Language Contact and Change

Spanish in Los Angeles

C A R M E N S IL V A - C O R V A L N

^ L L o z a a i t i a n i d a^Lo Q / t ^
bL COLEGIO DB M EXICO, 4 , C /

CLAREN D O N PRESS O XFO RD

 Oxford University Press, Walton Street, Oxford 0 x 2 6dp
Oxford New York
Athens Auckland Bangkok Bombay
Calcutta Cape Town Dar es Salaam Delhi
Florence Hong Kong Istanbul Karachi
Kuala Lumpur Madras Madrid Melbourne
Mexico City Nairobi Paris Singapore
Taipei Tokyo Toronto
and associated companies in
Berlin Ibadan

Oxford is a trade mark o f Oxford University Press

Published in the United States

by Oxford University Press Inc., New York

Carmen Silva-Corvaldn 1994

First issued in paperback 1996

A ll rights reserved. No part o f this publication may be reproduced,

stored in a retrieval system, or transmitted, in any form or by any means,
without the prior permission in writing o f Oxford University Press.
Within the UK, exceptions are allowed in respect o f any fa ir dealing for the
purpose o f research or private study, or criticism or review, as permitted
under the Copyright, Designs and Patents Act, 1988, or in the case o f
reprographic reproduction in accordance with the terms o f the licences
issued by the Copyright Licensing Agency. Enquiries concerning
reproduction outside these terms and in other countries should be
sent to the Rights Department, Oxford University Press,
at the address above

This book is sold subject to the condition that it shall not, by way
o f trade or otherwise, be lent, re~sold, hired out or otherwise circulated
without the publisher's prior consent in any form o f binding or cover
other than that in which it is published and without a similar condition
including this condition being imposed on the subsequent purchaser

British Library Cataloguing in Publication Data

Data available

Library o f Congress Cataloging in Publication Data

Language contact and change : Spanish in Los Angeles
Carmen Silva-Corvaldn.
(Oxford studies in language contact)
Includes bibliographical references and index.
1. Spanish language Dialects California Los Angeles County.
2. Spanish language Social aspects California Los Angeles County.
3. Languages in contact California Los Angeles County.
4. Bilingualism California Los Angeles County. I. Title.
II. Series.
PC4829.L58S55 1994 467.979493 dc20 93-6268
I S B N 0 -19 -8 2 3 6 4 4 -1

1 3 5 7 9 10 8 6 4 2

Printed in Great Britain on acid-free paper by

Bookeraft (Bath) Ltd., Midsomer Norton
 T o my brother

Hernn Silva-Corvaln
(19 35-19 8 7 )

Nuestras vidas crecieron junto al Maul

y el Maul va a dar a la mar .. .
&
 Acknowledgements

The research on which this book is based was financed primarily by two grants
from the National Science Foundation (BNS-8214733 and BNS-8721453) and by
grants from the University of Southern California Humanities Faculty Research
Support Grant (19823) and Faculty Research Innovation Fund (1985-6). Part of
the research and writing was carried out while I was a guest researcher at the Max-
Planck-Institut fur Psycholinguistik, Nijmegen (1986). The support of these three
institutions is gratefully acknowledged.
The research for this study would not have been possible without the co
operation of the Mexican-Americans who gave generously of their time, their
intuitions, and their attitudes about language and bilingualism, and who more than
once allowed me to invade the privacy of their homes with the ominous tape-
recorder. I wish to express my appreciation for their hospitality and their help.
The following previously published articles have been incorporated into various
sections of this book in an adapted and substantially revised or expanded form:
Bilingualism and Language Change: The Extension of Estar in Los Angeles
Spanish, Language, 62 (1986), 587-608; Tense-Mood-Aspect Across the Spanish-
English Bilingual Continuum, Variation in language: NWA V -X V at Stanford, in
K. Denning, S. Inkelas, F. Mc-Nair-Knox, and J . R. Rickford (eds.), (Stanford,
Calif, Stanford University Department of Linguistics, 1987), 395-410; Oral
Narrative Along the SpanishEnglish Bilingual Continuum, in John Staczek (ed.),
On Spanish, Portuguese and Catalan Linguistics (Washington, DC, Georgetown
University Press, 1988), 172-84; Past and Present Perspectives on Language
Change in US Spanish, InternationalJournalfor the Sociology o f Language, 79 (1989),
53-66; Current Issues in Studies of Language Contact, Hispania, 73 (1990) 162-
76; Cross-Generational Bilingualism: Theoretical Implications of Language At
trition, in T . Huebner and C. A. Ferguson (eds.), Cross-Currents in Second Language
Acquisition and Linguistic Theories (Amsterdam, John Benjamins, 1990), 325-45;
Spanish Language Attrition in a Contact Situation with English, in H. W. Seliger
and R. Vago (eds.), First Language Attrition: Structural and Theoretical Perspectives
(Cambridge, Cambridge University Press, 1991), 15 1- 7 1; On the Permeability of
Grammars: Evidence from Spanish and English Contact, in W. Ashby, M. Mithun,
G. Perissinotto, and E. Raposo (eds.), Linguistic Perspectives on the Romance
Languages (Amsterdam, John Benjamins, 1993), 19 -43.
I wish to thank the editors and publishers for granting permission to use this
material.
I am especially grateful to my colleagues Nancy Dorian, Suzanne Romaine, and
viii Acknowledgements
Roger Wright for their friendship, for reading the manuscript, and for offering
valuable critical observations and meticulous editorial suggestions which I have
done my best to take into account. Nancy and Suzanne deserve special recognition
for nurturing my initial interest in the topic of language contact, and for their
intellectual support during every phase of this project. It will be obvious to the
reader that Nancy Dorians work has had a profound influence on my own work.
I also very gratefully acknowledge Nancys careful scrutiny of each of the seven
chapters of this book, as well as her enlightening comments on practical and
theoretical issues.
M y indebtedness to many other colleagues and friends over the years is great.
William Labov and Benji Wald have inspired and encouraged my efforts to carry
out sociolinguistic work. For comments and discussion about earlier versions
of individual chapters, I give thanks to Roger Andersen, Edouard Beniak, Lucia
Elias-Olivares, Charles Ferguson, Susan Gal, Mary Ellen Garcia, Jane Hill, Thom
Huebner, Flora Klein-Andreu, Yakov Malkiel, Raymond Mougeon, Ricardo
Otheguy, Ellen Prince, and John Staczek. For thought-provoking questions about
some of the premisses and conclusions presented in Chapters i and 7, I convey
my thanks to Sarah Thomason. I also owe an extended thanks to Erica Garcia, who
offered incisive observations and invaluable advice on various papers which were
later incorporated into different sections of the book.
The research and writing of this study have taken several years. During this
time a number of students, too numerous to mention individually, have assisted
me with some of the research projects discussed here, have made helpful com
ments during seminars, and have at all times encouraged me with their enthusiasm
for the issues related to language contact. I give them special thanks.
I wish to recognize my indebtedness to my parents, Fernando and Felisa, and to
my sons Diego, Fernando, and Rodrigo, for giving me the love, encouragement,
and emotional support which I needed to pursue my professional life. M y brother
Fernando and my extended family (including Nicolas, the latest addition) have
also contributed to making my progress smoother. Finally, more than thanks are
due to my husband, Scott Dahlberg, who in the last few years has had to put up
with innumerable questions about English, and with linguistic issues as a frequent
topic of conversation. By dedicating this book to Hernn, I know they feel I am
dedicating it to all of them.

C.S.-C.
Pacific Palisades, California
Ju ly 1992
 Contents

Abbreviations xi
List o f Tables xii

1. Language Contact and Change i

1.1. Theoretical Preliminaries i
1.2. The Socio-Cultural Setting of Spanishin Los Angeles 7
1.3. Research Methodology 15

2. TenseMood-Aspect Across the Spanish Continuum 20

2.x. The Tense-Mood-Aspect Continuum 20
2.2. Compensatory Strategies 39
2.3. Universality of the Processes 48
2.4. Explanations and Conclusions 53

3. The Effect o f TenseMood-Aspect Simplification on Narrative and

Hypothetical Discourse 56
3.1. Introduction 56
3.2. Oral Narrative 56
3.3. Discourse about Hypothetical Situations 76
3.4. Summary ' 91

4. Exploring Internal Motivation for Change 92

4.1. Introduction 92
4.2. The Extension of Estar 92
4.3. Verbal Clitic Pronouns 121
4.4. Conclusions 130

5. Exploring External Motivation for Change 133

5.1. Introduction: On the Permeability of Grammar 133
5.2. Non-expression of a Complementizer or Relative
Pronoun Que That 136
5.3. Omission of Obligatory VerbalClitics 138
5.4. Obligatory SV X Order 141
5.5. Obligatory Expression of Subjects 145
5.6. Conclusion 166
x Contents
6. Conflicting Loyalties 168
6.1. Introduction 168
6.2. Acts of Shifting: Modelling 169
6.3. Modelling Across the Bilingual Continuum 184
6.4. Acts o f Conflicting Loyalties: Positive Attitudes
and Uncommitted Behaviours 186

7. Conclusion 207
7.x. Summary 207
7.2. Discussion and Theoretical Implications * 212

Appendix 1. Tape Contents and Speaker Coding 222

Appendix 2. Verb Test 227

Appendix 3. Factors Included in theCoding o f Subject Expression 230
Appendix 4. Attitude Questionnaires 232

References 240

Index 250
 Abbreviations

Cl verbal clitic pronoun PerCond periphrastic conditional

Cond conditional Perf perfect
exp. expected PerFut periphrastic future
Fut future PI present indicative
FutRef future reference Pluperf pluperfect indicative
HP historical present PluS pluperfect subjunctive
Imp imperfect indicative PPS present perfect subjunctive
Imper imperative Pres present
Ind indicative PresP present participle
Inf infinitive PresPerf present perfect indicative
IS imperfect subjunctive PresRef present reference
lit. literal(ly) Pret preterite indicative
P participle PS present subjunctive
PastP past participle Sub subjunctive
PastRef past reference
 List of Tables

2.1. Verb Forms Examined across the Continuum 21

2.2. Stages of Loss of Tense-Mood-Aspect Morphemes 27
2-3- Stages of Simplification (S) and Loss (L) 30
2.4. Tense Systems across the Spanish Continuum 3i
2-5- Responses to the Fill-in-the-Gap Verb Test 34
2.6. Tense Extensions in Group 1 40
2.7. Tense Extensions in Groups 2 and 3 41
2.8. Tense Substitutions in the Verb Test 44
2.9. Changes Affecting Ser To Be, Estar To Be, Ir To Go 46
2.IO. Summary of Tense Changes 48
2.II. Hierarchy of Markedness 5
3-i- Scale of Assertiveness Conveyed by Verb Morphology in
Reference to Non-past Situations 78
4.1. Parameters Explicative of the Choice between Ser and Estar in
Contrasting Contexts 95
4.2. Stages in the Extension of Estar 101
4-3- Responses to the Ser (S)/Estar (E) Questionnaire 104
4.4. Percentage of Innovative Estar (a) over All Tokens of Estar and
(b) over All Tokens of Innovative Estar plus All Tokens of Ser 105
4-5- Cross-tabulation of Innovative Use by Type of Adjective 107
4.6. Percentage of Innovative Uses of Estar 109
4.7. Cross-tabulation of Innovative Use by the Semantic Transparency
of the Choice between Ser and Estar hi
4.8. Innovative Estar by Spanish Proficiency and by Age of English
Acquisition 114
4.9. Proportion of Innovative Uses of Estar in Los Angeles and Morelia 116
4.10. Cross-tabulation of Innovative Use of Estar by Group and by Age
(Los Angeles) 116
4 .11. Innovative Estar in Three Families 118
4.12. Clitics Omitted in Required Contexts 124
4-i3- Clitics Omitted in Required Contexts, by Clitic Type and by Group 125
4.14. Quantification of Clitic Omission for Seven Speakers in Group 3 126
4-IS- Cross-tabulation of Clitic Position by the Semantics of the Matrix
Verb 128
4.16. Preverbal Clitic Placement in Verbal Periphrases 129
 List o f Tables xiii
4,17. Percentage of Preverbal Clitics in Verbal Periphrases in Los Angeles,
in Four Spanish-speaking Countries (Chile, Spain, Mexico,
Venezuela), and in Written Spanish 130
.1. Speakers Included in the Study of Word Order and Subject
Expression 143
5.2. Cross-tabulation of Subject Placement by Type of Subject and by
Speaker Group 144
5.3. Subject Noun Phrase Placement in Pragmatically Unmarked
Sentences in Rio Platense Spanish 145
5.4. Cross-tabulation of Subject Pronouns by Speaker Group in E L A
and Comparison with Madrid (Spain), Boston (Puerto Rico
Spanish), Santiago (Chile), and Caracas (Venezuela) 153
5.5. Cross-tabulation of Subject Expression by Speaker, by
Coreferentiality, and by Ambiguity of the Verb Form 155
5.6. Cross-tabulation of Subject Expression by Ambiguity of the
Verb Form 156
5.7. Cross-tabulation of Subject Pronoun Expression by Ambiguity
of the Verb Form 156
5.8. Percentage of Subject Expression by Syntactic Function of the
Coreferential Argument in the Preceding Sentence x57
5.9. Cross-tabulation of Subject Expression by Coreferentiality 158
5.10. Cross-tabulation of Subject Expression by Group, by
Coreferentiality, and by Ambiguity of Verb Form 160
5.11. Cross-tabulation of Coreferential Subjects by Type of Subject
and by Speaker Group 161
5.12. Cross-tabulation of Type of Subject by Speaker Group 163
f). 1. Types of Modelling Attested in Groups 2 and 3 187
6.2. Sample of Respondents to the Questionnaires 189
(>.3. Use of Spanish with Various Interlocutors 192
(.4. Use of Spanish in Relation to Topic or Activity 195
ft.5. Use of English in Relation to Topic or Activity 196
ft.6 Contact with Mexico 196
Most Important Dimension for Valuing Spanish Based on
Percentage of Respondents Selecting Reasons in Each 198
f).8. Percentage of Respondents Starring a Reason as Most
Important for Valuing Spanish 200
(1.9. Percentage of Respondents Crossing Out a Reason as Least
Important for Valuing Spanish 201
fi.io. Average Agreement with Positive (+) and Negative (-) Items
about Spanish and English 203
ft.11. Degree of Commitment to Act in Order to Maintain Ancestors
Culture and Language 204
xiv List o f Tables

7.1. Summary of Some Linguistic and Non-linguistic Phenomena

Examined 210
7.2. Types of Linguistic Changes and Associated Triggering Factors 215
 I

Language Contact and Change

1.1. Theoretical Preliminaries

I n t h e past thirty years, sociolinguistic research has been concerned, among other
questions, with the examination of language change (Romaine 1982). Labov (1972a;
1981a; ig8iF), in particular, has successfully challenged the traditional structur
alist position, which argued that change in progress could not be observed, by
developing the necessary methodological techniques to identify and study possible
changes in progress in apparent or real time. During this period, one of the most
debated questions within the field of historical and socio-historical linguistics has
been the interaction between internal linguistic factors and external social forces
(including as such a different language) in what Weinreich et al. (1968: 186-7) call
the actuation or motivation o f linguistic change, i.e. the role that both society and
the structure of language may play in stimulating or constraining linguistic change.
With respect to this dilemma, these authors observe that if the theory that linguis
tic change is change in social behaviour is taken seriously, then it may not be
possible to postulate predictive explanatory hypotheses. Rather, it is conceivable
that all explanations may have to be offered a posteriori.
Most neogrammarians and post-Saussurean structuralists, including generativists,
view change as motivated and governed by internal factors.1 Structuralism con
ceives language as a system whose elements are defined by the place they occupy
in opposition to other elements, or as a system controlled by language specific rules
and universal principles. This concept of language underlies the explanation of
change, on the one hand, in relation to the existence of structural spaces and of
incomplete or unbalanced correlations within the system and, on the other hand,
on the basis of processes of reanalysis essentially motivated by rule opacity. Also
considered to be internal are a number of cognitive factors which constrain pos
sible changes; for instance, it is suggested that changes which may lead to the
neutralization of important oppositions or which may cause comprehension diffi
culties will tend to be resisted (cf. Martinet 1962). In contrast, sociolinguistics
focuses on the social forces which shape language structure and use, as well as on
internally motivated variation. Sociolinguists have shown that language is inher
ently and systematically heterogeneous and variable, and that the seeds of change
lie precisely in the existence of this variation. In regard to change, therefore, most
1 I n the diffusion o f change, dialect geography, areal linguistics, and wave theory all consider the role
if social communication, but a strong structuralist point o f view disallows structural dialectology
(Anttila 1972: chs. 14 -15 ).
2 Language Contact and Change

analysts agree with one of the general principles postulated by Weinreich et al. (1968:
188):

Linguistic and social factors are closely interrelated in the development o f language change.
Explanations which are confined to one or the other aspect, no matter how well constructed,
will fail to account for the rich body o f regularities that can be observed in empirical studies
o f language behavior.

Despite changes of perspective and methodological advances, however, Labov

(1982) notes that we are still far from reaching a solution to the problem of
causality. The question of what forces, whether internal or external, have moti
vated a specific change at a given time and place continues to haunt sociolinguists
as well as structuralist historians. In this respect, Malkiel (1983: 251) justifiably
observes that the answers proposed depend to a large extent on the linguists
theoretical inclination. Accordingly, the linguist will favor either external or internal
factors . . . either an explanation allowing for the intervention of speakers, at varying
levels of consciousness, in the events affecting their speech or the rival explanation
operating with unguided clashes of blind forces.
In the context of this continuous controversy, increasing attention has been
focused on pidgin and creole languages, on the acquisition of first and second
languages, and on the linguistic phenomena which develop in situations of societal
bilingualism and multilingualism. In language-contact situations, developing lan
guages, receding languages, and maintenance of languages lend themselves to the
examination of hypotheses about linguistic change, because they are characterized
by constant and rapid changes which may be observed as they arise and spread in
the linguistic and social systems. Among the important theoretical issues dealt with
by those concerned with these studies have been the possible universality of the
linguistic processes characteristic of these situations, and the role that a model,
primary, or superordinate language may play in the shaping of the developing,
secondary, or subordinate language as against the possibility of autonomous de
velopments constrained by the linguistic system of each of the languages in question
and/or by human cognitive processes.
In regard to the issue of universality, it has been observed that some of the
phenomena characteristic of bilingualism and multilingualism, namely simplification,
overgeneralization, transfer, analysis, and convergence, are indeed attested across
different situations of linguistic stress. This observation has motivated the pro
posal of a number of corresponding theories (theory of transfer, theory of simpli
fication, etc.) valid for all these language situations (e.g. creolization, language
acquisition, language loss). Different scholars conceptualize and define these uni
versal phenomena in somewhat different ways. It becomes necessary, therefore, to
state explicitly how I understand them in the context of my work.
Researchers have not yet come to an agreement about the meaning of simplifi
cation. Indeed, Ferguson (1982: 58) has pointed out that the notions simplicity
and simplification are among the most elusive concepts used in the character
ization o f language. However, he adds that, despite this difficulty, there is general
 Language Contact and Change 3
consensus regarding what in fact constitutes linguistic simplification. Included
as such are, for instance, reduction of the inventory of linguistic forms, semantic
range, or language functions, and the elimination of alternative structures at cer
tain levels (p. 59).
Note that the phenomena referred to by Ferguson, reduction and elimination of
alternatives, are end-states that imply the existence of more complete varieties with
which the simplified ones are being compared. In a situation of languages in
contact, simplified varieties of a given language may develop as a result of incomplete
learning and/or language attrition. In the latter case, reduction and elimination of
alternatives must necessarily involve a classical stage of variation in the use of all
available forms, followed by the gradual disappearance of some forms, functions,
and/or alternatives. Variability and gradual simplification should in principle be
observable in the speech behaviour of individuals as well as across individuals.2
Accordingly, I consider reduction and elimination the result of a complex pro
cess of linguistic simplification which also implies rule generalization, in the sense
that the use of a form is being expanded to a larger number of contexts. Note that
my definition of simplification as a process differs from Fergusons (1982) view of
this concept as referring to an end-state. It also differs in some respects from
definitions of simplification offered by other authors (e.g. Meisel 1977; 1983a;
Mougeon et al. 1985; Mhlhusler 1981; Trudgill 1983: ch. 6).
Viewed as a process, then, simplification involves the higher frequency of use
of a form X in context Y (i.e. generalization) at the expense of a form Z, usually
in competition with and semantically closely related to X , where both X and Z
existed in the language prior to the initiation of simplification. Thus X is an
expanding form, while Z is a shrinking or contracting form. I f simplification reaches
completion, its final outcome is reduction or loss of forms and elimination of
alternatives, i.e. a simplified system with fewer forms and possibly, though not
necessarily, loss of meanings.
This definition of simplification appears to correspond to the notion of
overgeneralization (cf. Preston 1982; Silva-Corvaln 1990a) as the more extensive
use of a form than expected in ordinary practice. The only difference is that
simplification explicitly refers to contraction, that is, the less frequent use of a
competing form. Overgeneralization, on the other hand, may affect contexts where
no corresponding competing form exists, i.e. where we may have X Y extending to
X0 or vice versa (e.g. 0 + V se + V in Spanish examples such as 0 muri en un
accidente (he) 0 died in an accident > Se muri en un accidente (he) se died in an
accident). A related term is regularization, which I use to refer to those cases
where the forms extended or overgeneralized are those with a wider structural
distribution in the language in question.

2 Incomplete acquisition o f the mother tongue ( L i) or o f a second language (L2) in a contact

situation may also involve simplification, i.e. a period o f variation leading to reduction and elimination.
O'. Meisel (1983) for an in-depth discussion o f simplification as a strategy o f L 2 acquirers. I do not
concern m yself with the issue o f incomplete acquisition here.
4 Language Contact and Change
Transfer is undoubtedly a controversial notion as well. It may be simply defined
as the incorporation of language features from one language into another, with
consequent restructuring of the subsystems involved (cf. Weinreich 1974). This
definition accounts for the obvious cases of transfer, usually at the lexical, mor
phological, and phonological levels of analysis. By contrast, at the syntactic level
it may be more difficult to identify direct transfer and to prove permeability of a
grammatical system to influence from a diffe/ent system. For instance, even an
obvious candidate for an explanation based on direct transfer from English into
Spanish such as the non-expression of que that in complement clauses in Los
Angeles (LA) Spanish, as in Creo 0 te hubiera gustado I think (that) you wouldve
liked it, does not constitute an unquestionable case of incorporation of a foreign
feature. (The non-expression o f que is discussed in Chapter 5.)
Despite these difficulties, I consider that transfer may have occurred whenever
one or more of the following phenomena is present in the data;
(1) The replacement of a form in language S with a form from language F, or
the incorporation from language F into language S of a form previously absent
in S. In both cases, the form from language F may be incorporated with or with
out its full associated meaning, and it may undergo various degrees of
morphophonological adaptation to language S. Examples: In L A Spanish bye re
places adis, chao, and other forms of leave-taking; lonche is incorporated to refer
to a noon meal. This type of transfer is usually referred to in the literature as
borrowing; I consider it an instance of direct transfer.
(2) The incorporation of the meaning of a form R from language F, which may
be part of the meaning of a form P in S, into an already existing form, structurally
similar to R, in system S (cf. Weinreichs 1974: 30 extension or reduction of
function). This is also a case of direct transfer. Example: registrarse incorporates
the meaning of to register (in school), thus making obsolete the Spanish words
matricularse/inscribirse to register in a school/for a course.
(3) The higher frequency of use of a form in language S, determined on the
basis of a comparison with more conservative internal community norms, in contexts
where a partially corresponding form in language F is used either categorically or
preferentially. This constitutes an instance of indirect transfer from language F
into language S. Example: the more frequent use of progressives by Puerto Rican
bilinguals as compared to monolinguals (Klein-Andreu 1986; cf. also Morales 1986:
ch. 4; Mougeon et al. 1985; Schachter and Rutherford 1979; Silva-Corvaln
1986).
(4) The loss of a category or a form in language S which does not have a parallel
category or form in the system of F (cf. Weinreichs neglect or elimination of
obligatory categories). An example is the loss of adjective gender marking in some
varieties of Los Angeles Spanish. This is a form of indirect transfer.
Transfer leads to, but is not the single cause of, convergence, defined as the
achievement of greater structural similarity in a given aspect of the grammar of
 Language Contact and Change 5
two or more languages, assumed to be different at the onset of contact (cf. Gumperz
and Wilson 1977). Indeed, convergence may result as well from pre-existing inter
nally motivated changes in one of the languages, most likely accelerated by contact,
rather than as a consequence of direct interlingual influence (Silva-Corvalan 1986).
A change in a language L is considered to have accelerated in the speech of a.
group X when both the number of context types (as opposed to tokens of the same
context), and the frequency of use of the innovation in a token count in the various
contexts are higher as compared to an older age group Y in the same speech
community as X , and when this increase is in turn higher than any possible
increase of the same instance of change identified in the speech of a group P who
speak the same language as X, as compared to an older age group Q Jn the same
speech community as P.
This definition follows from the assumption that change-related variation ob
served in apparent time (Labov 1972a), i.e. across generations, may reflect stages
of diffusion in real time. Crucially, in the case of language contact, I consider the
language proficiency continuum (see section ii below) to stand for apparent time in
the manner in which different age groups are considered to represent apparent
time in sociolinguistic studies of variation. The claim that synchronic variation
across age groups may constitute evidence of change is debatable, since any vari
ation identified as a change on the basis of differential distribution of the variants
across age groups could also be interpreted as corresponding to age grading or
to stable variation in the different age groups. This problem does not affect our
study, however, because acceleration is established with respect to groups differ
entiated by length of contact with English and proficiency in Spanish (the pro
ficiency continuum) rather than by age.
Analysis is the process which underlies either the preferential use or the creation
of analytical or periphrastic constructions as opposed to synthetic ones. This
accounts, for instance, for the much more frequent use of the periphrastic future
and conditional forms (ir a go to + Inf.) rather than the corresponding synthetic
ones (-rd, -ria) in all varieties of Spanish.
At least three of these processes, simplification, overgeneralization, and analysis,
which characterize change in the languages o f monolingual communities as well,
may frequently be accounted for on the basis of cognitive and intralinguistic factors
(see Chapters 2-5), rather than as the result of direct transfer. There is a general
consensus, however, that intensive language contact is a powerful external promoter
of language change. Nevertheless, perhaps owing to the fact that the acknowledge
ment of contact-induced changes poses a threat to some of the methods of historical
linguistics (as discussed by Thomason and Kaufman 1988), it is only rather recently
that linguists have challenged the view that a grammatical system is impermeable
to direct transfer of elements which do not correspond to its internal structure
or tendencies of development (cf. Jakobson 1938). In this regard, Thomason and
Kaufman (1988) argue that there is evidence from a large number of contact
situations in support of their hypothesis that it is the sociolinguistic history of the
6 Language Contact and Change

speakers, and not the structure of their language, that is the primary determinant
of the linguistic outcome o f language contact (p. 35).
While I am basically in agreement with this hypothesis, the results of my
investigation of bilingualism in a situation of family and individual language shift
indicate that, even under conditions of intense contact and strong cultural pressure,
speakers of the receding language simplify or overgeneralize grammatical rules but
do not introduce elements which would cause "radical changes in the structure of
the language. Rather, these changes occur step by step in real time, and across the
proficiency continuum in the receding language. Since ultimately they may lead to
the development of a language which is essentially different from its non-contact
ancestor, I favour a slightly different hypothesis: that the structure of the lan
guages involved, to a large extent constrained by cognitive and interactional
processes, governs the introduction and diffusion of innovative elements in the
linguistic systems; the sociolinguistic history of the speakers is the primary deter
minant of the language direction and the degree of diffusion of the innovations as
well as of the more distant (in terms of time-span) linguistic outcome of language
contact (i.e. after several generations of normal language transmission; cf. Thomason
and Kaufman 1988). This hypothesis accounts for the changes attested in numerous
situations of language maintenance and/or shift involving normal transmission
across generations.
This book follows this tradition of language contact studies, dealing with some
o f the changes undergone by a particular variety of Spanish in a situation of
intensive contact with English in Los Angeles. Approximately 150 hours of re
corded Spanish language data from an intergenerational sample of 50 adult bilinguals
are examined. Whenever relevant, monolingual spoken Spanish is also studied for
comparison purposes.
The general hypothesis investigated is that, in language contact situations,
bilinguals develop strategies aimed at lightening the cognitive load of having to
remember and use two different linguistic systems. In the use of the subordinate
language, beyond phonology,3 the strategies suggested by the data include:
(1) Simplification of grammatical categories and lexical oppositions.
(2) Overgeneralization of forms, frequently following a regularizing pattern.
(3) Development of periphrastic constructions either to achieve paradigmatic
regularity or to replace less semantically transparent bound morphemes.
(4) Direct and indirect transfer of forms from the superordinate language.
(5) Code-switching, which involves the use of two or more languages by one
speaker in the same turn of speech or at turn-taking points.4

3 N o phonological phenomena are examined here.

4 Code-switching has been shown to be a characteristic communicative mode o f skilled bilinguals
(Poplack 1979). However, some forms o f code-switching also appear to function as a strategy to
compensate for diminished proficiency in a language, for memory lapses, etc. (Silva-Corvalan 1983A).
I do not deal with code-switching in this book.
 Language Contact and Change 7
The result is more or less massive changes in the secondary language. The
extent o f these changes correlates with the speakers level of bilingual proficiency
and with extralinguistic factors. The occurrence of the changes is further favoured
and accelerated by (a) absence of normative pressures on the subordinate language;
(b) restriction in the range of communicative uses of the subordinate language; and
(c) speakers positive attitudes towards the superordinate language combined with
either neutral or negative attitudes toward the subordinate one.
This book examines in depth the linguistic strategies listed in (i)-(4). I deter
mine the patterns of diffusion of a number of changes in the linguistic system, and
examine their correlation with the extralinguistic favouring factors postulated. I
give special attention to the issue of permeability at different stages of attrition or
distance from the Spanish dominant norm. By analysing and comparing the speech
behaviour of an intergenerational sample of speakers in apparent time and along
the Spanish proficiency continuum, I attempt to characterize and explain processes
and stages of language attrition. Thus, I expect to provide here one of the baselines
for the universal generalizations that may be possible when enough comparable
studies are available concerned with constraints on deviations from target or source
languages, about the relationship between the sequences of development/learning
and the sequences of loss, and about the strategies employed to compensate for
the lack of certain linguistic devices available to the users of the non-restricted
systems.

1.2. The Socio-Cultural Setting o f Spanish in L os Angeles

This study is based on data obtained from Mexican-American bilinguals living in

a mostly Hispanic area of Los Angeles County (inner rectangle in Fig. i . i ) between
1983 and 1988. According to the U S Census of 1980,5 Los Angeles County was a
multi-ethnic area of 7,477,503 persons, slightly over 30 per cent of the total popu
lation of California (23,667,902). There were 4,544,331 persons of Spanish origin
in this state; 3,597,065 persons five years old and over declared speaking Spanish
at home. Almost 50 per cent (2,066,103) of the Hispanic population resided in the
Los Angeles County, and of these, 80 per cent (1,650,934) were of Mexican origin.
The U S Census of 1990 indicates an overall increase in the population of Califor
nia, to 29,760,021. Of these, 25.4 per cent are of Hispanic origin. In 1990, a total
of 5,478,712 persons five years old and over speak Spanish at home in this state,
i.e. almost a 2 million increase in ten years. With respect to the density of Hispanic
population in the areas studied (inner rectangle in Fig. 1.1), the Los Angeles Times
of 25 July 19836 reports that it ranged from 30 per cent in Monterey Park to over

5 T h e figures from official Censuses are necessarily conservative, in that they do not include many
persons o f Spanish origin who are undocumented immigrants in California.
6 In all cases the statistics provided are the most recent available. Since I conducted most o f the
fieldwork in this community between 1983 and 1985, the figures from 1980 and 1983 are quite relevant
for the present research.
ft
F i g . i . i . Area of Residence of Speakers Studied
 Language Contact and Change 9
70 per cent in East Los Angeles, Commerce, Pico Rivera, and other suburbs
adjacent to East Los Angeles.
Contact between Spanish and English in the bilingual communities of Los
Angeles has been intensive and has involved large numbers of individuals (cur
rently it is estimated to include about 3 million persons in Los Angeles County).
Spanish was the language of prestige, associated with the Spanish conquerors,
until California was annexed to the United States in the late 1840s, i.e. for about
200 years. Since then, Spanish has become subordinate to English, not because of
linguistic reasons but because of the socio-political and historical factors which
have made English officially, socially, practically, and economically superordinate,
a status which this language to a large extent enjoys at the international level as
well. This unequal status accounts for the almost complete diffusion of English in
urban Hispanic communities, as well as (though to a lesser extent) in non-urban
areas.
Los Angeles is not easy to classify neatly into one category of bilingual commun
ity with reference to its genesis, boundedness, and stability (see Lewis 1978 for a
discussion of typologies of bilingualism). Nevertheless, certain features related to
the history and permeable boundaries of this community bear upon the linguistic
picture.
The city of Los Angeles was founded in 1777 by the first Spanish governor of
Alta California. The first public school, however, was opened only in 1855, seven
years after the end o f the Mexican/American War. The rapid change from a
Spanish-Mexican to an Anglo-controlled cultural, economic, and political environ
ment is reflected in the fact, among others, that in this public school only English
was permitted for instruction (Romo 1983: 25). During the second half of the
nineteenth century Los Angeles attracted a variety of immigrant groups, including
Anglos, Blacks, Jews, Chinese, and Germans. The Spanish-Mexican population
stayed at first in the original core area of the city, but increasing industrial con
struction and new waves of immigrants from Mexico created the need for moving
to the east. From a town of 100,000 people in 1900, Los Angeles grew to a metro
polis of over a million by 1930, of whom approximately 100,000 were Mexican
(Romo 1983: 5, 61), and to a megalopolis (Los Angeles County) of over 7 million
in the years following the Second World War. By the end of the 1960s, Los
Angeles had become the major destination for Mexican immigrants and Mexican
in-migrants from other areas of the southwest (Romo 1983: 170).
Although the Hispanic community of Los Angeles is basically a geographically
stable urban community, bound together by historical, social, linguistic, and cul
tural factors which differentiate it from other bilingual and English-only speaking
communities (e.g. of European or Oriental background), both its boundaries and
its internal structure are impermanent and undergoing constant change. Interstate
movements and changes from rural to urban centres by Spanish speakers in the
US quite often lead to a rapid shift to English, easily observable by comparing the
linguistic behaviour of these migrants with that of their offspring in the new
10 Language Contact and Change
environment. A similar situation of language shift is created by movement from
inner city to suburban areas, where ethnic and linguistic identity are more diffuse.
Other immigrant groups (e.g. Chinese, Vietnamese, Korean, Philippine) settle in
formerly exclusively Hispanic areas, for instance, causing some disintegration and
movement to other areas. By contrast, permanent and continuous migration of
groups of individuals from Hispano-America, as well as yearly seasonal migrations,
have been an important positive factor in the maintenance of Spanish. These
migrations reinforce links with non-restricted varieties of the language, and they
facilitate frequent intermarriage between bilinguals with reduced degrees of pro
ficiency in Spanish and nearly monolingual Spanish speakers. This social phenom
enon undoubtedly motivates increased conscious and unconscious efforts to maintain
Spanish as a community language.
Indeed, Spanish in the United States illustrates both maintenance and shift to
English. The maintenance of Spanish at the societal level is unquestionable. De
spite efforts to reassert the status of English as the dominant language through the
passing of state laws which establish English as the official language (e.g. in Cali
fornia in 1986 and Arizona, Colorado, and Florida in 1988), my direct experience
of the situation in the south-west indicates that Spanish is being spoken by ever
larger groups of individuals; new publications in Spanish appear on the market;
television programmes in Spanish increase their audience; important business
companies advertise in Spanish and provide customer services in this language
(e.g. telephone companies, law offices, hospitals, health centres, department stores).
Sixty per cent of the Hispanic population of the United States (close to 9
million according to the 1980 U S Census) resides in the five states referred to as
the Hispanic south-west, namely Arizona, California, Colorado, New Mexico,
and Texas. O f the 6.5 million people who speak Spanish at home in these five
states, over 3 million live in California, nearly 2.5 million in Texas, and under
500,000 in each of the remaining three states. An analysis based on these and other
1980 Census figures leads Bills (1989), however, to conclude that, although Spanish
speakers are the overwhelmingly dominant linguistic minority in each state, there
is no evidence that the Spanish language is being strongly maintained in the
United States Southwest (p. 26). Indeed, at the individual or family level, shift to
English is common in such urban centres as Los Angeles. The children of first-
generation Hispanic immigrants acquire Spanish at home, but most of them gradually
become dominant in English as they go through either a transitional bilingual
education programme or an English immersion programme.
The typical family situation is one in which the older child acquires only
Spanish at home, and maintains a good level of communicative competence in this
language throughout his life, with more or less attrition depending on a number of
extralinguistic factors, while the younger children acquire both Spanish and Eng
lish at home. These younger children are more likely to develop and maintain a
contact variety characterized by greater distance from the norms of first-generation
immigrants. When first-generation immigrants become grandparents and are close
 Language Contact and Change ii

to their grandchildren, these may acquire Spanish at home; but this case is rare
in large urban centres. At best, these third-generation children may be exposed
to Spanish from birth if they live in a Spanish-speaking community and are in
contact with Spanish-speaking family members and friends. They may develop
proficiency in understanding Spanish, but do not often speak this language with
any degree of fluency. However, there are exceptions to this average scenario. One
may occasionally come across a second-generation speaker who either never acquired
Spanish, or acquired it and lost it altogether, or stopped using it for years, and is
in the process of reactivating it, a phenomenon that I refer to as cyclic bilingualism
(cf. Torres 1989). Likewise, a third-generation speaker may exceptionally have
acquired Spanish from birth, and maintained it.
This dynamic situation of cyclic bilingualism, added to the fact that the socio-
historical information regarding such contact is recoverable, make Spanish-
English contact in Los Angeles a singular situation for the study of the processes
underlying both possible stable changes in language maintenance and changes
generally characteristic of language shift or loss. I will examine some of these
changes in. the following chapters.
It has been shown (e.g. Dorian 1981; Elias-Olivares 1979; Silva-Corvaln 1991a)
that in these situations of societal bilingualism an oral proficiency continuum may
develop between the two languages in contact. This continuum resembles in some
respects a creole continuum, inasmuch as one can identify a series of lects which
range from standard or unrestricted Spanish to an emblematic use of Spanish and,
vice versa, from unrestricted to emblematic English. At the individual level, these
lects represent a wide range of dynamic levels of proficiency in the subordinate
language. Speakers can be located at various points along this continuum depending
on their level of dominance in one or other of the languages or in both, but it is in
principle possible for an individual to move or be moving towards (hence dynamic
level) one or the other end of the continuum at any given stage of his life. The
following examples illustrate three different points in the Spanish continuum,
from full proficiency in Spanish to emblematic Spanish:

[1] Researcher = C; Silvia (2) = S7

C: Y tenan una casa grande para la familia o era chiquita? Cmo - ?
S: Bueno, cuando - muri mi pap - vivamos en una casa bastante, muy grande,
donde haba, este, muchas huertas - frutales, alrededor. Estaba en el centro la
casa y estaba rodeada de puros rboles, puros, puros rboles frutales. O sea que

7 See s. 1.3 .1 for an explanation o f speakers groups and numbering. In the examples, a series o f dots
stands for language material omitted because it is not relevant to the discussion; one or more dashes
represent short to longer pauses; a slanting line stands for interruption; a series o f xxx is inserted when
something said is incomprehensible to the transcriber. Square brackets in the English translation stand
for material (one sentence or longer) not included in the translation because it was originally said in
English. T h e information given in parentheses at the end o f the examples corresponds to speakers
name initial and number, sex, age, Group (1, 2, or 3), and tape (E L A no.) where the example occurs.
In [2], italicized words lack gender agreement.
 12 Language Contact and Change
salamos de all y - con una canasta o algo as. La llenbamos de duraznos,
manzanas, toda clase de, de fruta. Y verdura tambin que - se sembraba all.
Rbanos, lechuga, todas clases de verduras . .. (S2,f24,i,ELA37)
C: And did you have a large house for the family, or was it small? How - ?
S: Well, when - my father died - we lived in quite a, in a very big house, where
there were, uh, many orchards around. The house was in the center and it was
surrounded by just trees, just, just fruit trees. So we went out and - with a
basket or something. We filled it up with peaches, apples, all kinds of fruit.
And vegetables also - that they grew there. Radishes, lettuce, all kinds of
vegetables . . .
[2] Researcher = H; Robert (24) = R
H: Y tu tortuga cmo la conseguiste?
R: Un da yo y mi pap estbamos regresando de, de, de un parque con, con un
troque de mi to. Y estbamos cruzando la calle. Y nos paramos porque estaba
un s t o p s i g n . Y mi pap dijo, Ey, Roberto. Quita esa tortuga que est en la
calle. Y no le cre, y o u k n o w . Y mir. Y cre que era un piedra, pero grande.
Y no le hice caso. Entonces me dijo, Aprele. Quita esa tortuga, y o u k n o w .
Y me asom otra vez. Y s era tortuga. Estaba caminando ese piedra grande! frisa]
Pues me sal del carro, del troque. Y fui y consegu el tortuga. Y me lo llev pa
mi casa. (R24,m2o,2,ELA5o)
H: And your turtle, how did you get it?
R: One day my dad and I were coming back from, from, from a park in my uncles
truck. And we were crossing the street. And we stopped because there was a
s t o p s i g n . And my dad said, Hey, Roberto. Remove that turtle from the

street. And I didnt believe him, y o u k n o w . And I looked. And I thought it

was a stone, but big. And I didnt pay attention to it. So he said, Hurry up.
Remove that turtle y o u k n o w . And I looked again. And yes, it was a turtle.
That big stone was walking! [laughter] So I got out of the car, the truck. And I
went and got the turtle. And I took it home.
[3] Researcher = C; Nora (40) = N
C: Y cmo lo haces para poder entender todo en espaol y hablar en ingls?
Cmo lo haces?
N: Ay, ya no s, I dont know. Im surprised to be able to do that. (N40,f2i,
3,ELA48)
C: And how do you manage to understand everything in Spanish, and to speak in .
English? How do you manage?
N: Ay, I no longer know, [ ]

Note that in [2] a speaker from Group 2 uses the expressions you know and stop
sign in English, fails to establish feminine gender agreement in a few noun phrases
and in one clitic pronoun (all italicized in [2]), and uses only estar to be as an
auxiliary in all the progressive constructions where Mexico-born speakers would
most likely use other semi-auxiliaries (e.g. venir to come, ir to go). Despite these
 Language Contact and Change 13
deviations from the norms, his Spanish appears only slightly non-native, and it
is perfectly understandable. By contrast, in [3] the use of Spanish is mainly em
blematic o f the speakers ethnicity, and of her cultural ties with the Hispanic com
munity to which she belongs. This speaker is at the very bottom of the proficiency
continuum.
Studies which claim to be describing a given aspect of bilinguals competence
will necessarily be invalid if they do not take into account the place that bilinguals
occupy in the continuum illustrated in the examples given. Indeed, the absence of
this information constitutes a serious methodological problem which weakens the
validity of numerous group studies of the languages spoken by bilinguals.
The case of language attrition has been referred to as creolization in reverse by
Trudgill (1983) in his study of Arvanitika (a variety of Albanian spoken in Greece),
a term which captures the observation that while certain pidgins and creoles move
toward a higher number of grammaticalized distinctions, inlanguage loss the
reverse is true; for instance, speakers at the lower levelsof proficiency rely on
contextual and/or lexical strategies to communicate certain grammatical meanings
(e.g. temporal or aspectual distinctions).
The type and degree of reduction found in pidgins, however, is different from
that found in language attrition in situations of language shift (Romaine 1989a). As
observed by Dorian (1978, 1981) in East Sutherland Gaelic, in these situations
a language dies with its morphological boots on, i.e. with a certain degree of
complexity which pidgins do not seem to develop. I have observed the same in the
Los Angeles community. Indeed, speakers with the lowest levels of Spanish pro
ficiency are most fluent in English and do not need to use Spanish for any practical
purposes. However, the little Spanish they use, most frequently contained in long
passages of discourse in English, does retain some verbal inflections, gender, number,
case, and prepositions, as illustrated in [4]. This example, produced by Jos R., a
young man of seventeen, further illustrates the characteristic pattern observed for
US-bom youngsters: they would like to talk to their elders only in Spanish but in
fact use mostly English.
[4]Researcher (R), Jos R. (J, 44, Group 3)
R: Pero tu, tu abuelita est en un hospital ahora?
J: No. Mi abuelito st - t en (a) hospital de/
R: Tu abuelito? Y tu abuelita dnde est?
J: st (b) a la casa orita. No, st, mmhm, ella, ella no la, la, o, yo (c) vive con mi,
mmhm. Well, see. I, I want to speak Spanish, but like I dont know everything to,
you know, speak it, but I understand it.
R: . .. Pero con quin hablas en espaol t, a veces, digamos?
J: (d) Hable yo - yo, a ver - yo (d) hable con mi a, abue, abuela - (e) ms de ( f )
mi a, abuelo, porque cuando yo (d) hable con mi abuelo l no (g) entende, l

8 Letters (a), (b), (c), etc. indicate a point o f deviation from standard norms; the standard form is
given in the corresponding notes immediately following the example.
H Language Contact and Change
tiene (h) uno problema - eso - ears. So whenever I have a chance to speak, I speak
to my grandparents. So, I dont speak, I just - listen to what they're saying, and
then I, I - - hear it in my brain and, and - and try to understand instead o f speaking
back at them because I - they understand English as much.
R: But your, your grandmas in a hospital right now?
J: No. My grandpa is - is in (a) hospital.
R: Your grandpa? And your grandma, where is she?
J: Shes (b) at home now. No, she is, mmhm, she, she doesnt the, the, oh, I (c)
live with my, mmhm. [ ].
R: . . . But who do you speak Spanish with, sometimes?
J: I (d) speak - I, lets see - I (d) speak with my g, grandmother - more (e)than
with my g, grandfather, because when I (d) speak with my grandfather he
doesnt ( f ) understand, he has (g) a problem - that - [ ].

Notes: (a) Article el the is missing; (b) Preposition a to used instead o f en in, at; (c) vive
live-3sg used instead o f vivo live-isg; (d) hable speak used instead o f hablo speak-isg;
(e) mas de more of/than used instead o f mas que more than; ( f ) preposition con with is
missing; (g) Verb stem is not diphthongized; (h) numeral uno one is used instead o f
article/quantifier un a .

There are certainly many deviations from the standard norms in [4]. Yet the
passages in Spanish also contain the expected verb agreement with third-person
singular subjects, gender marking (la-fem casa the house), subordinate clauses,
and correctly used prepositions (porque cuandoyo hable con mi abuelo because when
I speak with my grandfather).
The most critical extralinguistic factors which seem to account for the amount
of attrition attested at the lowest levels of the proficiency continuum include the
highly infrequent or steadily decreasing use of Spanish; its restriction to the
domain of the family and close friends; and neutral or negative subjective attitudes
towards the maintenance of Spanish. These factors evolve quite naturally in urban
areas without a strong Hispanic presence where the population of Spanish origin
is more educated than in rural areas and has a higher income level (cf. Bills 1989).
In this respect, the situation in Los Angeles is rather complex. While the levels
of education and income remain on the whole below Anglo and Asian averages
(1980 Census), the strength of the Hispanic presence is indisputable. It is evid
ently almost impossible to forecast how strongly the Spanish language will be
maintained.
The long and sustained contact between English and Spanish in the U SA has
given rise to numerous changes in the system of Spanish (and no doubt also in the
English spoken in Spanish communities), some of which have been investigated
and reported in the literature. At first, research focused mainly on phonological
and lexical aspects. In contrast, since the mid-1970s morphological and syntactic
aspects have been explored, as well as the very interesting phenomenon of code
switching (e.g. various articles in Amastae and Elias-Olivares 1982; Bergen 1990;
 Language Contact and Change IS
Bills 1974; Bowen and Ornstein 1976; Elias-Olivares 1983; Fishman and Keller
1982; Hernandez-Chavez et al. 1975; Pealosa 1980; Snchez 1983; Wherritt and
Garcia 1989). This book aims at contributing to this tradition of studies of Spanish
in. the USA. In the following chapters, I will examine a number of lexical, semantic,
and syntactic changes at different stages of the Spanish proficiency continuum of
bilinguals within the broader frame of general language-contact and language-
change studies.

1.3. Research Methodology

1.3.1. Speakers and Language Data

The studies to be presented here are based on samples of data obtained through
recordings of conversations made mainly between 1983 and 1985. A few additional
recordings were made in 1987-8. Approximately 150 hours of audio-recorded
conversations with 50 Mexican-American bilinguals living in the eastern area of
Los Angeles were collected and transcribed (see Fig. 1.1). O f these, about 20 hours
were recorded by student assistants who are themselves members of the Mexican-
American community; the rest correspond to conversations between the author
and the subjects included in the study. In addition, two further techniques were
employed: (a) fill-in-the-gap questionnaires designed to obtain supplementary
information about the speakers use of various tenses, as well as of ser and estar; and
(b) a set of questionnaires which explore speakers attitudes towards English and
Spanish. These questionnaires were applied between 1987 and 1988.
The speakers for the study were chosen to fill in an intergenerational sample
including men and women o f different ages, categorized in three groups ( 1,2 , and
3) according to the length of time that the speakers families have lived in the
USA. Appendix 1 displays the distribution of the speakers in these three groups,
the sociolinguistic coding for each speaker, and the meaning of the codes used. It
should be obvious (see my discussion of the proficiency continuum above) that the
separation into three discrete groups does not correlate directly with three discrete
groups in terms of Spanish oral language proficiency. Indeed, because of their
social histories, different speakers in Groups 2 and 3 have acquired different levels
of fluency in Spanish.
Group 1 includes speakers born in Mexico, who immigrated to the U SA after
the age of eleven. This age cut-off point was established for two reasons. It is
usually considered that the critical age by which the structures of ones native
language are firmly acquired is around 1 1 - 1 2 (Lenneberg 1967). Secondly, since
the sample was to include adolescents, the cut-off point was made as low as pos
sible to allow for at least five years of residence in the USA for those in Group 1.
This length of residence seemed necessary for the development of some degree of
bilingualism and for adequate exposure to Los Angeles Spanish .
 i6 Language Contact and Change

Group 2 encompasses speakers born in the USA or those who have immigrated
from Mexico before the age of six. Group 3 also comprises speakers born in the
USA; in addition, at: least one parent responds to the definition of those in Group 2.
The places of origin in Mexico of the speakers in Group 1, and of the parents
of those in Group 2, include Mexico City (3 persons) and eight states north of it
(see Fig. 1.2), namely Guanajuato (2 persons), Jalisco (15), Zacatecas (1), Nuevo
Len (1), Durango (1), Chihuahua (3), Sonora (J), and Baja California (6).
Only those speakers in Group 1 have received more than three years of formal
instruction in Spanish, in Mexico, although only one person in this group com
pleted secondary education in Spanish (see Appendix 1). Most speakers in the
younger age group (15-29) have taken 2-4 semesters of Spanish in the USA to
fulfil their high-school foreign-language requirement. It is unlikely, however, that
this period of study had any significant effect on their proficiency in Spanish,
except in the case of Rod (no. 42, Group 3), who did not acquire productive
proficiency in Spanish before adolescence.
The level of education in English varies in every group: six speakers in Group
1 have received no formal education in this language, five have received their high-
school certificate in Los Angeles, two are attending high school, and three are
attending college. Only two of the sixteen speakers in Group 2 did not complete
secondary education. In Group 3, one speaker has achieved a doctoral degree, one
did not complete secondary education, three will soon complete this level of
schooling, seven have completed it, and the remaining six have continued studies
beyond high school.
The choice of speakers was restricted to include first-generation immigrants
from Mexico who had lived in Los Angeles for at least five years in the case of the
younger group, and fifteen years for the older age group. The same restriction
applied to speakers in Groups 2 and 3. In addition, US-born Mexican-Americans
should have acquired Spanish or at least been exposed to this language from birth.
On the whole, speakers in Group 2 were able to converse in Spanish with ease. In
contrast, seven of the eighteen speakers in Group 3 spoke Spanish with difficulty;
indeed, nos. 40, 44, and 46 (see App. 1) spoke with me mainly in English and
sporadically produced a few passages of spontaneous Spanish.
The eastern Los Angeles area was chosen because of its high concentration of
Hispanics, mostly of Mexican ancestry, and because approximately 40 per cent of
the population aged five years and over living in this area claimed Spanish as their
home language in the 1980 U S Census. M y initial contact with these communities
was facilitated by some of my students. I communicated with all the speakers
beforehand; they were aware of my interest in bilinguals and the life of Spanish
in the community. I established a friendly relationship with most of them (for
instance, by maintaining at least some telephone contact between the recording
sessions), especially so with a number of families, which included two and some
times three different generations of immigrants (see family groupings in Appendix
1). The speakers were recorded twice, with an interval of about six months between
F ig . 1.2. Mexico: States and Federal District
i8 Language Contact and Change

the two recordings, for periods ranging from 75 to 100 minutes each time. The
questionnaires were filled in after all the recordings were completed.
The goals of my studies required the use of careful sociolinguistic fieldwork
methodology in order to succeed in obtaining comparable data across speakers,
while at the same time maintaining an atmosphere of social conversation during
the recordings. Crucial to a study of the impact of language contact on both form
and function of such aspects as verb morphology and the serf estar opposition, for
instance, are spoken data ranging along a wide variety of topics and discourse
genres, which ensure the creation of contexts for the use of every form of the verb
system hypothesized to be available in the language of first-generation immigrants.
Despite a few drawbacks, the data obtained from every speaker represent the
variety of topics/discourse genres expected. These include, but are not limited to:
exposition of past events, narratives of personal experience, route directions, de
scriptions of people and places, argumentative discourse, and discourse about
hypothetical situations.
This variety of data was obtained quite naturally and with ease from first and
most second-generation immigrants, but it became a somewhat more difficult
enterprise with some of the speakers in Group 3. For instance, when talking about
counterfactual events, one of these speakers (no. 50) tended to switch to English
for long stretches of conversation, and had to be asked to switch back to Spanish.
Other speakers in the larger sample explicitly stated: No me gusta hablar hypo
thetically I dont like to speak hypothetically. The reasons for this dislike may not
have been exclusively linguistic, however. Indeed, when asked why, one of these
speakers explained to me that he only liked to talk about or say what he could be
certain of.
The analysis of hypothetical reference further requires topics which elicit both
past and non-past time reference, as well as the expression of various degrees of
possibility. These necessary data were obtained by introducing such topics as (a)
for past reference: life in Mexico if X had not come to the US, life in the USA
if X had chosen a different career/job/school; and (b) for non-past reference:
speakers response to the possibility of X dating/marrying someone from a different
race/religion/cultural background; situations which would justify abortion; or the
possibility of having a lot of money.
The recorded conversations were all transcribed by bilingual students at USC,
and edited by me and by some of my graduate students. The analyses of these data,
presented in the subsequent chapters, are based on these transcribed materials. I
have studied the speakers productive proficiency in Spanish. In a bilingual com
munity like the one illustrated by Los Angeles, however, the ability to understand
the minority language by far exceeds the ability to use it productively. My impres
sion is that all the speakers included in this study understood me, although it is
likely that at times I may have unconsciously modified my Spanish to adjust to my
interlocutors.
 Language Contact and Change 19
1.3.2. Analytic Framework
In the following chapters I will approach some specific questions motivated by the
data and by the hypotheses put forth in the preceding sections, through qualitative
and quantitative analyses of the oral corpora collected. In some instances, the
qualitative analysis will be supplemented by descriptive statistics (including tabu
lations, cross-tabulations, and analysis of variance), as well as by information stem
ming from ethnographic observations.
For the quantitative treatment of the data I have used the Statistical Package for
the Social Sciences (SPSS-X) and VARBRUL (explained by Rousseau and Sankoff
1978), computational programs which allow me to examine the statistical signifi
cance of the correlations proposed as existing between a given linguistic variable
and a number of internal and external variables. While quantification is relatively
simple when one is dealing with phonological variables, it is more delicate when
the phenomena involved are syntactic-semantic, as pointed out, among others,
by Garcia (1983), Lavandera (1978), Romaine (1981), and Silva-Corvaln (1989: ch.
4); it requires a much more careful contextual analysis, and the total linguistic
context from which the observation is drawn is more difficult to identify. Quan
tification is crucial, however, since, as I have shown elsewhere (Silva-Corvaln
1986), the influence of one language on another, and the discontinuities along the
bilingual proficiency continuum, may be more evident through differences in the
frequency of use of a certain structure rather than in its categorical occurrence or
non-occurrence.
I will explain the procedures followed in the qualitative linguistic analysis of
each study conducted as it becomes necessary. Suffice it to say here that this
analysis has been far from simple. On the one hand, it involved pioneering work
on the linguistic issues under examination as they occur in spoken Spanish. On the
other hand, the claim that a bilingual speaker no longer has available for productive
use a given lexical item or syntactic construction had to be based on careful com
parison of the linguistic behaviour of speakers across the continuum in clearly
similar discourse contexts. I have been performing this painstakingly slow scrutiny
of the data for seven years. This book is the result.
 2

Tense-Mood-Aspect Across the

Spanish Continuum

2.1. The Tense-Mood-Aspect Continuum

T h i s c h a p t e r focuses on simplification and loss of tense-mood-aspect mor

phology in the Spanish spoken by adult Spanish-English bilinguals representing

different degrees of Spanish language attrition. Specifically, it addresses three
research questions. What type of simplification and loss affects the Spanish verbal
system? What factors, linguistic or extralinguistic, account for the type and stages
of simplification and loss attested in the data? What is the universality of the
systems and processes identified in this study across other language-contact situ
ations, as well as in creolization and language acquisition?

2 .1.1. Tense in Spoken Los Angeles Spanish

Let us briefly examine the verb system of Spanish,1 presented in Table 2.1. Fol
lowing Comrie (1985: esp. ch. 1), I consider tense to be the grammaticalisation of
location in time, and include as such those forms discussed under the labels
of preterite , pluperfect, future', inon-finite forms, etc., regardless of the fact that
these forms are usually used to convey mood and aspect distinctions as well as
tense. Though it is not current practice, I have included the so-called periphrastic
future and the periphrastic conditional (formed with the PI and the Imp of ir to
go + Inf, respectively) as tenses, on the basis (a) that morphological boundness is
not in itself a necessary criterion (Comrie 1985: 11) to decide whether a given
category is grammaticalized and (b) that they are by far the most frequent forms
used to locate situations in the future (the PerFut, as in Va a llamar maana hes
going to call tomorrow), or in the future of the past (the PerCond, as in Dijo que
iba a llamar maana he said he was going to call tomorrow). I examine two uses
or functions of the morphological Fut, namely its use as tense: Estar aqu maana
hell be here tomorrow, and its modal use in reference to a hypothetical situation
in the present: Estar enfermo he must be ill [at the moment of speaking] . Like
wise, two functions of the morphological Cond are differentiated; tense: Dijo que
llamara ms tarde he said hed call later, and modal: Estara enfermo he must have
been ill (possibility in the past), S i tuviera tiempo llamara maana if I had time
Id call tomorrow (unlikely possibility in the future).
1 Statements about uses o f this verbal system do not necessarily apply to varieties o f Iberian Spanish.

L.
 Tense-Mood-Aspect Across the Continuum 21
T able 2.1. Verb Forms Examined Across the Continuum

Simple Compound

Non-finite forms P erf Inf: haber mirado

Inf: mirar look at (Perf PresP: habiendo mirado)
PresP: mirando
PastP: mirado

Finite forms
Indicative mood:
Pres: (l) mira (he) looks at PresPerf: ha mirado
Pret: mir (Past Perf: hubo mirado)
Imp: miraba Pluperf: haba mirado
Fut: mirar (Fut Perf: habr mirado)
PerFut: va a mirar
Cond: mirara Cond Perf: habra mirado
PerCond: iba a mirar
Subjunctive mood:
Pres: mire PresPerf: haya mirado
Imperf: mirara/mirase PluS: hubiera mirado
imperative mood:
Pres: (t) mira (you-sg) look
(usted) mire (you-sg-polite) look

1 also follow Comrie in the classification of tense forms into absolute, absolute-
relative, and relative. Comrie incorporates the notions of relative, absolute, and
absolute-relative tenses in his cross-linguistic study. Absolute tenses locate a situ
ation in time relative to the present moment; absolute-relative tenses locate a
situation in time relative to a reference point which is in turn established (abso
lutely) to be before or after the present moment; with pure relative tense, the
reference point is some point in time given by the context, not necessarily the
present moment. In Spanish, I consider absolute tenses the PI, Pret, Imp, Fut,
and PresPerf (semantically past with present relevance, Comrie 1985: 77-82).
Absolute-relative tenses include the Cond (as future in the past), the PS and IS,
and Ind and Sub compound forms. Relative tenses comprise the non-finite forms.
[1-4] illustrate Sub absolute-relative tenses.
[1] Dice/dijo que vengas. (PS, fut. in the present/past)
He says/said for you to come.2

2 Subject pronouns are not obligatorily expressed in Spanish. When it is important for the discus
sion to signal that there is a non-expressed subject in the Spanish example, I will indicate this by
placing the English subject in parentheses. Otherwise, the English translation will appear with the
corresponding pronoun.
22 Tense-Mood-Aspect Across the Continuum
[2] Dijo que vinieras. (IS, fut. in the past)
He said for you to come.
[3] Si me hubiera dicho, hubiera venido. (PluS, fut. perf. in the past)
I f he had asked, I wouldve come.
[4] Quiere que hayas cenado (PPS, past of fut. in the present) a las 7.
He wants for you to have eaten by 7.
The Perf PresP, the Past Perf, and the Fut Perf Ind are in parentheses in Table
2.1 to indicate that they are never used by any of the speakers in the sample. In
fact, the Past Perf is not used in any variety of spoken Spanish and is quite rare in
written Spanish, and the Perf PresP appears to be restricted to the written mode.
Their absence in spoken Mexican-American Spanish is, therefore, not surprising.
The Fut Perf is still used in spoken and written Spanish to refer to past hypothet
ical situations (as shown in [5];3 following Comrie (1976), I use situation as a tech
nical term to refer inclusively and indistinctly to actions, processes, events, states,
etc.), as well as to future situations viewed as completed after a reference point also
located in the future ([6]).
[5] A: Los Castro se han ido ya.
B: Habrn vendido su casa entonces.
A: The Castros have already left.
B: They must have sold their house then.
[6] Ya se habr ido cuando lleguemos ah.
He will have already left when we get there.
Furthermore, the morphological Fut (-ri) is used to refer to future time only in
a restricted number of registers in Hispano-American Spanish (e.g. threatening/
strong commands: Te quedars aqu hasta que yo diga (You)ll stay here until I decide
to the contrary; religious register). Likewise, the Cond is extremely rare in its
tense function, i.e. when it is used to refer to future in the past, as in reported
speech. The tense function is performed almost exclusively by the periphrastic
constructions. In the Mexican-American data, the Cond and the Fut occur with a
very low frequency in the speech of first generation immigrants,4 especially so in
their tense function. Given acceleration of change in languages in contact, the early
loss of these forms in the other two groups is predictable.
To identify patterns of simplification and loss in the verb system I have com
pared the linguistic behaviour of speakers in Groups 2 and 3 with that of those in
Group 1, i.e. with the speech of members of the same family, or at least of the
same bilingual community (see Appendix 1). In the qualitative analysis I have used
discourse analysis techniques for the identification of form functions and of the
expected changes along the continuum.
3 Examples without a reference to a speaker are made up by me. I am a native speaker o f Chilean
Spanish.
4 I refer to speakers in Group 1 as first-generation immigrants .
 Tense-MoodAspect Across the Continuum 23
This linguistic analysis is often complex. Indeed, it is well known that among
the important functions of verb morphology are those of indicating aspectual and
mood oppositions. Thus it would be possible to argue that, in syntactic contexts
which allow more than one form (as [7-8] illustrate), a speaker who uses Ind in
stead of Sub, for instance, has done so to communicate a higher degree of assert
iveness rather than because of lack of the Sub; or that a speaker who uses Pret
instead of Imp views the situation as a dynamic, unanalysed whole, and not as
temporally suspended in its existence (cf. Silva-Corvaln 1 gg i>).

[7] No te creo que viene (Ind)/venga (Sub) maana.

I dont believe you that hes coming/he may come tomorrow.
[8] Mi pais fue (Pret) ! era (Imp) hermoso antes.
M y country was/used to be beautiful before.
Since these are indeed possible arguments, I have tackled the questions they raise
by considering the possibility of simplification or loss in three different linguistic
contexts: obligatory contexts (i-iii; cf. Brown 1973: 255); discourse-pragmatic
contexts which favour a form (iv); and optional contexts (v), as follows:
(i) First, obligatory syntactic contexts, i.e. those requiring one or another verb
form, were considered, e.g. contexts which impose consecutio temporum constraints
in a number of subordinate clauses. Thus, the PluS provided by the researcher in
the protasis of [9A] limits the choice of verb form in the apodosis to PluS or
Pluperf or to Cond Perf. The choice of other verb forms (as in [9B]) is considered
to indicate simplification or loss of these perfect forms. Likewise, failure to use
Sub in certain subordinate clauses ([10]) which require it in Group 1 is ascribed to
simplification or loss.
[9] A: Y qu me dices de tu educacin si tus padres se hubieran quedado en
Mxico?
B: No estudiaba (Imp) mucho, yo creo.
A: So what can you tell me about your education if your parents had stayed
in Mexico?
B: lit.: I didn't study much, I think.
[10] Lo voy a guardar antes que Uega. (Pres)
Im going to put it away before he comes home.
(ii) Obligatory contexts created either by intrasentential linguistic material (e.g.
an adverbial expression) or by the verb form used in the preceding sentence were
taken into account to determine whether the form under examination was correct.
Accordingly, the use of Imp in [ 11- 12 ] and of Pret in [14] are regarded as signs
of simplification of the Pret/Imp opposition. This is because the adverbial expres
sion el ao 69 the year 69 and the complement un accidente an accident in [ii~
12] force an interpretation of the situation as singular and perfective so that the
Pret is required. Examples [13] and [14a], on the other hand, have created an
24 Tense-Mood-Aspect Across the Continuum
imperfective frame of repeated situations in the past which [13^] and [ 14], being
part of the same temporal-aspectual sequence, should but fail to maintain.
[11] El ao sesenta y nueve tenan (Imp) el World Cup en Toluca, en ftbol.
(D36,m 45,3,ELA 43)5
The year sixty-nine they had the World Cup in Toluca, in soccer.
[12] Iba a ser profesional, pero creo que tena (Imp) un accidente. (R.24,m2o,
2,ELA5o)
He was going to become professional, but I think he had an accident.
[13] a. Despus mi hermano era (Imp) el que iba (Imp) a misa
b. y entreg (Pret) el sobre. (838/19,3,ELA66)
a. Afterwards my brother was the one who went to church
b. and handed in the envelope.
[14] a. Porque este mejicano no saba (Imp) el ingls,
b. no ms habl (Pret) espaol,
c. pero era (Imp) muy bravo y muy macho. (D3,m45,3,ELA43)
a. Because this Mexican didnt know English,
b. he would only speak Spanish,
c. but he was very tough and very macho.

(iii) Discourse constraints on the occurrence of a form in a given context can

provide information needed to evaluate the correctness of the form used. For
instance, narrative abstracts and statements which orient or evaluate the narrative
events as a whole must be coded in the Pret in Spanish (Silva-Corvaln 1983a). This
rule, never broken by first-generation immigrants ([15]), is frequently violated by
speakers in Groups 2 and 3 ([16-17]), who use Imp instead.
[15] Dicen que fue (Pret) muy trgico. Yo no alcanc (Pret) a ver lo dems. Pero
estuvo (Pret) muy feo eso. (L3,f23,i,ELA45)
They say it was really tragic. I didnt have a chance to see the rest. But that
was real ugly.
[16] [produced at the end of a narrative about the speakers prom] Y estaba (Imp)
muy bonito el prom. (R24,m20,2,ELAso)
And the prom was real nice.
[17] Researcher = R; Hector (22) = H
R: Y la otra vez, qu pas (Pret)?
H: Era (Imp) en el seventh grade, con un muchacho del sixth grade. (H22,
m 2i,2,E L A n )
R: And the other time, what happenedV
H: It was in seventh grade, with a guy from sixth grade.

5 See Ch. 1, n. 7 for an explanation o f the information given in parentheses, and o f the symbols and
diacritics used in the examples.
 Tense-Mood-Aspect Across the Continuum 25
(iv) Favourable discourse-pragmatic contexts for the occurrence of a form were
identified in the speech of first-generation immigrants. If the same favouring
contexts failed to elicit the expected forms in the speech of those in Groups 2 and
3, the forms in question were considered to be undergoing simplification or to be
lost, depending on whether they did or did not appear in other contexts. This
methodology is illustrated through a comparison of [18] and [19] from a speaker in
Group 1 and in Group 3, respectively. Both passages are selected from conversa
tions dealing with racial discrimination and prejudices.

[18] C: Si, si viniera un da y te dijera, Mira, me voy a casar con un negro.

S: Con un negro, ah, bueno, yo pienso que yo s aceptara (Cond), porque
claro mi hermana no va - no se va a casar con cualquier negro simplemente.
Pues tomarlo como cualquier persona que sea (PS). . . . Y yo pienso que
tan slo por - por, cuando vengan (PS) los hijos, por ser hijo de mi hermana,
pos yo lo querra (Cond) bastante tambin, aunque sea (PS) negrito, no le
hace. (S2,f2S,i,ELA7)
C: Yes, if one day she told you, Look here, Im going to marry a black .
S: A black, ah, well, I think that I mould accept it, because of course my sister
is not- is not going to marry just any black. So take him like if he were any
other person.. .. And I think that just because - because - when children
came, because hes my sisters son, well I would love him well, even if he
were black, it doesnt matter.
[19] C: . . . Qu crees t que diras en este caso, qu, cul reaccin, cul sera tu
reaccin, verdad, qu - ?
D: Antes que me dice eso, yo creo que va a ir, vamos a tener mucho ms
oportunidades para yo conocer el hombre, de cualquier color, negro,
blanco, caf, y yo puedo decir algo de esa persona, si es bueno o no es
bueno. Y mismo tiempo ella va a hacer que ella quiere. (D36,m4S,3,
ELA90)
C: . . .. What do you think youd say in this case, what, what reaction, what
would your reaction be, right, what - ?
D: (Lit.) Before she says that to me, I think theres going to, w ere going to
have much more opportunity for me to know the man, of any colour,
black, white, brown, and I can say something about that person, if hes
good or if hes not good. And at the same time shes going to do what she
wants'

Note that [18 -19 ] are samples of hypothetical texts, i.e. texts which convey im
aginary, conjectural information, rather than facts stemming from perception and
memory. In texts of this type, one would expect frequent use of weak assertive
verb morphology, e.g. Cond and Sub. As expected, [18] (from Group 1) contains
two Cond and three Sub forms. By contrast, every finite form in [19] (from Group
3) is either PI or PerFut. Given a situation of this sort, one may postulate that the
26 TenseMood-Aspect Across the Continuum
speaker who produces [ig] has failed to produce a Cond and/or a Sub form.
Furthermore, if no favourable context (in 23 hours of transcribed conversation)
prompts the use of these forms, it seems safe to conclude that they are nonexistent
in this speakers verbal system. This statement applies to Spanish language production
only. Indeed, I have enough evidence to assume that most of the bilingual speakers
at the lower levels of the Spanish proficiency continuum understand the meaning
of these tense forms as used by their interlocutor?.
(v) Finally, optional contexts are defined as those where any of the variants
involved may occur, with or without associated meaning differences (cf. Silva-
Corvaln 1986: 605). For instance, the use of Ind or Sub in the example below is
considered to be optional, with a slight meaning difference associated with the
choice of either one.

[20] Quiz vengo (Pres)/venga (PS) maana.

Maybe 17/ come/ would come tomorrow.

The qualitative analytical techniques employed in the analysis of the verb system
used by the speakers allowed us to confirm both the existence of the proficiency
continuum referred to in Chapter 1 (clearly represented in even a small sample of
speakers, such as the one included here), as well as the presence of cross-overs and
overlaps among the speakers in the three groups.
Indeed, three Group 1 speakers (M6, F5, M8, all male in the younger age
group) appear to lack the PPS, a loss observable in most speakers in Group 2. This
makes their system almost identical with that of three speakers in Group 2 (E30,
M26, B27), US-born individuals in the older age group. Likewise, the system of
three speakers in the special Group 3 (R35, A37, B33, US-born parents) is more
fully developed than that of the three speakers at the lower end of the continuum
in Group 2 (V21, E18, M25, younger age group, Mexico-born parents). These
facts corroborate at least three of my impressionistic observations of the language
situation in this community: (a) that first-generation male immigrants who have
come to Los Angeles during their early teens are more prone than their female
counterparts to start losing the language of their ancestors; (b) that Group 2
Mexican-Americans born in Los Angeles before the rapid population growth ini
tiated in the 1950s have developed a higher level of proficiency in Spanish than
those born in the 1960s or early 1970s; (c) that Group 3 speakers whose parents
were born in rural areas of New Mexico, Texas, or Arizona (special Group 3) are
closer in proficiency and fluency to speakers in Group 2 than to those in the
regular Group 3.
There are no clear-cut qualitative differences between first-generation immi
grants and the rest of the speakers, as displayed in Table 2.2. However, while most
of those in the former group use the verb system in the manner predicted for a
basically monolingual variety of spoken Spanish, the majority of speakers in Groups
2 and 3 evidence a range of systems with more or less simplification and attrition.
 Tense-Mood-Aspect Across the Continuum 27
T able 2.2. Stages o f Loss o f Tense-M ood-A spect Morphemes

Group 1

Ei S2 E l2 P iS A9 L3 J16 R 11 Ji4 M6 F5 M8

Inf + 4- + 4* 4 4- + 4* + + 4- +
+ 4- 4~ + 4- + 4 4- 4- + 4-
PresP
PastP 4- + + + 4- + 4- + 4 4- 4- +
~h + 4 4- + 4- + + + 4- +
PI
PerFut + 4- + 4- 4. + 4- 4- + 4- 4- 4-
Pret + + + 4- 4- 4- 4- + + + 4- +
Imp + + + 4 4- + 4- 4- 4- 4- + 4-
Imper 4- + + 4- + 4- 4- + + n 4- +
PS 4- + + 4- 4 4- 4- 4- + + + +
PresPerf + 4- + 4- 4- 4- 4- 4- + 4- + +
4* 4- + 4- + + + + + 4- 4- +
IS
G tn d -Fu tR ef + + + 4- + 4- 4- + 4- + + +
Cond-PastRef + + 4- + 4- 4~ + - + 4* _

Pluperf + + + 4 4- 4. 4- 4- + 4- 4- +
PluS 4- X + 4- 4- + + + 4- + +
Perf In f + 4- 4. 4- + + 4- 4 + + - _

Fut-PresRef + 4- + + + 4- 4- 4- 4- _ - ~
F u t-F u tR ef 4- 4- 4- + 0 0 - 0 _ 0 0 _

PP S + + 4- + + + + X 4- - n n
Cond P erf 4- 4- _

Group 2

E30 M 26 B 27 L28 L 19 A29 A20 R 17 H22 V 21 E 18 M 25

In f + + + + + + + + 4- 4- 4- +
PresP + + + + + + + 4- + + + 4-
PastP + + + + + + + 4- 4- 4- + 4-
Pres + + + + + + 4- 4- 4- 4. 4.

