Language

Transfer
Cross-linguistic influence
in language learning

Terence Odlin

CAMBRIDGE
Applied Linguistics
Series Editors: Michael H. Long and Jack C. Richards
Language
Transfer
Cross-linguistic influence in
language learning

Terence Odlin
The Ohio State University

Cambridge University Press

Cambridge
New York New Rochelle
Melbourne Sydney
Published by the Press Syndicate of the University of Cambridge
The Pitt Building, Trumpington Street, Cambridge CB2 1RP
32 East 57th Street, New York, NY 10022, USA
10 Stamford Road, Oakleigh, Melbourne 3166, Australia

© Cambridge University Press 1989

First published 1989

Printed in the United States of America

Library of Congress Cataloging-in-Publication Data

Odlin, Terence.
Language transfer / Terence Odlin.
—
p. cm. (Cambridge applied linguistics series)
Bibliography: p.
Includes index.
ISBN 0 521 37168 6 hardcover. ISBN 0 521 37809 5 paperback
1. Language transfer (Language learning) I. Title. II. Series.
Pl 18.25.035 1989
418-dcl9 88-30760
CIP

British Library Cataloguing in Publication Data

Odlin, Terence
Language transfer. - (The Cambridge applied
linguistics series).
1. Foreign languages. Learning. Role of
native languages.
I. Title
418'.007
ISBN 0 521 37168 6 hardcover
ISBN 0 521 37809 5 paperback
Language Transfer
THE CAMBRIDGE APPLIED LINGUISTICS SERIES
Series editors: Michael H, Long and Jack C. Richards

This new series presents the finding of recent work in applied linguistics
which are of direct relevance to language teaching and learning and of
particular interest to applied linguists, researchers, language teachers, and
teacher trainers.
In this series:
Interactive Approaches to Second Language Reading edited by Patricia L.
Carrell, Joanne Devine, and David E. Eskey
—
Second Language Classrooms Research on teaching and learning
by Craig Chaudron
Language Learning and Deafness edited by Michael Strong
The Learner-Centred Curriculum by David Nunan
The Second Language Curriculum by Robert Keith Johnson
Language Transfer
by Terence Odlin
— Cross-linguistic influence in language learning
Linguistic Perspectives on Second Language Acquisition edited by
Susan M. Gass and Jacquelyn Schachter
In memory of Walter Odlin
1908-1985
Series editors' preface

Language transfer has been a central issue in applied hnguistics, set-

ond language acquisition, and language teaching forat least a cen¬
tury. Within the last few decades, however, its importance in second
language learning has been reassessed several times. In the 1950s it
was often deemed the most important factor to consider in theories
of second language learning as well as in approaches to second lan¬
guage teaching. In the 1960s its importance waned as learners’ errors
were seen not as evidence of language transfer but rather of “the cre¬
ative construction process.” Some researchers virtually denied the ex¬
istence of language transfer in their enthusiasm for universalist
explanations. In recent years, however, a more balanced perspective
has emerged in which the role of transfer is acknowledged and in
which transfer is seen to interact with a host of other factors in ways
not yet fully understood.
This reassessment of the significance of language transfer is lucidly
demonstrated in this new addition to the Cambridge Applied Linguistics
Series. In this timely book, Terry Odlin presents a comprehensive and
original account of the nature of language transfer and its role in second
language acquisition. Dr. Odlin documents the historical development
of the concept of language transfer, explores the role of transfer in
discourse, semantics, syntax, phonology, and writing systems, and ex¬
amines the way language transfer interacts with linguistic as well as
cultural, social, and personal factors in second language learning and
use. In the process, he surveys a large body of literature and examines
data from many different languages.
Dr. Odlin’s analysis challenges simplistic notions of language trans¬
fer and offers instead a convincing account of the process as a phe¬
nomenon that is fundamental to research in second language
acquisition and applied linguistics. This book will hence be invaluable
to students entering the field of second language acquisition, research¬
ers, language teachers, and anyone interested in the fundamental
question of how language systems interact during the process of sec-
x Series editors’ preface

ond language acquisition. We are therefore delighted to be able to

make Dr. Odlin’s research available to a wider audience through the
Cambridge Applied Linguistics Series.

