Spanish in the United States: Linguistic Contact and Diversity 9783110804973, 9783110165722

Citation preview

Spanish in the United States

W G DE

Studies in Anthropological Linguistics 6

Editors

Florian Coulmas Jacob L. Mey

Mouton de Gruyter Berlin · New York

Spanish in the United States Linguistic Contact and Diversity

Edited by

Ana Roca and John M. Lipski

Mouton de Gruyter Berlin · New York

1993

Mouton de Gruyter (formerly Mouton, The Hague) is a Division of Walter de Gruyter & Co., Berlin

© Printed on acid-free paper which falls within the guidelines of the ANSI to ensure permanence and durability.

Library of Congress Cataloging-in-Publication

Data

Spanish in the United States / edited by Ana Roca and John M. Lipski. p. cm. - (Studies in anthropological linguistics ; 6) Includes bibliographical references and index. Contents: Linguistic heterogeneity, civil strife, and per capita gross national product in inter-polity perspective / Joshua A. Fishman - A reconsideration of the notion of loan translation in the analysis of U.S Spanish / Ricardo Otheguy — The dialectics of Spanish language loyalty and maintenance on the US-Mexico border / Margarita Hidalgo - Spanish clitics in a contact situation / Manuel J. Gutierrez and Carmen Silva-Corvalan - Language choice in Hispanic-background junior high school students in Miami / Barbara Zurer Pearson and Arlene McGee - Literacy stories / Rene Cisneros and Betty Leone - Language maintenance institutions of the Isleno dialect of Spanish / Felice Anne Coles — Convergent conceptualizations as predictors of degree of contact in US Spanish / Ricardo Otheguy and Ofelia Garcia - Creoloid phenomena in the Spanish of transitional bilinguals / John M. Lipski - Diversification and Pan-Latinity / Frances R. Aparicio — Oral proficiency testing and the bilingual speaker / David Barnwell. ISBN 3-11-013204-4 (cloth : alk. paper) ISBN 3-11-016572-4 (pbk.) 1. Spanish language - 2. Social aspects - United States. 2. Sociolinguistics. 3. Languages in contact - United States. 4. Bilingualism - United States. I. Roca, Ana. II. Lipski, John Μ. III. Series. PC4826.A42 1993 306.4'4'0973-dc20 93-14956 CIP

Deutsche Bibliothek - Cataloging-in-Publication

Data

Spanish in the United States : linguistic contact and diversity / ed. by Ana Roca and John M. Lipski. - Berlin ; New York : Mouton de Gruyter, 1999 (Studies in anthropological linguistics; 6) ISBN 3-11-016572-4 NE: Roca, Ana [Hrsg.]; G T

© Copyright 1993 by Walter de Gruyter & Co., D-10785 Berlin. All rights reserved, including those of translation into foreign languages. No part of this book may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording or any information storage and retrieval system, without permission in writing from the publisher. Typesetting and Printing: Arthur Collignon GmbH, Berlin — Binding: Lüderitz & Bauer GmbH, Berlin. — Printed in Germany.

Acknowledgements

We wish to take this opportunity to thank the Florida Endowment for the Humanities and Florida International University for their generous support of the Ninth National Conference on Spanish in the United States, held in Miami in 1988. The support of the university and of the Florida Endowment for the Humanities enabled Professor Ana Roca to organize a national forum titled "English as the Official Language: Advance or Retreat?" as part of the conference. We would like to give our sincerest thanks to the collection's contributors, an excellent group of scholars with whom it has been a pleasure to interact. We thank both U. S. Hispanic bilingual students as well as non-native speakers of Spanish who make the effort and dedicate the time it takes to become more proficient and literate in the second most important language in their country.

Contents

Acknowledgements

V

Introduction

1

Joshua A. Fishman Linguistic heterogeneity, civil strife and per capita gross national product in inter-polity perspective

9

Ricardo Otheguy A reconsideration of the notion of loan translation in the analysis of U.S. Spanish

21

Margarita Hidalgo The dialectics of Spanish language loyalty and maintenance on the U.S.-Mexico border: A two-generation study

47

Manuel J. Gutierrez and Carmen Silva-Corvalän Spanish clitics in a contact situation

75

Barbara Zur er Pearson and Arlene McGee Language choice in Hispanic-background junior high school students in Miami: A 1988 update

91

Rene Cisneros and Elizabeth A. Leone Literacy stories: Features of unplanned oral discourse

103

Felice Anne Coles Language maintenance institutions of the Isleno dialect of Spanish 121 Ricardo Otheguy and Ofelia Garcia Convergent conceptualizations as predictors of degree of contact in U.S. Spanish 135 John M. Lipski Creoloid phenomena in the Spanish of transitional bilinguals . . 155

viii

Contents

Frances R. Apacicio Diversification and Pan-Latinity: Projections for the teaching of Spanish to bilinguals 183 David Barnwell Oral proficiency testing and the bilingual speaker

199

Index

211

Introduction

The idea for this collection stems from our research interest in the linguistic aspects of Spanish as used in the United States today. Tied to our academic interest has been our professional participation in El Espanol en los Estados Unidos, a national conference in linguistics begun by Professor Lucia Elias-Olivares in 1979 and first held at the University of Illinois-Chicago Circle. El Espanol en los Estados Unidos has become an annual event, hosted by universities around the country, including the University of Texas (Austin), the University of New Mexico (Albuquerque), the University of Iowa (Iowa City), Indiana University (Bloomington), Hunter College of the City University of New York, Florida International University (Miami), the University of Arizona (Tucson), and the University of Southern California (Los Angeles). El Espanol en los Estados Unidos has dealt with linguistic aspects of Spanish in the United States, as well as with selected cross-disciplinary issues. This meeting provides scholars with a professional forum for the exchange of ideas, as well as the presentation of completed studies and research in progress. Hard data, valuable insights, incisive discussions, and the dissemination of research findings (Amastae — Elias-Olivares 1982; Elias-Olivares 1983; Elias-Olivares et al 1985; Wherritt - Garcia 1988; Bergen 1990; Roca-Lipski 1992), are among the many achievements of this conference. The present volume makes an additional contribution to the growing research bibliography on linguistic aspects of Spanish in the United States. 1 Some of the contributors are established scholars in the field, while others are entering the field with promising studies. As the Hispanic population continues to expand, so has the body and scope of linguistic research on the Spanish spoken in the United States. The earliest studies concentrated on rural Mexican American Spanish in the Southwest. The parameters of the field have been extended in response to changing urban and demographic realities, and to paradigm shifts in linguistics itself. Attention is no longer confined to the traditional domains of phonology, morphology, syntax, and semantics. Today, a growing number of linguists are concerned with the intricate and complex issues relating language to society, education, ethnicity, the media, and politics.

2

Introduction

What has become clear in the last twenty years ist that the study of Spanish in the United States cannot and should not be divorced from a cross-disciplinary understanding of its context in time, place, and history. The Hispanic population continues to be the largest and fastest growing linguistic minority in the United States. We can no longer approach linguistic research along narrow lines, thinking only, for example, in terms of the Spanish spoken by Mexican Americans in the five Southwestern states where the majority reside (California, Texas, Arizona, Colorado, and New Mexico). Likewise, we cannot continue to limit the study of Puerto Rican Spanish to New York, nor of Cuban Spanish to Florida. With ever-increasing numbers of Hispanics arriving in the U. S. from many different nations, the demographic profile today is far more complicated than twenty, or even ten years ago. Just like the Latino population itself, the Spanish spoken in this country is not homogeneous. Spanish speakers in the United States come from different generations, nationalities, and cultures. The population also reflects varied levels of schooling, experience, economic power, and status in the work force. Linguistically, this population includes a wide range of receptive and productive skills in English and Spanish. Spanish was first used in what is now the United States as early as the 1530s, preceding the English-speaking settlers of New England. In much of the Southwest, Spanish has been used continuously since then, before and after this region was annexed to the United States. The first large-scale influx of Spanish speakers from outside the modern United States occurred at the beginning of the 20th century, as a result of the turmoil caused by the Mexican Revolution. Many found work in mines, agriculture, or in the railroad industry. However, when the depression of the 1930s hit the country and unemployment was at its peak, many Mexican Americans, including some who were U. S. citizens, were deported to Mexico. Puerto Rico became a part of the United States at the close of the Spanish-American War in 1898, and Puerto Ricans became U. S. citizens right before the U. S. entered World War I. Partly as a result of poor economic conditions on the island — increasingly under more U. S. control — the emmigration flow reached its peak after World War II. From 1945 to 1955 over 50,000 Puerto Ricans would leave the island annually to come to the United States. Although the majority of them reside in New York, there are Puerto Ricans in many other parts of the Northeast, as well as in other regions of the country. 2

Introduction

3

Although there were Cubans living in places like Key West, Tampa, and New York since the latter part of the 1800s, it was not until after Fidel Castro took over the government in 1959 that large waves of exiles began to arrive in South Florida in the early sixties. With over a million Cubans in the United States, this population forms the largest Hispanic group in the U. S. after Mexican Americans and Puerto Ricans. In comparison to these last two, though, much less has been published on the language situation of Cuban Americans, most likely because it is a much more recent phenomemon. Today, serious problems caused by overpopulation and poverty in Mexico, as well as the devastation and economic instabilities resulting from recent civil wars in Central America (as in Nicaragua and El Salvador, for example), have become social forces which have helped to increase the number of Hispanic immigrants entering the United States. The availability of legal and illegal seasonal agricultural work and other types of work, has served as economic bait for people who left their homeland for poltical or economic reasons. Additionally, the dream of a better life in "el Norte", where there already exists a growing population of Hispanics with whom immigrants could identify and communicate with, has offered psychological incentives for coming here. These and other factors have played a role in increasing legal and illegal immigration of Mexicans and other Latin Americans. At the same time, research is also expanding to include the study of linguistic aspects of the Spanish of the multiple subgroups which make up the more varied and complex Hispanic population, which now includes an increasing number of Spanish speakers from the Caribbean and Central America. Most recently, for example, the U. S. receives regular arrivals of "balseros" from Cuba, who are so-named because they sail to South Florida in small makeshift rafts, or balsas. Nicaraguans, Salvadorans, Guatemalans, Dominicans, and Colombians are among other groups which increasingly enter the U. S. with and without visas. Although Mexican Americans make up over 63 percent of the Hispanic population in the United States the most evident increase in the Hispanic population growth rate between 1982 and 1987 was due to the Central and South American immigrant groups coming into the country (Bureau of the Census 1987). Our collection opens with an insightful essay by Joshua A. Fishman called "Linguistic Heterogeneity, Civil Strife and Per Capita Gross National Product in Inter-Polity Perspective". In his essay (also the keynote address of the conference), Professor Fishman summarizes previous stud-

4

Introduction

ies and discredits some of the claims of major negative consequences that are commonly said to follow from linguistic heterogeneity and which we commonly find in the popular media and in political discussions: the claims that linguistic heterogeneity leads to or exaccerbates civil strife and lowers national productivity. Professor Fishman's essay examines methodological considerations and available data in the literature as he compares popular thinking with sociolinguistic research in this connection. Ricardo Otheguy's paper examines the notion of loan translation in the analysis of U. S. Spanish. He argues that loan translation is a defective construct which hinders rather than illuminates discussions on the process of language contact. Margarita Hidalgo's paper is an excellent sociolinguistic, two-generation study on the question of linguistic assimilation vs. linguistic loyalty and maintenance in Chula Vista, a border city in Southern California, which has now become the second fastest growing city of San Diego County. In "Spanish Clitics in a Contact Situation", Manuel J. Gutierrez and Carmen Silva-Corvalän examine clitics in Spanish in order to take a look at transfer phenomena in the bilingual context of the city of Los Angeles, California. Barbara Pearson and Arlene McGee's paper reports on language choice in Hispanic background junior high school students in Miami. In their report, Pearson and McGee offer a review of the literature and raise issues regarding the differences in the degree of proficiency and use of Spanish by generation and by domain. In "The Isleno Dialect of Spanish: Institutions for Language Maintenance", Felice Ann Coles examines an isolated and dying language spoken by only a few residents of St. Bernard Parish, Louisiana. Coles reports on the institutional language maintenance efforts which the Islenos have coordinated as a way of "revatilizing" their sense of ethnicity. In their incisive paper, Ricardo Otheguy and Ofelia Garcia discuss "Conceptualizations as Predictors of Degree of Contact in U. S. Spanish". They examine the idea of conceptual uniqueness of languages in contact phenomena, arguing that conceptualizations that are generally found among English-speakers in North America, but not among Spanishspeakers in Latin America, have been diffused into the culture of Latinos in the United States (page 135). In "Creoloid Phenomena in the Spanish of Transitional Bilinguals", John M. Lipski underlines the problems related to the term semi speakers

Introduction

5

and explains why explains why he opts for the more neutral term transitional bilinguals (TB). Lipski points out the lack of even a rough estimate of the proportional of TB Spanish speakers in the U. S., either in the schools or in society as a whole, and describes a combination of the features which would give a reasonable prediction of TB status. In "Diversification and Pan-Latinity: Projections for the Teaching of Spanish to Bilinguals", Frances R. Aparicio regards not only how the profile of the Hispanic student in the United States has been rapidly changing, but how this fact underlines the need to re-examine some of the controversial pedagogical questions in the teaching of Spanish to U. S. Hispanic bilingual students of diverse cultural and linguistic backgrounds. David Barnwell's essay, titled "Oral Proficiency Testing and the Bilingual Speaker", addresses questions regarding the use of the A C T F L Oral Proficiency interview to measure the language proficiency of U. S. Hispanic bilinguals. The final paper, "Literacy Stories: Features of Oral Discourse", by Rene Cisneros and Elizabeth A. Leone, analyzes a broad range of features of spoken, unplanned discourse, in order to gain insight into the ways U. S. Spanish language users actually speak. In putting together this volume, we aim to provide the reader with linguistic discussions and questions which reflect the current research trends in the field. To borrow a phrase coined by linguistics professor Garland Bills, we now must speak of "the many faces of U. S. Spanish". In order to acquire a truer picture of Hispanics in the United States today, in order to better understand the intricacies of the language in contact with English, we must have a more encompassing perspective of the cultural and linguistic varieties of the Spanish-speaking people in both urban and rural America. It is our hope that the perspectives presented in the volume's discussions will stimulate interested scholars to carry out research which remains to be done.

