Statistical Representation of Grammaticality Judgements: the Limits of N-Gram Models

TY - CHAP

T1 - Statistical Representation of Grammaticality Judgements

T2 - Fourth Annual Workshop on Cognitive Modeling and Computational Linguistics

AU - Clark, Alexander

AU - Giorgolo, Gianluca

AU - Lappin, Shalom

PY - 2013/8

Y1 - 2013/8

N2 - We use a set of enriched n-gram models to trackgrammaticality judgements for different sorts ofpassive sentences in English. We construct these models by specifying scoring functions to map thelog probabilities (logprobs) of an n-gram model fora test set of sentences onto scores which dependon properties of the string related to the parametersof the model. We test our models on classificationtasks for different kinds of passive sentences.Our experiments indicate that our n-gram modelsachieve high accuracy in identifying ill-formed passivesin which ill-formedness depends on local relationswithin the n-gram frame, but they are far lesssuccessful in detecting non-local relations that produceunacceptability in other types of passive construction.We take these results to indicate some ofthe strengths and the limitations of word and lexicalclass n-gram models as candidate representations ofspeakers’ grammatical knowledge.

AB - We use a set of enriched n-gram models to trackgrammaticality judgements for different sorts ofpassive sentences in English. We construct these models by specifying scoring functions to map thelog probabilities (logprobs) of an n-gram model fora test set of sentences onto scores which dependon properties of the string related to the parametersof the model. We test our models on classificationtasks for different kinds of passive sentences.Our experiments indicate that our n-gram modelsachieve high accuracy in identifying ill-formed passivesin which ill-formedness depends on local relationswithin the n-gram frame, but they are far lesssuccessful in detecting non-local relations that produceunacceptability in other types of passive construction.We take these results to indicate some ofthe strengths and the limitations of word and lexicalclass n-gram models as candidate representations ofspeakers’ grammatical knowledge.

M3 - Conference paper

SN - 978-1-937284-61-9

T3 - Proceedings of the Association of Computational Linguistics Workshop on Cognitive Modeling and Computational Linguistics

SP - 28

EP - 36

BT - Proceedings of the Fourth Annual Workshop on Cognitive Modeling and Computational Linguistics (CMCL)

PB - Association for Computational Linguistics

CY - Sofia

Y2 - 8 August 2013 through 8 August 2013

ER -