PerFut + + + + + + + + 4. + + 4.

Pret + + + * * * * * * * *

Imp + + + + + + +
@ @ @ @ @
Imper + + + + + + + n 4- 0 4. 4-
PS + + + + + + 4* 0 X 0 0 +
PresPerf + + + + + + 4- 4- 4- 4- 4-
IS + + + + X X X 0 0 0 O O
C ond-FutRef + + + + + 0 0 0 - - 0
Cond-PastRef + + + _ + + + - _ - -
Pluperf + + -f X - X + + _ _
PluS + + + + X + X _ - _ O
Perf In f + + + - + + + - _ _
28 TenseMood-Aspect Across the Continuum
T able 2 .2 (cont.)

Group 2

E 30 M 26 B27 L28 L 19 A29 A20 R 17 H 22 V2 1 E 18 M 25

Fut-PresRef + + _ _ _ _ _
]
F u t-Fu tR ef ] - - - - 0 ] - - -
PP S - - _ + + - n - n n n
Cond P e rf

G roup 3

R 35 3 7 B 33 A 34 R50 R49 M 47 H48 S38 D 39 J 43 M 41 D36 A46 R42 N 40

In f + + + + + + + 4. + + + + + + + +
PresP + + + + + + + + + + + + + +
PastP + + + + + + _ + + _ + + - _
PI + + + + + + + + + X + + 4. + + +
PerFut + 4- + + + + + + + + + 4. 4. + _ +
Pret X + * + * * * * * * * * X
*

Imp + + X + X @ 0 @ @ @ @ @ @ @ @ 4*
Imper + + + n + + + + 0 + - 0 - n
]
PS + + + + + 0 0 - 0 0 X X 0 ] 0 -
PresPerf + 0 + + + + + - 0 + X + - n _ n
IS + 0 + + X + 0 - _ X _ - _ -
C ond-FutRef 0 0 - _ 0 0 + _ - 0 0 _ _ - _ _
Cond-PastRef + +
Pluperf - 0 + - - - - + - 0 X - - _ 0 n
PluS 0 + n n - n n n
P erf In f + 4. n n n n n n
Fu t-PresR ef
F u t-F u tR ef + + ] ]
PPS 4. - - _ - n n n - - n n n n n n
Cond P erf n n n n n n

* Italics indicate speakers in special Group 3.

These systems may be ordered according to how different they are from that of
those in Group 1. Table 2.2 displays the variation characteristic of the continuum.
The speakers are ordered from left to right according to the decreasing number of
different tense forms attested in their data. On the vertical axis of the table, the
ordering of the tense forms follows as closely as possible their sequence of loss and
simplification in the sample of speakers.
 Tense-Mood-Aspect Across the Continuum 29

The symbols in Table 2.2 mean the following: + stands for a tense form which
is used according to the norms of general spoken Spanish, regardless of whether it
has acquired expanded uses as other forms are simplified and lost; - indicates that
the form is not part of the verb system underlying the speakers spontaneous use
of spoken Spanish; * indicates that the Pret/Imp opposition has become neutral
ized in a closed list of verbs which appear with imperfect morphology in both
perfective and imperfective contexts; @ signals that a speaker has used a number
of preterites instead of imperfects; x reflects that the form has failed to occur in a
number of obligatory contexts; 0 signals that the form has failed to occur in a high
number of obligatory contexts and appears to be non-productive, i.e. it is used
with just one or two verbs; ] indicates that the form occurs in one or a maximum
of three different frozen expressions (e.g. Que ser, ser! What will be, will be!,
Te dir que .. . Ill tell you that. . .); n indicates that no pragmatic context for the
use of a form has been identified in the data. The speakers in the table are
identified with their name initial and number (see Appendix 1).
I should also point out that no occurrences of the Cond in its tense function
(future in the past) are attested in the data. Cond FutRef and PastRef stand for
modal uses of the Cond in future and past contexts. Fut PresRef stands for the
modal use of the Fut to refer to hypothetical situations in the present; Fut FutRef
corresponds to the tense use of the Fut. Note that the tense versus modal or
aspectual function of tenses has been investigated exclusively for Fut and Cond. In
all other cases, the term tense may subsume tense, mood, and aspect meanings.
The two extreme ends of the continuum, from most developed to most simpli
fied, are represented by E i and S2 (Group 1), and by N40 (Group 3). The most
developed system used by Group 1 does not differ from the norms o f general
spoken Spanish.
With only three exceptions (L19, A29, R35), the absence of PPS clearly differ
entiates Groups 2 and 3 from most first-generation immigrants. The Fut, both as
tense and as modal with present reference, is absent in the data from most of the
speakers in these two groups as well. Further qualitative differences between
Group 1, on the one hand, and Groups 2 and 3 on the other, are established by
processes of simplification affecting all tenses from the Pret down on Table 2.2.6
Of the compound tenses, only the PresPerf is used with some consistency down
towards the lower stages of the continuum. Indeed, the Perf Inf and the PluS are
used only by six of the twelve speakers in Group 2, and by two of the sixteen in
Group 3. The Pluperf also undergoes early processes of simplification; note that
only two speakers in Group 3 use it with some degree of consistency.
With only minor exceptions, the progression of simplification and loss falls into
seven stages, implicationally ordered as shown in Table 2.3, such that if an indi
vidual uses the forms in (i), it may be assumed that the forms listed in (ii)-(vii) will
form part of his verb system and will not be affected by processes of simplification.

6 T h e Pret and the Imp undergo some interesting changes which I discuss later.
30 TenseMoodAspect Across the Continuum
T able 2 .3 . Stages o f Simplification (S) and Loss (L)

(i) L : Future Perfect Group i

Conditional (tense function)
(ii) L : Conditional Perfect Groups i, 2
Present Perfect Subjunctive
S: Future *
Perfect Infinitive
(iii) L : Future Groups 2, 3
Perfect Infinitive
S: Pluperfect Subjunctive
Pluperfect Indicative
Imperfect Subjunctive
Preterite (with closed list o f stative verbs)
Imperfect Indicative
(iv) L : Pluperfect Subjunctive Groups 2, 3
Pluperfect Indicative
S: Present Subjunctive
(V) L: Imperfect Subjunctive Group 3
Conditional (modal function)
S: Present Perfect Indicative
(Vi) L : Present Perfect Indicative Group 3
S: Imperative
(vii) L : Present Subjunctive Group 3

Conversely, loss of PS (stage vii) implies as well the absence of the forms listed in
(i)-(vi), and the simplification of the Pret Imp opposition, as indicated in (iii).
I have used a sort of apparent-time technique in establishing these seven
stages, assuming that the individual system most divergent from the norms of first-
generation immigrants reflects the most advanced stage of simplification and loss,
while intermediate degrees of divergence would reflect various intermediate stages
of simplification and loss. As expected, given the existence of a proficiency con
tinuum and the diverse social histories of the individuals in the sample, speakers
in Groups 1 - 3 do not fall neatly into each stage. The relationship between the
three groups and the different stages is as follows: two of the speakers in Group 1
are at stage (i) (E i, S2), and the three at the bottom of this Group (M6, F5, M8)
are at stage (ii); four in Group 2 (E30, M26, B27, L28) are at stage (ii) as well; the
other speakers in this group all share the loss of perfective marking on certain
stative verbs, but only five (including L28) appear to be at stage (iii), just as two
in Group 3 are (R35, A37, and possibly a third one: B33); the speaker with the
lowest proficiency in Group 2 (M25) is at stage (iv), which would otherwise be
characteristic of speakers in Group 3 only. Indeed, of the remaining speakers in
Group 3, six are at stage (iv) (A34, R50, R49, M47, D39, J43), three at stage (v)
 Tense-Mood-Aspect Across the Continuum 31
T able 2.4. Tense Systems across the Spanish Continuum

System U S born bilinguals

I II III IV V

Relative tenses
In f + + + + +
PresP + + + +
PastP + + + + -
Absolute tenses
P I (present) + + -f- + +
Pret (past) + + + + +
Imp (past) + + + + +
PerFut (future) + + + + +
PresPerf (past with present relevance) + + + + -
Fut (future) + - - -
Absolute-relative tenses
PerCond (future in the past) + + + + +
PS (future in the past/present) + + + + -
IS (future in the past) + + + - -
PluS (future perfect in the past) + + + - -
Pluperf (past o f past) + + + - -
PPS (past o f future in the present) + - - - -
Cond Perf (future perfect in the past) + - - - -
Cond (future in the past) - - - - -
Fut Perf (past o f future in future)

(H48, S38, M 41), three at stage (vi) (D36, A46, R42), and one (N40) is at the
lowest stage (vii).
What is remarkable about the manner in which the verb system appears to
change is its regularity. Patterns of simplification and loss are never random, but
at all stages conform to a predictable trend to develop a least grammaticalized
system within the constraints of universal grammar possibilities and preferences
(see Comrie 1985 for a discussion of these possibilities). Disregarding a few frozen
expressions, a small residue of forms undergoing advanced simplification, and the
neutralization of Pret and Imp with certain verbs (discussed later), it is possible to
distinguish five different systems which represent a steady progression toward a
less grammaticalized one. The least grammaticalized system in my data, system V
in Table 2.4, consists of two relative tenses, Inf and PresP, and three morpholo
gical oppositions in the Ind: PI, Pret, and Imp. In addition to PerFut and PerCond,
a number of periphrastic verbal constructions (discussed later in this chapter) as
well as lexical expressions complement this system for the verbalization of tense,
32 Tense-Mood-Aspect Across the Continuum
mood, and aspect meanings. The grammaticalized systems identified are shown in
Table 2.4. Given in parentheses is the time meaning of the form, established with
respect to the moment of speaking or present point of reference.
Systems I and II correspond to stages (i) and (ii), respectively, in Table 2.3,
while systems III-V are purified translations of stages (iii)-(vii), that is, they do
not take into account stages of simplification. Table 2.4 incorporates PerFut and
PerCond, given that these are by far the most frequently used forms in spoken
Latin American Spanish, including Group 1 Mexican-American Spanish, for ex
pressing future and future in the past time relations (see section 2.1.x).
Observe that system I has seven contrasting items in the past and only one in the
future (Fut/PerFut),7 in accordance with the general tendency of languages to
have a better developed past than future system (Comrie 1985: 85). System III,
the most developed one attested among US-born bilinguals, has lost one of the
past forms, the Cond Perf, and the Fut; system III retains five absolute-relative
forms, two for future in the past (Per Cond and IS); one for future in the past or
present (PS), one for future perfect in the past (PluS), and one for past of past
(Pluperf). O f these, system IV retains the PS and the PerCond, and system V
preserves only the PerCond for future in the past. The last items to disappear are
past with present relevance (PresPerf) and future in the past/present (PS).
As displayed in Table 2.4, the grammaticalized system developed by English
dominant bilinguals conforms to universal grammar constraints and preferences. It
is not the case, for instance, that a past of past is lost in a system with past of
future; the system also maintains the cross-linguistically more frequent future in
the past (PerCond) rather than future perfect in the past (PluS). Moreover, sys
tems III and IV have a grammaticalized future distinction (PerFut) not always
present in non-contact languages, e.g. in German and Finnish (Comrie 1985: 45),
and system III incorporates absolute-relative past distinctions not made in, for
instance, Russian, which has only absolute tenses (Present, Past, and Future). It
should in principle be possible, therefore, to find speakers at lower points along
the bilingual-proficiency continuum who would use even further simplified verb
systems. In addition, given widespread acceptance of the fact that simplified
languages (e.g. child language, pidgins, early stages of second-language acquisition)
rely heavily on context, it seems safe to assume that the most simplified tense
system could consist of basically non-finite forms with perhaps a small number of
auxiliary verbs, as is the case in many pidgins, but this stage is not attested in the
data analysed here.
The question might arise whether the loss of compound tenses could simply
derive from lack of the forms used in their composition. This is not the case,
however, since speakers who appear to have systems IV and V use the forms haber,
haba (certainly with a different meaning, there to be, there was, respectively)

7 These numbers do not include the Imp because this represents an exclusively aspectual opposition
with the Pret, nor the P S because this form may refer to future in the past or in the present.
 TenseMood-Aspect Across the Continuum 33
and PastPs (system IV only) in other types of construction. Note also that the
composition haber + PastP is retained in the Pres Perf in system IV. Indeed, the
different treatment of the PresPerf lends some support to Comries (1985: 78-82)
argument that this tense is in fact radically different from the absolute-relative
tenses, both conceptually and typologically. Some languages, for instance, have a
distinct PresPerf but no other perfect forms, while others have the reverse. The
former group is illustrated by Swahili and our bilingual system IV; Maltese ex
emplifies the latter.
One can further ask why languages in situations of linguistic stress develop these
grammaticalized systems to the exclusion of other logical possibilities. The answer,
it seems to me, must be sought in relation to all the four factors proposed as
playing a role in the case of languages in contact: namely cognitive, social,
intralinguistic, and interlinguistic. I return to the question of possible explanations
later in this chapter.

2 .1 .2 . Non-spontaneous Use o f Tense

A fill-in-the-gap verb test (see Appendix 2) was administered orally to 37 of the 40

speakers included in Table 2.2 (M6, A46, and N40 did not complete this task). At
the end of the second or third recording session with the speakers, I asked them to
complete the examples in the verb test with any verb in any form that they
considered appropriate. The speakers had the list of examples in front of them
while I read them, and were allowed as much time as necessary to complete the
task. They or I read some of the examples twice or more at their request.
The purpose of this questionnaire was to find out whether the absence of a form
in the spontaneous recordings represented the total loss of that particular form or
whether this form was still part of the speakers verb system, available for use
either in a more formal oral style or in the written mode. The test attempted to
replicate a situation in which the speakers would have enough time to think and
retrieve, presumably from their long-term memory, less frequently used forms.
Furthermore, it provided syntactic contexts not attested in the data from some
Group 2 and Group 3 speakers, which required a number of compound forms not
used in the conversations.
O f the 37 speakers who answered the verb test, Table 2.5 comprises only those
who provided a form which they had either not used at all or failed to use in some
required contexts in the spontaneous recordings. Thus, a + in Table 2.5 corre
sponds to a form not used at all in the required contexts or not used at all in the
spontaneous mode (cf. Table 2.2), while - x and o in Table 2.5 correspond to the
same lack of or reduced use of a verb form as indicated by these symbols in Table
2.2. 1 assume that forms with pluses in Table 2.2 are in the speakers verb system;
therefore, no further information about these forms is included in Table 2.5 (i.e.
the cells are left blank). Pluses in Table 2.5 are important; they allow us to
 T a b l e 2 .5 . Responses to the Fill-in-the-Gap Verb Test

Group i

E 12 P iS A9 L3 J16 R ii M8 A
J *4 FS a S C

PS
13 12 / 14 10 / 10 6/6
IS
18 12 / 14 9 /10 4/6
Cond-FutRef 29 11/ 14 10 / 10 6/6
Cond-PastRef - 2 / 14
IS 2 / 10 2/6
Pluperf 27 10/1:4 10 / 10 6/6
PluS 28 11/ 14 3 / 10 0/6
P e rf In f - + 20 11/ 14 * 9 /10 5/6
Fut-PresR ef - - 32 3/ i4 0 /10 3/6
Fut-F utR ef ++ ++ +- ++ ++ ++ 2, 17 19/28 11/ 2 0 6/6
PPS X - - 30 1/ 14 2 / 10 2/6
Cond Perf
. _ - - 5 3/ i 4 6 /10 6/6
Fut P e rfb _ - 2 / 14
9 8 /10 5/6

T able 2.5 (cont.)

Group 2

E30 M 26 B 27 L28 L 19 A29 A20 R 17 H 22 V21 E 18 M2S A

+ + _ + 3 /12
PS
IS + + + _ + + + 6 /12

Cond-FutRef + _ ' + - - - 2 /12

_ _ - - - 0 / 12
Cond-PastRef _
- + - 2 /12
Pluperf
+ + - + - 4 / 12
PluS +
P erf In f + + - + 3/12
_ - - - - 1/12
Fut-PresR ef _ + -
__ ++ +- -+ 13 / 12
Fu t-F u tR ef ++ -+ ++ +~ ++
_ _ - - - 0 / 12
PPS
0 / 12
Cond Perf
_ - - - - 0 / 12
Fut Perf
 not chosen by speakers in Groups 1 and 2, who tended to provide a Pret (a rather borderline alternative) even though the PresPerf is a productive
c Speakers A37 and J43 completed [9] in the verb test with a PresPerf form as an alternative to Fut Perf. This is an acceptable option, surprisingly
\ \ N S \ \ N S \ S S \ N
t ^ N < N r o O ' ?t ' ^ O ^ N O O O

1 1 1 + 1 1 T 1
?

vO
r*) 1 1 1 + 1 1 1 1 1 I* 1
Q

1 1 1 1 1 1 1 1 1

+ + 1 1 1 1 + 1 + 1 1 1

O'
ro + 1 1 1 1 1 1
Q

+ i + i 1 4 1 1 1 1 1

+ 1 1 1 1 1 1 1 1 1 1
(cont.)

1 1 1 + 1 1 + 1 1 1
2.5
able

1
bo .2
1 1 1 1 1 1
T

1 1 1
2
fi o
U
1 i 1 1 1 1 1 I I !
J3
Mn<3y
s i
*T3 U
1 + 1 + ! I I p p

m 1 + 1 1 I I I

UJ o
<
+ 4* I I i M S3
CL
S3
O *o
form in their Spanish.

v- + + J I T3
a pS Ph
73
c
9 Vi Uh
<3S
s cS 4J tM
(Jh Oh csJ
- CL C/3
I I T3 . Cu
g 3 S3 o . 3 J ctf
fC|l,HUU fcfr(PM fcfcD ^Ui4

I
 Tense-MoodAspect Across the Continuum 37
appreciate at a glance how many more verb forms are available to the speakers in
the non-spontaneous mode.
The number in column Q_ corresponds to the number of the example to be
completed in the verb test (see Appendix 2). Since the number of the example is
the same for all Groups, column Q_is included only in the results for Group 1.
Note that two examples in the verb test (2, 17 see Appendix 2) created a context
for the use of a Future form (PerFut, or Fut). I have included the responses to
both examples in Table 2.5, because it is interesting to observe that the morpho
logical Fut (Fut-FutRef), rare in spoken Latin American Spanish, is the form
preferred by Groups 1 and 2 in a school-type test. This may indicate that the
morphological Fut may also be preferred in a more formal oral register, or in
writing.
The numbers in column A refer to the number of speakers in each Group who
provided the relevant form in the example indicated in column Q , out of a total
of fourteen speakers who completed the verb test in Group 1, twelve in Group 2,
and fourteen in Group 3.
The same verb test was also administered to ten Spanish speakers from various
countries (including Argentina, Colombia, Costa Rica, Spain, and Venezuela) liv
ing in Los Angeles at the time (column S), as well as to six Chilean speakers in
Chile (column C). These sixteen people are all college-educated. Columns S and
C are included in the table for Group 1.
One may wonder if there is any correlation between the number of pluses in the
verb test that any individual has added to the number of forms used spontaneously
and the level of schooling in Spanish or in general. The answer appears to be
negative. Compare, for example, L 19 and E18 with V 21 and M25 in Group 2. The
first two speakers have received no formal instruction in Spanish and their level of
general education is in one case lower (L19) and in the other case the same or
higher than that of the latter two speakers. Yet L 19 and E18 add at least 50 per
cent more verb forms in the test situation (4/7 and 6/12, respectively), while V21
only provides one of the twelve forms that she had not produced spontaneously,
and M25 provides just two out of eleven forms not previously used.
Likewise in Group 3, the number of pluses added does not correlate with level
of formal schooling in Spanish or in general. Other factors may be playing a role
in the different behaviour observed in the two modes examined spontaneous oral
mode and test mode one of which must certainly be level of proficiency in
Spanish. One case where schooling may have had an effect concerns two speakers
in Group 3, D36 and R42. These speakers had not used any Cond forms in the
conversations, but completed an apodosis (Q_29), a modal non-past context, with
this form. These speakers have received some instruction in Spanish as a second
language (see Appendix 1), and question 29 is a typical school example for the
elicitation of Cond.
With regard to the speakers overall behaviour, the only notable difference in
Group 1 corresponds to the Fut. It seems to me, therefore, that the absence or
38 Tense-Mood-Aspect Across the Continuum

scarce occurrence of the morphological Fut in the conversations with speakers in

Group i responds to its being stylistically marked rather than to loss. This
observation may also apply to some speakers in Group 2 (eight out of twelve pro
vided a Fut form in the completion task). As for the other tenses, absent in both
modes, and especially in those examples which elicited the form expected from
most of the other non-Mexican-American speakers tested (e.g. the Cond Perf and
the Fut Perf), I feel more confident about asserting that they are either in an
advanced stage of simplification or absent from the verb system of Group 1
speakers.
In addition to the Fut, Group 2 provides a number of Sub forms in the test.
Indeed, the questionnaire further helped me confirm a diatopic difference between
Mexican and Chilean Spanish in regard to the Cond Perf and the PluS. While
Chilean Spanish (and some of the other varieties tested see column S in Table
2.5) retains the Cond Perf in its modal function in the apodosis of conditional
clauses (Q. 5 in the verb test), Mexican Spanish uses PluS instead. This modal
function is extended to contexts where other varieties would use a modal in the
Cond form plus a Perf Inf. Thus, Q. 28 in the verb test is completed with the PluS
by eleven out of fourteen Group 1 speakers, by three of the speakers in column S,
and by none of the Chileans. These different responses are illustrated in [21 a], [21 b],
[22a], and [22^].
[21] Si hubieras ido con nosotros al estadio, (a) habras visto (Cond Perf, Chileans)
/ (b) hubieras visto (PluS, Group 1) jugar a Fernando Valenzuela.
If you had gone to the stadium with us, you w ouldve seen Fernando Valenzuela
play.
[22] Antes de haber ido t a ese lugar horrible, (a) me podras haber preguntado
(Cond + Perf Inf, Chileans) / (b) hubieras preguntado (PluS, Group 1) cmo
era.
Before going to that horrible place, you couldve asked me what it was like.

The PluS is in general much more frequently used by Group 1 speakers than,
for instance, Chilean speakers. Frequency of exposure to this form, then, may
explain the maintenance of this tense in at least half of the sample in Group 2, as
well as its use in the test by five persons in Group 3. Most of the pluses for this
Group in Table 2.5 correspond to forms which are undergoing simplification in
the spontaneous mode, i.e. they are used only once in a while in contexts that
require them. Therefore, the systems displayed in Table 2.4, which disregard
forms undergoing advanced simplification, seem to be valid for both modes. The
only exception is the Fut, as noted above, which would form part of system III if
the answers given in the non-spontaneous mode were incorporated in Table 2.4.8
There is a further reason to shy away from incorporating the results of the verb
test: the kinds of knowledge tapped by a test of this sort may be quite different

8 Note that no examples in the verb test required the use o f Cond in its tense function.
 TemeMood-Aspect Across the Continuum 39
from the one that underlies the spontaneous use of a language. For instance, in a
test situation I could provide the second person plural form of verbs corresponding
to informal vosotros you, but it is not quite clear to me what the status of this form
is in the verb system of my variety of Spanish, which does not differentiate be
tween formal and informal second person plural morphologically. I certainly know
the forms, as I know many forms from other languages as well, but I would not
consider them part of the system which underlies my productive use of Spanish.
At this point, it becomes necessary to consider the strategies that Mexican-
American speakers are employing to compensate for the forms either lost or being
simplified. I examine this matter in the following section.

2.2. Com pensatory Strategies

As stated in Chapter i, the general hypothesis investigated in this book is that in

language-contact situations bilinguals develop strategies aimed at lightening the
cognitive load of having to remember and use two different linguistic systems.
This task of remembering and using two linguistic systems is not the same for all
bilinguals. Although frequency of use of the receding language may facilitate the
task, it is most likely heavier for those who have acquired two more complete
systems. Neither English-dominant bilinguals at the lower stages of the Spanish
proficiency continuum nor Spanish-dominant bilinguals with low English pro
ficiency bear the same linguistic load as more competent bilinguals. However, the
strategies they resort to when using Spanish include some of those utilized by
bilinguals placed at higher levels along both continua (English and Spanish). In
this section I discuss two of these strategies: semantic extensions of the verb tenses
still available to bilinguals in the recessive language, and the development of
periphrastic verbal constructions. As expected, both strategies occur with varying
degrees of intensity across the Spanish continuum.

2.2.1. Patterns o f Tense Extensions

Patterns of expansion appear to be as regular as patterns of loss. One may safely

say that the general strategy is to use either a retained form with a meaning close
to that of the one being simplified or lost, or the infinitive, a pure relative form.
In the data from Group i, in addition to the standard patterns of form shifts (cf.
Rojo 1974: 111-2 8 ) or expansions and the absence of some forms in the spontane
ous mode, both characteristic of many Spanish varieties (listed in (a)-(d) in Table
2.6), we attest simplification of PluS, illustrated in [23], as well as loss of Perf Inf,
and PPS ([24-5]). In Tables 2.6, 2.7, and 2.8 an asterisk indicates that the tense
used is not an alternative option in any standard or non-standard monolingual
Spanish variety that I know of; a question mark signals that it may be marginally
acceptable in some non-standard varieties.
40 TenseMoodAspect Across the Continuum
T able 2.6. Tense Extensions in Group 1

Tense expected Tense used

(a) Fut PerFut, PI

(b) Fut Perf PerFut + P erf In f
(c) Cond Imp, PerCond
(d) Cond P erf Pluperf, PluS
(e) PluS Pluperf, PPresPerf, PIS
(/) Perf In f In f
(?) pps ?IS, PPresPerf, PPret

[23-5] illustrate the extensions in (e), (/), and (g) respectively.

[23] Pero si, si me he casado (PresPerf) all sabr Dios cmo anduviera, (expected:
hubiera casado, PluS) (E12/44,1 ,ELAg2)
But if, if I had married (lit. have married) over there God only knows what
my life would have been like.
[24] El podia traerlo (Inf), pero no quiso, (exp.: haberlo trado, Perf Inf) (M8,
m i7,i,ELA59)
He could have brought (lit.: bring) him, but he didnt want to.
[25] No creo que agarrara (IS) mucho, (exp.: haya agarrado, PPS) (M8,mi7,
i,ELA59)
I dont think that he could have stolen a lot.9
An in-depth scrutiny and explanation of the semantic features shared by the
tenses used and the substituted or expected tenses is beyond the scope of this
book. I must note, however, that tenses are extended to substitute for others that
I deem to be close in meaning because they have in common their time reference,
their perfectivity, both time reference and perfectivity, or mood. Thus, in Group
1 the Perf Inf is substituted for by Inf and not by PI or Imp, for instance. Likewise,
the PluS is replaced by Pluperf, PresPerf, or IS, and not by an entirely unmotivated
Imp or PerFut. Continuing the cycle, once these expanding forms become lost,
they are in turn replaced by a semantically close existing form.
Any given tense tends to be quite regularly replaced by the same other tense or
tenses in the usage of speakers at the same or a similar stage in the continuum.
However, this regular pattern of substitution is occasionally disrupted in the sam
ples from speakers at the lower stages of the continuum in Groups 2 and 3. The
occurrence of an unanticipated form is illustrated in [26], where a PS substitutes
for a PI.

9 I do not give a literal (lit.) translation into English o f Spanish Subjunctive tenses given that this
language lacks most o f the forms corresponding to this mood.
 Tense-MoodAspect Across the Continuum 41
T able 2.7. Tense Extensions in Groups 2 and 3

Tense expected Tense used

(a) Cond-FutRef IS , *PerCond

(b) PS *P I
(c) IS *Pret, *Imp
(d) Pluperf *Pret
(?) PluS Imp, *Pret, *P I, PerCond
( / ) Cond Perf *PerCond
(g) PresPerf ?PI, ?Pret
(h) Imp *Pret

[26] Ella fue la que me habl, que Andale, Son, por qu no te metas (PS) en este
contest ? (exp.: metes, PI)10 (838/19,3,E L A 31)
She was the one who told me, Come on, Son, why dont you enter this
contest ?
In addition to those of Group 1, Groups 2 and 3 display further tense exten
sions, summarized in Table 2.7. The asterisks and question marks refer to contexts
where the expected form is required and it is not simply a pragmatic alternative as
it may be the case with some verb tenses. For instance, if a PI is chosen over a PS
in contexts where it could be interpreted to communicate a higher degree of
assertiveness than the subjunctive (jQuiz vengo (PI)/venga (PS) maana Perhaps
I will/may come tomorrow), it is obviously considered an expected use of PI.
Given the early loss of Cond Perf, the lack of Pluperf in some individuals in
Groups 2-3, and the later loss of PluS as well, the PerCond expands to substitute
for both Conditionals and for PluS ([27-28B]), while the Pret takes over the
functions of the Pluperf ([29]).
[27] No s yo cul tipo de vida iramos a tener (PerCond).11 Me imagino yo que
hubiera sido - - bueno, no me puedo yo imaginar, (exp.: habramos/hubiramos
tenido, Cond Perf/PluS) (M 26/51 ,2,ELA8s)
I dont know what type of life we wouldve had (lit.: would go to have). I imagine
it wouldve been - - well, I cant imagine.
[28] A: Cmo ves tu vida en Mxico si en vez de vivir aqu te hubiera tocado
vivir all?

10 It might also be possible to analyse [26] as a case o f mistaken type o f conjugation, i.e. that the
speaker has changed the verb meter put in, enter from the -er to the more frequent -ar conjugation,
in which case metas in [26] would be PI. I have no other evidence that the speaker may have changed
the conjugation type, however.
11 T his form o f the PerCond, with the verb ir go in the Cond (iramos) rather than in the Imp (bamos),
as in [28], is extremely rare.
42 TenseMood
Aspect Across the Continuum
B: Bueno, pues, iba a tener que (PerCond) acostumbrarme a las costumbres
de all, (exp.: habra/hubiera tenido que, CondPerf/PluS) (A29,m6o,
2,ELA73)
A: How do you see your life in Mexico if instead of living here you had had
to have lived there?
B: Well, I wouldve had to have (lit.: was going to have to) gotten used to the
way things are over there. "
[29] Y estbamos esperando a mi am - porque ellafue (Pret) a llevar mi hermano
a la dentista, (exp.: haba ido, Pluperf )12 (V 2i,fi8,2,ELAi7)
And we were waiting for my mom - because she had gone (lit.: went) to take
my brother to the dentist.

Towards the lower stages of the continuum, Imp and PI may also substitute for
PluS or Cond Perf ([30B]). In addition, PI and both Pret and Imp may replace the
Imperfect Subjunctive in certain contexts ([31-2]).
[30] A: Eso era lo que yo te preguntaba antes. Si t te podras imaginar cmo,
verdad?. Si, si en vez de casarte aqu te hubieras casado all [en Mxico].
B: . . . no s si poda (Imp), uhm, vivir all, (exp.: habra/hubiera podido,
CondPerf/PluS) (M47,f33,3,ELA8o)
A: Thats what I was asking you before. If you could imagine how, right? If,
if instead of having gotten married here youd gotten married over there
[in Mexico].
B: . . . I dont know if I couldve (lit.: could) lived over there.
[31] Era antes que compraron (Pret) el trailer, (exp.: compraran, IS) (S38/19,
3,ELA 3i)
It was before they bought the trailer.
[32] Se comunic con el p o l i c e d e p a r t m e n t a ver si tenan uno que estaba (Imp)
interesado en ser t e a c h e r , so me llamaron a m. (exp.: estuviera, IS) (Rgo,
m46,3,ELA 36)
He called the police department to see if they had anyone who might be (lit.:
was) interested in being a teacher, so they called me.

At the lower stages, PI expands to PS contexts ([33-4]), and it competes with

the Pret in contexts where Group 1 speakers use a PresPerf form. Interestingly,
the PresPerf (past with present relevance) tends to be substituted for by PI with
stative verbs ([35]), and by Pret with verbs of activity ([36]). The speakers differ
ing way of treating stative and non-stative verbs is reflected as well in the changes
which affect the Pret and Imp, an intriguing issue which I pursue later.

[33] 1 HOpE que no me toca (PI) la misma problema, (exp.: toque, PS) (D39,f28,
3,ELA42)
I hope I dont run into the same problem.

12 In English, these speakers use the Pluperf as expected (see s. 2.4).

 Tense-Mood-Aspect Across the Continuum 43
[34] Yo estoy encargado en, en el taller. Noms cuidando que salen (PI) los trabajos
en tiempo, (exp.: salgan, PS) (H48,m39,3,ELA23)
Im in charge of, of the shop. Just supervising to make sure that the jobs are
done (lit.: come out) on time.
[35] Esta fue la primera casa que compramos. Estamos (PI) como F i f t e e n y e a r s aqu,
(exp.: hemos estado, PresPerf) (R5o,m46,3,ELA3)
This was the first house we bought. W eve been (lit.: are) for about fifteen
years here.
[36] Si un castizo se casa otra vez con un espaol, ya noms lo llaman espaol, ya
se limpi (Pret) su sangre, (exp.: ha limpiado, PresPerf)13 (036,11145,3,
ELA43)
I f a castizo in turn marries a Spaniard, then they call him Spanish, because
his blood has gotten clean (lit.: got clean).
Finally, the six speakers at the lowest points of the continuum (Group 3)
sometimes use the Inf in place of some of the forms lacking in their system,14 as in
[37], or even more frequently simply switch to English, as in [38], which responds
to my question about imagining what the speakers life would have been like in
Mexico.
f37l Y no quera que otras - personas - hacer (Inf) cosas por ella, (exp.: hicieran,
IS) (A46/31,3,ELA2o)
And she didnt want other people to do (required lit.: that other people were
to do) things for her.
[38] Si crec (Pret, exp.: PluS) en Mxico, no creo que - uhm - - 1 w o u l d h a v e
GONE TO THE SAME SCHOOLS AS I DID HERE. (M41,m 18,3,ELA74)
If I had grown up (lit.: grew up) in Mexico, I dont think that - uhm - - i
would have gone to the same schools as I did here.

It must be noted also that no required syntactic contexts are attested in these
speakers samples for the use of a number of tenses (see Table 2.2). This is a clear
indication of the less fluent and simpler Spanish which they use.
The more tightly controlled situation provided by the verb test allowed me to
confirm most of the semantic extensions of the verb tenses still available to bilinguals
in the recessive language. This is seen by comparing Tables 2.6 and 2.7 with Table
2.8, which displays the most frequently used forms as alternatives to those expected
by the researcher in the verb test. A zero in Table 2.8 indicates that most of the
speakers in this group were unable to complete the example.

13 G iven the generic meaning o f this utterance, this seems to me to be a clear context for the use o f
a past with present relevance (as Spanish grammars note). On the other hand, the use o f Pret, especially
i f modified by the adverb y a already, is quite acceptable in specific situations (Y a se fu e H es gone ,
lit: already went).
14 Dislocated uses o f the In f in standard varieties o f Spanish include its use in apodoses (alternating
with Cond), and in direct commands (alternating with Pres Imperative).
44 TenseMoodAspect Across the Continuum
T able 2 .8. Tense Substitutions in the Verb Test

a Group 1 Group 2 Group 3

PS 13 *PI *P I *P I
PresPerf 9 *Pret *Pret *P ret/*In f
IS 18 *P S *Lm p/*PS *Im p /*P S
Cond-FutRef 29 Imp Imp Im p /*P S
Cond-PastRef IS Imp Imp Im p /PI/*P ret
P erf In f 20 In f In f In f/*Pret
P PS 30 IS / * P S Cond/*PS * P I/* Im p /* In f
PluS 28 IS IS *0 /IS
Pluperf 27 *Pret *Pret *Pret
Fut-PresRef 32 PI PI *0/ PI
Fut-F utR ef 2, 17 PerFut PerFut P erFu t/PI
Cond Perf 5 PluS P lu S /*IS P lu S/*Im p
Fut Perf 9 *Pret *Pret *P ret/*In f

Let us turn now to a somewhat more detailed look at the Pret and Imp forms.
Superficial scanning of the data appears to indicate confusion in their use, as
illustrated in [39-40].

[39] Imp for Pret:

Yo fui el nico hombre que tenan (Imp). (A29,m6o,2,ELA2)
I was the only son they had.
[40] Pret for Imp:
En la casa - mi mam era la nica que habl (Pret) espaol, y las dems hablaron
(Pret) en ingls. (036,11145,3,ELA43)
At home - my mom was the only one who spoke Spanish, and the others spoke
only English. (required lit.: used to speak)

A closer examination of the simplification of Pret and Imp morphology along

the proficiency continuum, however, reveals an interesting pattern of change. The
shrinking of the Pret (at stage (iii) in Table 2.3), which clearly differentiates Group
1 from every one of the US-born speakers in the younger age groups, and from
most in the older age groups, affects only a small number of verbs. These verbs,
which are stative and occur more frequently in imperfective contexts in the data
from Group 1, have neutralized the morphological perfective-imperfective dis
tinction in favour of the Imp. The list includes estaba was (Inf estar), era was (Inf
ser), tena had (Inf tener), and saba knew (Inf saber). Expectedly, speakers at higher
points in the continuum in Groups 2 and 3 retain the Pret forms of these verbs and
use them occasionally in perfective contexts. In contrast, four individuals in Group
 Tense-Mood-Aspect Across the Continuum 45
2 (R17, V 21, E18, M25), and the six speakers at the lowest end of the continuum
in Group 3 never use these four verbs in their Pret form. On the other hand, five
out of twelve speakers in Group 2, and nine out of sixteen in Group 3 wrongly use
the Pret in imperfective contexts (@ in Table 2).
It might be argued that neutralization has been caused by the morphophonological
irregularity of the Pret forms of the four stative verbs. This argument does not
hold, however, because the same speakers use the irregular form fue with the mean
ing of went, which is homophonous with the Pret of ser. In addition, these
speakers use other irregular Prets, puse put, hice did , vine came, dije said, di
gave, etc. What these retained irregular Prets have in common is that they are
non-stative verbs.
Thus, if restricted to a small number of verbs, prototypically stative and of
frequent occurrence in the data, Pret simplification is allowed to occur. Its ex
pansion, on the other hand, predictable on the basis of interactional (social) and
interlingual considerations, affects a large number of verbs, including statives and
non-statives (e.g. cambiar change, dormir sleep, enojarse get angry, hablar speak,
ir go, poner put, preocuparse worry, salir go out, usar use).
The different patterns of expansion of Pret and Imp (Pret expands to imperfect
contexts with stative and non-stative verbs, while Imp expands to perfective contexts
with statives only) may be interpreted to indicate that the location of situations in
the past, i.e. tense, is more crucial than signalling (at least by means of inflections)
certain aspectual distinctions. This seems natural, given that the lexical meaning
of the verb does not incorporate the notion of time, while all verbs have a certain
inherent lexical aspect.
Further interesting changes are attested in the data which appear to be related
to those affecting^, the homophonous Pret of ir to go, and ser to be. The majority
of the bilinguals in Groups 2 and 3 use the Pret fue exclusively with the dynamic
meaning of went. In addition, they use the Imp of serf estar, era! estaba was, in both
perfective and imperfective contexts, while the Imp form of ir, iba would go/was
going/went, has lost its lexical meaning and is used only as an auxiliary in the ir
a + Inf be going to + In f construction.
It is not possible for me to ascertain on the basis of the data studied whether all
these changes occur independently of one another or whether one change has
triggered some kind of chain reaction in the system. Whichever the case may be,
the trend is, as in monolingual language changes, toward a one-to-one relationship
between certain forms and their semantic content (fue past of to go, era = past
of to be, irrespective of aspect in both cases). The semantic content which
remains is the one which is most frequently associated with the form in commun
ication (interactional motivation). Further, when full lexical items (e.g. iba, Imp
of to go) reduce semantically into grammatical forms, the lexical and paradig
matic gap left is filled in by a periphrastic construction. I examine these developments
further in the next section.
46 TenseMoodAspect Across the Continuum
T able 2.9. Changes Affecting Ser T o B e, Estar T o Be, I r T o G o

Ser Estar Ir

Group i
Imp era was estaba was iba was going
Prt fue was estuvo <was fue went
Group 2, 3 a U u
Imp estaba + -ndo
era estaba
Prt fue

Gr. i, 2, 3 Auxiliary (past o f future) => Iba a + In f

2.2.2. The Development o f Verbal Periphrases

With respect to the development of periphrastic and auxiliary constructions, the

data give evidence of changes affecting ser be, estar be, and ir go. As stated above,
most of the bilinguals examined in Groups 2 and 3 use the Preterite fue went/was
exclusively with the dynamic meaning o f went. In addition, the Imps of ser/estar
are used in both perfective (i.e. instead of fu e ! estuvo Vas) and imperfective con
texts (see [12, 13, 41] below), while the Imp form of ir, iba would go/was going/
went, appears to have lost its lexical meaning of goal-oriented movement and is
used only as an auxiliary in the ir a + In fbe going to + In f construction (see [28]).

[41] [Beginning of a narrative; perfective context]

Ahhmm, pues una vez - a ver, era con un muchacho que pues a m no me
agradaba nada. (H22,m21 ,2,ELA 11)
Uhmm, well one time - lets see, it was with a guy that, well, I didnt like at
all.
In turn, the lexical meaning of iba is expressed by fue went, or seemingly more
frequently by means of the Imperfect Past Progressive (estaba + -ndo was + -ing,
where the PresP form corresponds to a verb of movement), as shown in Table 2.9,
and illustrated in [42]. In the sample from Group 1, the message communicated by
estaba caminando was walking in [42] would be expressed simply by iba went-
imperfective, or by the progressive form constructed with iba, iba caminando was
walking (lit.: went walking). Indeed, estar be replaces a number of verbs of
movement (e.g. andar go around, venir come) frequently used as semi-auxiliaries
in progressive constructions in Spanish. [43] illustrates the replacement of venir
come.
[42] y despus cuando estamos, yo estaba caminando a la casa con mi hermana - y
su nombre era Baca [a classmate], entonces - y le grit, Baca, tu madre es
una Baca ! (H 22,m 2i,2,ELAn)
 Tense-MoodAspect Across the Continuum 47
and then when we were, I was walking home with my sister - and his name
was Baca [a classmate whose name is homophonous with vaca cow], then -
and I shouted, Baca, your mother is a Baca !
[43] Y como cuarto, cuarta milla alguien estaba coming down at me. Nadie lo, lo
par. Y me, me machuc. (M 4i,mi8,3,ELA34)
And in about a quarter, a quarter of a mile someone was coming down at me.
No one - stopped him. And he - hit me.

M y general hypothesis is supported by this preliminary analysis; iba a + Inf

replaces the bound conditional morpheme -ra and achieves paradigmatic regu
larity, as follows: va a + Inf is going to + Inf = future, iba a + Inf was going to
+ In f = past of future.
Another auxiliary-type construction attested in the data involves the use of a
form of hacer to do/make as a carrier of tense-mood-aspect, followed by a
nominal which contributes the lexical meaning, as in [44] and [45].

[44] Lo hicieron rape a l. (L28,f37,2,ELA23)

Lit.: to him did rape to him
They raped him.
[45] Ella estaba haciendo disciplina (a todos en casa). (1150,11146,3,ELA36)
Lit.: she was doing discipline
She was disciplining (everyone at home).

These trends point in the direction of the following paradigm for past tenses, a
stage which has not been categorically reached in the data examined, however:

Imperfect Preterite
Non-statives Periphrastic Past Progressive with estar hacer + nominal
to be
Statives bound morpheme (-ba) p

The result is, then, that the perfective/imperfective opposition disappears with
stative verbs, which mark past and irrealis with Imp morphology only. Non-
statives, on the other hand, could develop an analytical paradigm which would
retain the perfective/imperfective aspectual opposition.

2.2.3. Summary o f Developments

Both spontaneous and non-spontaneous use of tense-mood-aspect morphology by

the bilingual individuals studied in Los Angeles reveal an evident trend towards
the development of a grammaticalized verb system which is simplified to comprise
four synthetic and three periphrastic forms, corresponding to system V in Table
2.4. In addition, it is intriguing to observe that speakers show a clear inclination to
split the verb lexicon in accordance with stative versus non-stative lexical aspect,
48 TenseMood-Aspect Across the Continuum
T able 2.10. Summary o f Tense Changes

Complete system Simplified morphological system with

innovative periphrastic constructions

In f In f
PresP PresP
PastP (retained as adjectives)
PI PI
Pret Pret/hacer + In f (non-stative verbs)
Imp Pret (non-stative verbs)
Im p/tar + PresP (stative verbs)
Fut (tense) PerFut/PI
Fut (modal) PI (modal verbs)
Cond (tense) PerCond/Imp
Cond (modal) IS ; Im p/Per Cond
PresPerf PI (stative verbs)
Pret (non-stative verbs)
Pluperf Pret/Im p
Fu t Perf PerFut; Pret
Cond Perf PerCond/PluS; Imp
PS PI
IS P S; Pret/Im p
PPS IS/Pret; Pret/Im p
PluS Pluperf/IS; Imp

3 Pret and Imp would coexist with their corresponding periphrastic constructions. Further
more, these forms would contrast in modality: Pret would be realis, and Imp irrealis.

both at the level of bound morphology (Imp for statives, Pret for non-statives) and
in the differing development of auxiliary constructions with hacer to make/do,
and with estar.
I f the variety of Spanish spoken by bilinguals in the mid-points o f the con
tinuum were to become stabilized given such necessary sociolinguistic conditions
as the continuous need to use this variety, normal transmission across generations,
and a large enough community of speakers the data examined provide evidence,
it seems to me, to substantiate the proposal that the outcome could be the result
of the changes summarized in Table 2.10.

2.3. U niversality o f the Processes

Regarding the question of the universality of the systems and processes identified
in this study, we note certain parallels with simplification and loss in other lan
guages in contact, such as those described by Dorian (1981) for East Sutherland
 TenseMoodAspect Across the Continuum 49

Gaelic, by Gal (1979; 1984) for Austrian Hungarian, by Mougeon and Beniak
(1991) and Mougeon et al. (1985) for Canadian French, and by Trudgill (1976-7;
1983) and Tsitsipis (1981) for Arvanitika in Greece. For instance, the early loss of
one of two same-meaning structures (or stylistic shrinkage) has been docu
mented in every one of these bilingual communities as well (cf. loss of Cond Perf,
in competition with IS; loss of Fut and extension of PI and PerFut).

2.j. 1. Emergence and Loss o f Tense

Specifically in regard to tense in Gaelic, Dorian (1981: 141) notes early simplifica
tion of Fut and Cond in the larger part of the less proficient speakers of Gaelic
and, even further, absence of these tenses in the lower levels of the proficiency
continuum. Indeed, she further notes that only five out of ten SSs [semi-speakers]
provided any recognizable conditionals at all, and only three consistently gave
recognizable conditionals (p. 140). Thus, the system of three morphophonologically
marked tenses15 past, future, and conditional is reduced to one, past. Similar
observations are made by Trudgill regarding Arvanitika. The parallels with Mexican-
American Spanish are striking.
With respect to creolization, Muysken (1981) claims that the emergence of
tense, mood, and aspect categories in Creole languages is each governed by a
theory of markedness. For the category tense, which we are interested in comparing
with that in the bilingual continuum, a markedness index based on (a) association
versus dissociation of three points (moment of speech, moment of the event,
theoretical reference point) and (b) the anteriority or posteriority of the event with
respect to the moment of speech justifies the hierarchy from least to most marked
in Table 2 .11 (cf. Muysken 1981: 191).
This markedness hierarchy, proposed to account for order of appearance of
tense morphemes in Creole languages, predicts that the first feature to emerge as
a tense category in early Creole systems would be anterior, which Muysken claims
to be indeed the case. The hierarchy seems to be roughly valid also as a predictor
of order of disappearance in the subordinate contact language that I have examined,
such that the most marked tense disappears first (note that Fut Perf is not even
attested in spontaneous production in Group 1), followed by past perfect (Pluperf)
and simple future (Fut), and lastly by present perfect (PresPerf) and simple past
(Pret). But clearly, it does not by itself make any predictions in the case of forms
with the same index of markedness. For instance, it does not predict whether
simple past or present perfect will emerge or disappear first; my research indicates
that PresPerf is lost earlier when the ancestor of the receding language has a strong
Pret. But I would expect the Pret to be lost first if the receding language were in
an advanced process of losing this form in the spoken mode (as could happen if

15 In Gaelic, only the verb for to be has a present tense. T h e present o f all other verbs, including
the habitual meaning, is formed with the present o f to be plus the corresponding gerund.
50 Tense-MoodAspect Across the Continuum
T able 2 . 1 1 . Hierarchy of Markedness

Tense Markedness
index

Simple present 0
Simple past * 1
Present perfect 1
Simple future 2
Past perfect 2
Future perfect 3

contemporary French or Northern Italian, for instance, became restricted languages

in a contact situation). That is, sociolinguistic evidence appears to be a better
predictor of patterns of loss. Furthermore, this type of theory of markedness
describes, but does not explain, the facts observed. Within a view of language as a
system of human communication, to be explanatory, a markedness hierarchy needs
to be justified with reference to factors which lie outside the linguistic system:
cognitive and interactional factors.
One must be cautious in proposing parallels across languages and social situa
tions, given that many scholars have agreed that simplification and other processes
leading to change appear to be constrained to a large extent by both intralinguistic
and social factors. In addition, as Dorian (1982: 36) warns us, Few if any languages
which are now dying have been used for acquisition or aphasia studies in their
earlier, more intact stages .. . Consequently, direct comparison of dissolution or
reduction with acquisition is difficult or impossible.
Nevertheless, it is arresting to note that some aspects of language loss appear to
a certain extent to be the mirror image of development in creolization, and in first-
and second-language acquisition. That is to say, in acquiring the verb system of
Spanish, and indeed of various other languages (Brown 1974; Klein 1986; Slobin
1986), learners go through stages of development which are in some respects the
reverse of the stages of loss identified in my data: the earlier tense forms to be
acquired are present and past (both perfective and imperfective), while future,
conditional, and compound tenses are acquired much in the same order in which
they are lost across the proficiency continuum. However, it is possible that this
correspondence may in fact reflect the freezing, at different levels of development
of grammatical proficiency, of the bilinguals less dominant or subordinate lan
guage. The possibility of loss at the individual level, i.e. starting from an individu
als more developed stage of Spanish in real time, is not documented in this study;
but I could not dismiss a priori the possibility that those speakers placed at lower
levels of Spanish proficiency at the time of this investigation had acquired the
 TenseMood-Aspect Across the Continuum Si
forms they now lack by the time they started using more and more English to the
detriment of their mother tongue.

2.3.2. Stative and Dynamic Lexical Aspect

There is a large degree of correspondence not only regarding emergence and
disappearance of tense-mood-aspect markers overall but also with respect to the
development and loss of verbal inflections with different types of verb. In his study
of Guyanese Creole, for instance, Bickerton (1975) gives evidence that stative and
non-stative verbs are treated differently with respect to tense-mood-aspect mark
ing. If unmarked, for instance, a non-stative verb is interpreted as referring to a
past situation, while a stative verb would be non-past.
In the specific case of Spanish, there are intriguing similarities between the
Mexican-American case and one particular instance of the acquisition of Spanish
as a second language, as documented by Andersen (n.d.; 1991). Indeed, the devel
opment of some Pret and Imp forms in the Spanish as a second language of two
English-speaking children (at ages 8 -10 and 12-14 ) parallels in a number of in
teresting ways the facts observed in the Los Angeles data with respect to stative
and dynamic verbs.
Sections 2.1 and 2.2 included discussions of the relationship between stativity
and the simplification of Pret and Imp morphology on the one hand and the
development of verbal periphrases on the other. It is striking to note that one of
the children studied by Andersen, Anthony, at an earlier stage of Spanish acqui
sition than his sister Annette, uses Pret morphology only with inherently punctual
verbs. All verbs in past imperfective contexts are stative, and appear with present-
tense morphology. These include vivir to live, ser to be, estar to be, saber to
know, valer to cost, and poder may. Annette, on the other hand, uses these
stative verbs in the Imp when required, but it is not known whether she would
extend their use to perfective contexts, as Los Angeles speakers do. Regardless,
there is sufficient coincidence between stages of acquisition and attrition to sup
port the existence of a basic distinction between statives and other types of verb,
and to suggest psychological and interactional explanations.
It is clear that speakers treat stative and non-stative verbs differently. What is
not so clear is whether this is due to the innateness of this distinction (as proposed
by Bickerton 1981 and Brown 1974, among others), or to the fact that statives
occur much more frequently in imperfective form (cf. Andersen n.d.: 9-10) in
everyday communication, or to both factors converging in the same direction. M y
study cannot throw any light upon the question of innateness, but it does suggest
that the inherent imperfective meaning of certain verbs and their high frequency
of occurrence in imperfective contexts favour the loss of the corresponding Prets,
and thus a consequent change in the meaning of the new Imp forms. At this
stage, the loss of Pret morphology is lexically constrained to occur with prototyp
ical statives, but it is in theory possible for this loss to continue spreading through
52 TenseMoodAspect Across the Continuum.
the verb lexicon in a manner which would reverse stages of development of Pret
and Imp morphology.
There is one further parallel between Andersens data and mine. It concerns the
use of file, the homophonous preterite of ir to go, and ser to be. Just as many of
the adult bilinguals in Table 2.2, Andersens children in the earlier stages of
acquisition use the Pret fue exclusively with the active meaning of went. The
trend appears to be towards a one-to-one relationship between certain forms and
their semantic content, an observation made by many in relation to diverse lan
guage situations.

2.j.j. Analysis

The development and preferential use o f periphrastic or analytical constructions is

also well documented, both across situations of linguistic stress and in the historical
progression of natural non-receding languages.16 The replacement of most of the
nominal case-marking by prepositions is a textbook example in the history of Latin,
which may have been spurred on by intensive contact with other languages during
the period of the Roman Empire. Speakers of Latin and the early Romance
languages substituted several analytical verbal constructions (perfect tenses, ser to
be + PastP passive) for inflectionally marked ones (cf. Fleischman 1982; Vincent
and Harris 1982).
In Mexican-American Spanish, existing periphrastic verbal constructions are
either preferred over their semantically close synthetic ones (all Groups), or new
ones are developed (see section 2.2.2) as a replacement strategy at the lower levels
of the continuum (Groups 2 and 3). In this latter case, tense-mood-aspect dis
tinctions are marked on one or two sorts of wild-card verb (hacer to do/make,
estar to be). This strategy reduces the burden of having to keep verb-stem
morphophonological variants under control, as well as having to maintain a pro
ductive command of tense-marking on the three conjugation types of Spanish
-ar verbs (cantar to sing), -er verbs (comer to eat), and -ir verbs (sufrir to
suffer). I do not concern myself with the changes which affect verb stems in Los
Angeles Spanish. Suffice it to say that most of these changes involve either not
making the morphophonological modifications which standard Spanish requires,
thus keeping the stem regular throughout all tenses and persons (e.g. hacieron they
did for hicieron, from hacer), or making them in tenses where they are not required
(as in puedta I could for podia, from poder). Likewise, maintenance of a tense in
this book means the use of any of the morphemes for that particular tense, regardless
of whether it is correctly marked for person and number or whether it is the
expected one for the conjugation to which the verb belongs.

16 In the historical progression o f natural non-receding languages analyticity and syntheticity appear
to be a continuing cycle (see Ch. 7), as is well attested e.g. by the development o f synthetical and
analytical conditionals and futures in the Romance languages.
 TenseMood-Aspect Across the Continuum 53
The wild-card auxiliary-type verb as a compensatory tactic is reminiscent of
that discussed by Dorian (1981: 150) for East Sutherland Gaelic semi-speakers.
These individuals compensate for their lack of control over morpho(phono)logical
tense-formation devices by relying more heavily on the progressive aspect, and all
progressive tenses make use of one single conjugated verb, that corresponding to
to be. It is also similar, as Dorian (p. 150) notes, to the overuse of mssen must/
have to among foreign workers in Germany, who use this modal as a substitute for
morphological tense-markers of the verb.
The very interesting parallels revealed by a good number of diverse studies give
evidence of the generality of the processes and phenomena which underlie changes
and restructuring of verbal systems in situations of language contact (attrition or
development) and acquisition. I will return to this topic in the last chapter. In the
following section, I examine some of the explanations offered to account for these
phenomena.

2.4. Explanations and Conclusions

With respect to the second research question motivating our study (see section
2.1), various explanations have been offered in the literature for the phenomena
here observed. These explanations take into account certain factors of a different
nature: on the one hand cognitive and social, on the other intra- and interlinguistic.
Thus it is usually the case that more than one factor, or even all of them, may be
motivating and constraining a specific process of change.
Regarding interlinguistic considerations, examination of the progression of sim
plification and loss in our data leads us to conclude that the impact of English is
only indirect. That is, evidence does not seem to be sufficient to conclude that
contact with a typologically different language would have resulted in a system
different from system IV or V. Note, in addition, that direct influence from
English does not justify the order in which the different tenses are lost: e.g. (a) the
early loss of Cond Perf; (b) the early simplification and loss of Pluperf, as discussed
below; (c) the loss of Perf Inf; (d) the retention of the PS down to system IV; (e)
the loss of PresPerf; nor, finally, ( / ) the retention of the morphological Pret-Imp
aspectual opposition down to very low points in the continuum.
In English, these speakers use the Pluperf as expected. Compare [46a, b] with
the literal translations of the Spanish utterances in [47a, b\. These examples are
from a narrative about a burglary told by V21 first in English [46a, b] and just a
few minutes later in Spanish [47a, b],

[46] a. We were waiting for my mom, cause she had taken my brother to the
dentist.
b. I had this piggy-bank, I hadfilled it up and they had broken it. (V21,
fi8,2,ELAi7)
54 TenseMood-Aspect Across the Continuum

[47] a- Y estbamos esperando a mi am - porque ella fue a llevar mi hermano

a la dentista.
Lit.: And (we) were waiting for my mom - because she went to take my
brother to the dentist.
b. Y yo tena una alcanca que tena mucho dinero adentro - y las personas
la agarraron, Y no tena dinero!
Lit.: And I had a piggy-bank which had a" lot of money inside - and the
persons took it, and it had no money inside! [i.e. the burglars had
broken it and taken the money] (V 21 ,f18,2,ELA 17)

Thus, as previous studies have shown (Dorian 1978; Silva-Corvaln 1986, among
others), the effect of English is only indirect, inasmuch as it happens to be the
superordinate language. Some changes occur rather as a result of reduction of both
exposure to and use of a complete variety of a subordinate language in contact with
a superordinate one. These changes affect closed inflectional paradigms. The more
open lexical semantic fields appear to be more permeable to direct influence by a
foreign system (see Chapter 6).
Earlier studies have given evidence that existing internal processes of change are
accelerated in contact situations. This observation is supported in the Mexican-
American data by the early loss of Cond/Perf and simple Fut. On the other hand,
though one may identify both intra- and interlinguistic favouring factors, these
alone predict neither the order nor the type of loss evidenced in the data examined.
Simplification and loss appear to be more appropriately accounted for by intra-
linguistic, cognitive, and interactional considerations (as discussed by Ferguson
1982: 59). Note that the simplest system of grammaticalized tense, system V,
appears to be cognitively less complex, and interactionally most justified in that the
forms retained seem to be the most frequently used in conversation and, one may
surmise, the most useful ones for the speakers communicative purposes.
I use the term cognitive complexity in relation to semantic transparency and to
the number of temporal anchoring points involved in a tense form. Indeed, there
seems to be widespread agreement that grammaticalized distinctions marked by
bound morphology are disfavoured in situations of linguistic stress (e.g. language
acquisition, pidginization; cf. Givn 1979) because of their low semantic transpar
ency and thus higher processing complexity, as compared with more or less cor
responding lexical and periphrastic constructions. In addition, with respect to the
number of temporal anchoring points, absolute-relative tenses are cognitively more
complex. They are like secondary constructions in that their meaning combines
two points of reference. Observe that no morphologically marked absolute-relative
tense is retained in system V, and the only one still present in system IV, the PS,
is undergoing advanced simplification at this stage.
Notice that interactional factors do not favour the retention of absolute-relative
tenses in a contact situation where the domains to which the subordinate language
is restricted call much more frequently for reference to immediate concrete worlds.
 Tense-MoodAspect Across the Continuum 55
In contrast, such factors favour the retention of Pret-Imp morphology, an opposi
tion which plays a crucial communicative role in reference to past situations (Silva-
Corvalan 1983), which in turn constitute a favourite topic of conversation.
Finally, intralinguistic motivations also favour the maintenance of bound Pret
and Imp morphemes as handy markers of realis versus irrealis, and dynamic versus
stative aspect, respectively. In contrast, intralinguistic preferences, i.e. those existing
in Group 1, as well as in Mexican varieties (according to the frequencies of verbal
forms provided by Moreno de Alba 1978: 196-2x6), favouring periphrastic con
structions over synthetic or compound ones account for the early loss o f Cond (as
tense), Fut, Fut Perf, and PPS.
All four factors, then, contribute in complex and interactive ways to the simpli
fication and loss of forms in the various stages of restructuring of the verb system
of a language under stress, in ways similar to those observed in the development
of non-receding languages.
 3

The Effect of Tense-Mood-Aspect

Simplification on Narrative and
Hypothetical Discourse

3.1. Introduction

T h e p r e c e d i n g chapter has documented the simplification and loss of verbal in

flections in the Spanish of Los Angeles bilinguals. This chapter proposes to exam
ine the effect simplification and loss may have on the expression of meaning complexes
(Klein and von Stutterheim n.d.) in hypothetical and narrative discourse. Klein
and von Stutterheim define a meaning complex (p. 1) as the set of temporally,
spatially, logically, etc. ordered information expressed in such texts as route de
scriptions, narratives, and reports. I explore the question of what it is that cannot
be said at the lower stages of the bilingual continuum or, at least, that cannot be
said with the same degree of stylistic flexibility or semantic subtlety as a result
of attrition of the verb system. In order to provide some answers, albeit far from
definitive, to this general question, I analyse and compare narratives and samples
of hypothetical discourse about hypothetical situations produced by Spanish- as
well as by English-dominant bilinguals.
In regard to narratives, the discussion in Chapter 2 allows us to predict that the
retention of Pret forms may permit speakers to communicate at least so-called
foregrounded information, while the loss of absolute-relative tenses as well as
simplification of Imp predict difficulties with the explicit establishment, by means
of verb morphology, of certain temporal sequences. With respect to hypothetical
discourse, simplification and loss of Cond and of Sub forms could have rather
drastic consequences in reference to hypothetical situations, which might be
misinterpreted as factual by an out-group individual.

3.2. Oral Narrative

3.2.1. Analysis

Conversational narratives have been chosen as a source of data because they constitute
clearly identifiable meaning complexes whose internal organization correlates with
some specific linguistic phenomena, e.g. the distribution of tenses in the various

E
 The Effect o f TMA Simplification 57

parts of the narrative.1 In addition, this organization lends itself to an analysis in

u-rms of adjunct, auxiliary, and necessary information in relation to these various
parts. This section focuses, therefore, on both the linguistic and content structures
of the narratives in order to explore the effect that the attrition of certain elements
of the Spanish language system may have on the semantic-pragmatic level. In the
.liialysis, I use a slightly modified version of the framework proposed by Labov
1972^) and Labov and Waletsky (1967). There has been much work on oral
narrative since these publications, some of which has pointed out weaknesses in
the analytical model put forth in them and in its applicability to the study of
narratives produced within different cultures. Despite this criticism, I decided
i!> use Labovs framework, with a few modifications, because it had proved to be
appropriate to explain different time and aspectual interpretations of verb
morphology in an earlier study of Spanish oral narrative that I had conducted
(Silva-Corvaln 1983).
Labov (1972b: 359-60) has defined narrative as one method of recapitulating
past experience by matching a verbal sequence of clauses to the sequence of events
which (it is inferred) actually occurred. The verbal sequence consists of independ
ent clauses which are temporally ordered, i.e. they have temporal juncture and
cannot be reshuffled without changing the semantic interpretation of the order of
the events. Labov posits that a fully developed narrative may show the following
elements: abstract, orientation, complicating action, evaluation, result or resolu
tion, and coda. To these, I propose to add three more elements which occur fairly
frequently in conversational narratives: pre-narrative, preface, and elaboration. Let
us consider them in example 1 from Group 1:

[1] A: A9,f62,iTELA3; B: A29,m6o,2,ELA32

1. A:Entonces, yo jugando con ella, jugando con ella, todo el tiempo.
2. Entonces, volvi en otra ocasin
3. y dice, Pues, sabes qu? dice No le hace lo que t pienses, t te vas a
casar con mi hijo.
4. B: Otra vez le volvi a repitir [s*V] la misma cosa.
5. A: Y le dije yo, Bueno, ndele pues.
6. Yo iba a llevarle la corriente, le llevaba la corriente;
7. todos los das salamos a comprar verdura o alguna cosa;
8. tonces nos encontrbamos, Buenos das, cmo est? Y, y todo esto, y as,
verdad?
9. Todava no le conoca, OK?
10. Pas.
11 . Entonces en una, en una tarde haba terminado . . . habamos terminado mi
hermana y yo de la cena.

' See Tsitsipis (1988) for an interesting ethnographic study o f narrative performance by speakers of
Arvanitika, a receding language.
2 See Ch. I , n. 7 for an explanation o f the abbreviations and symbols used in the examples.
 -T

58 The Effect o f TMA Simplification

12. Y lav los trastes y todo eso;

13. entonces el, los botes de los desperdicios estaban afuera.
14. Entonces, llevaba yo el, el traste de los desperdicios, iba . . . ya en la tarde
como . . . cosa de las seis de la tarde,
15. y sal afuera all a tirar los desperdicios al bote.
16. Y abr la puerta y el
17. . . . sal. *
18. Pero as, al salir la puerta haba ramas, verdad? Y ento -
19. haba una banqueta que era la que xxx hasta los ltimos xxx atrs.
20. Y en eso que voy saliendo
21. y va pasando una persona as,
22. ya mero que le echo los desperdicios,
23. Ay, perdone!
24. Y ya me, me detuve.
25. Oh, no tenga cuidado, dice.
26. Luego la persona sigui caminando.
27. Entonces ya fu i yo
28. y tir los desperdicios.
29. Quin era esa persona? [speaker laughs and looks at her husband]
30. B: Y, y, yo cuando llegu a la casa me qued pensando. Dije, Quin era esa
seorita que est ah, que vive enfrente? Yo nunca la haba visto aqu.
A: Yo no me di cuenta porque ya estaba oscuro.
B: Y yo le pregunt a mi mam.
A: Yo no me di cuenta. Yo no ms vi que era un hombre, pero no me di cuenta
quin era, ni saba tampoco quin era, no lo conoca.
1. A: So, me, just kidding with her, kidding with her, always.
2. So, once she said again
3. she says, You know what? she says, It doesnt matter what you think,
youre going to marry my son.
4. B: She said the same thing to her again.
5. A: And I said, Well, okay.
6. I just wanted to agree with her, I agreed with her;
7. every day we went to buy vegetables or something;
8. so we met Good morning, how are you? And this and that, you see?
9. I hadnt met him yet, OK?
10. and so it went.
11. So one, one evening I had finished, my sister and I had finished having
dinner.
12. And I did the dishes and all that;
13. and the, the garbage cans were outside,
14. And, I was carrying the, the garbage cans, I w as. . . in the evening like,
a t .. . about six in the evening,
15. and I went out there to throw away the garbage.

i
 The Effect o f TMA Simplification 59
16. And I opened the door and the
in, . . . I went out.
18. But, outside by the door there were branches, right? And the-
ig. there was a sidewalk that was the xxx to the end xxx in the back.
20. And as I am going out
21. someones going by like this,
22. I almost throw the garbage on him,
23. Oh, excuse me!
24. And so I, I stopped.
25. Oh, dont worry, he says.
26. And then the person kept walking.
27. So then I went
28. and threw away the garbage.
29. Who was that person? [speaker laughs and looks ather husband]
30. B: And, and me, when I got home I thought. I said, Who was that lady who
was there, who lives across from us? I had never seen her here.
A: I didnt see very well because it was dark.
B: And I asked my mother.
A: I didnt see very well. I just saw that it was a man, but I didnt see who it
was, nor did I know who it was. I hadnt met him.

Example [1] does not have an abstract. Indeed, I have noted before (Silva-
Corvaln, 1983) that abstracts, which summarize the story and occur toward the
beginning, frequently are either absent in conversational narratives or created
interactively by the participants. This section may be unnecessary in [1], given that
the narrative is part of a longer conversational passage about the events which led
to the speakers meeting and marrying her husband. Clause 11 in [1] illustrates
what I call the preface, a statement which announces that the ensuing speech event
is a narrative. Thus, the preface creates the frame within which the speaker expects
the listener to interpret the linguistic material. A number of expressions may serve
as the preface: entonces una tarde then one afternoon, fjate que un da you know,
one day, te voy a contar algo, Im going to tell you something, me acuerdo de una
vez I remember one time, etc.
The. pre-narrative, illustrated by clauses 1--10 in [1], precedes the narrative proper.
It consists of a minimal narrative or sub-narrative (notice the sequence of events
with temporal juncture in clauses 2, 3, and 5), which may provide orientation but
whose main function is evaluative: the pre-narrative builds up interest and curio
sity in the listener. This is what differentiates it from the abstract, which may be
considered to be a summarized pre-narrative with no evaluation function. In [1],
which narrates the speakers first encounter with her husband, the pre-narrative
emphasizes the feeling of the inevitability or predetermination of certain events (t
te vas a casar con mi hijo You are going to marry my son). Indeed, shortly before,
the speaker had said, Porque como te digo, ya estbamos en el plan de Dios que as
6o The Effect o f TMA Simplification

tenian que ser las cosas Because like Im telling you, we were already in Gods plan
that thats the way things had to be.
The counterpart of the pre-narrative is the elaboration, a passage which follows
the narrative proper. As its name indicates, in this section the speaker elaborates on
the events, makes comments, clarifies, and expands the narrative with further
details. Its main function appears to be evaluative: the narrators purpose is to
highlight the fact that the events were important and unusual. Example [i], which
ends with an elaboration section (30), is illustrative.
The evaluation is an important component of a narrative. It conveys the infor
mation that the story is worth reporting because the events were dangerous,
wonderful, hilarious, weird, amusing, or unusual (Labov 1972b: 371). The evalu
ation may be internal, or it may be external, that is to say, in clauses which are not
included in the narrative event structure. Example [1] uses both types of evalua
tion. Notice the use of direct speech in (25) as part of a narrative clause, the skilful
switch to the historical present (HP) in (20), both internal evaluation devices
which highlight the most climactic moments (cf. Silva-Corvalan 1983a), and the
use of a rhetorical question as an external evaluation mechanism in (29). (29) (plus
the speakers gesture toward her husband) also functions as the coda, since it brings
the participants back to the moment of speaking. The resolution, which answers the
question Whatfinally happened?, occurs in (24-8). These lines tell us how the series
of events narrated in [1], which occurred en una tarde one evening (n ), conclude.
The orientation element provides information about the time, place, participants
in the event, and situation in separate orientation sections ( 11, 13, 18, and 19), or
as part of the narrative clauses (14). The sequence of narrative clauses, which are
independent clauses with verbs in the Pret, the Imp, or the HP, in the simple or
progressive form, constitutes the complicating action. In [1], this sequence includes
(12), (14), (15-17), and (20-8), which illustrate all possible tense forms except the
Imp Progressive.
In sum, despite being fairly short, narrative [1] is complete in that it includes all
possible narrative elements but one (an abstract), and it provides auxiliary infor
mation (i.e. orientation and evaluation material), adjunct information (i.e. distinct
reference to entities and events not central to the comprehension of the sequence
of narrative events proper), and necessary information (i.e. information needed
to comprehend unequivocally the sequence of narrative events). For instance, al
though in [1] the information that it was dark is necessary, it is given explicitly
only later in the elaboration section; but at the crucial points the listener may infer
that it was difficult to see and perhaps also dark thanks to the auxiliary information
given in clauses 11 , 14, and 18.
The well-formed semantic and rhetorical structure of the narrative in [1] is
matched by the variety and well-formedness of the morphosyntactic structures.
For the purposes of this study, it is of particular interest to note that progressive
constructions are formed with semi-auxiliary verbs other than estar, namely ir to
go (20-1) and seguir to keep (26), since one of the changes discussed in Chapter
 The Effect o f TMA Simplification 61

2 pertains to the almost exclusive use of estar in progressives by speakers in Groups

2 and 3.
Similar analyses may be proposed for a number of narratives from Group 1
speakers, as [2] further illustrates. Narrative structures are complete; they include
preface (1 in [2]), evaluation, orientation, complicating action, resolution, and ela
boration ((19-24) in [2]). In sum, they provide necessary information, adjunct
information, and a wealth of auxiliary information with orienting and evaluative
function.

