See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/269101664

The Role of Primary Language Development in Promoting Educational Success

for Language Minority Students

Chapter · January 1981

DOI: 10.13140/2.1.1334.9449

CITATIONS READS

1,128 26,955

1 author:

Jim Cummins
University of Toronto
199 PUBLICATIONS 29,820 CITATIONS

SEE PROFILE

All content following this page was uploaded by Jim Cummins on 04 December 2014.

The user has requested enhancement of the downloaded file.

 SCHOOLING
AND
LANGUAGE.
MINORITY
STUDENTS:
A THEORETICAL
U.S. DEPARTMENT OF EDUCATION
FRAMEWORK
NATIONAL INSTITUTE Of EDUCATION "PERMISSION TO REPRODUCE THIS
I MA, A 110INA't SOUR(.ES INFORMATION MATERIAL HAS BEEN GRANTED BY
1,1NTFR DIRK:)
1h,, i5)s,,,14.4,1 has tststi, wpitedus,vs1 dti
otiMi 103313 3133 petson 1,r rogatt.tohon
totc.phal,,,s; I
hlmnt :31.331ye., t3 o. hews, c,,ese. f714)thy 'beet: crP. Ectuctibil
r4)13 flitp It n +Iudntt

Row Is 1,, 14113 41% SiAiPli ill !ill% 311,4 Al TO THE EDUCATIONAL RESOURCES
15 Pr`t 4,, 313,3 ',, 1,1,34ft.4.,11 01,1141 net INFORMA7 ION CENTER (ERIC
posthi, r I tY 410411 1,
Developed by
Office of Bilingual Bicultural Education
California State Department of Education
Sacramento, California .

Published by
Evaluation, Dissemination and Assessment Center
California State University, Los Angeles
Los Angeles, California
 s.

ISBN: 0-89755-011.0
Library of Congress Catalog Card Number: 81.71272

This publication was funded in whole or in part by a contract (G007902844) be-

tween the California State Department of Education and the United States Educa-
tion Department. The opinions expressed herein do not, however, necessarily
reflect the positions or policy of the United States Education Department; no of-
ficial endorsement by the United States Education Department should be
inferred.

Developed by
OFFICE OF BILINGUAL BICULTURAL EDUCATION
California State Department of Education"
Sacramento, California
Published and Disseminated by
EVALUATION, DISSEMINATION AND ASSESSMENT CENTER
,California State University, Los Angeles
Los Angeles, California

Printed in USA 1981

Second Printing 1982
Third Printing 1982
Fourth Printing 1982
Fifth Printing 1983
Sixth Printing 1983
Seventh Printing 1984
3
10.

HI

PREFACE
The growing interest in the problems of language minority
students In the United States has been accompanied by the
publication of an enormous number of books and articles. Often;
however, advice regarding approaches, methods, strategies, and
teithnlques for effectively educating language minority students
is Offered without any concern or explanation of empirical
evidence. With the possible exception of legal concerns, the Of-
lice of Bilingual Bicultural Education in the California State
Department of Education receives more Inquiries regarding
research evidence on the effectiveness of bilingual education
than on any other issue. Educators want to know which types of
programs actually work with non-English language background
students.
The Office of Bilingual Bicultural Education has identified
three major objectives for all instructional programs serving
language minority students. Regardless of the approach taken,
at the end of the treatment period, language minority students
should exhibit: (1) high levels of English language proficiency, (2)
appropriate levels of cognitive/academic development, and (3)
adequate psychosocial and cultural adjustment. 'The articles in-
cluded In this collection of papers explain the importance of
these goals and describe the likelihood of various types of in-
structional approaches to achieve such outcomes. Instead of
providing the reader with a series of unconnected suggestions
and recommendations, the authors collectively advance a
research-based theoretical framework for the design and im-
plementation of Instructional programs for language minority
students.
This publication is a progress report, not a collection of proven
answers. The theoretical framework implied in this volume is,
however, based on the best Information that science can provide
at this time. The research herein reported does not lead to
perfect progrpms with perfect outcomes, nor does it answer all
the questiod regarding language development, language ac-
quisition, and cognitive/academic development in bilingual con-
texts. But, taken collectively, these articles form the beginning of
a research-based theoretical framework for planning and improv-
ing bilingual education programs. We at the California State
Department of Education view this as substantive progress. We
iv

are encouraged by the potential practical applications of the

research presented in this collection and shall continue with the
refinement of this work. At the aame time, we not only invite
other researcherp, teacher trainers, and school district personnel
to put into practice the Ideas and Implications presented here
but also to improve and expand their programs to meet all of the
schooling needs of language minority students.

Guillermo.Lopez, Chief
Office of Bilingual Bicultural Education

5
0"

ACKNOWLEDGMENTS
Sincere appreciation is first expressed to the authors of the
five papers appear:ng in this collection: James Cummins,
Stephen Krashen, Dorothy Legarreta-Marcaida, Tracy Terrell, and
Eleanor Thonis. The contributors have not only astutely syn-
thesized their own research and the research of others but have
also skillfully presented the information In a manner especially
suitable to educators.
Special thanks are .1so due to the Project Team members in
the Office of Bilingual Bicultural Education: Maria Ortiz, Dennis
Parker, and Fred Tempes..These Individuals were responsible for
the design of the initial, outline questions, review of the interim
and final drafts, consultation with the authors, and preparation
of the articles for the publisher. In turn, the work of the Project
Team was ably promoted and supervised by several ad-
ministrators at the California State Department of Education, in-
cluding 'Ramiro Reyes, Guillermo Lopez, and Tomas Lopez.
Charles Leyba, Director, and his staff at the Evaluation,
Dissemination and Assessment Center, California State Univer-
any, Los Angeles are to be commended for their professional
preparation of this document. The Office of Bilingual Education
and Minority Language Affairs of the United States Education
Department provided the funds for this project.
Many more individuals and agencies, too many to UM here,
provided valuable assistance in the development of this collec-
tion; final .responsibility for this volume, nevertheless, rests with
the Office of Bilingual Bicultural Education, California State
Department of Education.

David P. Dotson
Project Team Leader
Office of Bilingual Bicultural Education
 vii

BIOGRAPHICAL SKETCHES

JAMES CUMMINS is a Visiting Professor in the Modern

Language Centre of the Ontario Institute for Studies on Educa-
tion. Re received his PhD from the University of Alberta in 1974.
His research interests include minority group achievement, bi-
lingualism, and reading disability. He is the recipient (with J. P.
Das) of the International Reading Association 1979 Albert J. Har-
ris Award for the best papers on detection and remediation of
reading disability.

STEPHEN KRASHEN is Professor of Linguistics at the Univer-

sity of Southern California. He has published extensively on
second language acquisition theory, second language teaching, c,

and neurolInguistics. He Is the author of Second Language Ac.

quisfccn and Second Language Learning (Pergamon Press,
:A.:1), co-author of The Human Brain (Prentice-Hall, 1977); and
-editor of Research in Second Language. Accp Onion (Newbury
House, 1980).

DOROTHY LEGARRETA-MARCAIDA Is Adjunct Professor of

Basque Studies at the University of Nevada, Reno, Dr. Legarreta
has published several articles on second language acquisition
and primiry language maintenance. Her earlier research was on
Black dialect speakers. Her present book-in-progress is a study
of ethnic identity and language maintenance among nearly
20,000 Basque children evacuated during the Spanish Civil War
and sent to France, England, Belgium, the U.S.S.R., and Mexico.

TRACY DAVID TEIWELL is the Chair of the Program in

Linguistics and a meraberof the Department of Spanish and Por-
tuguese at the University of California, Irvine. He has written
many articles on Spanish phonology, especially on the
phonology of Caribbean Spanish. His theoretical interest is
sociolinguistics and, in particular;language change. His recent
area of specialization is second language acquisition, especially
language teaching. He is perhaps best known as the !nitlator of a
-new" communicative approach tr; languagrl instruction, "The
 Natural Appfoach." He has published a text on applied
linguistics with Maruxa Satgues de Cargill called Linguistics
aplIcada (J. Wiley 3,nd Sons) and another will appear shortly on
Spanish phonetics with R. Barrutia, Fanatics y fonologla
espaialas (J. Wiley and Sons). He will also publish a book co-
authored with S. Krashen entitled The Natural Approach:
Language Acquisition In the Classroom (Pergemon Press).

ELEANOR WALL THONIS is the district psychologist for

Wheatland Elementary School District; a part-time instructor.
University of California, Berkeley; and the consultant for bi-
lingual education in the"Marysville Unified School District. She
has served as the director of the area Reading Center and Is the
author of several publications on reading for language minority
students.

a
 ix

INTRODUCTION
There are presently more than 375,000 students of limited
English proficiency in California public schools. There are an ad-
ditional 433,000 students of fluent English proficiency who have
a home language other than English. This means that
California's language 'minority student population in
kindergarten through grade twelve approximates 010,000.
As a group, language minority students tend to do poorly In
regular school programs. They da not acquire the language,
academic, and sociocultural skills necessary to meet the
challenges of vocational and higher education pursuits. Many
language minority students achieve only low levels of primary
language proficiency while acquiring less than native-like ability
In English.
Making decisions about instructional offerings for language
minority students has proven to be e. complex and demanding
task for school personnel and parents alike. Part of the difficulty
can be attributed to the absence of a theoretical framework upon
AO programs for language minority students can be based.
Without a framework, decision makers are often unable to focus
consistently upon the psychosocial and educational factors that
",_ influence the school achievement of language minority students.
While political and economic factors are alsO important, basing
educational programs solely on such grounds tends to affect
negatively the quality of the educational experience of language
minority students. Only by clearly understanding what educa-
tional attainments are possible for language minority students
can school personnel and parents judge the approprigteness of
the educational practices currently utilized by local schools.
Although political and economic compromises may be
necessary, they are best made when decision makers under-
stand as many of the pertinent dynamics as possible. This
publication offers information related specifically to the educa-
tional consequences of program decisions.on language minority
students.
In the past, mbst knowledge about programs for language
minority studtmts was based entirely on authority (laws and ex-
perts), the personal experiences of educators, and the "common
sense" reasoning of program designers and planners. Such infor-
mation may be important but is in Itself insufficient for making
-
critical educational decisions. Therefore, the Office of Bilingual
Bicultural Education of the California State Department of
Education has decided to turn to scientifically controlled studies
to establish the validity c I knowledge about Instructional pro-.
grams for language minority students. Empirical knowindge is
ceftain to Improve the ability of educators to predict individual
Student and program outcomes for specific types of students,
given certain types of Instructional treatments, and under dif-
ferent types of background conditions. Thus, the articles con-
tained In this collection represent an initial step in the develop-
ment of a research-based theoretical framework for the school-
ing of language minority students.
This collection of papers Is divided Into two major parts. The
first section, consisting of papers by James Cummins and
Stephen Krashen, addresses the theoretical underpinnings of
primary language development, second language acquisition,
and the relationship of both to normal school achievement. The
second section coigains a series of three papers, each ex-
panding upon the theoretical works In the first section and pro-
viding the reader with numerous instructional methods and
techniques, all consistent with each other and with the various
hypotheses posited by Cummins and Krashen.
No pedagogical issue relating to the instruction of language
minority students has been more vigorously debated than the
role of minority languages in bilingual education program e. In his
work, Cummins clarifies the role of the primary languagi, by: (1)
describing the nature of language proficiency and its connection
to academic and cognitive development, (2) identifying different
levels of bilingualism experienced by language minority students
and predicting the corresponding effects of each level on
academic achievement, and (3) 'suggesting a relationship be-
tween primary language development and eventual attainment In
the second language through tits notion of a common undeilying
dimension of language proficiency. Clearly, Cummins has
developed several important hypotheses and constructs that
help explain and reconcile the seemingly contradictory findings
of many other researchers.
While Cummins' article focuses on primary language cleivelop
ment and academic achievement, Krashen dedicates attention
to the acquisition of a second language, specifically English. The
author distinguishes between language acquisition and

10
 xi

language learning environments. The former, it is suggested,

leads to fluency, while the latter assists in the development of
what Krashen palls the "Monitor." Krashen also suggests that
the key to second language acquisition .1s exposure to "com-
prehensible Input" in substantial amounts and under Optimal
conditions. In describing the conditions necessary for second
language acquisition, Krashen analyzes the potential of `variJus
second language (grammar and communicative-based English-
as-a-Second-Language) and bilingual education (immersion,
transitional, and ideal) programs to meet the language and
academic needs of minority students.
Based on the assumption that high levels of primary language
proficiency promote adequate school achievement, Dorothy
Legarreta-Marcalda explores the effective use of primary
language In bilingual classrooms. The author addresses five key
questions related to the design, management, and implementa-
tion of bilingual classes:
1. To what extent should the child's primary language be used
overall in grades K-6?
2. In what manner should primary language Instruction be
delivered:
a. Concurrent translation?
b. Alternate immerton (direct method) usually through
language dominaht groupings?
3. What variety of the primary language should be used in the
classroom?
4. How can we ensure the prestige of the primary language via
vie the dominant language, English?
5. How can primary language use be monitored: a formative
evaluation process?
The responses to those questions are based on the recent find-
ings of empirical studies and the published material of ex-
perienced professionals.
Complementing Krashen's theoretical hypotheses, Tracy D.
Terrell presents an acquisition model called the "Natural Ap-
proach." This model Is one means of applying Krashen's
theoretical constructs. Terrell not only describes Vie model in
eelationship to Krashen's work and the work of others but also
addresses: (1) the principles of the Natural Approach, (2) natural
language acquisition situations, (3) appropriate teacher

11
xii

behaviors, (4) sample teaching techniques and strategies, (5) the

use of continua in the Natural Approach, and (8) student evalua-
tion. Terrell's article Is enhanced by the inclusion of many actual
classroom examples related to suggested techniques and
strategies.
The last paper, by Eleanor Thonis, deals with reading instruc-
tion in bilingual contexts. She expands upon Cummins' notion of
a Common. Underlying Proficiency (CUP) In the
cognitive/academic language skills area as it relates to literacy
acquisition among bilingual students. The potential for primary
language reading skill transfer to English is disCussed in detail.
Additionally, the author suggests appropriate methods and
techniques designed to promote primary language literacy in and
out of the classroom. Thonis concludes her article with a
description of the positive outcomes associated with billteracy.
Finally, the compendium concludes with an Appendix and a
Glossary. The appendix contains a sample copy of the 1981-82
version of the Bilingual Education Program Quality Review in-
strument, Kindergarten Through Grade Six. The use of this instru-
ment is one way the Office of Bilingual Bicultural Education pro-
motes the principles and standards of Implementation sug-
gested by the theoretical framework implied In this compendium.
Other promotional strategies Include the development of Asian
and minority group handbooks and periodic presentations at
regional technical assistance workshops for local *Khoo( district
personnel as well as coordination meetings with resource agen-
cies and county schools offices. The glossary of terms has been
included to assist the reader by promoting consistency in the use
of the key terms across articles, It Is suggested that the reader
become familiar with the entries in the glossary before attempt-
ing a thorough reading of any of the papers.
Most educators, government officials, parents, and communi-
ty members would agree that the goal of educational programs
designed for language minority stertiants Is to allow such
students to develop the highest degree possible of language,
academic, and social skills. necessary to participate fully In all
aspects of life. More specifically, as a result of an instructional
treatment, language minority students should attain: (1) high
levels of English language proficiency, (2) normal cognitive and
academic achievement, (3) adequate psychosocial and cultural
adjustment, and (4) sufficient levels of primary language develop-