Notes 1. A tangible reflection of research on Spanish in the United States over the past quarter century is the growing bibliography of seminal monographs and anthologies, including Bilingualism in the Barrio (Fishman, Cooper, Ma, et al.

6

Introduction 1971), El lenguaje de los chicanos (Chavez, Cohen, and Beltramo, eds. 1975), Studies in Southwest Spanish (Ornstein and Bowen, eds. 1976), Chicano sociolinguistics (Penalosa 1980), Spanish in the United States; Sociolinguistic aspects (Amastae and Elias-Olivares, eds. 1982), Chicano discourse (Sanchez 1983), Spanish in the United States; Beyond the Southwest (Elias-Olivares, ed. 1983), Spanish and English of United States Hispanics: A critical annotated bibliography (Teschner, Bills, and Craddock, eds. 1985), Spanish language use and public life in the U. S. A. (Elias-Olivares, et al. 1985), Research issues and problems in United States Spanish (Ornstein, Green, and Marquez, eds. 1988), Spanish in the U. S.: The language of Latinos (Wherritt and Garcia, eds. 1989), Spanish in the United States: Sociolinguistic aspects (Bergen 1990). Although not exclusively about Spanish in the U. S., Bilingualism and language contact: Spanish, English, and Native American languages (Barkin and Brandt, eds. 1982), provides insighful studies. Two additional collections include excellent overviews and articles on the subject. These are language diversity: Problem or resource? A social and educational perspective on language minorities in the United States (McKay and Wong, eds. 1988) and Sociolinguistics of the Spanishspeaking world: Iberia, Latin America, United States (Klee and Ramos-Garcia, eds. 1991).

2. For an excellent overview of "The Language Situation of Puerto Ricans", see Ana Celia Zentella's essay of the same title in Language diversity. Also, for an equally informative and useful overview on "The Language Situation of Cuban Americans", see Ofelia Garcia's essay in the same collection and Ana Roca's response on the issue of "Language Maintenance and Language Shift in the Cuban American Community of Miami", in Language planning (Marshall, ed. 1991). Language diversity also includes a valuable overview essay by Guadalupe Valdes on the language situation of Mexican Americans.

References Amastae, Jon—Lucia Elias-Olivares 1982 Spanish in the United States: Sociolinguistic aspects. Cambridge: Cambridge University Press. Barkin, Florence — Elizabeth A. Brandt —Jacob Ornstein-Galio (eds.) 1982 Bilingualism and language contact: Spanish, English, and Native American languages: New York: Teachers College Press, Columbia University. Barnach-Calbo Martinez, Ernesto 1980 La lengua espanola en los Estados Unidos. Madrid: Oficina de Education Iberoamericana.

Introduction

1

Bergen, John 1990 Spanish in the United States: Sociolinguistic aspects Washington: Georgetown University Press. Elias-Olivares, Lucia (ed.) 1983 Spanish in the U. S. setting: Beyond the Southwest. Rosslyn, VA: Clearinghouse for the National Association for Bilingual Education. Elias-Olivares, Lucia — Elizabeth A. Leone-Rene Cisneros — John Gutierrez (eds.) 1985 Spanish use and public life in the USA. (Contributions to the Sociology of Language 35.) Berlin: Mouton de Gruyter. Fishman, Joshua Α. —Robert L. Cooper — Roxana Ma 1971 Bilingualism in the Barrio. (Language Science Monographs 7). Bloomington: Indiana University Press. Garcia, Ofelia 1988 "The language situation of Cuban Americans", in: Sandra Lee McKay-Sau-ling Cynthia Wong (eds.), 166-192. Hernandez-Chavez, Eduardo — Andrew D. Cohen — Anthony D. Beltramo 1975 El Lenguaje de los Chicanos: Regional and social characteristics used by Mexican Americans. Arlington, VA: Center for Applied Linguistics. Klee, Carol —Luis A. Ramos-Garcia (eds.) 1991 Sociolinguistics of the Spanish-speaking world: Iberia, Latin America, United States. Tempe, Arizona: Bilingual Review Press/Editorial Bilingüe. McKay, Sandra Lee — Sau-ling Cynthia Wong, (eds.) 1988 Language diversity: Problem or resource? A social and educational perspective on language minorities in the United States. New York: Newbury House. Ornstein, Jacob —J. Donald Bowen 1976 Studies in Southwest Spanish. Rowley, MA: Newbury House. Ornstein, Jacob —George K. Green —Dennis J. Bixler-Marquez 1988 Research issues and problems in United States Spanish. Pan American University at Brownsville, Texas, in cooperation with the University of Texas at El Paso. Penalosa, Fernando 1980 Chicano sociolinguistics. Rowley, MA: Newbury House. Roca, Ana 1991 "Language maintenance and language shift in the Cuban American community in Miami: The 1990s and beyond", in: David F. Marshall (ed.), Language planning: Focusschrift in honor of Joshua A. Fishman. Amsterdam: John Benjamins. Sanchez, Rosaura 1983 Chicano discourse: Socio-historic perspectives. Rowley, MA: Newbury House.

8

Introduction

Teschner, Richard — Garland Bills—Jerry Craddock (eds.) 1985 Spanish and English of United States Hispanos: A critical, annotated, linguistic bibliography. Arlington, VA: Center for Applied Linguistics. Valdes, Guadalupe 1988 The Language Situation of Mexican Americans, in: Sandra Lee McKay-Sau-ling Cynthia Wong (eds.), 111-139. Wherritt, Irene —Ofelia Garcia (eds.) 1989 US Spanish: the language of Latinos. (International Journal of the Sociology of Language 79.) Zentella, Ana Celia 1988 "The language situation of Puerto Rican's", in: Sandra Lee McKay — Sau-ling Cynthia Wong (eds.), 140-165.

Linguistic heterogeneity, civil strife and per capita gross national product in inter-polity perspective Joshua A. Fishman

Among the major charges against linguistic heterogeneity that are encountered in the popular press in political discussions are the claims that it leads to or exacerbates civil strife, on the one hand, and that it lowers national productivity, on the other hand, in both cases because linguistic heterogeneity presumably counteracts rationality, civility, sensitive communication and the smooth operations of government and industry alike. These are charges to which some of my own recent research pertains 1 and it may prove instructive to compare popular thinking and sociolinguistic scholarship in this connection. Each of the foregoing charges can be translated into a formal interpolity hypothesis, namely: (a) The greater the degree of linguistic heterogeneity in a country ("degree of linguistic heterogeneity" being operationalized as the proportion of the population claiming as its own the major mother tongue of any given country, the smaller that proportion the greater the degree of linguistic heterogeneity, and, correspondingly, the larger that proportion the smaller the degree of linguistic heterogeneity), the greater the frequency and severity of civil strife in that country, and, similarly: (b) the greater the degree of linguistic heterogeneity in a country, the lower the per capita gross national product in that country.

Methodological considerations Until quite recently it would have been virtually impossible to do conclusive, worldwide, empirical research on hypotheses such as the above because of the large number of countries ( = polities) and the large number of additional variables that need to be examined in order to rigorously test these hypotheses. There are approximately 170 polities in

10

Joshua A. Fishman

the world today and if these were simply to be compared two at a time, in order to determine whether the linguistically more homogeneous one differs significantly (with respect to severity/frequency of civil strife and/ or with respect to per capita gross national product) from the linguistically less heterogeneous one, over 25,000 individual comparison would have to be made. Obviously, it would be both inordinately difficult to undertake and then to make sense out of so many comparisons. Actually, however, the methodological problem indicated above for two variables and 170 countries is compounded many times over if we realize that in order to test our hypotheses we also need to simultaneously consider many, many other variable that are descriptive of the countries of the world in addition to the two that we are focusing upon. What we really want to know is whether linguistically more heterogeneous and less heterogeneous polities differ in connection with civil strife and per capita gross national product over and above (independently of) the differences between such countries due to any and all other factors to which civil strife and per capita gross national product may be indirectly related. Civil strife, e. g., should be considered too when we are looking into the relationship between degree of linguistic heterogeneity and per capita gross national product. This is necessary so that we can tell whether any encountered relationship between degree of linguistic heterogeneity and per capita gross national product, whatever that may, is masked by or even due to the relationship between civil strife and per capita gross national product. And the same is also true, of course, with respect to degree of religious heterogeneity, degree of racial heterogeneity, proportion of the annual budget allocated to military expenses, etc., etc. Only if we can also consider all other possibly contributing variables can we tell whether linguistic heterogeneity per se really makes an independent (i. e. a non-redundant) contribution to per capita gross national product. However, there are an almost endless number of such other possibly contributing variables (indeed, political scientists have perfected 230-some different dimensions (238 to be exact), all in all, for describing countries) and all of these need to be utilized simultaneously, together with linguistic heterogeneity, when attempting to account for inter-polity differences in civil strife or in per capita gross national product. Thus, our task is to compare all countries simultaneously on all variables simultaneously if we really want to find out whether degree of linguistic heterogeneity is a truly independent (necessary, non-redundant) correlate of either civil strife or per capita gross national product. The price of bananas and the number of gloves sold on a particular day may

Linguistic

heterogeneity

11

correlate substantially. However, only if we include all other variables that might also possibly influence the cost of bananas on a particular day (e. g., average daily temperature in the banana groves, daily transportation costs between the groves and the markets, labor costs in the groves and in the markets, etc., etc.) can we safely avoid coming to the specious conclusion that the number of gloves sold is really genuinely (that is: independently) related to the cost of bananas. Does this sound like an impossibly tall order: to analyze hundreds of countries and hundreds of variables simultaneously? Do we have the necessary data in order to do that and do we have the necessary methods by means of which to do that?

Data and analytic methods Fortunately, the variety and even the quality of the data we need has been provided by the cross-polity databanks that American political scientists, both in government and in academia, have prepared, and repeatedly revised and expanded during the past quarter century. These databanks provide sifted, corrected and continually updated data on all the countries of the world in conjunction with over one hundred thirty different economic, political, social, cultural, historical, geographic and demographic variables. This data is not perfect, but it is the best available today anywhere in the world and since quite a bit of American economic, political and military planning and policy is based on this data it must at least be reasonably good on the whole and may even be quite a bit better than that. The analytic methods to do what needs to be done have been provided by statisticians and computer specialists who have relatively recently perfected approaches (primarily cumulative multiple correlation and factor analysis) that make it relatively easy, on the one hand, to examine huge amounts of multivariate data and, on the other hand, to parsimoniously zero in on the relatively few variables in any large data-set that are really the only independent (and, therefore, the only crucial) variables in explaining or in accounting for the variation in any given criterion variable. My co-workers and I are, I believe, the first to put both the exhaustive data-sets and the new statistical analytic methods to joint use in conjunction with determining the role of inter-polity variation in

12

Joshua A. Fishman

linguistic heterogeneity in so far as as accounting for cross-polity variation in civil strife and in per capita gross national product. Perhaps an apology is in order for the above brief detour into methodological issues, which no matter how brief it may be to the specialist inevitably seems overly long and insufficiently understandable to the nonspecialist. My concern is basically related to the usual scholarly preoccupation that findings, interpretations and conclusions rest upon foundations that are as firm as possible. It is also related to an attempt to get away from a contrasted approach which may be referred to as "the favorite country approach". We are all good, journalists, politicians and academics alike, at arguing from "preferred cases", that predictably provide negative answers to the questions we have initially posed about linguistic heterogeneity's possibly harmful consequences by examining them only in connection with Switzerland (where linguistic heterogeneity results in neither heightened civil strife nor in lowered per capita gross national product) or that answer the same questions in the positive by referring only to India, without even pausing to consider the many other dimensions (besides linguistic heterogeneity) on which Switzerland and India differ substantially and differ in ways that are directly related to civil strife and/or per capita gross national product. It is to escape from this more usual approach of arguing from "preferred (and biased) cases" that I have gone out of my way to study all variables and all countries simultaneously in order to clarify the true (i. e., the independent) relationship between linguistic heterogeneity, civil strife and per capita gross national product. 2

Civil strife3 Political scientists have kept records on civil strife in all of the countries of the world in four different ways: (a) magnitude and frequency of conspiracy against the established government, (b) magnitude and frequency of internal warfare due to revolution, sedition or secession, (c) magnitude and frequency of internal turmoil (riots, strikes, protests) and (d) a composite average of the above three. 4 The latter is the only measure of civil strife that will be discussed in this paper, although essentially identical results obtain from analyses of the contribution of linguistic