[2] L: L3,f23,i,ELA 45
1. L : Fjate que ahorita, con esto de que estamos hablando, me acuerdo de una
vez, no s si mi pap te cont;
2. Este, llegaron unos, unos negros all; en la casa,
3. tocaron, verdad?
4. y este, le dijeron a mi pap que saliera, que traan unas, unas televisiones
robadas, que si queran comprar o no s qu cosa, no s,
5. (a) no s, ahorita no me acuerdo si dijeron que eran robadas o no, no s. (b)
Pero el caso, yo, yo no estaba en la casa.
6. Y s, llegaron,
7. tocaron
8. y le dijeron a mi pap que saliera, que muy baratas.
9. Pues mira, sali.
10. Las, las televisiones estaban como, o sea, estaban en un, cmo te podra
decir?, en cartn, como si fueran nuevecitas y nada ms les podas mirar el
puro frente, el puro frente, o sea la pantalla, es todo, y pues limpiecitas y
todo. Y se miraban as bien empacaditas, as como nuevecitas, como sacadas
de la fbrica, me entiendes?
11. Entonces le pidieron a mi pap como cincuenta- no, cien dlares por dos.
12. Pues dijo mi pap, Pues esto es rpido, rpido.
13. Pas rpido la cosa,
14. que porque los andaban siguiendo, no s qu, verdad?, que rpido.
15. Pues fjate, (a) no les suelta los cien dlares mi pap!? (b) Y, y fue y, y
dndoselos.
16. Desaparecieron.
17. Y que empieza a sacarlas, este, a desempacarlas.
18. (a) No eran televis - , no eran ni televisiones! (b) Eran la pura pantalla!
/laughter/
19. Y le pusieron, no s qu le pondran, pues s, una s tena la pa- s creo una
de esas televisiones s tena la parte, una parte, pero viejsima.
20. T crees? /laughter/ T crees? Hjole!
21. Llegu a la casa y todos serios. Y dije yo, Qu pas?
22. Y yo mir esos mugreros all afuera, /laughter/ Y este, pa qu? Tienen
aqu, ya tenemos bastante, bastante junk [yonk] aqu - - Y este, mi pap
 The Effect o f TMA Simplification

muy serio verdad?, pues jams le haba pasado una cosa as. Y este, y todos
serios - -
Y me contaron. Y yo me di una enojada.
Dije, Ay, y pues cmo se ponen a comprar eso! Y parece, y era cuando
ms necesitaba el dinero uno. Imagnate que, que fueron y los soltaron! Y
yo dije, Dios mo, pero cmo!
Pero s, fjate. Le pasan a uno cosas as, exactamente.
L: You know now, because of what were talking about, I remember once, I
dont know if my dad told you about it;
Some, some blacks went there; to the house,
they knocked, right?
and, they asked my dad to come out, that they had some, some stolen
television sets, would he be interested in buying or something, I dont
know,
(a) I dont know, now I dont remember if they said that they were stolen
or what, I dont know, (b) But anyway, I, I wasnt home.
And so, they came,
and knocked
and they asked my dad to come out, that they were very cheap . . .
So, he went out.
The, the television sets were like, they were in a - how can I tell you? - in
a box, as if they were brand new, and you could only see the front, only the
front, the screen, thats all, and really clean and all. And they looked really
well packed, like really brand new, factory new, you understand?
So they asked my dad fifty - no, one hundred dollars for two.
So my dad said, Well this is fast, fast.
Everything happened very fast,
that they were after them, and so on, I dont know, right?, that it had to be
fast.
So, can you believe that my dad gives them the one hundred dollars!? He
went, and he gives them to them.
They vanished.
And he starts to take them out, uh, to unpack them.
(a) They werent telev - they werent even television sets! (b) They were
just the screen! /laughter/
And they put them, I dont know what they had put them, well one, one
did have the ser - I think one of those sets did have the part, one part, but
very, very old.
Can you believe it? /laughter/ Can you believe it? Gee!
I came home and everyone was very serious. And I said, I wonder whats
happened.
And I saw those thieves outside, /laughter/ And, Why? You have here,
we have here enough, enough junk. - - And uh, my dad was very serious,
 The Effect o f TMA Simplification 63
right?, because nothing like that had ever happened to him. And, everyone
was very serious - -
23. So they told me. And I got really angry.
24. I said, Ay! how can you buy such a thing! And it seems, and it was when
we most needed the money. Imagine! They just went and gave them the
money! And I said, My God! How could you!
25. But you see. These things happen to you, exactly like this.

None of the linguistic structures in the narratives from Group 1 deviate from
general standard spoken Spanish. Furthermore, they illustrate a variety of inde
pendent and subordinate clauses, Ind and Sub verbal forms including, among
others, the Pluperf, the Cond, and the IS and progressive constructions with
verbs other than estar. This variety of tenses and syntactic structures is incorpor
ated in an abundance of evaluation devices which include the following, among
others:
(a) Phatic expressions which stimulate listener involvement and interest, namely
imperative forms (e.g. fjate note, pues mira well, look, imagnate imagine), and
rhetorical questions (e.g. Me entiendes? You follow me?, T crees? Can you believe
it?, Cmo te podra decir? How could I put it to you?).
(b) Comparative clauses and superlatives (e.g. Tena mucho ms dinero que ella
He had a lot more money than her, Una de esas televisiones tena una parte, pero
viejsima One of those television sets had a part that was a very, very old part).
(c) Hypothetical manner clauses (e.g. Las televisiones estaban como.. . como si
fueran nuevecitas The television sets looked as i f . . . as if they were new).
(d) Adjectives qualified by adverbs of degree (e.g. muy baratas very cheap, bien
empacaditas very well packed).
(e) The HP, which frequently co-occurs with climactic events (see [1], clauses
20-2, and [2], clauses 15 and 17).
(/) Exclamations (e.g. Dios mo, pero cmo! M y God, how could you!).
(g) Negative constructions which call the interlocutors attention to a state of
affairs which appears to be contrary to what is expected (e.g. [2], clauses 5b and 18).
(h) Direct reported speech (e.g. [1], clauses 5, 23, 25; [2], clauses 12 and 21-4).
An examination of ten narratives by speakers in Groups 2 and 3 reveals some
interesting differences from those told by Group 1 speakers. These differences
concern mode of delivery, code-switching, evaluative techniques, clause complex
ity, the establishment of referents and topic reference continuity, the verb system,
and others which I do not discuss here (e.g. lexical differences, prepositions).
(a) Mode o f delivery. The most striking difference in this respect refers to the
fact that Group 1 narratives are delivered smoothly, with few self-corrections or
false starts, while the rest of the narratives contain numerous pauses, false starts,
self-corrections, and hesitation markers. In addition, these narratives have fewer of
the prosodic and kinesic features which characterize a performed narrative, namely
expressive sounds, sound effects (e.g. lengthening of phonetic segments, imitation
64 The Effect o f TAA Simplification
of noises made by entities involved in the narrative), motions, and gestures (cf.
Wolfson 1982: 24-9).
(b) Code-switching. Code-switching from Spanish to English does not occur
in the narratives examined from Group 1, although some of the speakers in this
Group code-switch once in a while during the recording sessions. In contrast,
code-switching characterizes nine of the ten narratives from Groups 2 and 3.
Switching to English appears to have two functions in these narratives: to fill in
lexical gaps in Spanish, and to serve as an evaluative mechanism (cf. Koike 1987;
Silva-Corvalan 1983^). Observe the switches (italicized) in [3] from Group 2. Note
that (1) is not italicized because I may have prompted B s switch to English by
introducing the word surfing, for which, to my knowledge, there is no Spanish
equivalent.

t3] B: A29,m6o,2,ELA 2; I: researcher; E: B s son, 0134

I: Si. Pero decan que estaba peligroso para hacer surfing.
1. B: Even for the surfers tambin. /S, s./ Vi uno que se subi en una, una ola
de esas.
2. Y luego quebr - it broke - arriba as,
3- y l estaba as co - , casi mero arriba - ,
4- y, y lo agarr
5- y, y lo, y lo voltio as
6. y se lo, se lo llev all the way in.
(E: Uh, I got a friend of mine at work, he - )
7- B: Yo en San Pedro me agarr una de esas - long time ago when I was younger.
8. Me garro all en-
9. You know where the breaker is in San Pedro? [addressed to the researcher]
I: Haciendo surfing 0 nadando?
10. B: No, nadando.
X I. I used to like to ride, ride the waves in.
(E: You know those guys that - )
12. B: Me agarr una de sas
13- y me voltio as, y - todo el cuerpo.
14. y me llev as -
15 - y iba yo tragando hasta agua.
16. Y - , me, me garro
17 - y me rasp todas las piernas.
18. I was bleeding up on my leg,
19. porque me, me garro abajo
20. y me llev hasta abajo, all the way down -
21. y - me rasp las, las piedras 0 conchas,
22. yo no s qu sera lo que haba abajo.
(E: You know those guys that follow a boat on a parachute?)
 The Effect o f TMA Simplification 65
23. B: M y legs were all bleeding.
(E: You know those guys - sometimes in Acapulco?)
I: Que tienen como unas alas - asi grandes no mas? [Researcher addresses
E]
E: Yeah, its a parachute - they got a rope.
24. B: Oh, that pulls you when you go up in the air.
I: Oh, si, si.
E: And then you release yourself.
25. I got a friend of mine - parachute dragged him - - [E tells his story in
English]3
I: Yes. But they were saying that it was dangerousfor surfing.
1. B: [Even for the surfers] also. /Yes, yes./ I saw one who got on one of those
waves.
2. And then it broke - [it broke] - above like this,
3. and he was like this li- almost on top-,
4. and, and it got him,
5. and, and it, it turned him over like this
6. and it dragged, it dragged him [all the way in].
(E: Uh, I got a friend of mine at work, he - )
7. B: Me in San Pedro one of those got me - [long time ago when I was younger],
8. It got me there in-
9. [ You know where the breaker is in San Pedro?] [addressed to the researcher]
I: Surfing or swimming?
10. B: No, swimming.
11. [I used to like to ride, ride the waves in.]
(E: You know those guys that - )
12. B: One of those got me
13. and turned me over like this, and - my whole body.
14. and it dragged me like this -
15. and I was even swallowing water.
16. And - it, it got me
17. and it scratched all my legs.
18. [I was bleeding up on my leg],
19. because it got me down
20. it dragged me down, [all the way down] -
21. and - they scratched me, the, the rocks or shells,
22. I dont know what was down below.
(E: You know those guys that follow a boat on a parachute?)

3 Note that E illustrates the typical phenomenon I referred to in ch. 1: this young man (approx.34
years old) understands the conversation I am having with his father in Spanish, but his participation is
in English.
66 The Effect o f TMA Simplification
23. B: [My legs were all bleeding].
(E: You know those guys - sometimes in Acapulco?)
I: Those who have like wings - like these big ones? [Researcher addresses E]
E: Yeah, its a parachute - they got a rope.
24. B: Oh, that pulls you when you go up in the air.
I: Oh, yeah, yeah.
E: And then you release yourself.
25. I got a friend of mine parachute dragged him------[E tells his story in
English]3

In [3] the speaker switches to English only to provide auxiliary information with
orientation and evaluative function, as in (7), (9), (11), (18), (20), and (23), i.e. no
complicating action clauses contain switches. As we move down the proficiency
continuum, however, speakers do not appear to code-switch to provide auxiliary
information only, but in addition they switch to English in narrative clauses proper.
This is most likely done to compensate for gaps in their knowledge of the Spanish
lexicon. These lexical gaps may have also prompted some of the switches in [3],
specifically those that relate to surfing. Indeed, the switches in (2), (9), and (11)
appear to indicate that in this speakers linguistic experience, aquatic sports are
talked about in English, thus causing either the non-acquisition or the loss of
Spanish lexical items in this semantic field.
[4-6] are illustrative of switches that occur in narrative clauses in the complicat
ing action. [4] occurs in the narrative presented in [7], which I discuss later in
relation to evaluation.

[4] Y luego, all o f a sudden fue tumbling down. (V 2i,fi8,2,ELA i7)

And then, all o f a sudden, it went tumbling down\
[5] . . . en cinco minutos me dijeron que podia go ahead, y so pues me fui. Y como
cuarta milla alguien estaba coming down at me. Nadie lo paro y me, me machuco
de frente. (M4i,m i8,3,ELA34)
. . . in five minutes they told me that I could go ahead, and so I went. And in
about a quarter of a mile someone was coming down at me. No one stopped him
and he, he hit me head on.
[6] Yo me puse ahi, a - - peliar con ellos, y eso es cuando me - - they tore me
up. (D36,m45,3,ELA43)
I started fighting with, ah, with them, and thats when they tore me up.

(c) Evaluative techniques. Several authors have shown (e.g. Schiffrin 1981; Silva-
Corvalan 1983a) that the HP co-occurs with the most climactic or dramatic events
in an oral narrative and, therefore, that it is an internal evaluation mechanism.
Likewise, it has been argued that direct speech is a kind of play-acting (Haiman
and Thompson 1984) serving to make the narrative more vivid and interest
ing. Furthermore, rhetorical questions and phatic expressions have the effect of
 The Effect o f TMA Simplification 67

involving the listener more forcefully with the various elements of the story.
Narratives from Group 1 incorporate these and other evaluation techniques (listed
in (a)-(h) above), as [1-2] show. In contrast, none of the ten narratives from Groups
2 and 3 include rhetorical questions, nor phatic expressions (see [3]); reported
direct speech occurs only in two of these narratives, and the HP in only one ([7],
clauses 2 and 3), but note that these HP forms are not used in the climactic part
of the complicating action, as they would be in a narrative from Group 1 (see [1],
clauses 20-2). These differences are also suitably illustrated by comparing English
dominant bilinguals narratives in both languages. [7] and [8] provide an appropriate
case; they correspond to a narrative told first in Spanish and immediately after
wards in English, at my request, by one of the speakers in Group 2 (V21).

[7] V: V 2i,fi8 ,2,E L A i7; I: researcher

I: Aj. Total que nos, nos interrumpimos cuando me stabas contando de - A
ver si me acuerdo, cuando tenas - - como cinco o seis aos, me de -
V: Como tres, cuatro aos.
1. Lego con - hermano nos st - tirando la - nos staba tirando cosas de all
arriba del lblo, rblo, - whatever.
2. Gara - un - un clock y lo tir -
3. y yo digo, no ms estaba viendo qu staban tirando.
4. Y lo vi que tir - el, el relol. Luego lo vi y luego,
5. all o f a sudden fue - tumbling down, y me cay aqu,
6. y sal mucha sangre! Y pa, pareca que staba - salendo de mi ojos, pero
era de aqu.
I: De la nariz.
7. V: S, pero muchos, tanto. Me cuerdo que - que gararon un - como una taza
muy grande, y - y all estaba, all staba toda la sangre!
8. Mi mami me agarr, me - me cuerdo que - staba las - perlas de mi mam
ah - y - y me llevaron al hospital. Y me recuerdo del doctor.
9. Me, no ms vi una - big machine de, de abajo mi - you know off - / Claro./
I was laying on the hospital bed. I just saw a big -
10. I dont know, I was scared.
11. Pero despus de eso I got un lollipop; remember the lollipop? Y luego -
como a couple of days after, tena - como stitches pero pareca como una
- roach, una cucaracha - aqu.
12. T ova tengo la cosa poquito, no puede ver porque - I have a lot o f you
know, other scars from the, athletes and like that. /I: Claro/
13. Me cuerdo deso; tambin mi hermano se cuerda deso. /I: No se
acuerda?/ Todava se cuerda. /I: Todava se acuerda, claro./ Todos nos
recuerdamos deso, tambin mi mami. Porque se, se sustaron mucho,
tambin mi hermano.
14. Mi hermano cre - thought that I was dying or something like that. Y - pos
todo eso. T ova me acuerdio de eso.
 The Effect o f TMA Simplification

Era mas mie- It was scary.

I: Aha. So we- we got interrupted when you were telling me- Lets see if I
can remember, when you were - about five or six years old, you we-
V: About three, four years old.
Then with - brother was - throwing - he was throwing things at us from
up a tlee, ree [she self-corrects her pronunciation of the Spanish word for
tree], - whatever. "
Grabs - a - a clock and he threw it -
and I say, I was just looking to see what he was throwing.
And I saw him throw - the, the clock. Then I saw him and then,
all o f a sudden it went - tumbling down, and it fell on me here,
and I bled a lot! And it lo, it looked as if my eyes were bleeding, but it was
from here.
I: From the nose.
V: Yes, but a lot, so much. I remember that - that they grabbed a - like a very
big cup, and - and there, there was all the blood!
My mom grabbed me, I - 1 remember that - - the pearls - my moms pearls
were there - - and - and they took me to the hospital. And I remember the
doctor.
Me, I just saw a - - big machine from, from under the - - you know off -
/Yeah./ I was laying on the hospital bed. I just saw a big -
I dont know, I was scared.
But after that I got a lollipop; remember the lollipop? And then - like a
couple of days after, I had - like stitches but it looked like a - roach, a roach
- here.
I still have that thing, you cant see because - I have a lot o f you know, other
scars from the, athletes and like that. /I: Sure/
I remember that; my brother also remembers that. /I: He doesnt re
member?/ He still remembers. /I: He still remembers, of course./ We all
remember that, my mom also. Because the-, they got very scared, my
brother also.
M y brother thin - thought that I was dying or something like that. And - all
that. I still remember that.
It was more sea - It was scary.

I: Yeah. Could, could you tell me the whole thing in English?

V: In English? /Yes./ Lets see - since the beginning? /Yes./
: When I was young I used to live in El Monte and over there in the front
yard we had this really big humungus tree.
And my brother when he was small he, you know, he had a couple of
friends over, and he was playing with the ball, and it got stuck on top of
the tree. And m, and I was under the tree playing Barbies with my friends
 The Effect o f TMA Simplification 69
3. and ah, he kept throwing stuff, you know, objects, different kind, you know,
big, small
4. and - he finally thought of getting this clock, you know, a bright idea o f his.
And he threw it while I was just, you know, I myself was just there sitting
and looking at him throwing stuff, you know, I was very curious. And then,
when he threw the clock I also watched him - do that.
5. And all of a sudden, before I knew it, it fell on my face.
6. And, you know, before I knew it, I wa-, in front of me there was a big bowl
of blood and everything, and, and then I had - ma, I remember being in
my moms arms in the car, you know, with her. Something ma, I know she
had something in my ma-, face. I dont know what it was. I think it was a
trapo. /Right./
7. And I got to the hosp-, I arrived at the hospital -
8. and I remember being at the hospital table - and a big machine over my
head overlooking- I was overlooking it. And then after, I guess, the opera
tion - I got a candy, a lollipop. /Yeah./
9. And after a couple of weeks - oh, not a couple of weeks, a couple of days
- the scar or whatever I had - looked like a big, you know, cockroach, that
had, you know, just standing there. I used to look in the mirror and used
to say. Oh, I used to hate it! And my brothers and sisters used to make fun
of me and everything. I remember looking at the mirror and I was going
like this [gesture]. /Yeah/

Note that in [7] evaluative statements occur almost exclusively in English switches
(italicized; see in particular (10), (14), and (15)). Switches to English occur also to
compensate for either lack of knowledge or memory lapses in Spanish (see (2), (9),
(11), (12)). By contrast, the speakers English version (in [8]) contains only a one-
word switch to Spanish {trapo rag, (6)), and a number of evaluative statements
which do not occur in the Spanish version (italicized in [8]), such as this really
big humungus tree (1), a bright idea of his (4), I was very curious (4), Oh, I
used to hate it! (9). In addition to switching, the speaker employs other evaluative
techniques in the Spanish version of the narrative which are also present in the
English version, namely adjectives, intensifiers (e.g. (6), (7) in [7], and (1), (3), (8),
(9) in [8]), and repetition. It appears that the speaker uses repetition in Spanish
to achieve rhetorical effect. Note that in [7], (13) she states, although almost as
if reading a list in terms of prosody, that she, her brother, everyone, and her
mom remember those events. This series of statements highlights the fact that the
experience was indeed scary (as she explicitly points out in an English switch in [7],
(15) )-
Furthermore, orientation is much more detailed in [8]: compare (1-3) with the
bare-bone orientation given in [7], (1). There is necessary and adjunct information
in the Spanish version in [7], but the auxiliary information is quite scant, and
prosodic effects (e.g. steep changes in pitch which accompany exclamations in
70 The Effect o f TMA Simplification
Spanish) are missing. Thus, the narrative loses liveliness as compared to its corre
sponding English version, despite the fact that the English version constituted a
retelling of the story. As such, one would not expect it to be as lively and interest
ing as a spontaneous narrative.
(d) Clause complexity. I refer here exclusively to the use of subordination. A
variety of types of subordinate clause occurs in all levels of the proficiency con
tinuum (as illustrated in [3], [7], [9], and [i2])."However, the frequency of use of
these constructions diminishes towards the lower end. Support for this observation
is provided by a more extensive quantitative study of the same speakers examined
in this book conducted by Gutirrez (1990). This author includes nominal, adjec
tival, and adverbial subordinate clauses and shows a decreasing use of subordina
tion, from 46 per cent to 29 per cent to 24 per cent in Groups x, 2, and 3 in the
sections which provide auxiliary information, i.e. where one would expect orient
ing and evaluative subordinate constructions to occur. This seems to be an indi
cation of syntactic simplification, especially so in the case of decreasing use and, in
some narratives, absence of relative clauses, given that these are considered to be
the prototypical exemplar of subordination (Haiman and Thompson 1984). Rela
tive clauses, for instance, do not occur at all in four of the ten narratives from
Groups 2 and 3, while three of four narratives examined for this specific purpose
in Group i4 contain this type of clause.5
(e) The establishment o f referents and topic reference continuity. It is well-known
that Spanish allows the non-expression of subjects and maintains subject reference
continuity by means of verb morphology. In most cases of subject switch refer
ence, however, the subject tends to be expressed, and indeed must be expressed if
the context does not clarify any possible ambiguities between competing referents
(see Chapter 5, and Silva-Corvaln 1982). These discourse/pragmatic rules are
broken only at the very low stages of the Spanish proficiency continuum.
Note that in Spanish only third-person plural verb forms may have a zero non-
referential subject, unless it is an impersonal construction marked with se (Se vive
bien aqu (one) lives well here) or a construction with a weather verb (Llueve mucho
aqu (it) rains a lot here) Thus, while [9] below is well formed and acceptable as
a discourse-initiating statement, [10] is not because non-third-person plural verb
morphology requires previous establishment of a subject referent.

4 Since one o f the narratives included from this Group had relative clauses and the other had none,
I randomly chose two additional ones (from S 2 ,f2 4 ,i,E L A 3 7 and from C 13 ,11145, i ,E L A 4). T h ey both
happened to include relative clauses.
5 On the basis o f an examination o f rates o f relativization in two Uto-Aztecan languages, Hill (1989)
suggests a different interpretation. She acknowledges that reduced rates o f relativization in substratum
languages in language shift may perhaps be due to the failure o f speakers to evaluate narrative in a
stigmatized language (p. 162), but feels much more strongly that, at least in Mexicano, this reduction
is not a reflection o f language attrition but may be associated with the role o f the substratum in coding
solidarity with Mexicano (as opposed to Spanish) by the speakers in her group o f narrow-honorific
men . I have not studied the functions o f relative clauses in m y data; therefore, at this time I can only
evaluate their diminishing occurrence as one further indication o f the overall decrease in the frequency
o f use o f subordinate constructions, as noted in Gutirrez (1990).
 The Effect o f TMA Simplification 7*
[g] Sabes? 0 entraron a robar a la casa de al lado.
You know? [They] broke into the house next door.
[10] *Sabes? 0 entr a robar a la casa de al lado.
You know? [He] broke into the house next door.

The discourse or intersentential rule that requires the previous establishment

of the subject referent of a non-third-person plural verb is violated in the following
narrative by a seventeen-year-old young man from Group 3, but examples of this
sort are rare in the general data and only this case is illustrated in the ten narratives
studied from Groups 2 and 3.

[11] J : j43,m i7,3,ELA 57; I: researcher

1. I: Y la pelea, a ver si me la cuentas.
2. J: Era un da, ea - era en Cal State.
3. I: Cal State LA.?
4. J: Cal State LA. /I: Ah, mira!/ Pero ah no, no me gusta pelear ah. No me
gu, - porque a era xxx era bad luck para ir pelear ah. /I: Fjate!/
5. So esa era lo que - fui a pelear ah,
6. pos la primera vez que pe-, pel ah me, me patiaron la cara y sangr de la
boca. Yo fui a limpiar y - y so - se - la segunda vez pel ah - - me patiaron
en el ojo - y me dejaron black eye. /I: Ah!/ En la tercer vez esoes cuando
me quebraron la nariz. So ah, ah, - ah no me gusta pelear. Ah no iba a
las peleas.
7. Bueno pues, a - ese da fui a peliar y peliamos.
8. Patiaron ah y,
9. y me sal de la pelea
10. y me quit mi nombre de la lista. /I: Aahh./
11. Es que es la - process of elimination, OK?/I: S/ Pelea - dos pelean. El que
pierde se va pa rear, hasta que llegan as un - un ladder.
12. I: Claro como una escalerilla as, s.
13. J: Aha, yeah. Y me ruin.
1. I: And the fight, lets see if you tell it tome.
2. J: It was a day, i- it was at Cal State.
3. I: Cal State LA.?
4. J: Cal State LA. /I: Oh, I see./ But not there, I dont like to fight there. No
I dont li- because it was - it was bad luck to go to fight there. /I: Imagine
that./
5. So that was what - I went to fight there.
6. Well the first time that I fou-, fought there, they kicked my face and I bled
from the mouth. I went to clean and- and so- se - the second time I fought
there - they kicked me in the eye - and they gave me a black eye. The third
time thats when they broke my nose. So there, there, - I dont like to fight
there. I didnt go to the fights there.
72 The Effect o f TMA Simplification
7. Well then, that day I went to fight and [we] fought.
8. They kicked there and,
9. and I withdrew from the fight.
10. and I removed my name from the list. /I: Aah!/
11. Its that its the - process of elimination, okay? /I: Yeah./ Fight - two fight.
The one that loses goes down, until they get like this a - a ladder.
12. I: Right, like a ladder.
13. J: Yeah. And [he] ruined me.
Observe now that in [11J, a story about a certain type of boxing competition,
lines 7 and 13 contain verb forms (p peliamos [we] fought, 0 me ''ruin [he] ruined
me) whose referents have been either only partly introduced (in the case of [we]
fought, the speaker (I) has been introduced, but not the opponent) or not pre
viously introduced in the narrative. It is obviously possible to infer that the speaker
fought against someone (in (7)) and that someone ruined him (in (13)). However,
since the hearer (I was the hearer, and later also the analyst) expects the explicit
introduction of subject referents in this structural and discourse context, the vio
lation of this rule causes a certain degree of puzzlement, i.e. there is loss of what
I call adjunct information.
(f) The verb system. I have shown elsewhere (Silva-Corvaln 1983) that different
tense forms correlate with the various sections of an oral narrative, such that the
specific narrative context in which the verb form occurs in large part determines
its tense and aspect meaning. In addition, our interpretation of the order in which
narrative events take place relative to one another depends to a large extent on the
verb form used to codify such events.
One of the most striking results of the present study is the observation that there
are no deviations from this established pattern in the tense forms of the complicat
ing action sequence in eleven of the twelve narratives analysed. Only one narrative,
which represents a very low point in the continuum (the position of speaker A46
with regard to loss of tense-mood-aspect morphology Table 2.2 is third from
the bottom), demonstrates one case of deviation which makes it difficult to un
derstand the order in which two of the events took place. I refer to the italicized
sentences, (7) and (13), in [12].
[12] A: A4,f3i,3, ELA20; I: researcher
1. A: . . . Y , y me llamaron. Mi mam me llam. Y from East LA to Pico Rivera,
I made it- [gesture to indicate she was really fast]
2. She called the paramedics first, then she calledme.
3. I: Ella- cuando ella llam a los paramedics, tuabu-, tu nana ya estaba - con
el nivel de azcar alto?
4. A: No, no. /I: No?/ No estaba en coma.
5. I: Estaba en coma?
6. A: No que-, no, no quera despertar. /I: Claro./
7. Okay. Cuando llegaron los paramedics - - yo llegu. /I: Claro/
 The Effect o f TMA Simplification 73
8. Ellos entraron y le, le dieron un, un sugar shot. /I: Claro./
9. I: Y t no estabas ah, o ya estabas ah?
10. A: No, no - no /I: Oh, no./
11. Llegu - como el mismo tiempo. /I: Ah, ya./
12. Pero habl con, con un seora que s, que estaba afuera. Un,un neighbor
/I: vecino/, ve, vecina.
13. Y - cuando entr ya, ya le pusieron el sugar shot. /I: Oh./
14. Y , y estaba ciendo, as twitches, /I: convulsiones/, con, convulsions.
1 . A: And, and they called me. M y mom called me. And from East L A to Pico
Rivera, I made it- [gesture to indicate she was really fast]
2. She called the paramedics first, then she called me.
3. I: She, when she called the paramedics, your gra-, your grandmother already
had a high sugar level?
4. A: No, no. /I: No?/ No, she wasnt in a coma.
5. I: She was in a coma?
6. A: No, no, she didnt want to wake up. /I: O f course./
7. Okay. When the paramedics arrived I arrived. /I: Yes./
8. They entered, and they gave her a, a sugar shot. /I: Yes./
9. I: And you werent there, or were you already there?
10. A: No, no - no. /I: Oh, no./
11. I arrived - at about the same time. /I: Oh, O K./
12. But I spoke with, with a lady that was outside. A, a neighbour. /I: Neigh
bour./ Neighbour.
13. And when I entered already, already they gave her the sugar shot. /I: Oh./
14. And, and she was already twitching, like this. /I: Convulsions/ Con-
convulsions.

(7), with a Pret form in both the main clause (llegu I arrived) and the temporal
subordinate one (llegaron they arrived), indicates that the two events were simul
taneous. However, the speaker explains later (note the verbal exchange in (9 -11)
in [12]) that she arrived after the paramedics. Therefore, the relative order of this
sequence of events would have been clearer coded either as in [13] (i.e. with the
order indicated by an adverb), or as in [14] (i.e. with the order signalled by the
verbal forms).

[13] Llegaron los paramedics y luego/despus llegu yo.

The paramedics arrived, and then/after that I arrived.
[14] Yo llegu cuando los paramedics (recin) haban llegado.
I arrived when the paramedics had (just) arrived.

With respect to (13) in [12], it is necessary to note that in Group 1, as in general

standard Spanish, cuando .. .ya with the meaning of when . . . already requires
Pluperf with activity verbs in the clause which contains y a already, as in [15].
74 The Effect o f TMA Simplification

[15] Cuando llam ya se haba ido.

When I called he had already left.

The occurrence of Pret instead of Pluperf prompts a preferred interpretation

(perhaps even the only one) of the sequence cuando .. . ya as when . .. thenfinally,
as in [16]:
[16] Cuando llam ya se fue.
When I called then he finally left.
This is, therefore, the preferred interpretation of (13) in [12]: And when I
entered then finally they gave her the sugar shot.6However, the correct temporal
sequence is precisely the reverse. The use of ya already among speakers at the
lower ends of the continuum needs to be investigated further. It is quite likely, for
instance, that ya may be developing into a tense-marker meaning past, just as in
some creole languages (e.g. Zamboangueo, which indicates past by preposing' ya
to the verbal lexeme, as inyakome (I) ate versus takome (I)m eating (Frake 1977)),
such that, if it occurs with a verb marked for past, this verb would be interpreted
to refer to a past of past situation. If this were the case, (13) in [12] would not be
misinterpreted by those who share these new norms.
In other elements of the narrative, however, tense-form deviations start occur
ring at very high points in the continuum. Notice, for instance, the use of the Imp
for Pret in the orientation in [11], (2) (era it was for fue it was), in the coda in
[17] (haba for hubo there was), and in an abstract in [18] (tenia for tuvo I had).
[17] Y vino y empezamos a pelear, pero ni l gan ni yo. Y despus de eso no haba
problemas. (H 22,m 2i,2,ELAn)
And he came and we started fighting, but neither he nor I won. And after
that there was never any problem.
[18] Tenia un accidente en mi, en mi carro hace como, hace ao, el treinta de abril
va a ser el ao. (M4i,mi8,3,ELA34)
I had an accident in my, in my car about, about a year ago, itll be a year on
the thirtieth of April.
The speakers in Group 1, on the other hand, never fail to use the Pret in these
types of examples. Here, the Pret is required because the examples occur within a
narrative frame, a complex event which is presented as a unit and must be referred
to as a single whole by means of a perfective form. It is interesting to note that
verbs which occur in the Imp instead of the Pret form are not event verbs. Indeed,
these verbs include ser to be, tener to have, and haber there to be, that is to say,
the simplification appears to be limited to stative verbs (cf. section 2.2); it does not

6 T h is was, by the way, the translation given by a graduate student assistant, member o f the
Mexican-American community. T h is student was bom in the U S A but, perhaps because o f her
university specialization in Spanish, has a high level o f proficiency in this language. She helped me
translate a number o f narratives into English.
 The Effect o f TMA Simplification 75
affect action verbs which are on the time line of the temporally sequenced events
in the complicating action.
In the preceding chapter I have shown that the IS and all compound forms start
to be replaced, variably at first, by simple Ind forms in the speech of some speakers
in Group 2. At the low levels of the continuum, both IS and all compound forms
disappear. This type of attrition is illustrated in the narratives studied here. Pret
replaces Pluperf, and it also replaces PluS in the auxiliary information of a narrative
from Group 2 ([19] and [20], respectively).

[19] Y el chofer nomas se paro y dijo si tenia el dinero para pagar su moto. Yo
mientras estaba todo asustado en medio de la calle con mi carro que se - me
apago. (M25,mi9,2,ELA55).
And the driver got up and asked me if I had the money to pay for his
motorcycle. In the meantime I was all scared in the middle of the street with
my car which stopped on me.
[20] I: Y, y el carro se - golpeo o algo, o casi nada?
M: Oh, muy poquito, casi nada, nada casi, ahm, como - si alguien se puso arriba
de mi carro, y le hizo poquito - un - un golpe. (M25,mi9,2,ELA55)
I: And the car got hit or something, or almost nothing?
M: Oh, just a little, almost nothing, almost nothing, ahm, as - if someone
had stood (lit.: stood) on my car, and had done (lit.: did) a little - a - a little
damage.

3.2.2. Conclusions

Let us return to one of our starting questions: what effect does language attrition
have both on the structure and the semantics of oral narratives? Or, what differ
entiates the narratives of a Spanish-dominant speaker from those of a non-Spanish
dominant speaker?
On the one hand, it seems fairly clear that linguistic simplification does
not affect the level of necessary information in the complicating action section. This
is because the Pret form of verbs is retained down to the lowest stages of the
continuum, at least in our sample, and this is precisely the form used to code the
pivotal events in the time line of the narrative. In addition, the content structure
of the narratives (preface, orientation, resolution, etc.) is not significantly reduced.
On the other hand, adjunct information, i.e. information which is in turn part of
auxiliary (evaluative and orienting) information, is lost to some extent, but it is
only at the low end of the Spanish proficiency continuum that instances of this loss
create some small degree of puzzlement in the analyst/interlocutor.
By contrast, there is some significant loss of internal and external evaluation
devices. We may conclude, then, that the narrative provides a frame within which
bilinguals of varying degrees of proficiency can successfully make a number of
statements about past events which may be fairly easily comprehended by an
76 The Effect o f TMA Simplification
interlocutor. However, if the point of telling a narrative lies in its evaluation, it is
likely that language attrition may result in the hearers failure to infer some signifi
cant messages. Assuming that bilingual speakers are aware of this possible loss, I
would predict that in more natural communicative situations, non-Spanish dominant
speakers will narrate their stories entirely in English or will switch to English to
provide auxiliary information as B does in [3], and V in [7]. Obviously, switching
would be possible only if the audience had the necessary competence to understand
spoken English, a type of competence which is fairly extended in the Mexican-
American community of East Los Angeles.

3.3. Discourse about Hypothetical Situations

I proceed now to consider the effect that simplification and loss of verbal morphol
ogy may have on the expression of meaning complexes in hypothetical oral dis
course, i.e. discourse which conveys imaginary, conjectural information, rather than
facts stemming from perception and memory (Klein and von Stutterheim, n.d.:
31). In passages of this type, one would expect frequent use of weak assertive verb
morphology, e.g. Cond and Sub, alongside the more assertive or factual PI and
Pret tenses. Indeed, the use of verb morphology is closely related to the various
manners in which speakers choose to express their attitude towards the possibility
that the contents of a proposition or a number of propositions constituting a
discourse unit may be true in a present or future world, or could have been true
in a past world given certain conditions. The choice among two or more seman
tically similar ways of communicating a piece of information determines a specific
style of communication, which in the case of the expression of possibility char
acterizes it along a scale of assertiveness. Thus, depending upon the linguistic
choices made by a speaker, and the manner in which he structures discourse about
hypothetical situations, his communicative style may appear to be assertive, cat
egorical, assured, defensive, committed, uncommitted, doubtful, or sceptical.
The analysis of hypothetical reference requires topics which elicit both past and
non-past time reference, as well as the expression of various degrees of possibility
and assertiveness. These necessary data were obtained during the recording ses
sions in the Mexican-American community by introducing such topics as (a) for
past reference: life in Mexico if X had not come to the US, life in the US if X had
chosen a different career, job, or school; and (b) for non-past reference: speakers
response to (i) the possibility of X dating or marrying someone from a different
race, religion, or cultural background, (ii) situations which would justify abortion,
(iii) the possibility of having a lot of money. Two or more of these topics were
discussed with every speaker included in the study. This section is based on an
analysis of samples from fourteen speakers, six in Group 1 (E l, L3, F5, C 13, J14 ,
R n ) , and four each in Groups 2 and 3 (A20, V 21, R24, A29, A34, D36, S38, R50).
Apparently not much attention has been paid to conversational discourse which
 The Effect o f TMA Simplification 77
refers to hypothetical situations nor to argumentation in studies of language at
trition. Although hypothesizing is part of argumentation, especially so in essays
and debates dealing with social sciences and humanities topics, not every hypo
thetical discourse is necessarily argumentative. The linguistic strategies which
play a role in the creation of discourse of this type include, on the one hand,
the structure of the hypothetical oral text7 itself and, on the other, such language
devices as verb morphology, the modal verbs poder can/may , deber must, etc.
First, I shall discuss verb morphology in relation to hypotheticality.

j. j. i. The Language o f Hypothetical Discourse

To examine hypothetical discourse it becomes necessary to discuss, even if only
briefly, the modal value of verb morphology, in addition to the specific contribution
of modal verbs and expressions. In this regard, I have proposed (Silva-Corvalan
1990b) that a crucial function of verb morphology is to contribute to the proposi
tion the meaning of more or less assertiveness, where assertiveness is defined as
speaker belief or confidence in the probability that the proposition may be true.
Degrees of assertiveness are pragmatically inferred to convey degrees of
hypotheticality. These two notions are in an inverse relation, such that more
assertiveness is inferred to convey stronger likelihood that the proposition is true,
i.e. less hypotheticality. Assertiveness and hypotheticality correlate with verb
morphology roughly as in the scale in Table 3 .1, which is suggested to apply in
reference to non-counterfactual situations in the data examined. A similar scale has
been argued for in the context of studies of conditional clauses by Klein-Andreu
(ig86/>). She presents a well-justified six-place scale along two axes, assertiveness and
actuality, which I incorporate in Table 3.1. In addition, counterfactuality in the
past is expressed by the PluS and by the Cond Perfect.
Both the discourse topic and verb morphology contribute to creating a hypo
thetical world, along with lexis, syntax, and prosody. The use of verbs which either
weaken the degree of assertiveness of the proposition (e.g. creer to think/to
believe, pensar to think, the modal verbs), or refer to wishes (e.g. querer to want/
to wish, tener ganas to want), as well as the use of such related adverbial and
adjectival items as posible possible, probable probable, quizds maybe, puede que
maybe serve to confirm the non-actuality of the situation referred to. Finally,
hypothetical worlds are created by syntactic means, e.g. conditional clauses, and
temporal clauses which establish a future time frame.

3.3.2. The Structure o f Hypothetical Discourse

The fact that these oral texts refer to situations which speakers have not previously
experienced, and which do not necessarily have an iconic time or spatial line of

7 Oral text and discourse are here used synonymously.

 The Effect o f TMA Simplification
3.1. Scale of Assertiveness Conveyed by Verb
T a b le
Morphology in Reference to Non-past Situations

Most assertiveness
Factual Present Indicative
Preterite
Least hypothetical future
Imperfect Indicative
Present Subjunctive
Most hypothetical Conditional
Imperfect Subjunctive
Least assertiveness

reference and development (as in narratives and descriptions, for instance, where
the order in which utterances are produced is said to mirror the order of events/
spaces in the real world) accounts for their rather loose structure both in terms of
the necessary sections which must compose them and with respect to the relative
order of these sections in the creation of the discourse.
I have examined eighteen samples of hypothetical discourse produced by the six
speakers in Group 1 referred to above. The analysis indicates that there are at least
two types of hypothetical discourse: strictly hypothetical, and hypothetical with
supportive argumentation. In the data examined, strictly hypothetical passages have
a future time perspective and typically occur when the speaker does not seem to
envision any cause for disagreement with his statements; rather, speakers are
speculating about their possible actions given a certain condition (e.g. having a lot
of money, having more free time, becoming ill, retiring). As one would expect in
conversational discourse, argumentation may be embedded in the development of
a strictly hypothetical text, but I will show that it is not part of its main or central
structure. Furthermore, although not attested in the data examined, strictly hypo
thetical discourse could also have a past time orientation; for instance, given a topic
dealing with what a speaker could have done when visiting a certain place, but did
not.
Hypothetical passages which refer to past counterfactual situations usually in
corporate argumentation in their central structure, as do passages where certain
specific issues are being discussed, either within a present or a future time perspec
tive. These oral texts involve the speaker in trying to justify a certain action or line
of reasoning, or weighing the positive or negative consequences of a situation. In
hypothetical texts with supportive argumentation it is always possible to infer that
the speaker expects some degree of disagreement with his stance.
A complete hypothetical discourse with supportive argumentation may comprise
the following elements: hypothetical macro-frame, hypothetical statement, argu
mentation, anchoring, qualification, disclaimer, and coda. Of these, the first two
 The Effect o f TMA Simplification 79

define what I call a minimal hypothetical passage. Strictly hypothetical discourse, on

the other hand, does not include argumentation nor anchoring as part of its central
structure. In what follows, I define and illustrate the various elements contained in
hypothetical discourse.

3.3.2.1. Establishment o f a hypothetical macro-frame

A macro-frame is a body of knowledge that is evoked in order to provide an

inferential base for the production and understanding of discourse.8 In the data
examined, this is in every case done through a question such as Te imaginas cmo
habra sido tu vida si te hubieras quedado en Mxico? Can you imagine what your life
would have been like if you had stayed in Mexico?, or through a statement
followed by a short prompting question, Mucha gente est en contra del aborto, pero
yo lo aceptara en algunos casos, Y t? Many people are against abortion, but I
would accept it in some cases, wouldnt you?9
Establishing a macro-frame, then, necessarily also means establishing a dis
course topic. This defines the sequence of utterances included in what we refer to
as the hypothetical discourse passage. In this sense, a macro-frame shares some
definitional characteristics with what Van Dijk (1980: 332-44) discusses in terms
of macro-speech acts and macro-structures.

3.3.2.2. Hypothetical statements

These consist of one or a series of utterances which present the speakers state
ments about his possible actions, attitudes, and beliefs given a certain state of
affairs. The most frequent construction used to encode hypothetical statements is,
expectedly, the conditional clause of the type if p then q\ If the state of affairs or
condition if p has been established in the hypothetical frame, the speakers state
ments contain in nearly every case only the second part of the conditional period,
then q\ i.e. the assumed consequence of p. However, throughout the discourse
speakers put forth a number of hypothetical statements dependent on certain
conditions presented by the speakers themselves, the content o f which is related to
the hypothetical macro-frame.

8 T h is definition is based on the one given for frame by Levinson (1983: 281): A frame . . . is a body
o f knowledge that is evoked in order to provide an inferential base for the understanding o f an
utterance.
9 All hypothetical macro-frames were created by the researcher. This may perhaps be due to the
nature o f our data-collection procedure, the sociolinguistic recorded dyadic conversation. However, it
seems to me that it may also reflect, i f not some reluctance on the part o f speakers to engage spontan
eously in conversation about hypothetical situations (even when they are not being recorded), at least
a marked preference for topics concerning factual situations.
8o The Effect o f TMA Simplification
3.3.2.3. Argumentation

These discourse passages consist of an utterance or a series of utterances which

present reasons offered in proof, rebuttal of, or as the motivation for a hypothesis,
statement, or position with respect to an issue.

3-3-2.4. Anchoring *

Anchoring refers to the strategy of linking the hypothetical statements to the real
world by means of illustrative narratives or exposition of relevant facts of which
the speaker has direct knowledge. It appears that anchoring has a double function:
communicative and cognitive. Its communicative function is argumentative and
entreating, inasmuch as it supports the speakers hypothesis and aims at convinc
ing the listener of the possibility that the speaker is correct in its assumptions.
Cognitively, anchoring hypothetical texts to experienced facts makes production
and processing easier. Indeed, we have already mentioned that hypothetical dis
course does not stem from perception and memory. This makes the problem of
linearization, conceptualization, and linguistic encoding a harder task as compared
with that of encoding narrative texts, for example (cf. Levelt 1979; 1982). Thus,
speakers frequently illustrate their hypothetical statements with anecdotes about
people they know to be in the sort of world which they imagine for themselves,
and either state explicitly or imply that this is what they would do, would have to
do, or would perhaps have done.

3-3-2-5 Qualifications

Qualifications consist of an utterance or a series of utterances which moderate or

restrict the hypothetical statements, making them less strong. Given a hypothetical
statement of the form if p then q (Si regresamos a Mxico estaremos mejor I f we
return to Mexico well be better o ff), qualifications take the form but i f p then
possibly also q (Pero si regresamos a Mxico quiz no estaremos mejor But if we
return to Mexico perhaps we wont be better o ff), or even a contrary hypothesis
though if p then possibly also q' (Aunque si no regresamos a Mxico quiz tambin
estaremos mejor Though if we dont return to Mexico perhaps well also be better
o ff), or are encoded as a proviso of the form lq, provided that x\ Qualifications are
also conveyed by such verbs as creer think, believe and pensar think, and they
may be reinforced by means of lexical expressions of the type o puede que maybe,
probablemente probably, etc.

3.3.2.6. Disclaimers

These are utterances which convey the speakers reluctance or refusal to accept
responsibility for the certainty of his conjectures about counterfactual and possible
 The Effect o f TMA Simplification 81

situations. Examples of disclaimers are such utterances as No puedo yo saber/asegurar

1 cant tell/be sure, No s, es difcil saber I dont know, its difficult to tell, As
me lo imagino, no?, pero no s This is how I imagine it, right?, but I dont know.

j.j.2 .7. Coda

As in oral narrative, the end of a hypothetical text may be overtly marked by means
of expressions like Eso es lo que yo pienso Thats what I think, As es como habra
sido This is how it wouldve been, Y todo eso habra hecho And I wouldve done
all that. However, as we have noted in the case of conversational narratives, codas
are quite infrequent in conversational hypothetical texts as well.
At this point, let us consider the analysis of the structure and language of
hypothetical discourse proposed through an examination of two examples from
Group 1. Example [21] is a case of hypothetical discourse with supportive argu
mentation. The issue is abortion.

[21] B: C 1 3,m54,1 ,ELA85; I: researcher

I: (a ). . . Pero en qu situaciones, por ejemplo, crees que aceptaras el aborto?
Le aconsejaras a alguien que se hiciera un aborto, en cualquier caso, un
aborto? (IIy/Fr)
B: (b) Bueno, no, no aceptara eso (Hy/St)
(c) porque me sentira yo como, como un asesino quitarle la vida a un
nio. (Argu)
(d) Ni en caso de que, digo yo, (Qual)
(e) muchos padres, pues si ven a sus hijos en problemas as, no los dejan
discurrir por ellos mismos, (H y/St)
( f ) pero creo yo que (Qual)
(g) no sera capaz de quitarle a un hijo, a uno de mis hijos o, o, a cualquiera de,
a mis herma - a mi hermana, o a los - en caso que hubiera alguna vez pasado
ese detalle, verdad? (Hy/St)
(h) Porque yo creo que (Qual)
() la vida, se la da a uno Dios. Y el ser que nace tiene que nacer. (Argu)
(j ) Pero, bueno, eso yo no s (Discl)
(k) si alguna vez en caso de, de que se vaya a morir la persona, verdad? (Hy/
St)
(!) Probablemente, eso yo creo que (Qual)
(m) ni Dios, ni Dios le permitira a nadie, verdad?, dejar morir a la madre por,
por el nio, verdad? (Hy/St)
(n) Pero est difcil tambin porque son dos, son dos seres que se quieren.
(Discl)
I: S, pero la madre ya est aqu ya y el -
B: (o) S. Y luego le hace ms falta a los demshijos,especialmente si tiene familia,
verdad? (Argu)
82 The Effect o f TMA Simplification
ip) En ese caso, pues, todava (jQual)
(q) podra yo, a, como le dicen los, los gringos, verdad?, bend, poquito. (Hy!
St)
I: ( ) ... But in what situations, for instance, do you think you could accept
abortion? Would you advise someone to have an abortion, in all cases,
abortion? (Hy/ Fr)
B: (b) No, I wouldnt accept that (Hy/Si) *
() because I would feel like, like a murderer, taking the life of a child, (Argu)
(d) And in the case of, I think, (Qual)
(e) many parents, well, if they see that their children are in this kind of trouble,
they dont let them decide by themselves, (Hy/St)
( f ) but I think (Qual)
(g) I wouldnt be able to take the child of one of my children, or, or any of, of
my sister, or of - in case something like that had ever happened, right? (Hy/
St)
(h) Because I think that (Qual)
(i) life, God gives it to you. And the being who is born has to be born. (Argu)
(j) But, well, I dont know, (Disci)
(k) if in a given case, if the persons going to die, right? (Hy/St)
(I) probably, that I think that (Qual)
(m) not even God, not even God would allow that to happen to anyone, right?,
to let the mother die for, for the child, right? (Hy/St)
(n) But its difficult also because they are two, they are two people one
loves. (Disci)
I: Yes, but the mother is already here and the -
B: (o) Yes. And then shes needed more by the rest of the children, especially if
she has a family, right? (Argu)
(p) In that case, well maybe (Qual)
(q) I could, as the, the gringos say, right?, bend a little. (Hy/St)

The structure of the hypothetical discourse in [21] illustrates the following

elements in its development:
a) Hypothetical macro-frame
b) Hypothetical statement 1
c) Argumentation
d) Qualification
e) Hypothetical statement 2
f) Qualification
g) Hypothetical statement 3
h) Qualification
i) Argumentation
j) Disclaimer
k,m) Hypothetical statement 4
 The Effect o f TMA Simplification 83

(/) Qualification
(n) Disclaimer
(0) Argumentation
(P) Qualification
(q) Hypothetical statement 5

Note that [21] contains five hypothetical statements, three supported by argu
mentation, and no anchoring. It also incorporates five qualifications, and two
disclaimers. It might be possible to develop a sophisticated method of quantifica
tion of the proportional number of the various elements, also taking verb morphol
ogy and lexicon into account as an index of degree of assertiveness. This complex
and time-consuming task is not attempted here, however, because it is beyond the
specific purpose of this chapter. At this point, based on the larger number of
qualifications and disclaimers over hypothetical statements and argumentation,
plus the absence of anchoring, I interpret the style of [21] as relatively less assertive
than a hypothetical text containing anchoring, and fewer disclaimers and qualifi
cations ([22] incorporates anchoring, for instance, and illustrates a more assertive
style). Thus, my interpretation of the pragmatics of [21] is that the speaker views
the issue of abortion as quite complex and is not prepared to commit himself to
a categorical position. Even though he acknowledges the fact that there may be
attenuating circumstances which could justify abortion (k), he qualifies the related
hypothetical statement (l,m), and proceeds to disclaim responsibility in (n). How
ever, his attitude changes slightly in response to the researchers implication that
the mother is more important than the baby, so that he ends the passage allowing
for the possibility that he could bend a little (q).
The language used in [21] supports my interpretation: four of the five hypo
thetical statements have Cond rather than PI tense morphology; two of the five
qualifications are reinforced (probablemente probably (/), and todava maybe (p)).
Note, furthermore, that hypothetical statement (q) is modalized (podra could), while
hypothetical statement (b) is not (no aceptara would not accept). This indicates
that it is more difficult for this speaker to justify abortion than to disapprove of it,
an attitude which is valued in his social group.
Example [22] presents a hypothetical discourse with anchoring. This passage
deals with the possibility of having to think about what one would do if a son or
a daughter wanted to marry someone from a different cultural and ethnic group.

[22] R: R n ,f4 3,i,E L A 77; C: researcher

C: (a) Qu haras si una de tus hijas o un hijo viniera un da y te dijera Me voy
a casar con un, alguien que es musulmn, o budista, o con un japons, o un
iran, verdad?, pues - (Hy/Fr)
R: (b) Pues no, no he pensado, (Discl)
(c) (Anchoring: illustrative narrative) (Ancho)
tengo la experiencia de una, una supervisora que tena antes.
84 The Effect o f TMA Simplification
(c 1) Ella es de religin tambin catlica y, y hace como unos dos aos, sali la
hija con un judo. (Frame)
Me dice exactamente as que, Parece, pero,
(c2) no creo que llegue a ninguna seriedad, no creo, no. (H y/St)
Y s, yo le digo a fulanita, me acuerdo que es la nica mujer que tiene, tiene tres
hombres, es mayor que R, como dos aos yo creo. Y hace unos tres, cuatro
meses volvi a la oficina a saludarme. Y le pregunt, le preguntaron por su
familia. Me dice,
(3) Pues, fjate que B., B. est haciendo planes ya de casarse. (jQual)
(4) Y le digo, iY cmo aceptan a este gero? (Frame)
(c$) Dice, No s, dice, yo creo que voy a tener que empezar a hacer a, a
convertirlo al catolicismo. (Hy/St)
(c6) No s cmo lo voy a hacer. (Discl)
(cj) Yo le digo a B. que va a ser un, un amigo muy difcil, (Hy/St)
(c8) porque l es un judo, judo, de religin bien, bien fundada, bien sedimentada,
como la de nosotros. (Argu)
( c q ) Deca, No s, voy a tener que, se, se son-, se ri, brome, voy a tener que,
hacerlo, hacerlo catlico. (Hy/St)
(co) De all si no lo logra es que, (Qual)
(rn) bueno, no s, no s, pero ya es cuestin de B, pero yo no he pensado
tampoco en eso. (Discl)
10
C: (d) Pero t crees que podras aceptarlo? T te, imagnate que un da venga
uno [de tus hijos]11 y te diga - (Hy/Fr)
R: (e) Yo pienso, Carmen, que, (Qual)
(f) yo tendra que [aceptarlo]. (Hy/St)
C: (a) What would you do if one of your daughters or a son came to you one day
and said: Im going to marry a, someone who is Muslim, or Buddhist, or a
Japanese, or an Iranian, right?, well - (Hy/Fr)
R: (b) Well no, I havent thought about it, (Disci)
(c) (Anchoring: illustrative narrative) (Ancho)
I know the case of a, a supervisor I had before.
(ci) She is also Catholic and, and about two years ago, her daughter started
going out with a Jew (Frame)
She says exactly this to me: It looks as if, but,
(c2) I dont think this will become serious, I dont think so, I dont. (Hy/
St)
And yes, I tell this woman, I remember she is the only daughter she has, she
has three sons, shes older than R, about two years I think. And about three,

10 T h e passage omitted here deals with the topic o f how couples from even the same ethnic and
cultural group find it extremely difficult to achieve a successful marriage.
" Language material within square brackets does not have phonetic realization, but is inferrable
from the context.
 The Effect o f T M A Simplification 85
four months ago she came by my office to say hello. And I asked her, they
asked her about her family. She says,
(3) Well, I tell you, B, B s making plans to get married. (jQual)
(4) And I say to her: And how come you accept this giiero (anglo)?
(Frame)
(5) She says: I dont know, she says, I think Im going to have to start, start
converting him to Catholicism. (Hy/St)
(c6) I dont know how Im going to do it. (Disci)
(cj) I tell B hes going to be a, a very difficult friend, (Hy/St)
(c8) because hes Jewish, Jewish, with a very strong religious background, very
strong, like us. (Argu)
(eg) She said: I dont know, Ill have to, to, she smi -, she laughed, kidding, Ill
have to, make him, make him Catholic. (Hy/St)
(c 10) And then if she doesnt succeed its because, (Qual)
(c 11) well, I dont know, I dont know, but its now in B s hands, but I havent
thought about that either. (Disci)
10

C: (d) But do you think you could accept it? Do you, imagine that one day one of
[your children]11 came to you
and said - (Hy/Fr)
R: (f) I think, Carmen, that (Qual)
( f ) I would have to [accept it]. (Hy/St)

The structure of 22 is as follows:

(a) Hypothetical macro-frame
(b) Disclaimer
(c) Anchoring
(ti) Real world frame
(c2) Hypothetical statement
(73) Qualification (mixed marriages do occur)
(74) Hypothetical frame reworded
(5) Hypothetical statement
(i'6) Disclaimer
(cj) Hypothetical statement
(<r8) Argumentation
(eg) Repetition of hypothetical statement in c$
(rio) Qualification
( ill) Disclaimer
(d) Hypothetical frame (recreated)
(e) Qualification
(/ ) Hypothetical statement 1
In [22] the speaker restates the researchers hypothetical frame in real world
terms in (ci). Having established that mixed relationships do occur and may in fact
86 The Effect o f TMA Simplification

lead to mixed marriages (<73), the speaker anchors the researchers hypothetical
frame (-4) and her own hypothetical statement (cj) to the comparable factual events
with which she seems to identify herself at the moment of the conversation. The
speakers answer to the hypothetical frame recreated in (d) supports my proposal
that anchoring does indeed fulfil the communicative function of presenting, through
real facts, the speakers own hypothetical statements, and supportive argumenta
tion. In the illustrative narrative, though realizing that mixed marriages cause
difficulties, the mother accepts her daughters decision. Likewise, R s final hypo
thetical statement (/ ) about her own possible actions is that she would have to
accept a mixed marriage as well.
Anchoring as well as verb morphology create a more assured communicative
style than that of [21].12 Note that the elements in the anchoring passage are never
coded with Cond morphology. All these factors, then, lead me to propose that in
[22] the speaker intends to assert quite positively that a son or daughter cannot be
prevented from marrying whoever they choose. But the speakers attitude towards
mixed marriages is one of forced acceptance. This is made clear in the final
hypothetical statement, which the speaker has chosen to modalize with tener que to
have to in the Cond. The alternative without the modal, yo lo aceptaria I would
accept it, would not convey the same implication of resignation to outside forces
compelling her to assent to a situation which might not be approved by her
immediate community or perhaps even by her interlocutor (whom the speaker may
have judged to be Catholic because of her Hispanic background). Furthermore,
despite the inclusion of anchoring and the more assertive verb morphology used in
[22] as compared to [21], the use of Cond and the occurrence of disclaimers and
qualifications make the style of communication somewhat doubtful, i.e. appropri
ately non-assertive.
It seems correct to say, then, that the hypothetical discourse produced by Group
1 speakers characteristically ranges, as expected by its very nature, from weakly
assertive to least assertive, i.e. the situations described are inferred to range from
least to most hypothetical.

3.3.3. Reference to Hypothetical Situations in Groups 2 and 3

By contrast with Group 1, those speakers in Groups 2 and 3 who show loss of
some lexical units related to modality and loss of least assertive verb morphology
give the impression of having an invariably assertive and assured style of commun
ication in reference to hypothetical situations. Furthermore, the structure of this
type of discourse appears to be simpler, in that there are fewer disclaimers and
qualifications. And these speakers produce much less language material (i.e. fewer
words) than those in Group 1, in response to the researchers establishment of a
hypothetical macro-frame. For instance, in a variety of samples of hypothetical
12 T h is comparison would perhaps be more reliable i f both discourse passages dealt with the same
issue.
 The Effect o f TMA Simplification 87
discourse from eight speakers in Groups 2 and 3, there is only one example of
jnchoring,13 there is little or no argumentation, and qualifications are conveyed
lexically rather than syntactically (e.g. creer to believe is used to qualify hypo
thetical statements, not conditional constructions of the type but if p then possibly
also -q \ or iq, provided that x). Consider [23], from the same speaker who told me
the narrative in [7].
[23] V: V 21 ,f 1 8,2,ELA67; C: researcher
C: Y, y, pero t, t piensas casarte un da, no es cierto?
V: S.
C: Y, e, ponte si conoces a alguien de Mxico.
V: Mero, mero Mxico?
C: S.
V: Yo creo que no me caso con una persona as [Hy/St]. Porque yo, e, los hombres
de mero Mxico creen que las seoras, muchachas deben estar en la casa
[.Argu\. Tienen en xxx mismo pensamiento de viejos, t sabes, que, como
seoras estn la casa y nios y no hacen nada, que no pueden hacer nada, no
deben hacer nada - afuera de la casa [Argu]. Y , um, yo no creo que [interrupted
Qual] - No quiero casarme con una persona as [factual statement].
C: And, and, but you, you plan to marry one day, dont you?
V: Yes.
C: And, uh, what if you meet someone from Mexico.
V: From Mexico itself?
C: Yes.
V: I think that I dont marry a person like that [Hy/St]. Because I, eh, men from
Mexico think that ladies, girls ought to be at home [Argu]. They have xxx the
same ideas as the old, you know, that, how ladies are at home, and children,
and they dont do anything, that they cant do anything, that they shouldnt do
anything - outside the home [Argu]. And, uh, I dont think that [interrupted
Qual] - I dont want to marry a person like that [factual statement].
Note one interrupted qualification, no disclaimers, and in fact a closing factual
statement in [23]. The style of communication appears assertive, categorical, de
cidedly against the possibility of marrying a Mexican man. The outright rejection
of this idea, however, may not have been what speaker V 21 wanted to convey. Yet
the absence of qualifications, disclaimers, and of weak assertive verb morphology
contribute to creating this effect.
Thus, the reduction in the amount of language material and consequent absence
of some of the elements of a hypothetical discourse, as well as the loss of Cond and
Sub morphology, block speakers in Groups 2 and 3 from expressing a high degree
of hypotheticality.

13 T h is example is produced by speaker A 34 (f40,3,ELA 83) when imagining her life had she lived
in Mexico.
88 The Effect o f TMA Simplification
Despite these losses, however, which result in a reduced range of styles of
communication (e.g. the more doubtful or sceptical style is absent), misinterpreta
tion rarely occurs. In the data examined, only once does the researcher give sign-,
of confusion about the factual status of the situation referred to. This is because
the speaker (S38, [24]) talks for a relatively long time in reaction to the question
whether being married to an Anglo male would be more or less difficult than, or
very different from, being married to a Mexican-American. Thus, after a while the
hypothetical frame is no longer obvious, given that the speaker uses mostly PI
forms. At this point, the researcher needs to ask for clarification:
[24] S: S38,fi9,3,ELA66; C: researcher
S: . . . Y es todo diferente, y siempre tengo que estar al pendiente que qu es lo
que l quiere, qu es lo que l est acostumbrado, o qu le gusta a l. Y todo
va a ser diferente, todo, todo. Como tengo que pensar en dos maneras en el
mismo tiempo, lo que yo quiero y lo que l quiere.
C: As te imaginas, claro, [clarifying]
S: . . . And everything is different, and I always have to be worrying about what
it is that he wants, what it is that he is used to, or what he likes. And everything
is going to be different, everything, everything. Like I have to think in two
ways at the same time, what I want and what he wants.
C: Thats how you imagine it, isnt it? [clarifying]
The relative successful communication of hypotheticality is achieved basically
because participants share a macro-point of reference or macro-frame for the
interpretation of the utterances as referring to non-factual situations. This frame
is usually created by an explicit or implicit question suggested by the researcher.
Given the assumption that speakers understand this question, their immediate
response is interpreted as referring to a hypothetical world, regardless of the
factuality conveyed by the verb forms used.
In this vein, compare [25] with [26], and [27] with [28]. Note that in [25-8] one
would expect frequent use of weak assertive verb morphology, e.g. Cond and Sub.
As expected, [25] and [27] (from Group 1) contain these forms. By contrast, every
finite form in [26] and [28] (from Group 3) are either in the PI or in the PerFut.
[25] Future, open possibility, A: A9,f2,i,ELA73; C: researcher
C: A ver, vamos a, a, vamos. Si tuvieras pero tanta plata qu vas a hacer, Ali, con
tanta plata? Qu has pensado?
A: S i tuviera tanta plata - No necesito plata para ser feliz.
C: Verdad que no, pero si la tuvieras qu?
A: Mhm. S i la tuviera, Carmen! /C: O qu sueos tienes?/ Bueno, primeramente
lo que haca, adoptaba - ese es mi sueo, si yo tuviera ahorita, si tuviera ese dinero,
adoptaba seminaristas----- Eso sera mi ilusin y all pondra todo mi dinero.
C: Lets see. Now, if you had so, so much money, what are you going to do, Ali,
with so much money? Have you thought - ?
 The Effect o f TMA Simplification 89
A: I f I had so much money - I dont need money to be happy.
C: Right, but if you had it, then what?
A: Mhm. I f I had it, Carmen! /C: Or what dreams do you have?/ Well, firstly,
what I would do, I would adopt - - thats my dream - - i f I had, right now, i f I
had that money, I would adopt seminarians---- that would be my ideal and I would
put all my money there.
[26] R: R24,m2o,2,ELA5o
R: Pues, ya que pago los taxes, me llevo mi dinero, verdad? /um, hmm/Y me compro
una casa y si ya tengo, pues, me compro otro, /um, hmm./ Entonces, busco a mis
hermanos o mi pap, cokay?, y si l tiene biles [pagos] muy grandes, como para
esta casa, pues, le ayudo. Y un carro nuevo para todos hasta que ya no tengo
dinero.
R: Well, after I pay taxes, I take my money, right? /uhm, uhm/ And I buy a house,
and if I have one, well, I buy another, /uhm, uhm/ Then, I look for my
brothers or my father, okay? and if he has big bills, like for this house, well, I
help him. And a new car for everyone until I have no money.
[27] Past, counterfactual; A: A9,f2,i,ELA73; C: researcher
C: .. . Cmo te imaginas tu vida en Mxico, si en vez de venirte para ac te
hubieras quedado en Mxico?
A: Bueno, yo como mujer, mi ilusin era el hogar, porque no tuve escuela. No
tuve, e, la educacin, verdad?, para- Tal vez que si hubiera tenido la oportunidad
s me hubiera preparado, pero la pobreza no lo permiti.
C: .. . How do you imagine your life in Mexico, uh, if instead of coming over here
you had stayed in Mexico?
A: Well, me being a woman, my ideal was the home, because I had no schooling-
Maybe i f I had had the opportunity, then I would have prepared myself, but
poverty didnt permit it.
[28] D: 036,0145,3,ELA90; C: researcher
C: Entonces si t por ejemplo en vez de vivir aqu hubieras vivido en Mxico,
cmo te imaginas tu vida all, de qu manera diferente crees t?
D: Mi primer cosa, que pude hablar el espaol mejor. Y tambin voy a sacar con
- a tener diferentes valores de Mxico en ideal del hombre y de mujer, el
machismo, ideas.
C: Cmo, cmo te imaginas diferente, que hubiera sido as?
D: Yo creo que vivo, yo crea que como hombres de ah, salir con la idea que las
mujeres son majas y son - mocionale - mocionale. /S, emocionales/.
C: Then if you for instance instead of living here had lived in Mexico, how do you
imagine your life there, in what way different do you think?
D: M y first thing, I was able to speak Spanish better. And Ym also going to get - have
different values in Mexico about ideas of man and woman, machismo, ideas.
C: How, how different do you imagine it would have been?
90 The Effect o f TMA Simplification
D: I think I live, I thought like the men there, get the idea that women are prett\
and are - motional - motional. /Yes, emotional./

The hypothetical frame created in [25] and [26] is one of open possibility in
the future: imagine what you would do if you won a lot of money. Within this
context, in [25] speaker A9 conveys different degrees of assertiveness by means of
PI, Imp, IS, and Cond morphology. By contrast, every verb in R.24s passage ([26])
appears in the PI. Thus, though his text is distinctly hypothetical, he conveys a
high level of self-assurance which may be far from the message he really intended
to communicate.
Examples [27-8] illustrate a past counterfactual hypothetical frame: imagine
what your life would have been like in Mexico. Expectedly for a speaker in Group
1, A9 uses PluS in [27], while D36s (Group 3) weakest assertive form is creta (Imp)
I would think. There is no doubt that D36 intends to convey hypotheticality
in [28]. However, his failure to use PluS makes it difficult to maintain the past
counterfactual frame, and leaves open the possibility that he may be presenting the
situation as possible in a future world. This speakers use of the Pret of poder can/
be able to (quepude hablar el espanol mejor that I was able to speak Spanish better)
in answer to the researchers first question is an indication, on the other hand, that
he understands that the frame created is past counterfactual.

3.3.4. Conclusion

The analysis shows that, in regard to the discourse communicative level, quantity
of expression is reduced toward the lower points in the proficiency continuum, in
that speakers in Groups 2 and 3 offer an ever-diminishing range of possibilities as
they speculate either about the past or about the non-past. This fact, however, may
not be related exclusively to the loss of verb morphology; it may also relate to lack
of lexical choices, and perhaps also to speakers reluctance to discuss hypothetical
situations. In turn, this reluctance may be rooted in cultural and cognitive factors.
Although cultural factors should not be dismissed a priori, it seems to me that, if
relevant, they would be so across the continuum, and would not be responsible for
the differences encountered across the three groups. Regarding cognitive factors,
I have already mentioned the problem of linearization, conceptualization, and
linguistic encoding of situations not stemming from perception and memory. In
this respect, it is clear that narration of past events is a cognitively simpler task
than hypothesizing, as explained by Levelt (1979; 1982). Support for this observa
tion comes from the fact that when hypothesizing, even Group 1 speakers continually
anchor their text to factuality by illustrating their statements with anecdotes about
people they know to be in the sort of world which they have been asked to imagine
for themselves.
Group 1 speakers have available to them a system of verb morphology which
allows them to convey different degrees of possibility, assertiveness, predictive
 The Effect o f TMA Simplification 9i

certainty, etc. A further outcome of this study has been to show that reduction of
ihis system restricts the number of possible choices in the field of modality, such
that at the lower levels of the continuum the use of almost exclusively PI morphol
ogy conveys a strong degree of assertiveness and predictive certainty, without
differentiating between more or less possible situations in the hypothetical world
created.

3.4. Summary

The form of an oral text is determined by a number of complex elements of the

environment in which the discourse takes place (e.g. speakers interrelationships,
degree of formality), by the topic of the discourse, and by the speakers belief
systems and the effect they intend to have on their interlocutors. The speakers
assessment of these factors leads him to make a number of choices from those
made available to him by two systems: a system of principles of discourse organ
ization, and the language system. These choices, paradigmatically and syntagmatically
consistent, concern all language levels (phonetics, prosody, lexicon, morphology,
syntax) and the manner in which a given oral text may be structured and developed.
These two types of element have constituted an objective reference point for my
comparison of narrative and hypothetical discourse across the Spanish proficiency
continuum. I have noted that language attrition rarely leads to misunderstanding
in these two types of discourse genre. Almost obviously, mutual comprehension
may be facilitated in the data-collection situations by the interlocutors high degree
of communicative co-operation. In situations of spontaneous verbal interaction,
speakers with low levels of proficiency in Spanish refrain from communicating
in this language at any length. They intersperse their discourse in English with
Spanish words, phrases, or sentences which they feel confident will be clearly
understood. When there arises the uncommon need to communicate in Spanish
with monolingual Spanish-speakers who have not yet developed the ability to
understand English (almost exclusively very recent immigrants), they will most
likely be required to produce relatively simple meaning complexes: route direc
tions, greetings, exposition of factual situations. In this respect, I have shown that
all speakers in the sample examined manage to convey the main events of a past
experience in Spanish. On the other hand, it is clear that language attrition affects
the encoding of auxiliary information in narratives, and leaves almost no stylistic
choices available to Group 2 and 3 speakers in reference to degrees of hypotheticality
in future and counterfactual situations.
 4

Exploring Internal Motivation for Change

4.1. Introduction

C h apter2 explored some of the changes that have affected the verb system of
Spanish as speakers within a number of family nuclei shift to English, the majority
language. In this respect, I showed that simplification and loss of tenses did not
appear to be the result of transfer from the superordinate language. On the other
hand, those tenses already undergoing simplification in the oral Spanish of first-
generation immigrants were lost even in the speech of US-born bilinguals with
high levels of Spanish proficiency. This observation lends support to the hypo
thesis that in language contact situations a number of changes have an internal
motivation, in that (a) they are in progress in the model monolingual variety
before intensive contact occurs and/or (b) they may be spurred by such features as
the semantic opaqueness of certain language specific forms or the relative complexity
of a given paradigm. When shown to be of one of these types, these changes have
been considered to be autonomous developments explicable in terms of the lin
guistic system involved, and quite probably responsive to such cognitive require
ments as the need to simplify and generalize rules, perhaps to make oral production
quicker and more automatic.
Chapter 4 deals with two independent changes that are internally motivated in
the sense of (a) and (b) above: the generalization of estar to be and the placement
of verbal clitic pronouns. I hope to demonstrate that language contact has the
effect of speeding up the diffusion of these changes despite their autonomous or
language-internal source. This chapter also examines a related question: the omission
of required clitic pronouns across the Spanish proficiency continuum.

4.2. The Extension o f Estar

4.2.1. The Ser/Estar Opposition

The particular change discussed here is the extension of the copula estar to be to
contexts previously limited to ser, also translated as to be .1 This type of study
1 One o f the characteristics o f both Rumanian and the Iberian Romance languages (Catalan, Galician,
Spanish, and Portuguese) is the existence o f two copulative forms: Rum. a f i and a sta, Cat. sser and
estar, G al., Sp ., and Port, ser and estar, which originate from Latin esse(re)/sedere and stre, respectively.
Stare disappeared in French (except as part o f the paradigms o f tre to be , a suppletive verb); it has
been retained in Italian with an expanded range o f lexical meanings and some copular and auxiliary
functions.
 Exploring Internal Motivation for Change 93
differs from most previous research on the effects of bilingualism insofar as it
examines a semantic change and its syntactic consequences in the grammar of the
minority language. That is, while trying to throw light on the issues raised by
language contact, I explore the complex question of semantic-syntactic variation
and change.
Two specific questions, raised by the linguistic problem at hand and by the
sociolinguistic situation, are investigated: (a) How does the extension of estar
proceed, i.e. what linguistic and social contexts are more or less favourable to the
innovation? (b) What effect does language contact have on the actuation of the
change?
The study of estar is based on language samples from thirty-three of the Mexican-
American bilinguals. They are all conversationally fluent in Spanish, but the range
of topics in which fluency is achieved is gradually more and more reduced towards
the lower end of the proficiency continuum. I focus exclusively on predicate
adjective constructions of the type illustrated in [13]:

[1] [Mi abuelita] es blanca. Ni es gorda ni es delgada. Esta bien. (M 8,mi7,i,

ELA59)
[My grandmother] is (ser) fair-skinned. Shes (ser) neither fat nor slim. Shes
(estar) okay.

[2] (a) El es bianco y - (b) no esta, ni esta; gordo ni esta flaco. Esta en medio.
(H22, m 2i,2,E L A n )
(a) Hes (ser) fair-skinned and - (b) hes not, hes (estar) neither fat nor skinny.
Hes medium.

[3] Y yo le dije patras: (a) Pero yo estoy inteligente y muy guapo y no te puedo
tener todo. (D36,m45,3,ELA43)
And I answered back to him: (a) But Im (estar) intelligent and very handsome
and I cant have everything.

Examples [1-3] are produced by three speakers from the three different groups
studied. [2b] and [3a] illustrate the diffusion of estar to contexts in which standard
varieties of modern Spanish allow only ser. The extension to new contexts repre
sents a more advanced stage in a continuous process of syntactic-semantic extension
of the copula estar throughout the history of Spanish. I will show here that the
more recent development in the context of predicate adjectives involves the
elimination of some selectional restrictions which apply to the choice of estar in this
environment. The ultimate result would be the loss of a semantic distinction in the
system of two contrasting copulas. At the stage of development characterized here,
the system is undergoing simplification, i.e. the progressive generalization of a
form X (estar) to a larger number of contexts, which in turn implies the loss or
reduction of the use of a competing form Y (ser).
As illustrated in [4-7], modern Spanish predicate adjective constructions fall
94 Exploring Internal Motivation for Change

into three groups: (a) those which allow both ser and estar ([4-5]); (b) those which
allow only estar ([6]); (c) those which allow only ser ([7]).