12
ment to promote normal school progress. Based on the empirical
evidence prasented in the five articles contained in this collec-
tion, properly designed and adequately implemented bilingual
education programs are one means to achieve such goals.
As a result of the controversy regarding-the legislation of state
and federal requirements, some educators today erroneously
believe that many parents and community members view bi-
lingual education with disfavor. This is not necessarily the case.
In a recent poll conducted for Newsweek by the Gai kip Organize
tion (March, 1981), 64 percent of the American public approved of
classes conducted In a foreign language as well as In English for
children who do not speak English. Another 14 percent did not
know enough about the value of these classes to make a judg-
ment. Contrary to the statements of a few political opportunists,
uninformed newspaper columnists, and some special interest
group representatives, the public Is generally supportive of
primary language instruction for language minority students,
even though there may not be an awareness of the strong scien-
tific case for such progranis.
The task of educating language minority students Is not sim-
ple. Nevert:' eless, creative and committed educators In coopera-
tion with r onm:ne7; parents and community members have
designed and WO wi,ited educational programs that result in
significantly Improved school performance on the part of such
students. In other words, under certain conditions, language-
related problems are no longer as likely to interfere with the
academic and vocational aspirations of language minority
students and their families. To accomplish this, educators must
rely upon empirical evidence rather than "folk remedies" as a
guide to professional decisions for selecting and Implementing
instructional programs for language minority children. This
publication is meant to be an important contribution toward this
end.

David P. Dotson
Project Team Leader
Office of Bilingual Bicultural Education

13
 CONTENTS

Page
PREFACE III
ACKNOWLEDGMENTS
BIOGRAPHICAL SKETCHES vu
INTRODUCTION Ix

PART ONE: THEORETICAL FOUNDATIONS

James Cummins
The Role of Primary Language Development in Promoting
Educational Success for Language Minority Students
3
Stephen D. Krashen
Bilingual Education and Second Language Acquisition Theory...
51

PART TWO: STRATEGIES FOR IMPLEMEN-

,TATION
Dorothy Legarreta-Marcaida
Effective Use of the Primary Language in the Classroom 83
Tracy D. Terrell
The Natural Approach in Bilingual Education 117
Eleanor W. Thonis
Reading Instruction for Language Minority Students 147
APPENDIX
Bilingual Education Program Quality Review Instrument,
Kindergarten Through Grade Six 183
GLOSSARY
215

14
 Part One
Theoretical Foundations

. 13
 The Role of Primary Language
IhvelopMent in Promoting Educational
Success for Languagi\Minority Students*
James Cummins

IN ORDER TO ASSESS the role of langtlage minority students'

primary language (LI) development in the acquisition-of English (L2)
academic skills, it is necessary to consider two questions: (1) What is
meant by "language proficiency"? and (2) What are the cross-lingual
dimensions of language proficiency, i.e., how does the development of.
proficiency in LI relate to the development of L2 proficiency? Confusion
concerning the rationale for bilingual education, assessment of bilingual
proficiency, and entry-exit criteria for bilingual programs stems from in-
adequate conceptualization of the nature of language proficiency and its
cross-lingual dimensions.
To account for the research data on bilingual education, it is necessary
to distinguish those aspects of language proficiency involved in the
development of literacy skills from other aspects of language proficien-
cy, and to note that these literacy-related aspects are interdependent
across languages, i.e., manifestations of a common underlying proficien-
cY
This paper is organized into three sections. First, the nature of
language proficiency and its relationship to academic and cognitive
development is considered. In the second section, the origins of current
misconceptions about bilingualism are examined, and a theoretical posi-
tion regarding the nature of bilingual proficiency is formulated in light of
the research data. The third section applies these theoretical positions
regarding the nature of language proficiency and its cross-lingual dimen-
sions to the current debate over the rationale for bilingual education, en-
try and exit criteria, and assessment of bilingual proficiency.

Many people have contributed to the present paper through comments on previous ver-
sions of the theoretical framework which it elaborates. I would like to thank Michael
Cana le, Steve Chetiarek, Lily Wong Fillmore, Fred Genesee, Steve Krashen, John Oiler
Jr., Muriel Saville-Troike, Bernard Spoisky, Merrill Swain, Rudolph 'Troike, and Beaji
Wald for their constructive criticisms. The suggestions of the editorial team for the pre-
sent volume have also been aureately useful and for this I would like to thank David
Dotson, Maria Ortiz, Dennis Parker, and Fred Tempes of the Office of Bilingual-
Bicultural Education, California State Department of Education.
4 Sch.lo ling and Language Minority Students:

The Nature of Language Proficien4

How Misconceptions About English Proficiency Create Academic
Deficits is Language Minority Students
The rationale for bilingual education in the United States (United
States.Commission on Civil Rights, 1975), as it is understood by most
policy makers and practitioners, can be stated nfollows:
Lack of English proficiency is the major reason for language
minority students' academic failure. Bilingual education is in-
tended to ensure that students do not fall behind in subject
matter content while-they are learning English, as they would
likely do in an all-English program. However, when students
have become proficient in English, then they can be exited to
an all-English program, since limited English proficiency will
no longer impede their academic progress.
Despite its intuitive appeal, there are serious problems with this ra-
tionale. First, it ignores the sociocultural determinants of minority
students' school failure which, it, will be argued. are more fundamental
than linguistic factors. Second, an inadequate understanding of what is
meant by "English proficiency" is likely to result in the creation of
academic deficits in language minority students.
Some concrete examples will help illustrate how this process operates.
These examples are taken from a Canadian study in which the teacher
referral forms and psychological assessments of over 400 language
minority students were analyzed (Cummins, 1980c). Throughout the
t ic hers' referral forms and psychologists' assessment reports are
ferences to the fact that children's English communicative skills appear
considerably better devakxd than their academic language skills. The
following examples illustrate this point:
PS (094). Referred for reading and arithmetic difficulties in
second grade, teacher commented that "since PS attended
grade one in Italy, I think his main problem is language,
although he understands and speaks English quite well."
GG (1114). Although he had been in Canada for less than a
year, in November of the grade one year, the teacher com-
mented that "he speaks Italian fluently and English as well."
However, she also referred him for psychological assespnent
because "he is having a great dew of difficulty with the grade
one program" and she wondered if he had "specifk learning
disabilities or if he is just a very long way behind children in
his age group."
DM (105). Arrived from Portugal at age 10 and was placed in
a second grade class; three years later in fifth grade, her

17
A Theoretical Framework 5

teacher commented that "her oral answering and comprehen-

sion is so much better than her written work that we feel a
severe learning problem is involved, not just her non-English
background."
These examples illustrate the influence of the environment in develop-
ing English communicative skills. In many instances in this study im-
migrant students were considered to have sufficient English proficiency
to take a verbal IQ test within about one year of arrival in Canada.
Similarly, in the United States, language minority students are often con-
sidered to have developed sufficient English proficiency to cope with the
demands of an all-English classroom after a relatively short amount of
time in a bilingual program (in some cases, as little as six months).
There is little doubt that many languaae minority students can develop
a relatively high degree of English communicative skills within about two
years of exposure to English-speaking peers, television, and schooling.
However, in extrapolating from the considerable English proficiency
that language minority students display in face-to-face communication
to their overall proficiency in English, we risk creating academic deficits
in these students.
Consider the following example: .

PR (289). PR was referred in first gra& by the school prin-

cipal who noted that "PR is experiencing considerable dif-
ficulty with grade one work. An intellectual assempent would .
help her teacher to set realistic learning expectations for her
and might provide some clues as to remedial assistance that
might be offered."
No mention was made of the child's ESL background; this only
emerged when the child was referred by the second grade teacher in the
following year. Thus, the psychologist does not consider this as a possi-
ble factor in accounting for the discrepancy between a verbal IQ of 64
and a performance IQ of 108. The assessment report read as follows:
Although overall, ability level appears to be within the low
average range, note the significant difference between verbal
and nonverbal scores....It would appear that PR's develop-
ment has not progressed at a normal rate and consequently
she is, and will continue to experience much difficulty in
school. Teacher's expectations at this time should be set ac-
cordingly,
What is interesting in this example is that the child's English com-
municative skills are presumably sufficiently well developed that the
psychologist (and possibly the teacher) is not alerted to the child's ESL
background. This leads the psychologist to infer from her low verbal 1Q
6 Schooling and Language Minority Students:

score that "her development has not progressed at a normal rate'" and to
advise the teacher to set low academic expectations for the child since she
"will continue to experience much difficulty in school." There is ample
evidence from many contexts (Mercer, 1973) of how the attribution of
deficient cognitive skills to language minority students can become self-
fulfilling,
Jn many of the referral forms and psychological assessments analyzed
in this study, the following line of reasoning was invoked:
Because language minority istudents are fluent in English,
thebr poor academic performance and/or test scores cannot
be attributed to lack of prolickluy in English. Therefore,
these students must either have defirient cognitive abilities or
be poorly motivated ("tazy,.
The trend to exit students to all- English programs as quickly as poisi-
ble in many' United States bilingual programs inevitably gives rise to a
similar line of reasoning. It is commonly observed that students classified
as "English proficient" after a relatively short stay in a bilingual pro-
gram and then exited to an all-English program often fall progressively
further behind grade norms in the development of English academic
skills. Because these students appear to be fluent in English, their poor
academic performance can no longer be explained by their English
language deficiency. Policymakers and educators are also reluctant to
blame the school for minority students' poor performance because the
school has accommodated the students by providing a bilingual pro-
gram. Once again, the academic deficiency will be attributed to factors
within the child.'
It is frequently assumed that language minority students have become
"English proficient" when they have acquired relatively fluent and peer-
appropriate face-to-face communicative skills. examples cited
above, as well as the research evidence reviewed in the remainder of this
paper, strongly suggest that this misconception operates to impede the
academic progress of language minority students. To understand the
nature of this misconception, it is necessary to consider the question of
what is meant by "English proficiency."

This process is, in wary respe cts. the opposite of the attribution of deficient cognitivear
linguistic ability an the basis of autface structure dialectal differences (Shay, 1977). In
the present situation. the presence of adequate surface structure leads teachers to
eliminate "lack of English proficiency" as an explanatory variable with the result that
low academic performance is attributed to deficient cognitive abilities In language
minority students.

19
A Theoretical Framework 7

What Is Meant By "English Profichincy"?

There is still little consensus among researchers as to the nature of
"language proficiency" or "communicative competence."a For exam-
ple, a model proposed by Hernandez-Chavez et al. (1978) comprised 64
separate prOriciencie' s, each of which, hypothetically, is independently
measurable. At the other extreme is Oiler's (1978; 1979) claim that
" "...there exists a global language proficiency factor which accounts for
the bulk of the reliable variance in a wide variety of language proficiency
measures" (1978, p. 413). This factor is strongly related to cognitive
abililty and academic achievement measures and is about equally well
measured by certain types of listening, speaking, reading, and writing
tasks.'
The communicative competence framework proposed by Canale
(1981), on the basis of the earlier Canale and Swain (1980) theory, adopts
an intermediate position ii distinguishing four components. These are:
1. Grammatical competence: Mastery of the language code (e.g., lex-
ical items and rules of word formation,' sentence formation, literal mean-
ing, pronunciation, and spelling).
2. Sociolinguistic competence: Mastery of appropriate language use in
different sociolinguistic contexts, with emphasis on appropriateness of
meanings and forms.
3. Discourse competence: Mastery of how to combine earrings and
forms to achieve a unified text in different modes (e. ., telephone in-
cluirrguitientative essay, and recipe) by using (a) /lesion devices to
relate utterance forms (e.g., pronouns and trans n words), and (b)
coherence rules to organize meanings (e.g.. repetition progression, con-
sistency, and relevance of ideas).
4. Strategic competence: Mastery of verbal and non-verbal strategies
(a) to compensate for breakdowns in communication due to insufficient
competence or performance limitations (e.g., strategies such as use of
dictionaries, paraphrase, and gestures), and (b) to enhance-communica-
tion effectiveness.

a Although language can be used for purposes not overtly communicative, e.g.:
problem-sabring (Canal! and Swain, 1980), these "analytic" (Bruner, 1975) language
skills develop within a matrix of human interaction; thus, for purposes of this paper,
the terms "language proficiency" and "communicative proficiency" are being
used synonymously.
if should be noted that Oiler (1979) leaves open the possibility that there may be smaller
specific components of language proficiency that are not encompassed by the global pro-
ficiency dimension.