Michael H. Long
Jack C. Richards
Preface

The significance of cross-linguistic influences has long been a contro¬

versial topic. As this book indicates, the controversy has had a long life
not only among second language teachers and researchers, but also
among linguists interested in questions of language contact and language
change. Although it would be too much to hope that this book will
cause such a long-standing controversy to die, the discussion of transfer
here may help to set to rest some dubious claims and to point the way
toward more productive thinking about cross-linguistic influences. While
I have tried hard to avoid the sweeping claims that unfortunately have
been frequent in discussions of transfer, 1 make no secret of my belief
that transfer is an extremely important factor in second language ac-
quisitionrTfie available evidence, I feel, warrants that belief. Thus, the
focus of this book is on empirical investigations of learners’ behavior
in many contexts. There is some discussion of the pedagogical impli¬
cations of certain investigations, but it seems to me that relatively little
is known about the best ways to make use of transfer research in the
—
classroom hopefully, more teachers and teacher trainers will begin to
think about what those ways are. There is also some discussion of
theoretical work in other areas of linguistics, but I have made efforts to
limit that discussion, which could go on interminably, and to limit the
jargon that usually accompanies such discussion. Readers familiar with
Government and Binding, Schema Theory, and Sprachbund will not find
those terms, though they will note allusions to research using those terms.
Some background in linguistics will be helpful in reading certain chapters
(especially Chapter 7), but the glossary provided should help with some
of the terminology that seemed impossible to avoid.
While this book has just one author, there are many people who have
helped bring about whatever may be praiseworthy in it. In my graduate
work I had the good fortune to take courses with Diana Natalicio, who
recognized the seriousness of challenges to contrastive analysis in the
—
1960s and 1970s but who also recognized that the most extreme albeit
—
fashionable criticisms of work on transfer were themselves open to
challenges. Some of the more novel ideas in this book owe a great deal

xi
xii Preface

to work by Jacquelyn Schachter, Sarah Grey Thomason, and Eric Kel¬

lerman, all of whom also provided valuable feedback on a number of
my ideas. As this work took shape, Jack Richards provided much en¬
—
couragement and support without his interest, this book might never
have been finished. Ellen Shaw and Linda Grossman of Cambridge Uni¬
versity Press helped in many ways to see the manuscript through the
final stages. I would also like to thank several people who made my
search for studies of transfer easier by sending me some of their work:
Christian Adjemian, David Birdsong, Susan Gass, Lynn Eubank, Marku
Filppula, John Hinds, Richard Schmidt, David Singleton, and Lydia
White. Many thanks are also due to Lisa Kiser, Alan Brown, and other
members of the Department of English at Ohio State who provided
valuable comments on earlier drafts of the manuscript. Moreover, 1
received many forms of assistance from friends and colleagues in the
Department of English, the Department of Linguistics, the programs in
English as a Second Language, and also from members of the Linguistics
Institute of Ireland. In addition, I would like to acknowledge the generous
support provided by other units at Ohio State, including the College of
Humanities, the Office of Research and Graduate Studies, and the In¬
structional and Research Computation Center. Finally, 1 would like to
thank my family for their encouragement not only with this project but
with much else besides.
Contents

Series editors’ preface ix

Preface xi

1 Introduction 1

2 Earlier thinking on transfer 6

2.1 Languages (and dialects) in contact 6
2.2 Transfer as a controversy in language teaching 15

3 Some fundamental problems in the study of transfer 25

3.1 Problems of definition 25
3.2 Problems of comparison 28
3.3 Problems of prediction 35
3.4 Problems of generalization 43

4 Discourse 48
4. 1 Politeness 49
4.2 Coherence 58
4.3 Discourse transfer and other factors 64
4.4 Summary and conclusion 69

5 Semantics 71
5.1 Propositional semantics 71
5.2 Lexical semantics 77
5.3 Summary and conclusion 83

6 Syntax 85
6.1 Word order 85
6.2 Relative clauses 97
6.3 Negation 104
6.4 Summary and conclusion 110

7 Phonetics, phonology, and writing systems 1 12

7.1 General versus specific predictions 112
vii
viii Contents

7.2 Phonetic and phonological transfer 113

7.3 Pronunciation, language universals, and typologies 120
7.4 Writing systems 124
7.5 Summary and conclusion 127

8 Nonstructural factors in transfer 129

8.1 Individual variation 130
8.2 Transfer and age of acquisition 136
8.3 Transfer, linguistic awareness, and social context 140