[4] Juan es/est alto.

John is (ser/estar) tall.
[5] La luna es/est redonda.
The moon is (ser/estar) round. *
[6] La taza est/*es vaca.
The cup is (estar/*ser) empty.
[7] El discurso es/*est breve.
The speech is (ser/*estar) brief.

The distribution of the copulas, especially in the context of predicate adjectives,

has been one of the most thoroughly studied and most debated issues in Spanish
linguistics. (Useful monographs on this subject are Falk 1979; Navas Ruiz 1963,
Va-Cerd 1982.) Indeed, the syntactic and semantic complexity of the question
is due in part to the fact that in Spanish the Latin meaning of estar (from Lat. stare)
to stand has undergone a series of modifications and has become more general,2
leading to the occurrence of estar with an increasing number of adjectives. This
process of change may be observed through comparison of some examples from
the Poema de Mi Cid in two versions: the original version, in Old Spanish (OS),
and a translation into Modern Spanish (MS). (The examples are taken from CopQeag
and Escudero, 1966: 342-3.)
[8] (a) OS: non se abre la puerta, - ca bien era cerrada
(b) M S: mas no cede la puerta, - que estaba bien cerrada.
but the door doesnt open, - because it was well locked.
[9] (a) OS: Armado es mi Cid - con quantos quel ha.
(b) M S: Armado est el Cid - y cuantos con l estn.
the Cid is armed - and (so) are all those who are with him.
[10] (a) OS: Dios, qu alegre era - tod cristianismo!
(b) M S: Dios, qu alegre estaba - todo el cristianismo!
God, how content Christians were!
In the (a) versions of [8--10], the copula employed is ser; in the (b) versions it is
estar. In Modern Spanish, estar is required in the contexts illustrated in [8] and [9].
On the other hand, the syntactic structure of [10] allows both ser and estar al
though, as is well known, with different semantic interpretations: in M S, [10a], with
ser, would describe the nature of Christianity as happy, while [ 1 o j, with estar, would
refer to a state of happiness, implying that Christianity may not be happy at all
2 As pointed out by Fountain (1982), the weakening o f stare to a point in which it is little more than
a copula, and its encroaching on some o f the functions originally fulfilled by *esse(re), have happened,
though to different extents, in all the Ibero-Romance languages. Copieag and Escudero (1966) suggest
that Rumanian may be following a development similar to that o f Spanish.
 Exploring Internal Motivation for Change 95
T able 4 .1 . Parameters Explicative o f the Choice between Ser and Estar in
Contrasting Contexts

Ser Estar

Inherent or essential Accidental or circumstantial

Tmperfective Perfective
Permanent Temporary
Defining, abstract, and independent of Dependent on concrete and/or immediate
immediate experience experience
Not susceptible to change Susceptible to change
Presented within a class frame o f Presented within an individual frame o f
reference reference

times. It is clear, then, that throughout the history of Spanish the contexts of use
of estar have increased to the detriment of those of ser. At the same time, the forms
contrast semantically and pragmatically in a number of syntactic contexts, with
more or less clear meaning differences according to the nature of the subject
referent and of the attribute assigned to it. Thus, with a large number of adjectives
both copulas are accepted (see [4-5]); but some adjectives co-occur only with estar
(e.g. lleno full, contento content), and some under normal conditions co-occur
only with ser (e.g. inodoro odourless, ambicioso ambitious, inteligente intelligent,
see also [7]). Significantly, this is so under normal conditions; in exclamatory or
sarcastic comments, when the speaker wishes to convey the message that the
characteristic attributed to the referent of the subject is somehow surprising or
unexpected, the copula employed is estar, as in /jQue inteligente estds! How clever
you are (today)! v. Eres inteligente You are intelligent. Indeed, Querido (cited in
Lujan 1981: 172) suggests that it is always possible to find a context for the use of
estar with an adjective which normally co-occurs with ser, but not vice versa.
Various parameters have been proposed to account for the choice between ser
and estar in contexts where either may occur; some of them are presented in Table
4.1. (See both Falk and Navas Ruiz for detailed evaluations of the parameters
included in this table.)
For the purposes of this study, it seemed to me that the concepts of class v.
individual frame of reference (based on Falk 1979),3 and susceptibility v. non
susceptibility to change, proved the most adequate to explain the choice between
ser and estar and the tilt toward estar in contexts where more conservative varieties
of Spanish choose ser either categorically or more frequently. The use of estar in
newer contexts implies a semantic change concerning the parameters that motivate
the choice of copula. This raises the interesting question of the meaning that these

3 Falk discusses at length the applicability o f these concepts, referred to by him as class norm and
individual norm, to the distribution o f ser and estar.
g6 Exploring Internal Motivation for Change

copulas may have such that these meanings allow for changes in the uses of the
forms and possibly also for changes of meaning. However, providing a definitive
answer to this question is beyond the scope of this chapter.
The parameter o f class v. individual frame of reference accounts for construc
tions like [4] and [5] as follows. The proposition Juan es alto attributes to Juan the
characteristic of being tall on the basis of a frame of reference which the speaker
has for a set of entities (masculine human beings) of which Juan is a member;
conversely, Juan estd alto credits Juan with tallness within an individual frame of
reference, which the speaker has established on the basis of his previous knowledge
of this specific human being. In [5], the attribute with ser predicates that the moon
is round in regard to the set of entities of which roundness may be predicated; with
estar, the attribute assigns roundness within an individual frame of reference
implying that it is not always so. Accordingly, it may be possible to imagine a
situation in which [5] with estar is not acceptable one when a speaker sees the
moon for the first time, during a full moon. In this specific (though rare) situation,
example [11] (with ser) is acceptable, whereas [12] (with estar) is not:
[11] Mira! La luna es redonda.
[12] Mira! *La luna esta redonda.
Look! The moon is round.
The individual frame of reference includes some of the other parameters in
Table 4.1. Indeed, the establishment of such a frame requires a concrete previous
experience of the referent to which the given attribute is assigned, and implies that
the relationship between referent and attribute is susceptible to change. Thus, [12]
is acceptable in the real world (since it may be safely assumed that speakers have
indeed seen the moon in its various changing shapes), but while [13a] is also ac
ceptable, the choice of estar in [13 b] is not. In fact, under a set of normal conditions
the entity building does not allow for the establishment of any individual frame
of reference with respect to the attribute roundness, given its exemption from
change in regard to that attribute.
[13] (a) El edificio es redondo.
(b) *E1 edificio esta redondo.
The building is round.
Perfectivity, temporariness, and circumstantiality may also be subsumed under
the individual frame of reference. Indeed, these three parameters imply suscep
tibility to change. In turn, the establishment that a given attribute is susceptible
to change requires knowledge of the entity to which it is being assigned so that it
may be compared with itself at different points in time. Note that the cases so
often quoted as counterexamples to the concepts of perfectivity, temporality, and
circumstantiality (Estd viejo/calvo!muerto Hes (estar) old/bald/dead) are nicely
accounted for by the notion o f individual frame, which allows for change but does
not specify that the new state be perfective or temporary.
 Exploring Internal Motivation for Change 97
Without entering into any detailed discussion of the explanations proposed for
the ser/estar opposition, I hope to have suggested the complexity of the question
before presenting the study undertaken in Los Angeles. Clearly, the choice be
tween ser and estar does not depend only on syntactic or lexical constraints. Indeed,
the acceptability of either copula may not be determined by the researcher on the
basis of the relevant sentence alone; the extended discourse and shared knowledge
among the interlocutors must be considered.
The data for this specific study were obtained through recordings of conversa
tions between the author and 33 of the Mexican-Americans under examination,
including men and women of different ages from the three basic groups established.
Approximately fifty hours of transcribed materials were analysed.4

4.2.2. Analysis and results

Let us look at [1-3] this time in the pertinent discourse passages [14-16].

[14] Mi abuelita es - (a) no es alta ni es baja. Es como la mitad, como - cinco pies,
algo as. Y es - de pelo negro. Lo tena largo, creo que se lo cort. Lo tiene
como hasta la cintura - ella. Y - usa lentes mi abuelita. Tiene - - (b) es blanca,
(r) Ni es gorda ni es delgada. Est bien. (M8,m 17 ,1 ,ELA59)
M y grandmother is - (a) Shes (ser) neither tall nor short. Shes like average,
like - five feet, something like that. And shes - black-haired. She used to have
it long, I think she had it cut. It goes down to her waist. And - she wears
eyeglasses, my grandmother. She has - - (b) shes (ser) white-skinned. (<;) Shes
(ser) neither fat nor slim. Shes just right.
[15] El es como cinco pies y - - diez pulgadas de altura. Ahmm, est un poco calvo
- - y su cabello es negro y as chino. Ahmm, tiene los ojos verdes. Est - -
(a) l es blanco y - - no est, (b) ni est gordo ni est flaco. Est en medio,
as, no s cmo decirlo. /I: Mediano?/ Mediano, s, (c) est mediano.
( 2 2 ,m 2 1 ,2 , E L A ii)
Hes about five foot and - - ten inches tall. Uhmm, hes a little bald - and his
hair is black and very curly. Uhmm, he has green eyes. Hes - (a) hes (ser)
white-skinned and - hes not, (b) hes (estar) neither fat nor skinny. Hes
medium, I dont know how to put it. /I: Average?/ Average, yes, (c) hes (estar)
[of] average [build].
[16] Mi pap era un hombre - muy alto, (a) Todos los Campas son altos, - como
me dijo mi to - menos usted, Daniel. /I: Te dijo?/ El cabrn! Y yo le dije

4 In an earlier study (Silva-Corvaln 1986), I examined estar in a smaller sample o f 27 speakers.

Another difference concerns the fact that the 1986 article quantified innovative uses o f estar as opposed
to conservative uses o f the same copula. M ore appropriately, the present chapter quantifies these
innovative cases o f estar in predicate adjective constructions as opposed to tokens o f ser, all conservative
uses, given that the former are encroaching into the realm o f ser, and not o f estar. I would like to thank
Elvira Behrend (cf. Behrend 1986) for bringing this crucial fact to my attention.
98 Exploring Internal Motivation for Change
patrs: (b) Pero yo estoy inteligente y muy guapo y no te puedo tener
todo. (036,11145,3,ELA43)
M y father was a very tall man. (a) All the Campas are (ser) tall, - as my unde
once told me - except you, Daniel. /I: He really told you so?/ The son of a
bitch! And I answered back to him: (b) But Im (estar) intelligent and very
handsome and I cant have everything.

These passages illustrate descriptive discourse, dealing here with people unknown
to the interlocutor. Having observed that estar appeared to be used innovatively in
the community under study, I tried to elicit examples with ser and estar in the flow
of conversation by leading the speakers toward describing people and places unknown
to me. By the discourse/pragmatic norms of standard general Spanish,5 this type
of description is carried out within a class frame. Even if the relationship between
referent and attribute is susceptible to change, the copula used is ser, unless an
individual frame has been explicitly established (e.g. through lexical means such as
the time adverbs now and before) Interesting exceptions in this respect are such
adjectives as calvo bald and viejo old, which in descriptions are far more fre
quently introduced with estar. (Note its use in [15]: est un poco calvo hes a little
bald.) The motivation appears to be pragmatic. In our society, being bald or old
is no asset; therefore, speakers tend to treat these not as defining features but as
circumstantial states.
In agreement with speakers of standard general Spanish, a member of Group 1
produces examples [14a-c], with ser. In the same communicative discourse con
text, a speaker in Group 2 extends the use of estar to introduce the adjectives gordo
fat\ flaco skinny, and mediano [of] average [build] in examples [15a-c]. In [i6|,
a speaker in Group 3 goes so far as to combine estar with the adjective inteligente
intelligent, even though intelligence is considered to be an inherent quality that
is not susceptible to change, and the proposition is clearly presented in a class
frame. These facts appear to indicate that the innovative uses of estar follow a path
of lexico-semantic diffusion within the linguistic system, and that in regard to
social factors, speakers whose families have been residing in the US longer are at
the most advanced stage. Similarly, Fernndez (1964: 277) observes that estar has
gradually become accepted with an increasing number of adjectives throughout the
historical development of Spanish. He points out that it extended to lleno full earlier
than it did to contento content, finally became obligatory with both, and it has
now timidly started to be used with feliz happy (my translation). I must add
that my impression is that Latin Americans are far from timid in this respect, i.e.
we attribute the quality of being feliz more frequently with estar than with ser.
With respect to the important question of the meaning of estar, one could make

5 Standard general Spanish here serves to refer to the Spanish spoken by college-educated native
Spanish speakers from different countries in the Hispanic world, e.g. Argentina, Chile, Colombia,
M exico, Spain, Venezuela. Various speakers from these countries were consulted; they all agreed with
my analysis o f a required class frame in the description o f people and places unknown to the inter
locutor.
 Exploring Internal Motivation for Change 99

the strong claim that a single invariant meaning pervades all its uses, including its
occurrence in locative predicates (Juan est en casa John is at home) and pro
gressive constructions (Juan est leyendo John is reading). Indeed, one can argue
that this abstract invariant meaning could be something like state of being . This
general meaning, however, would not account for the choice between ser and estar
in all cases.6 In this study I operate with a somewhat weaker claim: I assume that
linguistic forms may have a prototypical meaning which accounts for the core of
their uses; this prototypical meaning may include a composite of defining features
which, as Coleman and Kay (1981: 43) have argued in regard to the meaning of
Eng. lie, are not necessary conditions, and the evaluative logic according to which
these conditions are found to be satisfied, or not, is in general one of degree rather
than of simple truth and falsity . This notion of fuzzy boundaries and non
discreteness may allow us to understand the diachronic process of ever-extending
contexts in which estar has come to be used in Spanish. Indeed, the long disagree
ment on the meaning of the opposition ser! estar may in large part be due to the
discussants insistence on proposing a clear-cut dichotomy and, in many cases, one
explanatory parameter (e.g. imperfective v. perfective aspect) to the exclusion of
others (e.g. the pragmatic norm illustrated by [14]). Actually, it seems possible that
all or, at least, some o f the features proposed (see Table 4.2) are contained in the
prototypical meaning of ser and estar. Furthermore, the semantics of the attribute
and its relationship with the subject referent may also have fuzzy edges and lend
themselves to various degrees of acceptability with one or the other copula, as
illustrated by the various judgements of the following:
[17] Su nia est alta. Her girl is tall.
[18] Su nia est bonita. Her girl is pretty.
[19] ?Su nariz est grande. Her nose is big.
[20] PSus pestaas estn lacias. Her eyelashes are straight .
[21] ??Su piel est clara. Her skin is white.
[22] *Su nia est inteligente. Her girl is intelligent.
Outside a communicative context, where words seem to be associated with their
prototypical meanings, [17, 18] are acceptable; [19, 20] are slightly more acceptable
than [21]; and [22] is clearly unacceptable. But the prototype view suggests that,
in actual communication, speakers may choose to focus on one or another of the
elements contained in the prototype, thus opening the door to a semantic change.
Changes in the semantics of nouns and attributes may bring about changes in the
manner of conceptualizing their relationship and, consequently, in the selection of
a copula. However, I have no evidence that the adjectives used innovatively with
6 Strictly, it would not account for the unacceptability o f est in example (b) below, for instance: (a)
E l jarrn est sucio T h e vase is (estar) dirty; (b) El jarrn es/*est grande. T h e vase is (ser/*estar) big;
(c) Juanito est grande. Johnnys (estar) big . Interestingly, however, ()-type sentences with estar
constitute a high percentage o f the innovative tokens found in the data examined.
ioo Exploring Internal Motivation for Change
estar (e.g. gordo fat\joven young, inteligente intelligent, interesante interesting)
have undergone semantic changes relevant to the problem at hand. Rather, it
appears that the innovative uses reflect a process of loss of certain selectional
restrictions on estar.
The question of the meaning of nouns and adjectives is different from that of
copulas like ser and estar. By definition, a copula is a link, a connector, and carries
little or no lexical meaning. However, the fact that there exists an opposition
between ser and estar suggests that these forms do carry some semantic load. I wish
to propose that in the more conservative varieties of Spanish, the meaning of estar
in opposition to ser includes a composite of at least two elements: state of being
and perfectivity. This permits the selection of estar in three pragmatic contexts:
individual frame of reference, susceptibility to change, and circumstantiality. As
concerns change, I would like to suggest that the nature of everyday communica
tion (subject referents are normally known to the interlocutors and attributes are
therefore assigned within an individual frame) blurs the distinction between class
and individual frame and leads speakers to focus away from this parameter. At this
stage, the focus is on the use of estar to introduce attributes which are circum
stantial and/or susceptible to change. Thus, within the linguistic system, one
might identify three stages. At Stage I the meaning of estar involves both state of
being and perfectivity; its selection is constrained by (a) a discourse/pragmatic
norm which requires that in certain contexts the attribute be presented within an
individual frame and (b) the nature of the subject referent and of the attribute
assigned to it, such that the latter must be susceptible to change, and/or it must
be circumstantial. At Stage II, the individual frame restriction has been lost, and
the nature of the subject referent with respect to change and circumstantiality may
or may not be relevant.7 At Stage III, the only constraint on the selection of estar
is that it should introduce an attribute; the meaning of the copula is state of
being, i.e. the element of perfectivity is lost. At this stage, then, ser and estar
would no longer contrast in predicate adjectives. The occurrence of ser would be
lexically limited to a small group of homonymous adjectives of the type of listo: Juan
es listo Juan is (ser) smart, Juan estd listo Juan is (estar) ready . All this is sum
marized in Table 4.2.
In reality, the speech of individuals is characterized by a certain amount of
variation between a more restricted selection of estar, as in Stage I, and a less re
stricted one, as in Stage III. The placement of speakers in any one of these three
stages would not depend, therefore, on the presence or absence of a more or less
restricted selectional rule in their grammar, but on the frequency of use of estar in
constructions where the conditions of susceptibility to change and individual frame
are unfulfilled. Note, however, that Tables 4.5 and 4.7 below show that two
further linguistic factors appear to play a role in the diffusion of the change:

7 These possibilities are suggested by putting subject (S) in parentheses in Stage II in Table 4.3. A
stands for attribute .
 Exploring Internal Motivation for Change 10 1

T able 4.2. Stag e s in the E xten sion o f Estar

Stage I Stage II Stage III

S estar A (S) estar A estar A

Meaning
State o f being State o f being State o f being
Perfectivity Perfectivity
Pragmatics
Individual frame
Susceptible to change Susceptible to change
Circumstantial attribute Circumstantial attribute Attribute

S = subject; A = attribute.

semantic transparency of the choice between ser and estar, and type of adjective.
The effect that these two factors may have on the stages proposed in Table 4.2 has
not been studied quantitatively. Table 4.2 is based on my qualitative analysis of the
data, which indicates that only speakers in Group 3 variably reach Stage III, i.e.
they use estar with adjectives which under no conditions appear to refer to at
tributes either circumstantial or susceptible to change. The amount of data is
insufficient to quantify the stages displayed in Table 4.2 for each speaker, while at
the same time controlling for individual frame, susceptibility to change, and
circumstantiality. One would expect this quantitative analysis to show that speakers
with lower levels of Spanish proficiency evidence more frequent uses of estar
corresponding to Stage III.
At this synchronic state of the language, one would have to postulate that the
speakers grammar includes a variable rule of copula selection in contexts where ser
and estar seem synonymous. Indeed, this synonymy must be demonstrated for
every token of estar coded innovative. The variation is illustrated with the speech
of the group 3 speaker who produces [160], with the required copula ser, and [16b],
which presupposes the elimination of all constraints on the selection of estar except
that of introducing an attribute. Likewise, first- and second-group speakers evid
ence variable systems, as flows from the descriptive passages in [23-6]:
[23] Lo unico que tiene que (a) es chaparrita. Te -, tiene pelo corto. (b) No es muy
gorda ni flaca ni - De cara pues no se me hace fea. (c) Esta bonita. (d) Pa mi
tiene que estar bonita. (M6,11127,1 ,ELA 51)
The only thing is that (a) shes (ser) short. Sh -, she has short hair, (b) Shes
(ser) neither fat nor skinny nor - Her face, you see, I dont think shes ugly.
(c) Shes (estar) pretty, (d) Well, in my opinion, shes (estar) pretty.
[24] Rasgos de ella? Mira, (a) la nariz de ella no es como la mia; esta un poco
grande, pero - anchita. (b) Ah, sus ojos son chicos como los mios. Su cara, tu
sabes, es - bueno, (<r) era muy bonita mi madre. (E 12/44,1,ELA27)
102 Exploring Internal Motivation for Change
Her features? Well, (a) her nose is not like mine; its (estar) a little big, but
- wide, (b) Ah, her eyes are (ser) small like mine. Her face, you know, its
well, (c) she was very pretty, my mother.

[25] E: Tiene las pestaas - /A: chinas/ Si, chinitas. No las tiene muy tupidas,
verdad?, como hay veces unas personas tienen -
A: S. Hay gente que Ay, y qu chulas se Jes ven!, verdad?
E: Mm - no las, tiene - s. No las tiene as de tejavn, como dicen, verdad?
A: De becerro, (a) As he odo yo la forma cuando [las pestaas] estn
lacias, lacias - -
A: No, no, no, no. (b) Yo digo que la ma [la nariz] est chistosa. /E: No!/
Nunca me ha gustado mucho mi nariz. (A20, fi9, 2, ELA29)
E: His eyelashes are - /A: curly/ Yeah, curly. Theyre not thick, right?, like
some people sometimes have -
A: Yeah. There are people who - Gee, and they look real pretty!, right?
E: Uhm - He doesnt - yeah. He doesnt have them like eaves, as they say,
right?
A: Like a calfs, (a) Thats how Ive heard they call them when they [the
eye-lashes] are (estar) straight, straight - -
A: No, no, no, no. (b) I say that mine [the nose] is (estar) funny. /E: No!/
Ive never liked my nose much.

[26] R: (a) Est alta. Mide seis diez, (b) Est muy alta la muchacha.
C: Cunto mides t?
R: Seis uno.
C: Ella es ms alta que t!
R: No. Yo estoy seis uno y ella cinco; quiero decir cinco diez, no seis.
R: (c) Y est muy grande la, la muchacha; muy durita.
C: Y sus colores as?
R: (d) Est muy clara, como mi pap. (R24,m20,2,ELA50)
R: (a) Shes (estar) tall. Shes six ten. (b) Shes (estar) very tall, that girl.
C: How tall are you?
R: Six one.
C: Shes (ser) taller than you!
R: No. Im six one and shes five; I mean five ten, not six.
R: (c) And shes (estar) very big, this girl; very strong.
C: And her colour?
R: (d) Shes (estar) very fair, like my dad.

Examples [23-4] are from speakers in Group 1. [23a-b] follow the norms of
general Spanish; but [23^-^] are innovative because no individual frame has been
established for the use of estar. Observe that in [24c], in a similar context, the speaker
selects the more conservative copula ser, but in [24a] she uses estar to attribute a
non-circumstantial quality to her mother s nose. Examples [25-6] are from
 Exploring Internal Motivation for Change 103

speakers of Group 2; they illustrate similar cases of estar used innovatively in

discourse which requires a class frame, and with attributes of a non-circumstantial
nature. These examples give clear indication that the linguistic system of speakers
in all groups varies between Stages I, II, and III. Indeed, viewed without the con
straint imposed by an individual or class frame context, the relationship between
a person and the attribute tali or pretty is susceptible to change, so [23c] and
[26a-c\ appear to be at Stage II. However, susceptibility to change is slightly more
difficult to justify (though certainly not impossible) when the attributes big and
funny are predicated of a nose, straight of eyelashes, and fair of someones skin.
It would, of course, be possible to ask whether the very use of estar might not
create an individual frame of reference, with no need for a supporting context. In
contrast, I have adopted here the view that supporting context is required in
order for the speaker to be able to choose estar. To confirm this judgement, which
is crucial in my evaluation of estar as innovative in the various contexts, I prepared
a fill-in-the-gap questionnaire with examples presented in isolation: one could
not assume shared experience of the imagined subject referents, since sentences
were presented outside any continuous discourse context. Thus presented, the
examples conveyed defining descriptive statements.
I gave the questionnaires to three different groups: six graduate students (from
various Latin American countries and from Spain) in the Department of Spanish
and Portuguese at U SC; six college-educated Chileans in Chile; and thirty-one of
the thirty-three Mexican-American speakers included in the copula study (repre
senting the three Groups). The assumption was that, if the context were not
required to create an individual frame of reference, speakers in all groups would
use ser and estar randomly to complete the sentences, without any clear discrete
differences between the various groups. The examples to complete, slightly adapted
from examples produced by the speakers under study during the recordings, con
tained the following sentences:

1. Mi h ija____ buena para pelear.

My daughter____ good at fighting.
2. Cuando los nios____ chiquitos hablaban ms espaol.
When the children_____ small they spoke more Spanish.
3. Las orientales____ chaparritas, verdad? N o _____ altas.
Orientals____ short, right? They _ _ not tall.
4. S, _ _ _ fcil hablar espaol. [Not included in Chile]
Yes, i t ____ easy to speak Spanish.
5. Qu curiositos se ven cuando_____ chiquitos!
How pretty they look when they____ little!
6. Quera ver qu tan alto____ para poder llevar sus tacones altos.
She wanted to see how tall h e ____ to see if she could wear her high heels.
7. Bueno, Juan me poda ganar a correr, porque yo _ _ ms chica que l.
Well, John beat me when we used to run, because I ____ smaller than him.
104 Exploring Internal Motivation for Change
T a ble 4.3. Responses to the Ser (S)/Estar (E) Questionnaire

Example G C G r. i Gr. 2 G r. 3

5 E 5 E S E S E S

1 5 i 5 i ia i 8 1 10 i
2 6 0 6 0 3 8 4 5 2 9
3 6 0 6 0 11 0 9 0 5 6
2f> 6 0 6 0 10 1 9 0 5 6
4 6 0 - - 11 0 9 0 10 i
5 6 0 5 i 3 8 3 6 i 10
6 6 0 4 2 7 4 7 2 2 9
7 6 0 6 0 8 3 6 3 4 7
8 6 0 6 0 11 0 9 0 8 3
9 6 0 6 0 11 0 9 0 11 0
100 6 0 6 0 8 3 6 3 7 4
10 b 6 0 6 0 8 3 8 i 6 5
11 6 0 6 0 10 i 7 2 7 4
12 6 0 6 0 9 2 8 i 6 5
No. o f different
examples completed
with estar i 3 10 9 13

Note: G = 6 graduate students; C = 6 Chileans; Gr. 1 = 11 speakers; Gr. 2 = 9 speakers;

Gr. 3 = 11 speakers.

8. A m me gusta M ara;___ simptica.

I like Mary; she____ nice.
9. El nio de mi sobrina___ inteligente.
M y nieces child____ intelligent.
10. La casa de mi hijo__ grande y tambin____ bonita.
My sons house____ big, and i t _____also pretty.
11. Me gusta tu hermana.___ chaparrita, pero tiene bonito cuerpo.
I like your sister. S h e ____ short, but well-formed.
12. Yo tena un amigo suizo. N o ____ delgado ni gordo; pesaba unas 170 libras.
I had a Swiss friend. H e ___ neither slim nor fat; he weighed about 170 pounds.

The responses given by the Mexican-Americans, especially when compared

with their linguistic behaviour in conversation (note that the adjective inteligente
intelligent appears only with ser in the elicitation task), invite further study (cf.
s. 2. 1.2). For the purposes of this chapter, however, I limit myself to presenting,
in Table 4.3, the figures obtained with respect to the choice between ser (5 ) and
estar (E ).
 Exploring Internal Motivation for Change 105
T able 4.4. Percentage of Innovative Estar (a) over All Tokens of Estar and
(b) over All Tokens of Innovative Estar plus All Tokens of Ser

Conservative use Innovative use Total no.

No. % No. %

Ser 833 100 0 0 833

(a) Estar/estar 431/853 51 422/853 49 853
(b) Estar/estar + ser 422/1255 34

The responses to the fill-in-the-gap task, which visibly divide speakers into
conservative and innovative groups, confirm my analysis of the naturalistic data
collected in the Mexican-American community of East Los Angeles. Estar is being
used innovatively in this community: the message conveyed by estar in certain
contexts is the same as would have been conveyed by ser. Note that, in more
conservative dialects (G and C), ser is used almost categorically. In more advanced
dialects, although we find some variation, a clear trend exists toward obliteration
of the class individual frame opposition, with an increase in the use of estar and a
consequent reduction in the uses of ser by speakers with a longer history of family
bilingualism (Group 3).
Innovative estar is not restricted to Mexican-American Spanish, however. Its
extended use is attested in the materials for the study of Mexico City Spanish
published by the Universidad Nacional Autnoma de Mxico (UNAM 1971;
1976), and it has been examined in some detail in the Spanish spoken in Michoacn
(Mexico) by Gutirrez (1989). Furthermore, I have noticed innovative estar, es
pecially with evaluative adjectives, in spoken Madrid Spanish, and de Jonge (1987)
observes that it is quite frequent in Caracas Spanish in the context of expressions
of age.
For the purpose of identifying the linguistic and social contexts more or less
favourable to the innovative uses of estar, the data from the 33 Mexican-Americans
were analysed quantitatively by using the crosstabs program of SPSS (Nie et al.
1975). A total of 1,686 tokens of ser and estar were coded. As Table 4.4 illustrates,
only estar constructions represent innovations with predicate adjectives, with a
high frequency of 49 per cent of all uses of estar. This table also indicates that estar
has extended to 34 per cent of the cases where more conservative varieties of
Spanish would use ser.
The 422 cases of innovative use of estar were so classified on the well-tested
assumption that the speaker did not intend to convey any meaning difference by
choosing estar rather than ser. Thus the data analysis allowed me to assume that
[26a], Est alta, means the same as Es alta Shes tall (obviously in that specific
context); so [26] was codified as innovative. Had the same example occurred
io6 Exploring Internal Motivation for Change
with ser, it would have been codified as conservative.8 A total of 402 (95 per cent)
of the innovative cases correspond to constructions where estar functions in what
the discourse passage indicated should have been a class frame context. The copula
ser, on the other hand, occurs most frequently in a class frame context (690/833
cases, i.e. 83 per cent), or in contexts not codified with regard to frame of refer
ence.9 Examples [27-8] illustrate individual frame contexts where the selection of
estar is not innovative: '

[27] B: (a) Edward estaba muy gordo, verdad?

A: Como que engord un poco, verdad? (B33,fi8,3,ELA29)
B: (a) Edward was (estar) very fat, wasnt he?
A: He did gain a little weight, didnt he?
[28] (a) [Mi hija] es un poquito gordita, (b) aunque ya no est muy gordita ahora.
Especialmente, (c) ya est un poquito ms delgada. (R n ,f4 2,i,E L A i8 )
(a) [My daughter] is (ser) a little fat, (b) though shes (estar) not very fat right
now. Especially, (c) now shes (estar) a little slimmer.

Note that in [27] the subject referent is known to both speakers, who share an
individual frame of reference for Edward in regard to how fat he has been at
various stages of his life. Conversely, in [28a] the subject referent (the speakers
daughter) is unknown to the interlocutor. Thus, estar is conservative in [27a], while
ser (es is) is so in [28a]. In [28bc\r however, the speaker explicitly creates an
individual frame of reference through lexical means (the adverbs ya presently,
and ahora now) and contrasts her daughters present physique with an earlier
state. Accordingly, the copula estar selected in [28b-c\ is conservative.
Of various internal and external factors included in my quantitative analysis,
only those which show interesting correlations will be discussed here.10 One of these
factors is type o f adjective, included to test the hypothesis of lexical diffusion. In
deed, certain adjectives, amenable to classification into different semantic groups,
appeared to favour the use of innovative estar. The results of the cross-tabulation
of innovative use by type of adjective are presented in Table 4.5. (Three adjectives
of time, all used with ser, are not included in this table.)
Age, size, and colour include adjectives which refer to features of both

8 I cross-checked my codification o f innovative/conservative estar for most o f the 422 cases with
codifications done by 3 o f m y graduate students. There was total agreement. I would like to thank
M anuel Gutirrez (Chilean), and Alicia and Francisco Ocampo (Argentinians) for their help in this
respect.
9 Three types o f example were not codified with respect to frame o f reference, being instead
included as other cases in this factor group: (i) Subjectless constructions (Y a estaba/era oscuro It was
(estar/ser) dark already); (ii) constructions which, given a normal state o f affairs, do not allow ser in
standard general Spanish (Los tomates estn podridos T he tomatoes are rotten); (iii) Constructions with
sentential subjects, incompatible with the establishment o f an individual frame (Hablar bien es importante
T o speak well is important).
10 Quantified but omitted from discussion are the following factors: animacy o f the subject, adjective
grading (comparative, superlative), level o f proficiency in English, level o f schooling in English and in
Spanish.
 Exploring Internal Motivation for Change 107
T able 4.5. Cross-tabulation o f Innovative U se by Type o f Adjective

Type o f adjective Innovative use

No. %

Age 13 6 /17 4 78
Size 10 1/ 19 1 53
Sensory character 8 /17 47
Physical appearance (animate) 4 3 /1 3 8 3i
Description (non-animate) 12/48 25
Evaluation 90/394 23
Miscellany 10/56 18
Moral value 15 / 118 13
Colour 4 /3 8 11
Class 3/78 4
Total no. 4 22 /1,2 52

Note: p = <.00. Statistical Significance is given as a value o f p, where any value above .05
is considered to reflect a non-significant correlation.

animate and inanimate subject referents (e.g. alto tali, grande big, joven young,
mayor old, amarillo yellow); physical appearance and description refer to non-
measurable descriptive features of animate and inanimate referents, respectively
(e.g. curiosito cute, tosco coarse, liso smooth, sucio dirty); evaluation, sen
sory , and miscellany may predicate features of both animates and inanimates.
Evaluative adjectives judge both animate and inanimate entities (e.g. interesante
interesting, bueno good, fdcil easy); moral adjectives judge only animate re
ferents (e.g. honesto honest, respetable respectable, noble noble). Social status or
class is predicated only of animates (e.g. pobre poor, catolico catholic, mexicano
Mexican). Sensory adjectives comprise attributes which may be perceived by
tasting, smelling, or hearing (e.g. dulce sweet, caliente hot, silencioso quiet).
Miscellany includes various types of adjectives: of physical state (e.g. borracho
drunk, sano healthy), mental state (e.g. porfiado stubborn, celoso jealous), and
others not easily classifiable into any of the preceding types (e.g. inocente inno
cent/naive, femenina feminine, alergico allergic, sonador dreamer).
Interestingly, adjectives differ with respect to the innovation: age and size favour
it in over 50 per cent of the tokens; sensory adjectives also appear to favour it but
the number of tokens is rather low (17 in the samples from the 33 speakers). Next
come adjectives of physical appearance, with 31 per cent of innovative cases;
description, evaluation, and miscellany with 25, 23, and 18 per cent of innovation.
Included in the first four groups are adjectives which, independently of the com
municative context, refer to qualities which may be conceptualized as susceptible
io 8 Exploring Internal Motivation for Change

to change, especially so when they are assigned to animate beings. Animacy of the
subject appeared to me at first to be an entirely valid correlation. Indeed, note that
adjectives which describe non-animate entities do not favour the change (25 per
cent) as much as those which describe animate entities (31 per cent). However, size
(53 per cent) and age (78 per cent), highly favourable as well, include a good
number of adjectives which relate to inanimate entities, as illustrated in the ex
cerpts from house descriptions in [29-30],11 so the question of the subjects animac\
was not further pursued.

[29] Una de esas rec -, recmaras es el master bedroom, el ms grande. Y el otri 1

est pequeito. (H22,m2i,2,ELAso)
One of those bed -, bedrooms is the master bedroom, the larger one. And the
other one is (estar) small.
[30] Tenamos otro cabanete [sic] all arriba - pero estaba muy largo, y no caba la
hielera, y no estaba, no estaba ancho. (M47,f33,3,ELA52)
We had another cupboard up there - but it was (estar) too long, and the re
frigerator wouldnt fit, and it wasnt (estar), it wasnt (estar) wide.

Evaluative adjectives, including examples with such attributes as barato cheap,

caro expensive, dificil difficult, occur very frequently when judging entities found
in the immediate context and in expressions of subjective reactions in which estar
is the accepted copula (e.g. Qu dificil est esto! How difficult this is (estar)V, Est
buena esta sopa This soup is (estar) good). Absolute frequency of occurrence of
estar with certain individual attributes may be promoting its extension to less
frequent class frame contexts in which the same attributes are used. After all,
frequency, as Anttila (1972: 187) points out, is not a mere mechanistic concept, as
it has repercussions in association formation and memory.
The correlation between the remaining four groups of adjectives and innovative
estar is, interestingly, relatively much weaker. Clearly, the theory of lexical diffu
sion of phonological change (Wang 1969) may be extended to syntactic-semantic
changes, as the process of diffusion of estar within a speakers vocabulary appears
selectively to affect certain lexical items or groups of lexical items which together
constitute a semantic field (size, physical appearance, age, etc.).
Table 4.6 presents some of the quantitative results by group, by speaker, and by
type of adjective. Under this last factor I have listed the three most favourable
types that occurred more than ten times in the data from all the speakers, age, size
and physical appearance, and the three least favourable types to the innovation,
colour, moral value, and class. Observe that the totals for each group do not
correspond to the sum of the cases for the speakers. This is because Table 4.6 does
not list those speakers who had fewer than ten tokens of innovative estar in their
speech sample. Percentages are given only if there are more than five tokens in a
cell. Otherwise, raw numbers are given in parentheses.

11 House descriptions are also subject to the class frame o f reference pragmatic constraint.
 tokens. Phys = physical appearance;
\
COS<X \
H S
N NN V
0 0 0 0 0 0 0 ^ 0 0 0 00 00 bs. o o o n o so H

o
V
' ^ w O O O O O ' O O O t ^ m fr- OO O s O' o o o o oo o o II

yrs. old; N = total no. of tokens; % = % of innovative

T
f-

\
C* S MS N NM N
N V
OO s w O O O O O
l.\i,n

O O M O O O O O O O
U
I scs of
of In n o v im e

V
il
M
Ph OO^ M M
O n O CO O
Percentage

f N S S S N
.3 sO m sO t- O s >rs s O' f O Pi V r j ^ t ' O O ' O O o
co SO H M HH > ^ W
4.6.

yrs. old; O = 30-65

T able

V
be II
<
female; M = male; Y = 15-29

Vr > N O r> 00 00 OsTi-ThW m K. SO

sO to N r <! O' ifi t ^T
) N
*0 f'o

11 S
^ >-i >< s s foe 2S o 2 >< * fa 11
s S Cl, fe (Jh Z& *
& 5f ^ ^
$s ..........
H N m so O O O' N
*- ^i-1 U. ir> ^ ^ ^ J3 ^ I .5 "a<a
<T)
O W Cfl fa s s < w u O 02! ^ ^ Os U fflsi Q
110 Exploring Internal Motivation for Change
Assuming that the results of Group i reflect the path followed by the extension
of estar by adjective type, and noting also that the innovation is favoured by those
speakers who have acquired English from birth (see below, Table 4.8) and whose
speech diverges the most (impressionistically and by speakers own evaluation)
from Group x Spanish, it appears safe to suggest that those speakers who present
some innovative uses with colour, moral, or class adjectives represent more ad
vanced stages of the change, and seem more affected by the contact situation. A
serious problem, however, is raised by the scarcity of the data in a number of cells.
I feel that conclusions about Group 1 are safer, though, because despite individual
differences and empty cells for colour, moral, and class adjectives, what stands out
is that estar is never used innovatively with moral or with class adjectives, and only
once is employed in this way with a colour adjective, by F5. Interestingly, F s s
social history is one of intensive interaction with Group 2 type youngsters in El
Monte, the barrio where he lives (cf. Milroy and Milroy 1985 on social network
theory in relation to language variation and change).
A more serious problem is raised by the empty cells in Groups 2 and 3. For
instance, lack of data in some crucial cells makes the behaviour of R 17, S38, M47,
H48, and R50 similar to that of speakers in Group 1. Thus, we may compare only
six of the speakers in these two groups with respect to their distance from those in
Group 1, and on the basis of just a few tokens for each (bold numerals in Table
4.6). As compared to Group 1, then, the number of types of adjective affected by
the extension of estar in Groups 2 and 3 divide the speakers as follows:

One further type of adjective affected (colour or moral value): V 21, H22, B33,
and D36.
Two more types affected (moral value and colour or class): A20, and R24.

Note that A29 (Group 2) and R42 (Group 3) are not included in Table 4.6
because there are fewer than ten innovative tokens of estar in their speech samples.
One of six innovative tokens used by A 29 corresponds to a moral value adjective;
three of four innovative tokens used by R42 correspond to moral and class type
adjectives. This is further indication of the extension of estar to a larger number of
adjective types in these two groups.

In sum, group totals indicate more regularly the increase in innovation from
Group 1 to Groups 2 and 3. But, as with many other studies of syntactic variables,
we are faced with difficulties as soon as we set individuals apart, because of
deviations from the norm of the group but mostly because of lack of sufficient
spontaneously produced data. Although the interpretation of results for these
speakers must be tentative, the addition of new adjective-type contexts for the use
of estar seems reasonable proof of acceleration of this change in Mexican-American
Spanish (see definition of acceleration in Chapter 1).
Let us now examine the results of Table 4.7, which correlates innovative use
with the semantic transparency of the choice between ser and estar.
 Exploring Internal Motivation for Change 111
T able 4.7. Cross-tabulation o f Innovative Use by the Semantic Transparency o f the
Choice between Ser and Estar

Semantic transparency Innovative use

Total Group 1 Group 2 Group 3

No. % No. % No. % No. %

Different modality 367/846 43 126/362 35 147/262 56 94/222 42

.Apparent synonymy 2 8 / 17 1 16 16/79 20 6/50 12 6/42 14
Choice not allowed in
standard language 19 /17 4 11 11/ 10 2 11 2/24 8 6/48 12
Clear difference 8/64 13 3/32 9 4 /2 1 19 1/ 11 9
Significance p = <.00 p = <-00 p = <.00 p = <.00

These results also support the hypothesis of gradual diffusion. In this case, the
process appears to be controlled by what I have called the semantic transparency
of the choice between ser and estar to introduce an adjective in a given sentence.
The motivation for including this factor is provided by the observation, made in
historical linguistics, that complex systems are more likely to change than simple
ones; i.e. when the rules underlying an aspect of the grammar are opaque, we have
a potential source for re-interpretation.
As already explained, the coding of estar as conservative or innovative for every
token takes the sentence and the discourse/pragmatic context into account. The
decision as to whether the choice between ser and estar is transparent, however,
depends exclusively on an evaluation of how transparently or unquestionably one
or more of the parameters listed in Table 4.1 may account for the relationship
between subject referent and attribute in terms of meaning at the sentence level. Thus,
[23--/| are coded innovative because they violate the class frame pragmatic
norm, and they are also coded different modality because the two alternatives
(presented in [31] with estar and in [31^] with ser) may be accounted for by all the
parameters in Table 4.1. This context is the most favourable to the innovation in
every group.
[31] (a) [Ella] esta bonita. [She] is (estar) pretty.
(b) [Ella] es bonita.
The contrast between ser and estar is transparent (clear difference in Table 4.7)
with a group of adjectives which have different meanings associated with the
choice of copula, as shown in the translation of [32-3]:
[32] (a) Esta radio es buena. This radio is (ser) of good quality.
(b) Esta radio esta buena. This radio is (estar) functioning well.
112 Exploring Internal Motivation for Change

[33] (a) Pedro es callado/vivo. Peter is (ser) shy/smart.

(b) Pedro esta callado/vivo. Peter is (estar) silent/alive.
This context, as expected, is not favourable to diffusion (13 per cent tol.il
average). Interestingly, the three innovative tokens in Group x are produced b>
C 13, who is married to M26, who is in turn responsible for three of the four
innovative tokens in Group 2 (see Milroy and Milroy 1985 on social network
theory in relation to language variation and change). The remaining two tokens
occur in the sample from R24 (Group 2) and R50 (Group 3).
In the same vein, the innovation is disfavoured (11 per cent in Group 1 and total
average) by adjectives which in the standard language categorically require one or
the other copula, as in [34].
[34] (a) Eso es preferible. That is (ser) preferable.
(b) *Eso esta preferible. *That is (estar) preferable.
Contexts in which the ser/estar opposition is not associated with a semantic dif
ference in the adjective, and in which the choice of one or the other copula
represents what I have termed different modality, i.e. different and more or less
subtle ways of conceptualizing the relationship between the attribute and the
subject referent, correlate favourably with the innovative use of estar. Indeed, in
such contexts, illustrated by [35], innovative estar is used more frequently than ser
in Group 2 (56 per cent of the cases), and it is also quite frequent in the other two
groups. Also favourable (16 per cent total in the three groups) are a number of
cases quantified separately and included in the group labelled apparent synonymy .
In these examples, the opposition ser/estar appears to have almost no semantic
load: while it may be possible to propose a fairly undisputed explanation for the
alternatives in [35], it is much more difficult to do so for those in [36]:
[35] El mole poblano {(a) esta/(b) es} bueno ahi.
Puebla mole [a Mexican dish] is good there.
[36] Si el hombre {(a) esta/(b) es} soltero, puede hacer lo que quiera.
If the man is unmarried he can do whatever he pleases.
Alternative (a), with estar, is considered to be innovative in a discourse context
which does not support the establishment of susceptibility to change and/or an
individual frame of reference. Furthermore, while [350-^] represent different
modality, [36a-b] qualify as apparent synonymy . Indeed, in conservative varie
ties of Spanish the choice of ser or estar to link the attribute good with Puebla mole
is explained by the fact that the alternative with ser is defining, independent of
immediate experience, and presented within a class frame; but the alternative with
estar is dependent on immediate experience, and implies susceptibility to change.
Note that to qualify [35a] with the time adverb ahora now, thus implying change
with respect to another time, would make the use of estar conservative. In [36], the
choice of ser or estar to link the attribute unmarried to man does not seem clearly
linked to any of the parameters discussed. Example [36] is coded innovative in
 Exploring Internal Motivation for Change 113

order to be methodologically consistent with the discourse/pragmatic requirement

of class frame of reference; at the sentence level, however, it does not appear
possible to maintain that alternatives (a) and (b) unquestionably contrast with re
spect to any other explanatory parameter, namely circumstantiality, dependency
on concrete experience, or susceptibility to change . Note, for instance, that
neither alternative would become more or less acceptable if the examples were
qualified with ahora.
The preceding discussion indicates that we should expect apparent synonymy to
be more favourable to the extension of estar than different modality. This expec
tation is not fulfilled, as the percentages in Table 4.7 show. None the less, it is of
even more interest to see that diffusion of estar is least favoured both when the
opposition ser/estar is associated with clear semantic differences and when the
opposition is not allowed in the standard language, i.e. when use of estar would result
in blatant unacceptability.
Let us now examine the results of the cross-tabulation between innovative use
and both (a) level of proficiency in Spanish, and (b) age of acquisition of English.
Levels of proficiency in English and Spanish were established on the basis of the
speakers self-evaluation and the authors global and impressionistic assessment of
the speakers competence in each language. The speakers were evaluated relative
to one another; this assessment took into account at least the following elements:
pauses, hesitation, self-corrections, switches into the other language, questions
requiring information about how to say something, phonetic transfer, lexical trans
fer, and violations of rules of tense-mood-aspect, person, number, and gender
agreement. The three levels of Spanish proficiency do not necessarily correspond
to membership in Groups 1, 2, or 3, but all those in Group 1 are included in the
native proficiency level. Likewise, all speakers in Group 3 started acquiring
English from birth, and all those in Group 1 started acquiring it at age twelve or
older; although all speakers in Group 2 are born in the USA, some of them started
acquiring English between the ages of five and seven. Table 4.8 presents the
results.
The results displayed in Table 4.8 show that the level of oral proficiency in
Spanish does not establish differences with respect to the innovative uses of estar.
By contrast, it is arresting to note that, if the speaker acquires English before the
age of six (no speakers in the sample started acquiring English between the ages of
six and eleven), his speech will show a noticeably higher percentage of innovative
uses as compared to one who has acquired this language after the age of eleven.
This correlation may be interpreted in at least three ways: (a) early contact with a
linguistic system in which a single form, be, may correspond to two forms, ser and
estar, favours the evolution toward the use of a single form in contexts which allow
both (albeit contrastively); (b) early contact between two linguistic systems speeds
up processes of change in the system of the less used language; (c) lack of formal
education in a language speeds up processes of change. M y study does not dis
criminate among these three hypotheses: all three may be correct.
ii4 Exploring Internal Motivation for Change
T a ble 4.8. Innovative Estar by Spanish Proficiency and by Age o f English Acquisition

No. %

Spanish proficiency
Native 2 4 1/ 7 17 34 %
Good 156/463 * 34 %
Lower 2 5 /7 5 33%
Significance P = < -9 9
English acquisition
From birth 7 5 / 116 37%
3 - 5 years 164/265 40%
12-20 7 9 /1 9 5 29%
2 1+ 26/47 25%
Significance p = <.00

The hypothesis in (a), which acknowledges the possibility of influence from the
dominant language, could be explored by replicating this study in a situation of
language contact between two languages with dual systems of copulas. The results
of Table 4.8 support the possibilities in (b)-(c). Note, however, that in my subject
sample, acquisition of English before age six necessarily implies lack of formal
instruction in Spanish as a native, official language. I have not answered, therefore,
the question of what effect language contact per se has on the diffusion of innova
tive estar, i.e. at this point, one could argue that the change is accelerated by lack
of formal instruction, regardless of the situation of contact. To examine this
further, it would be necessary to study a population of illiterate Spanish mono
lingual families. This rather difficult task has not been attempted. A comparison
with the extension o f estar in Mexico, however, allows me to propose that one of
the consequences of language contact has been the acceleration of this change.

4.2.3. Acceleration o f Change

A comparison of the status of estar in Mexican-American Spanish with the results
of a similar study conducted by Gutirrez (1989) in Morelia (Michoacn, Mexico )
indicates that the diffusion of estar is at a more advanced stage in the USA, and
that other factors, in addition to lack of formal education, must be at work in
change acceleration. Gutirrez collected conversational Spanish data in Michoacn,
and attested the same type of diffusion of estar discussed in this chapter. Examples
[378] are taken from his doctoral dissertation:

[37] Y ahora vivimos all en Prados Verdes en las casas de Infonavit, estn
chiquitas, pero estn bonitas.
 Exploring Internal Motivation for Change
. . . and now we live in Prados Verdes in the Infonavit houses, they are sm all
(estar), but they are pretty (estar).

[38] . dicen que all [en USA] crecen mucho y aqu estamos muy enanos.
. . . they say that over there [in the USA] they grow a lot and here we are (estar)
extremely short.

Gutirrezs (1989: 116) results show that, as in Los Angeles, the adjective types
most favourable to the innovation are age (43 per cent), size (34 per cent), physical
description (33 per cent), and evaluation (12 per cent); no innovative uses occur
with moral value, colour, or class type adjectives (cf. Table 4.5 above).
As regards the social variables that may correlate positively with innovative
estar, Gutirrez observes that women appear more innovative than men; he also
notes that higher education correlates with lower percentages of innovation (6 per
cent and 21 per cent for university education and elementary education, respec
tively). The low number o f total cases does not permit h im to establish correlations
with the social variables while at the same time controlling for type of adjective;
necessarily, then, his results must be interpreted with caution, since it is in prin
ciple possible that education skewing may be a consequence of the types of adjec
tive used in each education group.
Acceleration of change seems to me to be supported by a comparison of both the
overall frequencies and the increase in linguistic contexts for the diffusion of estar
in Los Angeles with those of Morelia. Whether this speeding up is due specifically
to the situation of language contact remains to be explored further; the existing
evidence suggests multiple causation, i.e. language contact and lack of formal
instruction both contribute to the propagation of a change in the secondary lan
guage. Language contact manifests itself not exclusively as transfer from the con
tact language, but in such phenomena as simplification of alternatives, selective
acquisition, or freezing of the competence reached at a certain age. All these
strategies converge in the direction of alleviating the load of having to remember
and use two different linguistic systems.
Table 4.9 presents a few general comparative percentages.
Compared to Morelia speakers, percentages of innovation are higher even in
Group 1. For the three groups in Los Angeles, the overall percentage of innovative
estar is 34 per cent; it is only 16 per cent in Morelia. Furthermore, in each of the
three Groups studied in Los Angeles, the younger speakers were shown to have a
higher percentage of innovative uses of estar. Table 4.9 shows that this generational
difference does not seem to exist in Morelia: innovative uses of estar remain the
same across age groups (7 per cent/7 per cent; 26 per cent/27 per cent). Percent
ages increase along the education factor (from 7 per cent to 26-7 per cent), al
though not regularly; they are lower in the elementary education group than in
the secondary education group.
Let us observe, furthermore, that the hypothesis regarding acceleration of inter
nal changes also receives support from the results shown in Table 4.10. For the
n6 Exploring Internal Motivation for Change
Table 4 .9. Proportion of Innovative Uses of Estar in Los Angeles and Morelia

Los Angeles Morelia

------------------------------------------- (Michoacn)
Group 1 Group 2 Group 3 Total

No. % No. % No. % No. %

F 282 32 258 42 113 48 96 19

M 293 23 99 53 210 25 43 12

Age Education

Tertiary Secondary Elementary

No. % No. % No. %

3 0 -5 0 9 7 12 26 3 19
13 -2 9 15 7 70 27

T otal 4 22 /1,2 55 (34% ) 139/886 (16% )

Note: T h e results for Morelia are also broken down by age and by level o f education.

T able 4.10. Cross-tabulation o f Innovative Use o f Estar by Group and by Age

(Los Angeles)

15 -2 9 years 30 -6 5 years

No. % No. %

Group 1 7 speakers 6 speakers

82/295 28 74/280 26
Group 2 5 speakers 4 (3) speakers
140/234 60 19 / 12 3 15
(18/80) (23)
Group 3 5 speakers 6 speakers
3 9 /9 9 39 68/224 30

Significance p = <.00 p - <.00

 Exploring Internal Motivation for Change 117

older Group 2 speakers I have given results including and excluding L28 (latter
result in parentheses), who is quite exceptional with respect to estar; she has a very
low 2 per cent of innovation (1/43).
In each group, the younger speakers show a higher percentage of innovative
uses. Observe, however, that in Group 1 the difference between the younger and
the older speakers is only of 2 percentage points, similar to the situation reported
above for Morelia. That is, within a family of first-generation immigrant speakers,
we may assume that the speech of the older and the younger members will not
differ much in the use of estar.
Conversely, in a mixed family, the younger second- or third-group members
will show marked differences in regard to this feature when compared with older
first- or second-group members of this family. Note that the gap between the older
and the younger speakers in Group 2 widens to 45 percentage points. The differ
ence is not as steep within Group 3. Indeed, in terms of overall percentages of use
of a given linguistic form, my research demonstrates that in some cases Group 3
speakers may not be very distant from Group 1 (cf. quantification of preverbal
clitics, Table 4.17). Overall percentages in Group 3 must be interpreted with
caution, however. I have noted that speakers who are at the lower levels of the
Spanish proficiency continuum use a smaller lexicon, especially so in regard to
adjectives. This may account for their producing fewer innovative forms, while at
the same time extending the diffusion to new adjective-type contexts, as shown
above in Table 4.6.
Some of the tables show performance in terms of group behaviour; but there is
actually a certain range of variation within each group (cf. Table 4.6). Among those
in Group 1, place of origin in Mexico appeared to be a determining factor: impres
sionistically, innovative estar seemed more frequent among speakers from the
northernmost regions of Mexico such as Tijuana, and Ciudad Jurez. To examine
the possible effect of diatopical differences on those speakers whose parents have
come from different places, I have compared overall percentages in three families,
as shown in Table 4 .11. Tijuana borders California, Ciudad Jurez borders Texas;
Zacatecas, Jalisco, and Michoacn are neighbour states located further south (see
Fig. 1.2; Morelia is in the north-east of Michoacn).
It is clear that, despite differences in parents places of origin, those in the
younger age group have a uniformly higher percentage of innovation, character
ized by an increase of 7-42 percentage points from the older to the younger
generation within each family. The lowest difference corresponds to Family 1.
Interestingly, the mother (E12), who evidences the highest percentage of innova
tive estar in Group 1, came to the USA when she was about eighteen years old
from Tijuana, a border city characterized by intensive interaction with U S resid
ents. The same degree of interaction, however, characterizes Ciudad Jurez, yet R i 1
(the mother) evidences 33 per cent of innovative uses, only 7 percentage points higher
than Morelia speakers with secondary education, and her husband (J14) shows a
slightly lower percentage than both Jalisco/Zacatecas and Morelia speakers.
n8 Exploring Internal Motivation for Change
T able 4 . 1 i . Innovative Estar in Three Families

Parents origin No. % Offspring No. %

GROUP 1
Family 1:
E 12 Tijuana 34 44 II22 (son) 4i 5'
P iS Jalisco/Zacatecas 45 27 R24 (nephew) 39 67

Family 2:
R 11 C. Jurez 27 33 R 17 (daughter) 26 42
J14 C. Jurez 52 25 V 2 1 (daughter) 27 67

g r o u p s 2-3

Family 3*: (mixed)

(L /4 3,2, Tijuana) S38 (daughter) 21 52
R50 Los Angeles 64 28 R42 (son) 24 17

T he mothers samples o f speech (L /4 3) have not been analysed.

There is one exception in Family 3, R42. As I have pointed out in the preceding
chapters, R42 represents an interesting case. He did not develop productive com
petence in Spanish as a child, but was exposed to this language from birth both at
home and in the community. At the time of the data collection, he had completed
three semesters of Spanish in high school. He was able to converse with me with
quite a good degree of fluency and showed a very high level of oral comprehension.
It seems to me that, had he not had at least a fairly well-developed receptive
proficiency in Spanish, he would not have reached the degree of productive pro
ficiency he now demonstrates. However, his Spanish does appear to show the
effect of schooling, and this sets him apart from other youngsters who have
acquired this language in natural settings. Of these youngsters, four of the five
included in Table 4 .11 use innovative estar in more than 50 per cent of all cases.
The uniform speech behaviour of these younger speakers suggests that the change
affecting estar has become a stable feature of Mexican-American Spanish in Los
Angeles.

4.2.4. Conclusions
Numerous linguists (Meillet 1926, Bloomfield 1933, Tauli 1956, Benveniste 1968,
among many) have noted that semantic bleaching (i.e. the loss of lexical meaning)
goes hand in hand with an emphasis on grammatical function (or grammaticalization).
This phenomenon is characteristic of the development of modal and auxiliary
verbs, of determiners, prepositions, conjunctions, and of a number of bound
 Exploring Internal Motivation for Change 119

i-rammatical morphemes in natural languages. Spanish is no exception to this form

of growth. The development of estar from its original meaning of to stand to its
present rank as an auxiliary verb in progressive constructions, and as a copula
with predicate locatives and predicate adjectives, is a further illustration of this
universal tendency. In the specific context of predicate adjectives, as I have shown
in this chapter, the continuous extension of estar at the expense of ser is intensified
and diffused more rapidly in a stable and prolonged situation of bilingualism. This
observation offers a response to one of the questions which motivated the study
in the first place: what is the effect of language contact on language change? The
innovation examined here represents part of an evolutionary trend in Spanish and
other Romance languages. Given such a situation, the result of language contact
will be acceleration of the change.
It may be argued that at least one further condition is necessary for this rapid
diffusion: reduced access or lack of access to formal varieties of the language, as
well as to those institutions which maintain conservative and prescriptive language
norms, conditions that are present in Los Angeles Spanish. To examine the rela
tive impact of these and other factors which often coincide with intensive societal
bilingualism (e.g. reduction in domains of use, diminished concern for the acqui
sition and/or maintenance of a standard variety of the secondary language, lack of
formal instruction in this language), one could investigate internally motivated
changes in the superordinate language, English in our case. It seems to me that if
the hastening of changes in the direction of regularization of paradigms, and the
loss of one or more competing structures with closely related meanings is in any
way motivated by the bilinguals need to lessen his cognitive load when having to
communicate rather frequently in two or more languages, then we should expect
to find that internally motivated changes which involve the generalization of a form
X to a larger number of contexts are also accelerated in a situation of extended
language contact in the system of the superordinate language. Future research will
need to examine superordinate language variables to contribute to the examination
of this hypothesis.
With respect to specific influence from English on the fate of the competing
forms ser and estar, it appears reasonable to suggest that the extension of estar in
progressives (cf. Ch. 2) and its frequent association with be in these constructions
may be a factor favourable to the rapid diffusion of estar in the context of predicate
adjectives, where Spanish anyway evidences a slow process of change independ
ently from any language-specific influence. For instance, de Jonge (1987) shows
that estar is replacing ser in the context of adjectives of age (e.g. grande old/older,
pequeo small/young) in the standard varieties spoken in Caracas (Venezuela),
and in Mexico City.
There is nevertheless no evidence in Los Angeles that the dual system of
copulas is heading toward imminent collapse, despite contact with a linguistic
system in which ser and estar correspond to a single form, be, and despite the
aggressive behaviour of estar. There is a noticeable movement toward steady
 120 Exploring Internal Motivation for Change

functional specialization, but the continuous renovation of Spanish through the

arrival of new waves of immigrants secures the necessary input to keep I.os
Angeles Spanish from drifting away entirely from other Spanish dialects. Thu
possibility of continuous contact with complete and monolingual varieties of Span
ish distinguishes the Mexican-American community from other communities which
have undergone or are undergoing language shift and possible eventual language
death (such as those, among others, investigated^by Dorian 1981, Gal 1979, .ind
Hill 1989). Language shift occurs at the individual and family level in Los Angeles,
but it is far from obvious at the community level.
Clearly, although the ser/estar opposition with attributes is lost to a large extent
among speakers in Groups 2 and 3, the system described here does retain some
degree of meaningful opposition. Furthermore, in other syntactic contexts (nom
inal and prepositional predicates) the distribution of ser and estar persists as in the
standard norms. I have demonstrated also that it seems correct to assume the
occasional existence of semantically meaningless variation; that is, ser and esPtr
alternate in certain contexts without conveying meaning differences. Indeed, to
account for the possibility of change or for change in progress, variation theory
must assume that at a given stage the variants involved-phonological, syntactic,
or lexical are referentially and pragmatically synonymous, either in all or in .1
subset of possible contexts. The present study further supports the hypothesis that
change-related variation observed in apparent time (i.e. across generations) reflects
stages of diffusion in real time (cf. Labov 1972a: ch. 7).
Within the framework of variation theory, one further question raised here
concerns the path followed by innovation in the linguistic and social systems. One
important and stimulating outcome of this study, I hope, has been showing, both
qualitatively and quantitatively, the step-by-step movement of the process.12 Within
the linguistic system, the change is not abrupt; that is, it does not occur in all
possible contexts at once. Rather, the diffusion is gradual and context-selective; it
appears to depend on factors related to the semantic transparency of the choice
between ser and estar, as well as to changes in the manner of conceptualizing the
relationship between the subject referent and the attribute ascribed to it. These
changes, which become apparent in the skewed diffusion according to type of
adjective, trigger the loss of certain constraints imposed by such parameters as
susceptibility to change, and the class/individual frame. From this state of affairs,
one is free to draw the conclusion that the theory of lexical diffusion made explicit
by Wang (1969) may be extended to, and indeed finds support in, syntactic-
semantic changes.

12 Both Fernandezs observations and Vano-Cerdas revealing synchronic and diachronic study o f ser
and estar with adjectives imply the idea o f lexical diffusion. Vano-Cerda explains that his classification
o f adjectives into three groups (of extrinsic, intrinsic, and active sense) is motivated both by his
synchronic study o f the uses o f ser and esiar , and by the behaviour . . . o f the two verbs with
respect to the different classes o f adjectives throughout the historical evolution o f the language (p. 314 ,
m y translation). Neither linguist (nor anyone else as far as I know) has examined the process quanti
tatively or established any parallel with the lexical diffusion o f phonetic changes.
 Exploring Internal Motivation for Change 121

Finally, the external factors which accelerate innovation have been shown to be
related to the situation of language contact and shift in which Spanish happens to
be the less favoured language. The diffusion o f estar is most advanced among the
younger speakers in Group 2. This is interesting because this is the US-born
group which, if any, is likely to pass on Spanish to their descendants. Thus, it is
possible that estar may continue its expansion to all types of predicate adjective.

4.3. Verbal Clitic Pronouns

4 .3 .1. Verbal Clitics A long the Continuum

This section discusses verbal clitic pronouns (Cls),13 in small capitals in [39-40], an
area of the grammar of Spanish which appears to lend itself to the investigation of
the processes characteristic of a situation of language contact: simplification, transfer,
and overgeneralization.
[39] l a v i ayer
h er s a w - is g y e ste rd a y
I saw h er y e ste rd a y

[40] s e cort
H IM SELF CU t- 3 Sg
He cut himself
For this specific study I have analysed data from twenty of the speakers, who
represent the various degrees of Spanish and English proficiency characteristic of
the total sample under investigation. Fifteen of these speakers belong to four
different families; the remaining five are not related to anybody else in the sample.
The analysis allows me to conclude that this particular aspect of the morphosyntax
of Spanish is fairly impermeable to interlinguistic influence.
Even though it is not entirely clear to me how comparable Spanish clitics may
be to oblique pronouns in English, I assume cross-linguistic equivalence when, in
a given English sentence, the oblique pronoun translates the Spanish verbal clitic.14
Accordingly, lo is considered to be equivalent to him, unstressed, in [41-42]:
[ 4 1 ] l o conoc en la fiesta de Pepe
h immet-xsg in the party of Pepe
I met h im at Pepes party

13 T h is section is based on work started in collaboration with Manuel Gutirrez while he was a
graduate student at U S C (cf. Silva-Corvaln and Gutirrez, 1995). I would like to thank M anuel for his
contribution.
14 Note, however, that stressed oblique pronouns in English must be translated with non-verbal
stressed pronouns in Spanish in addition to the Cl, as in the following example. (See Garcia et at. (1987)
for an illuminating analysis o f the question o f inter-translatability and cross-linguistic equivalence.) I
saw him at the store (not her); Lo vi a l en la tienda (no A e lla ).
 122 Exploring Internal Motivation for Change
[42] Q uera verLO temprano
wanted-1 s g see h im early
I wanted to see him early
Overgeneralization, simplification, and transfer allow us to advance certain hy
potheses with respect to the behaviour of Cls along the continuum.
Simplification accounts for the extension of one of two or more structures which
have the same or similar meaning. This justifies expecting a trend towards the
categorical occurrence of one of the two possible alternative positions for Cls in
constructions with verbal periphrases, either preverbal ([43a]) or postverbal ([43^]).
Overgeneralization, on the other hand, might result in the extension of so-called
lexical reflexive or obligatory reflexive constructions, as in [44/]. This over
generalization of reflexive-type constructions is in fact a trend which appears to
characterize the diachronic development of many verbs in Spanish.15
[43] (a) l o puedo hacer maana (b) Puedo kacerto maana
it can-isg do tomorrow can-isg do i t tomorrow
I can do i t tomorrow
[44] (a) M i hermano no creci hasta cuando tena cuarenta! (Group 1:
standard non-reflexive form)
(b ) M i hermano no s e creci hasta cuando tena cuarenta! (Group 3: non-
standard reflexive form)
M y brother didnt ( s e ) grow up until he got to be forty!
Furthermore, if transfer from English affected this area of the grammar of
Spanish, the bilinguals propensity to make both languages structurally more
similar should result in omission of the Cl when the corresponding English
construction does not require an oblique pronoun, as in [45] (examples of this
type are further discussed in Chapter 5, which deals with the issue of syntactic
transfer).

[45] - y me dieron en la cara, y 0 quebraron mi, mi jaw (Gr. 3)

and me hit 3pl in the face, and 0 broke~3pl my, my jaw
. . . y me dieron en la cara, y m e quebraron la, la m andbula (Gr. 1)
t o m e broke~3pI the, the jaw

. .. and they hit me in the face, and broke my, my jaw

15 M y evidence for this statement, anecdotal and impressionistic, is based on the fact that native
speakers o f Latin American Spanish find it very difficult to explain the semantic difference between
certain pairs o f non-reflexive/reflexive verbs in a number o f examples (e.g. morir!morirse to die, ir !
irse to go), and use the reflexive form in cases which according to prescriptive grammars should be
constructed with the non-reflexive one (e.g. Ju a n se muri en un accidente Joh n se died in an accident,
Ju a n se fu e en auto a la escuela hoy John se went to school by car today). Furthermore, m y father used
to tell me that it was not necessary to use such verbs as demorar to take long , or comunicar to
communicate in the reflexive form, as in Ju a n se demor mucho John se took too long, ox Ju a n se comunic
con Marta John se communicated with M ary; but I, like most o f my Chilean peers, use them exclu
sively in the reflexive form.
 Exploring Internal Motivation for Change 123
Transfer should also result in the preference for postverbal placement of the Cl
([43 ]). This type of transfer, which at first evidences itself not in ungrammaticality
but in an increased frequency of use of parallel structures, is proposed by Klein-
Andreu (1986a: 7) as the most likely to occur, as well as the most likely to become
part of the community language norms. Furthermore, though previous research
has shown that, as compared to free morphemes, bound morphology is more
resistant to change (cf. Meisel 1983^, Pfaff and Portz 1979, Poplacks 1978 free
morpheme constraint, Weinreich 1974),16 the presence or absence of inflectional
markings in one language is also cited as one of the possible features to be affected
by the presence or absence of corresponding inflections in the contact language (cf.
Meisel, Weinreich). Thus, as inflections, Cls may be a plausible site for transfer
from English to affect at least both the position and the actual occurrence o f Cls.

4 .3 .1 .1 . Omission o f Cls

For the examination of Cl omission, in every sentence where a Cl occurred or

should have occurred the Cl was classified according to its functional relation to
the verb. Nine different groups or contexts were established, listed in [46-54]:
[46] Accusative: l o va a grabar Shes going to record h im

[47] Dative: l e mostr el libro a R o d I showed ( h im ) the book to Rod

[48] O bligatory reflexive: s e sali del equipo He (se) left the team
[49] T ru e reflexive: s e cort He cut h im s e lf

[50] Indirect reflexive: s e cort el dedo He (se ) cut his finger

[51] Reflexive with inanimate subject: s e quem e l p an The bread ( s e ) got burnt
[52] Reciprocal: n o s veamos a menudo We (n o s) often saw each other
[53] Im personal se : a h s e come bien One (se) eats well there
[54] Affective: y o t e corro 3 millas todos los das I run (yo u ) 3 m ile s every day.
Examples of the type of [54] are characteristically used in a colloquial or informal
style in Spanish.
Observe that in these nine contexts English may or may not be equivalent to
Spanish with respect to requiring a pronoun. The correspondences seem to be as
follows:
(a) Equivalent in English and Spanish: [46], [47], [49], and possibly also [52]
and [54],
(b) Non-equivalent contexts: (i) A reflexive verb form in Spanish corresponds to
a non-reflexive intransitive verb in English: [48], [51]. (ii) A so-called indirect
reflexive construction, in which the Cl in Spanish has the same referent as the

16 In agreement with a number o f researchers (Givn 1976, M eisel 1983^), I view Cls as verbal in
flections, i.e. as bound morphology (cf. Silva-Corvaln 1981).
124 Exploring Internal Motivation for Change
T able 4 .12 . Clitics Omitted in Required Contexts

Group i Group 2 Group 3

S2, 24 0/308 R 17 , 21 2/277 0.7% S38, fig 1/ 18 4 0.5%

A9, f2 0 / 3 17 V 2 1, fi8 6/148 4% D39, fe8 3 /2 9 5 1%
R 11, 42 0/296 H 22, m 21 0/320 A46, 31 0/77
E 12 , f44 0/296 R24, m20 0/207 R42, 11115 24/45 53%
OO
vO

M 8, m i7 1% A29, m6o 2/263 0.8% J4 3 , 11117 7 /2 5 7 2.7%

J14 , 1T141 0/247 H48, m39 0 / 13 1
P i 5, m 54 0/298 D 36, m45 2 3/255 9%
R50, 11146 3/292 1%
T otals 3/2063 0 .1% 10 /12 2 5 0.8% 6 1/15 9 7 3.8%

possessor in the direct object of a corresponding non-reflexive construction in

English: [50]. (iii) An impersonal use of se, marking the absence of any subject,
which may correspond to a construction with an impersonal subject pronoun (one,
you) in English: [53].
Regarding the possible omission of Cls, it is interesting to note that in a subset
sample of thirteen speakers in Groups 2 and 3, of a total of over 2,822 contexts for
the occurrence of an obligatory Cl (contexts [46-53]), only 71 are missing (2.5 per
cent). Seven speakers in Group 1 were also examined. Only one of these failed to
provide an obligatory Cl, three times in 298 cases. Examples [55-6] illustrate
omissions; Table 4.12 displays the quantitative results by speaker:
[55] . .. tenemos una{ y nosotros 0 llevamos [la moto] [/a] have-ipl one and we 0
take-ipl [the bike]
We have one and we take it (838/19,3,ELA30)
[56] . . . muy amarradas las tenan - - y ellas 0 rebelaron, 0 rebelaron []
. . . very controlled them had~3pl - - and they 0 rebelled, 0 rebelled
. . . they had them very controlled and they rebelled, they rebelled
(D36,m45,3,ELA43)

Observe in Table 4.12 that it is in Group 3 where omission is somewhat more

noticeable, and that R42 has an exceptionally high frequency of omission (53 per
cent). Of the other speakers in Group 3, the second highest percentage of omission
corresponds to D36: 9 per cent of a total of 255 contexts of occurrence. In Group
2, the highest percentage corresponds to V21: 4 per cent.
E xc ep tin g R42, these results show a very low percentage o f om issions. E x
pressed C ls , h ow ever, are not alw ays m arked correctly for gen der, n um ber, a n d /
or case, as illustrated in [5 7 -9 ]; bu t this is a different question, i.e. it concerns
processes o f loss o f gender, num ber, and case rather than o f w hatever ty p e o f loss
the absence o f a C l represents.
 Exploring Internal Motivation for Change 125
T a b l e 4 .13. Clitics Omitted in Required Contexts, by Clitic T ype and by Group

Clitic type Group I Group 2 Group 3

(1 speaker) (S speakers) (8 speakers)

[46] Accusative 0% 2/262 0.8% I IS / 4 3 3 3 -4 % l 2.7%

0.8%
[47] Dative 2 / 1 19 1.6 % 4/487 0.8% I 11/ 5 2 1 2% 1
[48] Obligatory reflexive 1 /10 0 1% 2/278 0.7% ' 13 /3 5 5 3.6 %
[49] True reflexive 0% 2/39 5% 6 /5 4 n%
[50] Indirect reflexive 0% 0% 10/77 13 %
>1.26% >6.6%
[51] Reflexive with 0% 0% 4/26 15.4 %
Inanimate Subject
[52] Reciprocal 0% 0% - 2 / 15 13 .3 %
[53] Impersonal 0% - -
[54] Affective -

[5 7 ] P ero la misat, el padre l o ; dijo muy pronto.

but the mass-femb the father rr-masc; said too soon
But the mass, the father celebrated i t too soon (V21 ,f 1 8,2,ELA67)
[58] Yo creo que no lo (los libros) usan.
I b e lie v e t h a t n o t i T - s g ( th e b o o k s ) u s e
I b e li e v e t h e y d o n t u s e th em ( j4 3 ,m i7 ,3 ,E L A s 7 )

[59] LE; iban a mandar a elr

H iM -d at, w e n t to s e n d a him-acq
They were going to send h i m (J43,m 1 7,3,ELA57)

Table 4.13 displays the percentages of omitted Cls by group according to type
of Cl. Table 4.13 indicates that in Groups 2 and 3 omission occurs more frequently
when it concerns a reflexive or reciprocal clitic, with a total omission of 1.26 per
cent and 6.6 per cent in Groups 2 and 3, respectively. By contrast, under 1 per
cent of dative and accusative Cls are omitted in Group 2, and only 2.7 per cent of
these Cls are omitted in Group 3. If we eliminate R42 from Group 3 (for the reasons
explained in section 4.2.3), the results still show higher omission with reflexives
and reciprocals, as flows from Table 4.14. Speaker R42 appears to show the effect
of schooling in his use of Spanish. Thus, even though another young man in
Group 3, J43 who acquired this language in natural settings and R42 appear to
have a similar overall level of proficiency in Spanish in terms of fluency, a closer
examination of their speech shows that R42s use of Cls is further removed from
the norms of Group 1 despite the fact that this speaker, but not J43, has received
formal instruction in Spanish in high school. This difference with respect to
omission of Cls (2.7 per cent omitted by J43; 53 per cent by R42) is also apparent
in regard to Cl position, as I show later.
1 26 Exploring Internal Motivation for Change
T a b l e 4 .1 4 . Quantification o f Clitic Omission for Seven Speakers in Group 3

Type o f clitic No. %

Accusative 6/408 1.5 %

Dative 9 /518 1.7 %
Reflexive and reciprocal 2 2/5 10 4 .3%

With respect to the omission of dative and accusative Cls, it may be concluded
that transfer from English does not play any role (if it did, we should have obtained
a higher percentage of omission of datives and accusatives when the coreferential
element is expressed; but this is not the case). As for reflexives, note that omission
is relatively high even in equivalent contexts ([49] true reflexives). Furthermore,
in most cases the factors which determine which verbs have a reflexive form in
Spanish are opaque. Here, then, we may have at least two motivations leading to
the same result: the complexity of the system in Spanish, and transfer of the
subcategorization of non-reflexive verbs in English to a number of verbs which are
obligatory reflexives in Spanish, as in [56]. This type of transfer, which may have
syntactic consequences, is explored in Chapter 6. The quantitative evidence, how
ever, does not support wholesale syntactic transfer. Furthermore, clitic omission is
almost non-existent in Group 2, which includes the US-born bilinguals who might
pass on their modified Spanish variety to younger generations.