20
 $ Schooling and Language Minority Students:

There are two major problems in applying this or any other theoretical
framework for communicative competence to minority students' acquisi-
tion of English proficience.r. First, these theories tend to be static since the
developmental aspects of communi ive competence in LI and 12.are
left vague; second, in general, little nsideration has been given to the
role of specific acquisition contexts in ermining the interrelationships
and development of different aspects or *communicative competence
(howeverp see Canale, 1981). In particular, the nature of the com-
municative demands of schooling (e.g., processing language outside of
one-to-one, face-to-face situations) has not been considered. The
relevance of these problems can be seen by examin* the development of
English proficiency among native English-speaking children.
The Development of English Proficiency in School Contexts. The
development of language proficiency can be considaKin two very dif-
ferent ways. First is the acquisition of what Bruner (197 as termed the
"species minimum" involving the phoncilogical, syntactic, and semantic
skills that most native speakers have acquired by age six (there is little
difference between the phonological competence of a six-year-old and a
fourteen-year-old). Similarly, mastery of basic syntax approaches
maturity by.age six, although the development of more sophisticated
rules and flexibility in grammatical control will continue into early
adolescence (Chomsky, 19:72). Also, semantic categories such as agent,
instrument, and recipient of action are present at a very early age.
However, in contrast to the acquisition of this "species minimum"
competence, other aspects of language proficiency continue to develop
throughout the school years and beyond. Obvious examples are literacy-
related language skills such as reading comprehension, writing ability,
and vocabulaiy/concept knowledge. Within each of the four com-
ponents oP communicative competence distinguished by Canale (1981),
4
nativeaspeakers achieve mastery levels in some subskills prior to others.
For example, within grammatical competence virtually all native
speakers master pronunciation before spelling. Similarly, some aspects 4
of sociolinguistic, discourse, and strategic competence will be mastered
at an early age and others much later, if at all.
However, within a second language context very different relation-
ships may exist among the various subskills, depending upon the specific
acquisition context, e.g.; formal L2 classroom vs. real life exposure, or
pre -schOol immigrant children vs. adolescent immigrant children whose
LI literacy skills are well developed. Also, the relationship of language
proficiency to cognitive and academic variables will vary both between
L I and 12 contexts and also vgithin L2 contexts, depending upon the con-

21
 A Thcoretital Framework
9

ditions of acquisition. Thus, almost by definition, the "species

minimum" will be attained by all native speakers regardless of academic
or cogitive abilities; however, this will pot necessarily be the case among
L2 ilearners. For example, prOnuncitftion skills may remain poorly
developed among many older L2 learners. Also,. cognitive and personali:
variables are likely to differentially influence the acquisition of dif-
ferent aspects of Li proficiency in different contexts. As Fillmore (1979)
suggests, personality variables (e. g., sociability) may be most influential
in determining the acquisition rate 'of L2 face-to-face conuiunication
skills in it peer interaction situation; however, cognitive skills in a peer in-
teraction situation; however, cognitive skills may be more involved in
determining the acquisition rate of L2 literacy skills in a classroom con-
text.
\ In short, current theories of communicative competence are not par-
ticularly helpful in elucidating issues related to the development of
English proficiency by language minority students. This is because these
theories (1) fail to incorporate a developmental perspective; (2) fail to
consider the development of communicative comietepce explicitly in
relation to specific contexts, in particular the school c6ntext; and (3) fail
to examine the developmental relationships between t I and L2. In other
words, the usefulness of most current theories is limited because they
either exist in a developmental and contextual vacuum or else have been
proposed in a very different context from that of bilingual education in
the United States.
The necessity for consickirinkilre question of what constitutes
language proficiency in school contexts from a developmental perspec-
tive is highlighted by a recent study which shows that immigrant students
arriving lifter age six take between six and seven years to approach grade
norms in English academic skills (Cummins, 1981). Results-of this study,
conducted among 1,210 immigrant students in the Toronto Board of
Education, are shown in Figure I. The Picture Vocabulary Text (PVT)
consisted of a group-administered vocabulary test, and results were
broken down by Age on Arrival (AOA) and Length of Residence (LOR).
Clearly, it takes considerably longer for immigrant students to develop
age-appropriate academic skills in English (five-seven yeari LOR) than it
does to develop certain aspects of age-appropriate English com-
municative skills (approximately two years). The reason is pot difficult to
see. Literacy-related language skills (such as vocabulary range) continue
to develop among native speakers throughout the school years, +whereas
some salient aspects of face-to-face communicative skills reach a plateau
by about age six. Clearly, many other aspects of face-to-face com-

42
 10 Schooling and Language Minority Students;

Figure 'I

AGE ON ARRIVAL, LENGTH OF RESIDENCE, AND PVT

STANDARD SCORES

AOA 0-1 10-11

11:9 12-13 1415

municative skills continue to develop throughout the school years; but

the data considered above suggest that these are not particularly salient
for teacher's and psychologists.
In a previous section, it was pointed out that failure to distinguish
these two dimensions of English proficiency can result in educational
deficits for language minority students. At this point, it may be helpful
to describe this distinction More completely and place it into a broader
theoretical framework so that it can be used to examine the developmen-
tal relationships between LI and L2 proficiency within bilingual educa-
tion programs.

23
A Theoretical Framework 11

A Theoretical Framework4
To recapitulate, three minimal requirements for a theoretical
framework of communicative proficiency relevant to bilingual education
in the United States have been outline* First, such a framework must in-
corporate a developmental' perspective so that those aspects of com-
municative proficiency mastered early by native speakers and 12 learners
can be distinguished from those varying across individuals as develop-
ment progresses; second, the framework must permit differences
between the linguistic demands of school and those'of interpersonal con-
texts outside the school to be described; and third, the framework must
allow for the developmental relationships between LI and L2 proficiency
to be described.
The framework developed in response to these requirements i
presented in Figure 2. The framework proposes that in the context of
United States bilingual education, communicative proficiency can be
conceptualized along two continuums. A continuum related to the range
of contextual support available for expressing or receiving meaning is
described in terms of "context-embedded" versus "context-reduced"
communication. The extremes of this continuum are distinguished by the
fact that in context-embedded communication the participants can ac-
tively negotiate meaning (e.g., by providing feedback that the message
has not been understood) said the language is supported by a wide range
of meaningful paralinguistic (gestures, intonation, etc.) anchituational
cues; context-reduced communication, on the other hand, relies primari-
ly (or at the extreme of the continuum, exclusively) on linguistic cues to
meaning and may, in some cases, involve suspending knowledge of the
"real" world in order to interpret (or manipulate) the logic of communi-
cation appropriately.°
In general, context-embedded communication derives from interper-
sonal involveMent in a shared reality that reduces the need for explicit
linguistic elaboration of the message. Context-reduced communication,
on the other hand, derives from the fact that this shared reality cannot be
assumed and. thus linguistic messages must be elaborated precisely and
explicitly so thit the risk of misinterpretation is minimized. It is impor-

°This theoretical framework should be viewed wilt, in a social context. The li.%iiguast profi-
ciencies described develop as a result of various types of communicative interactions in
home and school. The nature of these interactions is, in turn, determined by broader
societal factors. as described later in this paper.
The term "context-reduced" is used rather than "disembedded" (Donaldson. 1978) or
"decontextualizecl" because there is a large variety of contextual cues available to carry
out tasks even at the context-reduced end of the continuum. The difference, however, is
that these cues are exclusively him:is:iv in nature.

24
 12 Schooling and Language Minority Students:

Figure 2

RANGE OF CONTEXTUAL SUPPORT AND DEGREE OF

COGNITIVE INVOLVEMENT IN COMMUNICATIVE
ACTIVITIES

COGNITIVELY
UNDEMANDING

CONTEXT- CONTEXT.-
EsigEDDED kEDUCED

COGNMVELY
DEMANDING

tam to emphasize that this is a continuum and not a dichotomy. Thus,

examples of communicative behaviors going from left to right along the
continuum might be: engaging in a discussion, writing a letter to a close
friend, and writing (or reading) an academic article. Clearly, context-
embedded communication is more typical of the everyday world outside
the classroom, whereas many of the linguistic demands of the classroom
reflect communication that is closer to the context-reduced end of the
continuum. Recent research, reviewed by Tannen1(1980), suggests that
part of minority students' failure in mainstream classrooms may derive
from application of context- embedded strategies in the school setting
where context-reduced strategies (e.g., responding in terms of the logic
of the text rather than in terms of prior knowledge) are expected and
rewarded.
The vertical continuum is intended to address the developmental
aspects of communicative competence in terms of the degree of active
cognitive involvement in the task or activity. Cognitive involvement can
be conceptualized in terms of the amount of information that must b?.

25
 A Theoretical Framework 13

processed simultaneously or in close succession by the individual in order

to carry out the activity. .

How does this continuum incorporate a developmental perspective? If

we return to the four components of communicative competence (gram-
matical, sociolinguistic, discourse, and strategic) discussed by Canale
(1981), it is clear that within each one some subskills are mastered more
rapidly than others. In other words, some subskills (e.g., pronunciation
and syntax within LI grammatical competence).reach plateau levels at
which there are no longer significant differences in mastery between in-
dividuals (at least in context-embedded situations). Oder subskills con-
tinue to develop throughout the school years and beyond, depending ,
upon the individual's communicative needs. .

Thus, the upper parts of the vertical continuum consist of com-

municative tasks and activities in which the linguistic tools have become
largely automatized (mastered) and thus require little active cognitive in-
volvement for appropriate, performance. At the lower end 4 the con-
tinuum are tasks and activities in which the communicative tools have
not bicome automatized and thus require active cognitive Involvement.
Persuading other individuals that your point of view rather than theirs is
correct, or rriting an essay on a complex theme, are examples of such ac-
tivities. In these situations, it is necessary to stretch one's linguistic
resources (i.e., grammatical, sociolinguistic, discourse, and strategic
competencies) to the limit itrorder to achieve one's communicative goals.
Obviously, cognitive Involvement can be just as intense in context- e.
einbedded as in context-reduced activities.
AF masted, is developed, specific linguistic tasks and skills travel from
the bottom towards the top of the vertical continuum. In other words,
there tends to be a high level o( cognitive involvement in task or activity
performance until mastery has been achieVed or, alternatively, until a.
plateau level at less than mastery levels has been reached (e.g., L2 pro-
nunciation in many adult immigrants). Thus, learning Ifie phonology
and syntax of LI, for example, requires considerable cognitive involve-
ment for the two- and three-year-old child, and thus these tasks would be
placed in quadrant B (context-embedded, cognitively demanding).
However, as mastery of these .skills develops, tasks inviolving them would
move from quadrant B to quadrant A, since performance becomes
I gereiler and Seard,ainalia (M)) 4nt out that as children learn to write, the progressive
automatization of lower level skills (e.g., handwriting, spelling of common wards, pu -
tuation, common syntactic forms, etc.) releases increasingly more mental capacity
higher level planning of large chunks of discourse. To illustrate what writing must be ke
for a young child, they suggest trying to do some original writing with the wrong hand. It
is likely to be difficult to think much beyond the ward being written.

26-*
1
14 Schooling and Language Minority Students:
a

increasingly imtontatizel---and cognitively demanding. In a second

language context, the same type of developmental progression occurs. As
specific linguistic tasks and skills are mastered in L2, they move up the
vertical continuum.'
Literacy Development and Communicative Proficiency. Clearly,
within this theoretical framework, literacy is viewed as one aspect of
communicative proficiency. Although there are inherent characteristics
of literacy tasks that place them towards the context-reduced end of the
horizontal continuum, most theorists would agree that the more reading
and writing instruction can be embedded in a meaningful communicative
context (i.e., related to children's previous experience4 the more suc-
cessful it is likely to be. As the papers (this volume) by Krashen (1981)
and Terrell (1981) emphasize, the same principle holds for second
language instruction. The more context-embedded the initial L2 input,
the more comprehensible it will be and, paradoxically, the more suc-
cessful in ultimately developing 12 skills in context-reduced situations.
Thus, a major pedagogical principle for both LI and L2 teaching is that
language skills in context-reduced situations can be most successfully
developed on the basis of initial instruction which maximizes the degree
of context-embeddedness.
In terms of the vertical continuum, developmental relationships be-
tween cognitive ability and reading performance can be readily inter-
preted. Singer (19/7) reviews data that show a change between grades 1
*An impticat4.in of this theoretical framework for theories of communicative competence
is that there is likely to be different relationships among language tasks in a first
istallusae, compared to a second language cornett. This is because 12 learners are likely
to have lower levels of certain L2 skills as compared to native speakers. In other words,
tasks located close to the top of the vertical continuum for native speakers may be dose
to the bottom for 12 her s. Also, acquisition contexts may vary between 1.2 learners
and native 'peaky.,,. For example, skills acquired in context-embedded situations by
native speakers may have been learned in context-reduced situations (e.g., formal
classrooms) by L.2 learners. This would also result in variable relationship among
language skills between native speakers and L2 learners. Thus, an important
characteristic of the theoretical framework is that although communicative tasks and ac-
tivities can be mapped onto it in a general way (e.g., inherent test charactedstics make
reading and writing less context-embedded than facr-to-face communication), the exact
location of any particular task on the horizontal and vertical continuums will depend on
the individual's or group's proficiency Ind and acquisition context. Thus, for inunigrant
students in the host country for two years, acathmic tasks in 12 are likely to be more
cognitively demanding and context-reduced than for native speakers.
Space does not permit the question of individual differing= in learning styles among L2
learners to be discussed in detail. However, within the present framework, learning style
can be regarded as the way in which Individual learners define the degree of cognitive in-
volvement and context -em beckhedners of particular tasks. Thus, at least three factors
must be taken into account in .)eating any particular task in relation to the two con-
tinuums: (1) the task's inherent characteristics, (2) the learner's general level of proficien-
cy. and (3) the learner's individual learning style.

27
 A Theoretical Framework 15

and 5 in the amount of common variance between IQ and reading

achievement from 16 to 64 percent (correlations of .40 to .79). He inter-
prets this in terms of the nature of the component skills stressed in
reading instruction at different grade levels.
As reading achievement shifts from predominant em-
phasis on word recognition to stress on word meaning
and comprehension, the mental functions being assessed
by intelligence and reading tests have more in common.
(Singer, 1977, p. 48)
As development progresses, word meaning and reasoning-in-reading
(e.g., inferring and predicting text meaning) rather than word decoding
skills account for the variance between good and poor readers. In terms
of the present framework, word meaning and reasoning-in-reading skills
remain in the lower end of the vertical continuum (i.e., variance between
individuals in these skills remains large), whereas word recognition skills
tend to climb towards the upper end of the continuum as development
progresses. In other words, as fluency in reading is acquired, word
recognition skills are first automatized and then totally short-circuited,
since the proficient reader does not read individual words but engages in
a process of sampling from the text to confirm predictions (Smith, 1978).
Relevance of the Theoretical Framework to the Achievement of
Language Minority Students. A major aim of literacy instruction in
schools is to develop students' abilities to manipulate and interpret
context-reduced cognitively demanding texts (quadrant 1)). One reason
why language minority students have often failed to develop
high levels of academic skills is because their initial instruction has em-
phasized context-reduced communication, since instruction has been
through English and unrelated to their prior out-of-school experiences.
Attempts to teach English through context-reduced audiolingually-based
ESL may very well have been counter-productive in some respects
(Legarreta. 1979).
However, another contributing factor to minority students' academic
failure, and one which is still operating even in the context of bilingual
programs, is that many educators have a very confused notion of what it
means to be proficient in English. Ifianguage minority students manifest
proficiencies in some context-embedded aspects of English (quadrant A),
they are often regarded as having sufficient English proficiency both to
follow a regular English curriculum and to take psychological and educa-
tional tests in English. What is not realized by many educators is that
*Clearly. the relationships between IQ and early reading achirvement may vary as a funs -
Lion of the instructional approach.

