Syntax encodes information structure: Evidence from on-line reading comprehension

Although sentences are thought to be generally easier to process when given information precedes new information, closer examination reveals that these preferences only manifest within some syntactic structures. Here, we examine the consequences of the relative ordering of given and new information...

Full description

Bibliographic Details
Main Authors: Brown, Meredith (Author), Savova, Virginia (Contributor), Gibson, Edward A. (Contributor)
Other Authors: Massachusetts Institute of Technology. Department of Brain and Cognitive Sciences (Contributor), Massachusetts Institute of Technology. Department of Linguistics and Philosophy (Contributor)
Format: Article
Language:English
Published: Elsevier, 2016-05-12T13:45:55Z.
Subjects:
Online Access:Get fulltext
Description
Summary:Although sentences are thought to be generally easier to process when given information precedes new information, closer examination reveals that these preferences only manifest within some syntactic structures. Here, we examine the consequences of the relative ordering of given and new information (information structure) for the on-line comprehension of prepositional-object (PO) and double-object (DO) dative sentences. Experiment 1 demonstrated using self-paced reading that the on-line comprehension of DO structures, but not PO structures, is more difficult with new-before-given information structure. Experiment 2 assessed the comprehension of dative sentences with animate themes to evaluate two potential sources of this interaction: information-structural constraints encoded within syntactic representations (argument structure hypothesis) vs. word-to-word contingency statistics (linear position hypothesis). Despite experiment-wise differences in word-to-word contingency statistics, the interaction between syntactic structure and information structure persisted in Experiment 2. Taken together, these results suggest that syntactic representations can include information-structural constraints on their arguments.
Multidisciplinary University Research Initiative (Grant N00014-07-1-0937)
United States. Defense Advanced Research Projects Agency (Grant NBCH-D-03-0010)