4.3.1.2. Clitic position in verbal periphrases

The position of Cls in contemporary spoken Spanish may be either categorically or

variably pre- or postverbal depending on the type of verbal phrase, as explained in
(a)-(c). Examples [6o3] illustrate.
(a) Cls are categorically preverbal with finite verbal forms:
[60] lo compr/bo ha comprado ya
She bought rr/has bought i t already
(b) Cls are categorically postverbal with Infs, PresPs, and gerunds in complement
clauses:

[61] Vine p ara verTE

came-isg f o r see-Y O U
I c a m e s o a s to s e e y o u
[62] VindoiA te acordars de ella
se e in g -H E R y o u - r e f l r e m e m b e r - f u t - 2 s g o f h e r
S e e i n g h e r y o u ll r e m e m b e r h e r
 Exploring Internal Motivation for Change 127

(c) When Cls refer to an argument of an Inf or a PresP in a verbal periphrasis

with a finite semi-auxiliary verb, they may variably occur before thefinite verb. The
appearance of the Cl in front of the finite verb has been proposed as the conse
quence of clause union by Aissen and Perlmutter (1976), and of restructuring by
Rizzi (1978). This, then, is the only variable context for clitic placement in Spanish
(cf. Myhill 1988; 1989):

[63] Viene a verTE/TE viene a ver

c o m e s - 3 s g to s e e - Y O u / y o u c o m e s ~ 3 s g to see
H e s c o m in g to s e e yo u

It is interesting to note that none of the twenty speakers studied violates the
categorical pre- or postverbal placement constraints illustrated in [60-2]. In regard
to those utterances which allow one of two positions for Cls ([63]), on the other
hand, our study indicates that, contrary to what a naive view o f transfer might
predict, postverbal placement is less frequent in the speech of bilinguals. Further
more, the variables which simply favour preverbal placement in Spanish dominant
bilinguals (and in monolinguals cf. Landa 1989, Myhill 1988) appear as almost
categorical contexts for this order: the Spanish of bilinguals moves in the direction
of strengthening Spanish internal trends rather than English patterns (cf. the
extension of estar to be).
In his study of Cl placement in written Spanish, Myhill (1988) proves quanti
tatively that this phenomenon is constrained by at least two factors: the semantic
properties of the matrix or semi-auxiliary verb, and the relative topicality of the
subject and the Cl. The former factor favours restructuring (or clause union, i.e.
matrix plus Inf as one simple clause) more strongly than the latter. Thus preverbal
placement is strongly favoured when this verb retains little or none of its basic
meaning, and functions rather as a marker of tense, modality, or aspect, as illus
trated by a comparison of [64-5], In [65] the verbal periphrasis resembles a single
finite verb, and thus favours the preverbal placement o f the Cl.

[64] Matrix verb: venir to come (basic meaning)

Pepe viene a entrevistarmi hoy.
Petes coming to interview me today

[65] Matrix verb: venir keep (grammaticalized meaning)

Pepe me viene molestando por anos ya.
Petes been/kept bothering me for years

In [64-5] the Cl could have been placed in pre- or postverbal position. How
ever, the preverbal position is preferred when the matrix verb conveys epistemic
meanings, progressive aspect, and future tense, three meanings which Bybee (1985)
has shown to be expressed most frequently by means of bound morphology in
natural languages. The quantification of this preference for preverbal position in
128 Exploring Internal Motivation for Change
T a b l e 4.15. Cross-tabulation o f Clitic Position by the Semantics o f the Matrix Verb

Total Epistemic Deontic Progressive Future Basic meaning

N % (%) (%) (% ) (%) (%)

Chile 232 70 85 iS 89 91 33
Spain 3 27 70 79 60 91 94 41
Mexico 210 77 7S 63 80 92 41
Venezuela 302 62 75 56 75 74 32

Notes: Percentages correspond to preverbal position. N = number o f variable contexts for clitic place
ment. p = <.00.

data from four Spanish-speaking countries (Chile, Spain, Mexico, Venezuela) is

summarized in Table 4.15.17
The fact that Cls are a Romance creation, and that their position with respect to
the verb did not become fixed in Spanish before the fifteenth century, justifies
considering their variable position in verbal periphrases a syntactic-semantic change
in progress, which involves the gradual grammaticalization of a number of verbs.18
The trend to place the Cl in front of the matrix verb is, therefore, internally
motivated and related to the semantics of this verb.
The existence of the same tendency in bilingual Spanish is confirmed by our
study of Cl placement. Table 4.16 displays the quantification of preverbal v. post
verbal placement of Cls in verbal periphrases with those verbs which occurred at
least ten times in the sample of speech from the twenty speakers. Despite the
existence of a parallel construction with a postverbal pronoun in English, the
internally motivated opposite trend in the direction of preverbal placement appears
stronger in Los Angeles, as displayed in Table 4.16. No percentage is given when
there are fewer than five cases in a cell; the matrix or semi-auxiliary verbs are
ordered according to decreasing frequency of occurrence in Group 1. This order
ing roughly corresponds as well to decreasing frequency of preverbal placement.
The results indicate that future (ir a go to), progressive (estar be), and epistemic
(poder may) meanings clearly favour preverbal position, even more strongly than
in other varieties of Spanish (see Table 4.17). This preference for preverbal position
applies as well to the less favouring root modality (deber (de) must, tener que have

17 T h e data (taken from Esgueva and Cantarero 1981; Instituto de Filologa Andrs Bello 1979;
Rabanales and Contreras 1979; U N A M 19 7 1) were coded by a group o f my graduate students as part
o f a research methods course. I am grateful to these students for allowing me to use the coded data.
Some o f the matrix verbs examined are listed in Table 4.16.
18 Both in Spain and in Chile, speakers with a high level o f education have told me that, when they
were students in the 1960s, their instructors corrected them if they placed Cls preverbally in verbal
periphrases. T his preverbal position was allowed only with pronominal main verbs (e.g. asustarse to
get scared : S e va a asustar (He) s se going to get scared). T h is prescriptive attitude supports an
analysis o f preverbal placement with other types o f verbs as a change in progress.
 Exploring Internal Motivation for Change 129
T a b l e 4.16. Preverbal Clitic Placement in Verbal Periphrases

M atrix verb Group 1 Group 2 Group 3

(7 speakers) (5 speakers) (8 speakers)

No. % No. % No. %

Ir a go to 85/92 92 3 5 /3 6 97 61/66 92
Estar be 3 0 /3 3 91 19/20 95 3 1 /3 5 89
Poder may / can 23/38 60 18 / 19 95 3 6 /4 7 83
Tener que have to 17 /30 57 12 /16 75 8 / 13 65
Empezar a begin 8 / 11 73 0/1 2/2
Querer want to 6 /19 32 6 /11 55 12 /2 3 52
Deber (de) must i/ 6 17 3 /5 60 1/ 2

to), and to the inceptive class (empezar a begin). This trend increases among
second-generation bilinguals. Speakers in Group 3, on the other hand, though still
behaving according to this general tendency, have slightly lower percentages of
preverbal Cls than those in Group 2. This general result is not too surprising,
however. A comparison of other features of Group 3 Spanish with Group 2,
namely verb morphology and the extension of estar to be, reveals that, as a group,
third-generation immigrants do not seem to continue along the lines of a natural
historical development of their ancestors language. There is a break at this point:
the language of this group moves qualitatively further away from the norms of
Group 1 speakers, and also shows signs of convergence with the dominant contact
language.
There is, as we have seen, a certain amount of individual variation within each
group, but the only speaker who clearly departs from the strong general tendency
to place the Cl preverbally is R42 (Group 3). In only one of ten possible contexts
is the Cl preverbal in his data; i.e. R42 strongly favours postverbal position (90 per
cent) when this is a grammatical alternative in Spanish. I suggest that this is a
further consequence, in addition to Cl omission, of his formal learning of Spanish,
an experience which sets him apart from other members of his community, and
which appears to have made him more vulnerable to transfer from English. As
suming that formal learning increases the speakers awareness of form, this result
would support Meisels (1983#) observation that speakers who focus on form are
more likely to use transfer strategies than those who focus mostly on the messages
they wish to convey.

4.3.1.3. Acceleration o f change

As in the case of the extension of estar, the internally motivated trend to place Cls
preverbally in verbal periphrases, to a large extent as a consequence of a process of
130 Exploring Internal Motivation for Change
T a b l e 4.17. Percentage o f Preverbal Clitics in Verbal Periphrases in Los Angeles, in
Four Spanish-Speaking Countries (Chile, Spain, Mexico, Venezuela),
and in Written Spanish

Matrix verb Group Chile, Spain, Written Spanish

------------------------------- Mexico, V e n e z u e l a ----------------------
1 2 3
(%) (%) (%) (%) No. (% )

Ir a 92 97 92 '89 13 6 / 18 1 75
Estar 9i 95 89 83 7 5 /8 4 89
Poder 60 95 83 76 3 3 /17 8 19
Tener que 57 75 65 45 7/46 15
Querer 32 55 52 61 12/90 13
Empezar a 73 - - 46 3/32 9
Deber (de) 17 60 - 34 7/62 11

grammaticalization of a number of matrix verbs, is accelerated in a situation of

intensive language contact. This acceleration is clearly observed in Table 4.17,
which compares total percentages of preverbal Cls in Los Angeles with the cumu
lative percentages for four Latin American countries. Table 4.17 also includes
some of the results obtained for written Spanish by Myhill (1989), who examined
prose from five contemporary Latin American writers: Benedetti, Fuentes, Marquez,
Quiroga, and Rulfo.
As pointed out above, Group 2 has a higher percentage of preverbal Cls with
every semantic class of matrix verb. Indeed, with ir a go to, estar be, and poder
may, preverbal placement is almost categorical. In the other classes, the propor
tions in Groups 2 and 3 are in every case over 50 per cent. Note, furthermore, that
written Spanish displays much lower percentages of preverbal Q s with every verb
but estar. This difference is not unexpected: in general, the written mode proves
to be more conservative with respect to innovative language features. The various
quantitative data indicate, therefore, that this aspect of the grammar of Spanish
reflects an ongoing change which is more advanced, i.e. which has accelerated, in
a language-contact situation.

4.4. Conclusions

Section 4.3 offers evidence that some aspects of the morphosyntax of a language
(verbal clitics) are fairly impermeable to interlinguistic influence in a context of
societal bilingualism. A similar conclusion is reached by King (1989: 144-5), who
shows that Newfoundland French does not exhibit convergence with English, the
 Exploring Internal Motivation for Change
dominant language, with respect to either the omission or the placement of object
Cls. Furthermore, based on a study of Warlpiri-English bilinguals, Bavin (1989:
285) concludes that changes affecting the pronominal system of Warlpiri (includ
ing object Cls) cannot be attributed to direct influence from English, but rather to
a move- towards semantic transparency, which is generally internally motivated
change . This coincidence of results in three different language-contact situations
certainly strengthens the conclusion reached in the present study.
Indeed, the preverbal slot for Cls in Spanish appears to be as firmly imprinted
in speakers minds as tense-mood-aspect inflections are shown to be. Furthermore,
so-called reflexive Cls seem tightly associated with their verbal lexemes in the case
of obligatory reflexive verbs; in addition, there are a number of examples which
offer evidence of overgeneralization of this lexical pattern (as in [44^]). Further
evidence of the reality of Cl morphemes is provided by utterances where the Cl
appears in Spanish and the verb in English, as in [66-8].

[6 6 ] Mi m om quiere que lo s k eep - m y grad es u p

m y m o m w a n ts th a t th em k eep - m y grad es u p
M y mom wants me to keep my grades up 043,11117,3,ELA57)
[67] y lo que, queda - LO; invest in stock o algo asi
and what; remains - rrr acc invest in stock or something like that
And whatever remains you invest it in stock or something like that
(j43>m i7,3,ELA 57)
[68] No, uno - no s e quieren tie down
no, some - not s e want tie down
No, some dont want to get tied down [get tied down = amarrarii?]
(j43,m i7,3,ELAs7)

In the case of naturally developed bilingualism, this study does not offer evi
dence in favour of an overall transfer of parallel structures from the dominant
language. Regarding Klein-Andreus (1980; 1986a) hypothesis, then, my study
suggests that it should be modified to incorporate a condition on the type of
parallel structure likely to be transferred or preferred, as follows: in situations
of intensive bilingualism, the higher frequency of use in a subordinate language
S of morphological and/or syntactic parallel structures in a superordinate
language F will occur only if the parallel structure in F corresponds to a pre
ferred structure (i.e. more widely used variant) in S prior to the initiation of
contact with F.
It should be clear that this condition on parallel structure transfer corresponds
to what has been discussed in terms both of acceleration o f change and of
overgeneralization. Therefore, it seems that a theory of simplification, motivated
by interactional, cognitive, and intralinguistic factors (see Chs. 1 and 7), which
predicts the loss of certain morphosyntactic variables in subordinate contact lan
guages, accounts more appropriately than transfer for the preferred preverbal
132 Exploring Internal Motivation for Change
placement of Cls, as well as for the preference for progressive over simple forms
observed by Klein-Andreu in the Spanish of Puerto Ricans in New York, and the
preference for the prepositional phrase a la maison (de) over the simple preposi
tions chez, sur, and a verified by Mougeon et al. (1985) in Canadian French.
Accounts based on simplification and transfer must consider both the structures
of the languages involved and the type of contact situation. For instance, an
analytical construction in language F (generally regarded as cognitively simpler
than a corresponding synthetic one) which is rare or nonexistent in language G
may indeed be more complex for speakers of language G to acquire when learn
ing language F as a foreign language, or to transfer into language G from F. An
example is offered by Dorian (personal communication), who notes that imperfect
speakers of East Sutherland Gaelic generalize the more widely used synthetic
possessive construction rather than the free-standing possessive pronouns, even
though the latter structure finds a parallel in English, the contact language.
As for the role of the superordinate language in the shaping of the extension of
estar as well as the patterns of use of Cls, my conclusion approaches that of Dorian
(1978) in her study of East Sutherland Gaelic. We have dealt with different areas
of the grammar; yet in both studies direct influence from English, though possible,
is difficult to identify. This does not altogether preclude the possibility of transfer;
the convergence of certain grammatical features is certainly evident elsewhere (see
Chs. 5 and 6). In most cases, however, direct transfer seems to involve lexical
items, frozen/idiomatic expressions, and verb subcategorizations. This last type of
transfer motivates some omissions of Cls. None the less, ungrammaticality due to
estar and Cl usage is far from being widespread in Groups 2 and 3. Typically,
modifications regarding estar and Cls in these groups are for the most part more
appropriately described as more advanced stages of natural internal trends in the
ancestor variety.19 Conspicuous ungrammatical cases are very rare, and unlikely to
become norms, given that they occur in the speech of individuals who will prob
ably refrain from passing Spanish on to their descendants.
19 W ith respect to Cls, this statement could even apply to simplification o f case-, gender-, and
number-marking. Indeed, although this type o f simplification is not attested in Group 1, case-,
gender-, and number-marking o f Cls is lost to various degrees in different varieties o f Spanish. Fo r
example, case-marking with human object referents is lost to a large extent in spoken M adrid Spanish;
gender is not marked in the dative case in standard Spanish; plural marking with dative clitics is to a
large extent lost in many varieties o f Latin American Spanish (e.g. Ya le entregu el libro a los alumnos
I already gave (ler 3Sg) the book to the students;).
 Exploring External Motivation for Change

5.1. Introduction: On the Perm eability o f G ram m ar1

I n o t e d in Chapter 1 that the transfer of features from one language into another
is one of the strategies used by bilinguals to cope with having to communicate in
two different linguistic systems. Throughout this book, I have examined the pos
sibility that a number of changes affecting Spanish in Los Angeles may be the
result of (among others) specific processes of transfer of features from English, that
is, processes that imply what Martinet (in his Preface to Weinreichs Languages in
Contact) calls the permeability of linguistic cells.
Weinreich (1974: 4) states that the study of purely linguistic problems about
bilingualism may allow the linguist to see the cause of the susceptibility of a
language to foreign influence in its structural weaknesses (my emphasis), i.e. in the
existence of structural spaces and incomplete or unbalanced correlations within the
system. Weinreich (p. 44) further observes that foreign grammatical elements may
permeate the speech of a bilingual but are only very rarely incorporated in the
language as a code. Thus, he concurs with Meillet, Jakobson, and Sapir (all three
cited on p. 25) in supporting the impermeability of a grammatical system to
foreign structural elements except when they correspond to the systems internal
tendencies of development, an observation which has been shown to be dangerously
circular.
More recently, Thomason and Kaufman (1988) note that linguists have chal
lenged the view that a grammatical system is impermeable to direct transfer of
foreign elements. On the basis of evidence provided by a number of studies, they
argue that purely linguistic constraints on contact-induced change will not work
(p. 16), and go as far as to say that any linguistic feature can be transferred from
any language to any other language (p. 14). However, their conclusion that it is
the social context, not the structure of the languages involved, that determines the
direction and the degree of interference (p. 19) is not entirely in disagreement
with Weinreichs (1974: 4-5) proposal that the extent, direction, and nature of
interference of one language with another can be explained even more thoroughly
in terms of the speech behavior of bilingual individuals, which in turn is conditioned
by social relations in the community in which they live.

1 I would like to thank Enric Vallduvi for helpful comments on an earlier version o f this chapter, and
Francisco Ocampo for his help with word-order questions. I would also like to thank some o f my
graduate students for contributing in various manners to the progress o f this chapter: Alazne Landa,
Loli M artinez-Moya, Lu is Candia, and Gorka Elordieta.-
134 Exploring External Motivation for Change

The controversy concerns which specific factor associated with intensive lan
guage contact actually motivates change: transfer of elements from the contact
language, perhaps under pressure to make more similar the structures of the
languages involved? Lack of formal education in one of the languages? Reduced
use of one of the languages and consequent incomplete acquisition of this lan
guage? Transfer is undoubtedly a controversial notion, as I pointed out in Chapter
x. Some of the more obvious cases of transfer, usually at the lexical, morphological,
and phonetic levels of analysis, may be more easily identified and evaluated as to
their stability in the borrowing language in terms of frequency of occurrence and
numbers of users (cf. Poplack et al. 1988; Mougeon and Beniak 1991). By contrast,
at the syntactic level it is more difficult to identify transfer and to prove that there
has been permeability of a grammatical system to influence from a different one.
More often than not, this difficulty arises from the fact that a given phenomenon
may be considered to be syntactic by some analysts, but lexical (e.g. verb sub
categorization) or morphological (e.g. clitic pronoun usage, gender agreement)
by others. Thus, uncontroversially syntactic phenomena appear to be reduced to
such questions as word order, the possibility of null arguments, and patterns of
subordination.
In this chapter I examine four different phenomena which I hope will offer further
evidence that Weinreichs proposals appear to be correct, at least in a sociolinguistic
situation of intensive and extensive bilingualism in which two languages with
unequal social and functional status are in contact. Indeed, in agreement with Meillet,
Jakobson, Weinreich, and more recently Bickerton (1981), and Givon (1979), the
research considered in this chapter indicates that, even under conditions of intense
contact and strong cultural pressure, speakers of a secondary language do not
introduce elements which cause radical changes in the syntactic system of this
language under normal conditions of transmission (cf. Thomason and Kaufman
1988). It may be possible that any linguistic feature can be transferred from any
language to any other language as a nonce-borrowing in the speech of bilinguals
(cf. Weinreich 1974), but only those that are compatible (in the sense proposed in
the conclusion to this chapter) with the structure of the borrowing language at any
given stage will be adopted, disseminated, and passed on to new generations.
It must be kept in mind, however, that Spanish and English have been in
contact in the U SA for under 200 years, and that U S Spanish has enjoyed the
uninterrupted presence of monolingual or near-monolingual Spanish-speaking im
migrants. It is possible, that, given enough time depth and favourable socio-political
conditions, the changes allowed, which occur gradually, may lead to the development
of a language fundamentally different from non-contact standard Spanish. At this
stage in history, the evidence favours the hypothesis proposed in section 1.1 above,
that the structure of the languages in contact governs the introduction and diffusion
of innovative elements in the linguistic systems; while the sociolinguistic history of
the speakers is the primary determinant of the language direction and degree of
 Exploring External Motivation for Change 135
diffusion of the innovations, as well as of the more distant or remote linguistic
outcome of language contact.
In opposition to Weinreich, on the other hand, my research indicates that the
permeability of a grammar to foreign influence does not depend on its structural
weaknesses but rather on the existence of superficially (in terms of string order; cf.
Prince 1992) parallel structures in the languages in contact. Given a primary
language A and a secondary language B, the permeability of B will not be evident
in the incorporation of new syntactic structures on the model of A, but first and
foremost in the following: (a) the extension of the discourse-pragmatic functions of
a structure in B according to the model of the functions of the parallel structure
in A (cf. Weinreichs notion of modelling at the lexical level; this extension of
pragmatic functions of structures already existing in the borrowing language may
have motivated the statement that a language would change only in agreement with
its internal tendencies of development); (b) the preferential use or increase in the
frequency of use in B of an existing structure parallel to one in A to the detriment
of variants in B (which may or may not convey different discourse-pragmatic
meanings); and (c) the loss of semantic-pragmatic constraints governing the use of
the variants of a syntactic variable in B when the corresponding structure in A is
not sensitive to such constraints, (a), (b), (c) are instances of the types of transfer
(2)j (3); (4) discussed in Chapter 1 (p. 4).
The grammatical phenomena studied here do not allow me to provide a definitive
answer to whether (c) occurs independently of (b), or whether (c) is a consequence
of (b). However, the tendency to overgeneralize reflexive-type verbs (see Ch. 4),
thus losing the semantic content conveyed by the opposition reflexive/non-reflexive,
an opposition which finds no parallel in English, appears to indicate that (-) is not
necessarily dependent upon (b). Reduction of the range of social domains for the
use of a language appears to be a stronger cause of loss of semantic-pragmatic
variants than preferential choice of a structure that is parallel with one in the
contact language.
Within this wider definition of permeability (as compared to a narrower definition
in terms of introduction of new syntactic structures), let us consider four possible
cases of permeation of the syntax of Spanish by English grammatical rules: (1)
non-expression of the complementizer or relative pronoun que that; (2) omission
of an obligatory Cl when the corresponding English construction does not require
an oblique pronoun; (3) obligatory Subject-Verb-X order; and (4) obligatory
expression of a lexical subject. These cases may illustrate permeability of the type
of (b) and (c) above.
An example of permeability of type (a), the extension of the discourse-prag-
matic functions of a structure in B according to the model of the functions of the
parallel structure in A, is offered in Los Angeles Spanish by the extension of the
discourse-pragmatic functions of the deictic este/esta this (masc./fem.) in the speech
of a few bilinguals. Este/esta this appears to replicate the function of new-this' in
136 Exploring External Motivation for Change

English, a term proposed by Wald (1983) to refer to the use of unstressed this/these
to introduce into discourse indefinite specific referents which are being foregrounded
for extended attention. This type of construction with new-este this is not at
tested, to my knowledge, in Mexican Spanish, but it occurs in the data from Easi
Los Angeles (ELA), albeit infrequently. [1] and [2] are illustrative.

[x] Yo ahorita, yo termin un caso que, este negrito piat a cuatro hombres y a una
mujer. Y vivan muy, - como animales.
I recently, I finished with a case that, this black man killed four men and s
woman. They lived very, - like animals (1150,11146,3,ELA36)
[2] . . . haba dejado mi pap el carro ah y haba dejado las muas, ah, irse, ves? Pero
esta mua tena la, la idea de que si no coma algo en la casa, no se iba.
. . . my dad had left the cart there and he had let the mules, uhm,go, you see:
But this mule had the, the idea that if it didnt eat something in the house, it
wouldnt leave (A37,m57,3,ELA22)

In [1], a brand-new referent which is not present in the situation where the
communication takes place, negrito black man, is introduced in subject position
and modified by the deictic este this . In general spoken Spanish, however, deictics
are not used to modify nouns referring to brand-new entities, nor are brand-new
referents introduced in subject position (see discussion in section 5.5). The passage
that follows [1] concerns este negrito this black man. I must conclude, therefore,
that fi] illustrates transfer into Spanish of the pragmatic function of new-ife: the
introduction into discourse of an indefinite specific referent which is being
foregrounded for extended attention. A similar analysis in terms of transfer applies
to [2], the beginning of a narrative in which this mule plays an important role. In
this case the speaker has introduced all the mules (in object position), but the first
mention of one specific mule would require the use of the quantifier una one before
it could be referred to as esta mula this mule in the ensuing discourse.

5.2. Non-expression o f a Com plem entizer or Relative Pronoun Que

T h at

At first glance the non-expression of a complementizer in Spanish constitutes an

obvious candidate for an explanation based on direct transfer from English of the
possibility of a null complementizer in relative clauses and in complement clauses.
This section shows, however, that examples such as [3] correspond to permeability
of type (b), increase in the frequency of use of a parallel structure, rather than to
introduction of a new rule allowing the non-expression of the complementizer.

[3] Yo creo 0 inventaron el nombre.

I believe (that) they invented the name (A2o,f1 9,2,ELA46)
 Exploring External Motivation for Change 137
Note that the non-expression of que that is not attested in relative clauses in
LA Spanish, but only in complement clauses, i.e. no examples o f the type o f [4] are
found in the speech samples from any of the bilinguals studied.

[4] *E l nombre 0 (ellos) inventaron era extrao.

The name (that) they invented was strange

Interestingly, in the particular context of complement clauses the non-expres-

sion of que that is not a case of incorporation of a feature or rule foreign to the
Spanish system. Indeed, que is allowed to be unexpressed in verb complements in
formal and in written registers of Spanish, as shown in [5-6] (cf. Subirats-Riiggeberg
1987: 168-73). The same phenomenon is not allowed in relative clauses, however,
as seen in [7].

[5] Te ruego 0 me lo envies pronto

I pray (that) you send it to me promptly
[6] Deberan tomarse precauciones pues se cree 0 podran ir armados.
Precautions should be taken because it is believed (that) they could be armed
[7] *Te agradezco el regalo 0 me enviaste
Thak you for the present (that) you sent me

In L A Spanish, zero que appears almost exclusively with complements of the

estimative verb creer to believe/to think (estimative verbs are defined in section
5.5), and in the first person singular, yo creo I believe/think. This may be a re
flection of the data, however, in which creer is by far the most frequent estimative
verb. Other verbs of this type, such as pensar to think, used quite sporadically,
may also occur with a null complementizer ([10]). Furthermore, the speaker who
is at the most extreme end of Spanish language attrition in the sample, N40,
produces an example of zero complementizer with a non-estimative verb in Eng
lish contained in a Spanish sentence with no prosodic breaks: Esa noche we found
out 0 ella se fue That night we found out 0 shed left. Examples [8-13] further
illustrate the variation between expressed and zero complement:

[8] lNo la hallo muy entusiasmada1, yo pens entre m. Yo creo 0 no la quiere ver [la
pelcula] como yo.
I dont think shes very interested, I thought for myself. I believe 0 she doesnt
want to see it [the movie] as much as I (A20,fi9,2,ELA29)
[9] . . . no muy chiquito, pero yo creo 0 ha crecido como - unas seis pulgadas.
. . . not very small, but I believe 0 he has grown about - about six
inches (Li9,f22,2,ELA54)
[10] M i mam no quiere que hago eso. Ella piensa 0 si, si no voy full-time no voy a
terminar
M y mom doesnt want me to do that. She thinks 0 if, if I dont go full-time
I wont finish (S38,fi9,3,ELA66)
 Exploring External Motivation for Change
[ u ] St, creo que no ms esa vez fuimos.
Yes, I think that only that time we went (A2o,fi9,2,ELA29)
[12] Yo creo que son robadas [las radios]
I think that [the radios] are stolen goods (833/19,3,ELA29)
[13] . . . pero no s cmo, pues, no s cmo, pero pienso que eso va a ser.
. .. but I dont know how, well, I dont know how, but I think that thats
going to happen (S38,fi9,3,ELA66) '

Zero que occurs fairly frequently in the speech of 19 of 35 speakers examined.

These 19 speakers are distributed as follows. Group 1: two younger women and
two older women (4 o f a group of 13 speakers comprising 8 females and 5 males);
Group 2: four younger women and two older women (6 /11: 7f, 4m); Group 3:
three younger women, two older women, two younger men, and two older men
(9 /11: 6f, 5m).
The sex correlation is intriguing: 10 of 15 women show zero que in Groups 1
and 2, but none of the 9 men examined in these groups. Only the men in Group
3 (4/5) leave the complementizer unexpressed. I have not systematically examined
sex as a possible predictor of degree of language maintenance or loss. Impression
istically, however, it seems to me that women maintain the language of their
ancestors with a higher degree of proficiency than their brothers in any given
family; so I would not expect the women in my sample to show the effect of
English more strongly than men. The fact that only women are using null
complementizers in Groups 1 and 2, therefore, appears to be further evidence that
zero complementizer in English may have had only a favouring effect in the
diffusion of zero que in the L A bilingual community. The phenomenon has firm
roots in Spanish itself, and its extension to the oral mode may have had its origins
in ancestor Mexican varieties. The linguistic distributional restriction to a type of
complement clause where other Spanish registers also allow zero que argues in favour
of permeability as defined above, i.e. limited to already existing structures in the
language. It is possible that explanations based on the vague concept of latent
internal tendencies (Weinreich 1974: 25) may have been motivated by change that
has affected structures severely restricted according to register or style in any given
language.

5.3. Omission o f O bligatory Verbal Clitics

Consider at this point another candidate for syntactic transfer: the omission of an
obligatory Cl, specifically in constructions where the Cl is said to function as a
marker of possession of the entity referred to by the direct object (Sol and Sol
1977: 30). This is illustrated in [14d\ (the standard form), and in [14^), where the
first-person singular Cl me me-dative has been omitted and possession is marked
by a possessive determiner in the direct object nominal. I argue in this section that
 Exploring External Motivation for Change 139
[14b\ which replicates its English translation, corresponds to permeability of types
(b) and (c).
[14] (a) .. .y me dieron en la cara, y me quebraron la mandbula
and me hit-3pl in the face, and me broke-3pl the jaw
(b) . . .y me dieron en la cara, y 0 quebraron mi, mi jam
and me hit-3pl in the face, and 0 broke~3pl my, my jaw
. . . and they hit me in the face, and broke my, my jaw (036,^145,3,
ELA43)
Examples of the type of [14b\ though not frequent (these examples do not al
ways involve a switch to English), occur in the speech of many of the US-born
bilinguals in the sample, but do they represent a new syntactic construction? It
seems to me that the answer is negative. The structure of [14b] exists in Spanish,
with a variant that does not have a parallel in English, as illustrated by [15a, b], and
their single English translation.
[15] (a) Tiraron una piedra y quebraron mi portalpices (*a m).
(b) Tiraron una piedra y me quebraron el portalpices (a m).
They threw a stone and broke my pencil-holder
Constructions of the type of [15a, b] are characteristic of a number of verbs
subcategorized for a direct object and, under certain conditions, an indirect object
whose referent has the semantic role of beneficiary, as shown in [16]:
[16] quebrar: [__NPX (a NPy)],2 where x = theme, y = beneficiary
Other verbs in this class include: agarrar to grab, aplastar to crush, arreglar
to fix, dibujar to draw, lavar to wash, limpiar to clean, peinar to comb,
quemar to burn .
The condition requiring an indirect object is of a semantic-pragmatic nature: it
concerns the degree of detachment or inalienability viewed as existing between the
referents of the direct object and the indirect object, such that the more closely
related they are the higher the degree of active participation the beneficiary has
in the situation, and consequently the higher the likelihood of being coded as an
indirect object rather than simply as a possessor (see Garcia 1975, esp. ch. 4, on the
meanings of case roles). This is exemplified in [17, 18]:
1 17] (a) Le arregl la corbata a Juan\ (no hizo bien el nudo)
I fixed Johns tie; (he didnt tie a good knot)
(b) Arregl la corbata de Juan/su corbata; (*no hizo bien el nudo)
I fixed Johns tie/his tie; (he didnt tie a good knot)
[18] (a) Me lav el pelo (a m) (*mientras yo limpiaba la casa)
He washed my hair (while I cleaned the house)

2 B y convention, square brackets represent the subcategorization o f a lexical item, N P stands for
noun phrase, and parentheses indicate optional elements.
 140 Exploring External Motivation for Change
(b) Lavo mi pelo (mientras yo limpiaba la casa).
He washed my hair (while I cleaned the house)

Note that the interpretation of [17a, 18] is that the indirect object referent is
a relatively active participant in the situation (cf. Garca 1975: ch. 4) and therefore
closely related to the affected entity (the direct object; for instance, Juan was
wearing the tie in [17a]. By contrast, in [ 17/?, 18b], by virtue of the passive role of
the possessor, which is coded in the direct object, the affected entity must be
interpreted as being detached from its possessor. Thus, in [i8/>], mi pelo my hair
must refer to hair which is not attached to the speaker; for instance, hair that the
speaker has had cut and washed to make a wig.
The English translations given in [15, 17, 18] show that the two Spanish vari
ants correspond to one in English (although the distinction can be clarified in
English by other means). This interlinguistic difference is further indicated by the
fact that the Spanish sentences in [17, 18] may or may not be appropriately
continued by the language material given in parentheses, which highlights the
closeness of the affected entity to the beneficiary, while the English translation
accepts it in every case.
Indeed, the English verbs corresponding to those listed in [16] do not allow an
indirect object in this language, as exemplified in [19, 20].

[19] *1 grabbed the arm to John/*I grabbed John his/the arm

[20] *He washed the hair to me/*He washed me the hair

Bilinguals who use examples like [14^] (quebraron mi, mi jaw They broke my,
my jaw) appear to be violating a semantic-pragmatic constraint. A lexical expla
nation is also possible: the subcategorization of verbs as those listed in [16J is
changing along the English model such that they do not allow an indirect object.
This explanation, if correct, is compatible with the semantic-pragmatic one. In
fact, the semantic-pragmatic constraint is needed to characterize this particular set
of verbs undergoing change in L A Spanish. As [21] illustrates, other types o f verb,
subcategorized for an optional indirect object with a semantic role other than
beneficiary, do not appear to be changing.

[21] (Le) entrego la pistola (al policia)x i = goal

He handed in the pistol (to the policeman); (R.5o,m46,3,ELA38)

On the other hand, some speakers in Groups 2 and 3 redundantly mark posses
sion in Spanish constructions where the equivalent English construction would not
require it, as shown in [22], an indication that changes in the Spanish pattern may
be occurring independently of direct English influence. [22] contains three separate
markers of possession, the Cl me me, the stressed object pronoun in the indirect
object a mi to me, and the possessive determiner mi my. Possessive determiners
do occur with body parts in standard Spanish when the sentence does not include
an indirect object or dative Cl referring to the possessor (e.g. Ponlo debajo de t u s
pies Put it under your feet.
 Exploring External Motivation for Change 141

[22] .. .y me peg a mi en mi brazo

and me hit to me on my arm
. . . and he hit me on the arm (R24,m2o,2,ELA5o)
In conclusion, Spanish has not been permeated by a foreign syntactic structure,
i.e. a verb phrase with a direct object containing a possessor and a possessed entity
and no Cl correferential with the possessor, [ [0 V] [possessive det. N] ] (see [17 b,
18b). Rather, there seems to be a process of lexical change affecting the occurrence
of one of the possible arguments of a set of verbs. This occurrence is controlled by
a semantic-pragmatic constraint concerning the degree of closeness between the
entities involved and their degree of active participation in the situation. The loss
of this constraint is probably triggered by the bilinguals preference for equivalent
structures in the two languages (cf. Klein-Andreu 1986) and by the fact that the
parallel structure in English is not subject to the same semantic-pragmatic con
straint. These modifications correspond to the types of possible grammatical
permeability presented in (b) and (c) above, which involve neither incorporation
nor necessarily loss of syntactic structures.

5.4. Obligatory SV X Order

The order of major arguments in Spanish, sensitive to semantic-pragmatic rules,

may also offer a ground for possible modifications o f the type o f (b) and (c). For
instance, it is well known that Spanish allows both SV and VS order with single-
valency verbs (cf. Silva-Corvaln 1977; Ocampo 1989). Ocampo (pp. 17 1-2) has
shown that VS is the strategy selected to introduce new subject referents into
discourse. English, by contrast, does not have an exactly parallel VS option. In this
respect, compare the word order of [23a], produced by a Spanish monolingual,
with its English translation:
[23] . . . estuve una hora ah [a clinic] tocando timbre, (a) sali el mdico [VS]
(From Ocampo 1990: 97, ex. 14)
. . . I was there [a clinic] for an hour, ringing the bell, (a) the doctor came
out [SV]

In [23], the new NP subject referent el mdico the doctor is introduced in

postverbal position in Spanish, while the English version maintains its more rigid
SV order regardless of the informational status of the subject. Consider now [24],
produced by a Spanish-English bilingual:

[24] Una vez estaba en una gasolinera aqu (a) y una seora lleg ah [SV]. Yo estaba
ah esperando, estaban trabajando en el carro, (b) Una seora entr [SV] y me
pregunt si conoca . . .
Once (I) was at a gas station here (a) and a lady got there. (I) was waiting
there, (they) were fixing my car. (b) A lady came in and asked me if (I)
knew . . . (H48,m39,3,ELA24)
142 Exploring External Motivation for Change
In [24a\ the new NP subject referent una seora a lady is introduced in preverbal
position; it occurs in this same position in [24b], apparently reintroduced after a
parenthetical explanation. Both [24] and [24!] replicate the English versions. This
constitutes further illustration of the bilinguals choice of a structure which paral
lels one in his dominant language, thus violating not a syntactic rule but a semantic-
pragmatic constraint in Spanish, his secondary language.
It appears that modifications of semantic-pragmatic rules which lead to both loss
or overgeneralization of syntactic variants are frequent in language-contact situa
tions (see e.g. Dorian 1980; Landa 1995; Silva-Corvaln 1991a). But how likely are
modifications that affect word order in L A Spanish to spread and become stable?
An analysis of data from 15 speakers indicates that pragmatically controlled word-
order variation is still quite strong. Consider, among many found in the data,
[25a], produced by the same speaker as [24], where the new subject referent, mi
pap my dad, is introduced in postverbal position, as expected in Spanish. Fur
ther illustration of SV /V S variation is offered in [26, 27].

[25] [Eso] pas el otro da. (a) Vino mi pap [VS], Mira, aqu esta - necesitan un
precio aqu en estas formas'1
That happened the other day. (a) My dad went, Look here, this - they need
a price on these forms (H48,m39,3,ELA24)
[26] Pero yo me levanto a ver qu se va a hacer ese da. (a) S i el nio necesita esto
[SVO], (b) si la nia necesita lo otro [SVO] (c) y as va la rutina [VS] . . . ( d )
Y los nios realmente nos tienenjvenes [SVX], Porque H. dice de () cuando tena
l diecisis [VSX], diecisiete ( f ) y llegaba el mes de marzo [VS], le gustaba subirse
a una loma y volar papalotes.
But I get up to see what were going to do that day. (a) If the boy needs this,
(b) if the girl needs that (c) and the routine goes on like this . . . ( d ) And the
children really keep us young. Because H. says that (e) when he was sixteen,
seventeen ( / ) and March came, he liked to climb up a small hill to fly
kites. (L28,f37,2,ELA23)
[27] A: Naci en Colorado [tu pap], me dijiste?
D: No. Yo creo que (a) su pap naci en Colorado [SVPP]. Mi, mi abuelo.
A: Was he [your dad] born in Colorado, did you say?
D: No. I think (a) his dad was born in Colorado. My, my grandfather
(D36,m45,3,ELA43)

Examples [26, 27] illustrate the expected variation in word order: [SVO/X]
with evoked or given NP subjects ([26a, b, d])\ [VS] order with new NP subjects
([26c,/]); [VSX] with non-contrastive subject pronouns ([26?]); and [SVPP] with
a new but contrastive NP subject ([27]), where the speakers father is being con
trasted with his fathers father.
I have conducted a fairly rough quantitative analysis of word-order patterns in
data from 15 speakers (see Table 5.1), marking the preverbal/postverbal placement
 Exploring External Motivation for Change 143
T able 5 .1. Speakers Included in the Study o f Word Order and Subject Expression

Group i Group 2 Group 3

Sex Age Sex Age Sex Age

Ei F 20 V 21 F 18 S38 F 19
Sz F 24 A20 F 19 A46 F 31
Aq F 62 L 19 F 22 M 47 F 33
L28 F 37 R42 M 15
R24 M 20 H48 M 39
D36 M 45
R50 M 46

of the subject without considering differences in types of verb (e.g. transitives,

intransitives), the informational status of the subject (new or evoked information),
or the number of expressed constituents in the sentence. The results show higher
percentages of preverbal subjects in English dominant bilinguals (see Table 5.2),
which could be interpreted to reflect a trend towards a more fixed SV X order.
However, this possible outcome is still distant. Similar observations about the
resilience of word order under pressure from a different word-order system are
made by Mithun (1989: 245-6) in regard to Oklahoma Cayuga word order, in a
situation of contact in which English is also the dominant language.
The distribution of the speakers included in the quantitative analysis of word
order and in the study of subject expression in section 5.5 below is summarized in
Table 5.1: three speakers from Group 1; five from Group 2; and seven from Group
3-
A total of 643 sentences with expressed subjects were quantified. Overall results
for the 15 speakers indicate that 499/643 expressed subjects were placed preverbally,
i.e. 78 per cent. The percentages by group show a steady, though not large,
increase of preverbal subjects: Group 1, 74 per cent (99/133); Group 2, 77 per
cent (177/230); Group 3, 80 per cent (223/280).
The subjects were classified into three types: personal pronouns, NPs, and
other pronouns (e.g. demonstrative and indefinite pronouns like esos those, alguien
someone). The informational status of these three types of constituent, as well as
the observed correlation between word order and informational weight, seemed to
justify considering them separately: personal pronouns tend to refer to already
activated, given information, NPs tend to introduce new information into discourse.
The status of the other pronouns in this respect was not investigated; thus, in
principle, it appeared prudent to separate them from the other two categories. The
results, displayed in Table 5.2, confirmed the advisability o f the separation into
these three classes. With respect to word order, speakers in Groups 1 and 2 treat
144 Exploring External Motivation for Change
T a ble 5.2. Cross-tabulation o f Subject Placement by Type o f Subject and by
Speaker Group

Preverbal subjects

Personal pronouns Noun phrases Other pronouns

f
No. % No. % No. % P

Group 1 7 1/8 4 85 22/39 56 6 /10 60 <.00

Group 2 98/108 9i 5 9 /9 0 66 20/32 63 <.00
Group 3 I 3S / I44 94 7 3 / 118 62 15 / 18 83 <00

other pronouns in a manner similar to NPs. In Group 3, however, other pro

nouns favour preverbal position rather strongly.
With every type of subject, Groups 2 and 3 increase the percentage of preverbal
placement. This percentage is quite high with personal pronouns in Group 1, 85
per cent; and even higher in Groups 2 and 3, 91 per cent and 94 per cent. As
expected, NPs and other pronouns occur less frequently than personal pronouns in
preverbal position in the three groups of speakers; but the trend in favour of
preverbal position by speakers in Groups 2 and 3 is also apparent with these types
of subjects: from 56 per cent and 60 per cent in Group 1 to 66 per cent and 63 per
cent in Group 2; in Group 3, NPs have a slightly lower percentage of preverbal
position, 62 per cent, in comparison with Group 2, but a high 83 per cent of pre
verbal other pronouns.
Assuming that differences in communicative needs may be neutralized by the
amount of data and similarity of discourse topics, the attested increase of preverbal
subjects in Groups 2 and 3 may be interpreted, however, as the consequence of
processes o f loss o f semantic-pragmatic constraints on preverbal subject place
ment, as illustrated in [21]. Once again, this process of loss may be favoured by
contact with a more rigid SV-order language but does not constitute a radical
change in the system of Spanish.
Indeed, preverbal placement of personal-pronoun subjects is the most frequent
variant in monolingual varieties. Furthermore, Ocampo (1989) shows that in spo
ken Argentinian Spanish (from the Rio de La Plata region), NP subjects are almost
categorically preverbal in pragmatically unmarked SV X constructions (95 per cent,
see Table 5.3). With single-valency non-state verbs, on the other hand, SV is less
frequent than V S (40 per cent v. 60 per cent), but the subject can be predicted to
be almost categorically preverbal (96 per cent) when it conveys evoked information
in pragmatically unmarked constructions. These figures, adapted by me from
Ocampo (pers. comm.; see also Ocampo 1989), are summarized in Table 5.3.
As Table 5.3 clearly indicates, NP subjects are much more frequently placed
 Exploring External Motivation for Change 145

T able 5.3. Subject Noun Phrase Placement in Pragmatically Unmarked Sentences in

Rio Platense Spanish

SV X SV VS
(S mostly evoked) (S = evoked) (S = new)

252/264 95% 23/24 96% 33/34 97%

Totals S V /S V + V S 23/58 40%

SV(X) 275/322 85%

preverbally in Rio Platense Spanish, a variety not in intensive contact with Eng
lish, than in any of the samples of L A Spanish included in this study. The higher
percentages of preverbal subjects in English dominant bilinguals, as compared to
the Spanish dominant group, on the other hand, may perhaps be interpreted as
reflecting an incipient movement in the direction o f the SV(X) order o f the
dominant language.

5.5. Obligatory Expression o f Subjects

5.5.1. Analysis
%
The fourth potential case of permeation concerns the possibility that the expres
sion of subjects in the Spanish of bilinguals may be becoming obligatory, as one
might surmise on the basis of such examples as [28]. Note that in this example the
speaker repeats the subject of [28], un mexicano a Mexican, in [28^], where one
would normally expect a non-expressed subject (see [29b] below).
[28] En tiempos antiguos - (a) un mexicano case [se casaba] con un[a] mexicana y (b)
UN m e x i c a n o no puede [podia] hablar con - otros - nacionales [grupos etnicos]
In the past - (a) a Mexican married a Mexican and (b) a M e x i c a n couldnt talk
with other ethnic groups (R42,mi5,3,ELA3o)
In this section I will focus on the question of subject expression in some detail,
and will address two specific questions: (1) Does the almost categorical expression
of subjects in English favour a higher percentage of expression o f subjects in
bilingual Spanish as compared to monolingual dialects, i.e. is there evidence of
transfer of type (b)? (2) What accounts for the variable expression of lexical sub
jects at the various stages of Spanish-language attrition?
Most Romance languages, e.g. Spanish, Catalan, Italian, Portuguese, allow both
filled and empty subject syntactic positions, as shown in [29a, b\ respectively.
146 Exploring External Motivation for Change

[29] (a) Cuando Juana viva aqu, ellai/t siempre nos invitaba a su casa
When Juana lived here shei/t always invited us to her home
(b) Cuando Juanax viva aqu, 0i/t siempre nos invitaba a su casa
When Juana lived here *0/shei/t always invited us to her home

The phenomenon illustrated here, to which I refer as the variable expression o

the subject, lends itself to the investigation of permeability of type (b), i.e. increased
preference for subject expression, given that English requires the subject position
to be filled in contexts where Spanish does not, as the translations in [29] show.
Note that, with respect to this phenomenon, the incorporation of elements
foreign to the structure of the language would have to involve the occurrence of
non-referential, dummy subjects, as in [30, 31].
[30] *el est lloviendo
it is r a i n i n g

[31] * e l l o s me robaron el auto [said without knowing who stole the car]
TH EY v e s t o le n m y c a r

Not surprisingly, there is not a single case of such non-referential, dummy subjects
in the Spanish data examined.
If a higher percentage of expression of subjects is attested in bilingual Spanish,
this would be the result, then, of a rather insidious type of permeability; given that
it does not violate sentence-based grammatical rules, it is less salient and could
spread with relative ease. This type of transfer obviously requires a quantitative
examination of subject expression in the data.
It is not entirely clear how comparable subjects may be in English and Spanish.
Indeed, if the role of the subject pronoun in English is to indicate person and
number, then it is quite plausible that bilingual speakers may associate subject
pronouns in English with Spanish verb inflections. If this is the case, we would
expect a lower rather than higher Overall frequency of pronoun expression, as well
as a higher frequency of expression with morphologically neutralized verb forms
(e.g. cantaba I/he, she sang)3 in bilingual Spanish as compared to monolingual or
Spanish dominant varieties. By contrast, if bilinguals establish cross-linguistic
equivalence at a surface level, and if they truly tend to make both languages
structurally more similar (cf. Gumperz and Wilson 1977), then we would expect a
much higher frequency of subject pronouns as a result of intensive contact with
English.
I have argued, furthermore (Silva-Corvaln 1986; 1990^; and in Chapter 4 above),
that bilinguals tend to follow internal language preferences when simplifying
variables. This hypothesis does not predict which variant will be favoured by
bilinguals in the case of a stable linguistic variable such as subject expression seems

3 A morphologically neutralized verb form is any o f the undifferentiated forms for 1st and 3rd
person singular in all the tenses o f the subjunctive mood, in the imperfect o f the indicative, and in the
conditional (see T able 2.1).
 Exploring External Motivation for Change 147
to be. I f we consider the issue of analysis versus bound morphology, however, we
may conclude that overt subjects will be preferred over verb inflections as markers
of person and number.
On the other hand, many linguists (e.g. Dorian 1980; Landa 1995; Silva-Corvalan
1986; 1991a) have documented the loss of semantically close alternatives correlated
with different pragmatic functions in processes of simplification. Assuming this
type of loss and the pragmatic functions of subject expression (see below) in a pro
drop language (i.e. a language that allows null lexical subjects), we may expect a
trend towards fewer subjects. We are faced, then, with competing motivations
(cf. DuBois 1985): contact with a non-pro-drop language and analysis favouring
categorical subject expression; and loss of (marked) pragmatic options favouring
zero subjects.
It is certainly also possible to look at the issue at hand within the framework of
parametric syntax and its associated theory of markedness. In accordance with
traditional linguistic thought, some parametric studies (e.g. Roeper and Williams
1987, cited in Romaine 1992) make use of markedness theory to predict stages of
language acquisition, such that unmarked settings of parameters are said to be
acquired earlier than marked ones. In regard to the so-called pro-drop and non-
pro-drop phenomena, however, there is some controversy both over which of the
two settings is the marked one4 and over which is the possible interrelationship
between markedness theory and ease of acquisition (cf. Romaine 1992).
Despite evident disagreements, the parametric model of syntax has been applied
to the study of pidgins, creoles, and second-language acquisition. But here again
one finds contradictory evidence and claims, compounded by the unresolved ques
tions posed by transfer theory and by the hypothesis that L2 acquisition replicates
the stages of L i acquisition.
One could pursue this line of thinking and examine subject expression in a pro
drop language in contact with a dominant non-pro-drop language, as illustrated by
Spanish and English in Los Angeles, adopting the structural analysis proposed
by Adams (1987) in her study of the loss of pro-drop in French. Note, however,
that her theory forces her to assume a VOS underlying word order for pro-drop
Romance languages, and an SVO word order for non-pro-drop languages such as
French and English. Her well-argued analysis, however, leaves her at a loss to
explain why Yiddish is not pro-drop while Chinese languages are, and why Italian
requires the expression of the second-person singular subject pronoun with sub
junctives.
Having found no agreement on, or satisfactory answer to, the question of subject
expression, I have chosen to disregard the autonomous syntax option. Relying on
a view of language as a system of human communication whose structural features
will best be explained with reference to cognitive and social factors, I continue to

4 According to Romaine (1989: 3 9 -4 1), Hyams proposes that pro-drop is the unmarked case for
first-language acquisition, while White argues that pro-drop is the marked setting.
148 Exploring External Motivation for Change

adhere to a semantic-pragmatic approach to syntactic variation, and incorporan-

semantic-pragmatic variables as possible factors determining the expression or
non-expression of subjects.
Structurally, the observation that Spanish has the option of expressing or not
expressing a subject pronoun appears to be valid for most sentential contexts in a
discourse vacuum. Examined in the normal flow of speech, however, the complexity
of the question of variable subject expression becomes clear. It is evident on the
one hand that the phenomenon is not optional in every possible environment and,
on the other, that speakers may also have the option of expressing or not express
ing an NP subject, as illustrated in [32, 33]. Methodologically, then, it appeared
appropriate to me to consider NPs in at least some of the quantitative analyses of
this question.
[32] Pepe lleg temprano hoy a la oficina; *Pepe/?l/0 se vino en taxi.
Pepe arrived at the office early today; *Pepe/he/*0 came by taxi
[33] Me vine con Pepe hoy a la oficina; Pepe/l/0 vive cerca de mi casa
I came to the office with Pepe today; Pepe/he/*0 lives near my home
Note that if the subject is coreferential with the subject of the preceding sen
tence, as in [32], a full subject NP is not acceptable in Spanish (assuming an
informational intonation pattern, i.e. the pragmatic function is to communicate
information with no further connotations or implicatures). However, if the subject
is coreferential with the referent of an oblique constituent in the preceding sen
tence, the three alternatives are allowed: full NP, pronoun, or null subject.
Studies during the past decade have shown (cf. Bentivoglio 1987; Enriquez
1984; Silva-Corvaln 1982) that the variable expression of subjects is controlled by
discourse-sensitive factors, namely the establishment of an entity as the topic of
more than one sentence (a sentential topic is a referent about which new information
is being added in the proposition; Lambrecht 1987: 222) ([34]); the need to
express focal information ([35, 36]); the need to identify subject referents clearly
([37]).

Establishment o f topic
[34] S = S2,f24,i,ELA37; C = researcher
S: creo que si l lo hiciera - el sacrificio [studying in the evenings] - o yo lo hiciera,
sera algo bueno.
C: S. Para ayudarle al nio.
S: Al nio, s. /C: Um, hmm, s./ A veces s (a) yo me pongo a pensar y digo,
Bueno, (b) yo no estoy tan mayor todava . Porque para la edad que 0 tengo -
son veinticuatro aos. / C: Claro. / Entonces, () 0 me pongo a pensar que apenas
sera una edad como para - Yo todava, este, pudiera ser una mujer soltera.
/C: Um, hmm./ Y - 0 s que 0 tengo mi nio, entonces, con ms razn decir
uno, (d) 0 tengo el, el tiempo perdido, bah, 0 puedo aprovecharlo.
 Exploring External Motivation for Change 149

S: I think that if he made it - the sacrifice [studying in the evenings] - or I made

it, itd be good.
C: Yes. To help the boy.
S: The boy, yes. /C: Um, hmm, yes./ Sometimes (a) I start thinking and say,
Well, (b) P m not that old yet. Because for my age - Im only twenty four.
/C: Yeah./ So, (c) (I) start to think that it would be hardly the age to - I still,
eh, could be a single woman. /C: Um, hmm./ And - (I) know (I) have my son,
so, theres even a better reason for one to say, (d ) (/) have, have wasted time,
bah, (I) can use it well.

Focal information

[35] 0 dije entre m, 0 dije, Oye, Alb., eras t (*0) el que estabas hablando all
arriba? Yo (*0) mismo - estuve, a, sorprendido de m. - Yo (*0) era el que
estaba hablando all arriba? Primero 0 me encomend a Dios. (A29,m6o,2,
ELA2)
I (0) said to myself, I (0) said, Listen, Alb., Were you (*0) the one speaking
up there? / (*0) myself - was surprised at myself. - I (*0) was the one
speaking up there? First I (0) entrusted myself to God.

Contrast

[36] Y yo me fui a trabajar y l se qued ah un rato. Le hizo el reporte a la polica

y todo (R n ,f42,i,E L A i8)
*Y 0 me fui a trabajar pero 0 se qued ah un rato. Le hizo el reporte a la
polica y todo
And I (*0) went to work and he (*0) stayed a while longer. (He) gave the
report to the police and all

Clarification o f subject referent

[37] S = S2,f24,i,ELA37; C = researcher

S: Pues, se me grab tanto en la mente que cuando la sepultaron, yo de noche
miraba visiones, pero era la realidad. Porque yo despertaba gritando y mi
hermana tena que levantarse a verme. /C: Ah, fjate./ Y (a) ella iba a mi lado
y (b) yo estaba temblando, que hasta los dientes se oan que pegaban.
S: Well, it got so fixed in my mind that when they buried her, I saw visions at
night, but it was real. Because I would wake up screaming and my sister had to
get up to see what was happening to me. / C: Oh, gee!/ And (a) shed come up
to my bedside and I was trembling, you could even hear my teeth striking
together.
 Exploring External Motivation fo r Change
Thus, although the subject is not required in the second conjunct of [32, 3 3 1,
nor in any of the sentences in [34], it is required in the italicized positions in [35,
36/ 37]. Intralanguage analyses and cross-language comparisons must take these
facts into account, therefore, when examining the issue of pro-drop and of possible
interlinguistic transfer. Note that the English translations in [29, 32-7] require a
subject pronoun in every case.6
Clearly, then, discourse-related factors should be taken into consideration in
cross-group comparisons. Accordingly, this study adheres to the methodology
used in previous discourse-based studies. Basically, this consists in coding each
subject, expressed or not, of a finite verb by a number of factors hypothesized as
having some bearing on subject expression (cf. Bentivoglio 1987; Cifuentes 1980
1; Morales 1986: ch. 5; Silva-Corvaln 1977; 1982). (Excluded from quantification
are fixed phrases, e.g. lu sabes you know, eso es thats it, as well as other cat
egorical contexts such as impersonal constructions.) For this chapter, a total of
2,176 cases were examined and coded in the data from the three groups of speakers.
Building upon previous investigations, the internal variables listed below were
included in the analysis (see Appendix 3 for a list of all the factors included in each
independent variable group):
(i) Coreferentiality index. The subject in question is coded according to whether
it is coreferential with the subject of the preceding finite verb ([34 a, b~\ illustrate
coreferential subjects), with another argument of the preceding finite verb, or with
none of the preceding clause arguments (see [44-7]). Unless otherwise stated,
coreferentiality between subjects is opposed to switch reference between subjects
in the tables below, i.e. patterns of coreferentiality with other syntactic constitu
ents are disregarded in the columns labelled switch in the tables.
(ii) Morphological ambiguity of the verbal form. The morphologically ambiguous
forms include the first and third person singular of the Imp and the Cond, as well
as of the two Sub tenses ([37 a, b] illustrate ambiguous Imp forms).
(iii) Communicative value of the subject, with respect both (a) to its status as
focal or part of the focal information in its sentence ([35]), and (b) to its status as
a focus of contrast ([36]). Evidently, every contrastive subject is also focal informa
tion in its sentence, but not vice versa. Furthermore, I differentiated contrast from
focal information on the basis that their correlation with subject placement/word
order would not necessarily be the same, and word order was also being examined.
A subject is considered to convey focal information when it represents the most

5 Note that the requirement that a contrastive subject be expressed cannot be captured on the basis
of just the structural description; it must be specified that the subjects are non-coreferential. I f they are
coreferential, the contrastive situation does not obtain, regardless o f the structural description, and
expressing the subjects becomes optional : f/ y o hablo bien espanol y /pero el ingles 0/yo lo hablo muy
quebrado I speak good Spanish and/but English I speak poorly .
6 English allows null subjects in strings o f co-ordinated sentences with a coreferential subject when
these sentences present sequential situations as illustrated in (i) and (iii) (cf. [32]): (i) Pete came by taxi
and 0 arrived early at the office today; (ii) Pete came by taxi. *(2 arrived early at the office today, (iii) Pete
got up at six, 0 came by taxi, and 0 arrived early at the office today.
 Exploring External Motivation fo r Change
crucial piece of information in its sentence at the purely informational level. Note
that a subject may carry intonational prominence for pragmatic purposes other
than the communication of the newest or most crucial information; for instance, to
call the listeners attention towards the referent of the subject in relation to some
unusual or unexpected situation, as in [38]. In these cases of marked pragmatic
functions, the subject is not coded as focal (Rodrigo 1991). Indeed, the speaker can
make a communicative choice in examples of the type of [38]: to highlight the
subject referent or not (Ba or Bb); this choice is associated with the possibility of
expressing or not expressing the subject, respectively. In the case of informationally
focal subject referents, this choice is not available, as the unacceptability of [39, B
(b)\ indicates:
[38] A: Juan no puede ayudamos maana.
B: (a) e s e / j u a n no ayuda nunca/(b) 0 no ayuda nunca
A: Juan cant help us tomorrow
B: (a) t h a t o n e / j u a n never helps/() *(He) never helps
[39] A: Quin no puede ayudamos maana?
B: (a) j u a n no puede/(b) *0 no puede
A: Who cant help us tomorrow?
B: (a) j u a n cant /(b) *(He) cant
(iv) Semantics of the verb.7 It was hypothesized that some types of verb would
favour subject expression: estimative and verbs o f the type gustar to like were
therefore coded separately from other verbs.
Estimative verbs are those verbs that tend to present the speakers point of view
as implicitly opposed to that of others (e.g. to think, to believe, to assume, to agree
with-, cf. Enriquez 1984: 152, 235-45). This implicit idea of contrast may explain
their positive correlation with subject expression. As observed before, explicit
contrast is a categorical context for subject expression.
Verbs of the type gustar to like in Spanish are those that subcategorize a subject
with the semantic role of theme, and an indirect object with the role of experiencer,
as in [40].
[40] A Pepe le gust/le encant/se le olvid/se le perdi el vino
to Pete to-him liked/lacked/forgot/lost the wine
Pete liked/loved/forgot/lost the wine
With these types of verb the subject is frequently inanimate and conveys the
new information in the sentence; I considered it necessary, therefore, to code them
separately.

1 I also examined the effect that the reflexive/non-reflexive form o f the verb might have on subject
expression based on the hypothesis (suggested to me by Flora Klein-Andreu a few years ago) that
reflexive forms might favour non-expression given the disambiguating effect o f the reflexive pronoun.
Cross-tabulations showed that the correlation was random; i.e. given a reflexive/non-reflexive verb one
could expect a 50/50 chance o f an expressed subject.
152 Exploring External Motivation for Change
(v) Type of subject. Every subject was coded according to its category (e.g
pronoun, noun phrase) and the person and number of its referent.
(vi) Subjects were also coded to indicate if they occurred within the same turn
of speech, or at a new turn. New turns were considered those points at which a
different speaker initiated a turn, as well as those at which the speaker either
quoted himself or someone else within the same turn of speech.
(vii) Finally, I also investigated the possibility of a correlation between subject
expression and clause type: main or subordinate on the one hand and declarative,
interrogative, exclamative, and imperative on the other. However, only declarative
sentences were examined quantitatively due to the very few cases of interrogatives.
exclamatives, and imperatives attested in the data.

O f these independent variables, quantitative evidence provided by the cross

tabulations performed (with the statistical program crosstabs in SPSSX) indicated
that change in turn of speech and clause type (independent v. subordinate) did not
appear relevant. By contrast, coreferentiality, morphological ambiguity, verb se
mantics, type of subject, and communicative value (i.e. contrast and focus) showed
statistically valid correlations with subject expression. Contrast and focal information,
as expected, proved to be categorical contexts for the expression of a nominal or
pronominal subject. Thus, although examples of the type of [41], with a non
expressed contrastive subject, and [42], with a non-expressed focal subject, are in
principle possible errors, they are not attested in the data. That is to say, no
examples were found of non-expressed focal or contrastive subjects, even where
the verb inflection unmistakably identified the subject referent (as in [41, 42]).

[41] M i hermano estudi en USC, *y/pero 0 estudi-isg en Indiana

Lit.: M y brother studied at USC, *and/but 0 studied-isg at Indiana
[42] iQuin trajo los libros? *0 los traje-isg.
Lit.: Who brought the books? *0 them brought-isg

For the purpose of establishing whether English-dominant bilinguals show any

preference for pronoun expression, it may be sufficient to compare their linguistic
behaviour with that of Spanish-dominant bilinguals in the community. A stronger
case would be made, however, if the results were compared with Spanish varieties
not in intensive contact with English. This comparison is rendered difficult by the
fact that every study of this variable has incorporated slightly different constraints
in regard to what is to be included in the quantitative analysis. On the whole,
however, there are similar trends in the results obtained. Including NPs and
categorical contexts of occurrence (i.e. contrast, focus), the overall percentage of
subject expression in the three groups in East Los Angeles (ELA) is 41 per cent
(892/2176), remarkably close to the 43 per cent (457/795) attested in my study of
subject expression in West Los Angeles Spanish (Silva-Corvaln 1977).
To compare with non Mexican-American Spanish varieties, the results shown
in Table 5.4 exclude NPs, but they do retain focal and contrastive pronouns.
 Exploring External Motivation fo r Change 153
T a b l e 5.4 . Cross-tabulation o f Subject Pronouns by Speaker Group in E L A and
Comparison with Madrid (Spain), Boston (Puerto Rico Spanish), Santiago (Chile), and
Caracas (Venezuela)

Expressed pronouns/ % Yo I Nosotros we

total no. o f verbs
No. % No. %

Group I 118 /3 6 0 33 5 8 /13 7 42 6/39 15

Group 2 163/586 28 5 7 /19 7 29 11/ 7 6 15
Group 3 198/758 26 78/222 35 4 /6 1 7
Madrid 4 , 857/ 23,717 21 3,2 4 9 /10 ,18 5 32 2 53/2 ,4 31 10
P R S (Boston) - 37 573/ 1,333 43 35/208 17
Caracas - 3 2 9 /72 1 46 2 8 / 17 1 16
Santiago 1,58 7/4 ,18 2 38 752/2,238 34 9 9 /571 17

a T h is percentage has been adapted from Hochberg (1986) so as not to include the fixed phrase tu sabes
you know . Bentivoglio s (1987) and m y studies do not include fixed phrases.

Subsequent cross-tabulations examining the effect of coreferentiality and verb

ambiguity were performed both including and excluding these categorical contexts.
The percentages of expressed subject pronouns do not indicate any striking
differences overall among the three ELA groups, but there is a gradual decrease
of the frequency of expression in Groups 2 and 3 as compared to Group 1. ELA
percentages for all pronominal subjects are intermediate between those reported
for Madrid, 21 per cent (Enriquez 1984), and for Chilean, 38 per cent (Cifuentes
1980-1), and Puerto Rican Spanish (PRS) in Boston, 37 per cent (Hochberg 1986:
613). In regard to expression of the first-person pronouns yo I and nosotros we
only, the percentages of expression are lower than the frequency reported for PRS
and for Venezuela (Bentivoglio 1987: 36).
Note that Group 3 speakers, expected to show the effect of contact more than
any other group, have the lowest global percentage of subject expression in East
Los Angeles. Further, the percentage of usage in Madrid and among Groups 2 and
3 speakers is almost the same for yo (and fairly close for nosotros in Group 3), while
the highest percentages correspond to Caracas, a variety which is not in intensive
contact with English. Should it be concluded from these overall results that evid
ence is sufficient to bury the transfer from English hypothesis once and for all?8
Certainly not without controlling subject expression in its relation to semantic-
pragmatic factors, a methodology which should be incorporated in any comparison
across groups or individuals.9
8 See Morales (1986: 89 -100) for a brief reference to arguments in favour o f this hypothesis, and for
her own counter-argument based on a study o f subject expression in a sample o f Puerto Rican bilinguals.
9 Hochberg (1986) does not use this control, thus weakening the strength o f the cross-dialectal
evidence she discusses in support o f her functional hypothesis.
154 Exploring External Motivation for Change
Though overall group results in East Los Angeles show some decreasing fre
quency of expression in Group 3, cross-tabulation by each speaker indicates clear
crossovers between groups, as seen in Table 5.5, which orders the speakers from
highest to lowest percentage of expressed subjects (column 1). In addition to
speakers name and group membership, Table 5.5 includes the following: (a) In
column i, the number and percentage of occurrences of all types of subject (pronouns
and NPs) over total number of finite verbs for eadi speaker, (b) In column 2, the
percentages for pronouns only. (Unless specified, all tables include personal and
other types of pronoun.) In the other group (see Appendix 3, factor 9: Subject
type) I have included such deictic pronominals as ste this one and eso that".
(1c) In column 3, number, rounded-off percentages, and statistical significance for
all types of subject cross-tabulated by coreference. (d) In column 4, number,
rounded-off percentages, and significance for all types of subject cross-tabulated
by ambiguity of the verb form. Statistically significant correlations are italicized in
columns 3 and 4.
The results displayed in Table 5.5 show a wide range of different frequencies of
subject expression, but these may be motivated by diverse communicative needs.
More importantly, this table shows that the coreferentiality index has a stronger
correlation than ambiguity with the subject expression variable. While coreferentialit>
is statistically valid for 13 of the 15 speakers, the morphological constraint is valid
for only 5: 2 of the 3 speakers in Group 1, and 3 of the 5 speakers in Group 2.
Combining speakers into groups, the correlation between ambiguity of the ver
bal form and higher subject expression, shown to be valid in previous studies
(Cameron 1990; Hochberg 1986; Silva-Corvaln 1982), turns out to be statistically
valid in Groups 1 and 2, but non-existent in Group 3. The results are displayed
in Table 5.6.
Tables 5.4-6 include focal and contrastive subjects, which are obligatorily ex
pressed, and NPs. The exclusion of these types of subject obviously reduces total
percentages, but it does not alter the relative frequency patterns of expression
across groups or the correlation with the ambiguity variable, as illustrated in Table
5-7-
With regard to morphological ambiguity, then, it must be concluded (a) that it
exists as a weak constraint on subject expression among some speakers in Groups
1 (A9), and 2 (L28, A20), and (b) that it no longer exists in Group 3. In spite of
this loss, there are very few cases of confusion in the data; and most of them are
clarified almost immediately by the ensuing discourse (cf. Ranson 1991). Speakers
in Group 3 may be relying much more on contextual clues, therefore, to disambiguate
the subject referent, as in the following passage, in which the speaker is talking
about an aunt who had been burnt.

[43] Toda la cara - pelo - todo el pelo, eyebrows. Cuando fuimos, fuimos mi mam
y yo - a verla, 0 estaba hinchada, 0 parece [pareca] pumpkin, 0 estaba muy
hinchada de los ojos, (a) 0 no tena los ojos claros, estaban - pi - dark brown.
 V
N O' O o c o r - ' " t - o O' oo o m o so
II M (S O o r ^ w w o O O ' O O O m c o

M Tf M H
5.5. Cross-tabulation of Subject Expression by Speaker, by Coreferentiality, and by Ambiguity of the Verb Form

CO CO CO CO

0 1
i in O' r-* O' m 00 O' 0
w O N O 0 s O in in so m

13 /5 5
a M W M H W t-t O' O' O' tH w 00
o \ s N \ \ s s. \ \ \ \ \
O' 00 O' 00 t- so Tj in ^ 00 sO r- Tf- SO
Z m in CO m -** c o co CO CO co in co M

Oh

m n o m o ro o m so o >-* m 00
O

15
r r f - so r|- m in in h ro h ^ O O co
ec
C/J p
G o
3 st
bp
O O r- O O' so N CO O' rj- O CO CO r-
flO 15 Tj- SO N CO CO M N w CO N co

9/60
N Tf
8 o \ \ s X s. \ N N N \ N s. \
u I
( in hH Th in
On z m
w
O
N
SO
co
N
N .Cl-tO <s sO sO sO
M M l-f N >~i HH

v
11 O O ^ O O O O O s O f N O O O s O

.00
O O O O O O O O O O O O O O

'S 0 0 0 co o in in o r** so w o o rj- m o

40
O' h in in h n m ro t m ^ m in - t

00
M m 00 o O' ^ in so o O' so
O' O' h w in 00 m 00 w so i> 20/50
* s, O' O' ^ M OO O
cn Z so m rj- so m in t
Oh
Z
3

O' h ro h 1 - h
co m 00 m tJ- rt-
O rf- N rf sC
rj- in sO Q sO ^
2/64

S N N S , \ s

COOh o 0"0 \o O' O' O m m in

M >H N w co m o

C

u
8

N X O' N O N ^ OO
N Oh t J- CO CO CO CO N c o m n

m o O O' O' t~~- s C CO

J9

r j- t J- r f CO CO CO CO CO

c | CO N O ' O ^ OO OO S O w WOO'S
I co in co co co >-< -*f O'
22/117

CO s o s o 00
S 5/5 o
8 fe
able

sO Os O' m O O
z Z r> r> h 00 vO
CO C"
'3- in t'-'
CO
T

o- ^ fO N,
A46 (3)

s
o O00O 90
Q O ^ , 'V
O Oo OO N O ' ^ w I,.
N J 1 N r m N O' w CO Tf w M n Jif
1 56 Exploring External Motivation for Change
T able 5.6. Cross-tabulation o f Subject Expression by Ambiguity o f the Verb form

Ambiguous Non-ambiguous

No. % No. % p=<

Group 1 6 5 /111 59 10 1/39 7 34 .00

Group 2 8 3/12 9 64 221/598 37 .00
Group 3 95/256 37 271/670 40 39

a Pronouns and N Ps included.