Linguistic

heterogeneity

13

heterogeneity to civil strife as measured by each of the other measures of civil strife as well. A cumulative multiple correlation analysis which begins by aiming over 230 variables at the composite civil strife criterion variable reveals that the differences in composite-magnitude-and-frequency of civil strife across all of the countries of the world are both highly predictable and parsimoniously predictable. Out of 230 — some possible predictors of this criterion only a certain 13 make truly independent contributions that, taken together, yield the optimal multiple correlation of .82. Particularly noteworthy for our purposes is the fact that linguistic heterogeneity is not a member of this optimal subset of predictors. Of course, linguistic heterogeneity does have a certain correlation with (.21) with civil strife and, therefore, if we were to disregard, for a moment, its redundancy with other variables it would, at most, explain only 4% (or .21 squared) of the total variance in civil strife. However, when linguistic heterogeneity is confronted with all other possible predictors of civil strife it is eclipsed entirely and it becomes clear that it itself explains nothing at all in connection with civil strife that is not explained better (i. e., less redundantly or more independently) by other variables. What are the variables that displace linguistic heterogeneity from any consideration as an independent predictor of civil strife? They are variables such as (a) short term deprivation, (b) persistent deprivation, (c) the absence of coercive potential by the government and (d) the presence of anomic groups that are organized to exert pressure for their own particular benefit. Just a few variables such as these immediately come to the fore to form the relatively small optimal subset of 13 indispensible predictors that, when taken together, account for 68% ( = .82 squared) of the worldwide inter-polity variance in composite civil strife. A cumulative multiple correlation of .82 is quite an impressive accomplishement, but, of course, this still leaves about 32% of the variance in civil strife to be accounted for by variable not yet utilized (i. e., by variables not yet in the total 230 — some variable data set) or by better measures of the variables that have been utilized in this data-set. In either case, it is relatively certain that linguistic heterogeneity will not be among them because we have already utilized it in two different measurement modes 5 and the results have been identical: linguistic heterogeneity simply does not appear in the optimal subset of independent variables needed for the maximal and most parsimonious prediction of civil strife. We may have the conviction that in one country or another linguistic heterogeneity does make an important contribution to the explanation of civil strife, and perhaps that is indeed so. Civil strife

14

Joshua A. Fishman

in each country is overdetermined by its own historical circumstances, but in the world as a whole, across all countries and across all historical circumstances, once deprivation, central coercive potential and the presence of organized anomic groups are taken into consideration, it is entirely unwarranted to posit linguistic heterogeneity itself as a co-cause of civil strife. Whatever minor and redundant importance it may have is entirely attributable to its redundancy with the stronger independent predictors that do wind up in the optimal subset of predictors of civil strife.

Per capita gross national product6 Turning now to our second criterion variable, we proceed to ask whether worldwide inter-polity variance in linguistic heterogeneity makes an independent contribution to the explanation or prediction of worldwide inter-polity variance in per capita gross national product. Once again it becomes evident that the worldwide inter-polity variance in our second criterion variable is highly predictable on the basis of a relatively small subset of optimal predictors. Out of our original pool of 233 available variables only a certain 10 are needed in order to yield a cumulative prediction of .90 and, once again, linguistic heterogeneity is not a member of this optimal subset. Linguistic heterogeneity naturally has its own correlation with per capita gross national product, a correlation of — .32, which, at best (i. e., were it to be non-redundant) would indicate that linguistic heterogeneity could explain only 10% ( = .32 squared) of the worldwide inter-polity variation in per capita gross national product. However, linguistic heterogeneity is far from being a non-redundant predictor of per capita gross national product. When linguistic heterogeneity is confronted with all other possible predictors of per capita gross national product it is eclipsed entirely, indicating that in itself it really explains nothing at all about per capita gross national product. The cross-polity variables that do wind up in the optimal subset of predictors of worldwide inter-polity variation in per capita gross national product are primarily (a) governmantal modernization, (b) per capita newspaper circulation, (c) presence of a multi-party and parliamentaryrepublican government, and (d) industrial rather than agrigultural concentration of the workforce. All in all, these commulatively account for 81% ( = .90 squared) of the total worldwide inter-polity variation in per

Linguistic

heterogeneity

15

capita gross national product, leaving 18% still to be accounted for either by additional variables, above and beyond the 238 already conceptualized, operationalized and defined to date, or, alternatively, by the same variables measured more reliably and more validly. Even then it seems unlikely that linguistic heterogeneity would wind up in the optimal subset of predictors of per capita gross national product, because it is probably not sufficiently independent of the optimal subset of predictors, nor sufficiently powerful in comparison to them, in order to do so. There are simply much more immediate and much more independent predictors of cross-polity variation in per capita gross national product for us to be able to posit linguistic heterogeneity as a serious co-predictor, over and above what is contributed by governmental modernization, parliamentary democracy, popular literacy and large scale industrialization. 7

Linguistic heterogeneity as a criterion variable8 We have twice looked at linguistic heterogeneity as an independent (or non-redundant) predictor variable and have found it to be a distinctly weak one. We can now try to arrive at a clearer understanding of what inter-polity variation in linguistic heterogeneity is related to by viewing it itself as a criterion variable and then asking what other variables in our data set tend to predict inter-polity variation in linguistic heterogeneity. Once again we obtain a relatively small and parsimonious optimal subset of predictors which, when taken together via cumulative multiple correlation proceedures, yield a correlation of .88. The most powerful variables in the optimal subset, a subset which accounts for 78% ( = .88 squared) of the worldwide cross-polity variance in linguistic heterogeneity, are (a) Christianity or Islam as the dominant religion of a polity, (b) degree of religious and racial homogeneity, (c) whether a polity is historically a former Spanish colony and (d) whether a polity is Western or early-Westernized rather than late-Westernized under colonial auspices. Several comments are in order about the above optimal subset of predictors. First of all, they are all negatively correlated to the criterion of linguistic heterogeneity, i. e., they all tend to make for linguistic homogeneity rather than for heterogeneity. Secondly, we should note that neither civil strife nor per capita gross national product are in this subset, further confirming what we have reported in the two prior analyses,

16

Joshua A. Fishman

immediately above. Finally, it should be noted that if deprivation is the major underlying dimension of civil strife, and if modernization is the major underlying dimension of per capita gross national product, than homogenization, whether religious, demographic or political-developmental seems to be the major underlying (negative) predictor of linguistic heterogeneity. Obviously, at a deeper conceptual level, there are weak relationships between these three criteria as well as between their respective optimal subsets of predictors. This can be gleaned from the fact that civil strife and per capita gross national product correlate — .28 with each other, somewhere in between the correlation of linguistic heterogeneity with civil strife (.21) and the correlation of linguistic heterogeneity with per capita gross national product ( — .32). Nevertheless, these links are weak and no one of them is the best predictor of the other. This means that each can be best predicted without the other, because each is influenced by the other not only weakly but indirectly, via the influence of more primary and less redundant predictors. Accordingly, this implies that lingua francas and bilingualism enable polities to attain higher per capita gross national product and to avoid civil strife regardless of their degree of linguistic heterogeneity.

Some possible positive consequences of linguistic heterogeneity Thus far we have done no more than discredit the necessarily negative consequences that are commonly attributed to linguistic heterogeneity. However, the absence of negative consequences is not at all the same thing as the presence of positive consequences. The verifiability of any positive consequences of linguistic heterogeneity, if, indeed, there are any, is a question for future research to pursue. It may be that the pursuit of such consequences will be an equally thankless one, since in that connection too linguistic heterogeneity may be an exceedingly weak independent (i. e., non-redundant) inter-polity dimension. Perhaps what is involved, in so far as any possible positive consequences of linguistic heterogeneity are concerned, is not so much the degree of such heterogeneity as the type of policy adopted vis-a-vis linguistic heterogeneity. A serious problem in connection with the exploration of whatever positive correlates there

Linguistic

heterogeneity

17

may be of linguistic heterogeneity is the sad fact that political and other social scientists have not yet exhaustively studied the countries of the world with respect to those very dimension that seem to me to present the best prospects of yielding such findings. Taking these two considerations together, (a) positive policies (or positive languagestatus planning) with respect to linguistic heterogeneity and (b) possibly positive consequences of linguistic heterogeneity, I would opt to investigate linguistic heterogeneity and such variables as (a) proportion of the minority language population that has acquired the majority or dominant language [examining oralcy and literacy separately], (b) proportion of the minority language population that has completed secondary education and received higher education, (c) proportions of the minority language population that is employed and that is living above the poverty line, (d) some measure of positiveness of intergroup relations, (e) some measure of degree of bilingualism and biliteracy in the dominant language population. The above list of possibly positive consequences of a more positive (culturally pluralistic) policy toward linguistic heterogeneity are no more than hunches derived from the social science literature. Their exploration will require a combination of expanded inter-polity data, statistical sophistication and sociolinguistic expertise. Their exploration will get us away from constant "damage control" with respect to linguistic heterogeneity, where we are always concerned with either containing or disproving the alleged negative consequences of linguistic heterogeneity, and into the domain of "positive bilingualism", i. e., into the domain philosophically advanced by J. G. Herder, B. L. Whorf and H. Kallen. 9 Another lead that deserves to be investigated, which flows directly from the above discussion, is the possibility that where the policy toward linguistic heterogeneity is a non-accepting, negative or punitive one then the converse of the above tentative hypotheses might be confirmed as might some of the other negative consequences of linguistic heterogeneity that have long been suspected. It would be more than I dare hope for to believe that such further research (of either kind) might overcome the emotional and non-rational climate in which issues of linguistic heterogeneity have long been discussed. Emotional issues do not respond easily to empirical evidence. Nevertheless, having devoted much of my life to the empirical study of emotional sociolinguistic issues, I continue to believe that the possibly positive and the possibly negative correlates of linguistic heterogeneity have been too little studied contextually, and that those who are not afraid of being "confused by the facts" and not too

18

Joshua A. Fishman

impatient to pursue the facts wherever they may lead, have a great deal of work ahead of them before light can begin to counterbalance heat in this sensitive area.

Notes

1. This paper summarizes four previous studies, all of which were conducted with Frank R. Solano (Ferkauf Graduate School of Psychology, Yeshiva University, Bronx, Ν. Y. 10661) and one of which also benefited from the collaboration of Grant D. McConnell (International Center for Research on Bilingualism, Laval University, Quebec City, Quebec). 2. Prior to my own most recent research almost all worldwide interpolity studies of linguistic heterogeneity concentrated only on a very few variables at a time. See, e. g., Pool (1969), McRae (1983). My own early contribution to research on interpolity correlates of linguistic heterogeneity (see Fishman 1966) did examine all variables available at that time, but did so, by and large, one variable at a time rather than via statistical approaches that permit all variables to be examined simultaneously and contrastively with each other. 3. This section briefly summarizes Fishman and Solano (1990). 4. For full details as to separate as well as composite worldwide interpolity measurement of civil strife, see Feierabend and Feierabend (1972). 5. Two measures of linguistic heterogeneity were employed, a dichotomous measure and a continuous measure. The dichotomous measure characterized each polity as either having or not having at least 15% of its population claiming a mother tongue other than the dominant one of the country. The continuous measure characterized each polity by the exact proportion of its population claiming a mother tongue other than the dominant mother tongue of that country. The identical nature of the findings using these two different measures of linguistic heterogeneity is reported in Fishman, Solano and McConnell (1991). 6. This section briefly summarizes Fishman and Solano (1989 a). 7. While popular literacy may be as much a cause as a consequence of per capita gross national product, insofar as cumulative multiple correlation analysis is concerned, linguistic heterogeneity is apparently neither the one nor the other in any strong or non-redundant fashion. 8. This section briefly summarizes Fishman and Solano (1989 b). 9. For a discussion of "positive bilingualism", and a review of Herder's, Whorf s and Kallen's theoretical contributions to the topic, see Fishman (1978).

Linguistic heterogeneity

19

References Feierabend, I. K. and R. L. Feierabend (eds.) 1972 Anger, violence and politics: Theories and research. Englewood Cliffs: Prentice Hall. Fishman, Joshua A. 1966 "Some contrasts between linguistically homogeneous and linguistically heterogeneous polities", Sociological Inquiry 6: 146—158. (Reprinted [1968] in Joshua A. Fishman — Charles A. Ferguson —Jyotirindra Das Gupta (eds.) Language Problems of Developing Nations. New York, Wiley, 5 3 - 6 8 . ) 1978 "Positive pluralism: Some overlooked rationales and forefathers", in: James E. Alatis (ed.) International Dimensions in Bilingual Education. Washington: Georgetown University Press, 42 — 52. (Reprinted in Joshua A. Fishman, Rise and Fall of the Ethnic Revival: Perspectives on Language and Ethnicity. Berlin, Mouton de Gruyter, 445 — 455.) Fishman, Joshua A. — Frank R. Solano 1989 a "Cross-polity linguistic homogeneity/heterogeneity and per-capita gross national product: An empirical exploration", Language Problems and Language Planning 13: 103 — 118. 1989 b "Societal factors predictive of linguistic homogeneity/heterogeneity at the interpolity level", Cultural Dynamics 1: 414 — 437. 1990 "Crosspolity perspective on the importance of linguistic heterogeneity as a contributory factor in civil strife", Canadian Review of Studies in Nationalism 17: 131 — 146. Fishman, Joshua A. — Frank R. Solano — Grant D. McConnell 1991 "A methodological check on three cross-polity studies of linguistic homogeneity/heterogeneity", in: Mary E. McGroarty —Christian J. Faltis (eds.) Language in school and society: Policy and pedagogy. Berlin: Mouton de Gruyter, 21—29. McRae, Kenneth D. 1983 Conflict and Compromise in Multilingual Societies. Waterloo: Wilfrid Laurier University Press, 5 — 33. Pool, Jonathan 1969 "National development and language diversity", La Monda LingvoProblema 1: 140—156. (Reprinted, slightly revised (1972), in Joshua A. Fishman (ed.) Advances in the Sociology of Language, vol. 2, 213-230.)