28
16 Schooling and Language Minority Students:

because of language minority students' ESL background, the regular

English curriculum and psychological assessment procedures are con-
bly more context-reduced and cognitively demanding than they are
f English-background students. As was pointed out earlier, research
suggest that it takes much longer for language minority students
o approach commonly accepted age /grade norms in context-reduced
Aspects of English proficiency (five to seven years on the average) than it
in context-embedded aspects (approximately two years on the
Hypothetical curves representing these data are presented in

Fire 3
LENGTH OF TIME REQUIRED TO ACHIEVE AGE-
APPROPRIATE LEVELS OF CONTEXT-EMBEDDED AND
CONTEXT-REDUCED COMMUNICATIVE PROFICIENCY

----- Native English-Speakers

ESL Learners

Context-Embedded Face-tu-Facc Context-Reduced (Academic)

Communicative Proficiency Communicative Proficiency

'Native-speakers also, of course, take much longer to levelop proficiency in processing

language in context-reduced situations.

29
 A Theoretical Framework 17

In summary, I have tried to show how certain misconceptions re-

garding the notion of language proficiency are currently contributing to
thevacadank failure of language minority students, To more adFquately
address the issue of the acquisition of English proficiency in bilingual
programs, a theoretical framework has been developed in which two con-
tinuums are distinguished. One deals with the range of contextual
supports for the communicative activity while the other is concerned with
the degree of active cognitive involiement in the activity. Literate
cultures typically require their members to become proficient in com-
municative activities which are context-reduced and cognitively damn-
ding (e.g., reading and writing). There tends to be large individual dif-
lemon both within and between socio-economic groups in the extent to
which this dimension of communicative proficiency is developed. ur in
thevemainder of this paper, the dimension of language which is strongly
related to literacy skills will be termed "context-reduced language profi-
ciency.""
In the next section, several theoretical distinctions similar to those
developed in the present framework are briefly discussed, in order to
further elaborate the characteristics of context-reduced language
proficiency.
Related Theoretical Frameworks
Several theorists interested primarily in the development of first
language academic skills have similarly argued f6r the necessity to
distinguish between the processing of language in informal everyday in-
terpersonal situations and the language processing required in most
academic situations (Denim' and Scardamalia, 1981; Donaldson, 1978;
Olson, 1977). In concrete terms, it is argued that reading a difficult text
or writing an essay make fundamentally different information processing

"1 Wells (1979), in a ten-year longitudinal study, has hie:Wiled two broad types of home
communicative activities that strongly predict the acquisition of reading skills in school.
One is the extent to which there is "negotiation of mousing" (i.e., quality and qualidty
of communication) between aulults and children, the other is the extent to which literacy-
Mated activities are promoted in the home, ca., reading to children). There is no clear-
cut relationship between ado-econousie status (SES) and the former, but a strong rela-
tionship between SES and the latter.
"In previous articles I have contrasted cognftivs/ academic language proficiency (CALP)
with basic interpersonal communicative skills (BICS) in order to make the same point;
namely, academic deficits are often created by teachens and psychWogists who fail to
realize that it takes language minority students considers* longer to attain grade/age-
appropriate levels in English academic skills than it does in English face-to-face com-
municative skills. However, because this distinction was not explicitly integrated Into a
more general theoretical framework, niisinterproation occurred. Hence, the attempt to
define such a framework in this paper.

30
1I Schooling and Language Minority Students:

den: n the individual compared to engaging in a casual conversa-

tion h a friend.
Em , and Disembedded Thought and Language. Donaldson
(1978) distinguishes between embedded and disembedded thought and
'language from a developmental perspective and is especially concerned
with the implications for children's adjustment to formal schooling. She
points out that young children's early thought processes and use of
language develop within a "flow of meaningful context" in which the
logic of words is subjugated to perception of the speaker's intentions and
salient features of the situation. Thus, children's (and adults') normal
productive speech is embedded within a context of fairly immediate
goals, intentions, and familiar patterns of events. However, thinking and
language, which move beyond the. bounds of meaningful inter-
personal context, make entirely different demands on the individual in
that it is necessary to focus on the linguistic forms themselves for mean-
ing rather than on the intentions.
Donaldson (1978) offers a re-interpretation of Piaget's theory of
cognitive development from this perspective and reviews a large body of
research that supports the distinction between embedded and disembed-
ded thought and language. Her description of pre-school children's com-
prehension and production of language in embedded contexts is especial-
ly relevant to current practices in language proficiency assessment in
bilingual programs. She points out that:
...the ease with which preschool children often seem to
understand what is said to them is misleading if we take it as
an indication of skill with language per se. Certainly they
commonly understand us, but surely it is not our words alone
that they are understandingfor they may be shown to be
relying heavily on cues of other kinds. (Donaldson.
1978, p. 72)
Donaldson goes on to argue that children's facility in producing
language that is meaningful and appropriate in interpersonal contexts
can also give a misleading impression of overall language proficiency:
When you produce language, you are in control: you need on-
ly talk about what you choose to talk about-Jae child) is
never required, when he is himself producing language, to go
counter to his own preferred reading of the situationto the
way in which he himself spontaneously sees it. But this is no
longer necessarily true when he becomes the listener. And it is
frequently not true when he is the listener in the formal situa-
tion of a psychological experiment or' indeed when he
becomes a learner at school. (1978, pp. 73-74)

31
 A Theoretical Framework 19

The practical implications of this view will be discussed in 'the context of

current assessment practices in bilingual education.
Utterance and Ten. Olson's (1177) distinction between "utterance"
and "text" relates to whether meaning is largely extrinsic to language
(utterance) or intrinsic to language (text). In interpersonal oral situa-
tions, the listener has access to a wide range of contextual and para-
linguistic Information with which to interpret the speaker's intentions;
and, in this sense, the meaning is only partially dependent upon the
specific linguistic forms used by the speaker. However, in contrast to ut-
terance, written text:
...is an autonomous representation of meaning. Ideally, the
printed reader depends on no cues other than linguistic cues;
it represents no intentions other than those represented in the
text; it is addressed to no one in particular; its author is essen-
tial& anonymous; and its meaning is pecisely that
represented by the sentence meaning. (Olson, 1977, p. 276)
Olson explicitly differentiates the t of the ability to process
text from the development of the tongue (utterance) in the pre-
school years:
But language development is no simply a matter of pro-
gressively elaborating the oral mother tongui as a means of
sharing intentions. The developmental hypothesis offered
here is that the ability to assn a meaning to the sentence per
se, independent of its nonlinguistic interpretive context, is
aarkved only well into the school years. (Olson, 1977, p. 275)
Conversation and Composition. Bereiter and Scardamalia (1981) have
analyzed the problems of learning to write as problems of converting a
language production system geared to conversation over to a language
production system.capable of functioning by itself. Their studies suggest
that some major difficulties involved in this process are the following: (1)
learning to continue producing language without prompting from con-
versational partners; (2) learning to search one's own memory instead of
having memories triggered by what other people say; (") planning large
units of discourse instead of only what will be said next; and (4) learning
to function as both sender and receiver, the latter function being
necessary for revision.
Bereiter and Scardamtdia (1980) argue that the absence of normal con-
versational supports makes writing a radically different kind of task
from conversation.
We are proposing instead that the oral language production
system cannot be carried over intact into written composition,
that it must, in some way, be reconstructed to function
autonomously in -,toad of interactively. (p. 3)

32
20 Schooling and Language Minority Students:

Although the distinctions between "embedded-disembedded,"

"utterance-text," and "conversation- composition" were developed in-
dependently and in relation to a different set of data, they share the
essential characteristics of the distinctions outlined in the present
theoretical framework. The major difference is that the failure of other
frameworks to distinguish explicitly between the cognitive and contex-
tual aspects of communicative activities might incorrectly suggest that
context-reduced communication (literate tradition) is *Pinsk& ly more
cognitively demanding than 'context-embedded communication (oral
tradition).
Having described in some detail the nature of the academic tasks
students encounter in school, it is now possible to discuss the develop-
ment of bilingual proficiency among language minority students within
this context.
The Nature of Bilingual Proficiency
The Myth of Di lingual Handicaps
The image of bilingualism as a negative force in children's develop-
ment was especially common in the early part of this century when most
teachers of language minority children saw bilingualism almost as a
disease that not only caused confusion in children's thinking but also
prevented them from becoming "good Americans." Therefore, they felt
that a pre-conditior for teaching children the school language was the
eradication of their bilingualism. Thus, children were often punished for
speaking their first language in school and were made to feel ashamed of
their own language and cultural background. It is not surprising that
research studies conducted during this period (Darcy, 1953) often found
that bilingual children did poorly at school, many experiencing emo-
tional conflicts. Children were made to feel that it was necessary to reject
the home culture in order to belong to the majority culture, often ending
up unable to identify fully with either cultural group.
However, rather than considering the possibility that the school's
treatment of minority children might be a cause of their failure, teachers,
researchers, and administrators seized on the obvious scapegoat and
blamed the children's bilingualism. The research findings were inter-
preted to mean that there is only so much space or capacity available in
our brains for language; therefore, if we divide that space between two
languages, neither language will develop properly and intellectual confu-
sion will result (Jensen, 1%2). Table 1 outlines the interplay between
socio-political and psycho-educational considerations in establishing the
mIth of bilingual handicaps and the role of "scientific studies" in
perpetuating it.

33
MN
A Theoretical Framework 21

The soda- political and psyche-educational assumptions illustrated in

Table 1 are very much in evidence in the education
debate. The popular press frequently warns thlifiltealducation will
lead to social fragmentation and Quebec-style separatist movements.
This fear of bilingual education is often rationalized in psycho-
educational terms; namely, that if minority children are deficient in
English, then they need instruction in English, not in their first language.
Table 1

BLAMING THE VICTIM IN MINORITY LANGUAGE EDUCATION'

A. Overt aim Cotten aim I D. Outcomes

Teach English to Anglicize minority Even more intense Tice failure of these
minority children children because efforts by the efforts only serves
in order to create a linguistic and cal- school to eradicate to reinforce the
harmonious society tural diversity are the deficiencies in. myth of minority
with equal oppor- seen as a threat to Went in minority group deficiencies.
tunity for all. social cohesion. children.

"Seient(le*
B. Method ifiAtialtiOlf C. Results explimation

Prohibit use of LI in 1. Lt shooki be III. Shame in LI I. Bilingualism

schools and make eradicated because language and causes confusion
children reject their it will Interfere culture. in thinking. ento-
own adture and with the learning donal insecurity.
lassguage in order to of English. 2. Replacement of and school
identify with major- LI by 1.2. failures.
ity English Pout). 2. Identification
with Ll culture will 3. School failure 2. Minority group
reduce child% abil- among many children are
ity to identify with children. "culturally de-
English- speaking by definition
culture. since they are not
Anita's).
3. Some minority
language groups
are genetically in
Fedor (common
theory in the
United States in
the I920s and
19304.

"This table reflects the assumptions of North American school systems in the first half of
this century. However, similar assumptions have been made about minority language
children in the school systems of many other countries.

34
 22 Schooling and Language Minority Students;

Consider, for example, the view expressed by Bothell (1979):

Bilingual education IS an idea that appeals to teachers of
Vanish and other tongues, but also to those who never did
think that another idea, the United States of America, was a
particularly good one to begin with, and that the sooner it is
restored to its compownt "ethnic" parts the better off we
shall all be. Such people have been welcomed with open arms
into the upper mocha of the federal government In recent
years, giving rise t* the suspicion of a death wish. (p. 30)
The psycho-educational argumentiramars later when Bethel (1979)
approvingly quotes Congressman John' Ashbrook's opposition to bi-
lingual education:
The program is actually preventing children from learning
English. Someday somebody is going to have to teach those
young people to speak English or else they are going to
become public charges. Our educational system is finding it
increasingly difficult today to teachinglish-veaking children
to read their own language. When. children come out of the
Spanish- hanguage schools or Choctaw- language schools
which call- themselves bilingual, how is our educational
system going to make them literate in what will still be a com-
pktely alien tongue...? (pp. 32-33)
The argument that' defieiencies in English should be remediated by in-
tensive instruction in English appears at first sight much more intuitively
appealing, than the alternative argument that instruction in LI will be
more effective than instruction in English in promoting English skills.
This latter argument appears to invoke a "less equals more" type of logic
that is unlikely to convince skeptics. In order to evaluate these alternative
positions, it is necessary to make their propositions more explicit and
make empirical evidence rather than "common sense" the criterion of
validity. The issues revolve around two alternative conception% of bi-
lingual proficiency, termed the Separate Underlying Proficiency (SUP)
and Common Underlying Proficiency (CUP) models.
The SUP and CUP Models of ileingnal Proficiency
The argument that if minority children are deficient in English, then
they need instruction in English, not in their Li, implies: (a) that profi-
ciency in L I is separate from proficiency in English, and (t) that there is
a dirett relationship between exposure to a language (in home or school)
and achievement in that language. The SUP model is illustrated in Figure
4.

35 fr
 A Theoretical Framework 23

Flame 4
THE SEPARATE UNDERLYING PROFICIENCY (SUP) MODEL
OF BILINGUAL PROFICIENCY

0. The second implicatjon of the SUP model follows from the first, that
if LI and L2 proficiency are separate, then content and skills learned
through LI cannot transfer to L2 and vice versa. In terms of the balloon
metaphor illustrated in Figure 4, blowing into the LI balloon will succeed
in inflating LI but not L2. When bilingual education is approached with
these "common-sense" assumptions about bilingual proficiency, it is not
at all surprising that it appears illogical to argue that one can better in-
flate the L2 balloon by blowing into the LI balloon.
However, despite its intuitive appeal, there is not one shred of evidence
to support the SUP model."' In order to account for the evidence re-
viewed, we must posit a CUP model in which the literacy-related aspects
la Macnantara (1970) points out that a strict interpretation of a SUP model would leave the
bilingual in a curious predicament in that "...he would have great diffk-uhy in 'com-
municating' with himself. Whenever he switched languages he would have difficulty in
explaining in L2 what he had heard or said in LI" (pp. 25-26). It is not surprising that
the SUP model is not seriously proposed by any researcher. Nevertheless, it is important
to examine 'the research evidence in relation to this model, since many educators and
policy-makers espouse positions in regard to bilingual education which derive directly
from this Implicit model.

