Home

About

Advanced Search

Browse by Discipline

Scientific Societies

E-print Alerts

Add E-prints

E-print Network
FAQHELPSITE MAPCONTACT US


  Advanced Search  

 
Chunks and Dependencies: Bringing Processing Evidence to Bear on Syntax
 

Summary: Chunks and Dependencies:
Bringing Processing Evidence to Bear on Syntax
Steven Abney
University of Tšubingen
October 17, 1991
1 Introduction
At least some psycholinguists exploring how sentences are structured in
linguistic behavior have concluded that the "performance structures" that
emerge from experimental data differ from the syntactic structures hypothe-
sized by linguists. For example, one measure of structure is the location and
relative prominence of pauses when subjects read sentences aloud. Experi-
ments have indicated that the syntactic prominence of boundaries is only a
moderately good predictor of the prominence of pauses at those boundaries
(Grosjean, Grosjean, & Lane [9]). Other experiments have looked at parsing
by linguistically-naive subjects: when asked to group words together (Mar-
tin [12]), or to subdivide sentences at their natural joints [9]. Yet another set
of experiments examined the probability of errors in performance, the high-
est probability of error representing the most significant boundaries. Levelt
[11] investigated comprehension of spoken sentences mixed with noise, and
Dommergues & Grosjean [5] looked at errors in recall of sentences heard

  

Source: Abney, Steven P. - School of Information, University of Michigan

 

Collections: Computer Technologies and Information Sciences