9 Looking back and looking ahead 151

9.1 Some caveats 151
9.2 Some conclusions 152
9.3 Some areas for further research 155

10 Implications for teaching 157

Glossary 165
References 171
Language index 197
Author index 201
Subject index 207
 Introduction

When people hear a speaker with a “foreign accent,” they often try to
guess the speaker’s background. Sometimes racial features and some¬
times a style of clothing will help listeners guess correctly, but often the
only reliable clue seems to be how the individual talks. In such cases,
questions put to the speaker such as “Are you German?” or “Are you
Spanish?” suggest an intuition about the nature of language, an aware¬
ness, however unconscious, that the native language of a speaker can
somehow cause the individual to sound “foreign” in speaking another
language.
' The detection of foreign accents is just one example of the awareness
that people may often have, of cross-linguistic influence, which,is also
known as language transfer.' That awareness is also evident from time
to time in opmionsThat people have about foreign language study. Many
believe that the study of one language (e.g., Latin) will make easier the
study of a closely related language (e.g., French). Similarly, people often
believe that some languages are “easy” in comparison with others. For
example, many English-speaking university students see European lan¬
guages such as French as less difficult than Oriental languages such as
Chinese. Since the similarities between English and French seem to be
relatively great, French is often considered “easy.”
An awareness of language transfer is also evident in the mimicking of
foreigners. While the representation of foreigners in ethnic jokes is often
crude in more ways than one, stereotypes of the way foreigners talk are
sometimes highly developed among actors. The following passage comes
from a manual to train English-speaking actors in the use of different
foreign accents, in this case a Russian one:
Oh! I very good fellow! why? because I Cossack. 1 very big Cossack. Yah! I
—
captain of Royal Cossack Guard in Moscow in old country. Oh! I got fifty
— —
hundred five hundred Cossack they was under me. 1 be big mans. And
—
womens, they love me lots. Nastia Alexanderovna she big ballet dancer in
—
Czar ballet Countess Irina Balushkovna, she love me. All womens they love

1 A more extended definition and also a justification of the term transfer appear in
Chapter 3 (Section 3.1).

1
2 Language transfer
me. And men? Ach! they be ’fraid from me. They hating me. Why? because I
big Cossack. I ride big horse. Drink lots vodka. Oh! I very big mans.
(Herman and Herman 1943:340)

The manual provides a pronunciation guide for this passage so that

actors can make their phonetic mimicry seem plausible, but a number
of grammatical features in the passage also seem to be “typically Rus¬
sian,” such as the absence of an article and a copula in / very good
fellow. Another passage in the same manual provides a very different
— —
linguistic and ethnic stereotype. While the Irishwoman’s speech in
the following passage might be that of a monolingual speaker of English,
it is similar to stereotypical portrayals of Irish-English bilinguals by
modern Irish playwrights:
And what business is it of yours that I be awake or no? Be what right do you
come snooping after me, following me like a black shadow. Are youse never
going to leave me alone? Yous’d be after doing better minding your own
business and letting me for to mind mine. For 1 have an ache in me long-
suffering heart and lashin’s of pain cutting through me brain like a dull knife.
And me eyes is looking at a world that’s not of your living. For it’s a
—
revelation I’m after having a view into the banshee world of devils and
spirits and the dear departed dead now rotting their whitened bones under
the cold, black sod. Ah! sure, now, and it’s the likes of you and your friends
that call themselves sane, that disbelieves in what I’m after seeing and
knowing. (Herman and Herman 1943:100)
Analogous to the Russian passage, some of the grammatical features in
the Irishwoman’s speech appear to be stereotypically Irish: for example,
the syntactic pattern in what I’m after seeing and knowing, which in
standard English would be what I have seen and known. While these
portrayals of accents may seem exaggerated, they do typify the use of
special linguistic structures to characterize the speech of bilinguals.2
The distinctiveness of foreign accents often seems understandable in
light of cross-linguistic comparisons. For example, Russian does not have
present tense copula forms such as am or articles such as a, and so
omissions of the copula and indefinite article in / very good fellow may
seem to be clearly due to a difference in the grammatical systems of
Russian and English. The comparison of such differences, which is
known technically as contrastive analysis, has long been a part of second
language pedagogy, and in the twentieth century contrastive analyses
have become more and more detailed.’ Since such cross-linguistic com¬
parisons constitute an indispensable basis for the study of transfer, the