T able 5.7. Cross-tabulation o f Subject Pronoun Expression by Ambiguity o f the Verb

Form

Pronouns/total verbs Ambiguous Non-ambiguous

No. % No. % No. % p=<

Group 1 88/330 27 3 7 /8 3 45 5 1/2 4 7 21 .00

Group 2 10 9 /531 21 23/69 33 86/462 19 .00
Group 3 15 4 / 7 12 22 42/202 21 11 2 / 5 1 0 22 .80

1 Focal and contrastive pronouns excluded.

I dont know what color her eyes were, pero - with the fire - 0 tena como un
film over it, and they were light, almost light blue or light grey. /C: Oh!/
They were so light! Y - y estaban - tenan morphine / C: Claro/ para el pain
/ C: Para el dolor/ dolor. Y - hablamos con ella y 0 quera agua. Y, y 0 me dijo
que 0 quera agua, 0 quera agua. Pero - no pudamos - darle nada. (A46,
3i,3,ELA2o)
Her whole face - hair - all her hair, eyebrows. When (we) went, my mom and
I went - to see her, (she) was swollen, (she) looked like a pumpkin, (she) was
very swollen, her eyes, (a) (She) didnt have light-coloured eyes, (they) were
- - pi - dark brown. I dont know what colour her eyes were, but - with the
fire - (she) had like a film over it, and they were light, almost light blue or
light grey. /C: Oh!/ They were so light! And - and (they) were - (they) had
morphine /C: Yeah/ for the pain /C: For the pain/ pain. And- (we) spoke
with her and (she) wanted water. And, and (she) said that (she) wanted water,
(she) wanted water. But - (we) couldnt - give her anything.
In [43], nine singular verb forms occur without an expressed subject. The seven
morphologically ambiguous Imp forms have three possible grammatical subjects:
the speaker, her mother, and her aunt. The mother and the aunt are also possible
 Exploring External Motivation fo r Change 157
T able 5.8. Percentage o f Subject Expression by Syntactic Function o f the Coreferential
Argument in the Preceding Sentence3

No. %

Sentential subject 30/42 63

Switch reference with all arguments 624/1,080 57
Coreference with oblique argument 18/36 5o
Coreference with direct object 28/86 33
Coreference with indirect object 14/48 29
Coreference with subject 176/873 20

a All types o f subject included. The subjects o f 7 sentences which occurred after a short
pause in the conversation were excluded from the quantification o f coreferentiality because
their status in this respect was not quite clear to me.

subjects of the two non-ambiguous third-person singular forms (one Pres and one
Pret) in the passage. However, the content of the passage plainly indicates that the
subject referent of these nine verbs is the aunt. No subject needs to be expressed
given the unbroken topic chain, yet six Spanish speakers consulted agreed that
they would have tended to express the pronoun ella she in [43a] (Ella no tenia los
ojos claros she didnt have light-coloured eyes).
Previous studies of subject expression indicate that the most significant internal
variable determining variable subject expression is coreferentiality, such that if the
subject is coreferential with the subject of the preceding finite verb it is expressed
in only 20-5 per cent of cases, as opposed to 50-60 per cent subject expression
in cases of switch reference (cf. Bentivoglio 1987; Hochberg 1986; Silva-Corvalan
1982).
This coreferentiality constraint applies differently when the subject is coreferential
with a complement of the preceding sentence, or when it refers to the preced
ing proposition or propositions. Expression will be highest in this latter case
([44]),10 second highest when there is switch reference with all the arguments of
the preceding verb ([45]), lowest when the subjects are coreferential; and at least
in the sample from E L A speakers included here, coreferentiality with an indirect
object ([46]), a direct object, and an oblique argument ([47]) correlates with a
gradual increase in subject expression, as shown in Table 5.8. Let us first consider
some examples.
[44] Coreference with preceding proposition or propositions:
De vez en cuando pasa eso (Li9,f22,2,ELA54)
Sometimes that happens

10 T h is group also includes sentential subjects, as in Lo que tu quieres puede causar problemas. 0
puede agravar las cosas ('W hatyou want may cause problems. It (0) may make things worse ).
158 Exploring External Motivation for Change
T able 5.9. Cross-tabulation o f Subject Expression by Coreferentiality

Coreferential Switch

No. % No. % p=<

Group 1 4 1/ 16 5 25 125/343 Si .00

Group 2 50/290 17 2 5 4 /4 3 7 SB .00
Group 3 82/402 20 284/524 54 .00

a Pronouns and N Ps included.

[45] Switch reference with all arguments of the preceding verb:

Agarran los recibos y con ese dinero (j&-3pl) empiezan el da. Y cada vez que
l trabajaba les haca falta cinco o diez dlares. (Li9,22,2,ELA54)
They get the receipts and they start the day with that money. And whenever
he worked, they were missing five or ten dollars.
[46] Switch reference with subject - coreference with indirect object:
Le dijo al manejador que si (zS), le permita abrir el cuarto porque tena
miedo (R5o,m46,3,ELA3)
He asked the driver if he would let him open the room because he was scared.
[47] Switch reference with subject - coreference with oblique object:
Me estacion a cinco minutos de mi oficina. M i oficina es pequea.
(L i 9,f22,2,ELA54)
I parked five minutes away from my office. My office is small.

These correlations are very interesting because they appear to support hypotheses
which explain other phenomena (e.g. Cl duplication, Cl omission, relativization;
see e.g. Franco 1991; Givn 1976; Keenan and Comrie 1977; Landa 1995; Silva-
Corvaln 1981) on the basis of hierarchies of topicality. Furthermore, they bury
once and for all the idea that old/new information as an opposition between infor
mation that the speaker assumes to be present/not present in the interlocutors
consciousness is a factor with strong correlations with the subject expression
variable. I f presence in the interlocutors consciousness were the determining
factor, it would predict similar percentages in all cases of coreferentiality. Rather,
what appears to be crucial is whether the entity referred to by the subject has been
established as a discourse topic. If it has not, mentioning it in a complement does
not appear to be sufficient. Therefore, it tends to be referred to lexically in the
subject position before it can be left unexpressed in an ensuing sentence.
The quantitative results displayed in Table 5.9 oppose coreferentiality to switch
reference with the subject, regardless of possible coreferentiality with another
 Exploring External Motivation for Change 159
argument of the preceding finite verb, and include categorical factors, i.e. focal and
contrastive subjects. The results support the validity of the coreferentiality constraint
when the speakers are assembled into the three groups.
To examine coreferentiality and ambiguity further, Table 5.10 excludes categorical
factors and considers pronouns only in rows (a) and pronouns and NPs in rows (b).
When the speakers are grouped as in Table 5.10, the results show that, while
only Groups 1 and 2 are still sensitive to the ambiguity of the verbal form (cf.
Table 5.7), the three groups maintain the constraint on the expression of coreferential
subjects. Methodologically it seems appropriate to include NPs in the cross
tabulation, since speakers may be using them to code switch-reference subjects or
to disambiguate referents, as well as to exclude subjects that are obligatorily ex
pressed (i.e. focal and contrastive). Table 5.10 shows that the percentages obvi
ously change when NPs are included, but the significance of the correlations is not
much altered.
When individual results are examined excluding categorical factors, the results
remain quite similar to those of Table 5.5 in terms of relative frequencies and
statistical significance; therefore, I do not present them here.
Some speakers show a higher percentage of expressed coreferential subjects as
compared to E i (Group 1) and L 19 (Group 2), who have the lowest percentage
(excepting A46, a speaker at the bottom of the continuum who has an exceptionally
low percentage of expressed subject pronouns; see Table 5.5), and they are not
expressing as many switch-reference subjects. At the same time, they are not
expressing many pronouns when the verb form is ambiguous, as much as E i or
L 19 (70 per cent each), and they are expressing too many with non-ambiguous
forms (e.g. L28, H48, D36, R50, R42 in Table 5.5). What could explain this
differential degree of relevance between coreferentiality and identity of verbal form
in the first- and third-person singular?
Both factors, expression of subject with ambiguous verbal forms and in cases
of switch reference, converge towards a disambiguating function. In the case
of ambiguous verbal forms, many speakers, but especially those in Group 3 (see
Tables 5.5 and 5.10), seem to be relying on context for the disambiguation of
subject referents. The differential stages of loss of these constraints may reflect a
certain degree of direct transfer from English, inasmuch as English speakers may
not be as sensitive to person and number verb inflections as they would be to
matters of coreferentiality. This could favour the retention of the coreferentiality
constraint as compared to ambiguity.
Note that English allows coreferential subjects of independent sentences to
remain unexpressed even if there is an intervening subordinate clause with a
switch subject in cases of VP conjunction, as in [48], but requires the expression
of a coreferential subject in subordinate clauses, as in [49]. Spanish, by contrast,
does not favour an expressed subject in utterances like [49], as I indicate in the
translation.
 V
If o o
o o

p>
u
IS
B
c1t
oa ,-
2 '

V
O
c(D

Q N in
N -3* oc
*H w

s>

o LC tn
rt- O MO
?

>>->

c
3o V
u rO
o
II
o Oo O
o O
o o o
o o
>%
JD
a
o

Os T#-
cn

X>
4>
W
5 o
t
<u

Q
3
X>

4>

M ro
Cu o- CX
S 3 3
o O O
a \o o
 Exploring External Motivation fo r Change 16 1
T able 5 .11. Cross-tabulation o f Coreferential Subjects by T yp e o f Subject and by
Speaker Group*

Coreferential subjects

Personal pronouns Noun phrases

No. % No. %

Group 1 33/88 38 0/23 0

Group 2 26/109 24 8/60 13
Group 3 56/154 36 5/98 5

If
I

A
A
0

b
'j i

a Focal and contrastive subjects excluded.

[48] My mom opened the door carefully when I told her the light was on, 0
pushed a chair that was in the way, and suddenly 0 saw the thief

[49] M y mom always pays cash when she1*0 goes shopping

Mi mam siempre paga al contado cuando $/}ella va de compras
Further support for the sensitivity of English speakers to questions of corefer-
entiality is provided by the Hudson-Edwards (1986) study of the grammaticality
judgements offered by 303 college students on such complex sentences as those in
[50- 3]-
[S] Joe sawJane while taking his break
[51] Bob saw Sue while taking her break
[52] Arnold saw Grace while making a fool of himself
[53] Herb saw Susan while making a fool of herself
These authors conclude that three of four principles found to be operative in
judging the grammaticality of sixteen complex sentences of the type of [50-3] relate
to conditions of coreference between the subject of the subordinate clause and the
subject or another argument of the main clause.
I examined the possibility of direct transfer by cross-tabulating only coreferential
and non-focal subjects by type of clause, main or subordinate. Unfortunately, the
scarcity of the data does not allow valid quantification, but I did not observe any
trend towards expressing a higher percentage of subjects in subordinate clauses.
There is a trend, albeit weak, towards losing the coreferentiality restriction overall
(see S38 and M47 in Table 5.5), regardless of type o f clause.
Further support for this analysis, that there is a gradual loss of the coreferentiality
restriction, is provided by cross-tabulating expressed coreferential personal pro
nouns and NPs by speaker group. Table 5 .11 presents the results.
I2 Exploring External Motivation fo r Change

Table 5 .11 indicates that pronouns and NPs behave quite differently in regard
to the coreferentiality constraint. While Group 1 speakers may use a fairly high
percentage of coreferential as opposed to switch-reference subjects, they do not
produce a single coreferential, non-focal subject NP (0/23). By contrast, these
noun phrase subjects do occur in Groups 2 and 3 (13 per cent and 5 per cent).
Although it cannot be denied that knowledge of English may favour the retention
of the coreferentiality constraint and the loss of the*ambiguity constraint, internal
Spanish-language factors must be contemplated as well. Indeed, I am not aware of
any study of subject expression that has challenged the validity of the coreferentiality
constraint in the Spanish variety studied. By contrast, while the Real Academia
(1973: 422), Gili y Gaya (1970: 228), and research by Cameron (1990), Cantero
Sandoval (1976), Hochberg (1986), and Silva-Corvalan (1982) argue for a correla
tion between higher subject expression and verb-form ambiguity, two studies of
spoken Spanish, Barrenechea and Alonso (1977) in Buenos Aires and Enriquez
(1984) in Madrid find no statistical support for this correlation. Thus there may be
grounds for arguing that the ambiguity restriction is lost in a language-contact
situation because internal language factors do not strictly require it (note that it is
not statistically valid for A9 in Group 1 see Table 5.5).

5.5.2. Discussion
With respect to frequency of subject expression, then, English-dominant bilinguals
do not express a higher percentage of subjects overall. Crucially, however, when
they express these subjects they do not seem to follow some of the constraints that
Spanish dominant bilinguals adhere to.
As for the factors that account for the variable expression of subjects, only two
speakers in Group 3 appear to violate the coreferentiality restriction. Other factors,
though not discussed in detail in this chapter, remain significant as well. For
instance, both focal and contrastive subjects occur in Groups 2 and 3. Further
more, the semantics of the verb reveals statistically significant correlations in each
one of the three groups (p = <.oo). In all three groups the highest percentage of
subject expression correlates with estimative verbs (67 per cent, 53 per cent, 66 per
cent, mostly creer to believe and pensar to think), and with verbs of the type
gustar to like (56 per cent, 91 per cent, 80 per cent). These types of verb were
hypothesized to correlate with higher subject expression because of the implied
contrastive situation associated with estimative verbs and, in the case of gustar verbs,
because of their correlation with new information non-animate subjects.
It is also the case that across groups non-personal pronouns are much more
frequently expressed than any personal pronoun, and that plural personal pronouns
are expressed quite infrequendy (the 2pl form ustedes you was not quantified because
it occurred only five times in the nearly thirty hours of speech examined), as dis
played in Table 5.12. This table includes focal and contrastive subject pronouns,
as well as all types of sentence and clause (factors 1 1 and 12 in Appendix 3).
 Exploring External Motivation fo r Change 163
T able 5.12. Cross-tabulation of Type of Subject by Speaker Group

Group 1 Group 2 Group 3

No. % No. % No. '%

1 sg 58/139 42 58/204 28 78/229 34

2 sg 9/32 28 3/12 25 3/30 10
(tu + usted)
3 s8 24/101 24 35/112 3i 60/251 24
1 pi 6/39 15 12/79 15 4/63 6
3 P1 9/36 25 20/126 16 16/123 13
Other pronouns 19/32 59 46/90 5i 36/81 44
p <.00 P = < .00 p = <.00

The first-person singular pronoun yo I has a much higher percentage of

expression than any other pronoun in Group 1, a result which agrees with those
shown for other Spanish varieties (cf. Table 5.4). This higher percentage of occur
rences of yo I has been explained as a consequence of the egocentric nature of
verbal communication: by explicitly referring to himself, the speaker fulfils the
pragmatic need to keep himself overtly present in the verbal interaction (cf. Morales
1986: ch. 5; egocentricity would not apply in the case of nosotros we because in the
first person plural the speakers reference to himself is diffused). If the expression
of yo I has this pragmatic function, the much lower percentage occurrence of this
pronoun in Groups 2 and 3 would constitute evidence of the gradual loss of
semantic-pragmatic functions in language-contact situations.
Overall, then, it appears that the restrictions on subject expression shown to be
valid for other Spanish varieties remain to a large extent intact in Groups 2 and 3,
except for the loss of the morphological ambiguity factor. Otherwise, there is a
decrease in the strength of the coreferentiality restriction, and to a certain extent
also a weakening of the pragmatic need to refer explicitly to oneself and to the
interlocutor. Given the pragmatic functions of expressed subjects in Spanish, the
overall lower frequency of subjects in Groups 2 and 3, and the loss of the ambiguity
constraint in Group 3, converge to offer some evidence in favour of the hypothesis
that predicts the loss of semantic-pragmatic options in secondary languages.
By contrast, the possibility of an association of subject pronouns in English with
Spanish verb inflection is not supported (or we would find more subjects expressed
with neutralized verb inflections). This is also evidence against an across-the-board
type of transfer in a drive for structural parallelism. This specific study has examined
a syntactic variable which is pragmatically controlled in Spanish. Pragmatic
constraints which do not have contextual alternatives are maintained in East
Los Angeles, namely the effect of verb semantics, focality, and contrast. Such
164 Exploring External Motivation fo r Change

constraints as switch reference, the foregrounding of self, and morphological am

biguity, which may be compensated for by means other than subject expression
(for instance by the larger linguistic or extra-linguistic context), tend to be lost.
Is there evidence, then, of syntactic permeability in the data examined? The
answer is negative if permeability is defined narrowly as acts o f syntactic reception,
i.e. changes which involve not loss but additions to the grammar of a language
resulting from the external linguistic influence of a dominant language (Campbell
and Muntzel 1989: 190-1). On the other hand, a wider definition of syntactic
permeability would include the gradual loss of discourse-pragmatic restrictions,
which does produce structures with a slight foreign accent, comparable to what
happens at the phonetic level:
[54] Una noche me qued de acuerdo con una amiga. Me iba a ir a quedar a su
casa, (a) Yo no tena deseos de salir, (b) yo no tena nada que buscar en la
calle. Dej el nio encargado con una prima y (c) yo me sal [to go to her
friends]. (Li9,22,2,ELA82)
One night I agreed to go over to a friends. I was going to stay at her place,
(a) I didnt want to go out, (b) I had nothing to find in the street. I left my
son with a cousin and (<r) I went [over to her friends].
According to my own intuitions and those of six Spanish-speaking graduate
students whom I consulted, the subject yo T is necessary to clarify the subject
referent in [54a]; it is unnecesary in [54^], and unexpected in [54c']. [54] contains
a series of coreferential subjects. The situations are not explicitly contrastive, nor
is there evidence of any other pragmatic function which might justify the occur
rences of yo in [54b, c]. Thus, [54] seems to illustrate a less constrained use of
subject pronouns, a development which might ultimately lead to subjects becom
ing basically markers of person and number.
The trend towards a more fixed SV X word order, and at the same time towards
a lower frequency of subject expression, two opposing trends if English is taken
as the measure for comparison, is further proof that secondary languages do not
blindly caique the structures of a superordinate language. Rather, the explanation
for these opposing trends is more adequately found in the loss of semantic-pragmatic
constraints.
The scenario for the change would involve loss of pragmatic constraints leading
to a syntactic change. This may be comparable to phonological changes, which
may start with the spread or overgeneralization of one allophonic variant causing
the disappearance or loss of another, and eventual restructuring of the phonological
system. Weinreich (1974: 14) stated: Interference arises when a bilingual identifies
a phoneme of the secondary system with one in the primary system and, in
reproducing it, subjects it to the phonetic rules of the primary language. It seems
to me that my study offers evidence to extend this statement to the level of syntax,
as follows: Transfer arises when a bilingual (unconsciously) identifies a lexical or
 Exploring External Motivation fo r Change

syntactic structure o f the secondary system with one in the primary system and, in
reproducing it, subjects it to the semantic-pragmatic rules o f the primary language.
Consider one further example:

[55] Viajamos nosotros por tren y conocimos en el viaje bastante gente, (a) Y una
seora que - nos pusimos a platicar ella tena un beb. Y entonces el nio estaba
enfermo. Ya no me acuerdo exactamente. Se desmayaba y ella estaba
preocupada y lo tenia con medicinas. (1.28/37,2,ELA23)
We travelled by train and met a lot of people. And a lady that - we started
talking with she had a baby. And then the child was sick. I dont remember
exactly. He fainted and she was worried and was giving him medicines.

Is the italicized structure in [55] a case of syntactic transfer? Is it a violation of

a constraint on subject pronoun expression? Note that two coreferential constitu
ents, an NP, una seora a lady, and a pronoun, ella she, may both qualify as
subjects of tenia had in [55a]. This would certainly be a new construction in Spanish.
This language allows NPs that are prosodically clearly left dislocated to be
coreferential with the subject, but not two (as it were) subjects in the same sen
tence, French style: Mon pre il est venu M y dad he has arrived.
Note as well that Spanish does not introduce new referents into discourse in
subject position. The strategy consists in using haber there to be:

[56] Y haba una seora que nos pusimos a platicar, (y) ella tena un . . .
And there was a lady we started talking with; (and) she had a . . .

Indeed, Ocampo (1989: 171-2) has shown that brand-new and brand-new an
chored referents are introduced as postverbal NP arguments o f the single-valency
verbs haber there to be, estar to be, and existir to exist, a fact which, according
to Ocampo, responds to the maxim proposed by Lambrecht (cited in Ocampo, p.
172): do not introduce a new referent and talk about it in the same clause .
On the other hand, colloquial non-standard English has a similar French-style
construction:

[57] . . . and this lady she had a baby. And the baby was sick . . .

Based on the existence of examples like [57], could [55] be considered a case of
syntactic transfer, or is it rather an instance of discourse-function transfer, the
introduction of a non-salient new referent in the subject position? It appears to be
both: a new structure and its related function. Yet, what is the frequency of this
particular act of syntactic reception? In approximately twenty hours of speech
from fifteen speakers examined for this study, there are only two examples, both
produced by the same speaker (1.28/37,2; see [53] in Chapter 6). This is most
probably a nonce borrowing with no immediate consequences for the syntactic
system of Spanish.
1 66 Exploring External Motivation for Change

5.6. Conclusion

I summarize my observations about syntactic permeability as follows:

(i) Transfer may occur when bilinguals (unconsciously) identify a surface string
in the secondary system with one in the primary system and, in using it, subjet t
it to the discourse-pragmatic rules of the primjtry language. This shows th;it
languages are permeable at the discourse-pragmatic level. A similar conclusion is
reached by Prince (1992) based on a study of os-sentences in Yiddish. Accordiru
to Prince, this Yiddish construction does not alter Germanic syntax but replicate-,
the discourse function of Slavic io-sentences, i.e. the structuring of the proposi
tion into two parts, focus followed by presupposition (cf. it-cleft constructions in
English).
(ii) The intensity of discourse-pragmatic transfer may be primarily determined
by socio-cultural parameters rather than by language structure. This observation i%
based on the attested fact that longer duration of family residence in the U SA and
reduced domains of use of Spanish correlate positively with a higher degree
simplification or loss of discourse-pragmatic rules.
(iii) Transfer may occur at the lexico-syntactic level when bilinguals (uncon
sciously) identify a lexical item in the secondary language with one in the primary
language and subject it to the subcategorization of the item in the primary language.
I consider this an instance of lexico-syntactic permeability, and discuss it in Chapter
6.
(iv) The syntactic permeability of grammars is at first evident only in nonce
syntactic borrowings, i.e. impermanent and occasional instances of immediate
disturbances of the surface syntactic structures/strings of a language. Given suf
ficient time depth and the appropriate socio-cultural conditions surrounding contact,
nonce syntactic borrowings may become incorporated in the system of the borrowing
language. An instance of this type of outcome may be illustrated by the grammatical
convergence exhibited by four languages in contact in Kupwar (India), two Dravidian
languages (Kannada and Telugu), and two Indo-Aryan languages (Urdu and
Marathi) (Gumperz and Wilson 1977).
(v) Under conditions of normal transmission (Thomason and Kaufman 1988),
nonce syntactic transfer will spread with enormous difficulty in the language
system and across speakers (if at all; see my brief discussion of Gumperz and
Wilsons study below and also Chapter 7). The syntactic ystem of grammars is
remarkably impermeable to foreign influence.
Considering the fact that I have dealt with two European languages which have
many surface similarities, and that the contact situation is uniquely favourable to
the immigrant language (the uninterrupted flow of large numbers of Spanish
speaking immigrants, for instance), my conclusions would appear rather bold.
Indeed, in order to reach conclusions of convincing generality, it would be necessary
to carry out close studies of dissimilar as well as similar languages in comparable
 Exploring External Motivation for Change

l ontact situations. One such study has been conducted by Mougeon and Beniak
(1991); they also note that direct (overt in their terminology) grammatical transfer
is rare, but the situation they examine also involves two European languages,
I .nglish and French in Ontario.
My observations find some encouraging support, furthermore, in Weinreichs
remark, based on a survey of a considerable number of unrelated bilingual and
multilingual situations, that foreign grammatical elements are only very rarely
incorporated in the language as a code (1974: 44).
On the other hand, Gumperz and Wilson (1977: 151) argue that their study of
Indo-Aryan/Dravidian language contact in India offers evidence that borrowing
extends to all aspects of the grammatical systems. Note, however, that these
authors study the outcome of contact that has existed for more than six centuries
in the case of Kannada and Marathi, and that Urdu-speakers have been in the
region of Kupwar for more than three centuries. In addition, Gumperz and Wilson
conclude that almost all the changes can be interpreted as reductions or general
izations that simplify surface structure in relation to underlying categories and
relationships (164; my emphasis) It appears, then, that the creole-like present
varieties of Kannada, Marathi, Urdu, and Telugu spoken in Kupwar have not
really borrowed or introduced new syntactic rules into their systems, but rather
essentially lost existing variants and retained or extended those that paralleled
structures in the other languages. If this were so, the Kupwar contact situation
would offer strong evidence from dissimilar languages in support of my observations.
If, on the other hand, the grammar of the creole-like varieties is radically differ
ent from that of the corresponding standard varieties, it would be necessary to
investigate at what point in the centuries of development of these varieties these
radical changes were introduced.
In sum, it seems to me that, if Thomason and Kaufmans (1988: 14) statement
that any linguistic feature can be transferred from any language to any other
language refers to the introduction of foreign elements into a language code, it
needs to be qualified with respect to syntactic features as follows: under conditions
of normal transmission, any linguistic feature from language S may end up being
present in language F after a long process of step-by-step changes in language F
such that at any given step the change involved will not constitute a radical
modification of the system of F but rather the extension of an already existing
syntactic variant (i.e. the modification is compatible with the structure of the
borrowing language). It seems to me that this chain-like process may advance
when a syntactic variant loses its status of marked (in terms of frequency and/or
pragmatic function), thus opening the door for further structural changes to take
place. I must note, however, that I have examined a rather narrow range of syn
tactic phenomena the possibility of null arguments and word order and that I
consider verb subcategorization a lexico-semantic question, albeit with syntactic
consequences (see Chapter 6).
 6

Conflicting Loyalties

6.1. Introduction

W e i n r e i c h (1974: 68) proposes a definition of language shift as the change from

the everyday, traditional use of one language to that of another. He also proposes
the term language loyalty to describe a state of mind in which a language, in
opposition to other languages, is assigned a high value by its speakers, who art-
driven to defend this language from foreign interference and/or from an imminent
language shift. Language loyalty is defined as a principle . . . in the name of which
people will rally themselves and their fellow speakers consciously and explicitly 10
resist changes in either the functions of their language (as a result of a language
shift) or in the structure or vocabulary (as a consequence of interference) (99).
I have no evidence (as could have been provided by extended observation or
direct questioning, for instance) that the speakers included in this study con
sciously or explicitly resist transfer from English or shift to English. Most of them
have expressed, however, spontaneously in conversation or in response to direct
questions, positive attitudes towards Spanish as well as willingness to maintain it
and pass it on to their descendants. I refer to these expressions as acts of loyalty
and examine them in this chapter. This scrutiny shows that acts of loyalty are in
conflict with uncommitted behaviour (i.e. lack of commitment to turn positive
attitudes into action) in most of the speakers in Groups 2 and 3. In addition, the
preceding chapters have offered evidence of transfer, although mainly of the indi
rect type. In this chapter, I present and briefly discuss uses of lexical items by
bilinguals in Groups 2 and 3 which deviate from the norms of Group 1 speakers.
I apply the phrase acts of shifting to these uses because they appear to be the
consequence of direct transfer of forms and/or form meanings from English. Acts
of shifting are apparently in conflict with acts of loyalty, but in fact they seem to
me to represent a successful strategy for the maintenance of Spanish among
second- and third-generation immigrants in Los Angeles.
In Chapter 1 (p. 4) I listed four types of phenomenon as illustrations of potential
transfer, not all of which are necessarily present in every case of contact. Briefly,
these include: (1) The replacement of a form in language S with a form from
language F, or the incorporation from language F into language S of a form pre
viously absent in S. This constitutes an instance of direct transfer. (2) The in
corporation of the meaning of a form R from language F, which may be part
of the meaning of a form P in S, into another form, structurally or semantically
similar to R, in system S. This is an instance of direct semantic transfer. (3) The
 Conflicting Loyalties
higher frequency of use of a form in language S, determined on the basis of a
comparison with more conservative internal community norms, in contexts where
a partially corresponding form in language F is used either categorically or pre
ferentially. (4) The loss of a category or a form in language S which does not have
a parallel category or form in the system of F. (3) and (4) are instances of indirect
transfer.
The preceding chapters, however, have shown evidence against direct influence
from English on a number of aspects of Spanish grammar: patterns of simplification
of the tense-mood-aspect system, omission and simplification of verbal clitic pro
nouns, non-expression of the complementizer que, extension of estar, expression of
subjects, subject position relative to the verb. None of these phenomena constitute
a clear illustration of transfer of types (1) or (4), i.e. no replacement or incorpor
ation of forms and no loss of categories are explainable as the result of contact
specifically with the English language. Rather, these modifications, when they
occur, seem to be the consequence of processes of simplification constrained by
the grammar as well as by the social environment and reduced communicative
functions of the affected language.
The extension of discourse-pragmatic functions discussed in Chapter 5 (syntac
tic permeability of type a, p. 135), on the other hand, corresponds to the type of
transfer (*2) that I describe as the incorporation of the meaning of a form R from
language F into a structurally or semantically similar form in system S. At the
syntactic level, this type of transfer does not appear to be frequent. By contrast, the
loss of semantic and pragmatic constraints (which in some cases may be argued to
be of type 4) is evident in the linguistic phenomena examined, but the motivation
for this loss cannot be definitely ascribed to direct influence from English. Indirect
transfer, apparent in the higher frequency of use of parallel forms (3), is the only
type decidedly supported by the data examined in the previous chapters.
Where, then, do we find evidence of direct transfer of types 1 and 2? As already
suggested in Chapter 1, evidence of direct transfer abounds at the level of the
lexicon, which is affected in various manners. In sections 6.2 and 6.3, 1 discuss the
different types of replacement and incorporation o f lexical forms and/or meanings
and their occurrence across the Spanish proficiency continuum. Sections 6.4 and
6.5 discuss acts of conflicting loyalties, i.e. both spontaneous and elicited attitudes
towards Spanish.

6.2. Acts o f Shifting: Modelling

Weinreich (1974: 47) identifies three types of lexical interference: transfer o f form
(presumably with their corresponding meaning), transfer of meaning only, and
transfer of form and meaning in the case of compound lexical items. He further
subdivides this third type (pp. 50-3) into three more groups, one of which he
refers to as semantic extensions or loan translations, defined as reproduction in terms
170 Conflicting Loyalties
of equivalent native words [which] can be carried out with compounds, phrase-,,
and even larger units such as proverbs (p. 50). Both Otheguy (1993), and Otheguv
et al. (1989) (OGF) show the terminological and conceptual problems involved
in Weinreichs classification. Furthermore, Otheguy (1993) forcibly argues that
Weinreichs notion of loan translation was not properly developed to account for
modelling of compound lexical items: by defining loan translations as unusual
combinations of words, Weinreich created a theoretical impasse in which it became
impossible to distinguish between unusual things to say and unusual linguistic
means with which to say them (p. 36).
In view of the difficulties posed in particular by Weinreichs analysis of possible
cases of multiple-word modelling, O GF propose to set these aside, and to deal
with individual words only. Single-word transfers must further be differentiated
from switches to English. O GF do this by applying a simpler version of the model
developed for this purpose by Poplack (1987) and Poplack et al. (1989). These
authors have established a set of criteria to distinguish switches from borrowings,
e.g. level of phonological assimilation, level of social integration, discourse function,
frequency of occurrence. Of these, O GF take into account the phonological factor
only (they state that the investigation of the other criteria is beyond the scope of
their paper), and count all phonologically integrated single-word items as cases of
transferring.
Thus, their classification distinguishes between items that preserve English
phonology (single-word switches), and those that have been adapted to Spanish
phonology (single-word borrowings). In this latter group, they differentiate
between loans, the transferring of forms with their meanings (e.g. puchar, which
replaces empujar to push), and caiques, the transferring of meanings only (e.g.
grados degrees extends its meaning to incorporate one of the meanings of English
grades marks showing a students level of achievement).
O G Fs is a well thought-out and useful taxonomy of borrowings that allows
them to quantify the rate of increase of anglicisms between two generations of
Cuban-Americans (residing in the town of West New York, New Jersey), as well
as to conclude that second-generation Cubans use numerous duplicating
calquewords that displace existing, traditional Spanish message formulations (grados
displaces notas, for instance), an indication that their command of the Spanish
lexicon is diminishing. O G Fs classification into loans and caiques corresponds
to Otheguys (1993) distinction in terms of transferring and modelling, and it
corresponds also to the possible types of transfer phenomenon that I describe
under types 1 and 2 above (see also Chapter 1). Therefore, I adopt the distinction
between additions to the inventory of the impacted variety (i.e. loans, transferring,
type 1 transfer) and (semantic) modifications of the existing inventory (caiques,
modelling, type 2 transfer) in my discussion of lexical transfer, and extend it, with
some adjustments, to apply to multiple-word cases of loans and caiques. My main
interest is to identify and discuss some types of single or multiple-word caique that
are attested at different points in the bilingual proficiency continuum.
 Conflicting Loyalties 17 1

Otheguy (1993) correctly notes that the unusual combinations of words that
are generally considered loan translations form a heterogeneous set including (a)
items that do not constitute a linguistic innovation of any kind, but rather com
binations which at most represent a cultural or pragmatic innovation (e.g. E l da de
dar gracias Thanksgiving Day and mquina de contestar answering machine,
instead of contestador automtico, lit.: automatic answerer), as well as (b) items in
which the innovation results from the establishment of novel metaphorical con
ceptualizations. An example of this last group is llamar para atrs to call back,
where atrs behind/to the back/backwards is claimed to have been metaphorically
extended to incorporate the concept of repetition. Nevertheless, Otheguy admits
that the English construction (to go/pay/send/ back, etc.) may have motivated the
metaphorical extension of the concept of behindness to the temporal notion of
repetition in Spanish. I return to this example later.
Based on an analysis of the Los Angeles data, I distinguish the following types
of what are generally considered to be borrowings:

6.2.1. Single-Word Loans

This category consists of the transferring of forms with their meanings (e.g. bil from
bill, in the sense of payment owed, mapear or mopear from to mop, puchar from
to push, sinque from sink, so from so (so is a loan that replaces the Spanish
conjunction as que even in the speech of Group 1 speakers. It is a stable, wide
spread loan in L A Spanish), suiche from switch, troca or troque from truck, all
attested in Groups 1, 2, and 3).

6.2.2. Single-Word Caiques

This category consists of the transferring of meanings into an already existing
lexical item (e.g. parientes relatives extends its meaning to incorporate the mean
ing of English parents, attested in Group 3). Many other single-word caiques are
attested in the three groups, for instance aplicacin application, including making
a request (Span.1 solicitud application, in the sense of making a request), grados
grades (marks showing a students level of achievement) (Span, notas), papel
(news)paper (Span, diario/peridico), moverse to move [to another house] (Span.
cambiarse/mudarse de casa), carpeta carpet (Span, alfombra/moqueta), etc.

6.2.j . Multiple-Word Calques that Do Not Alter Semantic and/or

Grammatical Features

Examples are das de semana weekdays (Span, das de trabajo), mquina de contestar
answering machine (Span, contestador automtico), patio de juegos playground

1 T h e abbreviation Span, used in the examples stands for general standard or popular Latin Ameri
can Spanish.
172 Conflicting Loyalties

(Span, patio (de escuela) ), tarjeta de plstico plastic (i.e. credit card, Span, tarjeta/
tarjeta de crdito).
In what follows, I have chosen to disregard single-word borrowings and the
types of unusual combination that do not alter the semantic and syntactic re
sources of Spanish (cf. Otheguy, 1993), except to note here that they are attested
in the speech samples from all the speakers under study. Instead, I focus on two
types of caique: caiques of bound collocations, idioms, and proverbs, and lexico-
syntactic caiques (defined in Sections 6.2, 4 -5 below), that is, on multiple-word
caiques that have lexico-semantic and grammatical consequences on the replica (or
borrowing) language.
It will be obvious to the reader that some of the combinations included as lexico-
syntactic caiques could perhaps have been considered a bound collocation or at
times even an idiom. Indeed, the borderline between all these forms of collocations
is often rather tenuous, as is also the type of transfer and the effect that it may have
on the replica language.2 This section is, therefore, more an approximation to the
difficult question of multiple-word modelling than a definitive proposal for its
solution.

6.2.4. Caiques o f Bound Collocations, Idioms, and Proverbs

A bound collocation, idiom, or proverb from the source language is reproduced
exactly with lexical units from the replica language. The caique may alter semantic
and/or collocational features of the replica language, but these modifications re
main restricted to the phrase in question. (This section does not discuss proverbs,
which are not attested in the Los Angeles data.)
1 am acutely aware of the difficulties involved in the definition and identification
of collocations and idioms (see e.g. lengthy discussions in Cruse 1986; Jackson
1988; Ruhl 1989). Cruse defines the term idiom as a lexical complex which
is semantically simplex (p. 37), and illustrates it with to kick the bucket, to pull
someones leg, to spill the beans. The term collocation, on the other hand, is seman
tically transparent (i.e. divisible into semantic constituents) and it refers to strings
of lexical items which regularly co-occur. Bound collocations, like foot the bill and
curry favour, are idiom-like in that their constituents do not like to be separated
(Cruse 1986: 41), yet they are not idioms because they are, like collocations,
semantically transparent. However, the semantic transparency of foot or curry
is debatable, since it seems that neither is otherwise used in the way in which they
are in the collocations foot the bill or curry favour.
Many word combinations may not be classified neatly into one or another type,
but the question of precise classification does not concern me, since the borderline
between a bound collocation and an idiom seems to me difficult to resolve. What
I consider important about the calquing of either bound collocations or idioms
2 Similar difficulties have probably stopped Otheguy et al. (1989: 43-4) from examining multiple-
word transfers, and may have also prompted their observation that the only items in which clear
evidence o f modeling can be discerned are individual words.
 Conflicting Loyalties 173
is that they are comparable to a single-word caique. This is because, even though
they consist of more than one word, in practice they display the internal cohesion
expected of single words. Thus, the internal arrangement of words in a bound
collocation or idiom calqued by a replica language might in principle violate the
syntactic rules of this language, yet I do not consider it a syntactic transfer or
innovation unless there is clear evidence that the new pattern is being extended
beyond the specific context of the caique. It is quite likely in principle, however,
that innovative syntactic patterns introduced in bound collocations or idioms may
in time become context-free and productive, in a manner similar to what has been
proposed for the borrowing of bound morphemes (cf. Weinreich 1974: 31).
Consider [1-3 ], attested in Groups 2 and 3, which illustrate what may be bound
collocations or idioms:

[1] Estoy quebrada Im broke (Li9,f22,2,ELAs4)

Span.: No tengo un peso/Estoy en bancarrota
lit.: (I) dont have one peso/(I) am in bankrupt
[2] OK. Eso est bien conmigo (Ri7,f2i,2,ELA3)
OK. Thats fine with me
Span.: OK. (Eso) est bien
OK. Me parece bien
lit.: OK. (It) to-me seems fine
[3] So l sabr si se cambia su mente. (D39,f28,3,ELA42)
lit.: so hell know if (he) se changes his mind1
Span.: Asi que l sabr si cambia de opinin.
lit.: so that hell know if (he) changes o f opinion

In Spanish, the verb quebrar may mean to go bankrupt, and it may be used as
in [4]. The corresponding adjectival expression is en bancarrota or en quiebra (Mara
est en quiebra/en bancarrota Mary is bankrupt). The adjective quebrada broken
refers to fracture, not to financial status as it does in [1], which replicates the
English collocation. It has been brought to my attention, however, that in Mexican
Spanish the state of being broke may be attributed to a human entity with the verb
andar to go around (Ando quebrado Im broke). Thus, it is possible that the
combination in [1] may involve the substitution of estar to be for andar (another
instance of the extension of estar discussed in Chapter 4) rather than the caique of
a bound collocation.

[4] Esa empresa/Mara quebr/ha quebrado.

That company/Mary broke (went bankrupt)/has broken
As indicated by the general Spanish interpretations, the unusual elements in
[2] are conmigo with me and the expression of the subject eso that, which though
3 T h e use o f se in [3] reflects an independent phenomenon o f extension o f reflexive-type verbs (see
Ch. 4).
174 Conflicting Loyalties

possible, is highly unlikely to be said. Example [2] is a word-for-word translation

of the corresponding English construction.
The word mente mind in [3] may be said to have extended its meaning from
memory, recollection, intelligence to incorporate one of the senses of mind in
English, namely opinion, view, intention. No examples are attested of the type of
[5], for instance, another possible context for the use of mente mind with its
extended meaning. *

[5] Soy de mente diferente

Im of a different mind

In general Spanish, the sense of Im of a different mind would be expressed

with tener ideas/ opiniones diferentes to have different ideas/opinions or pensar de
manera diferente to think differently. The semantic extension of mente mind, then,
occurs only in the bound collocation cambiar su mente to change ones mind .

6.2.5. Lexico-syntactic Caiques

These are multiple-word caiques that alter semantic and/or grammatical features
of the replica language. Furthermore, these caiques seem clearly separable into six
classes on the basis of the semantic and/or grammatical effect of the transfer or
Spanish.

6.2.5./. Type 1

A multiple-word unit which reproduces one in the source language brings about a
change in the meaning of a word in the replica language. The modified word with
its extended meaning or meanings does not remain restricted to be used in only
one set combination:

[6] Es un modo de tener un buen tiempo. (039/28,3,ELA42)

lit.: (it)s one way to have a good time
Span.: Es un modo de pasar un buen momento/pasarlo bien
lit.: (it)s one way to pass a good moment/pass it well
[7] . . . pero cuando lleg el tiempo que ellos ya queran sus carritos (H48,
m39,3,ELA 23)
lit.: but when arrived the time that they already wanted their cars
Span.: pero cuando lleg el momento/la hora que ellos ya queran sus carritos
lit.: but when arrived the moment/the hour. . .
[8] porque otro tiempo - - ando en el carro - - y empec a notar que no estaba
bien (Li9,f22,2,ELAs4)
lit.: because another time - - (I) go in the car - - and (I) started to notice that
(it) was not fine
T
Conflicting Loyalties 175
Span.: porque otra vez/en otra ocasin . . .
lit.: because another instance/on another occasion . . .

The word tiempo is quite vulnerable in a SpanishEnglish contact situation

because many but not all the uses o f English time overlap with it. This situation
probably favours the extension of tiempo, basically a durative mass concept in
Spanish, to incorporate the notion of one specific point: one of several instances,
an occasion, an hour, a moment ([6-8]). In Spanish the phrase buen/mal tiempo,
in the singular, can only mean good/bad weather. The word tiempo may occur in
the plural form, as in buenos/ malos tiempos, with the meaning of good/bad periods
of time, i.e. it retains the meaning o f duration.
The often quoted case of para atrs is also an illustration of a lxico-syntactic
caique of type 1. The phrase para atrs collocates with a number of verbs of
movement in Spanish: mirar para atrs look back, pasar (X ) para atrs pass (X)
to the back, caminar para atrs to walk backwards . In these collocations, atrs
conveys its general, standard, traditional locational meaning of to the back, be
hind, back, backwards.
In Los Angeles Spanish, as in all other varieties of U S Spanish, para atrs has
extended its meaning to include not strictly repetition as Otheguy (1993) suggests,
but rather in reply or in return, which back has in such English verb phrases
as to call back, to come back, to send back. The extension of para atrs to include
the concept of repetition or return allows it to occur with verbs that did not allow
this combinational possibility. Examples attested in my data include dar para atrs
to give back, llamar para atrs to call back, pagar para atrs to pay back, decir
para atrs to answer back, regresar a alguien para atrs to make someone go home,
ir para atrs to go back, traer a alguien para atrs to bring someone back. I f the
extension of the meaning of para atrs were attested in Spanish varieties clearly not
in contact with English (to my knowledge, it has not been attested), one could
more forcefully argue that the mechanism for the change was purely a metaphori
cal extension. If not, transfer of meaning from English would seem to be a more
appropriate explanation.

6.2.5.2. Type 2
A multiple-word unit which reproduces one in the source language brings about
a change in combinational restrictions related to, among other things, animacy of
the constituents, tense or aspect of the verb, and other semantic constraints on the
constituents of the corresponding construction in the replica language. The caique
affects the meaning of a word included in this construction. Consider [9, 10],
produced by speakers in Group 2:
[9] Y tu carro que compraste, cmo te gustai (H22,m2i,2,ELA54)
lit.: and your car that (you) bought, how to-you pleases?
Span.: Y tu carro que compraste, te gustai
 Conflicting Loyalties
lit.: and your car that (you) bought, to-you pleases?
and the car you bought, how do you like it?
[10] Como te gusto [la pelicula]? (A2o,fi9,2,ELA29)
lit.: how to-you pleased [the movie]?
Span.: Te gusto [la pelicula]?
lit.: to-you pleased [the movie]?
how did you like it [the movie]?

In Spanish the interrogative word of manner cbmo does not have the meaning
of to what extent, amount, or degree that how may have in English. Therefore,
as indicated in the general Spanish versions, questions about the extent to which
something is or was liked, as in [9, 10], should not include the interrogative word
como, since como-questions must be answered with a description of the manner in
which the situation obtains. This interlinguistic difference in the meaning of conm
and how accounts for the acceptability of [11b, c] as possible answers to [110] in
English, for the unacceptability of the corresponding exchange, [12a, b\, in
Spanish, and for the acceptability of [12*:], which describes in what manner or
condition soup is liked, as an answer to [12a].

[II] (a) How do you like the soup?

(b) I like it, its good
to I like it very hot
[12] (a) Cmo te gusta la sopa?
ib) *Me gusta, est buena
I like it, its good
to Me gusta bien caliente
I like it very hot

It is clear, then, that the syntactic structure cmo X gustar Y exists in Spanish,
but since cmo does not include the meaning to what extent, its distribution is
more restricted than English in this respect. Furthermore, neither English nor
Spanish may ask about the manner in which a specific Y is or was liked. When
cmo or how is used with the sense of in what manner, certain restrictions are
imposed on the semantics of the subject (Y), which must be non-specific, and on
the aspect of the verb, which must be imperfective (see the unacceptability of [13c]
in both languages). The less restricted meaning of cmo in L A Spanish results in
less restricted syntax, but the appropriate answer refers to the extent of the liking
([13^]), not to the manner ([13c]). The answer given to the question in [9], ex
panded here as [13a, b], illustrates; [ 1 3 , unacceptable in English, is not attested
in L A Spanish either.

[13] (a) Y tu carro que compraste, cmo te gusta? (H22,m21 ,2,ELA54)

and the car you bought, how do you like it?
 Conflicting Loyalties 177

(b) Mi carro me encanta (L 19,f22,2,ELA54)

I love my car
(c) *Me gusta amplio
*1 like it spacious
The semantically less restricted meaning of cmo how in L A Spanish is also
extended to a declarative context, to the complement of saber to know. Here again,
the cmo in the complement of saber to know refers to the manner of a specific
situation, as in [14], which could be appropriately followed up by instructions on
how to, as in [14b]. Note, however, that the wider meaning of how in English allows
it to occur with to know when only a general knowledge or skill required to do
something is being referred to, as in [15]. While in general Spanish [15] must be
translated without cmo how, this complementizer is acceptable in L A Spanish,
as shown by [16].
[14] (a) Mara no sabe cmo irse a la universidad
Mary doesnt know how to go to the university
(b) Dile que tome el bus 73
Tell her to take bus 73
[15] He doesnt know how to read yet
No sabe 0 leer todava
[16] S saba como hablar espaol (039/28,3,ELA42)
lit.: yes knew-3sg how speak-Inf Spanish
Span.: S saba 0 hablar espaol
He did know how to speak Spanish
Another example of a lexico-syntactic caique of type 2 is offered by sentences
related to measurements of weight, age, etc.:
[17] Mi padre es seis pies [de altura]
lit.: M y dad is six feet [tall]
Span.: Mi padre mide seis pies
lit.: M y dad measures six feet
In Spanish ser may select a predicate nominal with the feature [+ specific unit
of measurement] only if the subject encodes the measured category (El largo es cinco
pies The length is five feet, Los aos de estudio son cinco The years of study are
five, La edad para empezar es seis aos The age to begin is six years old). Calquing
the use of to be to attribute a specific unit of measurement to any type of entity
brings about the loss of a selectional restriction on the subject NP. Thus in LA
Spanish one finds such examples as [17], with the copula ser to be relating
a specific unit of weight, age, or height to a human subject referent. Examples
[18 -19 ] illustrate further:
[18] [la casa] ahora es cien mil o ciento veinte mil dlares (H22,m 21,2,E L A i 1)
Span.:. . . cuesta costs'
 Conflicting Loyalties

Now it [the house] is one hundred or one hundred and twenty thousand
dollars
[19] [la planta] ya es como dos, tres aos (V2i,fi8,2,ELA67)
Span.:. .. tiene has
By now it [the plant] is about two, three years old

6.2.5. j. Type 3
A prepositional phrase in the source language is reproduced with lexical units of
the replica language, but the preposition is not the one used in the replica language
(e.g. in popular or standard Latin American Spanish). Rather, it corresponds tu
the preposition used in the parallel source-language phrase (e.g. in English).
Cases of missing prepositions and of uses of prepositions that do not comph
with the norm in popular or standard Latin American Spanish are fairly frequent
in the data from Groups 2 and 3, an expected condition given the semantic
complexity of prepositions in Spanish (as well as in English). The fact that almost
every sentence contains a preposition, and thus a possible site for an error or
deviation from the norm, appears to be a strong factor in creating the impression,
exaggerated in my view, that the syntax of those in Groups 2 and 3 is quite
different from those in Group 1. In any case, in most cases these modification
cannot be attributed to direct influence from English, as shown by [20-2].
[20] y 0 la maana habl 0 el trabajo (A4,f3i,3,ELA2o)
lit.: and 0 the morning spoke~3sg 0 the work
Span.: y en la maana llam del trabajo
and in the morning he called from work
[21] Qu son tus planes del futuro? (R24,m20,2,ELA50)
lit.: what are your plans of-the future?
Span.: Cules son tus planes para el futuro?
What are your plans for the future?
[22] ya empez su tercer ao de medicina con U C LA (R24,m2o,2,ELAso)
lit.: just started his third year of medicine with U C LA
Span.: ya empez su tercer ao de medicina en U C LA
hes just started his third year of medicine at U C LA 4
There are, on the other hand, a number of instances in which the [preposition
+ NP] in English may be considered a collocation (cf. Cruse 1986: 40), in the sense
that either only one preposition is allowed to occur in that particular construction
or one of several possible prepositions is the one that most frequently co-occurs with
the noun or NP in question.5 If the corresponding construction in L A Spanish
4 A native speaker o f English informs me that it would be possible to say ''with U C L A in a different
discourse context, i.e. i f the person had been studying at another university and had changed to U C L A .
T h is is not the case in [22].
5 I conducted an informal survey o f 9 native non-Hispanic speakers o f English to examine preposi
tional phrase collocations. These native speakers were told that prepositions included such little words
 Conflicting Loyalties 179
reproduces the English preposition either instead of the expected one in Spanish
or when no preposition is required in Spanish, I consider it an instance of a lexico-
syntactic caique:
[23] para llegar all en tiempo (B33,fi9,3,ELA29)
lit.: to arrive there on time
Span.: para llegar all a tiempo
[24] ella estaba en sus rodillas (A2o,fi9,2,ELA29)
lit.: she was on her knees
Span.: ella estaba de rodillas
lit.: she was o f knees
[25] Yo voy a una parte en los jueves (V 2i,fi8,2,ELAi6)
lit.: I go to a place on the Thursdays
Span. : Yo tengo un compromiso $ los jueves
I have a commitment on Thursdays
[26] no haba nios a cuidar (M47,f33,3,ELA52)
lit.: no there were children to take care
Span.: no haba nios para/que cuidar
there were no children to take care of
[27] y le echaron como veinte pescados a una vez (R24m20,2,ELA50)
lit.:. . . at one time
Span.:. . . de una vez
lit.: o f one time
and they threw him like about twenty fish at once

(>.2.5.4. Type 4
The subcategorization of a verb in the source language is transferred to the corre
sponding one in the replica language. Changes in subcategorization occur when
the replica language {a) reproduces the syntactic-semantic relationship of the
arguments of the verb in the source language ([29-34]); (b) leaves the required
preposition out when none is required in the source language ([35]); (c) reproduces
the preposition that collocates with the source-language verb with a formally and/

as to, at, under, over, in, on, from, for, above, with, against, without, etc. . I then asked them to tell me
the first preposition that came to mind after hearing the examples I would give them, and illustrated
the test with a couple o f examples. T he phrases given as stimulus and the prepositions supplied most
frequently by the speakers consulted were the following (number o f times is given in parentheses):
in (9) the room, on (8) the road, between (4) you and me, on (6) time, with (4)/to (4) care, with (4) the
baby, in (6) school, on (5) his knees, in (8) January, on (5) his neck, on (4) the corner, for (5) you, to (4)/
on (4) the left, with (8) sugar, on (8) vacation, in (4) the stars, on (8) Sunday. I was particularly interested
in the 5 examples italicized. T his informal survey supports m y claim that these words are strongly
associated with the preposition on in English, while in Spanish they collocate more frequently with
prepositions that do not usually translate as on (e.g. a at, de o f ). Group 2 and 3 speakers, however,
frequently use en on/in rather than a or de.
 Conflicting Loyalties
or semantically similar one in the replica language, whether or not a preposition is
required in this language ([36, 37]); or (d) reproduces the valency of a given verb
in the source language ([38]).
The most quoted example of this type of lexico-syntactic caique in books and
articles dealing with English influence on Spanish concerns the verb gustar to like.
In Spanish, gustar to like is subcategorized for a subject with the semantic role of
theme or patient and an indirect object (introduced by a and with an obligatory
coreferential verbal clitic le!les) which has the semantic role of experiencer. The
opposite syntactic-semantic relationship holds in modern English,6 as the examples
below show.

[28] A nadie le gusta ellaq where i = IO = experiencer

q = subject = theme
lit.: to no one him likes sheq
No one likes herq where i = subject = experiencer
q = DO = theme
Speakers in Groups 2 and 3 caique the syntactic-semantic structure of to like,
although not exactly. Indeed, while the experiencer is coded in the nominative as
a subject, in many of the examples attested in the data the theme is not in the
accusative as in English. Instead, it is coded as an indirect object, i.e. it is intro
duced by a (variably in the case of infinitival clauses), and it may appear with a
coreferential dative Cl, as illustrated in [29-32]. The only instances of exact calquing,
i.e. the theme coded as a direct object, with no coreferential Cl and no , are
produced by the two speakers at the bottom of the continuum ([33, 34]).
[29] Se llama la Sra X , pero naden [sic] le gusta - a ella7 (V 2i,fi8,2,ELAi6)
lit.: se calls the Mrs X , but no one him/her - likes to her
Span.: Se llama la Sra X, pero a nadie le gusta - ella
Her name is Mrs X , but no one likes her
[30] Los cocodrilos les gustaron a matar (R.24,m2o,2,ELA5o)
lit.: the crocodiles them-liked-3pl to kill
Span.: A los cocodrilos les gustaba-3sg matar
Crocodiles liked to kill
[31] No nos gustbamos trabajar (D3,m45,3,ELA43)
lit.: not us-liked-1 pi work
Span.: no nos gustaba~3sg trabajar
we didnt like to work
[32] Ahorita le gusta a - a todo (M47,f33,3,ELA52)
lit.: now her - likes to - to everything

6 To like and other verbs (e.g. to think) patterned like modern Spanish in medieval English.
7 T he speaker pauses briefly after gusta. Otherwise, it would have been impossible to tell whether an
indirect object marker a had been produced in this example.
 Conflicting Loyalties

Span.: Ahorita le gusta todo

Now she likes everything
[3 3 ] Y gusta golf mucho8 (R42,mi5,3,ELA3o)
lit.: and likes golf much/a lot
Span.: y le gusta mucho el golf
and he likes golf a lot
[34] Yo gusto eso (N4o,f2i,3,ELA48)
lit.: I like-isg that
Span.: A m me gusta-3sg eso
I like that
Further examples of changes in subcategorization are offered by [35, 36]. [35]
omits the required preposition, thus changing the subcategorization of entrar to go
into, to become a member of, intransitive in Spanish, from verb followed by
prepositional phrase, [___ + PP], to the transitive structure verb followed by NP,
[___ + NP]. I have no evidence that this extension allows this verb to occur with
a Cl, as in lo entras por la izquierda you enter it from the left, but the transitive
structure opens up this possibility.
[35] y entras 0 el washroom (838/19,3,E LA 31)
lit.: and enter-2sg the washroom
Span.: y entras a/en el washroom
and you enter the washroom
[36] reproduces the preposition that collocates with the English verb to have respect
for someone, roughly with the structure [V NP(DO) PP]. This construction translates
into Spanish tenerle respeto a alguien [V NP(DO) PP(IO)], with an indirect object
and its corresponding dative clitic. The use of para instead of a coincides with a
change in syntax in Spanish, inasmuch as prepositional phrases with para for do
not require a coreferential clitic (0 tengo unos libros para ellos I have some books
for them as against les tengo unos libros a ellos I have some books for them). Cal-
quing the preposition para in an example like [36] opens up the possibility for
this syntactic change in LA Spanish. Note that the preposition change illustrated
by [36], produced by a speaker in Group 2, appears to have only a lexical conse
quence at this stage since the Cl is not blocked. The seed has been planted,
however, for a consequent lexico-syntactic modification which could involve either
blocking the Cl or allowing Cls coreferential with the NP in a para for preposi
tional phrase.
[36] Nosotros les tenamos respeto para ellos y todo eso (R 17 /2 1 ,2,ELA3)
lit.: we them had respect for them and all that
Span.: Nosotros les tenamos respeto (a ellos) y todo eso
We had respect for them and all that
8 Because this person speaks slowly, I feel confident that there was no a after gusta.
182 Conflicting Loyalties

Examples involving the substitution of por for for a to that bring about a lexico-
syntactic modification and consequent blocking of the Cl have been reported
(Franco, n.d.) as occurring with such verbs as buscar a to look for (someone), and
esperar a to wait for (someone). In US Spanish these verbs have changed from a
structure [V + NP] to [V + PP], with the consequent blocking of Cl formation:

[37] 0 estoy esperando por ella

lit.: am waiting for her
Span.: /estoy esperando (a ella)
Im waiting for her

Finally, in [38] visitar to visit, a two-argument verb in Spanish, reproduces the

valency of the verb that translates it in English, which may be used intransiti vely
in this language.

[38] El me dijo - cuando mi familia fue all al Texas a visitar (036,11145,3.

ELA43)
lit.: he told me - when my family went there to the Texas to visit
Span.: El me dijo - cuando mi familia fue all a Texas a visitar a los parientes/
de visita.
He told me - when my family went there to Texas to visit (relatives) / on a
visit

6.2.,5.5. Type 5

A pragmatically neutral word order in the source language is reproduced with

lexical units of the replica language. (A construction is considered to be pragmat
ically neutral when its function involves the conveying of information with an
informational intonation contour (Silva-Corvaln 1983^), without any special con
notations of surprise, contrast, sarcasm, etc.) The word order in question is either
not allowed in any non-literary style or occurs under very marked pragmatic
conditions in the oral mode in the replica language (e.g. contrast, contrary to
expectation, sarcasm). In the context where the caique occurs, there is no indica
tion that these pragmatic conditions obtain. Word-order caiques are illustrated in
[3943] (see also section 5.4 above).

[39] Ella hablaba como yo ms o menos, machucado espaol, mitad las palabras ingls
y mitad, palabras espaol (R5o,m46,3,ELA3)
Span.: . . . espaol machucado . ..
She spoke like me more or less, chopped up Spanish, half in English and half
in Spanish
[40] es el nmero uno gastador de petrleo (039/28,3,ELA42)
Span.:. . . gastador nmero uno .. .
its the number one user o f oil
 Conflicting Loyalties 183
[41] si cuatro otros alumnos (838/19,3,ELA30)
Span.: si otros cuatro alumnos
if four other students
[42] la ms importante persona (D3,m45,3,ELA43)
Span.: la persona ms importante
the most important person
[43] esa es una diferente generacin (036,1045,3,ELA43)
Span.:. . .generacin diferente
thats a different generation
It is interesting to note that type 5 caiques occur mostly in Group 3. In the
samples examined from eleven speakers in Group 2 only two cases were found:
[44], produced by V 21, a speaker with lower proficiency than most in this group,
and [45], provided by L 19 , who is by contrast quite proficient in Spanish (see
Table 2.2) Observe that [45] differs from all the others in that it incorporates a
switch to English, king-size, a unit of measurement that does not have an exact
correspondence in the Hispanic world (a cultural borrowing). This fact would
perhaps justify placing [45] in a category other than word-order caique, for in
stance, a switch to English with a Spanish borrowing (cama bed) included in the
switch. Since I have only one example, I have chosen to consider it a type 5 lxico-
syntactic caique, with an adjective-noun (or noun-noun) order.
[44] tengo dos ms meses (V 2i,fi8,2,ELAi7)
Span.: tengo dos meses ms
I have two more months
[45] pos tengo tambin king, compr una king-size cama (L 1 9,f22,2,ELA82)
Span.: pos tengo tambin king, compr una cama de tamao king
Well I have also a king, I bought a king-size bed

6.2.5.6. Type 6
This type of caique is a form of relexification: by reproducing the source model
almost word for word, the caique creates a syntactic structure that is non-existent
in the replica language. The lexical units are from the replica language, with
possible switches contained in the combination. Examples [46, 47] are illustrative.
The genitive marker s is in parentheses in [46] because I cannot discern if it is
present or not, given that the following word starts with an s and the speaker is
talking fairly rapidly. Be that as it may, the construction possessor-possessed does
not exist in Spanish.
[46] [Estaba trabajando]. . . abajo de un hermano(s) social security (D39,
f28,3,ELA42)
lit.:. . . under of a brother(s) social security
Span.:. .. con la seguridad social de un hermano
184 Conflicting Loyalties
lit.: .. . with the security social of a brother
[He was working].. . under a brothers social security (number)
[47] Yo nac diez millas afuera de la ciudad de Santa Fe (.437,0157,3,ELA22)
lit.: I was born ten miles away from the city of Santa Fe
Span.: Yo nac a diez millas de la ciudad de Santa Fe
lit.: I was born at ten miles from the city of Santa Fe
s>
Finally, I have considered as an instance of relexification the calquing of the
sequence that is why. Even though it may be argued that this is a bound collocation
in English, I decided to include its calquing into Spanish as type 6, because ii
creates a syntactic structure that does not exist in this language:
[48] Eso es por qu yo quiero un hijo (R42,mi5,3,ELA3o)
lit.: that is why I want a child
Span.: Por eso (es que) yo quiero un hijo
lit.: for that (is that) I want a child
[49] y eso es por qu nosotros fuimos pall (S38,fi9,3,ELA3i)
lit.: and that is why we went there
Span.: y por eso (es que) nosotros fuimos pall
lit.: and for that (is that) we went there
In the following section, I estimate how widespread multiple-word modelling is
across the bilingual continuum.

6.3. Modelling Across the Bilingual Continuum

This section is based on an examination of one- to one-and-a-half-hour samples

of speech from each of 4 speakers in Group 1, 1 1 speakers in Group 2, and 13 in
Group 3, i.e. on a total of approximately 33 hours of taped conversations. Interest -
ingly, the analysis shows that the two types of multiple-word modelling that I am
concerned with, caiques of bound collocations and lexico-syntactic caiques (sec
tions 6.2.4 and 6.2.5), distribute differently across the continuum. Group 1 is set
apart from the other two groups in that multiple-word modelling is practical]}
non-existent in their speech. Multiple-word modelling is far from being extensive
in Group 2, and only a slight increase is noted in Group 3. (I use modelling as
a cover term to refer both to caiques of bound collocations and to lexico-syntactic
caiques; note that this last type may include the transfer of form and meaning (type
6 in section 6.2.5).)
In their study of the Spanish of two generations of Cuban Americans in New.
York, Otheguy el al. (1989: table 1) show that, while single-word modelling (in
their terminology, modelling means calquing of meaning only) amounts to onl>
0.1 per cent in a corpus of 7,162 words in the first generation (six Cuban-born
adults), it increases considerably to 1.6 per cent of 5,724 words in the second
generation (six US-born late adolescents). Of the three types of contact feature
 Conflicting Loyalties
they study, loanwords, modelling, and switching, they observe that modelling is
what most distinguishes the speech of the second generation from that of the first.
Loanwords average 0.8 per cent and the difference between the two groups is of
only 0.2 percentage points. Switching is the most prevalent type of phenomenon,
but it does not increase as much as modelling from one generation to the next.
I have noted and analysed only cases of multiple-word modelling in my data.
The assessment of how extensive modelling might be is based on a simple count
of the instances of caiques attested in comparable amounts o f recorded conversa
tion across speakers. This is obviously not a rigorous assessment of the preponder
ance of this phenomenon. However, it is an approximation which is more reliable
than a completely impressionistic evaluation: it offers some form of quantified
information based on the assumption that, given similar topics of conversation
across speakers, one may expect the discourses produced to be comparable in
terms of the linguistic devices employed. Comparability of discourse topics applies
in the case of most speakers except those at the bottom stages of the proficiency
continuum (e.g. N40, R42, A46). These speakers are quite aware of the limitations
of their linguistic resources in Spanish; they confine their conversation in this
language to a number of very reduced concrete topics which they can handle
without deviating too much from Spanish norms, and they switch to English if
necessary to continue conversing.
This strategy most probably explains the relatively low number of caiques used
by these speakers. One might even submit that speakers with a very low degree of
proficiency in a language do not have the ability to be linguistically creative in this
language. More proficient speakers, on the other hand, display the same type of
linguistic creativity that characterizes innovations in fully fledged monolingual
varieties, e.g. adoption of loanwords, semantic extensions. Observe, in this respect,
the skilful extension of the meaning of descansar to rest in [50].
[50] They were laying off. So, I didnt get laid off. Ramn, Ramn got laid off.
And I quit because he got laid off. Because I was working, and he was
working at nights . . .
Dije, No, si LO v a n A d e s c a n s a r a l (lit.: if him are-3pl gonna rest him), pa
qu me quedo yo, especial yo? Yo, de aqu, como, onde puedo agarrar trabajo. El,
es ms difcil, porque hes not reglado para garrar trabajo. (039/28,3,
ELA42)
I said, No, i f t h e y r e g o i n g t o l a y h i m o f f , why should I stay, especially
me? Im from here, so I can get a job anywhere. As for him, its more
difficult, because hes not fixed (legalized) to get a job.
As a transitive verb, descansar to rest/lean/lay takes an inanimate direct object
in Spanish (descansa tu cabeza en mi hombro rest/lean/lay your head on my
shoulder). Most likely based on the connection between descansar and lay with the
meaning o f rest/lean, D39 establishes the further link with lay off. This has lxico-
syntactic consequences because both the meaning of lay off and its subcategorization
are transferred to descansar, which now incorporates the meaning of to fire (an
 18 6 Conflicting Loyalties

employee) and allows an animate direct object, i.e. it represents a lexico-syntactic.

caique of type 4.
Only two unquestionable caiques are attested in the data examined from the
four speakers in Group 1 (Ei, R n , C 13, J14). One involves the use of volver para
atrds to return, the other is the substitution of a at for en in/on/at (Me recogin
A la biblioteca he picked me up at the library). Both examples occur in E l s speech.
The results for Groups 2 and 3 are presented in Table 6.1. The mean is based on
an average of the cases attested for every ten minutes of conversation.
Table 6.1 indicates that modelling increases in Group 3 (those in the special
Group 3 are italicized). One person in Group 2 does not produce any cases of
modelling, and only one evidences an average of more than 1.4 cases every ten
minutes. This speaker is V 21, whose Spanish has been shown to be further from
the norms of Group 1 in other respects as well when compared to other speakers
in her group (see Table 2.2). By contrast, all speakers in Group 3 incorporate
modelling in their speech, ranging from an average of 0.6 to 2.1 cases per ten
minutes of conversation.
I have not focused on loanwords, but I see no reason to believe that loanwords
would amount to a much higher percentage in L A Spanish than in Cuban Spanish.
Thus, if the major lexical differences between general Spanish and LA Spanish art-
due to the calquing of meaning only, then one might expect that intelligibility
between these varieties would not be badly impaired, especially if the interlocutor
has some knowledge of English. This is indeed the case, as my analysis of narrative
and hypothetical discourse indicates (see Chapter 3). Regular inferential processes
allow both Group 1 speakers and out-group members to draw the necessary
semantic connections between the meaning that the items in question have in the
more conservative varieties and the usually extended meanings that these items
acquire as they caique those in the corresponding items in the source language. It
seems to me that the same inferential processes allow speakers of, for instance,
Castilian and Chilean Spanish to understand one another even though their lexi
cons do not exactly correspond. Words are not used in isolation; they are embed
ded in contexts which usually offer the necessary clues to their interpretation.

6.4. Acts o f Conflicting Loyalties: Positive Attitudes and

Uncom m itted Behaviours9

6.4.1. The Study o f Attitudes

The attitudes that monolinguals and bilinguals have towards such aspects of societal
bilingualism as the issue of the maintenance of two languages, the public and
9 This section owes much to the help o f Consuelo Siguenza, a graduate student and m y research
assistant in the language attitudes project. Consuelo participated in all stages o f the research, and
applied the questionnaires in person to almost every one o f the respondents. Analysis, generalizations,
and conclusions are, however, m y sole responsibility.
 Conflicting Loyalties 187
T able 6 .1 . T ypes o f Modelling Attested in Groups 2 and 3

Modelling types

Bound Lxico-syntactic caiques: type

i 2 3 4 5 6
Tim e (min) Na N N N N N N Total N/io min.

Group 2
B27 60 0 0.0
A29 60 i i 0.2
L28 60 i i 2 0.3
L 19 90 i 3 i 1 6 0.6
M26 90 i 6 7 0.7
A20 90 2 , 3 2 i 8 0.8
R 17 60 3 3 i 7 1.2
R24 60 i i i 3 2 8 i -3
E 18 60 3 3 2 8 i -3
H22 90 i i 4 4 3 13 M
V2X 90 i 2 9 3 5 i 5 26 2.9

Total 910 7 12 19 13 22 2 11 86 0.9

Group 3
H48 90 i i 2 i 5 0.6
R so 90 5 / 6 0.6
A34 60 1 4 5 0.8
A37 45 2 i 1 4 0.9
N40 40 3 i 4 1.0
A46 45 X 3 i 5 1.1
60 2 4 i 7 1.2
833
M 41 60 i i 2 i i i 7 1.2
M 47 60 2 2 2 2 8 i -3
S38 90 i 2 i 2 i 2 4 13 1.4
R42 45 2 i i 3 i 8 1.8
D39 90 i 3 2 3 4 2 4 19 2.1
D3 6 90 3 2 4 4 6 19 2.1

Total 865 2 25 7 18 28 10 20 no i -3

a N = no. o f cases per 10 mins.

 Conflicting Loyalties
private status assigned to each language, or the linguistic adequacy or correctness
of the minority language can throw some light upon the motivations that may
promote either the maintenance or the fading away of the minority language.
Thus, the study of attitudes has constituted the focus of many research efforts that
attempt to predict the future of (usually) a minority language in a bilingual con
text, as well as to elucidate the symbolic status that the languages involved may
have for particular social groups.
The mental position (feelings, emotions) that individuals have, or are willing to
disclose, towards a specific fact or state of affairs is not easily measurable. Romaine
(1989& 257) notes this problem in her discussion of attitudes towards bilingualism:
The translation of the notion of attitude from the subjective domain into
something objectively measurable, and therefore more easily comparable, is a
common problem in any research that involves social categorization and perceptual
judgements.
In terms of direct methods, the attitude questionnaire has been one of the
devices most used to elicit measurable information about specific aspects of bilin
gualism, e.g. code-switching, the appropriate domains for use of the languages,
self-reports concerning language usage patterns, and level of language proficiency.
Questionnaires have a number of undeniable advantages over direct observation or
the personal interview: a higher number of people may be surveyed in a shorter
period of time, the answers to closed questions make measurement and comparison
across individuals easier. Unfortunately, questionnaires are also riddled with
pitfalls.
Dorian (1981: 157-60), for instance, notes that some of the problems include
not knowing who in fact filled out the questionnaire, and not being able to clarify
or expand on the instructions. (Supposedly she refers to questionnaires filled out
when the investigator is not present or when they are given to a large group of
people at the same time.) Furthermore, the issue of the predictive validity of the
instrument used to elicit language attitudes has attracted attention and criticism.
According to some researchers (Agheyisi and Fishman 1970; Edwards 1985; Ryan
et al. 1982), most types of questionnaire are inadequate to obtain data relating to
anticipatory behaviour, and therefore lack predictive validity because they focus
exclusively on the cognitive and/or affective components of attitude.
Attitudes, however, involve not only a mentalist perspective, which implies that
they would have to be inferred from the subjects introspection and/or the re
searchers interpretation of overt behaviour, but also actual, overt, observable be
haviour (Agheyisi and Fishman 1970). This multiple-component view (cognitive,
affective, and conative or action components) dictates the inclusion in the ques
tionnaire of items which elicit information about the various dimensions of lan
guage attitudes.
Accordingly, in my study I employed a number of questionnaires modelled
mostly on questionnaires used by Dorian (1981), Fishman et al. (1971: ch. 5), and
Mejias and Andersons (1988), (MA) and applied them in 1988 -9 (i.e. long after
 Conflicting Loyalties 189
T able 6.2. Sample o f Respondents to the Questionnaires

Group i Group 2 Group 3 Total sample

Age F M F M F M F M
2 1-3 0 2 2 5 3 3 i 10 6
3 1- 7 0 4 4 S 3 3 4 12 1 1
Total 6 6 10 6 6 5 22 17 = 39

I had completed the recording of conversations). By using a battery of different

questionnaires already tested and applied in bilingual contexts I expected to obtain
data on the various components of attitudes which could be used for comparative
purposes. Appendix 4 contains the English version of the five questionnaires
administered (there was also a Spanish version of all the questionnaires except No.
Ill, applied only to Groups 2 and 3).
The questionnaires included mainly closed-question items. This eliminates the
problem of respondents failing to focus on the expected dimension, since all they
have to do is choose from a set of provided categories, and it also facilitates the
quantification and comparison of responses. The questionnaires were administered
in an interview setting, with the respondents either writing their answers in the
presence of the researcher10 or asking her to do so. The researcher was there to
help facilitate the completion of the questionnaire, and interacted with the re
spondents in a friendly, interview-like fashion, questioning, explaining, clarifying,
and expanding on certain items. Such a method was especially helpful with the
older respondents, who either could not read in Spanish or were prevented by
poor eyesight from completing the questionnaire without help. Thirty-three ques
tionnaires were applied in person and 9 had to be mailed;11 of these 9, 4 persons
responded. Not all the respondents completed every questionnaire; this is noted
when relevant in the ensuing discussion. In addition to these 37 individuals, all
included in the sample of 50 bilinguals studied in this book, two additional ones
volunteered to participate: B 51, f43 (included in the results for Group 2), married
to R50; and R52, m45 (included in the results for Group 3), married to E45. Table
6.2 summarizes the information.
Personal contact in a combination of the face-to-face (rather than telephone)
interview and questionnaire format offers the interviewer the possibility to focus
the attention of the respondent on the desired cue, and to assess and influence the
mood of the respondents (thus reducing the chance of boring or irritating them).

