Automaton Theories of Human Sentence Comprehension (Studies in Computational Linguistics) (Paperback)
By relating grammar to cognitive architecture, John T. Hale shows step-by-step how incremental parsing works in models of perceptual processing and how specific learning rules might lead to frequency-sensitive preferences. Along the way, Hale reconsiders garden-pathing, the parallel/serial distinction, and information-theoretical complexity metrics, such as surprisal. This book is a must for cognitive scientists of language.
John T. Hale is associate professor in the Department of Linguistics at Cornell University.
“Hale is a genuine computational psycholinguist: a rare type of cognitive scientist who uses computational models of language processing to explain what sorts of structures are hard and easy for humans to produce and/or comprehend. In this monograph, he presents the formal foundations underlying this type of research, drawing on results in computer science, linguistics, psychology, and information theory. The presentation is clear and systematic, making this an extremely useful book for scholars in one of these fields seeking to understand work in this highly interdisciplinary area.”
— Tom Wasow, Stanford University
— Tom Wasow, Stanford University