36
24 Schooling and Language ivlinoiity Students:

of a bilingual's proficiency in LI and L2 are seen as common or in-

terdependent across languages. Two ways of illustrating the CUP model
(the Interdependence Hypothesis) are shown in Figures 5 and 6.
Figure 5
THE COMMON UNDERLYING PROFICIENCY MODEL (CUP)
OF BILINGUAL PROFICIENCY

Common Underlying
Proficiency

LI
Channel 12
Channel

Figure 6
THE 4. DU ALu-IC REPRESENTATION OF
BILINGUALG

Surface Features Surface Features

of I I of L2

Common Underlying
Proficiency

37
 A Theoretical Framework 25

Figure S expresses the point that experience with either language

can promote development of the proficiency underlying both
languages, given adequate motivation and exposure to both
either in school or in the wider environment. In Figure 6 bilingual profi-
ciency is represented by means of a "dual iceberg" in which common
cross-lingual proficiencies underlie the obviously different surface
manifestations of each language. In genval the surface features of LI
and L2 are those that have become rklatively automatized or less
cognitively demanding whereas the underlying proficiency is that in-
yolved in cognitively demanding communicative tasks."
There are five major sources of evidence for the CUP model: (I)
results of bilingual education programs, (2) studies relating age on arrival
and immigrant students' L2 acquisition, (3) studies relating bilingual
language use in the home to academic achievenient, (4) studies of the
relationship between LI and 1.2 prtifidency, and (5) experimental studies
of bilingual information processing. The first three sources will be con-
sidered in more detail than the latter two because of their direct relevance
to current concerns of bilingual educators in the United States'.
Evaluations of Ilignsuid Progranis
Although there is a widespread perception that bilingual educa-
tion has yet to its effectiveness (Trombley, 1980), findings of the
available, ntrolled research are strongly supportive of the basic
principle underlying bilingual education, i.e., the CUP model of bi -
lingual proficiency. For example, Troike (19711) reviewed 12 evaluations
and several research studies in which bilingual instruction was found to
be more effective than English-only instruction in promoting English
academic skills. Two of these evaluations are outlined here as well as
several other evaluations in the United States and elsewhere that clearly
relute the SUP model.
Rock Point Navajo Study. Before the bilingual program was started
in 1971, Children were two years behind United States norms in
"The data used to support the CUP model primarily involve "commit-reduced language
proficiency" because the model Is developed explicitly in retainer to the development of
bilingual academic skills. It is probable, however, that many aspects of "context-
embedded language proficiency" may also be interdependent across languages. As far as
context-reduced language proficiency Is concerned, the transfemiellltY ACratelangualles
of the proficiencies Involved fe reads* (e4. inferring and predicting meaning based on
sanspling from the test) sad inking (e.g. planning lane chunks of &mune) b obvienni.
However, even where the task &meads are langssage specific fr. g., decoding or
sprains), a strong reiftionddp may be obtained between skills fa LI and 1.2 as a result of
a mcwe meridiem! proftchincy (and motivation) to handle coenitively demanding
content-reduced language task& Similarly, on the camext-engsedded side, many sacks-
linguistic rules of face -to -face commualcatkin are language-specific, but I. I and Li
sockginerlisdc skills may be related as a result of a possible gem:razed sensitivity to
sociolinguistic.rules of discoune.

e 30
26 Schooling and Language Minority Students:

English reading by the end of sixth grade despite intensive teaching of

English as a second language. The bilingual program used Navajo
as the major initial medium of instruction and continued its use
throughout elementary school. English reading instruction was delayed
until Navajo reading skills were well established (mid-second grade). By
the end of the sixth grade, children in the bilingual program
were performing slightly above United States grad norms in English
readiitg despite considerably less exposure to English than previously
(Rosier and Fertile, 1976).
Santa Fe Bilingual Program. In the schools involved in this program,
Spanish was used for between 30 and 30 percent of the school day
throughout elementary school. It was found that children enrolled in the
bilingual program consistently performed significantly better than the
control group (in an English-only program) in both reading and
mathematics. Children enrolled continuously in the bilingual program
from second grade caught up with United States norms in English
reading by fifth grade and stayed close in sixth grade. In math this group
surpassed the national average in fourth grille and maintained an equal
or superior status through sixth grade (Leyba, 1978).
Legarreta Study: Direct ESL-Bilingual Comparison. 4 study carried
out by Legarreta (1979) in California compared the effectiveness of three
types of bilingual treatments with two types of English-only treatments
in facilitating the development of English communicative competence in
Spanish,background kindergarten children. The three bilingual
treatments were found to be significantly superior to the two English-
only treatments in developing English language skills. The most effective
program was one with balanced bilingual usage (50 percent English, 50
percent Spanish).
Nestor School Bilingual Program Evaluation. The Nestor program in
San Diego involved both Spanish- and English-background students and
used a team teaching approach in which instruction in the early grades
was pridierily through the children's LI , The proportion of instruction in
Limas gradually increased until, by fourth grade, approximately 50 per-
cent of instruction was through each language. The evaluation of the
program (Evaluation Associates, 1978) showed that Spanish-background
students gained an additional .36 of a year's growth in English reading
for each successive year they spent in the bilingual program. Spanish-
background students who had spent five years or more in the bilingual
program at the elementary level tended to perform slightly better in
English reading.than the school average at the junior high school level,
despite the fact that at least 37 percent of the comparison group were
originally native English speakers. In mathematics, the sixth grade

39.
 A Theoretical Framework 27

Spanish-background children in the program were over a year ahead of

the Spanish speakers in the comparison district and only one month
behind grade level. The English-background participants in the Nagor
bilingual program performed at a higher levee than the comparison
groups on a large majority of measures; however, this may be due to a
selection bias.
The Colorado Bilingual PWg,umr Evaluation. Egan and Goldsmith
(1981) and Egan (1981) report on the "overwhelming success" of bi-
lingual programs in Colorado for both language minority and Anglo
students. Over 90 percent of the 39 programs for which data were
available reported that "limited-English-proficient" students showed a
rate of academic progress at least as good as that normally expected for
all students. More surprising, however, was the fact that 50 percent of
the programs showed growth rates in English academic skills for
language minority students well beyond the normal expected growth
rates for all students. These results are especially significant in view of
previous research in Colorado (Egan and Goldsmith, 1981) showing that
411ispanic students tended to fall progressively further behind grade
norms during the elementary school years.
Sodertatie Program for Finnish Immigrant Children in Sweden. The
findings of this evaluation are very similar to those of the Rock Point
Navajo evaluation. Finnish children in Swedish-only prognuni were
afbund to perform worse in Finnish than 90 percent of equivalent socio-
economic status Finnish children in Finland and worse in Swedish than
about 90 percent of Swedish children (Skutnabb-Kangas and Toukomaa,
1976). The Sodertalje prokram, however, used Finnish as the major in-
itial language of instruction and continued its use throughout dealt:Mary
school. Swedish became the major language of instruction from third
grade. By sixth grade, children's performances in this program in both
Finnish and Swedish were almost at the same level as that of Swedish-
speaking children in Finland, a considerable improvement in both
languages compared to peir performances in Swedish-only programs
(Hanson, 1979).
Manitoba-Francophone Study. A large-scale study carried out by
Hebert et al. (1976) alhong third, sixth. and ninth grades, in. which
minority fmncophorte students in Manitoba °were receiving varying
mounts of instruction through the medium of French, found that
the amount of French-medium instruction showed no relationship to
children's achievement in English. In other words, francophonetstudents
receiving 80 percent instruction in French and 20 percent instruction in
English did just as well in English as students receiving 80 percent in-
struction in English and 20 percent instruction in French. However,

40
28 Schooling and Language Minority Students;

amount of instruction in French was positively related to achievement in

French. In other words, students' French benefited at no cost to their
progress in English.
Edmonton Ukrainian-English Bilingual Program. This program has
existed in eight Edmonton elementary schools since 1972 and is financial-
ly supported by the Alberta government. In 1978-1979 there were 697
students enrolled between kindergarten and fifth grades. Ukrainian is
used as a medium of instruction for 50 percent of the regular school day
throughout elementary school. Only about 15 percent of the students are
fluent in Ukrainian on entry to the program. A study carried out with
first and third grade students (Cummins and Mulcahy, 1978) found that
students who were relatively fluent in Ukrainian as a result of parents
using it consistently in the home were significantly better able to detect
ambiguities in English sentence structure than either equivalent mono-
lingual English-speaking children not in the program or children in the
program who came from predominantly English-speaking homes. The
evaluations of the program have shown no detrimental effects on the
development of children's English or other acackmic skills. In fact, by
the end of fifth grade children in the program had pulled ahead of the
comparison group in English reading comprehension skills (Edmonton
Public School Board, 1979).
In summary, the results of research on bilingual pro:warns show that
minority children's LI can be promoted in school at no cost to the
development of proficiency in the majority language. b other words, the
educational argument against bilingual education Is invalid; in order to
explain the findings, it is necessary to posit a common proficiency dimen-
sion that underlies the development of academic skills in both languages.
The data clearly show that well- implemented bilingual programs have
had remarkable success in developing English academic skills and hive
proved superior to ESL-only programs in situations where direct com-
parisons have been carried out.
How do we reconcile the success of L1- medium programs for minority
children with the fact that majority language children fare very well
acaclanica4 in French or Spanish immersion programs (Cummins,
1979b; Swain, 1978)7" There are many differences between these
situations, e.g., prestige of Ll, security of children's identity and self-
concept, and level of support for LI development in home and environ-

"A French immersion . program involves teaching students from English home
backgrounds through the medium of French for a major part of the school day from
kindergarten through high school. The goal is bilingualism in French and English. These
programs are now extremely commix' in Canada_ and evaluations show that students
gain high levelsof French proficiency at no cost to proficiency in English (Swain, 19711).

41
 A Theoretical Framework 29

ment. Thus, it is not surprising that different forms of educational pro-

grams should be appropriate for children with very different background
characteristics. The apparent contradiction between findings in minority
and majority contexts completely disappears when we stop thinking in
terms of "linguistic mismatch" or "home-school language switch." In
immersion programs for majority language children, as well as in bi-
lingual programs for minority children, instruction through the minority
language has been effective in promoting proficiency in both languages.
These findings, which have been replicated in an enormous number of
studies, support the following "Interdependence" Hypothesis: To the
event that instruction in Lx is effective in promoting proficiency in Lx,
trumfer of this proficiency to Ly will occur provided there is adequate
exposure to Ly (either in school or environment) and adequate Milli %I-
lion to learn Ly. In other words, far from being contradictory, th. -ante
theoretical principle, the CUP model, underlies immersion programs for
majority language students as well as bilingual programs for language
minority students.
Age on Arrival and L2 Acquisition
It would be predicted on the basis of the Interdependence Hypothesis
that older learners who are more cognitively mature and whose LI profi-
ciency is better developed would acquire cognitively demanding aspects
of L2 proficiency more rapidly than younpr learners. Recent reviews of
research on the age issue confirm this prediction (Cummins. !Ma;
Cummins, 1981; Ekstrand, 1977; Genesee, 1978; ICrashen et al., 1979).
The only area where research suggests older learners may not have an ad-
vantage is pronunciation, witichl, significantly, appears to be one of the
least cognitively demanding aspects of both Li and L2 proficiency. In
terms of the model presented in Figure 3, we would expect the advantage
of older learners to be especially apparent in context-reduced aspects of
9 L2 proficiency because of their greater amount of experience in process-
ing context-reduced aspects of Lt.
The extent of the advantage older learners have in acquiring context-
reduced cognitively demanding aspects of L2 is illustrated by the data in
Figure 7. The test, a group adaptation of the Ammoris Picture
Vocabulary Test (Ramsey and Wright, 1972), and subjects (1,210 fifth,
seventh, and ninth grade immigrant students in the Toronto Board of
Education) are the same as in Figure 1. However, the data are presented
in terms of absolute scores on the test rather than in terms of grade
norms. In Figure 1, older and younger L2 learners appeared to approach
grade norms at a generally comparable rate. However, because older
learners have further to go in order to reach grade-appropriate levels of

42.
 30 Schooling and Language Minority Students:

Figure 7

AGE ON ARRIVAL, LENGTH OF RESIDENCE, AND PVT RAW

SCORES

PVT

40

35

010
..-------
...--/
AOA 0-1 2-3 4-5 6-7 8-9 10-11 12-13 14-15

L2 academic proficiency (consider, for example, the difference between

the vocabulary knowledge of a twelve and six-year-old monolingual
English child), we would expect them to acquire more L2 than younger
learners b absolute terms in the same amount of time. This is clearly the
case in Figure 7. In this study, it was possible to make 90 comparisons
between older and younger learners on context-reduced cognitively
demanding aspects of L2. In 89 of these, older learners performed
better.'5

"it may appear surprisi that older learners make more rapid progress in acquiring L2 in
view of the popular myth that there is an optimal pre-puStrtal age for L2 acquisi-
don. However, a jor reason for the advantage is obvious when the data are viewed
from within the of the CUP model. For example, in learning the term
"democracy" the for a 14-year-old immigrant child consists of acquiring a new
label for a concept y developed in Li; for a 6-year-old immigrant child the term
will not be acquired until the conorpt has been developed. The advantage of older
learners lies in the interdependence of conceptual knowledge across languages.

43
 A Theoretical Framework - 31

The relationship between Li and L2 proficiency in immigrant students

was explicitly investigated in two studies. Cummins et al. (1981) reporte
that olden. Japanese immigrant students, whose Li literacy skills were
better developed, acquired English proficiency significantly faster than
younger immigrant students. It was also found that students who im-
migrated at younger ages developed significantly lower proficiency in
Japanese compared to students who immigrated at older ages and who
had been in Canada for the same amount of time. All the students in this
study were from upper-class backgrounds.
Skutnabb-Kangas and Toukomaa (1976) also report that among Fin-
nish immigrant children in Sweden, the extent to which LI had been
developed prior to contact with Swedish was strongly related to how well
Swedish was learned. Children who migrated at age 10-12 maintained a
level of Finnish close to Finnish students in Finland and achieved
Swedish language skills comparable to those of Swedes. By contrast,
children who migrated at younger age levels or who were born in Sweden
tended to reach a developmental plateau at a low level in both Finnish
and Swedish academic proficiency.
Consistent with the Skutnabb- Kangas and Toukomaa findings (1976),
there is considerable anecdotal evidence that immigrant students from
Mexico fare better educationally than native-born Mexican - Americans.
For example, Troike (1978) stated that:
It is a common experience that... children who immigrate to
the United States...qfter grade six... rather quickly acquire
English and soon out-perform Chicano students who have
been in United States schodis since grade one. (p. 15)
Based on a survey of school personnel in four southwestern states, Caner
(1970) similirly reported that many teachers and administrators believe
that older immigrant students achieved better than native-born Chicano
students. '41
In summary, considerable research supports prediction derived
from the Interdependence Hypothesis that older immigrant children
"Two empirical studies (Kimball, 1968; Anderson and Johnson, 1971) support these
teacher perceptions. However, a recent study (Banal, 1979) reports that immigrant
students who had had at least`two years of schooling in Mexico performed significantly
lower in Academic skills than naive -born Mexican students. Two factors are important
in interpreting these results: first, the immkarant students came front significantly lower
socio-etonomic backgrounds than the native-born students: second, they had been in
United States schools only between two and five years. The Canadian findings Ceported
earlier (Cummins, 1931) mistiest that it can take up to seven years for immigrant students
to approach Brink non= in English academic skills. Students who were in Canada for
three years were still approximately one standard deviation below grade norms. Thus,
the relatively short length of residence and the socio-ecanomic differences between im-
migrant and native-born students can account for Baral's (1979) findings.

