Table Of ContentISLAND CONSTRAINTS
STUDIES IN THEORETICAL PSYCHOLINGUISTICS
VOLUME 15
Managing Editors
Thomas Roeper, Dept. of Linguistics, University of Massachusetts at Amherst
Kenneth Wexler, Dept. of Brain and Cognitive Science, MIT, Cambridge, Mass.
Editorial Board
Robert Berwick, Artifical Intelligence Laboratory, MIT, Cambridge, Mass.
Manfred Bierwisch, Zentralinstitut fur Sprachwissenschaft, Akademie der
Wissenschaften, Berlin
Merrill Garrett, University of Arizona, Tucson
Lila Gleitman, School of Education, University of Pennsylvania
Mary-Louise Kean, University of California at Irvine
Howard Lasnik, University of Connecticut at Storrs
John Marshall, Neuropsychology Unit, Radcliffe Infirmary, Oxford
Daniel Osherson, M.I.T., Cambridge, Mass.
Yukio Otsu, Keio University, Tokyo
Edwin Williams, Princeton University
The titles published in this series are listed at the end of this volume.
ISLAND
CONSTRAINTS
Theory, Acquisition and Processing
Edited by
HELEN GOODLUCK
Dept. of Linguistics, University of Ottawa
and
MICHAEL ROCHEMONT
Dept. of Linguistics, University ofB ritish Columbia
SPRINGER-SCIENCE+BUSINESS MEDIA, B.V.
Library of Congress Cataloging-in-Publication Data
Island constralnts : theory, acqulsltion, and processing / edited by
Helen Goodiuck and Michael Rochemont.
p. cm. -- (Studies in theoretical psycholinguistics ; v. 15)
Inc I udes index.
ISBN 978-90-481-4148-7 ISBN 978-94-017-1980-3 (eBook)
DOI 10.1007/978-94-017-1980-3
1. Grammar. Comparative and gensral--Syntax. 2. Language
acqulsition. 3. Generative grammar. I. Goodluck, Helen.
II. Rochemont, Mlchael S. (Michael Shaun), 1950- III. Series.
P291.184 1992
425--dc20 92-7222
ISBN 978-90-481-4148-7
Printed on acid-free paper
All Rights Reserved
© 1992 Springer Science+Business Media Dordrecht
Originally published by Kluwer Academic Publishers in 1992
Softcover reprint of the hardcover 1st edition 1992
No part of the material protected by this copyright notice may be reproduced or
utilized in any form or by any means, electronic or mechanical,
including photocopying, recording or by any information storage and
retrieval system, without written permission from the copyright owner.
TABLE OF CONTENTS
ACKNOWLEDGEMENTS vii
HELEN GOODLUCK and MICHAEL ROCHEMONT / Island
Constraints: an Introduction 1
ROBERT C. BERWICK / No Variable Is an Island: Computa-
tional Complexity and Island Constraints 35
JOHANNE S. BOURDAGES / Parsing Complex NPs in French 61
NOMI ERTESCHIK-SHIR / Resumptive Pronouns in Islands 89
JANET DEAN FODOR / Islands, Learnability and the Lexicon 109
HELEN GOODLUCK, MICHELE FOLEY and JULIE SEDI-
VY / Adjunct Islands and Acquisition 181
PAUL HIRSCHBUHLER and DANIEL VALOIS / Argument
Extraction out of Indirect Questions in French 195
ROBERT KLUENDER / Deriving Island Constraints from
Principles of Predication 223
DIANE LILLO-MARTIN / Sentences as Islands: on the Boun-
dedness of A' -Movement in American Sign Language 259
DANA McDANIEL and CECILE McKEE / Which Children
Did They Show Obey Strong Crossover? 275
MARTIN PICKERING and RICHARD SHILLCOCK / Process-
ing Subject Extractions 295
BRADLEY L. PRITCHETT / Parsing with Grammar: Islands,
Heads, and Garden Paths 321
v
vi TABLE OF CONTENTS
CLIFTON PYE, HINTAT CHEUNG, and SUSAN KEMPER /
Island Constraints at Eighty 351
MICHAEL ROCHEMONT / Bounding Rightward A-dependen-
cies 373
DOUGLAS SADDY / Sensitivity to Islands in an Aphasic In-
dividual 399
LAURIE A. STOWE / The Processing Implementation of
Syntactic Constraints: the Sentence Matching Debate 419
LYDIA WHITE / Subjacency Violations and Empty Categories
in Second Language Acquisition 445
LIST OF CONTRIBUTORS 465
NAME INDEX 469
SUBJECT INDEX 475
ACKNOWLEDGEMENTS
The conference on island constraints out of which this volume grows
was supported by a grant from the School of Graduate Studies at the
University of Ottawa. We would like to thank Kenneth Forster, Nigel
Duffield, Y ounghee Na and Geoffrey Huck, and David Swinney for
their presentations at the conference. Faculty and students in the
Linguistics Department at Ottawa, most especially Philip Hauptman,
Janet Benger and Claudia Iulianella, generously contributed their time
and energy to organizational matters.