A reconsideration of the notion of loan translation in the analysis of U. S. Spanish1 Ricardo Otheguy

0. Introduction The notion of loan translation, under a variety of such largely synonymous names as caique, loanshift, and semantic loan, is a staple in any discussion of Spanish-English contact in the United States and, more generally, in any treatment of the linguistic consequences of bilingualism. 2 This paper attempts to show that, under any of its many names, the notion of loan translation tends to obscure rather than illuminate the nature of English influence over U. S. Spanish and, more generally, to misrepresent the process of language contact in any setting. It is proposed here that loan translations are either: (a) phrases revealing contact-induced systemic linguistic changes already encompassed under other constructs, or (b) phrases that do not reveal any systemic linguistic changes at all. Further, it will be argued that (c) because of this flawed construct, the amount of linguistic interpenetration taking place in contact communities is highly overstated. In the specific case of U. S. Hispanics, it will be argued that (d) the notion of loan translation contributes to mistaking as Anglicization of their language what is simply the Americanization of their culture. 3 In recent theoretical work, Thomason-Kaufman (1988: 35 ff) have advanced new proposals regarding correlations between sociolinguistic processes and types of linguistic interference. But within the newly proposed interference types (borrowing, for communities maintaining their language; substratum, for those shifting from it), it is still useful to distinguish, as in Weinreich (1953), between transferring and modeling. In transferring, the alterations result in additions to the inventory of the impacted variety; in modeling, alterations occur in its existing inventory. Within modeling, those alterations that affect only a single word are generally known as semantic extensions; those that encompass entire phrases, which are the focus of attention here, are usually called loan

22

Ricardo

Otheguy

translations or caiques. 4 The illustrations of both types of modeling in (1) and (2) are taken from the speech of a six year-old Spanish-English bilingual girl living in New York. (1)

[After seeing another child yell across the street to say hello to her brother, she asks:] Mami, icbmo ese nino sabe a Eric? [RAO 2/90]. 'Mami, how does that child know Eric?'

(2)

[She is mounting what she knows is a difficult case for taking a prized pen from her father's desk] Papi, til me prestas esa pluma y yo te la doy para aträs; please, please, prestamela y yo te la doy para aträs. [RAO 7/89]. 'Daddy, you lend me that pen and I'll give it back to you, please, please, lend it to me and I'll give it back to you.'

The saber of (1) appears where speakers of general Spanish would almost certainly use conocer. It is said to be a semantic extension because, on the model of English know, the meaning of saber has been extended to areas that in unimpacted varieties are covered by conocer. The dar para aträs of (2) appears where speakers of general Spanish would almost certainly use devolver 'to return', and is said to be a loan translation or caique modeled on English give back.5 There is every indication that in works on language contact the term loan translation is not simply a /αςοη de parier about contact settings, but that it is intended as a reference to systemic changes occurring in the structure of the impacted languages. Almost without exception, linguistic studies of bilingual settings include discussions of loan translation. All the major survey or theoretical works on language contact of the past fifty years make use of the construct. It is included, for example, in Weinreich (1953), Appel-Muysken (1987), Holm (1988), and Thomason —Kaufman (1988). In studies that correlate micro social-interaction factors with types of contact phenomena, one of the types is almost always loan translation (Oksaar 1979: 106). The same is true of areal studies. In a recent analysis of the languages of Meso-America, the authors separate out those Sprechbund features "that are perhaps better considered part of an ethnography of communication than of a formal grammar" (Campbell — Kaufman — Smith-Stark 1986: 558). Significantly, loan translations and caiques are not in this Sprechbund section, but rather in the Sprachbund section dealing with "formal grammar" and with "diffused structural traits" (1986: 558). The notion of loan transla-

A reconsideration

of the notion of loan translation

23

tion is also standard fare in discussions of the influence of English on Spanish (Craddock 1976, Dillard 1975, Espinosa 1975, Milan 1982, Montes Giraldo 1973, 1985). As shown below, these authors leave no doubt that they conceive of loan translation as an alteration in the structure of U. S. Spanish. Similarly, with reference to the particular construction that will become the focus of attention below, namely verbs in contact Spanish followed by para aträs, calquing is taken seriously as a manifestion of contactinduced systemic changes. The thorough and carefully nuanced treatment by Lipsky (1985: 91 ff) presents this construction as "an apparent caique on back,'''' and places it in the context of "syntactic Anglicisms." After correctly pointing out the many ways in which the contact usage of para aträs exploits inherent Spanish possibilities, Lipsky nevertheless concludes that "[b]ilingual interpenetration is a necessary component of any proposed explanation" (1985: 99). Contrary to these prevailing views, then, this paper will take an alternative approach to items described under the label loan translation/ caique. It is an approach that rests on three generalizations that are standard in the study of language, but whose consequences for an understanding of contact-induced phenomena have not been sufficiently appreciated. This paper will stress the familiar assumptions (a) that different speech communities operate with conceptual inventories that reveal only some areas of overlap; 6 (b) that under the pressure of contact many of the non-congruent areas in these inventories tend to converge; 7 and (c) that this conceptual convergence is distinct from, and can occur independently of, linguistic convergence. 8 Based on these assumptions, and using data from the Spanish of the United States, this paper will argue that items described as loan translations are simply either: (a) semantic extensions (or in some cases grammatical alterations) that happen to be reported in phrasal contexts; or (b) phrases that adhere to a non-contact linguistic system (in this case general Spanish) but that reveal the impacted community's adaptation to cultural and conceptual patterns prevalent in the dominant group, in this case the broader North American society. It is in these latter cases that the notion of loan translation overstates the amount of linguistic contact, presenting as adaptations occurring in linguistic structure what are simply cultural or conceptual adaptations expressed through language. 9

24

Ricardo

Otheguy

1. Loan translations as mislabeled semantic extensions The root of the problem lies in the criterion that is used to define loan translation, especially to distinguish it from semantic extension. Because it continues to serve as model for all subsequent definitions, the relevant source to examine the distinction remains Weinreich (1953), who defined semantic extensions as contact-induced changes in the meanings of individual words. 10 This definition seems workable in principle, since the locus of contact is squarely within the linguistic system, in this case within lexical semantics. But Weinreich then defined loan translations as "unusual combinations of words" (1953 [1974]: 51), giving rise to the uncertainty as to whether the combination was unusual for cultural or linguistic reasons. Among Weinreich's examples of semantic extension was the use in U. S. Spanish of ministro to mean a minister of a Protestant church; among those of loan translation, Canadian French escalier de feu to refer to a fire escape.11 Weinreich proposed that what he called the semantemes of English minister allowed it to be used for officers of both government and church, while those of Spanish ministro only allowed its use for officers of a government. (For an officer of a Protestant church, general Spanish would have had pastor.) In the Spanish of the United States, and under pressure from English, the distinction between ministro and pastor collapsed, so that one could find phrases such as ministro pentecostal 'pentecostal minister'. The trouble with this account of semantic extensions is that exactly the same analysis can be provided for loan translations. In non-contact French a semanteme presumably distinguished feu from incendie, so that feu refered to burning in general while incendie was specialized for accidental fires. Under the influence of English fire, which has both the general and the specialized senses, the meaning of feu would have been extended, thus making it possible to say, not escalier d'incendie but escalier de feu. It would seem, then, that even assuming the facts to be as Weinreich saw them, escalier de feu would be readily explainable simply in terms of semantic extensions, without having to appeal to a separate notion of loan translation. Granted, the combination escalier de feu may have been at some point "an unusual combination"; but also unusual was at some point the combination ministro pentecostal, and unusual in fact will necessarily be any combination of words that arises by the process of

A reconsideration of the notion of loan translation

25

semantic extension. (If there were no unusual combinations, no semantic extensions would be detected in the first place.) On Weinreich's definition, there seems to be no way to sustain a construct loan translation that would be different from, and additional to, semantic extension. Other definitions have been proferred both before and after Weinreich, some using the term loan translation itself, others using the synonymous semantic loan, loanshift, and caique. Most center on the idea that loan translations involve structural or functional changes. Haugen (1950 [1972]: 85) characterizes these items as resulting from "functional shifts of native morphemes"; Craddock (1981: 202 ff) sees them as "constructions" of the source language that have come into the recipient variety; and Dillard (1975: 194, 202) describes them as items reflecting a "structural breakdown." Sometimes the process is described as involving "wordfor-word translations" (Craddock 1981: 208, Holm 1988, 86) or "literal translations" from the source language (Dillard 1975: 194, Milan 1982: 197). But if these definitions are applied to some of the typical examples of loan translation discussed in the literature, difficulties arise immediately. The structure of escalier de feu, for example, would appear to be [NounPreposition-Noun], which is a familiar French collocation, as is the specific structure [Noun-de-Noun]. The construction is, moreover, different from that of the model (fire escape or fire stairs) which is not a [Noun-Preposition-Noun] construction, but rather a [Noun-Noun] construction. One is hard pressed to imagine in what sense it can be said that escalier de feu is a loan translation because it reproduces the structural or constructional features of fire escape, when escalier de feu is constructed in a manner different from fire escape and in accordance with structural principles that are indigenous to French. Insofar as one can come to any conclusions from this example, whatever it is that resembles the model phrase in a loan translation — over and beyond the semantic extensions it may contain — it is certainly not its structure. There should be no misunderstanding regarding what is, and what is not, being called into question here. It seems clear that contact settings can and do give rise to crosslinguistic influences in grammar that go well beyond the long-recognized borrowing of inflectional and derivational affixes (already pointed out by Whitney 1881). Such influences are well documented in the work of prior generations of scholars (Casagrande 1954, Hoijer 1948) as well as in that of current ones (Daiuta 1984, Dvorak 1983, Gal 1989, Hill-Hill 1980, Lindenfeld 1971, Thomason - Kaufman 1988). The point is not to deny the soundness of postulating contact-

26

Ricardo

Otheguy

induced systemic changes. The point, rather, is that such deviations cannot be postulated for the kinds of items generally called loan translations and caiques. Similar difficulties attend definitions of loan translations as "wordfor-word" substitutions or "literal translations" of the model. First, if the Quebecois had translated fire escape word for word they obviously wouldn't have come up with escalier de feu. They wouldn't have come up with the word escalier at all, since there is no mention of ladders or stairs in the model phrase fire escape. Only if the model was fire stairs, which is not at all clear, could one even consider the notion of word-forword substitution for escalier de feu. More generally, the question must be raised whether it even makes sense to think that there can be such a thing as word-for-word substitution. Presumably, a word-for-word substitution would require the existence, prior to contact, of words in the impacted language that are direct counterparts of words in the model. But it is a generally recognized fact of lexical structure that words in one language do not have these "direct counterparts" in another (Culler 1976). If the vocabularies of languages provided the perfectly equivalent crosslinguistic pairs that the notion of word-for-word substitution presupposes, then speakers of contact varieties would show no tendency to make meanings more congruent, for the simple reason that they would already be so. In sum, neither the notion of structural deviance nor that of literal or word-for-word translation appears as a likely criterion for establishing a construct loan translation that would be in any way distinguishable from the independently necessary one of semantic extension (or from the also independently necessary one of grammatical contact). A final possibility remains for establishing loan translation as a separate contact type. In Weinreich's presentation of examples, feu is reported within a larger phrase while ministro is presented by itself. A case might be made that the semantic extensions that appear within loan translations will, in every other context, adhere to traditional usage, that what defines a loan translation is being the unique environment where a semantic extension is found. In the case at hand, the argument would be that escalier de feu is the only context in which feu is used with the extended meaning. But once the meaning of a word is extended, the same pressures that made for the extension in some initial context are likely to make for its operation in other contexts as well. In point of fact, not only does it appear to be the case that the "accidental" sense of feu predates irtcendie,

A reconsideration

of the notion of loan translation

27

but dictionaries of Canadian French (Narcisse-Eutrope 1974, Turenne 1962) list the phrases station de feu 'fire station', assurance contre le feu 'fire insurance', and aller au feu 'to go to the fire', all of which appear to contain the extended meaning of feu. The same is true of saber mentioned above. The child who produced saber a Eric 'to know Eric' in (1) once heard someone make a reference to a soap-opera character with whom she was familiar and reacted in surprise by saying itu sabes esa noxelai 'do you know that soap opera?' Again here saber has encroached onto the area of conocer, a usage also readily observable among many adult Latinos in the United States. It would seem, then, that contact varieties undergo semantic extensions, of which the saber of (1) may very well be a good example (and of which ministro and feu would also be good examples if the facts were as presented by Weinreich). And they also undergo foreign-inspired grammatical changes. But there is no evidence that they produce an additional type of contact item called loan translation.