44
32 Schooling and Language Minority Students:

make more rapid progress than younger children in acquiring 1.2 profi-
ciency. It should be noted that these relationships between LI and
L2 do not operate in a sociocultural vacuum. ilfhe role of sociocultural
factors in relation to cognitive and linguistic factors will be considered in
a later section.
Primary Leagues*, Development hs the Home
Several studies show that the use of a minority language in the home is
not a handicap to children's academic progress." This was evident in the
Cummins and Mulcahy (1978) study of the Ukrainian bilingual program
where first and third grade students who used Ukrainian consistently in
the home were better able to detect ambiguities in English sentence struc-
ture. Two other studies (Bhatnagtr, 198U; Chesarek, 1981) suggest that,
under certain conditions, a switch to the use of the majority language in
the home is associated with poor academic progress in the majority
langua8c.
Chesarek (1981) carried out a long?udinal study among elementary
students on a Crow, reservation in Montana in which he identified a sub-
group of students who had one or more Crow-speaking parents bat were
raised as Enflish speakers. This group of students scored significantly
lower on a non-verbal ability test at school entry than either native Crow-
speaking children or English-speaking children of two English-speaking
parents. In a longitudinal follow-up at third grade in one of the reserva-
tion schools that utilized a bilingual instructional program, it was found
that this group performed worse on several aspects of English achieve-
ment than the native Crow-speaking-group." Chesarek (1981) sums up
these findings as follows:
In other words, children who had only three years exposure to
English in a bilingual program context were surpassing
children for whom English was the only language. (p. 14)
A very similar pattern of findings emerges from a recent study carried
out by Bhatnager (1980) in Montreal, Canada. in this study, the

"In addition to the studies considered in the text, studies carried out by Carry and Cum-
mins (1979), Ramirez and Pulitzer (1976), and Yee and La Forge (1974) with minority
francophone, Hispanic, and Chinese stet s, respectively, show that, In itself, the use
of a minority LI in the home is not an impediment to the acquisition of 1.2 acadendc
shills In school. These findings, of course, create problems of de "Bevis* mismatch"
rationale for bilingual education, namely, the minority students fail in school because
their home language is diffeent from that of the school.
lochesarek (1991) points out there was very little bilingual activity in the classroom since
the major efforts were being devoted to developing an onhography and teaching
materials as well as training Rees to assume instructional activities.

45
A. Theoretical Framework 33

academic progress of 171 Italian immigrant children in English language

elementary schools and 102 in Rem* language schools was examined in
relation to language spoken at home and with friends and siblings. Bhat-
nager sums up his findings as follows:
The results reported here do not support the popular assump-
tion that the more immigrant children speak the local
language the better their adjustment to the host culture. It is
interesting to note that immigrant chikIrtm who used Italian
and a Camxlian language interchangeabb, were better even at
English or French, of both the spoken and written variety,
than children who used English or Fl-ench all the
time....Language retentkm...should lead to higher academic
adjustment, better facility in the host language, and better
social rektions of immigrant Chi en. (1980, pp. 153-155)"
In all these instances, the SUP model would have predicted that
students exposed exclusively to the majority language at home would
perform better than students who used a minority language at home.
This prediction receives no support from the research findings; instead,
the research sum the predicdon derived from the CUP model, that
experience with either language is capable of promoting the proficiency
that underlies the development of academic skills in both languages.
Thus, whether English or a minority hinguage is used in the home is, in
itself, relatively unimportant for studesne academic development. As
Wells' (1979) study has shown, what is inspmant for future academic
success is the quality of interaction children experience with adults.
Viewed from this perspective, encouraging minority parents to com-
municate in English with their children in the home can have very detri-
mental consequences. If parents are not conifortable In English, the
quality of their interaction with their children in English is likely to be
Is than in LI. Thus, the lower academic achievement of minority
children who used L2 exclusively with their parents and friends in Bhat-
nagar's (1980) and Chesarek's (1981) studies may be httributable to the
lower quality of communication their parents were capable of providing
in their second language.a°
'ilibatnager OM) reports that imm*rant students who used Li exclusively with parents
and siblings also palkomed siwdficantly worse than those who used both LI and
However, it Reins Wm* that A's finding can be attributed to the fact that only
those students who had bmnigtated rdodvdy recently would use LI exclusively.
Length of !IMMO= Is not coasideseel in nemeses study, but the data in F I
suggest that It takes lit:mignon students at Iciest five Imam to approach grade norms in L2 ty

academie skills.
Data from two other SCUMS also support the CUP model. Theses are corxelitiopal
studies of the relationship between LI sod 1.2 proficiency and everimenial studies of bi-
Ultra, information processing.

46
 34 Schooling and Language Minority Students:

in summary, the research fladings from evaluations of bilingual pro-

grams, studies of immigrant aildren's academic progress, and studies
that examined the consequences of different patterns of home language
use, are consistent with predictions derived from the CUP model.
However, the observed relationships between LI and L2 do not operate
indepemlently of the sociocultural context. In the next section the role of
sociocultural factors in determining minority students' academic
development is considered.
Sociocultural Determinants of Minority Students' AL-Wyman
Linguistic, cognitive, or educational factors by themselves cannot ac-
count for the school failure of minority students because there are large
individual and group differeices, in academic' achievement of minority
students exposed to the same educational conditions (e.g., home-school
language switch). Consider, for example, the fact that immigrant
students who arrived in Canada before age six achieved grade norms in
L2 academic skills (see Figure 1), whereas Finnish students who im-
migrated to Sweden at an early age attained only a low level plateau in
Swedish academic skills. This latter pattern also appears to characterize
Hispanic students who immigrate at an early age or who are born in the
United States.
What sociocultural factors account for this pattern of deferential
achievement by minority students in different contexts7Socio-econornic
status (SES) cannot account for4the differences because all groups were
low SES. Acculturation, or the degree to which minority students adopt
the language and cultural values of the majority, likewise fails to account
for the data. If acculturation were the major factor at work, we would
expect those minorky students who used only English at home to per-
form better academically than those who maintained the use of LI at
home. In fact, as the studies by Chesarek (1981) and Rhatnagar (1980)
demonstrate, such "acculturated" students often (but not necessarily
Many studies have shown Mghly signifiaun ocarelailons between LI and L2 proficiency
(Cununins, 1979a) and it haw been reported that Sganish reading proficiency developed in
a bilingual program is the mast stable predictor of English reading proficiency levels
students develop after transferring; from the bilingual prose= (Fischer aqd Cabello.
1978).
Experimental studies of bilingual information ptercessh;g have condstendy shown that
bilinguals process semantic memory information in the maw way in their two languages
and in the same way as monolinguals (Quantum and Stelae*, 1980 Enrique*, ISSO;
Kolas, 190; Landry, ISM 19110; McCormack, 1974). In other words, Mammals have
only one senses* memory system that can be accessed via two languages. The studies
cited above have been carried out with adult bilinguals; however, a recent study (Chit-
Chana, 1981) carried out with Chinese elenumtary school students has reported simfhr
results. She condddm that, at the input and conceptual level, the two languages of the
bilingual are in one storage.

47
 A Theoretical Framework 35

always) show lower levels of English academic achievement than students

who continue to use their LI at home and maintain their allegiance to the
home odium."
An examination of the sociocultural characteristics of minority groups
that tend to perform poorly in L2-only school situations suggests that the
attitudes of these groups towards, their own identity may be an important
factor in interaction with educational treatment. Specifically, groups
such as Finns in Sweden, North American Indians, Spanish- speakers in
the U.S., and Franco-Ontarians in Canads all tend to have ambivalent or
negative feelings towards the majority culture and often also towards
their own culture. This pattern has been Marti documented for Finnish
immigrants in Sweden by Skutnabb-Kangas and Toukomaa (1976). For
example, Heyman (1973) concludes:
Many Finns in Sweden feel an fiVerSiOn, and sometimes even
hostility, towards the Swedish knguage and refuse to learn
it... under protest. There Lc re/awed evidence of this, as there
is, on the other hand, of Finnish Fectalechildren and
adultswho are ashamed of their Finnish language and do
not allow it to live and develop. (p. 131)
The same pattern of ambivalence or hostility towards the majority
cultural group and insecurity about one's own language and culture is
found, to a greater or lesser extent, in other minority groups that have
tended to perform poorly in school. For itumple, many Franco-
Ontarians tend to regard their own dialect of French as inferior and to
show low aspirations for social and atomic mobility in the majority
anglophone culture. In contrast, minority groups that do well in school
and to be highly motivated to learn the majority language and often
(though not always) have a strong sense of pride in their own cultural
backgrounds.
According to this interpretation, part of the reason bilingual education
is successful in promoting minority students' academic progress is that
by validating the cultural identity of the students (as well as that of the -
community), it reduces their ambivalence towards the majority language
and culture. Older immigrant students often fare better than minority
students born in the lost country because they have not been subject to
the same ambivalence towards both cultural groups in their pm-school
and early school years and, hence, approach the task of learning L2 with
a secure identity and academic self-concept. Similarly, the exclusive use
of L2 rather than LI in the home is likely both to reflect and contribute
to minority students' ambivalence towards L2.
' "I am grateful to Steve Chesarek for pointing this out to Inc.

48
 36 Schooling and Language Minority Students:

Clearly, at this stage, these suggestions in regard to the operation of

",bicultural ambivalence" ay.. speculative. However, they appear to ac-
count for the data better than a simple "acculturation" explanation and
also provide the basis for a more adequate rationale for bilingual educa:
tion than "linguistic mismatch" between home and school.
How does the operation of sociocultural factors relate to the linguistic
factors (e.g., interdependence between LI and L2) described earlier? The
development of communicative proficiency in Li and II can be regarded
as an intervening variable mediating the effects of the sociocultural con-
text on achievement. For example, sociocultural factors are likely to af-
fect patterns of parent/child interaction that will influence the develop-
ment of communicative proficiency (as described in Figure 2) in LI
and/or 1.1 that will, in turn, influence children's ability to benefit from
instruction. Thus, if parents are ambivalent about the value of their
cultural background or fed that they speak an inferior dialect of LI , they
may not strongly encourage children to develop LI skills in the home.
They may tolerate (or even encourage) children to watch television far a
considerable portion of the day on the grounds that this will help them to
learn English and do well at school. This attitude may be encouraged by
some teachers who believe that children should be exposed to as little LI
as possible.
Compare this situation to that of language minority parents who feel a
strong sense of pride In their cultural background and are eager to
transmit this cultural heritage to their children. They are likely to spend
more time " "negotiating meaning" (in LI) with their children, which ac-
cording to Wells' (1999) findings, is a strong predictor of future
academic success. If we assume that those aspects of communicative pro-
ficiemy most relevant to academic success develop largely as a result of
quality and quantity of communication with adults, then children in the
second situation will come to school better prepared to handle the
context-reduced communicative demands of school than children in the
first situation, despite the fact that they may know little or no English
(Chesarek, 1981). As the research reviewed in the context of the CUP
model clearly shows, communicative proficiency already developed in LI
can readily be transferred to L2, given motivation to learn L2 and ex-
posure to L2.
How do school programs interact with sociocultural and linguistic fac-
tors? As outlined in Table I, schools have contributed directly to minori-
ty children's academic difficulties by undermining their cultural identity,
attempting to eradicate their L 1, and exposing them to incomprehensible
context-reduced input in English. Recent evaluations of bilingual educe-

49
A Theoretical Framework 37

tion, however, have shown that when schools reinforce minority

children's cultural identity, promote the development of the L1 com-
municative proficiency children bring to school, and make instruction in
English comprehensibk, by embedding it in a context that is meaningful
in relation to students' previous experience, then minority students mi-
mic= academic success and develop high English literacy skills, in
spite of sodoculnual impediments.
In summary, although both sociocultural and educational factors con-
tribute directly to the development of communicative proficiency in
minority students, a large majority of academic and communicative
de hits (e.g., low reading achievement) are developed in these students
only as a result of failure by educators to respond appropriately to the
sociocultural and communicative characteristics children. bring to school.
In this section, bilingual communicative proficiawy has been con-
sidered as a dependent variable in relation to sociocultural and educa-
tional factors. Bilingual communicative proficiency can also be regarded
as an intervening variable, which in turn influences the further develop-
ment of cognitive and academic skills. In other words, how do different
patterns of bilingual proficiency influence students' ability to benefit
from interaction with their scholastic environment? This issue is con-
sidered in the next section.
Bifingual Pro th:Wm as Educational Enrichment: The Threshold
Hypothesis
It was pointed out in a previous section that because bilingual children
performed more poorly than monolingual children on a variety of
verbal-academic tasks in early studies, bilingualism was often regarded
as a cause of language handicaps and cognitive confusion. However,
more recent findings refute this interpretation. A large number of studies
have reported that bilingual children are more cognitively flexible in cer-
tain respects and bear able to analyze linguistic meaning than are
monolingual children (Cummins, 1979b). Albert' and Obler (1978) con-
clude on the basis of neuropsychological research findings that:
Bilinguals mature earlier than monolinguals both In terms of
cerebral lateralization for language and in acquiring skills for
linguistic abstraction. Bilinguals have better developed
auditory language skills than monolinguals. but there Li no
clear evidence that they differ from monolinguals in written
skills. (p. 248)
These findings are not at all surprising when one considers that
bilingual children have been exposed to considerably more "training"
in analyzing and interpreting language than monolingyal children.

