Minimization of dependency length in written English q

Gibson’s Dependency Locality Theory (DLT) [Gibson, E. 1998. Linguistic complexity: locality of syntactic dependencies. Cognition, 68, 1–76; Gibson, E. 2000. The dependency locality theory: A distance-based theory of linguistic complexity. In A. Marantz, Y. Miyashita, & W. O’Neil (Eds.), Image, Language, Brain (pp. 95–126). Cambridge, MA: MIT Press.] proposes that the processing complexity of a sentence is related to the length of its syntactic dependencies: longer dependencies are more difficult to process. The DLT is supported by a variety of phenomena in language comprehension. This raises the question: Does language production reflect a preference for shorter dependencies as well? I examine this question in a corpus study of written English, using the Wall Street Journal portion of the Penn Treebank. The DLT makes a number of predictions regarding the length of constituents in different contexts; these predictions were tested in a series of statistical tests. A number of findings support the theory: the greater length of subject noun phrases in inverted versus uninverted quotation constructions, the greater length of direct-object versus subject NPs, the greater length of postmodifying versus premodifying adverbial clauses, the greater length of relative-clause subjects within direct-object NPs versus subject NPs, the tendency towards ‘‘short-long’’ ordering of postmodifying adjuncts and coordinated conjuncts, and the shorter length of subject NPs (but not direct-object NPs) in clauses with premodifying adjuncts versus those without. ! 2006 Elsevier B.V. All rights reserved.