12.09.2014 Views

The Category P Features, Projections, Interpretation

The Category P Features, Projections, Interpretation

The Category P Features, Projections, Interpretation

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

26<br />

sharp contrast to the core lexical heads, in these languages prepositions pied-pipe<br />

obligatorily under wh-movement (e.g. al ma dan siper? ‘about what [did] Dan tell’?)<br />

(Webelhuth 1992, Grosu 1994, Kayne 1994, Koopman 2000, Horvath 2001).<br />

According to the criteria discussed so far, there is a full alignment between the<br />

properties of P and the functional properties. This by itself is sufficient to maintain the<br />

hypothesis that P is a functional category. To complete the picture, I turn now to the<br />

more controversial property of P, labeled in (2) as head-complement relation.<br />

2.1.3 Head-complement relation<br />

<strong>The</strong> standard assumption in linguistic theory is that only lexical heads can assign<br />

theta-roles. However, given a certain variety of prepositions (e.g. because, after,<br />

under, above) referred to informally as semantically contentful (see chapter 1), and<br />

certain contexts (e.g. locative), it is rather common to describe the relation between<br />

prepositions and their complements as theta-assignment (cf. Emonds 1985). 19 <strong>The</strong><br />

classification of P as functional seems inconsistent with its alleged theta-assigning<br />

ability. In what follows I will show that the inconsistency is only apparent, since even<br />

when the relation between P and its complement is a predicate-argument relation, it is<br />

not a theta-relation.<br />

Note first that there are syntactic contexts (6) where theta-assignment clearly<br />

does not seem to be appropriate to describe the P-complement relation (as will be<br />

discussed in details in chapter 3). <strong>The</strong>se contexts do not present a problem for my<br />

hypothesis.<br />

(6) a. Dan relied on Mary.<br />

b. Marge believes in love.<br />

Let us focus then only on the contexts where the discussed relation is often<br />

assumed to be a theta-relation (e.g. locative PPs), illustrated below:<br />

(7) a. Dan found a coin in/near the garden.<br />

b. Lisa put the pen on/under the table.<br />

19 I use the familiar notion ‘a semantically contentful P’ just as a convenient label for the present<br />

discussion. In the approach to P developed here, semantic contentfulness does not refer to an inherent<br />

property of individual Ps, but rather follows from the function of P (see 2.2.2).

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

Saved successfully!

Ooh no, something went wrong!