are<br />

opaque for government, thus allowing PRO.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!