Restructuring Parameters and Complex Predicates, a Transformational Approach
H. S. Choe, 1988
This thesis discusses the two types of complex predicate -- morphologically- complex predicates (Type I) and 'restructured' predicates composed of morphologically-independent predicates (cf. Rizzi (1982); Type 11). It introduces and develops a theory of complex predicates/words. The theoretical framework is a recent development of the transformational grammar-- the principles-parameters approach to Universal Grammar (UG) (cf. Chomsky (1986a and b)).
We view complex predicates as derived through transformations operating on minimal elements. Type I1 complex predicates are obtained through a head-to-head transformation, which we call The Restructuring Rule (RR). The RR affects categories in a certain way, linking them, and thereby giving rise to Type I1 complex words, whose properties and syntax characterize the 'restructuring' phenomenon discussed in the literature. Type I complex predicates are obtained through head-movement; for both empirical. and theoretical reasons, head-movement is reunderstood as having a more restricted function than the one assumed in the literature: Head-movement is instantiated as adjunction but not substitution. Head-adjunction creates the segments of both target and trigger minimal categories, forming chains called H-chains. Empirical data suggest that Type I complex predicates are also formed through RR accompanied by head-movement, which we call overt RR. Consequently, complex predicates are obtained through head movement,(move-head), RR (affect-category), or overt RR (move-category); complex predicates form three different types of complex word according to their syntactic derivations. Hove-head operates on terminal strings (heads), affect-category on heads In X-bar theory's sense (X-heads), and move-category on both heads and X-heads. Those head-to-head transformations imply a certain concept of tree structure (that differs from the usual concept of tree structure), under which a head and the X-head projected from the head are independent entities of each other, subject to different principles of grammar. The proposed head-to-bead transformations assume the projection principle on X-heads and the theta-criterion on X heads, which apply to both lexical (X-)heads and functional (X-)heads such as C and I, and which apply to categories but not to segments derived through head-movement.