50
 38 Schooling and Language Minority Students:

The greater analytic orientation to language of bilingual children is con-

sistent with the view of Viaotskii (1962), who argues that being able
to express the same thought in different languages will enable the child
to "see his language as one particular System among many, to view
its phenomena under more general categories, and this leads
to awareness of his linguistic operations" (p. 110). Lambert and
'nicker (1972) argued that a similar process was likely to ?mate among
children in bilingual programs. They suggested that, as children develop
high level bilingual skills, they an likely to practice a form of "incipient
contrastive linguistics" by comparing the syntax and vocabulary of their
two languages.
How do we resolve the apparent inconsistency that bilingualism is
associated with both positive and negative coitnitive and academic ef-
fects? An analysis of the characteristics of subjects in these two types of
studies suggests that the level of bilingualism children attain is an impor-
tant factor in mediating the effects of bilingualism on their educational
development (Cummins, 1979b). Specifically, a large majority of the
"negative" studies were carried out with language minority children
whose LI was gradually being replaced by a more dominant and
prestigious L2. Under these conditions, these children developed relative-
ly low levels of academic proficiency in both languages. In contrast, the
majority of studies that have reported coanitiVe advantages associated
o with bilingualism have involved students whose LI proficiency has con-
tinued to develop while L2 is being acquired. Consequently, these
students have been characterized by relatively high levels of proficiency
in both languages.
These data have led to the hypothesis that there may be threshold
levels of linguistic proficiency bilingual children must attain in order to
avoid cognitive deficits and allow the potentially beneficial aspects of
becoming bilingual to influence cognitive growth. The Threshold
Hypothesis assumes that those aspects of bilingualism that might
positively influence cognitive growth are unlikely to come into effect un-
til children have attained a certain minimum or threshold level of profi-
ciency in the second language. Similarly, if bilingual children attain only
a very low level of proficiency in one or both of theirlanguagefz, their in-
teraction with the environment through these Itutguages both in terms of
input and output, is likely to be impoverished.
The form of the Threshold Hypothesis that seems to be most consis-
tent with the available data is that there are two thresholds (Cummins,
1976; Toukomaa and Skutnabh-lCangas, 1977). The attainment of a
lower threshold level of bilingual proficiency would be sufficient to avoid

51
 A Theoretical Frames lurk 39

any negative cognitive effects; but the attainment of a second, higher

level of bilingual proficiency might be necessary to lead to accelerated
cognitive growth. The Threshold Hypothesis is illustrated in Figure 8.
Since this hypotheils was originally formulated (tummins, 1976),
several studies have reported findings consistent with its general tenets
(Cummins and Mulcahy, 1978; Duncan and DeAvila, 1979; Kessler and
Quinn, 1980). Duncan and De Avila (1979), for example, found that
language minority students who had &wok:wed high levels of LI aid
.proficiency (proficient bilinguals) performed significantly better than
monolingnals and other sub-groups of bilinguals (partial and limited bi-
linguals) on a battery of cognitive tasks. Kessler and Quinn (1980) found
that Hispanic bilingual students who had been in a bilingtial program
performed significantly better than monolinguals on a science problem-
solving task, while Cummins and Mulcahy (1978) found that Ukrainian-
English bilingual students who spoke Ukrainian at home and received 50
percent instnicdon through Ukrainian were better able to detect am-
biguities in English sentence structure than were monolingual English-
speaking students.
Figure
COGNITTVE EFFECTS OF immiewr TYPES OF BILINGUALISM*
1)pe off Bitfrigudism Clognithe Meer

Thetfinient bifingsonnn Positive

High ievds in both cognitive
111121Pones effects

Higher ducilsold
B. Avast biktswasksm Neitinr positive lewd of bilingual
Native he level in = negative proficiency
o®e of tbe bloWsne5 cognitive effects

Lona thiaihold
level of bilingual
C. Lintieed live
cognitive effects
pendency
Low level in both
beguolie4 OW be
bahmoed err dominant)

'Adapted non Tonkin= and Slannabb-Kangas, 1977, p.

52
 Schooling and Language Minority Students:
w

In summary, far from impeding English language and general

academic skills development, as the SUP model would predict, bilingual
Mangold! appears to offer students apotentially enriching educational
environment. For language minority students, this potential appears to
be realized only when their LI continues to develop as they are acquiring

Andication of liteareticat Analysis to Bilingual Education

In this section, the implications for bilingual education of the research
and theory outlined earliir will be made explicit. The four major implica-
tions relate to the rationale for bilingual education, entry criteria,
reclassification and exit criteria, and assessment considerations.
The Rationale for Bilingual Edinition
The failure of 1.2-only programs to protnote 1.2 literacy skills effec-
tively among some groups of language minority children was interpreted
by many academics as support for the hypothesis that mismatch between
the language of home and language of school Is a major cause of
academic retardation among minority children (Downing, 197$;
UNESCO, 1953; United States Commission on Civil Rights, 1975). This
Linguistic Mismatch Hypothesis is exemplified in the well-known
UNESCO statement that "it is axiomatic that the best medium for
teaching a child is his =they tongue" (UNESCO, 1953, p. 11).
The Linguistic Mismatch Hypothesis has come to be the main
theoretical rationale for bilingual education in the United.States. This is
unfortunate because it greatly over-simplifies the complexity of the issues
and as a general principle has title validity. The success of majority
languaw students in Fixmch immersion programs and of some minority
children in 12-only programs show clearly that "linguistic mismatch"
has limited explanatory power.
The transitional form of bilingual education operating in most states
derives directly from the linguistic mismatch hypothesis. The focus on in-
itial mismatch between the "visible" surface forms of LI and L2 implies
that children can be switched to an English-only program when they have
acquired basic fluency in English. Thus, in mo, transitional programs,
the role of LI instruction in developing English academic proficiency is
inadequately understood. LI is viewed only as an interim carrier of sub-
ject matter content until 12 can take over, rather than as the means
through which children "negotiate meaning" with significant adults in
their world, thereby laying the foundation for overall academic and
cognitive development.
There are several major differences between the linguistic mismatch
rationale and that developed in this paper. First, the present rationale
 A Theoretical Frameviork
41

/
emphasizes the sociocultural determinants of minority students'
academic difficulties. A major reason for the success of quality bilingual
programs is that they encourage minority students (and probably the
minority community) to take pride in their cultural background. A pro-
gram that continues to promote students' LI throughout ekknentary
school is much more likely to reinforce children's cultural identity than
one that aims to remove children as quickly as possible from any contact
with, of use of, LI iI3 School.
A secoorway in which the present rationale differs from the linguistic
I mismatch rationale is that it takes account of the difference between
context-embedded and context-reduced conununicativi proficiency. The
linguistic mismatch rationale leaves undefined the name of the "English
proficiency" required to survive in an all-English classroom; but by
default, relatively superficial aspects of context-embedded com-
municative proficiency have usually been regarded as adequate. This
assumption ignores the fact that it takes L2 learners considerably longer
to achieve grade-appropriate levels of L2 context-reduced com-
municative proficiency than it does to achieve peer-appropriate levels of
face-to-face context-embedded communicative proficiency. Thus, the
present analysis suggests that a realistic reclassification threshold of
"English proficiency" is unlikely to be attained by most language
minority students until the later trades of elementary school.
A third difference between the linguistic mismatch rationale and that
developed in this paper relates to the role assigned to minority students'
LI proficiency in the acquisition of English acadanic skills. Instruction
through LI is regarded as much more than an interim carrier of subject
mum content; rather, h is the means thrnugh which the conceptual anti
communicative proficiency that underlies botia I and English literacy is
developed. The elaboration of the CUP Model provides,a rationale for
continuing the promotion of Ll literacy development throughout
elementary school as a means 'of simultaneously contributing to'' the
development of both English and LI literacy skills.
A fourth difference is the fact that, unlike the linguistic mismatch ra-
tionale, the present rationale emphasizes the additional cognitive and
linguistic advantages (beyond the obvious advantage of being bilingual)
that research suggests are associated with the attainment of proficient
,'
bilingual skills.
Finally, within the present framework, the language spoken by the
child in the home is, in itself, essentially irrelevant. Mutt should be much
more important in determining thelasponse of the school are the
sociocultural characteristics and overall level of communicative profi-
ciency of children on entry. The school program should in every case at-

54
42 Schooling and Language Minority Students:

tempt to build an (rather than replace) the entry characteristics of

children.
Who Should Eater IR Squid Programs?
The research evidence reviewed above strongly suggests that programs
that aim to develop a high level of proficiency in two languages provide
greater potential for academic development for all children than educa-
tion through the medium of only one language. Whether or not this
greater potential is realized in any particular bilingual program will, of
count, depend on the quality of the program. Research has failed to
identify any category of student for whom a bilingual education would
be less suitable than a monolingual education. This issue has been exten-
sively researched in Canada in the context of French/English bilingual
programs. Students with learning disabilities, low academic ability, and
non-English or non-French home backgrounds have all been found to
perform at least as well in French/English bilingual programs as
equivalent students in English-only programs (Cummins, 1910. In
other _words, the enrichment potential of bilingual education is accessible
to all students.
This conclusion is also clearly supported by the recent large-scale
evaluation of bilingual education programs in the state of Colorado
(Egan and Goldsmith, 1981), which found that students from English
language backgrounds gained just as much from bilingual education as
"linguistically different" students. Both groups of students are reported
to have made significant gains in bilingual programs compared to what
would have been expected in regular English programs. For language
minority students who fail in L2-only school programs, bilingual educa-
tion offers a very basic form of enrichment, i.e., the possibility of educa-
tional survival.
There has been considerable debate in recent years about which
categories of language minority students should enter bilingual pro-
grams. Much of this debate has been political in nature and only Dulay
and Burt (1980) have advanced any serivas educational argument in
favor of limiting access to bilingual education by Limited Engligh Profi-
cient (LEP) students. Arguing on the basis of the Linguistic Mismatch
Hypothesis, Dulay and Burt suggest that "English-superior" LEP
students should receive instruction primarily Would, English, "primary-
language superior" LEP students should receive bilingual education,
while "limited balanced" (i.e., equally limited in LI and L2) students
should be taught through whichever language is spoken at home. The
analysis and research reviewed in this paper shows that this suggestion
has no educational support, either empirical or theoretical.

55
 A Theoretical Framework 43

Reclassification and Exiting Coadderadons

It should be clear by now that there is no educational justification for
exiting students from a successful bilingual program. The CUP model
provides an interpretation of why students in bilingual programs per-
form well in English academic skills despite much less instruction
through English. Furthermore, many studies show cognitive and
academic advantages as a result of attaining literacy and fluency in two
languages. Exiting students from bilingual programs in the early grades
of elementary school is likely to short-circuit these academic advantages;
the rationale for a quick -exit policy is either socio-political in nature or
else based on an ill-conceived SUP model of bilingual proficiency.
It is instructive to examine the confused logic of transitional bilingual
education as currently practiced in many school districts. Minority
students in transitional programs are expected to make so much progress
in the cognitive and amides* skills underlying English literacy in the ear-
ly grades that after two or three years they should be able to compete on
an equal footing with their monolingual English-speaking peers. In other
words, a CUP model of bilingual proficiency is implicitly endorsed in the
early grades. Yet proponents of a quick-exit policy revert to a SUP model
by assuming (contrary to their earlier assumption and the research data)
that children's English skills will not develop adequately unless they are
mainstreamed as soon as possible to an English-only program. It is ironic
that the earlier they want the child mainstreamed, the more effectivegthey
must assume the LI instruction to have been in pfbmothig L2 proficiency
(Cummins, 198(1d).
Assessment Conalderadons
The lack of a theoretical framework that would allow the relationship
between "communicative competence" and academic achievement to be
considered is especially obvious in the confusion surrounding ap-
propriate ways of assessing language proficiency and dominance for en-
try and exit purposes in bilingual education. Some measures are intended
specifically not to relate to academic achievement [e.g., the Bilingual
Syntax Measure (Burt et al., 1975)1, while others are intended to show a
moderate relationship [e.g., the Language Assessment Scales (DeAvila
and Duncan, 1976)).
Given that the purpose of language proficiency assessment is place-
ment of students in classes taught through the language which, it is
assumed, will best promote the development of academic skills, it is im-
perative that the test have predictive validity for academic achievement.
In other words, the test must assess aspects of language proficiency
related to the development of literacy. If it does not, then its relevance to

56
44 Schooling and Language Minority Students:

the placement of bilingual students is highly questionable (Cummins,

1980b).
For entry at the kindergarten level, assessment should probably in-
volve cognitively demanding context-embedded measures, while for exit
purposes, cognitively demanding context-reduced measures should be
used (see Figure 3). The rationale for this suggestion is that context-
embedded measures are necessary to reflect children's pre-school
language experiences, but context-reduced measures are more ap-
propriate for reclassification purposes because they more accurately
reflect the communicative demands of an all-English classroom.
Condutdon
Although further research is required to specify in detail what con-
stitutes "sufficient" English proficiency for reclassification purposes,
there is considerable evidence regarding conditions necessary for English
literacy 6evelopment among students traditionally performing poorly in
English-only school programs. The research suggests that achievement in
English literacy skills is strongly related to the extent of development of
LI literacy skills. Thus, rather than reclassifying and exiting minority
students as soon as possible, teachers and administrators should be con-
cerned with providing students with sufficient time in the bilingual pro-
gram to develop "threshold" levels of biliteracy.
How much time is sufficing? The evidence reviewed earlier suggests`
that school districts should aim to provide at least 50 percent of instruc-
tion in the early grades through the child's LI, and instruction in and
through the LI should be continued throughout elementary school.
Although there are no exact formulas as to how much LI and 1.2 instruc-
tion ought to be provided at any particular grade level, it seems
reasonable to suggest that it would be appropriate to provide more
English input in school in situations where exposure to English outside
school is limited. However, this increased exposure should not come in
the early grades where the instructional emphasis should be on LI, in
d.

order to develop the conceptual apparatus required to make English

context-reduced input comprehensible. Where there is little or no ex-
posure to English outside school, between 50 and 75 percent of the in-
structional time could be through English from third grade.
It is critically important, however, that decisions made by teachers, ad-
ministrators, and policy-makers regarding bilingual education take ac-
count of the nature of language proficiency and its cross-lingual dimen-
sions. The rationale for bilingual education and the specific program
suggestions made in this paper and others in this volume can be ap-
preciated only when it is realized that context-reduced communicative

57
 A Theoretical Framework
45

proficiency is different from context-embedded ,communicative profi-

ciency and that most academically important aspects of LI and 12 profi-
ciency are manifestations of the 611111e underlying dimension.

