Modeling sentence processing difficulty with a conditional probability calculator

Zhong Chen	Cornell University
Tim Hunter	University of Minnesota
Jiwon Yun	Stony Brook University
John Hale	Cornell University

Abstract

We present the conditional probability calculator CCPC for predicting word-by-word processing difficulties in human sen- tence comprehension. This system, in conjunction with weighted grammars and the linking hypothesis Entropy Reduction (Hale, 2006), derives the subject-object asymmetry in Italian relative clauses, including the animacy effect of head nouns.

Files

Modeling sentence processing difficulty with a conditional probability calculator (219 KB)