10 T h is researcher was in all cases Consuelo Sigiienza, who is a member o f the Mexican-American
community investigated.
11 Questionnaires were mailed to those speakers whose schedule did not allow us to meet with them.
Seven o f the speakers included in the sample o f 50 had moved and could not be contacted; one had
died.
 Conflicting Loyalties
Although the researchers presence risks biasing the respondents, the fact that they
are all exposed to the same application method offsets this possible bias, since in
principle it should affect all speakers in the same manner. Thus, differences and
similarities in the results obtained across speakers can be interpreted as responding
to factors other than the researchers presence. Despite these efforts, the responses
to the questionnaires must still be treated with caution, and should be regarded as
useful only in their general gestalt (Dorian 1981:.? 160).

6.4.2. Questionnaire I

This questionnaire, admininistered to speakers in Groups 1, 2, and 3, was partly

based on and adapted from Dorian (1981), Fishman et al. (1971), Gal (1979), and
Zentella (1981). Questionnaire I is reproduced in Appendix 4, followed by an
arrangement of its 80 items, according to the type of dimension being investigated,
into 9 groups (I-IX , A or B, depending on the language about which information
is sought).
The 79 closed questions were responded to by 34 persons surveyed, but only 10
of these answered 4 open questions included at the end of the questionnaire
(80-3), fairly briefly. The comments with respect to item 80 agree that the inter
locutor is the most crucial factor in determining what language to speak. Most of
the respondents agree that they switch languages if the topic calls for the other
language, when they get stuck and cannot promptly remember how to say
something in one or the other language, or when a word [in the other language]
better describes a situation (M25). As for their feelings towards switching, they
are reflected in S38s (fi9, 3) response. She states that she switches when she is
among down-to-earth bilingual people I feel comfortable with [because] some
thoughts are better expressed in Spanish.
Item 82 was also commented on briefly. Groups 2 and 3 concurred that when
they used Spanish in front of an English-only speaker they did not sense any
hostility. Even if they did, however, this would not stop them from using whatever
language they cared to speak. One speaker from Group 1, on the other hand, stated
that he had once been told that he should learn to speak English, to which he
responded, in English, I know English. The topic addressed by item 82 came up
spontaneously during some of the conversations recorded earlier. Example [51] is
illustrative.

[51] R: Rso,m4,3,ELA36; C: researcher

C: Fjate. T poco a poco has ido viendo que ha llegado ms y ms gente a
la polica que son latinos.
R: Latinos. Como ahora, estaba en el catering wagon y, y, y estaba hablando
- Un mecnico mejicano le dijo una a la - al que est cocinando en el
catering wagon, le dijo una de doble sentido, una palabra de doble sentido
noms. Nothing serious, nothing serious, you know, just a - No me acuerdo
 Conflicting Loyalties 191

qu era ni nada. So le habl yo patrs en espaol. Ya te agarr la

movida, le dije, Ya te, ya te estoy escuchando. Y luego este gringo
estaba a un lado y luego Eh, dice, don t speak thatforeign language around
here? Es lo que me dice a mi, you know. What do you mean foreign
language? That sucker was around here before the English were!' [R laughs].
And he says, Man, youre right!, he says, Youre right! OK? [R and C
laugh]
C: He had to accept it, uh?
R: Yeah, he had to accept it. Two reasons: I m right, and two, I m the boss. [R
and C laugh]
C: There you are. Little by little youve been seeing that more and more
people have joined the police who are latinos.
R: Latinos. Like now, I was at the catering wagon and, and, and was speaking
- A Mexican mechanic said one to the - to the one whos cooking in the
catering wagon, he told him a double-sense one, just one double-sense
word. Nothing serious, nothing serious, you know, just a - I dont re
member what it was. So I talked to him in Spanish. I got what you said,
I said, Im, Im listening to you. And then this gringo was to one side
and then Eh, he says, dont speak that foreign language around here.
Thats what he says to me, you know. What do you mean foreign
language ? That sucker was around here before the English were! [R
laughs]. And he says, Man, youre right!, he says, Youre right! OK?
[R and C laugh]
C: He had to accept it, uh?
R: Yeah, he had to accept it. Two reasons: Im right, and two, Im the boss!
[R and C laugh]

This passage from a speaker in Group 3 supports in part the opinion expressed
by those in Groups 2 and 3. The question remains, however, whether the English
monolingual would have agreed with R s statement if the second reason (R is the
boss) had not applied.
Table 6.3 displays the results of selected items in the sections (IA and IB; see
Appendix 4) probing the respondents assessment of their use of Spanish with
various interlocutors.
The decline in the use of Spanish across groups is evident. This situation seems
to apply not only in work and church domains but also within the family environs.
While Group 1 speakers report almost exclusive use of Spanish with parents,
grandparents, and siblings, some speakers in Groups 2 and 3 speak only frequently
or sometimes to their parents and siblings in this language, and four respondents
in Group 3 report never using Spanish to communicate with their siblings.
R 17 (f2i,2,ELA3) spontaneously assessed this situation during the time of
the recorded conversations. She told me that when they were small their parents
asked them to speak only Spanish at home. However, this rule was frequently
 T able 6.3. Use of Spanish with Various Interlocutors

c

i*
3

+>
C

3O
3O
P*
CX

3U
'
a .
2 ,
O -

O *
s
a
ao
*0
<
o

<
<

z
z

z
D

8
B

vn

-O

ns
co

bc>

$ *
c i>
G O 5.
co

o
o
J-
co

x>

)
8 = & *
W

O
&
v-

r-

d
w
Cfl

't4> | -* .2
g1
?r*3 jS
< 2 fc frO K ip <
' t a ' ,tN O O '000\0

S ^ r<U =3 & ,fa * C

S3 .5
00

CO GO
G

r- <s
*2
N

cn
0 3
tn . b

u
lx

w _ U CJ >,
<u be V
G

0 *s O
7 O
>

13
W

<L> *rt

'TT'
N

a
CO CO
N

U ^ U ,
N
CO

N^O vO t N N
N

<
>
CO

"cl
o1
u

II
!-j C
S c3

5T
g
M

&

6
o
r't">Noc<''se*i>

_s

*rt
*-*
mu
s
r'j 't ro to
<T> rf rf ^ ** "$

>->
A = always; F = frequently; UO = usually or sometimes; N = never; NA = does not apply. In addition to ticking the NA box when the situation
in fact did not apply in the case of a specific respondent, this box was also ticked when the respondent did not feel comfortable about giving an
answer.
b Nos. and letters in Tables 6.3-6 correspond to those in the questionnaire (see Appendix 4).
 Conflicting Loyalties 193
disregarded by the time her younger brother and sister were born. Consequently,
the youngest did not learn enough Spanish to interact comfortably in Spanish; the
siblings use English among themselves. R herself explains:

[52] . . . y mi hermanita chiquita si habla espaol, pero no creo que tiene la voca,
el voca, el vocabulario? / Researcher S, el vocabulario./ S, no tiene mucha
vocabulario para, para estar en una conversacin. Y, y yo estaba diciendo,
Pos otra vez necesitamos de esa regla.
.. . and my little sister does speak Spanish, but I dont think she has the voc,
the voc, the vocabulary? /Researcher: Yes, the vocabulary./ Yes, she doesnt
have much vocabulary to, to be in a conversation. And, and I was saying,
Well, we need that rule again.

In the work domain, Spanish tends to be used more with fellow workers than
with supervisors. Dorians (1981: 162-3) respondents report the same with respect
to Gaelic, a situation which she deems to be a reflection of East Sutherland society,
where there are simply very few bilinguals in supervisory positions. This expla
nation possibly applies in Los Angeles, where those in supervisory positions may
be bilinguals, but most frequently English-dominant. It may also be the case,
however, that the large numbers of recent Spanish-speaking immigrants working
in Los Angeles may stimulate Mexican-Americans to revive their lost linguistic
skills, even if the English-dominant bilingual is in the supervisory position. This
situation is well reflected in the spontaneous description volunteered by L28 (f37,
2) of her husbands (H48, m38, 3) proficiency in Spanish. L s husband is the son
of the owner of the printing company to which [53] refers.

[53] Y digo yo que este seor [her husband] su espaol era horrible, verdad?,
porque lo haba perdido cuando se mudaron a ese vecindad. S, pero cuando
nos casamos entonces ya no era el estudiante, ya se puso a trabajar en la
planta. Y en la planta, los hombres que trabajan las imprentas casi todos son
hispanoamericanos. Vienen de distintos pases, pero todos hablan espaol.
Entonces por su amistad, en la imprenta l pudo aprender espaol.
And I say that this man his Spanish was horrible, right?, because he had lost
it when they moved to that neighborhood. Yeah, but when we got married
then he was no longer a student, he started working in the company. And in
the company, the men who work at the printing presses are almost all
hispanoamericans. They come from different countries, but they all speak
Spanish. So then because of their friendship, in the company he learnt
Spanish.

In addition to the negative effect that the different neighbourhood (mostly non-
Hispanic) had on Hs Spanish, his wife surmises that he had also stopped using it
because his grandparents had died by then.
In regard to the use of Spanish with strangers, Group 1 speakers hesitate to use
194 Conflicting Loyalties

this language with individuals who are not necessarily in-group members (e.g.
clerk, teacher, police officer). This trend is strengthened in Groups 2 and 3.
No inconsistencies are reported in the corresponding items that concern the use
of English in relation to interlocutors. Respondents in Groups 1 and 2 tend to
report more reduced use of English in the situations examined (see questionnaire
I, section IB) as compared to speakers in Group 3.
Table 6.4 displays the results of questionnaire I, section IIA, followed by Table
6.5, corresponding to section IIB. The two items in Table 6.5, 29 and 32, concern
ing English, are roughly conversely comparable with 30 and 31 respectively in
Table 6.4, concerning Spanish. Note that this comparison does not indicate incon
sistencies in the responses. On the other hand, the lack of fitness between the
responses given to items 12 and 3 1, for which I expected nearly exact results, raises
the usual questions concerning the reliability of self-report. Overall, however, I am
confident that the questionnaires are an accurate reflection of the gradual reduction
in use that affects Spanish in this community. This is also revealed by self-report
with respect to the use of Spanish according to topic or type of activity in Groups
2 and 3.
Writing and reading in Spanish are reported to be much less frequent than
listening activities in this language. For instance, 6 of 12 respondents in Group 1
report reading Spanish publications either always or frequently, 5 write letters
in Spanish either always or frequently, but 9 of these 12 speakers listen to Spanish
radio programme and music broadcasts either always or frequently. Although
there is less reading, writing, and listening in Spanish in the other two groups, 9
of 13 speakers in Group 2 and 4 of 9 in Group 3 report reading Spanish publica
tions sometimes . Letter-writing is also an activity that occurs, though only some
times, among Group 3 respondents. The maintenance of some degree of literacy in
Spanish in this group is encouraging. (No speakers were illiterate in both languages.)
As would be expected given the differing levels of proficiency in Spanish across
the continuum, Group 1 speakers evidence strong preferential use of Spanish
in response to items 56, 59, 68, and 69 by marking mostly always; speakers
in Groups 3 show preferential use of English by selecting mostly sometimes or
never; Group 2 falls between these two groups, with responses distributed almost
equally between frequently, sometimes, and never. Preferential use of English
increases accordingly across the three groups.
The responses to the items in section VIIIA of questionnaire I show that contact
with Mexico remains strong in Groups 1 and 2, and quite solid in Group 3 as well.
Table 6.6 displays the answers.
Close contact with Mexico certainly facilitates the maintenance of Spanish, and
visits to Mexico motivate Hispanics to learn or improve their knowledge of this
language. But regardless of how much reduction Spanish has undergone in terms
of domains of use and linguistic resources in Groups 2 and 3, almost every speaker
in these groups, as well as in Group 1, responds with an emphatic yes to question
45: When I have children, I want them to be able to speak Spanish fluently. Only
 <

o
D

Cu
3
O

>% <
Z
>
m tn N w o**) ro >-t co
o
<J a
*6L 8-
H
o
O o
c
r^so a o ov'so
o P 03
a 8
H
3
.2 i!
rt U-< <
13 a
3 2
G
O8
A
cn <
z

NO
w
< o
H D

in m <*) rf in \o Tf-m o
Oh D
rj- r^- ^ vo ^ vo '4*
s
g
0)-
>
u fc "cS I!
<u I H-*
'2 t
us Ch U> be
*H
c/T
3
O bo rs O 'h
.5 i .
5b >> .~ ia 8 a jS
"3 2 T" 3SISu oo 2 o
"es
q^PhP^Hh-1EU il
Oh
N w "3*l0'0 S* CO <
i-* f O f O V N O >0 N SO
196 Conflicting Loyalties
T able 6.5. Use of English in Relation to Topic or Activity*

Topic Group i Group 2 Group 3

A F UO N NA A F UO N NA A F UO N NA

29. Praying 3 i i 7 7 3 2 i 6 1 2
32. Religious
service i i 5 4 i 3 6 3 i 3 2 3 i

a A = always, F = frequently, UO = usually or sometimes, N = never, NA = does not apply.

T a ble 6.6. Contact with Mexico

Activity Group 1 Group 2 Group 3

A N NA A N NA A N NA

5. I visited Mexico 9 2 i 10 3 3 5 i
6. Family visited Mexico 8 2 2 11 2 4 4 i
14. Visited from Mexico 9 3 11 2 7 i 1

1 A = yes; F = frequently; UO = usually or sometimes; N = no; NA = does not apply.

three speakers respond that full fluency is not crucial, a little Spanish would be
fine (Group 3: B33, D36, E45). Similarly, all speakers but two in Group 1 (J14,
J16 ) respond yes to question 47: When I have children, I want them to be able
to speak English fluently.

6.4.3. Questionnaires I I and I I I

Questionnaires II, administered to Group 1, and III, given to Groups 2 and 3,

were based on Dorians (1981) two versions of her questionnaire II, and on another
questionnaire devised by Mejias and Anderson (1988) (MA). Questionnaire II was
answered by 12 speakers (6 females, 6 males) and questionnaire III was completed
by 16 speakers in Group 2 (10 females, 6 males), and by 11 in Group 3 (6 females,
5 males). Group 1 speakers were asked to select the main reasons they have to be
happy to know Spanish, while Groups 2 and 3 were asked to choose the main
reasons they would have if they were ever to try to improve or maintain their
Spanish (see questionnaires in Appendix 4).
When presented with the questionnaire, the respondents were instructed to
circle the reasons that seemed important to them, draw a line through the least
 Conflicting Loyalties 197

important reasons, put a star next to the most important one or two reasons, and
to leave unmarked those that were neutral to them. They were also instructed to
write in their main reasons if these were not included in the questionnaire. No
respondents added reasons to the ones given (Dorian reports the same for her
respondents).
The sixteen-item questionnaire corresponds to the four dimensions of attitude
applied by MA in their study of the language attitudes of Rio Grande Valley
residents in Texas. The group-identity (G) (value V in MA) dimension refers
to the use of Spanish as a public symbol of group worth and lasting values and
traditions, with ethnic identity at its core, as illustrated by item 6: [If I were ever
to try to improve, or maintain my Spanish, my main reasons would be:] It would
make me feel more a part of the community I live in. The sentimental (S) di
mension refers to the private feelings, emotions, and personal satisfaction the
bilingual experiences in the use of Spanish, as illustrated by item 4: Spanish is
a very rich and expressive language. The communication (C) dimension refers to
the use of Spanish for public understanding and interpersonal communication, i.e.
to the actual transmission and communication of information between people in
Spanish, as illustrated by item 5: It is necessary for daily communication. Finally,
the instrumental (I) dimension refers to the use of Spanish for personal benefits,
material profits, and gains, as illustrated by item 14: It helps me make money at
my job. The communication and group-identity dimensions represent extrinsic
perspectives, through which bilinguals view themselves as identifying the use of
Spanish with belonging to a group or community, whereas the instrumental and
sentimental dimensions represent intrinsic perspectives, through which bilinguals
view themselves as individuals.
These dimensions roughly correspond to some of Dorians (1981: 165-71) rea
sons for valuing Gaelic: G = local integrative and tradition, I = operational,
S = subjective aesthetic. For purposes of comparison with MA, in the quantitative
analysis I have classified items 2 and 16, borrowed from Dorians questionnaire II,
as instrumental rather than abstract principle and exclusionary , respectively, as
Dorian does.
M A report that, for their entire sample of 293 subjects, using Spanish for
communication is strongest among the four attitude dimensions (endorsed by an
average of 45.9 per cent of their sample). This is followed by value (22.6 per cent),
sentimental (15.2 per cent), and instrumental (13.2 per cent) reasons. On this basis
they hypothesize that Spanish will be maintained along with English in Rio Grande
Valley, Texas, rather than suffer a shift to English. But they find a contradiction
to strong Spanish language maintenance when viewing the variable of generational
group. They note that although third- and fourth-generation respondents, whose
residence in the U SA has been longer, chose more communicative reasons for
maintaining Spanish, first- and second-generation respondents chose more senti
mental reasons for maintaining Spanish. MA interpret this to indicate a tendency
toward language shift away from Spanish among newer residents, i.e. first- and
xg8 Conflicting Loyalties
T able 6.7. Most Important Dimension for Valuing Spanish Based on Percentage of
Respondents Selecting Reasons in Each

Group i Group 2 Group 3 Total sample

(X2 f (16) (ix) (39)
% % % %
P
Group identity 25 12 6 17
Sentimental 27 27 22 26
Communication 21 27 22 23
Instrumental 27 33 So 34

a Figures in parentheses indicate no. of respondents.

second-generation respondents. They attribute this to a temporary desire to shift

from Spanish which will moderate as residency is more established (p. 406).
In our sample, although much smaller than M As, we find a different outcome:
a higher percentage of our respondents select the instrumental dimension as the
most important (34 per cent), with communicative reasons being in third place,
selected by only 23 per cent of the sample. As I show below, this difference may
be due in part to the inclusion of item 7 (reading in Spanish), a communicative
reason, in our questionnaire II but not in M As, and to the classification of item 2,
strongly supported by most respondents, as instrumental. Table 6.7 summarizes
the results. For the total sample of respondents, the attitude toward the use of
Spanish for instrumental reasons is strongest among the four attitude dimensions.
The attitude toward the use of Spanish for sentimental reasons is the next strong
est (26 per cent of the sample), followed by communication reasons. Group-
identity reasons are the weakest among the four attitude dimensions, with only an
average of 17 per cent of the respondents choosing them.
It is evident in Table 6.7 that the communication dimension of attitude in
creases in importance from Group 1 (21 per cent) to Group 2 (27 per cent). The
lower weight given to communicative reasons in Group 1, compared with group-
identity and sentimental reasons, may indicate that recent immigrants are still
nostalgic about their native land and culture. A different explanation is needed to
account for the decline in the value assigned to communicative reasons in Group
3 (22 per cent), however, as compared to Group 2. Indeed, the weakening of the
communicative motive as Spanish-language proficiency decreases further also makes
sense, given that a person with reduced Spanish productive proficiency will tend
not to use this language to read or for oral communication, unless absolutely
necessary. For our sample, absolutely necessary cases involve those in which the
low-functioning bilingual must communicate with recent Spanish-speaking immi
grants or visitors who do not understand English. Otherwise, interaction takes
 Conflicting Loyalties m
place in the two languages: each interlocutor uses his dominant language knowing
that the listener understands both English and Spanish.
In terms of intergenerational and language-proficiency differences, in Group i,
the attitudes toward the use of Spanish for instrumental and sentimental reasons
are strongest among the four dimensions (27 per cent). Although a well-balanced
group overall, the private, intrinsic attitude dimensions (S and I) are endorsed by
Group 1 respondents as the most important for maintaining Spanish, with com
municative reasons as the weakest. These results agree in part with those obtained
by M A for first-generation immigrants, who also chose more sentimental reasons
for maintaining Spanish.
Even though differing in degrees, Group 2 and Group 3s attitude dimension
relationships are identical. Instrumental reasons are the most important in the
maintenance of Spanish, followed by S and C reasons, while group-identity rea
sons are the weakest. These overall results seem to point in the direction of
possible diversity of bilingual communities across the U SA in terms of attitude
and motivations toward maintaining Spanish, although many more in-depth and
extensive studies using identical questionnaires and methodologies need to be
carried out before any firm generalizations may be established. Therefore, the
comparisons across communities presented in this chapter are necessarily only
tentative.
In the Rio Grande Valley, M As respondents chose communicative reasons
those that deal with public understanding and interpersonal communication as the
most important for maintaining Spanish in their community, with instrumental
reasons being the weakest. In Los Angeles, our sample of respondents chose
instrumental and sentimental reasons, those dealing with personal benefits and
emotional satisfaction, as the most important reasons for using or maintaining
Spanish in their community. And even though the item dealing with reading
makes a difference, an examination of the percentage of respondents who starred
each item confirms that of the communicative reasons offered (items 5, 11 , 12),
only talking to people from other Spanish speaking countries is starred equally
across the three groups, but it does not receive the strongest support. Table 6.8
presents the percentage of respondents who starred each of the 16 items in ques
tionnaire II. Within each group I have italicized the two items starred by most
respondents.
Conspicuous in Table 6.8 is the high value assigned to item 2 in Groups 1 and
2 (its broadening to have more than one language). Dorian proposes this reason
as an abstract principle, supported only weakly by her respondents. Item 16 (it
is useful to know a second language) is strictly comparable with item 2 only in the
version given to Groups 2 and 3. It is somewhat surprising that the responses to
these two items are not the same in these two groups: the utilitarian reason for
knowing more than one language is favoured by Group 3 (item 16), while the
purely intellectual reason (item 2) receives a warmer response from Group 2. The
most starred items in the three groups, then, do not pertain specifically to Spanish,
200 Conflicting Loyalties
T able 6.8. Percentage o f Respondents Starring a Reason as
M ost Important for Valuing Spanish

Dimension Group 1 Group 2 Group 3 Total sample

(I2)a (16) (11) (3 9 )
% % % %

(G) !. 41.7 12.5 0,0 17-95

( I) 2. 75- 43-7 9.1 43-59
(S) 3- 8-3 0.0 9.1 5-13
(S ) 4- 33-3 18.7 0.0 17-95
(C) 5- 33-3 18.7 9.1 20.51
(G) 6. 16.7 0.0 0.0 5 -i 3
(C) 7- 8-3 12-5 0.0 7.69
(I) 8. 16.7 0.0 9.1 7.69
(G) 9. 16.7 12.5 0.0 10.26
(G) 10. 16.7 0.0 9.1 769
(C) 11. 8-3 0.0 0.0 2.56
(C) 12. 25.0 25.0 25.0 25.64
(S ) 13. 25.0 12.5 0.0 18.32
(I) 14. 0.0 0.0 18.2 5-13
( S ) IS- 33-3 25.0 27-3 2 8 .21
( I ) 16. 8-3 25.0 45-4 25.64

3 Figures in parentheses indicate nos. of respondents.

but to the more general or non-specific notion of another language. This elicited
attitude concurs with the spontaneous opinion expressed by one of the speakers in
Group 2:

[54] Insisten que se - claro, las dos [idiomas], Somos asi. Es una cosa - bonito hablar
dos idiomas. Its something that enriches the person. Its something that en
riches the society. Its, its education, its culture. It gives you two systems of
thinking, you know. (Ei8,f27,2,ELA32)
They insist that - of course, the two [languages]. Were like that. Its - nice
to speak two languages.

There is a difference in the number of items that speakers starred which also
points to a fairly disheartening situation. Although all respondents were asked to
star one or two most important reasons for maintaining or improving their know
ledge of Spanish, in Group 1, four respondents marked three reasons; in Group
2, two respondents marked three reasons; while no one in Group 3 starred more
than two reasons (nor did they offer other possible reasons in the section of the
questionnaire which encouraged them to do so). This selectional pattern seems
to indicate that Group 1 speakers have more very important reasons to maintain
 Conflicting Loyalties 201
T a b l e 6.9. Percentage o f Respondents Crossing Out a Reason as
Least Important for Valuing Spanish

Group 1 Group 2 Group 3 Total sample

(12) (16) ( 11) (3 9 )
% % % %

(G) 1. 0.0 12-5 9.1 7-7

(I) 2. 0.0 0.0 0.0 0.0
(S) 3- 8.3 43-7 27-3 28.2
(S ) 4. 8.3 12.5 9.1 10.3
(C) 5- 8.3 12-5 18.2 12.8
(G) 6. 33-3 31.2 36-4 33-3
(Q 7. 0.0 3 1.2 27-3 20.0
(I) 8. 4 1.7 68.7 3^-4 S i. 1
(G) 9. 0.0 3 1.2 18.2 17.9
(G) 10. 16.7 43-7 27-3 30.8
(C) 1 1 . 0.0 3 1.2 27-3 20.0
(C) 12. 0.0 0.0 9.1 2.6
( S ) 13. 8-3 6.2 9.1 7-7
( I ) 14. 58-3 43-7 45-4 48.7
( S ) i 5. 25.0 0.0 0.0 7-7
( I ) 16. 25.0 0.0 0.0 7-7

Spanish, and results in more items being thus marked in the questionnaires from
this group, where only item 14 is never starred; fewer items are starred in Group
2 (10), and even fewer in Group 3 (9). This may reflect a weakening of support
toward the maintenance of Spanish in Groups 2 and 3.
Furthermore, US-born bilinguals do not consider, overall, that the preservation
or protection of tradition and group integration is the most important motive to
maintain their ancestors language. Two G reasons are selected as the most impor
tant by only 12.5 per cent of the sample in Group 2, and a low 9.1 per cent of those
in Group 3 (i.e. one of xi respondents) star only one G reason, item 10: Its the
language of my friends and neighbours. This item, however, is also crossed out as
the least important by three speakers (27.3 per cent) in this same group, as illus
trated in Table 6.9. This table shows the percentage of respondents who crossed
out each of 16 prompts as the least important reasons why they would like to
improve or maintain Spanish (in the case of Groups 2 and 3) and why they are
happy to know Spanish (in the case of Group 1). Italicized in Table 6.9 are the
reasons marked by a higher percentage of respondents as least important for
valuing Spanish.
The three groups agree in their high percentage of rejection of two instrumental
motivations which value Spanish as a secret language (item 8), and as a means to
202 Conflicting Loyalties

make money at a job (item 14). Although this latter choice obviously depends 011
the type of occupation, it seems to be on the whole an accurate assessment, given
the secondary and unofficial status of Spanish in Los Angeles and the primary
status of English, the single official language in California, in education and in the
business world. Likewise, the three groups concur in never crossing out item 2
(Its broadening to have more than one language), an intellectual motivation
which was also starred by respondents in the three groups.
Generally, then, the results of questionnaires I, II, and III appear to indicate a
correlation between decrease in the number and strength of the motivations or
reasons for maintaining or improving knowledge of Spanish, reduction in the
domains of use of Spanish in the community, and linguistic attrition of this
secondary language.

6.4.4. Questionnaire I V

Questionnaire IV (see Appendix 4), administered to the three groups, was adapted
from Dorians (1981: 184-5) questionnaire III. For purposes of analysis and quan
tification, item 4 in section A (Spanish is a difficult language to learn) was left out:
because this was a fact question (see Questionnaire IV) I was uncertain as to the
attitude that the answer to this item could be interpreted as reflecting. The re
maining 29 items were divided into 4 categories: positive attitudes towards Spanish
(11 items); negative attitudes towards Spanish (9 items); positive attitudes towards
English (7 items); and negative attitudes towards English (2 items). The responses
were scored with 1- 5 points, decreasing from strongly agree to strongly dis
agree. Table 6.10 displays the results.
Table 6.10 shows clearly that attitudes towards both English and Spanish re
main on the whole quite positive. It is not surprising that Group 1 should favour
Spanish slightly more than the other two groups. Mean strength of agreement with
positive statements about Spanish decreases from 4.1 in Group x to 3.7 in Group
2 and to 3.6 in Group 3. Within each subgroup, women and men do not seem to
differ significantly in their support of Spanish. Some minor division of sentiment
appears to occur at the age cut-off point in Groups 2 and 3, where negative
statements about Spanish receive a stronger rejection from the younger respond
ents (1.7 per cent). This favourable trend may be due to societal changes of
attitude towards bilingualism in general in the last thirty years. Indeed, while older
bilinguals report having been disciplined for using Spanish in school, younger
bilinguals have not had such experience. By contrast, some of them have attended
schools where some form of bilingual education was being promoted.
The favourable responses given to questionnaire IV are somewhat in contrast to
the relatively low value placed by Group 3 on the existence of strong reasons for
maintaining or improving their Spanish (see Table 6.8). These differences do not
indicate inconsistencies, however, since not selecting a reason as most important
 Conflicting Loyalties 203
T a b l e 6 .10 . Average Agreement with Positive (+) and Negative (-)
Items about Spanish and English

+ Span. Span. + Eng. - Eng.

N* % % % %

Group i
15 -3 F 2 4-5 1.9 3-0 3-2
M 2 4.2 2.0 2.7 2-5
3 1- 7 0 F 4 4.0 2.0 3-2 i -3
M 3 4-1 2.4 3-4 1.8
Total XI 4-<b 2 .1 3-1 2.0

Group 2
15 -3 0 F 5 3-7 i -7 2.4 i -7
M 3 3-7 i -7 2-5 1.8
3 1- 7 0 F 4 3-6 2 .1 31 i -7
M 2 4.0 2.0 3-4 i -5
Total 14 3-7 1.8 2.7 i -7
Group 3
15 -3 0 F 3 3-5 i -7 3-1 1.2
M i 3-6 i -7 3-7 i -5
3 1- 7 0 F 3 3-6 2.4 3-5 2.5
M 4 3-7 2.2 3-2 2.0
Total 11 3-6 2.0 3-4 1.8

a N = no. o f respondents in each subgroup.

b italics indicate total per cent in each group for each category o f attitude.

for maintaining Spanish (questionnaires II and III) does not imply a negative
sentiment towards it.

6.4.5 Questionnaire V
Questionnaire V, administered last to all three groups, was based on Fishmans (in
Fishman et al. 1971: ch. 5) ten-item commitment scale. This commitment scale
was used by Fishman to examine the possible greater relevance to subsequent
pertinent behavior of commitments than of attitudes or other self-reports (p. 108).
Sophisticated statistics allowed him to confirm this greater relevance. More impor
tant for my purpose, given that I did not incorporate an overt behaviour criterion
in my study (such as the invitation to participate in a Puerto Rican evening
function sent by Fishman to part of his sample), is his observation that language-
attitude and language-usage self-reports do not appear to be guaranteed predictors
o f language behaviours unless these self-reports concern commitment-type
204 Conflicting Loyalties
T a b l e 6 . i i . Degree o f Commitment to Act in Order to
Maintain Ancestors Culture and Language

N* Total Mean Questions receiving Questions receiving

yesb the 2 highest numbers the 2 highest number-,
o f yes answers o f no answers
Item (N) Item (N)

Group 1
15 - 3 0 F 2 r4 77
M 2 9 .50
3 1- 7 0 F 4 16 44
M 3 21 77
Total 11 .62s 1 (10), 2, 5, 6 (8) 9 (6), 7 (5 ).
Group 2
15 -3 0 F 5 33 .48
M 3 5 .18
3 1-7 0 F 4 26 .72
M 2 15 83
Total 14 55 6 (10), 2, 8 (9) 7 (13). 4 (7 )
Group j
15 -3 0 F 3 6 .22
M 1 1 .1 1
3 1-7 0 F 3 6 .22
M 4 16 44
Total 11 .24 6 (7 ), 8 (5 ) 7, 9 (10), 4 (9)

a N = no. o f respondents.
b Yes = 1 point, no = 0 point.
c Italics indicate total mean in each group.

attitude items. Similarly, Dorian (1981) notes a gap between relatively positive
attitudes toward Gaelic . . . and willingness to see something done to implement those
attitudes (p. 174).
Despite possible methodological weaknesses involved in the application of an
instrument such as questionnaire V, it seems to me of value to compare the res
ponses obtained from the speakers in my sample, and to venture some explanations
for their differing degrees of behavioural commitment to supporting the maintenance
and strengthening of Mexican culture and of Spanish as a community language.
Table 6 .11 displays the results. The mean value is calculated by scoring yes
responses with 1 point, and no responses with zero. Thus, the higher the mean,
the stronger the commitment may be interpreted to be. Six speakers (distributed
equally in the three groups) responded no to all questions and did not sign the
questionnaire. The rest o f the sample all signed it, regardless of the number of
positive or negative responses.
 Conflicting Loyalties 205
The results displayed in Table 6 .11 indicate clearly that degree o f commitment
varies within each of the three groups. It is only when global results for each group
are compared that the willingness to participate actively in programme-oriented
activities organized to promote Spanish language and Mexican culture is seen to
decrease steadily and rather dramatically as family length of residence in the USA
increases. These results contrast with the positive attitudes towards Spanish that
speakers showed in their responses to questionnaire IV. At the same time, they
support Dorians and Fishmans observations that bilingual communities show a
dismaying gap between positive attitudes towards a receding or secondary lan
guage and willingness to do something or have something done to implement those
attitudes.
The commitment scale clearly shows that more respondents in Group 1 agreed
in answering yes to a larger number of items. The number of items agreed upon
diminishes, however, as we move down the proficiency continuum. This concurs
with the results of questionnaires II and III (see Table 6.8), which reveal a descent
in the number of most important reasons to maintain Spanish. Notable is the fact
that the more intellectual type of activity (question 6: Would you agree to attend
a lecture or conference on the topic of how persons of Mexican ancestry in Los
Angeles can improve their command of Spanish language and Mexican culture?)
receives the strongest support in Groups 2 and 3, and also considerable support in
Group 1. This response also accords with the support that item 2, the abstract
principle, received in questionnaires II and III. Consistent with this support is the
rejection by most speakers of the more militant proposal to join a protest meeting
against Los Angeles persons of Mexican ancestry who have stopped speaking and
reading the Spanish language.
Otherwise noteworthy in these responses is the fact that degree of commitment
can be almost exactly inverse between siblings or spouses. For instance, J7 (mi7,
1) responds no to all 9 questions, while his twin brother, M8 (mi7, 1), answers
yes to all. Likewise, R i7 s (f2i, 2) sole negative answer is to participating in a
protest meeting (item 7), but her sister, V21 (fi8, 2), does not commit herself to
4 (5, 7, 8, 9) of the 9 items. One further example comes from L 19 (f22, 2), who
answers yes only to item 9 (contribution of $15), while her brother, H (m2i, 2),
answers no to item 9, but yes to items 1, 2, 5, 6, and 8.

6.4.6. Conclusion
On the whole, then, the results of questionnaires I, II, and III indicate a correlation
among the following: decrease in the number and strength of the motivations
or reasons for maintaining or improving knowledge of Spanish, reduction in
the domains of use of Spanish in the community, and linguistic attrition of this
secondary language as reflected in the differing varieties characteristic of Groups
1, 2, and 3. Attitudes towards Spanish language and Mexican culture (question
naire IV, Table 6.10) remain generally very positive. However, the loyalty implicit
206 Conflicting Loyalties

in these positive attitudes seems to conflict with the uncommitted behaviour re

vealed by the responses to questionnaire V given by respondents in Group 2 and
especially by those in Group 3.
It is necessary to bear in mind, however, Fishmans suggestion (1971: 114) thai
certain populations [language-politicized and language-ideologized populations]
can report their language behaviors validly but others [the general population |
cannot. Our sample fits generally the category of general population.
 7

Conclusion

7.1. Summary

T h r e e of the main purposes of this final chapter are: to provide a summary of the
linguistic changes which have been examined in this book; in relation to these
individual changes, to review how distant the speech of Groups 2 and 3 is from the
Spanish dominant norm of Group 1; and to discuss some of the theoretical impli
cations of this books findings.
Let us first restate briefly that the general hypothesis investigated is that in
language-contact situations bilinguals develop strategies aimed at lightening the
cognitive load of having to remember and use two different linguistic systems. In
the use of the subordinate language, these strategies include: simplification of
grammatical categories and lexical oppositions; overgeneralization of forms, fre
quently following a regularizing pattern; development of periphrastic construc
tions, either to achieve paradigmatic regularity or to replace less semantically
transparent bound morphemes; and direct and indirect transfer of forms from the
superordinate language.1 These strategies facilitate the maintenance of the less
used language, they converge towards rendering communication more efficient (cf.
Maher 1991: 81), and indeed respond, as proposed by Thomason (1986: 250), to
the same factors that make simplifying internal changes natural, namely ease of
perception, ease of production, and overall ease of learning .
The linguistic phenomena analysed have illustrated the four strategies referred
to. Chapters 2 and 4 have given clear evidence of processes of simplification
affecting the verbal system, and the opposition between the copulas ser and estar,
as well as of the concomitant overgeneralization of estar, of certain simple tense
forms, and of the lexical reflexive verb pattern. Chapter 2 has also demonstrated
the trend to develop verbal periphrases (e.g. with hacer to do/make + nominal)
as a means to facilitate the production of conjugated forms by using simply one
conjugational paradigm. The extension of estar, as well as the phenomena examined
in Chapter 4 the omission of the complementizer que, the overall trend to place
subjects preverbally, and the differing loss of constraints on subject expression
were offered as possibly illustrating indirect transfer from English. In addition
to single-word loans and caiques, widely discussed in the existing literature on
contact phenomena, direct transfer from English was plainly attested in the cases
of multiple-word modelling (with possible syntactic implications) examined in
Chapter 6.
1 A fifth strategy, code-switching, has been dealt with only very briefly in Ch. 3.
208 Conclusion

One further important goal of this book has been the examination of the per
meability of Spanish to interference or transfer from English. I have examined
innovations in the speech of Groups 2 and 3, identified as such on the basis of a
comparison with the norms of the Spanish-dominant bilinguals in Group 1, in an
effort to clarify which types of change may be more appropriately explained n-,
intra- or interlinguistically motivated.
In.this respect, both spontaneous and non-spentaneous use of tense-mood-
aspect morphology reveals a clear trend towards the development of a simplified
grammaticalized verb system, independently of English influence (see Table 2.4).
In addition, it is intriguing to observe that speakers show a clear inclination to split
the verb lexicon in accordance with stative/non-stative lexical aspect, both at the
level of bound morphology (Imp for statives, Pret for non-statives) and in the dif
fering development of auxiliary constructions with hacer to make/do, and with estar.
These processes find interesting parallels in a good number of diverse studies deal
ing with language contact (attrition or development) and with language acquisition.
Regarding interlinguistic considerations, the generality of the processes identified
and the progression of simplification and loss of the different tenses in my data
agree in indicating that postulating direct influence from English is not justified.
Chapter 2 has argued that the simplest system of grammaticalized tense appears
to be more appropriately accounted for by intralinguistic, cognitive, and interactional
considerations (cf. Ferguson 1982: 59). This observation supports the hypothesis
that in language-contact situations a number of changes affecting the secondary
language have an internal motivation in that (a) they are in progress in the model
monolingual variety before intensive contact with another language occurs and/or
(b) they may be spurred by such language-specific features of the secondary
language as the semantic opaqueness of certain forms or the relative complexity of
a given paradigm. A similar point is made and supported by Mougeon and Beniak
(1991: ch. 5) with reference to changes in Ontarian French.
Chapter 4 deals with two independent changes that are also internally motivated
in the sense of (a) and (b) above: the generalization of estar to be and Cl place
ment. Furthermore, transfer from English does not appear to play a role in the
various patterns of omission of dative, accusative, and reflexive Cls. The complex
ity of the Spanish system and reduced use of and exposure to this language emerge
as more adequate reasons for these modifications.
Four possible cases of permeation of the syntax of Spanish by English gram
matical rules investigated in Chapter 5 lead also to the conclusion that transfer
from English is mostly indirect. For instance, knowledge of English probably
favours (in the sense that these phenomena are more frequent in the speech of
bilinguals in Groups 2 and 3) the non-expression of the complementizer que that,
the slight tendency to place subjects preverbally, and the selective retention of
some constraints on subject expression. However, none of these modifications
produces ungrammatical structures in Spanish, nor does the omission of so-called
 Conclusion 209
dative clitics of possession, which corresponds at best to what Mougeon and
Beniak ( 19 9 1: 18 8 - 9 ) ca^ an ambiguous change, i.e. one that may be predictable
both intra- and intersystemically . Crucially, however, with the exception of the
variable omission of the complementizer, the explanation for these modifications is
more adequately found in the loss of semantic-pragmatic constraints. The trend
towards a more fixed SV X word order, and at the same time towards a lower
frequency of subject expression in Groups 2 and 3, two opposing trends if English
is taken as the measure for comparison, is further proof that secondary languages
do not blindly caique the structures of a superordinate language.
As noted before by numerous linguists, calquing may be more easily identified
in the lexicon. A number of single-word and multiple-word items used by bilinguals
in Groups 1, 2, and 3, which deviate from the norms of more conservative Group
1 speakers, are shown to be the consequence of direct transfer of forms and/or
form meanings from English (Chapter 6).
The comparison of diverse linguistic phenomena across speakers who represent
varying lengths of residence in the USA, different degrees of decline of Spanish
language use, 'and dissimilar overall attitudes towards the maintenance of this
language reveals the points at which these linguistic phenomena manifest them
selves in relation to the extralinguistic factors associated with the bilingual pro
ficiency continuum. Table 7.x summarizes the linguistic phenomena analysed in
the three groups of speakers, as well as extralinguistic information about degree
of use of Spanish and degree of commitment to its maintenance as one of the
community languages.
The various sections in Table 7.1 have been adapted from tables or text in the
preceding chapters as follows: section I is based on Table 2.3. A plus sign between
columns in sections I and II is used to indicate that the content of the columns to
the left is added to the content of the columns on the right. Sections II, III, and
IV are adapted from Tables 4.6, 4.13, and 4.17 respectively. The matrix verbs
included in section IV are those that occur frequently enough to allow a comparison
across the three groups. Sections V, VI, and VII are based on Tables 5.2, 5.4, and
5.10 respectively. Sections III, V, and VII include the following abbreviations:
Refl = reflexive; inan = inanimate; S = subject; PerPro = personal pronoun; NP =
noun phrase; OthPro = other pronoun; MorphAmb = morphological ambiguity of
the verbal form; Coref = coreferentiality with the preceding subject. Section VIII
is from Table 6.1. Section IX comes from section 5.2. Section X contains three
subsections. The value for degree of use of Spanish (a) is based on the responses
to items 71, 72, 74 of Table 6.3. It is determined by assigning one point if the
response is either A (always) or F (frequently), and calculating the percentage o f
the total number of possible points in each group that the actual points assigned
represent. The value for agreement with positive items about Spanish (b) is based
on Table 6.10 (scale = 1-5 points). Finally, degree of commitment is based on
Table 6 .11.
 T able 7.1. Summary of Some Linguistic and Non-linguistic Phenomena Examined

I. Loss and simplification o f tense morphemes

II. Extension o f estar by adjective type
G roup i G ro up 2 G ro up 3 G ro up i G ro ups 2, 3
L: Fut P e rf + Fu t + IS Age + Moral
Cond (tense) P erf In f Cond (modal) Size Class
Cond Perf PluS PresPerf Physical
PPS Pluperf PS Colour
S: IS S: Imper
Pret
Imp
PS
III. Clitics omitted by clitic type IV. Preverbal clitics
G ro up i G ro up 2 G ro up 3 Matrix verb G ro u p i G ro up 2 G ro up 3
(No.) (No.) (No.)
(%) (%f (% )
Dative 2 4 15 Ir (a) 92 97 92
Obligatory Refi 1 2 h Estar 91 95 89
Accusative 2 I"!J Poder 60 95 83
True Refi 2 6 Tener que 57 75
Indirect Refi 65
10 Querer 32 55
Refi with inan S 52
4
Reciprocal 2
Total (% ) 0.1 0.8 3-8

V. Preverbal subjects VI. Subject pronoun expression

G ro up 3 G ro u p i G ro up 2 G ro up 3
G ro u p i G ro up 2
(%) (%) (%) (%) (%) (%)
28 26
PerPro. 85 9i 94 33
NP S6 66 62
OthPro 60 63 83

V II. Valid constraints on subject expression VIII. % modelling in every 10 minutes o f speech

G ro up 3 G ro up i G ro up 2 G ro up 3
G ro up i G ro u p 2

Yes Yes No (No.) (% ) (No.) (%) (No.) (%)

MorphAmb
Yes Yes 2 0.0 5 86 0.9 no 1 .3
Coref Yes

IX . % speakers omitting the complementizer

G ro up i G ro up 2 G ro u p 3
4 / 13 3 0 .7 % 6 /11 54 .5 % 9 /11 8 1.8 %

X . (a) Degree o f use o f Spanish with fam ily members (b) Degree o f agreement with positive items about Spanish (<r) Degree o f commitment to maintain
ancestors language and culture
G ro up i G ro up 2 G ro up 3
.29
(a) Use .69 30
(b) Positive items 4 .1 3-7 3-6
.62 .24
(f) Commitment 55
212 Conclusion

7.2. Discussion and Theoretical Im plications

7.2.1. External factors

Conspicuous in Table 7.x is the correlation between gradual distancing of Group*
2 and 3 from some of the linguistic norms of Group 1 speakers (sections I ~ I \ ), j
sharp reduction of Spanish language use from Group 1 to 2 and 3, a slight descent
in the strength of positive attitudes, and a significant decline in the degree of
commitment to act in order to maintain Spanish culture and language. Generally,
then, the extralinguistic factors illustrated in section X (Table 7.1) appear to be
good predictors of the degree o f linguistic attrition o f a secondary language in a
situation of societal bilingualism, as well as of the stages (according to the differing
length of family residence in the U SA represented in Groups 1, 2, and 3) at a hich
such strategies as simplification, overgeneralization, and transfer manifest them
selves with respect to different linguistic phenomena.
Table 7.1 shows that certain linguistic modifications occur only in the speech of
speakers in Group 3. Others are attested in Groups 2 and 3, but never (or onh
rarely) in the speech of those in Group 1. Interestingly, this is so regardless of
whether there has been a certain degree of decline in Spanish language use in this
group.
These results support in part our initial hypothesis regarding the correlation
between extralinguistic and linguistic factors (see Chapter 1). This hypothesis
proposed that changes affecting bilinguals Spanish were favoured and accelerated
by (a) absence of normative pressures on the subordinate language, (b) restriction
in the range of communicative uses of the subordinate language, and (c) speakers'
positive attitudes towards the superordinate language in contrast to either neutral
or negative attitudes toward the subordinate one.
In light of the results, proposals (b) and (c) may need to be slightly modified. In
the case of (b), given the absence of changes with respect to the ancestor variety in
some speakers in Group 1 who have evidently reduced Spanish usage, the degree
of restriction should be determined with more precision. This requirement may
also be justified by the very small difference in degree of use of Spanish which
emerges from the reports of Groups 2 and 3, although these reports must be taken
with caution. Furthermore, it is likely that reduction in use is to a certain extent
counterbalanced by the fact that Group 1 speakers have come to the U SA after age
eleven i.e. one may assume that they have been exposed to normative pressures
in elementary school in Mexico. As for (c), the attitudes towards English are
positive overall, but no negative attitudes towards Spanish are uncovered in any
group. Commitment, by contrast, declines in Group 2 and even more clearly in
Group 3. Thus, it appears that changes away from the norms of Group 1 are
favoured by or at least correlate with either neutral attitudes or weak com
mitment to maintaining the subordinate language.
These findings support Mougeon and Beniaks (1991: 221) hypothesis that it is
possible to account in a predictable and principled way for the association between
 Conclusion 213

individual instances of change attested in a contact setting and specific thresh

olds of bilingualism and frequency of use of the minority language. On the basis
of the specific outcome of my research regarding some speakers in Group x, I
would like to suggest that frequency of use in adolescence or adulthood should be
considered in relation to the possibility of a stage of pre-adolescent monolingualism
in the minority language.

7.2.2. Universality o f the processes

A number o f language processes, including first- and second language-acquisition,
foreigner talk, pidginization, creolization, and language loss, have been observed to
give rise to language varieties which share reduced or incompletely developed
features frequently described in terms of simplification1 (see several contributions
in Andersen 1983; Ferguson 1977; 1982; Meisel 1977; 1983/; Mougeon and Beniak
1991; Mhlhusler 1981; Romaine 1988; Trudgill 1983: ch. 6). Furthermore, it has
been claimed that these similarities are motivated by psycholinguistic universals
which underlie inter alia the order of acquisition and possibly also the order of loss
o f linguistic items (see e.g. Andersen 1989; Menn 1989).
With respect to the issue of universality, it has been observed that the four
strategies identified in the preceding chapters simplification, overgeneralization,
transfer, and analysis are indeed attested across different situations of linguistic
stress, and may justifiably be accounted for with reference to cognitive processes
which govern language acquisition and use in general.
Assuming that two of Slobins forty operational principles (cited in Andersen
1989: 387) for first-language acquisition are correct frequency and the separate
mental storage of different units the outcome of simplification, i.e. a simplified
system with fewer categories and oppositions, should obviouly be simpler to store,
remember, and use. A simpler system must also make more manageable the rapid
retrieval of less frequently used linguistic forms. Likewise, overgeneralization does
away with exceptions and restricted subclasses; analytical constructions are seman
tically more transparent or paradigmatically more regular than synthetic ones; and
convergence (when two or more languages are involved) through indirect or direct
transfer reduces the task of having to store, remember, and use different linguistic
rules. All these strategies converge, then, to lighten the cognitive load of having to
use two or more linguistic systems in daily communication and at the same time
contribute to rendering communication more efficient under the special circum
stances involved (see Chapter 3).
In the context of my work, I have considered linguistic simplification to be a
complex process which also implies rule generalization, in the sense that a given
form is being expanded to a larger number of contexts at the expense of a receding
2 T h is observation is not incompatible with the attested existence o f contact-induced complicating
changes (sometimes alongside simplifying ones), esp. in situations where the affected language is not
receding.
214 Conclusion

form. So defined, simplification of forms, of grammatical categories, and also

frequently, though not necessarily, of meanings is widespread in the data exam
ined. It could be easily demonstrated, however, that it is also prevalent in function
ally unreduced superordinate languages.3 While this is indeed the case, there are
two important differences between the two situations. One is that receding varie
ties do not develop linguistic alternatives for reduced or eliminated grammatical
categories (although they may for some lexical items). Simplification thus results in
simplified linguistic systems or subsystems (e.g. no gender-marking, fewer tense
forms, reduced case-marking)4 (see Taylor 1989: 179, who observes that the speech
of not-Mly-competent speakers of Gros Ventre shows the unreplaced loss of
some features, with the result that some things either cannot be expressed, or must
be expressed in roundabout ways).
The second difference concerns the speed of simplification in a contracting
secondary language. One generation in the life cycle of these languages is equiva
lent to many more than one in the non-receding corresponding language. The
same point about speed of change has been made by Markey (cited by Romaine
1989a: 380) in connection with the evolutionary rate of creoles: thirty years in the
life cycle of a creole might well be equivalent to three centuries in the life cycle of
a non-creole. In the case of creoles, the rapid changes leading to the development
of more complex linguistic systems (a contrast that has led linguists to view
attrition as a case of creolization in reverse) may be considered as processes of
language creation. Once a creole has reached maturity, so to speak, its rate of
change would not necessarily be different from any other language changes.
There seems sufficient evidence to maintain that language contact has the effect
of speeding up simplification and diffusion of changes. In regard to minority
languages, these tend to be changes already present in the non-contact ancestor
variety, i.e. they have intralinguistic roots (such as the rapid extension of estar and
the internally motivated trend to place Cls preverbally in verbal periphrases).
Maandi (1989) argues similarly that contact with Swedish has had the effect of
accelerating the internally motivated simplification of the objective case-marking
system of Estonian among Estonian immigrants in Sweden.
Among the most debated theoretical issues dealt with by those concerned with
language change (in growth, attrition, or normal life of languages) has been the
impact that one linguistic system may have on that of another. Concern with the
question of transfer (or interference) and substratum or superstratum influence is
evident in a number of recent publications (e.g. Thomason and Kaufman 1988;

3 Regarding Spanish in particular, the simplification processes attested in Groups 2 and 3 which
affect the verbal system and the omission o f Q s (as well as multiple-word modelling with syntactic
implications) have no counterparts in monolingual communities using the language; i.e. they probably
require some degree o f language-use (functional) restriction in a context o f daily use o f 2 or more
languages.
4 T h is is the sense in which the term simplification appears to be used most frequently in linguistic
writings. Observe that in m y framework simplification is viewed as a process which may lead to a
simplified system as a result.
 Conclusion 215
T a b le 7.2. Types of Linguistic Changes and Associated Triggering Factors

Group 1 Group 2 Group 3

Direct transfer
(a) Lexical: with syntactic implications No Yes Yes
without syntactic implications Yes Yes Yes
(b) Semantic-pragmatic Yes Yes Yes
Indirect transfer
Higher frequency of parallel structures Yes Yes Yes
Forgetting or incomplete learning
Loss of discourse-pragmatic constraints No Yes Yes
Loss of tense-mood-aspect inflections No Yes Yes
Loss of agreement markers (e.g. gender, case) No Yes Yes
Acceleration o f intralinguistic processes
Diffusion o f preferred variants of either Yes Yes Yes
stable or changing variables in the
ancestor varieties

Dorian 1989; Mougeon and Beniak 1991; Seliger and Vago 1991) and in many
works of the last hundred years or so (concerning the formation of Latin American
Spanish, for instance: Alonso 1939; Catalan 1958; Fontanella de Weinberg 1980;
de Granda 1968; 1971; 1991; Henriquez Urena 1921; Lapesa 1964; Lenz 1893;
Lope Blanch 1967; Lopez Morales 1980; Suarez 1966).
In planning her volume, Dorian (1989: 9) explicitly proposed as one of the
possible lines of investigation the search for an answer to the following question:
Is change in dying speech forms the result of interference, convergence, inde
pendent autogenetic processes or all of these? As may be evident by now, my
answer would be all of these . More importantly, however, my own research has
contributed in this respect by identifying in three groups of speakers some types
of change that may be expected to result from direct or indirect transfer, some that
probably result from forgetting or incomplete learning caused by reduction in
language use, and yet others that are the outcome of acceleration of independent
intralinguistic processes (see Table 7.2).
It is clear from Table 7.2 that Groups 2 and 3 are not qualitatively distinct with
regard to these broad classes of linguistic change. The dissimilarities are discov
ered when finer subclasses are distinguished and quantitative analyses performed.
Thomason and Kaufman (1988) and Campbell and Muntzel (1989) (CM) have
challenged the view that a grammatical system is impermeable to direct transfer of
foreign elements. On the basis of evidence provided by a number of studies,
Thomason and Kaufman argue that any linguistic feature can be transferred from
any language to any other language (p. 14), but their evidence stems from old
2l6 Conclusion
contact situations, not ongoing ones. In their consideration of structural changi-s it1
obsolescing languages, CM also surmise that several grammatical changes in
American Finnish can only be attributed to the external influence of English'
(p. I90).s
A brief examination of the examples provided, however, leads me to a different
conclusion. For instance, CM note that Standard Finnish (SF) does not permit
agentive phrases, but American Finnish (AF) does, replicating the English Lby phr.u.-
model. However, the assumed corresponding agentive phrases are not headed In
a by-type preposition; they contain case inflections or postpositions (translated by
CM as from (within), from (without), and through) that do not appear stried}
to mark agentivity. It is unclear, then, whether the examples provided do in faci
represent acts of syntactic reception. The same reservation is raised by another
proposed case of syntactic permeation of AF: the occurrence o f subjects with
infinitival complements of nouns or adjectives. But S F is said to not normalh
permit this structure, which implies that the structure does exist in S F under
certain perhaps more restricted conditions. Could this not be, then, an instance of
loss of semantic-pragmatic constraints leading to an extended use of the structure
in question?
These reservations and the outcome of my own research lead me to reassert that
it may be possible for any linguistic feature to be transferred from any language to
any other language as a nonce borrowing in the speech of bilinguals (cf. Weinreich
1974), but only those that are compatible with the structure of the borrowing
language (i.e. they correspond to an already existing syntactic variant in this
language, see Chapter 5) at any given stage will be adopted, disseminated, and
passed on to new generations.
The modifications that occur in a receding language under conditions of intense
contact and strong cultural pressure are first and foremost the result of simplifica
tion or overgeneralization of grammatical rules that do not introduce elements
which cause radical changes in the structure of the language. Since it seems
possible that these gradual changes may in time lead to the development of a
language fundamentally different from its non-contact ancestor, I would like to
reaffirm my proposal that the sociolinguistic history of the speakers is the primary
determinant of the language direction and degree of diffusion of the innovations,
as well as of the more remote linguistic outcome o f language contact; the structure
of the languages involved, to a large extent constrained by cognitive and interactional
processes, governs the introduction and diffusion of the innovative elements in the
linguistic systems. As suggested in Chapter 1, this hypothesis probably accounts
for the changes attested in numerous situations of language maintenance and/or
shift after several generations of normal language transmission (Thomason and
Kaufman 1988).
Weinreich (1974: 4) has stated that the cause of the susceptibility of a language

5 C M (p. 190) follow Jane H . H ill (whom they cite) in calling these changes acts o f reception .
 Conclusion 217
to foreign influence [lies] in its structural weaknesses. By contrast, my research
indicates that the permeability of a grammar to foreign influence does not depend
on its structural weaknesses but rather on the existence o f superficially parallel
structures in the languages in contact. It seems to me that my study has offered
evidence, on the other hand, to extend Weinreichs (1974: 14) statement about
phonological interference to the level of syntax, as follows. Transfer arises when a
bilingual identifies a lexical or syntactic structure of the secondary system with one
in the primary system and, in reproducing it, subjects it to the semantic-pragmatic
rules o f the primary language (see Chapters 5 and 6).
While no evidence of direct transfer of productive syntactic forms has been
uncovered, the speech o f US-born bilinguals gives evidence of incipient transfer of
discourse-pragmatic functions from the primary to the secondary language, and of
loss of semantic and pragmatic constraints that might be spurred on by the domi
nant language (indirect transfer). This loss of semantic or pragmatic constraints
leads to the overgeneralization of syntactic structures (e.g. certain word order
patterns).
I have found evidence of direct transfer (i.e. replacement and incorporation of
forms and/or meanings) at the level of the lexicon, which is affected in various
manners. Reproductions of multiple-word units, for instance, may have syntactic
implications, in that verb subcategorization or combinational possibilities may be
modified. Otherwise, the only additions to the syntax of the replica language come
through the internal arrangement of words in bound collocations. It is quite likely,
however, that innovative syntactic patterns introduced in bound collocations (or in
idioms) may in time become context-free and productive, in a manner similar to
what has been proposed for the borrowing of bound morphemes (Weinreich 1974:
3*)-
One question that arises in this connection, therefore, is the justification for the
common affirmation that the Spanish of US-born bilinguals is very different from
the Spanish of those in Group 1. It seems to me that what creates this impression,
exaggerated in my view, is basically simplification o f tense - mood - aspect mor
phology and gender agreement, and confusions in the use of prepositions. The fact
that almost every sentence contains one or more of these phenomena, and thus a
possible site for deviation from the norms of Group 1, appears to be a strong factor
in creating stereotypes o f widespread incorrectness and lack of systematicity.
The changes that characterize language loss, the manner in which languages
develop ontogenetically, and the evolution of languages in historical time have
been proposed as comparable in some respects (see e.g. Slobin 1977). Numerous
attempts have been made to investigate child language acquisition and language
loss in aphasie subjects in order to ascertain whether one would mirror the other
(see Gleason 1982). Likewise, non-pathological language loss (loss resulting from
forgetting a less-used language) has been proposed as a special kind of language
acquisition (Andersen 1982; 1989; Preston 1982; Seliger 1991) and, further, to be
very probably its reverse in terms of the order of retention of a number of language
 2l8 Conclusion

items. In this vein, for instance, and in the context of her study o f East Sutherland
Gaelic, a dying language, Dorian (1982) asks herself whether the loss of language
follows a recognizable pattern according to the complexity or frequency of the
items in question, and whether this pattern in any way parallels the pattern of
acquisition, two very complex questions which she can answer only partially.
One should heed Mougeon and Beniaks (1991: 185) remark that simple solu
tions to complex problems are simply wrong . However, it is tempting to consider
Menns (1989: 340) four broad levels of acquisition of any language rule or pat
tern (O) ignorance; (1) rote or formulaic knowledge; (2) pre-conventional or
overgenerative knowledge; (3) full conventionalized knowledge the mirror image
o f stages o f loss or de-acquisition of the rules of a minority language. Indeed, the
extreme degree of functional reduction that characterizes the Spanish of speakers
at the lowest levels of the continuum could be seen as a stage of rote or formulaic
knowledge. The Spanish they have retained appears to be pinned to a small
number of topics and situations which can trigger the use of Spanish in an almost
formulaic fashion.
Assuming this parallelism, and assuming that it is due to the workings of similar
psychological principles and/or to the freezing, as it were, of the process of
acquisition of the minority language at level 1 or 2 of those proposed by Menn,
then the linguistic structures retained by those in Group 3 may represent those
aspects of a language system that are most regular, frequent, and perceptually
salient (cf. Campbell and Muntzel 1989; Menn 1989; Mougeon and Beniak 1991).
I have argued and given evidence that four types of factor cognitive, social,
intralinguistic, and interlinguisticcontribute in complex and interactive ways to
the simplification, loss, and transfer of forms or meanings in the various stages of
restructuring of a receding language. This result also has implications for general
linguistic theory, inasmuch as it gives further evidence of the fact that the struc
ture of language is to a large extent constrained by cognitive and interactional
processes. This supports, then, Slobins (1977) assertions that the study of lan
guage during its changing phases is an excellent tool for discovering the essence
of language itself (p. 185), and that both change and structure are bound by the
same psycholinguistic and sociolinguistic constraints imposed by the processing of
speech in real time and in social settings (p. 186).
It is no wonder, therefore, that in language attrition in contact situations we
observe the same principles which characterize change in unstressed languages.
Our study gives evidence of the workings of at least the following: generality,
frequence, distance, and semantic transparency.
The principle o f generality predicts that, given two forms with at least one shared
structural context and with closely related meanings, the form with a wider struc
tural and pragmatic distribution, or for which there are no other alternatives in the
language, will be acquired first, retained, or lost later (see e.g. Maandi 1989; Prince
1992; Seliger 1991).
With respect to the verb system, for instance, this principle correctly predicts
 Conclusion 219

the earlier loss of the Cond Perf as opposed to the PluS in Mexican-American
Spanish (see Chapter 2). Observe (a) that, while the PluS may occur in the protasis
and apodosis of past counterfactual conditional clauses, the Cond Perf may occur
only in the apodosis,6 and no other tense may substitute for the PluS in the pro
tasis without a change in meaning ([1]); and (b) that, in contexts which refer to
possibility in the past, where the PluS does not occur, there are a number of
structures which may substitute for the Cond Perf ([2a-c.

[1] Si hubiera jugado, hubiera ganado (PluS)

* habra .. . habra (Cond Perf)
I f (he) had played, (he) wouldve won
[2] (a) Se habra ido ya cuando Uds. llegaron (Cond Perf)
* hubiera (PluS)
(He) must've already left when you arrived
(b) Se deba haber ido ya cuando Uds. llegaron
(He) mustve already left when you arrived
(c) Seguramente se haba ido ya cuando Uds. llegaron
Surely (he) had left when you arrived
The principle of generality refers to intralinguistic structural factors. I feel,
however, that it is related to circumstances which have to do with the nature of
communication. On the one hand, generality has frequency implications. More
general forms occur more frequently in discourse and tend, for this reason, to be
retained (longer) in receding languages (cf. Dorian 1983: 161), while infrequent
forms are either lost or never acquired (cf. Mougeon & Beniak 1991: 220).
On the other hand, consideration of the nature of communication leads me
to propose a further related principle, the principle o f distance (cf. Silva-Corvaln
1985: 565). According to this principle, if a language system has several forms in
the same syntactic-semantic sphere, the form which is farthest away from the
speaker, in the sense that it refers to objects or events which are the farthest from
him in his objective (e.g. actual distance) or subjective (e.g. possibility o f actual
ization) world, will tend to be lost first and acquired later. Interactional, and
perhaps also memory factors underlie this principle: since speakers tend to speak
about themselves and their immediate objective world rather than about distant
and hypothetical situations, the infrequently used forms, if any, will disappear
first, while the more frequently used ones will be acquired earlier or retained
longer.
Thus, consideration of the principle of distance makes it reasonable to expect
that, regarding the possible extension of a copula, estar rather than ser would be
extended because estar is the copula associated with concrete and/or immediate
experience (v. Ch. 4). M y research gives evidence that this is in fact so.

6 T h e Cond P e rf may occur in the protasis o f some varieties o f Spanish spoken in northern Spain.
It has not, to my knowledge, been reported in this context in any M exican Spanish variety.
220 Conclusion

morphology

F i g . 7 .1. Cycle o f changes in unstressed languages.

morphology --------------- lexicon --------- (syntax (periphrasis))

F i g . 7.2. Linear changes in language attrition

The development of periphrases and simplification support the cognitively

motivated principle o f semantic transparency, defined by Slobin (1977: 186) as a
tendency to maintain a one-to-one mapping between underlying semantic struc
tures and surface forms, with the goal of making messages easily retrievable for
listeners. The ideal iconic stage is never reached, however, because linguistic units
are not isolated; so changes in one item lead to other changes which frequently
result in polysemy. The general trend in language attrition seems to be toward
simplification and loss of forms, but the remaining ones may become semantically
more complex (although there is also greater reliance on contextual cues).
In conclusion, even though the same principles seem to be at work, there is one
major difference between attrition in language contact and complexification and
development in child language, creolization, and historical language change. In the
latter group of situations, language goes through cycles of simplification, loss, and
complexification (see the evolution of future tense forms in the history of Romance
in Fleischman 1982). By contrast, in language attrition the changes are linear and
lead to extensive reduction. Thus, while Slobins (1977) four charges to Lan
guage to be clear, processible, quick, and expressive appear to account in part
for the cycle of changes in unstressed languages illustrated in Fig. 7.1, in attrition
the cycle is broken and remorphologization does not occur, as shown in Fig. 7.2.
The linear changes in Fig. 7.2 are intended to represent the facts associated with
attrition and final language death. I have shown in this book that bilingualism
involving Spanish and English is not dying in Los Angeles. Rather, Hispanic
communities throughout this city give evidence of the wondrously complex socio-
linguistic phenomenon of societal bilingualism: Spanish illustrates a continuum
 Conclusion 221

of levels of proficiency along which speakers move, up or down, either in their life
time or across generations. (There is also a clear proficiency continuum in English,
characteristic of speakers in Group i.) This is a dynamic situation which, given
enough time depth and favourable social conditions, may lead to a cycle of changes
as pictured in Fig. 7.1.
Societal bilingualism offers a wealth of research possibilities. Essential relation
ships between language and society, between language and mind, between lan
guage and language can be found and examined in a situation of this type. Studies
of SpanishEnglish contact have made, and will most certainly continue to make
valuable contributions to our knowledge o f the nature o f the linguistic, cognitive,
and social processes underlying both possible changes in language maintenance
and changes characteristic of language shift and loss. This book is one more stage
in the advancement of our knowledge of the topic, and my impression is that in
some areas it has provided some firm foundations upon which we can continue to
build.
 Appendix i
Tape Contents and Speaker Coding

T ape Contents

Tape No. Speaker ID No. Coding

1 2 3 4 5 6 7 8 9 10

Group i
E L A 26, 64 Elisa V. i A B F A C I J N T P
E LA 37, 76 Silvia P. 2 A F F A C Z G N V P
E L A 45, 72 Lourdes C. 3 A F F A C Z J N TP
E L A 56 Mara G. 4 A B F A C I G N T P
E L A 13, 90 Fortino V. 5 A A M A C I G Z T P
ELA 51 Manuel C. 6 A D M A C Z J N T P
E L A 58, 91 Jos Luis A. 7 A B M A C G G K T P
ELA 59, 91 Moiss A. 8 A B M A C G G K T P
EL A 2, 3, 73 Alicia C. 9 A C F B C J G N V P
EL A 9, 10, 88 Consuelo S. 10 A D F B C J J N TP
E L A 16, 17, 18, 19, 77 R osaJ. 11 A C F B C J H N T P
E L A 27, 28, 44, 92 Eva R. 12 A C F B C J J N T P
E L A 4, 5, 6, 85 Carlos I. 13 A D M B C Z J N V P
ELA 17, 18, 77 Juan J. 14 A F M B C Z G N V P
E L A 28, 44, 95 Phil R. 15 A B M B C H J N T P
E L A 4 1, 81 Julin C. 16 A F M B C Z G N V P
Group 2
EL A 3, 4, 67, 68 R o syJ. 17 C A F A D I N K R R
E L A 32, 33, 70 Elsa 0 . 18 C A F A D I N Z R P
EL A 54, 82 Laura R. 19 A A F A D I N Z P P
EL A 26, 29, 61B, 64 Araceli C. 20 A A F A D I N K R P
E LA 16, 17 67 VirgieJ. 21 D A F A D I N K R P
EL A 11 , 50, 69 Hector R. 22 A A M A D I N K P P
E L A 49, 71 Hector Z. 23 C A M A D I N L R P
ELA 50 Robert R. 24 C A M A D I N L P R
ELA 55, 87 Mark Anth. R. 25 B A M A D I N L P R
EL A 4, 5, 6, 85 Mercedes I. 26 B A F B D J N L R P
 Appendix i 223

Tape No. Speaker ID No. Coding

i 2 3 4 5 6 7 8 9 xo

ELA 21, 22, 94, 22B Beatrice 0 . 27 B A F B D J N Z R P

E LA 23, 24, 75 Lucy A. 28 C A F B D J N Z R P
E LA 2, 3, 73 Alberto C. 29 B A M B D G N Z P P
E LA 33, 35, 36, 79 Enrique M. 30 A A M B D H N Z R P
E LA 39, 40 Ed C. 31 A A M B D H N Z P P
E LA 47, 48, 86 Gilberto S. 32 A A M B D G N Z R P
Group 3 ( special')
E L A 29 Belinda P. 33 C A F A M I N L R P
E L A 46, 83 Ana Mara G. 34 C A F B M J N Z P P
E L A 7, 8, 94 Rufina 0 . 35 A B F B M G N Z R P
E L A 43, 90 Diego V. 36 C A M B M P N K P S
E L A 21, 22, 94, 95 Alejo O. 37 B A M B M J N Z R P
Group 3
E LA 30, 3 1, 66 Sonia R. 38 C A F A E I N P R R
E L A 42, 61 Dolores Y. 39 C A F A E J N Z R S
E L A 48 Nora S. 40 F A F A E J N Z P P
E L A 34, 74 Marc Andr. A. 41 D A M A E I N Z R P
E L A 30, 63 Rodney R. 42 D A M A E G N L P R
E L A 57, 84 Joe R. 43 D A M A E G N P P R
E L A 62 Jos Rich. M. 44 F A M A E G N Z P R
ELA i, 25, 89 Esther Z. 45 D A F B E J N K P S
ELA 19, 20 Alma Ali. C. 46 F A F B E H N Z R P
E LA 52, 53, 80 Maria 0 . 47 C A F B E I N Z R P
E L A 23, 24, 75 Henry A. 48 C A M B E J N Z R P
E LA 31, 32, 87 Richard E. 49 B A M B E J N Z P P
E L A 36, 38, 78 Robert Rd. 50 B A M B E H N Z P R

F am ily Groupings

Parents Offspring Other relatives

1. Alicia (G ri, No. 9) Alma Ali. (Gr3, No. 46)

Alberto (Gr2, No. 29)
2. Rosa (G ri, No. xi) Rosy (Gr2, No. 17)
Juan (G ri, No. 14) Virgie (Gr2, No. 21)
3. Eva (G ri, No. 12) Laura (Gr2, No. 19) Robert (Gr2, No. 24)
Phil (G ri, No. 15) Hector (Gr2, No. 22) Mark (Gr2, No. 25)
224 Appendix i

Parents Offspring Other relatives

4. Carlos (G ri, No. 13)

Mercedes (Gr2, No. 26)
5. Julin (G ri, No. 16) Araceli (Gr2, No. 20) Hector (Gr2, No. 23
Belinda (Cr3, No. 33)
Ana M. (Cr3, No. 34)
6. Lourdes (Grx, No. 3) Elisa (G ri, No. 1)
Manuel (G ri, No. 6)
7- Jos L. (G ri, No. 7)
Moiss (G ri, No. 8)

8. Beatrice (Gr2, No. 27)

Alejo (Gr3, No. 37)
9. Lucy (Gr2, No. 28)
Henry (Gr3, No. 48
10. Gilberto (Gr2, No. 32) Nora (Gr3, No. 40)
11. Robert Rd. (Gr3, No. 50) Sonia (Gr3, No. 38) Richard (Gr3, No. 49)
Rodney (Gr3, No. 42)

Explanation o f Speaker Codings

I. Impressionistic level o f language proficiency or fluency

M y evaluation is based on such features as: pauses, self-corrections, smoothness

and speed of delivery, switching to English, requests for help with vocabulary, and
the speakers own subjective evaluation of the degree of difficulty or ease he has in
communicating in Spanish.
1 . Spanish: A - native; B = near native; C = good; D = fair; F = very poor.
2. English: A = native; B = near native; C = good; D = fair; F = very poor.

II. Ascribed social features

3. Sex: F = female; M = male.

4. Age: A = 15-29 years; B = 30--65 years.

III. Group by family length o f stay in the USA

5. Group 1:
C = both parents from Mexico; speaker came to the USA after age 11.
 Appendix i 225

Group 2:
D = both parents from Mexico, or one parent from Mexico and the other from
New Mexico or Texas; speaker was born in or came to Los Angeles before age
XI.
Group 3 (special):
M = parents and grandparents from Arizona, New Mexico, or Texas; speaker
has lived in Los Angeles for at least 19 years.
Group 3:
E both parents from California or one parent came to the U SA before age 11 ;
speaker has lived in Los Angeles all his life.
Note: Both subgroups (M and E) are referred to as Group 3, and treated as one
group unless otherwise stated in the text.

IV. Level o f formal schooling

6. English:
G = attending/incomplete high school
J = high school diploma
H = technical school beyond high school
I attending college
P = Ph.D. degree
Z = no formal schooling
7. Spanish as L i:
G = incomplete elementary
J = incomplete secondary
H technical before completion of secondary
/ = college
N = not applicable
8. Spanish as L2 in High School and/or College
Z = zero semesters; K = 2 semesters; L 3 - 4 semesters;
P 5-6 semesters; N = not applicable.