2. Loan translations as cultural, not linguistic, contact The incorrect postulation of loan translation does more harm than simply provide an inaccurate taxonomy of language contact phenomena. As impacted varieties develop features of foreign provenience, they also retain, unchanged, a sizable portion of their traditional make up. It is in this sense that from a strictly linguistic point of view there are no contact varieties, only contact features (Hudson 1980: 48 ff). Distinguishing between features that show evidence of contact and those that do not is thus essential for the study of linguistic hybridization. But misled by the notion of loan translation, scholars consistently misidentify as cases of systemic contact many phrases that reveal no systemic traits that would differentiate them from phrases used in non-contact varieties of the same language. The impression that these linguistically unremarkable usages are instances of language contact comes from the fact that either: (a) reference is being made to an object or cultural item that is little known to speakers of non-contact varieties; or (b) reference is being made to an item which, though familiar, is being conceptualized in a manner different from the one usually found in non-contact varieties. In both cases these spurious

28

Ricardo Otheguy

loan translations give evidence, not of language contact, but of the familiar phenomenon of synchronic creativity. 12 Each case will be taken up in turn. (a) The phrase El dia de dar gracias, used by U. S. Hispanics to refer to the holiday that is celebrated in the United States on the fourth Thursday of November, illustrates a new usage of the first kind. The phrase would have no referent and serve no purpose in Spain or Latin America. Yet despite its unmistakable Anglo-American cultural flavor, nothing in it is linguistically Anglicized, as it adheres to a familiar Spanish grammatical pattern and uses lexical items in their familiar senses. And it cannot even be said that the phrase constitutes "a literal translation" of an English model. While its cultural model is Thanksgiving Day, its linguistic model cannot be the English phrase Thanksgiving Day, where word order and the choice of verb form are different. Rather, its linguistic models are such Spanish phrases as El dia de Navidad 'Christmas Day', El dia de Reyes 'Three Kings Day', etc. To be sure, in using the phrase El dia de dar gracias, and in taking the day off from work and gathering with relatives to eat turkey, U. S. Latinos are certainly giving strong evidence of cultural adaptation. But they are not engaged in linguistic modeling, since the lexical and grammatical systems through which the cultural adaptation is being expressed have themselves remained unchanged. 13 (b) A new usage of the second kind is illustrated by the name often given in U. S. Spanish to telephone answering machines, namely mäquina de contestar, a formulation based on the verb contestar 'to answer' and on mäquina 'machine'. In contrast, the general Spanish designation for this appliance is contestador, or contestador automätico, a formulation based on a noun derived from the same verb contestar 'to answer', combined with the word automätico 'automatic'. 14 In this case the United States and Latin America have few or no differences when it comes to the referent (an answering machine); they differ, rather, in the culturally dictated conceptualization of this referent, and in the resulting linguistic formulation used to name it (an automated answerer in traditional usage versus a machine that answers in the contact usage). 15 Since we only know about the difference in conceptualization between English answering machine and traditional Spanish contestador automätico through the different linguistic formulations, it is of considerable relevance that these formulations form part of much broader patterns. For speakers of standard Spanish, other appliances besides the answering machine are conceptualized in terms of the pre-appliance function having

A reconsideration of the notion of loan translation 29 become automatic (e. g., cajero automätico 'teller machine' from cajero 'teller'; lavadero automätico 'laundromat' from lavadero 'sink, washer'). Similarly, for speakers of English the formulation in answering machine, though occasionally abandoned for other patterns, is also broad and not limited to a single item. Thus the cajero automätico coined by Spanish speakers from cajero is paralleled by the English speakers' teller machine coined from teller.16 The position that identical referents can be conceptualized differently accords well with the long-standing view that the linguistic value of an expression is not the same as the object it names (Frege 1892). Two expressions can refer to the same thing and yet have very different meanings, even within the same language. To put a Latin American twist on the familiar example by Alston (1970), the expression Gabriel Garcia Märquez does not mean the same as the author of A Hundred Years of Solitude, and yet they both name the same referent. The same is true when expressions are compared, not within, but across languages. Even though contestador automätico and answering machine do refer to the same thing, they, like Alston's Walter Scott and the author of Waverly, do not mean the same, and do not express the same conceptualization. Assimilating populations, then, are not only exposed through the donor language to new conceptualizations connected to new referents, but also to new conceptualizations connected to familiar referents. A day set aside to be thankful to God (new conceptualization) is connected to the feast on the fourth Thursday of November (new referent); a machine that answers (new conceptualization) is connected to the telephone appliance (familiar referent). In both cases, the tendency among impacted populations is to adopt the influencing group's conceptualizations. But as in the cases discussed above, in mäquina de contestar the words of the new usage, the senses in which they are utilized, and the construction in which they appear, [Noun-Preposition-Noun], are all familiar in noncontact varieties of Spanish. 17 And just as the model for El dia de dar gracias was not the linguistic item Thanksgiving Day, but rather the notion of a day to be thankful to God, the model for mäquina de contestar is not the linguistic item answering machine but the notion of a machine that answers. The usage mäquina de contestar is thus another case of conceptual Americanization without linguistic Anglicization. 18 The standard assumption is made here that the concepts expressed by means of combinations of semantic and structural units are not language but the product of language. 19 Once one draws sharply the distinction between linguistic resources and their use — a distinction which has long

30

Ricardo

Otheguy

commanded consensus among linguists — the only conceptual elements that can properly be said to be part of language are those that are encoded in lexical or structural form. All other conceptualizations are derived from, but do not constitute part of, language. The point being made here is that in loan translations all encoded structural and conceptual elements have remained unchanged, and that the contact alterations all reside in derived, non-linguistic conceptualizations. 20 New usages produced to express new conceptualizations of foreign origin are in fact parallel to new usages produced to express novel conceptualizations of indigenous origin. The usage Teologia de la liberation 'Liberation theology' arose in Latin America to give expression to a new religious concept developed by Spanish speakers, which is exactly the description of how El dia de dar gracias arose in the United States. Both must have been at some point "unusual combinations of words". But the combinations were unusual from a cultural, not a linguistic standpoint. The reason no one had produced Teologia de la liberation was that the concept it expresses did not exist for speakers of Spanish before. And this is exactly the reason no one had produced El dia de dar gracias, which also did not exist for speakers of Spanish before. In both cases, the new usages were allowed by the Spanish system all along. Thus El dia de dar gracias and mäquina de contestar, no less than Teologia de la liberation, are instances of synchronic creativity without diachronic change. From a linguistic point of view, it does not matter that the concept that prompted a new Latin American Spanish usage arose within the ancestral home of the language, whereas the concept that prompted a new U. S. Spanish usage was adopted from foreigners. What matters to determine whether there is language contact is not the cultural origin of new conceptualizations, but the lexical, semantic, and structural resources with which they are expressed. In works on areal phenomena, the position that loan translations involve conceptual but not linguistic convergence receives an interesting sort of confirmation by omission. In the process of marshalling an impressive array of evidence for their contention that Meso-America is a linguistic area, Campbell —Kaufman —Smith-Stark (1986) carefully document all cases of contact through use of the actual phonological forms of the lexical and grammatical items under study. The only exception is the list of the loan translations shared by many of the languages of the area. For these, only an English gloss is given (1986: 553). Thus we learn that in many Meso-American languages wrists and thumbs are referred to by items glossed as "head of hand" and "mother of hand"

A reconsideration

of the notion of loan translation

31

respectively. But the actual phrases, the words of the Meso-American languages themselves, are not offered. This is a most reasonable and sensible omission, and therein lies a lesson. Contrary to what is suggested by placing the discussion of "head of hand" and "mother of hand" in the linguistic rather than the cultural section of the study, these items do not owe their status as loan translations to their linguistic characteristics. What makes them instances of modeling is what they say, not the words, affixes, or constructions with which they say it. It is the mere fact of calling the thumb a "head of hand" that turns that particular item into a loan translation, no matter how this notion is rendered in any particular language. What these MesoAmerican communities shared at the time of contact was not a set of linguistic traits, but a set of concepts. 21 And it was these concepts that diffused through the area in order to produce the current array of phrases. No information other than the common conceptualization reflected in the English gloss is relevant to the analysis of these items as loan translations, and thus, correctly, no other information is offered. Perhaps no clearer demonstration of the irrelevance of linguistic form need be given than its complete absence. 22 The generalizations in this section can be summarized in the following points, (i) The significant motivating novelty for modeling phenomena lies in the changing conceptual inventory to which impacted populations are exposed. It is a minor matter whether there is also a parallel inventory of new objects or institutions, (ii) The prompt for these new usages is not a newly adopted referent, since new usages arise for familiar referents too (and since one cannot always tell the difference between new and old referents anyway), (iii) The prompt for the new usages is not a new linguistic expression, since we have seen that it is not the form of new expressions that is being copied. The simplest generalization, therefore, is that loan translations are motivated by the adoption of what impacted speakers perceive as new conceptualizations of the donor group. 2 3 (iv) One cannot simply assume, as do many proponents of loan translation, that the mere expression of a foreign conceptualization is a case of language contact. 24 In the absence of systemic evidence to the contrary, such usages are better regarded as instances of synchronic creativity.

32

Ricardo Otheguy

3. The analysis of modeling in U. S. Spanish The criticism of the construct loan translation is not intended to minimize the impact of English on the Spanish of the United States. The point, rather, is that finding the specific areas in which such impact has taken place is made impossible by reliance on the construct. The kind of linguistically expressed conceptual modeling embodied in El dia de dar gracias and mäquina de contestar is rampant among U. S. Hispanics. But if one is to distinguish linguistic creativity pressed in the service of cultural imitation from true linguistic imitation, a case-by-case approach must be used. Under analysis, utterances that would first pass unnoticed may turn out to be clear manifestations of linguistic mimesis, while highly stigmatized loan translations will turn out to be innovative exploitations of an untouched traditional system. For instance, most analysis of Spanish agree that, in general Spanish, calificar can be transitive in both verbal and adjectival usages, as in numbers (3) and (4): (3)

Califico a Carlos de incapaz de desempenar ese cargo. 'He judged Carlos incapable of occupying that position.'

(4)

Carlos no es un hombre calificado para ese cargo. 'Carlos is not a man who qualifies for that position.'

But most analyses point out (e. g. Torrents dels Prats 1976) that speakers of general Spanish do not use calificar intransitively, as in (5), which is, however, quite common in the Spanish of the U. S.: (5)

Carlos no califica para ese cargo. 'Carlos does not qualify for that position.'

For any theory that includes the distinction between transitive and intransitive verbs as part of the language system, number (5), which is seldom remarked upon, would appear to be a clear case of language contact within the area of modeling, like the saber in the first example given above. 25 In contrast to (5), it can plausibly be argued that there is no case of language contact in (6), or in (2), even though llamar para aträs, modeled on English call back, is regarded by many as a clear instance of loan translation.

A reconsideration

(6)

of the notion of loan translation

33

Le dije a Carlos que me llamara para aträs. Ί told Carlos to call me back.'

It must be noted, first, that llamar para aträs, like so many other spurious loan translations, is not structurally parallel with the presumed model. Depending on the analysis, the Spanish phrase consists of a preposition followed by an adverb or, if no such combination is to be allowed, of a complex adverbial. The simplex adverb back in English is clearly not structurally parallel under either analysis. And as in all the previous examples, the structure of llamar para aträs is quite familiar in the noncontact dialects. Verbs followed by adverbial prepositional phrases are commonplace in general Spanish, as is the specific sequence of a transitive verb followed by para aträs. (7)

No me quiero meter por esa calle, porque mäs adelante te encuentras que estän en obras, y tienes que volver para aträs. Ί don't want to go down that street, because further up you find that there are men working and you have to turn around.'

In addition to differing syntactically from the presumed English model, llamar para aträs and other phrases with para aträs differ also in the semantic content of their components. Analysts of different theoretical persuasions agree that para should be analyzed in opposition to the preposition por, and that its meaning involves the notion of spatial or temporal movement toward a goal. For example, in the European Structuralist analysis by Lopez (1970: 138), the meaning of para is "future union," and is represented by a diagram in which, from a point labeled 'V', an arrow moves toward a target: V —> | . And in the analysis by Lunn (1988: 169), performed under the theory of Cognitive Grammar, the meaning of para is similarly said to involve a trajector moving toward a landmark. To be sure, these notions are not entirely equivalent in the different schools. (The target for the Structuralist Lopez, for example, is not theoretically equivalent to Lunn's landmark, which forms part of the figure-ground approach to meaning within Cognitive Grammar). Still, there is a striking similarity between the analyses and considerable agreement on the examples chosen to illustrate them. Lunn exemplifies her landmark with Venezuela in (8) while Lopez exemplifies her target or point of future union with tu casa in (9). (8)

Mi primo se fue para Venezuela. 'My cousin left for Venezuela.'

34 (9)

Ricardo Otheguy Voy par α tu casa. 'I'm going to your house.'

All these analysts take cognizance of the well-known fact that, in many languages, items that once in their etymological origins may have described only physical movement or location have long since come to be used to describe movement and location of a more abstract nature, such as a temporal one (Ullmann 1951). Examples of this in English are words like within and behind, in phrases such as He'll get it done within the hour and They put that problem behind them long ago. For Spanish, analysts routinely offer examples of para in which the target or goal has to do with time rather than with space. Example (10) is from Lunn: (10)

Nos habiamos citado para las tres. 'We had made an appointment for three o'clock.'

