Computational models of human sentence processing: am model comparison approach

• Applicant: Professor Dr. Shravan Vasishth
• Subject Area: General and Comparative Linguistics, Experimental Linguistics, Typology, Non-European Languages
• Term: from 2007 to 2011
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 38541091

• Final Report Year: 2011

Final Report Abstract

To summarize, we developed several scaled-up models of parsing using different grammar formalisms and different parsing architectures. We implemented two distinct complexity metrics, surprisal and retrieval difficulty, in these parsers, and showed that these metrics are good predictors of eyetracking data. We also demonstrated how parallelism in parsing architectures can be investigated empirically, and provided the first results suggesting that a certain amount of parallelism in parsing may yield better empirical results than a strictly serial parser. We also found that the early-late distinction in reading time measures may not be well-motivated; even early measure such as single-fixation durations might reflect post-lexical (e.g., syntactic) processing difficulty. Several unexpected results emerged in this project. First, grammatical knowledge of head-finality in German versus English was shown to be an important determinant of how well native speakers can maintain predictions of upcoming phrases; we also presented a computational model which demonstrated how this preference emerges as a function of experience. Second, a new theory of interference was developed to account for surprising data that emerged from a series of self-paced reading and eyetracking studies. Third, we developed a novel algorithm for characterizing scanpaths in reading, an idea that was very well-received in the field.

Publications

• 2008. Parsing costs as predictors of reading difficulty: An evaluation using the Potsdam Sentence Corpus. Journal of Eye Movement Research, 2(1):1-12
  Marisa Ferrara Boston, John T. Hale, Umesh Patil, Reinhold Kliegl, and Shravan Vasishth
  
• Compound effect of probabilistic disambiguation and memory retrievals on sentence processing: Evidence from an eye-tracking corpus. In: A. Howes, D. Peebles, and R. Cooper, editors, Proceedings of 9th International Conference on Cognitive Modeling, Manchester, UK, 2009.
  Umesh Patil, Shravan Vasishth, and Reinhold Kliegl
  
• Processing grammatical and ungrammatical center embeddings in English and German: A computational model. In: A. Howes, D. Peebles, and R. Cooper, editors, Proceedings of 9th International Conference on Cognitive Modeling, Manchester, UK, 2009
  Felix Engelmann and Shravan Vasishth
  
• 2010. Short-term forgetting in sentence comprehension: Crosslinguistic evidence from head-final structures. Language and Cognitive Processes, 25(4):533-567
  Shravan Vasishth, Katja Suckow, Richard L. Lewis, and Sabine Kern
  
• Morphological ambiguity and working memory. In: Peter de Swart and Monique Lamers, editors, Case, Word Order, and Prominence: Psycholinguistic and theoretical approaches to argument structure, volume 2 of Studies in Theoretical Psycholinguistics. Springer, 2010
  Pavel Logačev and Shravan Vasishth
  
• 2011. Parallel processing and sentence comprehension difficulty. Language and Cognitive Processes, 26(3):301-349
  Marisa F. Boston, John T. Hale, Shravan Vasishth, and Reinhold Kliegl