2 The Irishwoman’s speech is a more accurate characterization than what is often

found in so-called Stage Irish (cf. Bliss 1978; Sullivan 1980).
3 Technical terms that appear in the glossary (see page 165) are indicated by boldface
at their first occurrence.
 Introduction 3

discussion of second language research in this book will frequently in¬

clude contrastive observations.
In light of such everyday abilities as the recognition and mimicry of
foreign accents and in light of common beliefs about cross-linguistic
similarities and differences, there appears to be a widespread assumption
that language transfer is an important characteristic of second language
acquisition. It might seem obvious that many characteristics of a learner’s
linguistic behavior will closely approximate or greatly differ from the
actual characteristics of the second language because of similarities and
differences predicted by a contrastive analysis. In fact, however, the role
of language transfer in second language acquisition has long been a very
controversial topic.4 Some scholars have indeed argued for the impor¬
tance of transfer; some have gone so far as to consider it the paramount
fact of second language acquisition. Yet other scholars have been very
skeptical about the importance of transfer. Among linguists and language
teachers today, there is still no consensus about the nature or the sig¬
nificance of cross-linguistic influences.
Much of the discussion in the next chapter will review the reasons
for the skepticism about transfer, but a brief consideration of one of the
most important reasons is appropriate now. As already noted, charac¬
teristics of the Russian language seem to explain sentences such as I very
good fellow. A contrastive explanation, however, seems less than com¬
pelling in light of other facts. For example, speakers of Spanish, which,
like English, has copula verb forms, frequently omit forms such as am
and is (cf. Section 2.2). Moreover, such errors are found not only among
Russian and Spanish speakers but also among speakers of other lan¬
—
guages and also among children learning English as their native lan¬
guage. Thus, while a contrastive analysis might explain a Russian
speaker’s omission of copula forms, a Spanish-English contrastive anal¬
ysis would not explain the same error, and a contrastive analysis is
irrelevant for monolingual children who make this same error as they
acquire English. The pervasiveness of certain types of errors has thus
been among the most significant counterarguments against the impor¬
tance of transfer.
Despite the counterarguments, however, there is a large and growing

4 The terms acquisition and learning will be used interchangeably throughout this
work even though much of the writing on second language acquisition (e.g.,
Krashen 1981) distinguishes between the two terms. I agree with Krashen and
others that the outcomes of acquisition can differ depending on the awareness of
language that individuals have (cf. Section 8.3). However, 1 strongly disagree with
Krashen’s analysis of transfer and with much else in his interpretation of second
language acquisition (cf. Sections 2.2, 3.1). Since his characterization of acquisition
and learning is questionable in several respects, I see no reason to use his
terminological distinctions (cf. Gregg 1984; Odlin 1986).
4 Language transfer