V. Age at which language acquisition started

9. English:
P = at birth
R = 3-6 years
S = 6 - 11 years
T = 12 -19 years
V = after 19 years
226 Appendix i
io. Spanish:
P at birth
R = reactivated at adolescence
5 = reactivated after 20 years of age
Note: Reactivated means that the speaker states that he made a conscious
decision and effort to use Spanish more frequently, and to improve his fluency in
this language. '
 Appendix 2
Verb Test

Ejemplos

1. Creo que no (voy a poder ir) para Mxico el prximo mes.

2. Cuando fui a visitarlos, me (llevaron) a ver una pelcula muy bonita.
3. Al nio le (hablan) puro ingls.
1. Algunas veces.........al parque el domingo en la tarde.
2. Eso l o .........el prximo ao.
3. Tus grados no estn buenos, hijo ,.........estudiar ms.
4. Bueno, pues, ella extraaba porque nunca.........lejos de la casa.
5. Si hubieras ido con nosotros al estadio,.........jugar a Fernando Valenzuela.
6. Es posible que y a .........de hacer los tamales.
7. Yo haca los dibujos como y o ...........
8. Cuando lo conoc.........unos 20 aos, pero no estoy seguro.
9. Nos invitaron a cenar, pero estamos tan atrasados. Seguro que ellos ya . . . .
cuando lleguemos a su casa.
10. La grabacin te qued muy m ala.........de hacerla otra vez.
11. A los 20 aos mi cuada ya s e .........con mi hermano.
12. Me dijo que yo tena que contestar la pregunta en la mejor forma que
13. No quiero q u e.........con esos muchachos.
14. Estando en la escuela, t .........tomado clases de espaol.
15. No s q u .......................lo que haba en la caja.
16. Cuando lleg el polica, le dijeron que alguien.........a robar.
17. Dicen q u e........... el precio de la carne maana.
18. Lo llev a la escuela para q u e.........con los otros chamaquitos.
19. Decan que l tena otra mujer, pero no s s i ......... verdad o no.
20. Ahora me arrepiento d e .........al party.
21. T tienes muchas caries,......... ir al dentista.
22. En mi aplicacin miraron que y o .........que saba hablar espaol.
23. El patio de mi casa.........bastante amplio, pero la yarda de enfrente
chiquita.
24. Cuando cumplas 5 aos de residente,.........ciudadano.
25. Qu edad tiene Pedro?
.........unos 35, pero no estoy seguro.
228 Appendix 2
26. Lo que mis padres hicieron por m, lo ........por mis hijos tambin, pero ya no
lo hice.
27. Entonces lleg y le pregunt que s i .........y me dijo, S; me com unos tacos.
28. Antes de haber ido t a ese lugar horrible, m e.........preguntado cmo era.
29. Bueno, si tuviera mucho dinero, m e......... un Cadillac.
30. No creo que......... para Mxico sin que nadie supiera.
31. Si te sientes m al.........descansar.
32. No s qu ........lo que hay en esta caja.
33. Es difcil describir mi nariz; cmo l a .........................t?

Approximate English translation

Examples

1. I dont think I (am going to be able to go) to Mexico next month.

2. When I went to visit them, they (took) me to see a movie.
3. To the child, they (speak) only in English to him.
1. Sometimes I ........ to the park on Sunday afternoons.
2. That I .........next year.
3. Your grades are not good, son,.........study harder.
4. Well, then, she felt homesick because she.........never away from home before.
5. If youd gone to the stadium with us, yo u .........Fernando Valenzuela play.
6. Its likely that they.........doing the tamales.
7. I made the drawing as I .........
8. When I met him, h e .........about 20 years old, but Im not sure.
9. They invited us to dinner, but were so late. Surely they...already when
we get to their place.
xo. The recording didnt turn out good........... do it again.
11. When she was 20, my sister-in-law.........already my brother.
12. He told me that I had to answer the question as best as I .........
13. I dont want you . . . . . with those guys.
14. Being in school, y o u .........taken Spanish lessons.
15. I dont know what....... that was in the box.
16. When the policeman arrived, they told him that someone....... to rob the
house.
17. They say that the price of m eat.........tomorrow.
18. I took him to school so h e .........with the other children.
19. They say he had another woman, but I dont know whether i t .........true or
not.
20. Now Im sorry fo r .........to the party.
21. You have a lot of cavities,......... go to the dentist.
 Appendix 2 229
22. They saw in my application that I .........that I knew how to speak Spanish.
23. M y backyard.........quite big, but the front garden.......... small.
24. When youve been a resident for five years, yo u .........a citizen.
25. How old is Peter?
.........about 35, but Im not sure.
26. What my parents did for me, I .........for my children as well, but I didnt do
it.
27. So then he came home and I asked him if h e ...........and he told me, Yes, I ate
some tacos .
28. Before going to that horrible place, yo u .........asked me what it was like.
29. Well, if I had a lot of money, I .........me a Cadillac.
30. I dont think that.........to Mexico without having told anyone.
31. I f you dont feel well, yo u .........rest.
32. I dont know what.........that is in this box.
33. Its difficult to describe my nose. H ow .........it?
 Appendix 3
Factors Included in the Coding of
Subject Expression

1. Dependent variable
1. Expressed subject
2. Non-expressed subject
2. Coreferentiality index with preceding sentence
1. Coreference with subject
2. Coreference with indirect object, switch reference with subject
3. Coreference with direct object, switch reference with subject
4. Switch reference with subject, direct object, and indirect object
5. Sentential subject
6. After pause
7. Coreference with oblique complement, switch reference with subject
8. Non-identifiable subject
3. Morphological ambiguity o f the verbal form
1. Ambiguous
2. Non-ambiguous
4. Contextual ambiguity*
1. Ambiguous subject referent despite discourse context
2. Non-ambiguous subject referent in discourse context
3. Subject referent identified by the verbal inflection
5. Contrast
1. Subject focus of contrast
2. Subject non-contrastive
6. Verb type
1. Reflexive form
2. Non-reflexive form
7. Verb syntax/semantics
1. Estimative (e.g. pensar to think, creer to believe, dear to say)
2. Gustar to like type verbs (e.g.faltarle to lack, ocurrirsele to occur to one)
3. Stative (e.g. estar to be locative, vivir to live, quedarse to stay, encontrarse
to be located)
4. Intransitive of activity (e.g. salir go out, pasar/ ocurrir to happen, irse to
leave, correr to run)
5. Copulative verb with nominal or adjectival predicate
 Appendix 3 231

6. Presentational (e.g. aparecer to appear, llegar to arrive, asomarse to show

up)
7. All others
8. Dear to say/tell as a verb of reporting
8. Second dependent variable: subject word order
1. Preverbal: transitive verb with direct or indirect object NP
2. Preverbal: transitive verb with verbal clitic and no object NP
3. Postverbal: transitive verb with direct or indirect object NP
4. Postverbal: transitive verb with verbal clitic and no object NP
5. Preverbal: intransitive verb
6. Postverbal: intransitive verb
7. Non-expressed subject
8. Copulative verb
9. Subject type
1. Y o Y
2. Tu/usted you-sg. familiar/formal
3. El/ella he/she/it
4. Nosotros/nosotras we-masc./fem.
5. Ustedes you-pl.
6. Elios/ellas they-masc./fem.
7. Noun phrase
8. Other pronouns
9. Tu you (interlocutor + others)
10. Ambiguous subject
10. Turn o f speech
1. New
2. Same
11. Clause type
1. Independent
2. Relative clause
3. Other subordinate clauses
4. Infinitival clause
12. Sentence type
1. Declarative
2. Interrogative
3. Exclamative
4. Imperative
13. Communicative value
1. Focal subject
2. Non-focal subject
3. Subject and another focal argument
* Although included in the coding chart, the data were not coded for this factor
because the very few ambiguous subjects that occurred (four cases in group 3)
were captured in 2.8: non-identifiable subject.
 Appendix 4
Attitude Questionnaires

Questionnaire I

This questionnaire grouped answers under the following headings: ALWS = always;
U S U L Y = usually; O F T = often; S M T M S = sometimes; N V R = never;
NA = does not apply. The instruction at the head of the questionnaire read:
Please tell me which column is correct for you.
1. I speak both Spanish and English to most Mexican friends my age.
2. Non-Mexicans visit me at my home.
3. I eat typically Mexican foods.
4. I belong to an organization primarily for Mexicans.
5. I have visited Mexico during the past year or two.
6. A member of my household has visited Mexico during the past year or two.
7. I think that being Mexican is different from being another kind of American.
8. I think I might be happier living in Mexico.
9. Most of my good friends are of Mexican ancestry.
10. I speak English to my mother and to other Mexican female adults.
11. I think that being Mexican is different from being another kind of Hispano.
12. When I go to church, I attend a Spanish service.
13. I speak primarily in Spanish to any Mexican friends who know both English
and Spanish.
14. Someone from Mexico has visited me or some member of my household
during the past year or two.
15. I speak Spanish when I get emotional or upset with a Mexican friend or
relative.
16. Education is one of my major interests.
17. Occupational success is one of my major interests.
18. Literature-art-music-drama are an area of prime interest to me.
19. I speak English to them when Iwantmy parents to do me a
20. I speak English to them when I wantmy grandparents to do
21. I speak English to them when Iwantmy children to do me
22. I think the husband should havethe final word on most problems that come
up in the family.
 Appendix 4 233
23. Finding non-Mexican friends is one of my major interests.
24. I read Spanish publications (newspapers, magazines, church bulletins).
25. Many Americans I have met are prejudiced against Mexicans.
26. Religion is one of my major interests.
27. I read the Bible in Spanish.
28. When I talk to the priest/minister at church I speak to her/him in Spanish.
29. In private I pray in English.
30. In public (or in a group of people) I pray in Spanish.
31. I attend worship services in Spanish.
32. I attend worship services in English.
33. I write poems, songs, or stories in Spanish.
34. I am a comadre or compadre to someone.
35. I have a madrina and/or a padrino.
36. I listen to Spanish radio programmes.
37. I feel as much at home among Americans as among Mexicans.
38. I am interested in travelling to places I have never visited before.
39. I write poems, songs, or stories in English.
40. I speak Spanish when I become very friendly or familiar with another Mexican.
41. I like (or would like) to watch Spanish T V programmes.
42. There are some people with whom I try to speak a better kind of Spanish.
43. I think it is important that Mexicans living in Los Angeles preserve their
customs and traditions.
44. I go to see Spanish movies or shows.
45. When I have children, I want them to be able to speak Spanish fluently.
46. Some Mexicans/Mexican-Americans give too much emphasis to being Mexi
can.
47. When I have children, I want them to be able to speak English fluently.
48. There are some people to whom I try to speak a better kind of English.
49. There are some Mexicans/Mexican-Americans who try to act too American.
50. I receive letters from relatives living in Mexico.
51. I rent videos in Spanish.
52. I rent videos in English.
53. I watch the news on T V on a Spanish channel.
54. I listen to Spanish music broadcasts.
55. I listen to radio stations in English.
56. I prefer to speak Spanish.
57. I prefer to speak English.
58. I feel more comfortable speaking English.
59. I feel more comfortable speaking Spanish.
60. I write letters in Spanish.
61. I use Spanish with my fellow workers.
62. I use Spanish with my boss or supervisor.
63. I count in Spanish.
234 Appendix 4
64. I make telephone calls in Spanish.
65. I discuss local affairs in Spanish.
66. I discuss national affairs in Spanish.
67. I discuss health in Spanish.
68. I prefer to speak Spanish with local people younger than myself.
69. I prefer to speak Spanish only if the other personaddresses me in Spanish.
70. I use Spanish when I speakwith my parents^
71. I use Spanish when I speakwith my grandparents.
72. I use Spanish when Ispeakwith my in-laws.
73. I use Spanish when I speakwith mybrothers and sisters.
74. I use Spanish when I speakwith my sisters-in-law.
75. I use Spanish when I speakwith mybrothers-in-law.
76. I am close with my cousins, uncles, and aunts.
77. I have friends that speak only English.
78. I have friends that speak only Spanish.
79. I f I came across a Hispanic-looking individual for the first time, I would speak
to her/him in Spanish if the person were:
(a) my age or younger
(b) a child
(c) older than me
(d) a salesperson
(e) a clerk (bank, post office, etc.)
( / ) a teacher
(f) a restaurant employee
(h) a police officer
(i) a car maintenance employee
( j ) someone on the street wearing a suit
(k) someone on the street wearing jeans and a T-shirt

Open-ended questions
80. Are there times when you should speak only Spanish or only English? Tell me
about them.
81. Do you switch from English to Spanish or vice versa? Why do you think you
do that? How do you feel about doing that?
82. What do the English speakers who dont know Spanish say when you use
Spanish? How do you feel about that?
83. Where do you usually buy your groceries?

Questionnaire I

The items are grouped according to dimensions investigated. Numbers in paren

theses correspond to parallel statements about Spanish and English.
 Appendix 4 235
IA. Use of Spanish in relation to interlocutors: 1, 13, 15, 40, 42 (48), 61, 62, 70
(19), 71 (20), 72, 73, 74, 75, 28, 79.
IB. Use of English in relation to interlocutors: 10, 19 (70), 20 (71),21, 48 (42).
IIA. Use of Spanish in relation to topic or type of oral activity: 12, 30,31 (32,
12), 63, 64, 65, 66, 67.
IIB. Use of English in relation to topic or type of oral activity: 29, 32 (12).
IIIA. Writing and reading in Spanish: 24, 33 (39), 60, 50.
M B. Writing and reading in English: 39 (33).
IVA. Preferential use of Spanish: 56 (57), 59 (58), 68, 69.
IVB. Preferential use of English: 57 (56), 58 (59).
VA. Listening in Spanish: 36 (55), 41, 44, 51 (52), 53, 54.
VB. Listening in English: 52 (51), 55 (36).
VIA. Positive attitudes toward Mexican culture: 3, 4, 7, 8, 9, ix, 22, 34, 35, 43,
45, 49, 76, 78-
VIB. Positive attitudes toward mainstream culture: 2, 23, 37, 47, 77.
VIIA. Negative attitudes toward Mexican culture: 46 (25).
VIIB. Negative attitudes toward mainstream culture: 25 (46).
VIIIA. Contact with Mexico: 5, 6, 14.
IX. General attitudes: 16, 17, 18, 26, 38.

Questionnaire II: Spanish D om inants Form

Instructions

When answering the questions below, circle the ones that seem important to you,
put a star by the most important one or two, and draw a line through any that are
particularly unimportant to you.

Abrviations used in Questionaires I I and I I I

The letters in front of the numbers correspond to the classification of Mejias

and Anderson (1988) (MA): G = group identifying (V or value in MA);
I = instrumental; S = sentimental; C = communicative. These roughly correspond
to four of Dorians (1981: 165-71) reasons: G = local integrative and tradition; I =
operational; S = aesthetic. For purposes of comparison with MA, I have consid
ered items 2 and 16 instrumental rather than abstract principle and exclusionary,
as in Dorian.

I m glad I know Spanish because:

G i. Its the language of my ancestors.

I 2. Its broadening to have more than one language.
S3. I can enjoy Mexican music better.
236 Appendix 4
S 4. Spanish is a very rich and expressive language.
C 5. It is necessary for daily communication.
G 6. It makes me feel more a part of the community I live in.
C 7. I can read in Spanish, for example, the Bible or newspaper columns.
I 8. Its useful to have a secret language that not everyone else understands.
G 9. This part of the world was always Spanish-speaking and Im keeping that
tradition alive.
G 10. Its the language of my friends and neighbours.
C 11. I can understand the Spanish programmes on T V and radio.
C 12. I can talk to people from other Spanish-speaking countries.
S 13. Spanish is a beautiful language to hear and speak.
I 14. It helps me make money at my job.
S 15. It makes me feel good about myself.
I 16. I use it at my job.
If your main reasons are not given above, please tell me what they are or write
them here:

Questionnaire III: English dominants form

Instructions

When answering the questions below, circle the ones that seem important to you,
put a star by the most important one or two, and draw a line through any that are
particularly unimportant to you.

I f I were ever to try to improve or maintain my Spanish,

my main reasons would be:

G 1. Some or all of my ancestors were Spanish-speaking.

I 2. Its broadening to have more than one language.
S 3 . I would be able to enjoy Spanish music better.
S 4. Spanish is a very rich and expressive language.
C 5. Because it is necessary for daily communication.
G 6. It would make me feel more a part of the community I live in.
C 7. I would be able to read in Spanish, for example, the Bible or newspaper
columns.
I 8. Its useful to have a secret language that not everyone else understands.
G 9. This part of the world has always been Spanish-speaking in the past, and I
would be proud to keep that tradition alive.
G 10. Its the language of my friends and neighbours.
C 11. I would understand the Spanish programmes on T V and radio.
C 12. I would be able to talk with Spanish speakers from other parts of the
Spanish-speaking world in Spanish.
 Appendix 4 237
S 13. Spanish is a beautiful language to hear and speak.
I 14. It would help me make money at my job.
S 15. Because I do not want to lose my language.
I 16. It is useful to know a second language.
I f your main reasons are not given above, please tell me what they are or write
them here:

Questionnaire IV

Instructions

For the statements below, please check the box which expresses best how you feel:
whether you agree with the statement, are uncertain about it, or disagree with it.

Abbreviations used in Questionaire I V

The letters in front of the numbers correspond to the six groups identified in
Dorians questionnaire III (1981: 172-8). A = fact question; B = attitudes (Span
ish); C = goal-directed questions; D = programme-oriented; E = attitudes (Eng
lish); F = special attitude towards English. A + after these letters indicates that
the item conveys a positive attitude, a indicates that the item conveys a negative
attitude. The questionnaire grouped answers under the following headings:
S T N G L Y A G R = strongly agree; A G R = agree; U N C R T N = uncertain;
D ISA G R = disagree; S T N G L Y D ISA G R = strongly disagree.

Section A

B + x. I like to hear Spanish spoken.

B + 2. We should work hard to save the Spanish language.
B - 3- As all Mexican people speak English, it is a waste of time to keep up with
Spanish.
A 4 . Spanish is a difficult language to learn.
B - S - There are far more useful things to spend time on than Spanish.
B + 6 . Spanish is a language worth learning.
B - 7- Spanish has little value in the modem world.
C + 8. I should like to be able to read books in Spanish.
B + 9. Anyone who learns Spanish will have plenty of chances to use it.
B - 10. There is no need to keep up Spanish for the sake of tradition.

Section B

B + 1. I want to keep up Spanish in order to help the Mexican people.

B - 2. Speaking Spanish wont help you get a job.
238 Appendix 4
B+ You cant be a real Mexican/Mexican-American without Spanish.
3-
D- Learning Spanish or not should be left to a persons own choice.
4-
B+ We owe it to our ancestors to keep Spanish alive.
5-
C+ 6.
I should like to be able to understand the Spanish songs on the radio and
television.
D - 7- School time should be used for more practical subjects than Spanish.
B + 8. Spanish has a beauty all its own. *
B - 9- Its looking backward instead of forward to try to keep Spanish alive.
D + 10. More radio and television time should be given in Spanish.

Section C
E+ 1. English should be taught in all countries.
E- 2. Mexicans in the U SA should speak Spanish and not a foreign language.
E+ 3. English will take you further than Spanish.
E+ 4. Its wrong to teach in Spanish in the elementary schools when Spanish
speaking children should be learning English.
F+ 5. You are considered to belong to a higher class if you speak English.
E+ 6. Those who dont want to learn English shouldnt come to live in Los
Angeles.
E+ 7. English is a beautiful language.
E+ 8. English is better for studying scientific subjects than Spanish.
C- 9. Spanish will become less important in Los Angeles in the future.
E- 10. English will become less important in Los Angeles in the future.

Questionnaire V

Instruction: Please answer my questions with either Y ES or NO.

1. Would you agree to participate in a small-group discussion, with other persons
of Mexican origin in Los Angeles, on the topic of improving your command
of Spanish language and Mexican culture?
2. Would you agree to have as your room-mate in college a person of Mexican
origin who preferred to speak in Spanish?
3. Would you agree to spend a weekend at the home of another person of
Mexican ancestry in Los Angeles who wanted to discuss with you how to
improve your command of Spanish language and Mexican culture?
4. Would you agree to invite another person of Mexican ancestry to spend a
weekend at your home in order to discuss with her (him) how to improve your
command of Spanish language and Mexican culture?
5. Would you agree to join a club for people of Mexican origin in Los Angeles
who are interested in improving their command of Spanish language and
Mexican culture?
 Appendix 4 239

6. Would you agree to attend a lecture or conference on the topic of how persons
of Mexican ancestry in Los Angeles can improve their command of Spanish
language and Mexican culture?
7. Would you agree to join a protest meeting against Los Angeles persons of
Mexican ancestry who have stopped speaking and reading the Spanish lan
guage?
8. Would you agree to attend a meeting of a local chapter of a Mexican-American
organization for the strengthening of the use of Spanish in Los Angeles?
9. Would you, if asked, agree to contribute $15.00 to help finance the activities
of a Mexican-American organization for the strengthening of the use of Spanish
in Los Angeles?
10. I f you have answered yes to any of the above please give your name, address,
and telephone number.
 References

A d a m s, M. (1987), From Old French to the Theory of Pro-drop. Natural Language and
Linguistic Theory, 5: 1-32.
A g h e y is i, R., and F is h m a n , J. A. (1970), Language Attitude Studies: A Brief Survey of
Methodological Approaches. Anthropological Linguistics, 12: 137-57.
A is se n , J., and P e r l m u t t e r , D. (1976), Clause reduction in Spanish, in H. Thompson et
al. (eds.), Proceedings o f the Second Annual Meeting o f the Berkeley Linguistics Society, 1-
30 (Berkeley, Calif.: Berkeley Linguistics Society).
A l o n s o , A . (1939), Examen de la teora indigenista de Rodolfo Lenz, Revista de Filologa
Hispnica, 1: 331-50.
A m a s ta e , J . , and E l a s - O l i v a r e s , L. (eds.) (1982), Spanish in the United States: Socio-
linguistic aspects (New York: Cambridge Univ. Press).
A n d e r s e n , R. W. (n.d.), Interpreting Data: Second Language Acquisition of Verbal
Aspect (MS, Univ. of California at Los Angeles).
----- (1982), Determining the Linguistic Attributes of Language Attrition, in R. D. Lambert
and B. F, Freed (eds.), The Loss o f Language Skills, 83-118 (Rowley, Mass.: Newbury House).
----- (ed.) (1983), Pidginization and Creolization as Language Acquisition (Rowley, Mass.:
Newbury House).
----- (1989), The up and down Staircase in Secondary Language Development, in
Dorian (1989: 385-94)-
----- (1991), Developmental Sequences: The Emergence of Aspect Marking in Second
Language Acquisition, in T. Huebner and C. A. Ferguson (eds.), Crosscurrents in Second
Language Acquisition and Linguistic Theories, 305-24 (Amsterdam: Benjamins).
A n t t i l a , R. (1972), An Introduction to Historical and Comparative Linguistics (New York:
Macmillan).
B a r r e n e c h e a , A ., and A l o n s o , A . (1977), Los pronombres personales sujetos en el espaol
hablado de Buenos Aires, in J. L. Blanch (ed.), Estudios sobre el espaol hablado en las
principales ciudades de Amrica, 333-49 (Mexico, DF: Universidad Nacional Autnoma
de Mxico).
B a v in , E . L. (1989), Some Lexical and Morphological Changes in Warlpiri, in Dorian
(1989: 267-86).
B e h r e n d , E. (1986), The Use of Ser and Estar by Bilingual Mexican Americans in the
Chicago Area: A Languages-in-Contact Study (MA thesis, Hamburg Univ.).
B e n t iv o g li o , P. (1987), Los sujetos pronominales de primera persona en el habla de Caracas
(Caracas: Universidad Central de Venezuela).
B e n v e n is te , E. (1968), Mutations of Linguistic Categories, in Lehmann and Malkiel
(1968: 83-94).
B e r g e n , J . (ed.) (1990), Spanish in the United States: Sociolinguistic Issues (Washington, DC:
Georgetown Univ. Press).
B ic k e r t o n , D. (1975), Dynamics o f a Creole System (Cambridge: Cambridge Univ. Press).
 References 241
----- (1981), Roots o f Language (Ann Arbor, Mich.: Karoma).
B ills , G. (ed.) (1974), Southwest A real Linguistics (San Diego, Calif.: Institute for Cultural
Pluralism).
----- (1989), The US Census of 1980 and Spanish in the Southwest, in Wherritt and
Garcia (1989: 11-28).
B lo o m f i e ld , L. (1933), Language (New York: Holt, Rinehart & Winston).
B o w e n , J. D., and O r n s t e in , J. (eds.) (1976), Studies in Southwest Spanish (Rowley, Mass.:
Newbury House).
B r o w n , R. (1973), A First Language (Cambridge, Mass.: Harvard Univ. Press).
B y b e e , J . (19 8 5 ), Morphology: A Study o f the Relation between Meaning and Form (A m sterd am :
B e n jam in s).
C a m e ro n , R. (1990), Variable Constraints on the Alternation of Subject Pronouns and
Empty Subjects in Puerto Rican Spanish (thesis proposal, Univ. of Pennsylvania).
C a m p b e ll, L., and M u n t z e l , M . C. (1989), The Structural Consequences of Language
Death, in Dorian (1989: 181-96).
C a n t e r o S a n d o v a l, G. (19 7 6 ), Peculiaridades en el empleo del pronombre personal y o en
el habla culta de la ciudad de Mxico, Anuario de Letras, 14 : 2 3 3 - 7 .
C a t a l n , D. ( 19 5 8 ), Gnesis del espaol atlntico: Ondas varias a travs del Ocano,
Revista de Historia Canaria (La Laguna), 24: 1 - 1 0 .
C i f u e n t e s , H. (1980-1), Presencia y ausencia del pronombre personal sujeto en el habla
culta de Santiago de Chile, in Homenaje a Ambrosio Rabanales: Boletn de Filologa de la
Universidad de Chile, 31: 743-52.
C o le m a n , L., and K a y , P. (1981), Prototype Semantics: The English Word Lie, Language,
57: 26-44.
C o m rie , B. (1976), Aspect (Cambridge: Cambridge Univ. Press).
----- (1985), Tense (Cambridge: Cambridge Univ. Press).
CoppsAG, D., and E s c u d e r o , G. (1966), Ser y estar en espaol y en rumano, R em e
roumainc de linguistique, 11: 339-49.
C r u s e , D. A. (1986), Lexical Semantics (Cambridge: Cambridge Univ. Press).
D o r ia n , N. (1973), Grammatical Change in a Dying Dialect, Language, 49: 414-38.
----- (1978), The Fate of Morphological Complexity in Language Death, Language, 54:
590-609.
----- (1980), Maintenance and Loss of Same-Meaning Structures in Language Death,
Word, 31: 39-45.
----- (1981), Language Death (Philadelphia: Univ. of Pennsylvania Press).
----- (1982), Linguistic Models and Language Death Evidence, in L. Obler and L. Menn
(eds.), Exceptional Language and Linguistics, 31-48 (New York: Academic).
----- (1983), Natural Second Language Acquisition from the Perspective of the Study of
Language Death, in Andersen (1983: 158-67).
----- (ed.) (1989), Investigating Obsolescence: Studies in Language Contraction and Death
(Cambridge: Cambridge Univ. Press).
DuBois, J. W. (1985), Competing Motivations, in J. Haiman (ed.) Iconicity in Syntax, 343-
65 (Amsterdam: Benjamins).
E d w a r d s , J. (1985), Language, Society and Identity (Oxford: Blackwell).
E l a s - O l i v a r e s , L. (1979), Language Use in a Chicano Community: A Sociolinguistic
Approach, in J. B. Pride (ed.), Sociolinguistic Aspects o f Language Learning and Teaching,
120-34 (Oxford: Oxford Univ. Press).
242 References
L. (ed.) (1983), Spanish in the United States: Beyond the Southwest
E la s -O liv a r e s ,
(Washington, DC: National Center for Bilingual Education).
E n r q u e z , E . V . (19 8 4 ), E l pronombre personal sujeto en la lengua espaola hablada en Madrid
(Madrid: Consejo Superior de Investigaciones Cientficas).
F a l k , J. (1979), SER y ESTAR con atributos adjetivales (Uppsala: Almqvist & Wiksell).
F e r g u s o n , C. A. (1977), Absence of Copula and the Notion of Simplicity: A Study of
Normal Speech, Baby Talk, Foreigner Talk and Pidgin, in D. Hymes (ed.), Pidginization
and Creolization o f Language, 141-50 (Cambridge: Cambridge Univ. Press).
----- (1982), Simplified Registers and Linguistic Theory, in L. Obler and L. Menn (eds.),
Exceptional Language and Linguistics, 49-66 (New York: Academic).
F e r n n d e z , S. (1964), Un proceso lingstico en marcha, in OFINES (1964: ii. 277-85).
F is h m a n , J. A. (1971), Bilingual Attitudes and Behaviors, in Fishman et al. (1971: 105-
16).
------- C o o p e r, R. L., and M a , R. ( 19 7 1) , Bilingualism in the Barrio (Bloomington, Ind.: Indiana
Univ. Press).
------- and K e l l e r , G. D. (eds.) (1982), Bilingual Education for Hispanic Students in the United
States (New York: Teachers College Press).
F l e i s c h m a n , S. (1982), The Future in Thought and Language: Diachronic Evidence from
Romance (Cambridge: Cambridge Univ. Press).
F o n t a n e l l a d e W e in b e rg , M. B. (1980), Espaol del Caribe: rasgos peninsulares, contacto
lingstico o innovacin?, Lingstica Espaola Actual, 2: 189-201.
F r a k e , C. O. (1977), Lexical Origins and Semantic Structure in Philippine Creole Span
ish, in D. Hymes (ed.), Pidginization and Creolization o f Languages, 223-42 (Cambridge:
Cambridge Univ. Press).
F r a n c o , J. (n.d.), Loan Translation in Prepositions Subcategorized by Verbs in LA Span
ish (MS, Univ. of Southern California).
----- (1991), Spanish Object Clitics as Verbal Agreement Morphemes, in J. D. Bobaljik
and T. Bures (eds.), M IT Working Papers in Linguistics, 14: 99-114.
G a l , S. (1979), Language Shift: Social Determinants o f Linguistic Change in Bilingual Austria
(New York: Academic).
----- (1984), Phonological style in bilingualism: The Interaction of Structure and Use, in
D. Schiffrin (ed.), Meaning, Form and Use in Context, 290-302 (Washington, DC:
Georgetown Univ. Press).
G a r c a , E . ( 19 7 5 ) , The Role o f Theory in Linguistic Analysis: The Spanish Pronoun System
(Amsterdam: North-Holland).
----- (1983) Context Dependence of Language and of Linguistic Analysis, in F. Klein-
Andreu (ed.), Discourse Perspectives on Syntax, 181-207 (New York: Academic).
----- v a n P u t t e , F., and T o b in , Y. (1987), Cross-linguistic Equivalence, Translatability,
and Contrastive Analysis, Folia Lingistica, 21: 373-405.
G i l y G a y a , S. (1970), Curso superior de sintaxis espaola (Barcelona: Bibliograf).
G rv N , T. (1976) Topic, Pronoun and Grammatical Agreement, in C. N. Li (ed.), Subject
and Topic, 149-88 (New York: Academic).
----- (1979), Prolegomena to Any Sane Creology, in I. F. Hancock (ed.), Readings in Creole
Studies, 3-35 (Ghent: Story-Scientia).
G l e a s o n , J. B. (1982), Converging Evidence for Linguistic Theory from the Study of
Aphasia and Child Language, in L. Obler and L. Menn (eds.), Exceptional Language and
Linguistics, 347-56 (New York: Academic).
 References 243
G ran d a, G . de (1968), Transculturacin e interferencia lingstica en el Puerto Rico contemporneo
(j 8 8- i 6) (Bogota: Instituto Caro y Cuervo).
----- (1971), Algunos datos sobre la pervivencia del criollo en Cuba. Boletn de la Real
Academia Espaola, 51: 48191.
------- ( 1 9 9 1 ) , E l espaol en tres mundos: Retenciones y contactos lingsticos en Amrica y Africa
(Valladolid: Secretariado de Publicaciones, Universidad de Valladolid).
G u m p erz, J., and W i l s o n , R. (1977), Convergence and Creolization, in D. Hymes (ed.),
Pidginization and Creolization o f Languages, 151-67 (Cambridge: Cambridge Univ. Press).
G u t i r r e z , M. (1989), Michoacan Spanish/Los Angeles Spanish: Trends in a Process of
Linguistic Change (Ph.D. diss., Univ. of Southern California).
----- (1990), Sobre el mantenimiento de las clusulas subordinadas en el espaol de Los
Angeles, in J. J. Bergen (ed.), Spanish in the United States: Sociolinguistic Issues, 318
(Washington, DC: Georgetown Univ. Press).
H a i m a n , J., and T h o m p s o n , S. (1984), Subordination in Universal Grammar, in C.
Brugman and M. Macaulay (eds.), Proceedings o f the Tenth Annual Meeting o f the Berkeley
Linguistics Society, 510-23 (Berkeley, Calif.: Berkeley Linguistics Society).
H e n r q u e z U r e a , P. (1921), Observaciones sobre el espaol de Amrica, Revista de Filologa

Espaola, 8: 357-90.
H e r n n d e z - C h v e z , E. et al. (eds.) (1975), E l lenguaje de los chcanos (Arlington, Va.: Center
for Applied Linguistics).
H i l l , J. H . (1989), The Social Functions of Relativization in Obsolescent and Non-
obsolescent Languages, in Dorian (1989: 149-64),
H o c h b e r g , J. G. (1986), Functional Compensation for /s/ Deletion in Puerto Rican

Spanish, Language, 62: 609-21.

H u d s o n - E d w a r d s , A., and H u d s o n - E d w a r d s , A. (1986), Syntactic, Semantic, and Prag
matic Influences on Judgments of Grammaticality, in K. Ferrara et al. (eds.), Linguistic
Change and Contact, 137-42 (Austin: Univ. of Texas Press).
J a c k s o n , H. (1988), Words and Their Meaning (London: Longman).
J a k o b s o n , R. (1938), Sur la thorie des affinits phonologiques entre des langues, Actes du
Quatrime Congrs International de Linguistes, 48-59 (Copenhagen: Munksgaard).
J o n g e , B. d e (1987), Estar comes of age, in F. Beukema and P. Coopmans (eds.), Linguistics
in the Netherlands, 101-10 (Dordrecht: Foris).
K e e n a n , E., and C o m r i e , B. (1977), Noun Phrase Accessibility and Universal Grammar,
Linguistic Inquiry, 8: 63-99.
K i n g , R. (1989), On the Social Meaning of Linguistic Variability in Language Death
Situations: Variation in Newfoundland French, in Dorian (1989: 139-48).
K l e i n , F . (19 8 0 ), A Quantitative Study of Syntactic and Pragmatic Indications of Change
in the Spanish of Bilinguals in the US, in W. Labov (ed.), Locating Language in Time and
Space, 6 9 - 8 2 (New York: Academic).
K l e i n - A n d r e u , F. (1986a), La cuestin del anglicismo: apriorismos y mtodos, Thesaurus:
Boletn del Instituto Caro y Cuervo, 40: 1-16.
----- (19861!), Speaker-Based and Reference-Based Factors in Language: Non-past Con
ditional Sentences in Spanish, in O. Jaeggli and C. Silva-Corvaln (eds.), Studies in
Romance Linguistics, 99-119 (Amsterdam: Foris).
K l e i n , W. (1986), Second Language Acquisition (Cambridge: Cambridge Univ. Press).
----- and v o n S t u t t e r h e i m , C. (n.d.), Text Structure and Referential Movement (MS,
Max-Planck-Institut fr Psycholinguistik, Nijmegen).
244 References
K o ik e , D. A. (1987), Code Switching in the Bilingual Chicano Narrative, Hispania, 70:
148-54.
L a b o v , W. (1972a), Sociolinguistic Patterns (Philadelphia: Univ. of Pennsylvania Press).
----- (1972^), Language in the Inner City (Philadelphia: Univ. of Pennsylvania Press).
----- (ed.) (1980), Locating Language in Time and Space (New York: Academic).
----- (1981), What Can Be Learned about Change in Progress from Synchronic Descrip
tion? in D, Sankoff and H. Cedergren (eds.), Variation Omnibus, 177-99 (Edmonton:
Linguistic Research).
----- (iqHib), Resolving the Neogrammarian Controversy, Language, 57: 267-308.
----- (1982), Building on Empirical Foundations, in W. P. Lehmann and Y. Malkiel
(eds.), Current Issues in Linguistic Theory, xxiv: 17-92 (Amsterdam: Benjamins).
----- and W a l e t s k y , J. (1967), Narrative Analysis: Oral Versions of Personal Experience, in
J. Helm (ed.), Essays on the Verbal and VisualArts, 12-44 (Seattle: Univ. of Washington Press).
L a m b r e c h t , K. (1987), On the Status of SVO Sentences in French Discourse, in R. S.
Tomlin (ed.), Coherence and Grounding in Discourse, 217-61 (Amsterdam: Benjamins).
L a n d a , A. (1989), Posicin de los clticos en construcciones con perfrasis verbales en el
habla culta de Caracas (MS, Univ. of Southern California).
----- (r99 S)> Conditions on Null Objects in Basque Spanish and Their Relation to Lesmo
and Clitic Doubling (Ph.D. diss., Univ. of Southern California).
L a p e s a , R. (1964), El andaluz y el espaol de Amrica, in OFINES (1964: ii. 173-82).
L a v a n d e r a , B. (1978), Where Does the Sociolinguistic Variable Stop?, Language in
Society, 7: 171-83.
L e h m a n n , W., and M a l k i e l , Y. (eds.) (1968), Directions fo r Historical Linguistics (Austin:
Univ. of Texas Press).
L e n n e b e r g , E. H. (1967), Biological Foundations o f Language (New York: Wiley).
L e n z , R. (1893), Beitrge zur Kenntnis des Amerikanospanisch, Zeitschrift f r Romanische
Philologie, 17: 188-214.
L e v e l t , W. (1979), Linearization in Discourse (MS, Max-Planck-Institut fr Psycho
linguistik, Nijmegen).
----- (1982), Linearization in Describing Spatial Networks, in S. Peters and E. Saarinen
(eds.), Processes, Beliefs, and Questions, 199-220 (Lancaster: Reidel).
L e v i n s o n , S. C. (1983), Pragmatics (Cambridge: Cambridge Univ. Press).
L e w i s , E. G. (1978), Types of Bilingual Communities, in J. E. Alatis (ed.), International
Dimensions o f Bilingual Education, 19-34 (Washington, DC: Georgetown Univ. Press).
L o p e B l a n c h , J. (1967), L a i n f l u e n c i a d e l s u s t r a t o e n l a f o n t i c a d e l e s p a o l d e M x i c o ,
Revista de Filologa Espaola, 50: 145-60.
L p e z M o r a l e s , H. (1980), Sobre la pretendida existencia y pervivencia del criollo en
Cuba, Anuario de Letras, 18: 85-116.
L u j n , M. (1981), The Spanish Copulas as Aspectual Indicators, Lingua , 54: 165-209.
M a a n d i , K. (1989), Estonian among Immigrants in Sweden, in Dorian (1989: 227-41).
M a h e r , J. (1991), A Crosslinguistic Study of Language Contact and Language Attrition,
in Seliger and Vago (1991: 67-84).
M a l k i e l , Y. (1983), Multiple versus Simple Causation in Linguistic Change, in From
Particular to General Linguistics, 251-68 (Amsterdam: Benjamins).
M a r t i n e t , A. (1962), A Functional View o f Language (Oxford: Oxford Univ. Press).
M e i l l e t , A. (1926), Linguistique historique et linguistique genrale (2 vols., Paris: Champion).
M e i s e l , J. M. (ed.) (1977), Langues en contact: pidgins, creoles. Languages in contact (Tbingen:
TBL Verlag Narr).
 References 245
------ (1983a), Strategies o f Second Language Acquisition: More than One Kind o f Simpli
fication , in Andersen (1983: 120-57).
------ (1983), Transfer as a Second Language Strategy, Language and Communication, 3:
11- 4 6 .
M H. A., and A n d e r s o n , P. L . (1988), Attitude toward Use o f Spanish on the South
e jia s ,

Texas Border, Hispania, 7 1: 4 0 1-7.

M en n , L . (1989), Some People Who Dont Talk Right: Universal and Particular in Child
Language, Aphasia, and Language Obsolescence, in Dorian (1989: 335-45).
M il r o y , J., and M il r o y , L . (1985), Linguistic Change, Social Network and Speaker Inno
vation, Journal o f Linguistics, 2 1: 33984.
M it h u n , M . (1989), The Incipient Obsolescence o f Polysynthesis: Cayuga in Ontario and
Oklahoma, in Dorian (1989: 243-57).
M o r a l e s , A. (1986), Gramticas en contacto: Anlisis sintcticos sobre el espaol de Puerto Rico
(Madrid: Playor).
M o r en o d e A l b a , J . (1978), Valores de las formas verbales en el espaol de Mxico (Mexico:
Universidad Nacional Autnoma de Mxico).
M o u g e o n , R ., and B e n ia k , E. (1991), Linguistic Consequences o f Language Contact and Re
striction: The Case o f French in Ontario, Canada (Oxford: Clarendon Press).
------ , ------ , and V a l o is , D. (1985), A Sociolinguistic Study o f Language Contact, Shift,
and Change, Linguistics, 23: 455-87.
M u h l h Au s l e r , P. (1981), Pidginization and Simplification o f Language, Pacific Linguistics, ser.
B , No. 26.
M u y s k e n , P. (1981), Creole Tense/M ood/Aspect Systems: The Unmarked Case?, in
P. Muysken (ed.), Generative Studies on Creole Languages, 18 1-9 9 (Dordrecht: Foris).
M y h il l , J. (1988), The Grammaticalization o f Auxiliaries: Spanish Clitic Climbing, in
S. Axmaker, A. Jaisser, and H. Singmaster (eds.), Proceedings o f the Fourteenth Annual
Meeting o f the Berkeley Linguistics Society, 352-6 3 (Berkeley, Calif.: Berkeley Linguistics
Society).
------ (1989), Variation in Spanish Clitic Climbing, in T . J. Walsh (ed.), Synchronic and
Diachronic Approaches to Linguistic Variation and Change, 22750 (Washington, D C :
Georgetown Univ. Press).
N a v a s R u iz , R . (1963), Ser y estar: Estudio sobre el sistema atributivo del espaol (Salamanca:
Acta Salmanticensia).
Nie, N . H. et al. (1975), Statistical Package fo r the Social Sciences, 2nd edn. (New York:
M cGraw-Hill).
O c a m po , F. (1989), The Pragmatics o f Word Order in Spoken Rioplatense Spanish (Ph.D.
diss., Univ. o f Southern California).
------ (1990), The Pragmatics o f Word Order in Constructions with a Verb and a Subject ,
Hispanic Linguistics, 4: 87-128.
O f in e s (1964), Presente y futuro de la lengua espaola (2 vols., Madrid: Ediciones Cultura
Hispnica).
O t h e g u y , R . (1993), A Reconsideration o f the Notion o f Loan Translation in the Analysis
o f U S Spanish , in A. Roca and J. Lipski, Spanish in the United States, 2 1- 4 5 (Berlin:
Mouton).
------ G a r c a , O., and F e r n n d e z , M . (1989), Transferring, Switching, and Modeling in
West New York Spanish: An Intergenerational Study, in Wherritt and Garcia (1989:
4 1-52).
P e a l o s a , F . (1980), Chicano Sociolinguistics (Rowley, Mass.: Newbury House).
246 References

P f a f f , C., and P o r t z , R. (1979), Foreign Childrens Acquisition o f German: Universals vs.

Interference, paper presented at the L S A Annual Meeting, L os Angeles.
P o p l a c k , S. (1978), Quantitative Analysis o f Constraints on Code-Switching (New York: Center
for Puerto Rican Studies).
------ (1979), ^Sometimes I 'll start a sentence in Spanish y termino en espaol (New York: Center
for Puerto Rican Studies); repr. in Amastae and Elias-Olivares (1982: 230-63).
------ (1987), Contrasting Patterns o f Code-Switching in Tw o Communities , in E . Wande
et al. (eds.), Aspects o f Multilingualism: Proceedings from the Fourth Nordic Symposium on
Bilingualism, iq 84, 5 1- 7 7 (Uppsala: Borgstroms).
------ S a n k o f f , D ., and M i l l e r , C. (1988), The Social Correlates and Linguistic Processes
o f Lexical Borrowing and Assimilation, Linguistics, 26: 47-104.
P o u N T A iN , C. (1982), *Essere/Stare as a Romance Phenomenon, in N . Vincent and
M . Harris (eds.), Studies in the Romance Verb, 13 9 -6 0 (London: Croom Helm).
P r e s t o n , D. R. (1982), How to Lose a Language, Interlanguage Studies Bulletin, 6: 64-87.
P r in c e , E . (19 9 2 ), O n Syntax in Discourse, in Language Contact Situations , in C. Kramsch
and S. M cConnell-Ginet (eds.), Text and Context: Cross-Disciplinary Perspectives on Lan
guage Study, 9 8 - 1 1 2 (Boston: Heath).
R a n s o n , D. L . (1991), Person Marking in the Wake o f / s / Deletion in Andalusian Span
ish , Language Variation and Change, 3: 13 3 -5 2 .
R e a l A c a d e m ia d e l a L e n g u a (1973), Esbozo de una nueva gramtica de la lengua espaola
(Madrid: Espasa Calpe).
R iz z i , L . (1978), A Restructuring Rule in Italian Syntax, in S. J . Keyser (ed.), Recent
Transformational Studies in European Languages, 11 3 - 5 8 (Cambridge, M ass.: M IT
Press).
R o d r ig o , V. (1991), Subject Expression and /%/ Elision in East Andalusia Spanish (M S,
Univ. o f Southern California).
R o e p e r , T ., and W il l i a m s , E. (eds.) (1987), Parameter Setting (Dordrecht: Reidel).
Rojo, G . (1974), L a temporalidad verbal en espaol, Verba, 1 (Universidad de Santiago de
Compostela).
R o m a in e , S. (1981), On the Problem o f Syntactic Variation: A Reply to Beatriz Lavandera and
William Labov (Austin, T ex.: Southwest Educational Development Laboratory).
------ (1982), Socio-historical Linguistics: Its Status and Methodology (Cambridge: Cambridge
Univ. Press).
------ (1988), Pidgin and Creole Languages (London and New York: Longman).
------ (1989a), Pidgins, Creoles, Immigrant, and Dying Languages , in Dorian (1989: 36 9-
83).
------ (1989^), Bilingualism (Oxford: Blackwell).
------ (1992), The Evolution o f Linguistic Complexity in Pidgin and Creole Languages, in
J. A. Hawkins and M . Gell-M ann (eds.), The Evolution o f Human Languages, 2 13 -3 8
(Reading, Mass.: Addison Wesley).
R o m o , R . (1983), East Los Angeles: History o f a Barrio (Austin: Univ. o f Texas Press).
R o u s s e a u , P., and S a n k o f f , D. (1978), Advances in Variable Rule Methodology, in D.
Sankoff (ed.), Linguistic Variation: Models and Methods, 57-69 (New York: Academic).
R u h l , C. (1989), On Monosemy: A Study in Linguistic Semantics (Albany: State Univ. o f
New York Press).
R y a n , E. B ., G i l e s , H., and S e b a s t ia n , R . J . (1982), An Integrative Perspective for the
Study o f Attitudes toward Language Variation, in E. B. Ryan and H. Giles (eds.),
Attitudes towards Language Variation, 1 - 1 9 (London: Edward Arnold).
 References 247

S nchez, R. (1983), Chicano Discourse (Rowley, Mass.: Newbury House).

S c h c h t e r , J., and R u t h e r f o r d , W. (197g), Discourse Function and Language Transfer,
1 - 1 2 (Toronto: Ontario Institute for Studies in Education).
S c h if f r in , D. (1981), Tense Variation in Narrative, Language, 57: 45-62.
S e l ig e r , H. W. (1991), Language Attrition, Reduced Redundancy, and Creativity, in
Seliger and Vago (1991: 227-40).
------- an d V a g o , R. M . (eds.) ( 19 9 1 ) , First Language Attrition (Cambridge: Cambridge Univ.
Press).
S il v a - C o r v a l n , C. (1977), A Discourse Study o f Some Aspects o f Word Order in the
Spanish Spoken by Mexican-Americans in West L os Angeles (M A thesis, Univ. o f
California at Los Angeles).
------ (1981), The Diffusion o f Object-Verb Agreement in Spanish, Papers in Romance, 3:
163-76.
------ (1982), Subject Expression and Placement in Mexican-American Spanish, in Amastae
and Elias-Olivares (1982: 93-120).
------ (1983), Tense and Aspect in Oral Spanish Narrative: Context and Meaning , Lan
guage, 59: 6o-8o.
------ (1983#), Code-Shifting Patterns in Chicano Spanish , in L . Elias-Olivares (ed.), Spanish
in the U S Setting: Beyond the Southwest, 69-87 (Rosslyn, Va.: National Center for Bi
lingual Education).
------ (1983^), On the Interaction o f Word Order and Intonation: Some OV Constructions
in Spanish, in F . Klein-Andreu (ed.), Discourse Perspectives on Syntax, 117 - 4 0 (New York:
Academic).
------ (1985), Modality and Semantic Change, in J . Fisiak (ed.), Historical Semantics:
Historical Word Formation, 547-72 (Berlin: Mouton).
------ (1986), Bilingualism and Language Change: The Extension o f Estar in Los Angeles
Spanish , Language, 62: 587-608.
------ (1989), Socio lingstica: Teora y anlisis (Madrid: Alhambra).
------ (1990a), Cross-Generational Bilingualism: Theoretical Implications o f Language
Attrition, in T . Huebner and C. A. Ferguson (eds.), Cross-Currents in Second Language
Acquisition and Linguistic Theories, 325-45 (Amsterdam: Benjamins).
------ (1990#), The Pragmastylistics o f Hypothetical Discourse, in L . Hickey (ed.), The
Pragmatics o f Style, 8 7-105 (London: Routledge).
------ (1990c), Current Issues in Studies o f Language Contact , Hispania, 73: 162-76.
------ (1991a), Spanish Language Attrition in a Contact Situation with English, in Seliger
and Vago (1991: 15 1- 7 1) .
------ (i99i>), Basic Meanings and Context-Bound Functions o f Tense in Spanish, in J.
Gvozdanovic et al. (eds.), The Function o f Tense in Texts, 255-70 (Amsterdam: North-
Holland).
------ and G u t i r r e z , M . (1995), On Transfer and Simplification: Verbal Clitics in Mexican-
American Spanish, in P. Hashemipour et al. (eds.), Studies in Language Learning and
Spanish Linguistics in Honor o f Tracy D. Terrell, 30212 (San Francisco: McGraw-Hill).
S l o b in , D. I. (1977), Language Change in Childhood and in History, in J. Macnamara
(ed.), Language Learning and Thought, 18 5 -2 14 (New York: Academic).
------ (ed.) (1986), The Cross-Linguistic Study o f Language Acquisition (Hillsdale, N J: Erlbaum).
S o l , Y ., and S o l , C. (19 7 7 ) , Modern Spanish Syntax (Lexington, Mass.: Heath).
S u r e z , J . A. (1966), Indigenismos e hispanismos vistos desde la Argentina, Romance
Philology, 20: 68-90.
248 References

S u b ir a t s - R g g e b e r g , C. (1987), Sentential Complementation in Spanish (Amsterdam:

Benjamins).
T a u l i, V. (1956), The Origin o f Affixes, Finnisch-Ugrische Forschungen, 32: 170 -225.
T a y l o r , A. R. (1989), Problems in Obsolescence Research: The Gros Ventres o f Montana,
in Dorian (1989: 167-79).
T h o m a so n , S. G . (1986), On Establishing External Causes o f Language Change, in S.
Choi et al. (eds.), Proceedings o f the Second Eastern States Conference on Linguistics, 243
51 (Columbus: Dept, o f Linguistics, Ohio State ifniv.).
------ and K a u f m a n , T . (1988), Language Contact, Creolization, and Genetic Linguistics
(Berkeley: Univ. o f California Press).
T o r r e s , L . (1989), Mood Selection among New York Puerto Ricans, in Wherritt and
Garcia (1989: 67-77).
T r u d g il l , P. (1976-7), Creolization in Reverse: Reduction and Simplification in the Alba
nian Dialects o f Greece, Transactions o f the Philological Society (1976-7): 32-50.
------ (1983), On Dialect (Oxford: Blackwell).
T s it s ip is , L . (1981), Arvanitika Language Change in Speech Communities in Greece, in
H. I. Aronson and B. J. Dafden (eds.), Folia Slavica: Studies in Balkan Linguistics to
Honor E. P. Hamp on his Sixtieth Birthday, 378 -8 3 (Columbus, OH: Slavica).
------ (1988), Language Shift and Narrative Performance: On the Structure and Function
o f Arvanitika Narratives, Language in Society, 17: 6 1-8 6 .
V a n D i j k , T . A. (1980), Texto y contexto (first published in English); Spanish translation by
j . Domingo Moyano (Madrid: Ctedra).
V a - C e r d , A. (1982), Ser y estar + adjetivos: Un estudio sincrnico y diacrnico (Tbingen:
Narr).
V in c e n t , N ., an d H a r r is , M . (19 8 2 ), Studies in the,Romance Verb (L o n d o n : C ro o m H elm ).
W a l d , B. (1583), Referents and Topic Within and Across Discourse Units: Observations
from Current Vernacular English, in F. Klein 4Andreu (ed.), Discourse Perspectives on
Syntax, 9 1 - 1 1 6 (New York: Academic). .*
W a n g , W . S .- Y . (19 6 9 ), Competing Changes as a:Cause o f Residue , Language, 4 5 : 9 - 2 5 .
W e in r e ic h , U. (1974), Languages in Contact, 8th printing (The Hague: Mouton).
------- L a bo v , W ., and H e r z o g , M . (1968), Empirical Foundations for a Theory o f Language
Change, in Lehmann and Malkiel (1968: 95-195).
W h e r r it t , I., and G a r c a , O. (eds.) (1989), U S Spanish: The Language o f Latinos (Inter
national Journal o f the Sociology o f Language, special issue No. 79).
W o l f s o n , N. (1982), The Conversational Historical Present in American English Narrative
(Dordrecht: Foris).
Z e n t e l l a , A. C . (1981), Hablamos los dos, we speak both : Growing U p Bilingual in El
Barrio (Ph.D. dissertation, Univ. o f Pennsylvania).

Sources o f written data examined

E s g u e v a , M ., and C a n t a r e r o , M . (eds.) (1981), E l habla de la ciudad de Madrid: Materiales

para su estudio (Madrid: Consejo Superior de Investigaciones Gentlicas, Instituto Miguel
de Cervantes).
I n s t it u t o d e F il o l o g a A n d r s B e l l o (19 7 9 ), E l habla culta de Caracas: Materiales para
su estudio (Caracas: Universidad Central de Venezuela).
 References 249

R A., and C o n t r e r a s , L. (eds.) (1979)1E l habla culta de Santiago de Chile: Materiales

a ba n a les,

para su estudio (Boletn de Filologa 2: Universidad de Chile).

U n a m (19 7 1), E l habla de la ciudad de Mxico: Materiales para su estudio (Mexico, D F :

Universidad Nacional Autnoma de Mxico).

------ (1976), E l habla popular de la ciudad de Mxico: Materiales para su estudio (Mexico, D F :
Universidad Nacional Autnoma de Mxico).

EL COLEGIO DE MEXICO

* 3 905 0642903 %*
 Index

actuation i bilingual communities 9, 18, 49, 199, 205
see also change, motivation o f diversity o f 199
Adams, M . 147 bilingualism:
Aissen, J . 127 cyclic 1 1
Alonso, A. 162, 2 15 typical family situation o f 1 0 - 1 1
Amastae, J . 14 Bills, G . 10, 14, 15, 89
analysis 2, 5 2 - 3 Bloomfield, L . 118
definition o f 5 borrowing 4, 172, 173, 2 17
analytical constructions 52, 2 13 cultural 183
Andersen, R. W. 5 1, 52, 2 13 , 2 17 nonce- 134, 165, 216
Anderson, P. L . 188, 196, 197, 198 nonce syntactic 166
Anttila, R. 108 see also caiques; loan; modelling; transfer
assertiveness 23, 76, 83, 90, 91 Bowen, J . D . i j
scale o f 76, 77, 78 Brown, R. 23, 25, 50, 5 1, 154, 156
attitudes, language 18 6 -206 Bybee, J . 128
attitude questionnaire 188
dimensions o f 197, 198 caiques;
negative 7, 14, 193, 202, 203, 204, 205, 2 12 o f collocations and idioms 1 7 2 - 4
neutral 7, 14, 197, 2 12 definition o f 170
positive 7, 168, 186, 2 0 2 -6 , 2 12 ; see also frequency o f 18 5 - 7
uncommitted behaviour lexico-syn tactic 17 2 , 17 4 -8 4
subjective 14,. 108, 188, 197, 219 multiple-word 170
towards English 15 ,2 0 2 , 2 12 single-word 17 1 , 17 3 , 207
towards Spanish 168, 169, 202, 205, 2 12 o f subcategorization 17 9 -8 2
see also commitment; language loyalty see also borrowing; loan; modelling; transfer
attrition, language 3, 7, 10, 20, 26, 5 1, 56, 77, Cameron, R. 154, 162
202, 205 Campbell, L . 164, 2 15 , 218
and complexification 220 Cantero Sandoval, G . 162
and creolization in reverse 13 , 2 14 Catalan, D . 2 15
effect on hypothetical discourse 9 1; see also change, linguistic 1 - 2
hypothetical oral discourse acceleration o f 22, n o , 114 , 1 1 5 , 119 , 129,
effect on structure and semantics of 130, 131215
narratives 7 5 - 6 , 9 1; see also narrative, oral definition o f 5
predictors o f 2 12 diffusion o f 1 n.
principles characterizing 2 18 - 2 0 ; see also extralinguistic factors in 1, 2, 5, 7, 10, 13,
principles o f change 20, 106, 1 2 1, 164, 209, 2 12
see also loss; reduction; simplification internal processes o f 54
intralinguistic factors in 1 , 2, 50, 92, 162,
Barrenechea, A. 162 208, 214, 2 15
Bavin, E . L . 1 3 1 motivation o f 1 ; see also actuation
Behrend, E. 97 n. multiple causation o f 1 1 5
Beniak, E. 3, 49, 132 , 134, 167, 207, 208, 209, processes o f 5
2 12 , 2 13 , 2 15 , 2 18 , 219 triggering factors 2 15
Bentivoglio, P. 148, 150, 15 3 , 157 and variation 1 , 5
Benveniste, E . 118 see hypothesis, general; principles o f
Bergen, J . 14 change
Bickerton, D . 5 1, 134 Cifuentes, H. 150, 153
 Index 251
code-switching 6, 14, 63, 188 specific influence from English on 119
see also conversational narratives stages o f change o f 10 0 - 1
cognitive load 6, 207 and type o f adjective io i, 106--110, 1 1 5 , 120
Coleman, L . 99 and variation by age 1 1 7
commitment used innovatively in L os Angeles 105
scale 203, 205 see also ser and estar
degree o f 205, 209, 2 12 este/esta:
see also attitudes extension o f discourse-pragmatic functions
communicative style 76, 86 o f 135
como: and 'nevi-this 1 3 5 - 6
meaning extension o f 1 7 6 - 7 and permeability 13 5
complementizer, see que estimative verb 137
Comrie, B. 20, 2 1, 22, 3 1 , 32, 33, 158 definition o f 15 1
convergence 2, 129, 130, 13 2 , 166, 2 13 , 2 15 expanding form, definition o f 3
definition o f 4-5
grammatical 166; see also permeability o f Falk, J . 94, 95 n.
grammar Ferguson, C . A. 2, 3, 54, 208, 2 1 3
in Kupw ar 166 Fernndez, S. 98, 12 0 n., 170, 184
conversational narratives 57, 59, 81 Fishman, J . A. 15, 188, 190, 203, 205, 206
clause complexity in 63 Fleischman, S . 52, 220
code-switching in 64, 66 Fontanella de Weinberg, M . B. 2 15
effect o f language attrition on 7 5-6 Frake, C. O, 74
establishment o f referents 63, 70 Franco, J . 158, 182
evaluative techniques 63, 69
mode o f delivery 63 G al, S . 49, 120, 190
tense forms in complicating action 72 Garca, E. 15, 19, 1 2 1 n., 139, 140, 170, 184
see also narrative, oral G ili y Gaya, S. 162
Copeag, D . 94, 94 n. Givon, T . 12 3 n.
Cruse, D . A. 172, 178 Gleason, J . B. 2 17
Granda, G . de 2 15
diffusion: Gumperz, J . 5, 146, 166, 167
lexical 108, 120 n. gustar:
lexico-semantic 98; see also estar, gradual and changes in subcategorization 1 7 9 - 8 1
diffusion of and subject expression 1 5 1 , 176, 180
Dorian, N . 1 1 , 13 , 48, 49, 50, 53, 54, 120, 132, Gutirrez, M . 70, 105, 11 4 , 1x5
142, 147, 188, 190, 193, 196, 197, 199,
202, 204, 205, 2 15 , 2 18 , 2 19 Haiman, J . 66, 70
Harris, M . 52
Edwards, J . 188 Henrquez Urea, P. 2 15
EMas-Olivares, L . 1 1 , 14, 15 H ill, J . H . 120, 142
Enriquez, E . V . 148, 1 5 1 , 15 3 , 162 Hochberg, J . G . 153, 153 n., 154, 157 , 162
Escudero, G . 94, 94 n Hudson-Edwards, A. 16 1
estar: Hyams, N . 147 n.
and acceleration o f change 1 1 4 - 1 7 hypothesis, general 6, 39, 47, 207
and age o f acquisition o f English 1 1 3 - 1 4 about change acceleration in the
direct influence from English on 53, 1 3 1 , superordinate language 1 1 9
132, 169, 178, 208 about internal motivation o f change in
effect o f diatopical differences on 1 1 7 - 1 8 language contact 92
effect o f formal learning o f Spanish on 129 about introduction and diffusion o f
effect o f language contact on 114 , 119 innovations 6
gradual diffusion o f n o , 120; see also see also estar, direct influence from English on
diffusion hypothetical oral discourse 82, 83, 86, 87, 9 1,
innovative use o f 105, 106, 1 1 3 186
and lexical diffusion 106, 108, 120 anchoring, definition o f 80
in M orelia (Michoacn, Mexico) 105, 114 , argumentation, definition o f 80
i 15-17 coda, definition o f 81
and Spanish proficiency 1 1 3 - 1 4 definition o f 76
252 Index
hypothetical oral discourse (com.) Lenz, R. 2 15
disclaimers, definition o f 8 0 -1 Levelt, W. 80, 90
elements contained in 7 9 - 8 1 Levinson, S. C . 79 n.
in Groups 2 and 3 8 6 -9 0 Lewis, E. G . 9
hypothetical statements, definition o f 97 linearization 80, 90
macro-frame, definition o f 79 loan:
qualifications, definition o f 80 definition o f 170
strictly 78, 79 single-word 207
structure o f 7 7 - 8 1 see also borrowing; caiques; modelling; transfer
topics to elicit 76 loan translation 16 9 -7 0 , 1 7 1
with supportive argumentation 78 loanwords 185
in New York Cuban Spanish 185
immigrants: Lope Blanch, J . 2 15
differences between first-generation and other Lopez Morales, H. 2 15
groups 26 Los Angeles, demographic information 9 - 10
included in this study 1 5 - 1 6 loss, linguistic:
interference 13 3 , 168, 169, 208, 2 14 , 2 15 , 2 17 at the individual level 50
arisal o f 164 o f semantic-pragmatic constraints 13 5 , 144,
inter-linguistic 2 18 164, 209, 216
lexical 169 see also attrition; language; reduction
see also transfer Lujan, M . 94

Jackson, H . 172 Maandi, K . 2 14 , 2 18

Jakobson, R. 5, 13 3 , 134 Maher, J . 207
Jonge, B. de 105, 1 1 9 Martinet, A. 1, 133
meaning complexes 56, 76, 91
Kaufman, T . 5, 6, 13 3 , 134, 166, 167, 2 14 , 2 15, see also hypothetical oral discourse; narrative
2 16 Meillet, A. 118 , 13 3, 134
Keenan, E. 158 M eisel, J . M . 3 and n., 12 3 and n., 129, 2 13
K ing, R. 130 Mejias, H. A. 188, 196, 197, 198
Klein, W. 56, 76 M enn, L . 2 13 , 218
Klein-Andreu, F . 4, 77, 12 3 , 1 3 1 , 132, 14 1, Mexican culture 204, 205
1 5 1 n. M ilroy, J . 110 , H 2
Koike, D . A. 64 Mithun, M . 143
modal value o f verb morphology 77
Labov, W. 1, 2, 5, 57, 60, 120 modelling 170
Lambreeht, K . 148, 165 as a cover term 184
Landa, A. 12 7, 142, 147, 158 in Los Angeles Spanish 1 8 5 -7 , 2 1 1
language acquisition 2, 20, 32, 54, 147, 208, multiple-word 170, 172
2 13 , 2 17 in N ew York Cuban Spanish 18 4 -5
and language loss 50 see also borrowing; caiques; loan; transfer
levels of, 218 Morales, A. 4, 150, 15 3 n., 16 3, 2 15
see also estar, and age o f acquisition o f Moreno de Alba, J . 55
English; Spanish as a second language Mougeon, R. 3, 4, 49, 132 , 134, 167, 207, 208,
language contact, intensive: 209, 2 12 , 2 13 , 2 15 , 2 18 , 219
factors associated with 134 Muhlhausler, P. 3, 2 13
language loyalty 205 Muysken, P. 49
acts o f 168 M yhill, J . 12 7, 130
definition o f 168
see also shifting, acts of; attitudes; narrative, oral:
uncommitted behaviour abstract 60, 74, 94, 99, 197, 199, 205
language shift 6, 10, 1 1 , 13 , 120, 221 adjunct information 60, 6 1, 69, 72, 75
definition o f 168 auxiliary information 60, 6 1, 66, 69, 70, 75,
to English 9 - 1 0 , 168, 197 76, 91
Lapesa, R. 2 15 coda 60, 74, 78
Lavandera, B. 19 complicating action 60, 66, 67, 72, 75
Lenneberg, E . H . 15 defined 57
 Index 253
elaboration 60, 61 and societal bilingualism 1 1 , 221
elements o f 57 and tense systems 3 1 , 32; see also tense-
evaluation 60, 6 1, 63, 66, 67, 75, 76, 103, mood-aspect, simplification and loss of
107, n o , 1 1 2 , 1 1 5 , 185 in the use o f Gaelic tense 49
necessary information 60
orientation 60, 6 1, 66, 69, 74, 75, 78 que:
performed 63 in relative clauses 13 7
preface 59, 75, 133 sex correlation with use o f 138
pre-narrative 59, 60 variation between expressed and zero 1 3 7 - 8
resolution 6o, 6 1, 75
see also conversational narratives reduction 13 , 50, 54, 9 1, 93, 167, 2 15 , 2 18 , 220
Navas Ruiz, R. 94 in amount o f language 87
N ie, N . H . 105 o f social domains o f use o f a language 135
normative pressures, absence o f 7, 2 12 in use o f Spanish; see Spanish, use o f
and variation 3
Ocampo, F . 14 1, 144, 165 see also attrition, language; loss; ser, reduction
Omstein, J . 15 in uses of; simplification
Otheguy, R. 170, 1 7 1 , 17 2 and n., 175, 184 regularization 119
overgeneralization 2, 5, 6, 1 2 1, 122, 164, 207, definition o f 3
2 12 , 2 13 , 2 16 , 2 17 relexification 18 3 -4
definition o f 3 see also caiques, lexico-syntactic
o f obligatory reflexive verbs 13 1 Rizzi, L . 127
Rodrigo, V . 1 5 1
para atrs 1 7 1 , 175, 186 Roeper, T . 147
permeability (of grammar) 4, 7, 13 3 Rojo, G . 39
definition o f 135 Romaine, S. 1, 13 , 19, 147 and n., 188, 2 13 , 2 14
and internal tendencies o f development 133, Romo, R . 9
,
135 138
o f parallel structures 13 5 , 136, 14 1, 142, 167,
Rousseau, P. 19
Ruhl, C. 172
2 17 Rutherford, W. 4
and structural weaknesses 13 3 , 135, 2 17 Ryan, F . B . 188
syntactic 16 4 -7
see also interference; transfer; loss, language; Sanchez, R. 15
o f semantic-pragmatic constraints Schchter, J . 4
Pfaff, C. 12 3 Schiffrin, D . 66
Poplack, S. 6 n., 12 3 , 134, 170 Seliger, H . W. 2 15 , 2 17 , 2 18
Portz, R. 12 3 ser:
Pountain, C. 94 n. reduction in uses o f 105
prepositions 14, 178, 2 17 ser and estar:
in Canadian French 132 and class v. individual frame 9 5 -6 , 98
in collocations 178 n. meaning o f 100, h i
effect o f English on 17 8 -9 in opposition 9 3-5, 120
subcategorized by verbs 17 9 -8 2 and prototypical meaning 99
Preston, D . R. 3, 2 17 questionnaire eliciting 10 3 -4
Prince, E. 13 5 , 166, 2 18 in Romance languages 92 n.
principles o f change: semantic and pragmatic contrast o f 94
distance 219 semantic transparency o f the choice
generality 219 between i n - 1 2
semantic transparency 220 throughout the history o f Spanish 94
proficiency continuum 5, 6, 7, 15, 169, 170 shifting, acts o f 168
as apparent time 5, 30 as a maintenance strategy 169
attitudes across, see attitudes, language see also language shift
and data quantification 19 Silva-Corvaln, C. 1, 3, 4, 5, 6 n ., 1 1 , 19,
definition o f 1 1 23, 24, 26, 54, 55, 57, 59, 6o> f>4, 66, 70,
and emblematic use o f Spanish 1 1 , 13 72, 77, 97 n., 12 3 n., 14 1 , 142, 146, 147,
factors accounting for lowest levels o f 14 148, 150, 152 , 154, 157 , 158, 162, 182,
and linguistic load 39 219
254 Index
simplification 5, 6, 2 13 , 2 14 development o f verbal periphrases 4 6 -7 , 52
definition o f 2 -3 non spontaneous use o f 3 3 -9
identification o f 22 patterns o f extension o f 39-45
see also attrition, language; loss; reduction; patterns o f tense extensions 39
tense-mood-aspect, simplification and loss semantic extensions o f 43
of in spoken Spanish 2 1 - 3
simplified system 2 13 see also verbal forms
definition o f 3 tense-mood-aspect, analysis of:
Slobin, D . I. 50, 2 13 , 2 17 , 2 18 , 220 discourse-pragmatic contexts for 2 3-5
societal bilingualism, see proficiency continuum tense-mood-aspect, simplification and loss of:
Sol, Y . 138 in different linguistic contexts 23
Spanish, standard general 98 effect o f English on 53, 54, 138
Spanish, use of: factors accounting for 54, 208
and correlation with Spanish maintenance and markedness 49, 50, 147
and attrition 205 and parallels with other languages 48-50
reduction in 1 9 1 - 5 , 2 12 and stative vs. dynamic lexical aspect 5 1
in Rio Grande Valley, T exas, 197 Thomason, S. G . 5, 6, 13 3 , 134, 166, 16 7, 207,
Spanish as a second language: 214, 2 15 , 216
acquisition o f Pret and Imp in 5 1 - 2 Thompson, S. 66, 70
see also estar, effect o f formal learning o f tiempo:
Spanish on; verbal clitics, effect o f formal meaning extension o f 17 4 -5
learning o f Spanish on Torres, L . 1 1
Spanish in L os Angeles: transfer 2, 92, 1 1 5 , 1 2 1 , 13 2 , 133, 134, 145, 2 14
U S Census information 7 arisal o f 16 4-5, 2 17
Spanish maintenance: definition o f 4
reasons for 10, 197, 199 direct 4, 5, 6, 133, 16 1, 168, 169, 207, 209,
structural weaknesses 13 4 3 , 5, M7
Stutterheim, C . von 56, 76 o f discourse-pragmatic function 136, 165,
style o f communication 76, 86, 87 166, 2 17
reduced ranges o f or stylistic shrinkage 49, from English 122, 12 3, 126, 1 3 1 , 136, 15 3 ,
88, 91 159, 208
Surez, J . A. 2 15 indirect 4, 6, 169, 207, 2 15 , 2 17
Subirats-Riiggeberg, C. 13 7 o f parallel structures 12 3, 1 3 1 2; see also
subject expression: permeability
competing motivations favouring 147 as a result o f formal learning 129
and cross-linguistic equivalence 146 syntactic 122, 126, 134, 138, 165, 167, 173;
discourse-pragmatic factors favouring see also permeability
14 8 -5 0 , 163 see also borrowing; caiques; interference; loan;
and dummy subjects 146 modelling
frequency o f 153 Trudgill, P. 3, 13 , 49, 2 13
and given or new status 14 1 - 2 , 158 Tsitsipis, L . 49, 57 n.
internal variables constraining 15 0 -2
and loss o f coreferentiality constraint 1 6 1 - 2 uncommitted behaviour 2 0 3 -6
and pro-drop language 147
variable 146, 148, 157 Vago, R. M . 2 15
see also estimative verb; gustar, word order, Van Dijk, T . A. 79
subject placement Van-Cerd, A. 94, 120 n.
syntactic reception, acts o f 216 verbal clitic pronouns 92, 134, 135, 169
in American Finnish 2 16 and acceleration o f change 12 9 -3 0
see also caiques, lexico-syntactic; permeability, cross-linguistic equivalence o f 12 1
syntactic; transfer, syntactic direct English influence on 13 2 , 140
effect o f formal learning o f Spanish on 129
Tauli, V. 118 as markers o f possession 1 3 8 - 4 1, 209
Taylor, A. R. 2 1 omission o f 122, 1 2 3 - 6
tense: and parallel structures, see transfer
and cognitive complexity 54 in periphrases 12 7 - 9 , 2 14
definition o f 2 1 position o f 126
 Index 255
verbal forms: w ild-card verb 52
ambiguous 146, 159 auxiliary-type 47, 53
stative and dynamic lexical aspect 5 1 Williams, E . 147
see also wild-card verb Wolfson, N . 64
Vincent, N , 52 word order:
caiques 18 2 -3
Wald, B. 136 pragmatically neutral 182
Waletsky, J . 57 in Rio Platense Spanish 145
Wang, W. S .-Y . 108, 120 and subject placement 1 4 1 - 4
Weinreich, U . 1, 2, 4, 12 3 , 133, 134, 135, 138, see also verbal clitic pronouns
164, 167, 170, 173, 2 16 , 2 17
Wherritt, I. 15 Zentella, A. C. 190