This direct linkage to a target by means of the meaning of para, in our case the target aträs, suggests that llamar para aträs is structured semantically in a very different way from call back, where no connection between a trajector and a goal, and no directional or movement sense of any kind, appears to be operative. It turns out, then, that quite independently of any English influence, para can have aträs as a goal or point of future union, as in (7); that this landmark can be a point in space, as in (8) and (9); and that it can also be a point in time, as in (10). Thus neither para nor aträs has undergone any semantic extension. Each expresses in llamar para aträs what it expresses everywhere else: in the case of para, the notion of movement toward a goal; in the case of aträs, the notion of a point situated behind the speaker in either time or space. The innovation in llamar para aträs is thus neither in lexis nor in grammar, and resides outside of language, in the new conceptualization of repetition in terms of a physical "return" metaphor. What distinguishes U. S. Spanish speakers from speakers in non-contract areas in this case is that the U. S. ones have chosen to express the repetition of an action (the calling again in response to a first call) through the metaphor of spatial revisiting, of returning physically to the original point of the call. Thus llamar para aträs is no different from mäquina de contestar, whose only novelty resided outside of language, in the decision made by speakers of Spanish, in the manner of speakers of English, to conceptualize the appliance as a machine that answers. To place the new conceptualization of llamar para aträs in its proper context, it is instructive to look at varieties of general Spanish spoken

A reconsideration

of the notion of loan translation

35

outside of, and very distant from, the United States, where similar metaphorical extensions from space to time are at work, and where a similar connection has been established between "return" and "repetition." In the River Plate dialects of Spanish, as well as in many other areas of South America, temporal returns and repetitions are commonly expressed using the phrase de vuelta, which relies on a preposition plus the word vuelta, commonly defined as "turn, turning" (de Gamez 1973) or "movement around a point" (Real Academia Espanola 1984). This movement can range from the very concrete ir de vuelta, 'to go back', to the more abstract, entregar el trabajo de vuelta, 'hand the job/paper in again', or, more to our point, llamar de vuelta, 'to call back'. (Serrana Caviglia, Mirta Groppi, and Marisa Malcuori, Departamento de Lingüistica, Universidad de la Repüblica, Montevideo, personal communication, 1988). In the analysis in terms of prototypes that is favored by the more recent proponents of the old idea of space-time metaphors (Lakoff— Johnson 1980, Lakoff 1987), both llamar de vuelta and llamar para aträs would be explained by seeing repetition of an action in time as a metaphorical extension of repeated physical passage through a point in space. In this account, the prototypical meaning of traditional de vuelta would be a physical turning around and looking backwards, while the prototypical meaning of para aträs would be a physical movement toward a landmark in space. Both would be extended to time, so that either by facing a point left behind in time, as in River Plate de vuelta, or by reaching it after covering a trajectory, as in United States para aträs, the current act of calling would be related to a prior act of calling. What makes River Plate Spanish llamar de vuelta a repetition is exactly what makes U. S. Spanish llamar para aträs a repetition, namely that the second call is being conceived in terms of a first call that is located, metaphorically, in a prior space-time spot. To be sure, as Lipski (1985) and other analysts have insisted, the English model call back very likely did play some role in the creation of llamar para aträs. But it was not a linguistic or structural role. Rather, it was a role similar to that played by answering machine in the creation of mäquina de contestar. Speakers of Spanish in the U. S. could very well have gotten the idea from speakers of English that the concept of "behindness" in space could be applied metaphorically to the temporal notion of repetition. But they then deployed the resources of their language in a manner that, to repeat, is syntactically and semantically different from that of English, and that, furthermore, appears to involve no alteration of any systemic area of Spanish lexis or grammar.

36

Ricardo Otheguy

4. Summary The problem of contact-induced modeling has been framed in terms of the need to distinguish, (a) those utterances by speakers of impacted varieties that express simple reconceptualizations, that is, that reflect a cultural or conceptual break with the ancestral society, from (b) those that embody a true linguistic break with the ancestral dialects. It has been proposed that this distinction between linguistic and cultural modeling — between contact and creativity — can only be drawn on a caseby-case basis. The case-by-case approach is particularly important because 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. By proliferating the examples of loan translation, this approach tends inevitably toward overstating, by a considerable margin, the actual amount of language contact. It has been shown that some so-called loan translations do not display any constructional features that can be abscribed to the influencing language, and that their only diverging items are the semantic extensions of component words. It has been shown, further, that other so-called loan translations are entirely traditional from a linguistic point of view, containing no modeled elements whatever, not even semantic extensions. In dealing with modeling phenomena, a sharp distinction has been drawn between cultural or conceptual modeling, on the one hand, and linguistic modeling on the other. Under examination, loan translations seem to be cases of the former, while true linguistic modeling only takes the form of semantic extensions. Though it is quite likely that semantic extensions (as well as loanwords and other contact phenomena) are found in greater proportions in phrases expressing conceptual or cultural modeling, our point has been that many such phrases remain free of such interference and are unhybridized from a linguistic point of view. In previous works on U. S. Spanish, when dealing with semantic extensions of the type illustrated by saber and calificar we have essayed the term calqueword, precisely to stress that the phenomenon is found at a specific level of linguistic analysis (Otheguy — Garcia 1988; Otheguy — Garcia — Fernandez 1989). For items like mäquina de contestar and llamar para aträs, it would be useful to always speak of cultural or conceptual

A reconsideration of the notion of loan translation

37

modeling. 26 This will remind us that these ^tems represent, not diachronic developments in the lexicon or grammar of the Spanish language, but synchronic adaptations to the largely non-Hispanic cultural and conceptual environment of U. S. Latinos. In this view, contact settings illustrate not only the diachronic instability of linguistic systems, but also their synchronic stability and expressive adaptability to ever changing cultural and conceptual needs.

Notes

1. Comments on earlier versions of this paper from Edward Bendix, Joseph Davis, William Diver, Susan Gal, Alan Huffman, Robert Kirsner, Raymond Mougeon, Wallis Reid and William Stewart were extremely useful and are gratefully acknowledged. All errors, of course, remain mine. 2. Hope (1971: 639 n) suggests that the proliferating terminology surrounding this phenomenon is mostly a feature of North American scholarship. But these terms do not seem to have sharply differentiated definitions anywhere. For a summary of their use by scholars writing in English, French, German, Italian, and Spanish, see Lazaro Carreter (1974: 77). 3. Changes that take place in contact settings due to internal developments within the impacted variety and without direct intervention of the donor language are beyond the scope of this presentation. For discussion of this topic with regard to U. S. Spanish, see Klein-Andreu (1985) and SilvaCorvalan (1990). 4. There is nearly complete agreement in the use of "loan translation" as synonymous with "caique" (e. g. Jensen 1912: 116, Lehiste 1988: 20, Hudson 1980: 59). But scholars disagree in many other aspects of terminology. Following the use originally found in Whitney (1881: 18) and Jensen (1912: 116), loan translation/caique is used only for lexical phenomena in, for example, Casagrande (1954) and Campbell — Kaufman — Smith-Stark (1986). But following Haugen (1950 [1972: 85]), the term is used also for morphosyntactic contact in, for example, Dillard (1975), Hudson (1980), and Montes Giraldo (1985). Furthermore, only some authors adhere to a distinction between "semantic extension" or "loanshift" for modeling in single words and "loan translation/caique" for modeling in phrases (Henzl 1981, Lehiste 1988, Weinreich 1953). Appel - Muysken (1987: 165) regard both "loan translation" and "caique" as interchangeable with "loanshift"; and so do Craddock (1976: 46, 1981: 208) and Daiuta (1984: 70). And Haugen (1950 [1972: 85]), who coined the term loanshift, considered it synonymous not only with "loan

38

5.

6.

7.

8.

9.

10.

11.

Ricardo Otheguy translation" but with "semantic loan" as well. For ease of presentation, henceforth we will dispense with the doublet "loan translation/caique" and will use "loan translation" alone, but will continue the practice of applying it only to phrasal phenomena, reserving "semantic extension" for modeling in single items. For analyses of dar para aträs as a loan translation, see for example Varela (1974). For thorough documentation on the use of this form in many varieties of English-influenced Spanish, see Lipski (1985). The partial conceptual uniqueness of linguistic communities is discussed in Sapir (1929 [1949: 162]), Culler (1976: 114), Hudson (1980: 85), Hymes (1967: 16) and Ullmann (1951 [1957: 217]). That contact produces conceptual convergence, which in turn can give rise to linguistic convergence, is maintained by most students of contact, such as for example Appel - Muysken (1987: 165), Bloomfield (1933: 445 ff), Casagrande (1954: 140), Haugen (1938 [1972: 22ff|), Hockett (1973: 119ff), Weinreich (1953 [1974: 59 ff, 91 ff|), Pratt (1980: 62), and Whitney (1881: 10 ff). The distinction between linguistic and cultural or conceptual convergence is drawn with clarity in, for example, the work on contact between different Northwestern California languages by Bright — Bright (1965: 251). While the discussion here is primarily focused on U. S. Spanish, the critique of loan translation is applicable to any environment where the construct is used to describe systemic change. This includes not only contact settings, where the roles of donor and recipient language are fairly clearly defined, but also the more complex situations of creole and post-creole communities. For the use of this construct in such settings, see for example Holm (1988: 86 ff). Weinreich (1953) was both a continuation of such prior analyses as Bloomfield (1933: 455 ff), Haugen (1938 [1972: 19 ff], 1950 [1972: 85 ff]) and Ullmann (1951 [1957: 162 ff]), and the precursor of subsequent ones such as Craddock (1981: 202), Henzl (1981: 310), Lehiste (1988: 20), and Milan (1982: 197). Exceptions to this treatment of the distinction between semantic extension and loan translation can be found in Hope (1971: 637) and Ferguson — Heath (1981: 527), whose classificatory scheme recognizes instances of modeling only in individual lexical items, lacking, in effect, any reference to loan translations. Weinreich may have been hasty in attributing escalier de feu to English influence. Raymond Mougeon (personal communication) points out that Le dictionnaire Robert gives the use of feu to refer to an accidental fire as preceding the rise of the word incendie. Similarly, the Spanish Academy's Diccionario de la lengua espanola lists several ecclesiastical senses of ministro that are common in Latin America now, suggesting that this example too may not have been, even in Weinreich's time, an instance of English influence. But in order to focus on Weinreich's analysis, it will be necessary to accept

A reconsideration of the notion of loan translation

12.

13.

14.

15.

16.

17.

18.

39

for the moment his contention (1953 [1974: 47 — 51]) that these usages were peculiar to the contact dialects and were not found in general French or Spanish. Synchronic creativity is called productivity by Lyons (1977: 77) in order to stress its more mechanical aspects, as also found in Chomsky (1965: 15 and 198 n) or in Weinreich (1969 [1980: 260]). It is called versatility by Garcia (1975: 41) in order to stress its more conceptual or instrumental features, as also found in Bolinger (1965: 568 ff) or in Diver (1980: 5). And it is called openness by Hockett (1973: 108) to stress the role of analogy. By whatever name, creativity is widely recognized as a systemically non-invasive factor. The fact of cultural without linguistic hybridization is even more patent in those speakers who refer to Thanksgiving Day as Dia de accion de gracias, distancing themselves even more from the North American model. Given that answering machines are found in traditional Spanish-speaking areas, the referent here is not as new or as uniquely Anglo-American as Thanksgiving Day; and given the availability of a traditional formulation {contestador automätico), U. S. Spanish usage competes in this case with traditional usage in a way that it did not in the case of Thanksgiving Day. Of course, for speakers of U. S. Spanish whose experience of answering machines is limited to the United States the referent will be new, and the contact usage will not compete with the general Spanish one. Relevant here is the observation in Haugen (1938) that apparently identical referents (such as cellars and rivers in Wisconsin versus cellars and rivers in Norway) may feel different to contact speakers simply by virtue of their different location. This means that in many instances it is impossible to tell whether referents in the new society (answering machines in the United States) are in fact the same as those in the ancestral one (answering machines in Latin America). Speakers of English often name new appliances by means of derived lexicalizations, as in computer. The Spanish pattern is also at times abandoned in favor of derivation, as in nevera 'refrigerator'. The point is that the patterns under discussion are widespread in each language yet different form each other, not that they hold exclusive sway in the naming of new appliances. Significantly, the names for an older generation of appliances were often approached by speakers of Spanish in a manner similar to that used for the answering machine by contact speakers. So, for example, in many traditional varieties of Spanish the conventional formulation for the lawnmover is mäquina de cortar hierba and for the typewriter mäquina de escribir. The points made here about contact settings can be applied as well to creole or post-creole environments. For example, the phrases that translate into English as "greedy" in several African-influenced varieties {big eye in Bahamian Creole, gwo ze in Haitian Creole, olho grande in Brazilian Portuguese) do not necessarily constitute contact features induced in these languages by

40

19.

20.

21.

22.

23.

24.

Ricardo Otheguy Twi or Ibo, as seems to be suggested by Holm (1988: 86 ff). These Bahamian, Haitian, and Portugese phrases may be instances of contact if they constitute new lexical units, or if the new usage cannot be made to fit into the existing meaning of the component items. But they may be simply new usages of existing units that adhere to each of the languages' lexical and grammatical principles; all that may be remarkable in these items is their reflecting a conceptualization of a referent (greed) in the manner of speakers of Twi and Ibo (by calling it the big eye). Only detailed analysis can decide, and in its absence there is no reason to assume hybridization. Under some interpretations, the maximalist view of language adopted by cognitive grammarians such as Langacker (1988: 150) may constitute a significant challenge to this assumption. Even assuming that the encoded meanings of words include denotation and potentiality for appearing only in certain collocations (Lyons 1977: 237 ff, 613 ff), loan translations such as the ones examined here would show no structural or semantic convergence with English. The denotations and collocations of the individual words are all familiar in traditional Spanish. Only their combined applicability to the expression of a new conceptualization is new. To the extent that these concepts were expressed by compounds, and to the extent that these compounds have to be regarded as new dictionary units, then these Meso-American items could be said to involve contact-induced changes in the lexicons of these languages. But this question is beyond the scope of the present discussion. One might argue that what diffused was more than a set of conceptualizations; it was a set of conceptualizations of a set of referents. Alternatively, one might argue that what diffused was a selection of words with similar meanings to use for the same referent. In either case it would still be non-linguistic diffusion. Selection of items for the purpose of making a reference is not part of language, and diffusion of such patterns of selection is not linguistic diffusion. The combinations of words formed to name a referent respond to how the referent is conceptualized, which is a fact of culture, not of language. Only under a very broad notion of Pragmatics could such selections start to be examined as possible members of some peripheral area of the use of the language system. Their inclusion under structural change would be highly inaccurate in any event. This position is thus in sharp contrast to that assumed by Nash (1980) in discussing Spanish-influenced English in Puerto Rico, in which reconceptualization is seen as something that happens in the language, that is, as itself a form of language contact. In cases like mäquina de contestar, a possible area where linguistic contact may be detected involves lexical loss. If as a result of using this strictly Spanish phrase these speakers lose, or fail to acquire, the use of contestador,

A reconsideration of the notion of loan translation

41

for example, then it may be argued that linguistic contact is at work. But it would be contact of a different kind, falling under the category of simplification (Klein-Andreu 1985, Silva-Corvalan 1990), and not of loan translation or modeling. 25. The issue whether this kind of modeling should be regarded as a semantic, or as a grammatical or syntactic, extension is beyond the scope of this paper. Scholars who in such cases prefer to see a syntactic change may also regard cases like calificar as instances of transferring (of a syntactic feature) rather than of modeling. This issues too is beyond the scope of the present work. 26. The term "caique phrase" which we ourselves have used in the past to refer to these items itself partakes of some of the flaws being pointed out now, and does not represent a clear enough separation from the approaches being criticized here.