body of research that indicates that transfer is indeed a very important

factor in second language acquisition. Accordingly, the primary aim of
this book is to reconsider the problem of transfer in light of recent second
language research. While the research to be reviewed points to the im¬
portance of transfer, it also frequently points to the importance of other
significant factors in second language acquisition. Thus, even though a
comprehensive review of second language research is beyond the scope
of this book, there will be frequent discussion of cases in which transfer
is either not a significant influence or an influence that interacts with
other influences.
There are a number of reasons for language teachers and linguists to
consider more closely the problem of transfer. Teaching may become
more effective through a consideration of differences between languages
and between cultures. An English teacher aware of Spanish-based and
Korean-based transfer errors, for example, will be able to pinpoint prob¬
lems of Spanish-speaking and Korean-speaking ESL students better, and
in the process, communicate the very important message to students that
their linguistic and cultural background is important to the teacher.5
Also, consideration of the research showing similarities in errors made
by learners of different backgrounds will help teachers to see better what
may be difficult or easy for anyone learning the language they are
teaching.
There are yet other reasons to know about research on transfer. For
historical linguists, such knowledge can lead to insights about the re¬
lation between language contact and language change. Although lan¬
guages change for a variety of reasons, the bilingualism that often results
from language contact situations can be a major factor. For example,
Hiberno-English, the dialect spoken in parts of rural Ireland, does have
several of the unusual characteristics of the Irishwoman’s speech cited
earlier, and a number of those characteristics appear to result from the
influence of Irish. Research on transfer is also important for a better
understanding of the nature of language acquisition in any context and
is thus of interest to anyone curious about what is common to all lan¬
guages, that is, language universals. As Comrie (1984) has noted, second
language research can provide a valuable empirical check on the merit
of universalist theories, and the issue of transfer is likely to figure prom¬
inently in such research.
This book consists of ten chapters. The next two provide an overview
of the issues: Chapter 2 is a historical survey of the controversy sur-
5 Throughout this book, the term ESL (English as a Second Language) will be used
even in cases in which EFL (English as a Foreign Language) might be more
appropriate. While such a terminological distinction can be crucial for those
developing syllabi or preparing pedagogical materials, the distinction is less
important for researchers studying cross-linguistic influence.
 Introduction 5

rounding language transfer, and Chapter 3 is a discussion of four types

of problems especially important in the investigation of transfer. The
next four chapters survey second language research on transfer and
universals in relation to linguistic subsystems: discourse (Chapter 4);
—
semantics including a discussion of morphology (Chapter 5); syntax
(Chapter 6); and phonetics, phonology, and writing systems (Chapter
7). Chapter 8 discusses in more detail some aspects of language transfer
which structural descriptions cannot always account for, such as the
effects of individual variation in second language acquisition. Chapter
9 reviews important currents in the research discussed in earlier chapters,
and Chapter 10 considers some of the implications that the research
may have for teaching.

Further reading
Most studies of transfer appear in a wide variety of journals, but they
sometimes appear in special collections. One of the best collections is edited
by Gass and Seiinker (1983). A recent book-length study by Ringbom (1987)
combines a review of many of the controversies about transfer with a
detailed empirical study. Ellis (1985) has written a remarkably comprehensive
and judicious survey of research on second language acquisition, including
work on transfer. For more discussion of linguistic analyses of the literary
treatment of foreign accents, a text by Traugott and Pratt (1980) is useful.
Recent introductions to linguistics include texts by Bolinger and Sears (1981)
and Fromkin and Rodman (1983).
2 Earlier thinking on transfer

Discussions of transfer often begin with the work of American linguists

in the 1940s and 1950s. Yet while the work of Charles Fries, Robert
Lado, and others was clearly a major catalyst for subsequent research,
serious thinking about cross-linguistic influences dates back to a con¬
troversy in historical linguistics in the nineteenth century. Accordingly,
this chapter begins with a discussion of that controversy among scholars
whose primary interests were not second language acquisition or lan¬
guage teaching but rather language classification and language change.
The controversy promoted work on language contact that overlaps con¬
siderably with more recent studies of second language acquisition. Be¬
cause the thinking of Fries, Lado, and others prompted much of the
growth of research in second language acquisition, their views receive
considerable attention, as do the views of some who have been very
critical of their work. While this chapter can give only a suggestion of
the historical context of the polemics on transfer, it provides important
background for some fundamental issues discussed in subsequent
chapters.

2.1 Languages (and dialects) in contact

Historical change and language mixing
Language contact situations arise whenever there is a meeting of speakers
who do not all share the same language and who need to communicate.'
When the communicative needs of people go beyond what gestures and
other paralinguistic signals can achieve, some use of a second language
becomes necessary. The languages learned in contact situations may or
may not show some kind of language mixing, that is, the merging of
characteristics of two or more languages in any verbal communication.

1 There are other kinds of language contact besides those discussed in this chapter,
as, for example, when a French scholar deciphers a text in ancient Egyptian. Such
cases, though, are exceptional in a number of ways.