REFERENCES
Albert, Martin L, and Loraine K. Obler. The Bilingual Efrain. New York: Academic
Press, 1978.
Anderson, James 0., and William H. Johnson. "Stability and Change Among Three
Generations of Mexican-Americans: Factors Affecting Achievement," As:senora
Educational Research Marital, VIII, No. 2 (March, 1971). 285-309.
Baral, David P. "Academic Achievment of Recent Immigrants from Mexico," MADE
Journal, III (1979). 1-13.
Seedier, Carl, and Marlene Scardamalia. "Does Learning to Write Have to Be So Dif-
ficult?" Unpublished manuscript. Ontario Institute for Studies in Education. 980.
. and . "From Conversation to Composition: The Role of Instruction in
Developmental Process," Advances in Instructional Psychology, ed. Robert Glaser.
Val. 2. Hillsdale, New Jefligr. Lawrexce Fulham Assoc, 1981. .
Bethell, Tom. "Against Bilingual Education: Why Johnny Can't Speak English."
Harper's, CCLVIII, (Felnuary, 1979). 30.33.
Ilharnager, Joel. "Linguistic Behavior and Adjustment of Immigrant Chien in French
and English Schools in Montreal," Interrooloord loom& of APPlied Psychology.
XXIX (1960), 141-158.
Bruner, Jerome Seymour. "Language as an Instrument of Thought," Prnbkins of
Language and Learning, ed. A. Davies. London, England. Heimann, 1975.
Bun, Martha K., Heidi C. Dulay, and Eduardo Hemandea-C'havez. Bilingual Syntax
Memo, MAO. New York: The Psychological Corporation, 1975.
Canak, iffichael. "From Commtmkative Competence to Connnunicadve Language
Pedagogy," Language and COMMIU00711011, arcs., J. Richard, and R. Schmidt. New
York: Longman, 1981.
and Merrill %min. "Theoretical Bases of Communicative Approaches to
Second Language Teaching and Testing," Applied Linguistics, 1 (1980), 1.47.
Cisramazza, Alfonso. and lsabd Brows. "Semantic Classification By Bilinguals,"
Canadian Journal of Psychology, XXXIV, No. I (March, 1980), 7741.
Carey, Steven T., and James Cummins. "English and French Achievement of Chute 3
Children from English, French and Mixed French-English Home Backgrounds
Attending the Edmonton Separate School System English-French Immersion
Provram." Report submitted to the Edmonton Separate School System, 1979.
Caner, Thomas P. Atexican-Amerkans in School: A Ms:my of Educational Neglect. New
York: College Entrance Examination Board, 1970.
Chesarek, Steve. "Cognitive Consequences of Home or School Education in a Limited
Second Language: A Case Study in the Crow Indian Bilingual Community." Paper
presented at the Language Proficiency Assessment Symposium, Airily House,
Virginia, March, 1981.
Chomsky, Carol. "Stages in Language Development and Reading Exposure," Harvard
Educational Review, XLII, No. 1 (February, 2972), 1-32.
Chu-Chang, Mae. "The Dependency Relation Between Oral Language and Reading in
Bilingual Education," Journal of Education, CLXIII (1981), 30.57.
Cummins, James. "Age on Arrival and Immigrant Second Language Learning in Cpnada:
A Reassessment." Applied Linguistics, II, No. 2 (1981), 132-149.

58
46 Schooling and Language Minority Students:

"Cogve/Academic Language Proficiency. Linguistic Interdependeme, the

Optimal Age Question and Some Other Manus," Wor*ing Papers on Bilingualism,
No. 19 (1979a).
. "'The Cioss-Lingual Dimensions of Language Proficiency: Implications for
Bilingual Education and the Optimal Age Issue," TESOL Quarterly, XIV.
No. 2 (June, 1980), 175-187.
"The Exit and Entry Fallacy in &lingual Eduadon," NABE Journal, IV
(1980d), 23-60.
"The Influence of Bilingualism on Cognitive Growth: A Synthesis of Research
Findings and Explanatory Hypothesis," Working Pc gPers on Bilingualism,
No. 9 (1976), 1-43.
"Language Proficiency, &literacy, and French Immersion," Canadian Journal
of Education , in press.
"Linguistic Interdependence and the Educational Development of Bilingual
Children," Bilingual Education Paper Serif, Vol. 3, No. 2. Los Amides, California:
National Dissemination and Assessment Center, California State Univenhy,
Los Aniteles. September, 1979b.
"PsYdiological Assesament of Immigrant Children: Logic or Intuition?"
Journal of Multilingual and Multicultural Development, I (1980c), 97-111.
, and Robert Mulcahy. "Orientation to Language in Ukrainian-English Bilingual
Children," Child Development. XLIX, No. 4 (Member, 1978), 1,239-1,242.
Merrill Swain, Kazuko Nakaihna, Jean liandscombe, and Daina Green.
"Linguistic Interdependence Among Japanese Immigrant Students? Paper presented
at the Language Proficiency Assessment Symposium, Able House, Virginia,
March, 1981.
Darcy, Natalie T. "A Review of the Literature on the Effects Of BWngualism Upon the
Measurement of Intelligence," Journal of Genetic Psychology, LXXXII
(March, 1953).
DeAviht, Ediard A., and Raton E. Duncan. Langsmge Assessment Sages /LAS). Corte
Madera, California: Unguametrics Groups, 1976.
Donaldson, Margaret. Children's Mindy. Glasgow: Collins, 1978.
Downing, John. "Strategies of Bilingual Teaching," International Review of Education,
XXIV, No. 3 (1978) 329-346.
Dulay, Heidi C., and Marina K. Bun. "The Relative Proficiency of Limited English
Profkient Students," NABE Journal, IV (1980), 1-23.
Duncan. Sharon E., and Edward A. DeAvila. "Bilingualism and Cognition: Souse Recent
Findings," NABE Journal, IY(1979), 15-50
Edmonton Public School Board. "Evaluation of the Bilingual (English- Ukrainian)
Program, Fifth Year." Research report. 1979.
Egan, Lawrence A. "Bilingual Education: A Challenge for the Future," NABE News,
(March, 1981).
and Ross Goldsmith. "Bilingual-Bicultural Education: The Colorado Success
Story," NABE News, (January 1981).
Ekstrand. Lan Henri. "Social and Individual Fraie Factors in L2 Learning: Comparative
Aspects," Papers from the First Nordic Cooference on Bilingusdism, ed., Tove
Skutnabh-Kannas. Helsinki. Finland: Helsinfors Universitetet, 1977.
Enriquez, Miguel. "Semantic Integration in Bilingual Langone Processing." Paper
presetned at the Summer Institute cm Bilingual Education. University of
Arizona, 1980.
Evaluation Associates. "Nestor School Bilingual Education Program Evaluation."
Unpublished research report, San Diego. California, 1978.

59
 a

A Theoretical Framework 47

Filleleft, Lily Wong "lndlvidud Differences in Second Language Acquisition,"

ludisidusd Ddiferentve IR Language Ability and Language Behavior, eds.,
Charles 3. Fillmore, Daniel Kemple', and William S-Y. Wang. New York: Academic
Press, 1979, pp. 203428.
Fischer, K. B., and B. Cabello. "Prediethm Student Success Fallowing Transition from
Bilingual Programs." Paper presented at AERA Madan, Toronto, Canada.
Los Avneles: University of California, Los Angeles, Center for the Study of
Evaluation, 1978.
Genesee, Fred. "Is There an Optimal Age for Starting Second Language Instruction?"
AfrGgi Journal of Echication, XIII (1978), 145-154.
Hanson. Gag. "The Position of the Second Genennion of Finnish IllaniattultS in Sweden:
The lamonance of U114.4003 is the Home Language to the Welfare of Second
Generation Immigrants.' Paper presented at a aymposium on the position of tin
second generation of Yugoslav immigrants in Sweden, Split, Yugoslavia,
October, 1979.
Hebert, Raymond at al, Rendearent Academism, et Langer, Thinseignentent chez lea
Elms hutwo-Alatitobains. Saint Boniface, Manitoba: Centre de ,ItediaChgli du
College Universkaire de Saint-Bcadface, 1976.
Hernandet-Chava, Eduardo, Merino K. BUM and Heidi C. Dulay. "Language
Dominance and Proficiency Testing: Sone General Considerations,"
NABE .1081,1101. Ill (1979), 41-54.
Heyman, Anna-Ciretta. lnvandrwharrr: Skin t. Stockholm, Sweden: StOckhahns
Invandrarnamd, 1973.
Jensen, 3. Vernon. "Effects of Childhood Bilinnualism, 1," nementary English,
)(MIX (Februar, 1962), 132-143.
Kessler, Cardin. and Mary Ellen Quinn. "Positive Effects of Bilingualism on Science
Problem- Solving Abilities," 31st Annual Georgetown Chriversity Round Table
on Language, wed Linguistics, ed., James E. Muds. Washington, D.C.:
Georgetown University Press, 1980, pp. 295 -308.
MAIM, W. 3. "Parental and Family Influences on Academic Achievement Among
Medcan-Anurkan Students." Unpublislud PhD dissertation, University of
California at Los Angeles, 1968.
Kolas, Paul A. "Bilingualism and Information Processing" Scient(jk American.
CCXVII, No. 3 (Manch, 1968), 7946.
Krasbea, Ste: D. "Bilingual Education and Second janguage Acquisition Theory,"
Schooling and Language Mbutrio: Students: A 77reoretical Framework
Los Angeles, California: Evaluation, Dissemination and Amt Center,
California State Universit, Los Anples, 1981.
Mkhati A. Lois, and Robin C. Scarceila. "Age, Rate and Eventual Attainment
in Second Language Acquisition." TESOL Quarterly, XIII, No. 4
(December. 1979), 571-$82- 6

Lambert, Wallace E., and G. Richard Tait's. Educathur of Chilobetr.- The

St. Lambert Etpitiment. Rowley, heassachusestk Newbury Home, 1972.
Landry. Rodrigue 3. "Apprentissane darn des Commies Bilinnues." Unpublished
research report. Udiversity of Moncton, New Brunswick, Canada, 1990.
. "Le Bilinguisme: Fact= Repetition," Canadian Modem Language Review,
XXXIV (1978), 548-576.
Legureta, Dorothy. "The Effects of Program Models on Language Acquisition
by Spanish Speaking Children," TESOL Quarterly. XIII, No. 4
(December, 1979), 521-534.

Go
4$ Schooling and tanguage Minority Students:

Leyba, Charles F. Longitudinal Stu*, Title VII Bilingual Pmgrarn, Santa F

Public Schools, Santa fir, New Mexico. Loa Anggits, C:alifornia: National Dis-
semination and Assessment Center, California State University, Los Angeles, 1978.
Macnamara, John. "Wawa lista and Thought," 21st *MOW Round Tie* on Lamptager
and Lirsvastks, 1970, ed., Janata E. Abatis. Washhqtton, D.C.: Georgetown Uni-
versity Pass, 1970, pp. 25-45.
McConnsda, Peter D. "Bilingual Linguistic Memory: Independence or Interdependence?"
Ren8tnillsm, Bicultural's:en and Edgar:lion, ed., S. T. Caw/. Edmonton, Alberta,
Canada: University of Alberta Press, 1974.
MUM. Jane R. Labe Eng the Men MO &tended.- Clinical and Social System Perspectives
on Mental ROWAIIita9. Berkek7, California: University of California Press, 1973.
Ca Ler. Johan W., Jr. "Mae Language Factor in the Evaluation of Bilingual Education,"
Georgetown University Round Table on Languages and Linguistics
e. James E. Abatis. Washington. D.C.: Geometown University Press,
1978, pp. 410-422.
. Longbow rests at School A tic Ammar& New York: Lcmgman, 1979.
Oistm, David R. "From Utterance to Text: The Bias of Language in Speech and Writing,"
flarvan1 Edircational Review, XLVII, No. 3 (August, 1977). 257-281.
Ramirez. Arnulfo G., and Robert L. Pulitzer. "The Acquisition of English
and Maintenance of Spanish in a Bilinsual Education Program," English as a
Second Language in Bilingual E.Arcation, eds., James E. Mails, and K. nvaddell.
Washington, D. C.: TESOL, 1976.
Ramsey, Craig A., and Edward N. Wright. "A &cup, English Language Vocabulary
Knowledge Test Derived from the MIMEO Full- Range Picture Vocabulary Test,"
Psychological Reports, XXXI (1972), 103-109.
Rosier, Paul, and Marilyn Fard4. "Bilingual Education at Rock PointSome Early
Results," TESOL Quarterly, X, No. 4 (December, 1976), 379-388.
Shim Roger W. "How Misconceptions About Lang map Affect Judaments About
Intelligence," Issas in Etwasatimr Reading. ed., S. F. Wanat. Arlington, Virginia:
Center for Applied Lingssistics, 1977.
Singer, Harry. "IQ Is and Is Not Related to Reading," Isms In &ablating Reading,
ed.. S. F. Wow- Al en, Virginia: Canter for Applied Linguistics, 1977.
Skutnabb-kangas, Tave, and Plods ankeens*. Tinxtring Migrant anithvetas Mother
Tongue raid Learning the Language if the Host Country in the Constar
of the Saris -Cultsinzi Situation of the Migrant Family. Helsinki: The Finnbh National
Commission for UNESCO, 1976.
Smith, Frank. Underseanclink Reuling: A Psychalhrguistic Analpis of Reading and
Learning to Read. New York: Holt, Rinehart sad Wlnsion, 1978.
Swain, Mervin. "French Immersion: Early, Late or Partial?" 77re Canadian Marva
Lange Review. ed., S. T. CareY, MOM (May, 1978), 577-585.
Tamen, Deborah. "Implications of the Oral/L Continuum for Cross-Cultural
COMMUniCatilla," Georgetown University Round Tat* on Languager and
Linguistics, 1980, ad., James E. Mids. Washington. D.C.: Georgetown
University Press, 1980.
Terrell, Tracy D. "The Natural Approach in Bilingual Education," Sehooling and
Langumie Minority Students; A TlieoretioW Framework. Lon Angeles, California:
Evaluation, Dissaninatkan and Assessment Center, California State University,
Los Angeles, 1981.
Toukomaa, Patti, and Tove Skuntabh-ilangas. The Intensive reaching of the Mother
Tongue to Migrant Chimer of Preschool Age and Children in the Lower Level of
Compri:maim School. Research Reports 26. Tampere: Department of
Sociology and Social Psychology, University of Tampere, Fhd.and, 1971.
 A Theoretical Framework 49

Troike, Ruda Itia C. "Research Evidence for the Effectiveness of Biliggual Eelimeon,
Rifingssol Ethstetton Rsper Series, Vol. 2, No. 5. Los Angeles, California: National
Dissemination and Assessment Center, California State University,
Los Angeles, December. 1978.
Trembley, William. "Is Bilingual Education Able to Do Its Job?" Los Angeles Times,
September 4. 1980.
UNESCO. The Use 41 Vernacular Lenguages in Education. Monographs of fundamental
education. No. 8. Paris: UNESCO, 1953.
United States Commist.:on on Civil Rights. A Defter Chance to Learn: Bilingual &mature!
Education. Clearinghouse Publication No. 51. Washington, D.C.:
OQVCITIMent Printing Office, 1975.
Vightska, Lev Sensonovids. Thought and Language. Cambridge, Massachusetts: M.I.T.
Press, 1%2.
Wells, Gordon. "Describing Children's Linguistic Development at Home and at School,"
British Educational Restate* Jotanal, V (1979), 75-89.
Yee, Leland Y., and Rolfe La Forge. "R Between Mental Abilities, Social Class.
and Exposure to English in Chinese ourth Graders," Journal of Educational
Psycholcer. LXVI, No. 6 (December, 1974), 826434.

62
View publication stats