VB
HELEN GOODLUCK AND MICHAEL ROCHEMONT
ISLAND CONSTRAINTS: AN INTRODUCTION!
Island constraints have formed a central component of grammatical
theory since the ground breaking work of Ross (1967) developing a
proposal in Chomsky (1964). The papers collected in this volume
address island constraints from a variety of theoretical linguistic and
psycho linguistic perspectives. They result from a conference held at the
University of Ottawa during the fall of 1989, at which specialists from
several fields met to engage in interdisciplinary dialogue over the nature
and manifestations of island constraints. In this brief introduction, we
will attempt to summarize the intellectual setting that serves as back
drop to the investigations reported on here. To do so both fully and
briefly presents a difficult challenge. But we think the inevitable risk of
misrepresentation and offense is worth the effort. For it is only by
making results accessible outside their domains of specialization that we
can hope to foster the cross-disciplinary dialogue that we believe so
essential to the growth of our knowledge and understanding of the
mechanisms of language.
I. ISLANDS AND GRAMMATICAL THEORY
Syntactic accounts
Many languages, including English, display syntactic constructions in
which a phrase in non-argument position is associated with a vacant
argument position elsewhere in the sentence for purposes of semantic
interpretation. Well studied English examples include matrix and em-
1
Helen Goodluck and Michael Rochemont (eds.), Island Constraints, 1-33.
© 1992 Kluwer Academic Publishers.
2 HELEN GOODLUCK AND MICHAEL ROCHEMONT
bedded wh questions, topicalization, and relative clauses, as illustrated
in the respective sentences below.
(1) (a) In which room did you decide to put it_?
(b) I never realized how happy he makes her _.
(c) John, I don't think I like _ very much.
(d) This is the man who I told you about _.
In (a), the wh PP is an understood complement of put, in (b) the wh
AP is an understood complement of make, in (c) the sentence initial
NP is the understood object of like, and in (d) the relative pronoun is
understood as the object of about. In all these cases, the dislocated
phrase (the antecedent) is in a non-thematic non-argument position,
and is understood to fill a thematic function elsewhere in the sentence
(signalled by '_' in these examples). (In general, an argument position
expresses a thematic role designated by the verb.) Moreover, the
thematic (argument) position it is interpretively linked to is not filled
with any lexical material-it is a gap.
A central feature of the constructions in (1) is that the relation of
antecedent to gap is apparently unbounded, in the sense that there is no
theoretical limit (though there is certainly a practical one) to the
amount of material that may intervene. In (2) for example, the dis
located sentence initial phrase is related to a gap across at least three
levels of clausal embedding.
(2) John, I think you told me (that) Bill said (that) Mary claimed
(that) Sally is planning to recommend __ for a job.
But the lack of bounding in such cases is only apparent. This is perhaps
best appreciated when such cases are contrasted with comparable
examples in which there is no bounding effect whatsoever. For instance,
consider the as for example in (3a), and the syntactically and inter
pretively equivalent example of topicalization in (3b).
(3) (a) As for John, I have a good friend who likes him.
(b) *John, I have a good friend who likes_.
The contexts in which this bounding of antecedent-gap relations sur
faces have been characterized since Ross (1967) in terms of 'island
ISLAND CONSTRAINTS: AN INTRODUCTION 3
constraints', which serve to block the association of antecedent to gap
under specific syntactic conditions. Of the restrictions identified by
Ross and others, the ones we will discuss here are the Complex NP
Constraint, exemplified with a relative clause in (3b) and with a
nominal complement in (4a), the Subject and wh Island Conditions
(Chomsky, 1973) in (4b, c) respectively, and the Adjunct Island Condi
tion (see Huang, 1982's Condition on Extraction Domain), illustrated
in (4d, e).
(4) (a) *John, Mary made the claim that Sally plans to recommend_
for ajob.
John, Mary claimed that Sally plans to recommend _ for a
job.
As for John, Mary heard the rumor that Sally intends to
marry him.
(b) *John, an article about _just appeared in the newspaper.
As for John, an article about him just appeared in the news
paper.
(c) *Bill, I wonder who likes_.
As for Bill, I wonder who likes him.
(d) *The heat, we left early because of _.
As for the heat, we left early because of it.
(e) *The money, I lied so that I could keep_.
As for the money, I lied so that I could keep it.
Chomsky (1973) proposes to reduce a number of these restrictions to a
single grammatical principle, Subjacency, governing the association of
dislocated phrases and their respective gaps. Subjacency imposes a
locality restriction on this association to the effect that a dislocated
phrase may be related to a gap across no more than a single 'bounding
node', which may be defined for English as any instance of NP or S that
dominates the gap and not the antecedent. Chomsky proposes a further
principle, that the COMP position of a clause may mediate between a
higher antecedent and a lower gap only if the COMP is neither
occupied by a wh phrase or wh complementizer nor is used to mediate
between some other antecedent and gap. Assuming further that poten
tial violations may be saved only through the provision of available
mediating COMP positions, these principles together subsume the