References

Alston, William 1970 "Theories of meaning", in: A. Lehrer — Κ. Lehrer (eds.) Theory of meaning. Englewood Cliffs: Prentice-Hall. Appel, Rene —Pieter Muysken 1987 Language contact and bilingualism. London: Edward Arnold. Bloomfield, Leonard 1933 Language. New York: Holt. Bolinger, Dwight 1965 The atomization of meaning, Language 41: 555 — 573. Bright, Jane —William Bright 1965 "Semantic structures in Northwestern California and the SapirWhorf hypothesis", American Anthropologist 67: 249 — 258. Campbell, Lyle — Terrence Kaufman—Thomas Smith-Stark 1986 Meso-America as a linguistic area. Language 62: 530 — 571. Casagrande, Joseph 1954 "Comanche linguistic acculturation", I and II, International Journal of American Linguistics 20: 140-151 and 20: 217-237. Chomsky, Noam 1965 Aspects of the theory of syntax. Cambridge: MIT. Craddock, Jerry 1976 "Lexical analysis of Southwest Spanish", in: Donald Bowen —Jacob Ornstein (eds.) Studies in Southwest Spanish. Rowley, Mass: Newbury House.

42

Ricardo Otheguy 1981

"New World Spanish", in: Charles Ferguson - Shirley Brice Heath (eds.), New York: Cambridge. Culler, Jonathan 1976 Ferdinand de Saussure. New York: Penguin. Daiuta, Amy 1984 "Remarks on calquing", CUNYForum 10: 7 0 - 9 0 . (City University of New York.) de Gamez, Tana 1973 Simon and Schuster's International Dictionary English/Spanish, Spanish/English. New York: Simon and Schuster. Dillard, J. L. 1975 All-American English: A history of the English language in America. New York. Vintage Books. Diver, William 1980 "The ecology of language", Columbia University Working Papers in Linguistics 5: 1—6. Dvorak, Trisha 1983 "Subject-object reversals in the use of gustar among New York Hispanics", in: Lucia Elias-Olivares (ed.), Spanish in the U. S. setting: Beyond the Southwest. Rosslyn, VA: National Clearinghouse for Bilingual Education. Espinosa, Aurelio 1975 "Problemas lexicograficos del espanol del Sudoeste", in: Eduardo Hernandez-Chavez — Andrew Cohen — Anthony Beltramo (eds.), El lenguaje de los chicanos. Arlington, VA: Center for Applied Linguistics. Ferguson, Charles — Shirley Brice Heath (eds.) 1981 Language in the USA. New York: Cambridge. Frege, Gottlob 1892 "On sense and reference". (Translation of "Über Sinn und Bedeutung".) In: P. Geach —Μ. Black (eds.), Translation form the philosophical writings of Gottlob Frege. Oxford: Blackwell. [1974] [Reprinted in Readings in semantics, ed. by F. Zabeeh. University of Illinois Press.] Gal, Susan 1989 "Lexical innovation and loss: The use and value of restricted Hungarian", in: Nancy Dorian (ed.), Investigating obsolescence: Studies in language contraction and death. Cambridge University Press. Garcia, Erica 1975 The role of theory in linguistic analysis: The Spanish pronoun system. Amsterdam: North-Holland. Haugen, Einar 1938 "Language and immigration", Norwegian-American Studies and Records 10: 1 - 4 3 .

A reconsideration of the notion of loan translation

43

1950 "The analysis of linguistic borrowing", Language 26: 210 — 231. [1972] [Reprinted in The ecology of language. Stanford University Press.] Henzl, Vera 1981 "Slavic languages in the new environment", in: Charles Ferguson — Shirley Brice Heath (eds.), Language in the USA. New York: Cambridge. Hill, Jane-Kenneth Hill 1980 "Mixed grammar, purist grammar, and language attitudes", Language in Society 9: 321—348. Hockett, Charles 1973 Man's place in nature. New York: McGraw-Hill. Hoijer, Henry 1948 "Linguistic and cultural change". Language 24: 335 — 346. Holm, John 1988 Pidgins and Creoles. New York: Cambridge. Hope, Τ. E. 1971 Lexical borrowing in the Romance languages. New York: New York University Press. Hudson, R. A. 1980 Sociolinguistic. New York: Cambridge. Hymes, Dell 1967 "Linguistic problems in defining the concept of tribe", in: John Baugh—Joel Sherzer (eds.), Language in use: Readings in sociolinguistics. Englewood Cliffs, NJ: Prentice-Hall. Jensen, Kristian Sandfeld 1912 "Notes sur les caiques linguistiques", in: Festschrift Vilhelm Thomsen. Leipzig, 166—173. Klein-Andreu, Flora 1985 "La cuestion del anglicismo: Apriorismos y metodos", Thesaurus 40: 533-548. Lakoff, George 1987 Women, fire, and dangerous things: What categories reveal about the mind. Chicago: University of Chicago Press. Lakoff, George — Mark Johnson 1980 Metaphors we live by. Chicago: University of Chicago Press. Langacker, Ronald 1988 "Autonomy, agreement, and cognitive grammar", in: Diane Brentani — Gary Larson —Lynn McLaod (eds.), Parassession on agreement in grammatical theory. Chicago Linguistic Society 24: 147 — 180.

Lazaro Carreter, Fernando 1974 Diccionario de tirminos filologicos. Madrid: Editorial Gredos. Lehiste, Ilse 1988 Lectures on language contact. Cambridge: MIT Press.

44

Ricardo Otheguy

Lindenfeld, Jacqueline 1971 "Semantic categorization as a deterrent to grammatical borrowing", International Journal of American Linguistics 37: 6 — 14. Lipski, John 1985 "The construction Pa(ra) Atras among Spanish-English bilinguals: Parallel structures and universal patterns", Revista Interamericana 15: 91-102. Lopez, Maria Luisa 1970 Problemas y metodos en el anälisis de las preposiciones. Madrid: Editorial Gredos. Lunn, Patricia 1988 "How por and para mean", in: John Staczek (ed.), On Spanish, Portuguese, and Catalan linguistics. Washington: Georgetown University Press. Lyons, John 1977 Semantics. Cambridge. Milan, William 1982 "Spanish in the inner city", in: Joshua Fishman —Gary Keller (eds.), Bilingual education for Hispanic students in the United States. New York: Teachers College Press. Montes Giraldo, Jose Joaquin 1973 "La politica y las 'politicas'", Thesaurus 28: 86-105. 1985 "Calcos recientes del ingles en espanol", Thesaurus 40: 17 — 49. Narcisse-Eutrope, Dionne 1974 Le parier populaire de Canadiens Frangais. (3eme section.) Quebec: Laval. Nash, Rose 1980 "Reconceptualization as a form of language contact", in: William McCormack — Herbert Izzo (eds.), The Sixth LACUSforum. Columbia, SC: Hornbeam Press. Oksaar, Els 1979 "Models of competence in bilingual interaction", in: William Mackey—Jacob Ornstein (eds.), Sociolinguistic studies in language contact: Methods and cases, The Hague: Mouton. Otheguy, Ricardo — Ofelia Garcia 1988 "Diffusion of lexical innovations in the Spanish of Cuban Americans", in: Jacob Ornstein-Galicia — Dennis Bixler-Marquez (eds.), Research issues and problems in United States Spanish. Brownsville, Texas: Pan American University. Otheguy, Ricardo — Ofelia Garcia — Mariela Fernandez 1989 "Transferring, switching, and modeling in West New York Spanish: An intergenerational study", International Journal of the Sociology of Language 79: 41 — 52.

A reconsideration of the notion of loan translation

45

Pratt, Chris 1980 El anglicismo en el espanol peninsular contemporäneo. Madrid: Editorial Gredos. Real Academia Espanola 1984 Diccionario de la lengua espanola. Madrid: Espasa-Calpe. Sapir, Edward 1929 "The status of linguistics as a science". Language 5: 207 — 214. [1949] [Reprinted in: David G. Mandelbaum (ed.) Selected writings of Edward Sapir. Berkeley: University of California Press.] Saussure, Ferdinand de 1916 Cours de linguistique generale. [1972] Cours de linguistique generale. Paris: Payot. Silva-Corvalan, Carmen 1990 Current issues in studies in language contact. Hispania 73: 162 — 177. Thomason, Sarah—Terrence Kaufman 1988 Language contact, creolization, and genetic linguistics. Berkeley: University of California Press. Torrents dels Prats, Alfonso 1976 Diccionario de dificultades del ingles. Barcelona: Editorial Juventud. Turenne, Augustin 1962 Petit dictionnaire du 'Joual' en Francais. Ottawa: Les Editions de l'Homme. Ullmann, Stephen 1951 The principles of semantics. New York: Philosophical Library. [1957] The principles of semantics. New York: Philosophical Library. Second Edition. Varela, Beatriz 1974 "La influencia del ingles en los cubanos de Miami y Nueva Orleans". Espanol Actual 26: 16 — 25. Weinreich, Uriel 1953 Languages in contact. New York: Publications of the Linguistic Circle of New York. [1974] [Languages in contact. The Hague: Mouton], 1969 "Problems in the analysis of idioms", in: Joan Puhvel (ed.), Substance and structure of language. Berkeley: University of California Press. [1980] [Reprinted in William Labov — Beatrice Weinreich (eds.), Uriel Weinreich on Semantics. Philadelphia: University of Pennsylvania Press.] Whitney, W. D. 1881 "On mixture in language", Transactions of the American Philological Association 12: 5 — 26.

The dialectics of Spanish language loyalty and maintenance on the U.S.-Mexico border: A two-generation study* Margarita

Hidalgo

As a topic of linguistic and sociolinguistic inquiry, the diversity of the Mexican-American population of the Southwest (SW) does not appear to be immaterial, for it has not been relegated to playing a marginal role in the intricate drama of Mexicans' adaptation and pseudo-assimilation to the American mainstream. On the contrary, since the late 1960s and early 1970s, observers of Mexican-American linguistic reality have produced an impressive bulk of research dealing with an interrelated series of issues that range from mere descriptions of language varieties to political and legislative implications for bilingual schooling. A subtopic of obvious concern in Mexican-American sociolinguistics is that of language maintenance, loyalty and shift. 1 The limited studies available to date, report findings that, although not always comparable, show convergent trends and patterns, especially with respect to language maintenance. In her review of literature on the topic, Floyd (1985) discusses the various social and other non-linguistic factors impinging on maintenance and shift. "The most frequently considered factors have been generation and age of speaker, with less attention paid to factors such as speakers' language use at home, childhood residence, education and sex" (Floyd 1985: 21). Most studies report greater use of Spanish for first-generation speakers, use that gradually decreases with successive generations of individuals, and in some cases, as early as in second-generation families. 2 As important as the factor of generation is the factor of age, which is strongly linked to mother tongue retention, that is, the older the individual, the greater the degree of Spanish language use seems to be. Moreover, some studies report that speakers at the two age extremes (i. e., oldest adults and youngest preschool children) favor monolingual Spanish use, while school-age children and younger adults engage in bilingualism and favor less use of Spanish. Greater maintenance of Spanish among older speakers (and consequently greater use of English among younger speakers) appears to be a constant trend in the SW, regardless of the method employed to draw out the data:

48

Margarita

Hidalgo

mother-tongue claiming, self-reported proficiency in Spanish, listening comprehension, or self-report of language use in the home. 3 Floyd (1985: 22) suggests that the differences between patterns of language use according to age and generation indicate that among subjects in the areas studied, language shift away from the ancestral language and in the direction of the dominant language, English, is in process. "While the shift may be progressing at a faster pace in the large urban areas such as Austin and Los Angeles, the literature suggests that even in more remote areas like Arroyo Seco, New Mexico, language shift is in evidence" (Floyd 1985: 22). Use of Spanish at home, formal study of the language, and childhood residence have been studied as extensively as age and generation, but the research at hand shows that Spanish use in the household (between adults and children and between children themselves) is a moderate predictor of the language to be used by future generation speakers. 4 Also, selfreported literate skills in Spanish appear to be significantly associated with formal instruction. Contradictory reports on area of childhood residence (urban vs. rural) indicate that dwellers of a large city may preserve Spanish as well as those coming from the countryside. 5 Another demographic factor influencing language use is gender. Research indicates that teenage girls tend to identify more closely with Anglo-American culture and are therefore more accepting to the school language, English, while boys choose to speak more Spanish to express rebelliousness against authority. It is also believed that various psychosocial mechanisms — perceived value orientation, role expectations, attitudes towards learning and reward identification — come into play in the language choices of males and females. However, Mexican-American women are less likely than men to experience conflicts emanating from mainstream society and are more likely to conform to outgroup values. Also, in a community like Rio Grand City, Texas, where 95% of the women are employed as professionals or as white-collar workers, adult females feel the pressures to shift to the English language. 6

1. Are U. S. Mexicans like all the others? The question that reseachers have been raising since the mid-1960s is whether Mexican-Americans are apt to follow the path of linguistic assimilation trod by earlier waves of immigrants. By the same token, the