6
 Earlier thinking on transfer 7

If mixing does occur, native language influence is only one of the possible
forms it can take. Another kind of mixing is in the form of borrowings
from a second language into the native language (e.g., the use by English
speakers of the loanword croissant from French to describe a certain
kind of pastry), and still another kind is code-switching, in which there
is a systematic interchange of words, phrases, and sentences of two or
more languages (cf. Sections 8.2, 8.3).
People often show some awareness of language mixing, even though
they are usually not familiar with terms such as code-switching and
transfer (cf. Chapter 1 and Section 8.3). Among Indians in the Vaupes
Tegion of the Amazon rain forest, for example, there is a keen awareness
of the mixing that arises in their multilingual villages (Sorenson 1967;
Jackson 1974). Such awareness probably reflects a consciousness going
far back into prehistory. Whether in the rain forest or elsewhere, humans
have often seen themselves as belonging to different social groups, and
they have often considered language to be an important distinguisher of
their own group from others; it is no accident that the names of languages
frequently designate ethnic groups (e.g., Chinese, Navajo, and English').
Accordingly, any introduction of loanwords or other kinds of language
mixing may be viewed either as a kind of linguistic intrusion or as a
“foreign import,” sometimes welcome, sometimes not (cf. Le Page and
Tabouret Keller 1985). It is significant that the Indians of the Vaupes
do not look upon language mixing favorably; although that attitude is
probably not universal, people in many communities do have similarly
unfavorable attitudes toward various kinds of mixing. For example,
English loanwords in French and other languages have frequently been
a target of language purists (cf. Section 8.3).
From antiquity onwards there is a historical record of people asso¬
ciating language contact and mixing with “contamination” (Silvestri
1977; Thomason 1981). Typical of such associations were scholarly
discussions in Renaissance Europe about the link between Latin and the
vernacular languages related to it. With regard to the origins of French,
for example, scholars speculated about how speakers of other languages
may have “corrupted” the language brought to Gaul by the Romans
(Silvestri 1977). Although some scholarly work before the nineteenth
century did make specific claims, most of the discussion about language
contact and mixing was rather nebulous. Apart from occasional remarks
about loanwords, few discussions included either detailed characteri¬
zations of the nature of cross-linguistic influences or specific examples
of such influences (Silvestri 1977).
In the nineteenth century, debate about the importance of language
contact and mixing intensified. The question of mixing had major im¬
plications for two interrelated problems that interested many nineteenth¬
century linguists: language classification and language change. The
8 Language transfer

steady accumulation of grammars of languages in every part of the world

made ever clearer the diversity of human languages and the scientific
challenge of classifying them (Robins 1979). Many scholars came to
believe that grammar was the soundest basis on which to construct
classifications. Aware that lexical borrowings (i.e., loanwords) could
often make classification decisions difficult, scholars often expected to
find in grammar a linguistic subsystem unaffected by language contact
and thus a key to distinguish any language. Muller (1861/1965:75), for
example, was well aware of the large number of loanwords from Latin,
French, and other languages in English, but considered English grammar
to be immune from cross-linguistic influences:
The grammar, the blood and soul of the language, is as pure and unmixed in
English as spoken in the British Isles, as it was when spoken on the shores
of the German ocean by the Angles, Saxons, and Juts [sic] of the continent.2
The beliefs of Muller and others about the uniqueness of grammar were
usually related to assumptions about the tree model of language change
in which languages are viewed as parts of a “family tree.” In that model,
Latin, for example, is characterized as the “parent” language and French,
Spanish, Rumanian, and other Romance tongues as the “daughter” lan¬
guages.3 The pattern of change in the tree model is primarily one of
internal development, in which characteristics of the parent language
undergo changes that are systematically manifested in the daughter lan¬
guages. For example, in French, Spanish, and Rumanian, noun phrases
commonly have definite articles, as in the following translations of “the
mountain”:
French: le mont
Spanish: el monte
Rumanian: munte-le
In Latin there were no articles, but in virtually all of the daughter lan¬
guages there are definite and indefinite articles. The development of such
an innovation in each of the daughter languages suggests the existence
of what Whitney (in a slightly different connection) called “forces which
are slowly and almost insensibly determining the growth of a language”
(1881:25). Like Muller, Whitney took a dim view of the notion of
mixing, and their attitude was shared by other scholars aware of the
massive evidence of internal change not only in morphology and syntax
but also in phonetics and phonology (e.g., Meiller 1948). If internal

2 Despite the racist connotations of this passage, Muller was an outstanding advocate
of racial understanding in the nineteenth century.
3 As Robins (1979) observes, such characterizations of historical relationships are
metaphoric and only partially revealing. Nevertheless, the family tree metaphor had
an enormous impact on thinking about language change.
 Earlier thinking on transfer 9

development could explain so much about language change, were ex¬

planations involving language mixture really necessary?

