A Probabilistic Parser as a Model of Global Processing Difficulty
Frank Keller · eScholarship (California Digital Library) · 2003
We present a model of global processing difficulty in human parsing.This model is based on a probabilistic context-free grammar and is trained on a realistic corpus sample.It achieves broad coverage and good parsing accuracy on unseen text, and its predictions are significantly correlated with experimental data on word order preferences in German.The model makes predictions about the differential behavior of verb final and verb initial sentences and provides evidence for the importance of lexical information in sentence processing.