The dialectics of Spanish language loyalty

49

issue of loyalty or disloyalty to Spanish is consistently linked to a greater ability or proficiency in the English language and to a higher socioeconomic status. In this vein, Macias (1985) offers several explanations that oppose the predictions of the demise of Spanish in the U. S. Macias sees a pattern of Spanish language increase over time tantamount to an increase in the numbers of English monolinguals and bilinguals. These trends result from the interaction of retention factors such as the large size of the MexicanAmerican community, its relative linguistic homogeneity, continued inmigration from Mexico, relative social isolation, Spanish language media use, and intergenerational language maintenance. Simultaneously, patterns of bilingualism emerge in the SW, since 54% of all MexicanAmericans born in Mexico become bilingual and 63% of all the Mexicanorigin population born in the U. S. is bilingual. Likewise, attainment of higher education correlates positively and strongly with English dominance, bilingualism, and higher income. Hence, shift to English or Spanish language loss are questionable because monolingualism for some and bilingualism for others co-exist. While higher income, better schooling, and English proficiency foster bilingualism for upwardly mobile, lifetime residents of the SW, incessant inmigration, relative marginalization and distance from the mainstream, as well as ethnolinguistic homogeneity, enhance language maintenance. Moreover, the discussions by Sanchez (1983) and Sole (1985) confirm Macias' hypothesis. Whereas Sanchez's Marxist analysis indicates that income and consumption of capitalist goods and values determine a shift towards English dominance, using the 1980 Census Bureau data and a sample of 164 Mexican-American students at Austin, Sole demonstrates that English is more actively used by those individuals whose fathers have achieved professional status. Nevertheless, both authors firmly state that Mexican-Americans in the lower socioeconomic strata are more Spanish-retentive. In another vein, the loose connection between attitudes towards mother-tongue and mother-tongue use for communities in which Spanish and other ethnic languages are spoken is tackled by Fishman (1985) in reference to the 1960 and 1970 census data. In his analysis, Fishman equates mother-tongue claiming with attitudes towards ethnic languages (1985: 108) and points out that most of those who claim non-English mother-tongues no longer currently use them. Some indications of the extent of this attrition are gleaned from a comparison among the "big six" non-English mother-tongues (i. e. French, German, Italian, Polish,

50

Margarita

Hidalgo

Spanish and Yiddish) (1985: 144). He then proceeds to give attrition rates for each language among foreign and native born individuals, showing that Spanish has a low attrition rate of 19% among the former and 36% among the latter. Fishman's data indicate that at least some U. S. Spanish speakers behave in ways similar to speakers of other ethnic languages.

2. Are Border Mexicans like all the others from the Southwest? Two pieces of work on language maintenance and shift in border areas clearly contradict the general patterns of the SW. In the Rio Grande Valley of Texas, Amastae (1982) observes that even though MexicanAmerican speakers shift from Spanish to English outside the home — a displacement normally associated with upward mobility — a complementary shift from English to Spanish takes place once the desired socioeconomic status is attained. In like manner, Aguirre's (1982) study among students in a California border town reveals that there is a reported balance between the use of the two languages for various social events and persons; Aguirre discounts the clear existence of a diglossic element in the sociolinguistic behavior of border residents. In the same border area where Amastae (1982) carried out his study, Mejias and Anderson (1984/1987 and 1988) advanced the following hypotheses: a) Attitudes towards language maintenance are significantly more positive among older individuals and first and second-generation Mexican-Americans than among younger individuals and third and fourth-generation bilinguals, but border people generally prefer maintenance of Spanish rather than shift to English, b) Instrumental attitudes toward Spanish are not as strong as sentimental and communicative attitudes, c) When instrumental attitudes are present, men express them more openly than women, d) Older generations of Mexican-Americans attach symbolic values to Spanish, whereas recent arrivals do not. The uniqueness of the border has also been addressed by Hidalgo (1983, 1984, 1986, 1988), who proposes a strong correlation between language attitudes and language use and the independence of language behavior of demographic factors. 7 From a microsociolinguistic perspective and taking into consideration the unusual setting of the border, I

The dialectics of Spanish language loyalty

51

have looked into demographic and linguistic dimensions influencing language maintenance in Chula Vista, a border city in southern California, and have hypothesized that the former factors cannot entirely predict language maintenance. On the contrary, I have assumed that loyalty toward the ancestral language is indeed linked to language maintenance (that is, language attitudes are related to language use), because the U. S.Mexico border is a unique southwestern setting, where neither English nor Spanish completely prevails in the minds or the daily lives of its residents. Border inhabitants (from both sides of the border) oftentimes verbalize their loyalty to the ancestral language and even feel the need to justify it because it has become a central component of their identify. Whereas on the Mexican side of the border, loyalty to Spanish stands out as an expected form of defensive behavior in virtually all individuals (Hidalgo 1983, 1986), on the U. S. side, it may come into conflict with the American value system. The existence of loyalty (as another form of attitude) has thus been inferred from observable behavior and may be useful in the prediction of actual use and maintenance of Spanish in the community.

The setting With more than 100,000 inhabitants, Chula Vista is the second largest and second fastest growing city of San Diego County. It lies seven miles north of the Mexican border and seven miles south of downtown San Diego. It has recently renovated its town center in order to recreate the village atmosphere it possessed in the 1900's when it was the site of a major famer's market. Today the economy of Chula Vista is based on a mixture of diversified activities such as construction, manufacturing, transportation, communication, wholesale trade, retail trade, finance, insurance, real estate, services and government. Chula Vista belongs to the region commonly known as the South Bay area (which also includes Imperial Beach, San Ysidro, South San Diego, Otay Mesa and National City). It is estimated that over 30% of the people (legally) living in the South Bay are of Hispanic origin, whereas 40% are non-Hispanic white. In Chula Vista itself, the population of Hispanic origin is about 27%. The large Hispanic influence observed in the South Bay area in general and in Chula Vista in particular is due to the proximity to the International border with Mexico. Much of the impact and spinoff from cross-

52

Margarita

Hidalgo

border economic activities — retail trade, shopping, tourist travel, and coordinated manufacturing programs (assembly plants or twin plants) occurs in the South Bay. The economy of the area, thus, depends to an extent on the stability and well-being of the Mexican economy. The median household income in the South Bay area is about $ 18,000.00 per year (considerably below the San Diego median income). However, the median family income in Chula Vista is about $ 32,000.00. The survey conducted by me in the spring of 1988 intended to reach the subsection of Chula Vista known as Castle Park, located to the southeast of the city, where the population is predominantly of Mexican descent. It is assumed that Castle Park represents families of lower and lowermiddle incomes with its south sector reporting a disposable income of $25,000.00 or less and its northeast section reporting from $25,000.00 to $ 28,000.00 per year.

The school The survey, which collected sociodemographic and sociolinguistic data, was conducted in English among seven groups of Castle Park High School students enrolled in grades ninth through twelfth. Castle Park was chosen because of the conspicuous parental involvement in educational activities as well as for the uniqueness of its bilingual program. The students were selected according to three criteria: 1) density of Mexican-origin population in the classroom; 2) instructor's willingness to grant her/his period time to me, and 3) student's willingness to participate. Castle Park High School is located in one of the least affluent sectors of Chula Vista. The Mexican-origin population in the school reaches 54%; the school district has therefore implemented an effective bilingual program consisting of four levels of English as a Second Language plus the offerings of all requirements for graduation (i. e., science, mathematics) in both Spanish and English. In view of the proportion of Mexicanorigin students, the percentage of children with limited proficiency in the English language (LEPs) reaches 26%. The support given to MexicanAmerican and other Hispanic students is additionally reflected in the hiring of bilingual (certified and non-certified) personnel. Also, both the school administration and the students' parents appear to be involved in social (but de-ideologized) activities conducive to maintenance of language and ethnicity.8

The dialectics

of Spanish language loyalty

53

The children The survey, which lasted from 20 to 40 minutes, was conducted by me personally. In all cases I was able to monitor students and responded to questions raised by them. Seventy-six females and sixty males between 14 and 19 years of age participated in the study. Of the one-hundred and thirty-six Subjects (Ss), 20% reported as a place of birth an urban community in the Mexican border area and 59% an urban center in the U.S. border area. The remaining 21% were either born in the interior of Mexico, the interior of the U. S., or a foreign country other than Mexico. Ss had been residing in Chula Vista for a period of time ranging from five months to eighteen years. The mean number of years of residence was 8.7, whereas the mode was 4 and the median was 8. A substantial percentage of Ss had previously resided elsewhere in the Southwest (61%) or in a Mexican border city (25%). In their previous residence, they had lived an average of 7 years with a range of 6 months to 16 years. In their second previous residence — most frequently a southwestern or Mexican community — they had lived an average of about three years. All in all, about one-half of all the Ss in the sample appeared to have experienced geographic mobility before they settled in Chula Vista. About three fourths of the Ss' fathers (79%) and mothers (75%) were born in Mexico and had come to the U. S. at 19 or 20 years of age. The fathers' occupations were primarily unskilled or semi-skilled manual labor (76%), whereas the rest worked in services, managerial occupations, or professional activities (24%). About one-half of the Ss (45%) claimed that their fathers had completed high school and about one third (29%) reported that their fathers had completed 8 years of formal education. In addition, only about one fifth of all the Ss claimed their fathers had one or more years of college education. In contrast, more than 60% of the Ss' mothers were not gainfully employed but apparently had the same level of education as the fathers.

Contact with Mexico The Ss' contact with Mexico is substantial, since about two thirds of them claimed to cross the border from several times per week to several times per month. Students do not typically visit Mexico for business, work, shopping, or commercial transactions but sometimes they engage

54

Margarita

Hidalgo

in visits to relatives and friends; by and large, the most frequently reported activity carried out in Mexico is disco dancing. Contact with Mexico also takes place when relatives cross the border and visit from once a week to every summer or special occasions, the most frequent visit occurring weekly or biweekly. According to the children, about 67% of the Mexican relatives visiting Chula Vista normally stay one day or one weekend. Another form of contact with Mexico is the use of services such as auto repair, health care, tailoring or hairstyling. However, children claimed to use these services with rare frequency. On the other hand, about one-half of all of them regularly go to night clubs, discotheques and restaurants. Other forms of entertainment — dog racing, horse racing, water sports, picnics, bullfights and the like, which are rather typical of Anglo-American tourists — are not all procured by Chula Vista teenagers.

The parents Students were asked to give the names and telephone numbers of their parents (mother, father or both) in order to contact an adult in every household. About one-half of the students provided this information and eighty-one parents (48 mothers and 33 fathers) were interviewed in their homes a few weeks after the surveys were collected at Castle Park. 9 About three fourths (74%) of all the parents were born in Mexico in either a border area (21%) or in an urban or semi-urban community of the interior (53%). About one-fifth of all them (18%) were born in or near the U. S.-Mexico border. The mean number of years of residence in Chula Vista was 16 with a modal length of 10 years and a median of 14. The previous residence of more than one-half of them (59%) had been a southwestern community, and about one fourth (27%) had resided in a Mexican border city. In their previous residence parents had lived a mean of 10 years with a modal length of residence of 3 and a median of 8. The residential pattern of the second previous residence was similar to the first (i. e., the U. S. Southwest [35%] or a border area in Mexico [35%] with the only difference that about one in five parents [21 %] had resided in the interior of Mexico). In their second previous residence, the parents had lived less than one year. In sum, the data on residence show that Mexican adults settle down in Chula Vista only after living in either a Mexican border city or a southwestern community.

The dialectics of Spanish language loyalty

55

Contact with Mexico Parents' contact with Mexico is by no means intense, but it occurs more often than children's contact. More than one-half of the parents claim to regularly go to Mexico for shopping (54%), to visit relatives (56%), or to eat at restaurants (59%); also parents tend to use services (30%) that children hardly ever use, attend movies (16%) or prepare picnics (11%) in which their children do not seem to participate. When they cross the border, parents claim to stay in Mexico from a few hours to one day. In addition, as part of their contact, parents receive in their homes relatives from Mexico every week (13%), every month (27%), every other month (13%), every summer or on special occasions (40%). According to the parents, Mexican relatives may spend from one day to several weeks when the visit Chula Vista.

The results The American side of the U. S.-Mexico border provides a very useful scene for the study of attitudes towards Spanish, for Spanish is the mother tongue of most Mexican-origin migrants. In the ancestral country, Spanish is accepted as fully official and loyalty to the national language is linked to national loyalty. In the north of Mexico — where native indigenous populations have not impacted the Spanish-speaking population — people have amalgamated the notion of being Mexican with the speaking of Spanish. The questionnaire items elaborated for this study thus attempt to explore how people from Chula Vista feel about their ancestral language. The questions do not call on respondents to claim a behavioral commitment to Spanish; they are rather intended to elicit a reaction justified upon personal, emotional, and practical motives. Attitudes towards Spanish were approached in three dimensions: 1) Spanish as an instrumental language; 2) Spanish as an integrative language (i. e., as an identifier with Mexico); 3) Spanish as a language that serves personal and developmental purposes. The attitudinal variables consist of several items that are presented in Tables 1.1, 1.2 and 1.3 with their corresponding scale definitions and response percentages. Table 1.1 shows that parents are slightly more positive than their children in their evaluation of Spanish as a useful language for making good deals with the Spanish-speaking clientele in Chula Vista and when traveling in Spanish-speaking countries. However,

56

Margarita Hidalgo

τη •rt oo r- O

ο

m -rt ο-rt ο

o

cd 2 C CO « W · ^ kl ε νη u δ

4) Ό ε υ ε -ö es c u. α C ο c d dΟ ~ 3 > 3 ο α -2 εcΠ cd 2 cd < D ο c o υ -C Ο Ο α ο Ο CQ b u, co tu Ρη £ 2