Areal linguistics
In the judgment of many scholars in the nineteenth and twentieth cen¬
turies, the answer to this question is affirmative, despite the cogent
arguments by Whitney and others. Even while the evidence supporting
the tree model of change is strong, there is good evidence supporting a
very different model, the wave model (cf. Bloomfield 1933; Bailey 1973;
Bynon 1977; Zobl 1984). First developed in the nineteenth century, the
wave model has long been recognized as a useful complement to the tree
model, especially for an understanding of dialect change. While char¬
acterizations of the wave model have changed over the years, the model
still posits that linguistic patterns in one dialect can affect another dialect
considerably, especially if the dialects are spoken in adjoining regions.
For example, the English spoken in Chicago and that spoken in nearby
towns in northern Illinois are similar, but recent changes in pronunci¬
ation that have appeared in Chicago make that variety different from
other Illinois dialects. The changes seem to be slowly spreading from
Chicago into other Illinois towns, largely through contacts between Chi¬
cagoans and people in the larger towns (cf. Callary 1975; Chambers
and Trudgill 1980).
There is now a considerable body of scholarship pointing to the im¬
portance of dialect mixture (e.g., Trudgill 1986). Yet the significance of
the wave model is not limited to dialect contact. As most linguists ac¬
knowledge, the difference between languages and dialects is often fuzzy,
as the linguistic situation in parts of Spanish-speaking Uruguay and
Portuguese-speaking Brazil shows. While Spanish and Portuguese are
distinct in many ways, they might well be considered two Romance
dialects instead of two languages were it not for political facts. On the
border between Brazil and Uruguay there have been frequent contacts
between people of both nations, and, not surprisingly, the similarity of
Spanish and Portuguese has encouraged a great deal of mixing which
one might call either dialect mixing or language mixing (Rona 1965).
Even when the differences between two languages are greater than is
the case with Spanish and Portuguese, there is a possibility of language
mixing. For example, Rumanian and Bulgarian have somewhat different
“genetic” classifications (the former is a Romance language and the latter
a Slavic language), but centuries of contact between speakers of Ru¬
manian, Bulgarian, and other languages in the Balkans have led to many
areal (i.e., regional) similarities not due to internal changes (Sandfeld
1930; Joseph 1983a). The definite article, for instance, follows the noun
10 Language transfer

in Rumanian, Bulgarian, and Albanian, as in the following translation

equivalents of “the village”:
Rumanian: sat-ul
Bulgarian: selo-to
Albanian: fshat-i
As noted earlier, French, Spanish, and Rumanian all have definite ar¬
ticles, but in the Romance languages of Western Europe the definite
article precedes the noun, whereas in Rumanian the definite article fol¬
lows the noun. This divergence of Rumanian from the general Romance
pattern can be best explained in terms of areal contact: The position of
the article appears to reflect centuries of bilingualism in the border
regions between Rumania and Bulgaria.4

Pidgins and creoles

The areal similarity of languages in the Balkans is among the best known
examples of the long-term effects of language contact, but there is also
a great deal of evidence for the importance of language contact in his¬
torical change in other areas, such as India (Gumperz and Wilson 1971;
Emeneau 1980) and Ethiopia (Leslau 1945, 1952). Not all contact will
lead to transfer, however. The importance not only of transfer but of
other explanations for contact phenomena became clear as the study of
pidgin and creole languages intensified toward the end of the nineteenth
century. In that period Hugo Schuchardt, a German linguist who had
noted the likely effects of transfer in certain contact situations in Europe,
became interested in the so-called trade languages spawned in Africa,
Asia, and elsewhere from encounters between the local inhabitants and
Europeans. In some encounters, as in the case of the dealings between
European and Chinese merchants on the coast of China, contacts were
limited and the trade jargons used had only the status of “marginal
languages,” which are usually called pidgins. In other contact situations,
however, the trade languages became more extensively used and often
became languages acquired by young children; those languages are usu¬
ally considered creoles.5 Initially, Schuchardt thought that transfer was

