Structural Connections in Syntax and Processing: Studies in Russian and Japanese

This thesis consists of two parts. Part I provides an analysis of the locative inversion construction, the conjunction agreement construction, and the genitive of negation construction in Russian. These constmctions are argued to share one formal property: they contain an element other than the highest nominal argument satisfying the Extended 3rojection Principle. The EPP is shown to be independent not only of morphological features (such as Case or agreement), but also of categorial features (such as N or D); movement to the EPP position is shown to be subject to the Minimal Link Condition for all calegories, so that non-canonical subjects can move to the EPP position only when they are as "close" to it as the highest NP argument. Given our assumptions about VP structure, this happens only in sentences containing unaccusative verbs. The syntactic properties of locative inversion, conjunction agreement, and genitive of negation are shown to follow from the manner in which various principles operating in Russian syntax, such as'the discourse principles, the properties of covert (as opposed to overt) feature-checking, the morphological Case system of Russian, and the existential closure applying to all !P-internal positions, operate on sentences whose nominal "subjects" remain in their base-generated positions in overt syntax. . Part I1 investigates the processing complexity of unambiguous Japanese sentences. The investigation utilizes the theory of processing complexity developed in Gibson & Thomas (1996a), within which the memory cost associated with an incomplete syntactic dependency increases as a fbnction of the number of lexical items that are processed between the point where the relationship is posited and the point where it is satisfied. Two types of processing complexity contrasts found in Japanese are discussed: those associated with the nun~ber of "stacked" sentence-initial NPs and those associated with the degree and type of center-embedding present in a sentence. Both experimental and intuitive data are provided as evidence for the existence of the complexity contrasts. It is shown that the contrasts are accounted for within the Locality Theory of Gibson & Thomas. On a more general level, :his tvork describes the properties that any theory of processing complexity must have to successfUlly deal with the available Japanese data and offers a constrained and principled explanatibn of apparent variation in the processing complexity of similar structures across lang sges. Thesis supervisors: Dr David Pesetsky and Dr. Edward Gibson Acknowledgments I owe much to many people, but among the people I owe most to (linguistically speaking) are my two advisors, Dhvid Pesetsky and 'Ted Gibson. It goes without saying that this work would never (ever) have bccri began, continued and (most importantly) finished, were it not for them. I appreciate their insights, ,pidance, patience, enthusiasm, pushing and pulling, superb management, and timely reminders that dissertations not finished by the deadline are no use at all. They have made work in their respective fields look exciting, challenging, and occasionally beautihl in fact, so much so, that it was impossible not to join in tke fin. David has taught me syntax when I was an undergraduate at MIT, and the fact that I s:uck with it for so inany years probably has a lot to do with him. One of the first problem sets I remember doing for his class was on genitive of negation (see chapter 4), one of the first papers I remember writing for him was on word order in Russian (see chapter 2), and he was also the one who lent me the book by Crockett, in which I discovered a wealth of material on conjunction agreement (see chapter 3). Thank you, David (and I swear I don't have your Peshkovsky). Ted has introduced me to the field of language processing and has demonstrated, by personal example, the attraction and excitement of having one's theories confirmed (or disconfirmed) by tangible experimental data. If it weren't for his interest and encouragement, my notions about Japanese processing would never have become explicit enough to be testable and would definitely never have been tested. His habit of saying: "We'll jump off this bridge when we come to it" hi appropriate moments has also helped. I particularly appreciate the fact that he had trusted me with his newborn theo~y, on which the processing part of this work is based. The other member of my committee, Alec Marantz, Shigeru Miyagawa, and Ken Wexler, also deserve enormous thanks. Alec has introduced me to Distributed Morphology and has offered a wealth of suggestions, constructive criticism, direction, and encouragement. I shudder to think how much clarity this work would have lost without his input. Shiger~ has acquainted me with Japanese syntax and shown me how rewarding and fascinating it can be to try to discover the structure of a free word order language. Conversations with him have inspired a lot of projects and a lot of ideas and it has always been a joy to work with him. Ken's insights, moral support, and enthusiasm about a good theory in any branch of linguistics have been a tremendous help, too. Over the years, his excitement about the field has been an inspiring example to me. I am truly gratehl to all those who have taught me at MI'f: Takako Aikawa, Bob Berwick, Noam Chomsky, Suzanne Flynn, Ken Hale, Morris Halle, Jim Harris, Irene Him, Jim Higginbot ham, Michael Kenstowicz, Utpal Lahiri, Wayne O'Neil, and Steve Pinker. Whatever professional accomplishments I possess, I owe to them. Morris deserves special gratitude for bringing his knowledge, experience, and intuition to the task of helping me shape my ideas about Russian morphology. My experience of the Iast five years has been made richer, more bearable, and, occasionally, even fbn, by the invaluable discussions, company, and support of my fellow students: Judy Baek, Chris Bader, Pilar Barbosa, Jonathan Bobaljik, Marie Claude Boivin, David Braun, Ben Bruening, Tony Bures, Andrew Carnie, Chris Collins, Marie-Helen Cote, Diana Cresti, Hamida Demirdash, Danny Fox, Tom Green, Martin Hackl, Paul Hagstrom, Heidi Harley, Doug Jones, Yoonjung Kang, Shawn Kern, Miori Kubo, Idan Landau, Andrea Leszek, Ingvar Lofstedt, Gaurav Mathur, Martha McGinnis, David McKay, Seth Minkoff, Kumiko Murasugi, Renate Musan, Jon Nissenbaum, Toshi Oka, Rob Pensalfini, Orin Percus, Colin Phillips, Norvin Richards, Taylor Roberts, V.J. Sarma, Uli Sauerland, Carson Schutze, Hooi-Ling Soh, Luciana Storto, Hiroaki Tad?, Chris Tancredi, Hubert Truckenbrodt, Dylan Tsai, Hiro Ura, Fleur Veraart, Akira Watanabe, Kathleen Williams, Dag Wold, Susie Wurmbrand, and Brian Yeager. Thank you, all I would also like to thank those who have taught me Linguistics at Brandeis: Jane Grimshaw, Ray Jackendoff, and Moira Yip. They have shown me the excitement and beauty of their fields, for which I am very gratefill. My Brandeis classmates, Feargal Murphy, Vieri Samek-Lodovici, and Hubert Tnickcnbrodt, have shown me the warmth and humor that can be found in a Linguistics community. I would also like to mention my gratitude to Bob Whitman, who taught me Introduction to Linguistics 11 years ago at Wesleyan University. I also extremely grateful to the countless native speakers of Russian and Japanese who I have harassed kith requests for judgements, and who have helped me gather the experimental data on which the processing part of this work is based. I am especially gratehl to Yasuaki Abe, Takako Aikawa, Yuko Maekawa, and Lisa Shimizu for their help and their patience. My life in the department has been made easier and smoother by the help and support of Mary Grenham, Jennifer Purdy, Bev Stohl, Linda Thomas. Special thanks go to Bev for making Nadia's visits to the department fun. Finally and most importantly, this work would never have been completed without the patience, love, and support of my family. My grandmother Sarra, my father Alexander, my mother Marina, and my daughter Nadia have taught me more than anyone about living and working with integrity and a sense of fun. I hope that some portion of these qualities has managed to find its way into this work. I dedicate this thesis to them, with much love and gratitude. Table of