4 Explanations about the development of the Rumanian article do not agree in all
details, yet whatever the correct explanation is, language contact is probably
involved (cf. Joseph 1983a).
5 Todd (1983), Miihlhausler (1986), and others have argued that it is an
oversimplification to equate pidgins with the language of one generation of adults
and creoles with the language of children of the subsequent generation. Singler
(1988) suggests that the distinction should refer to whether a group has adopted a
language as the language of ethnic identification, in which case it is a creole and not
a pidgin. If Singler’s view is correct, the adoption of a pidgin by children may be a
necessary but not a sufficient condition for terming the new language a creole.
 Earlier thinking on transfer 11

a likely explanation for many of the features of pidgins and creoles, and
in his earlier articles on the subject he presented cases that he viewed
as probable or at least possible instances of transfer: for example, the
word order of a Spanish creole in the Philippines and the word order
of a Portuguese creole in India (Schuchardt 1883a, b). However, other
scholars’ arguments and Schuchardt’s own investigations of other con¬
tact situations led him to speculate about alternative explanations for
many characteristics in pidgins, creoles, and other contact languages. In
examining the pidginlike English spoken by servants and colonizers in
India, Schuchardt (1891/1980) wondered how much the rudimentary
structure of the language might be due to attempts by the colonizers to
make English more comprehensible by simplifying structures; such at¬
tempts have subsequently become known as Foreigner Talk (e.g., Fer¬
guson 1975). His growing awareness of the similarities between different
trade languages led Schuchardt (1909/1980) to consider various types
of simplification: the simplification of grammatical features by children
learning their native language, the simplification that adults use for the
benefit of children, the simplification seen in Foreigner Talk, and the
simplification found in the grammars of so many pidgins and creoles
(cf. Slobin 1977). As Gilbert (1980) suggests, Schuchardt became more
and more aware of apparently universal tendencies toward simplicity in
situations involving language contact and language acquisition. A similar
awareness motivates much of the current research on second language
acquisition (cf. Meisel 1983).
In the study of pidgins and creoles in the twentieth century, there has
been considerable disagreement about the relative importance of transfer
and universals in the creation of contact languages (cf. Andersen 1983a;
Muysken and Smith 1986). Explanations based on transfer and those
based on universalist notions are not always mutually exclusive, and
there are other theoretical positions besides those two (cf. Miihlhausler
1986). However, universalist and transfer approaches sometimes do
conflict, and both approaches have many clear attractions (cf. Section
3.4). Miihlhausler argues persuasively that multicausal explanations are
the most satisfactory, but weighing the relative importance of transfer,
universals, and other factors in such explanations remains difficult. Most
probably, transfer will play a relatively minor role in some situations,
but in other cases it will play a major role. Certain pidgins and creoles
in New Guinea seem to illustrate the former possibility; perhaps the
contact between speakers of many different languages keeps any native¬
language influence from greatly affecting the newer creoles of the region
such as Tok Pisin (Miihlhausler 1 986). On the other hand, the Hawaiian
Pidgin English spoken by many Japanese shows unmistakable influences
of Japanese word order and other structures, as Nagara (1972) and
others have shown (Section 6.1). This, along with other evidence, sug-
12 Language transfer

gests that when only a small number of languages are spoken in a contact
situation, a pidgin will show more transfer effects (cf. Singler 1988).

Borrowing and substratum transfers

Aside from research on pidgins and creoles, scholarship on language
contact in general has increased considerably in the nineteenth and twen¬
tieth centuries, and with that increase has come much evidence for the
importance of cross-linguistic influences. With the growth of this evi¬
dence, the diversity of situations in which transfer occurs has also become
clearer, as research by Weinreich (1953/1968) and others shows. It
should be noted that Weinreich used the term interference to cover any
case of transfer. Nevertheless, his survey of bilingualism shows that the
effects of cross-linguistic influence are not monolithic but instead vary
considerably according to the social context of the language contact
situation. These effects can often be distinguished through the use of
the terms borrowing transfer or substratum transfer (cf. Thomason and
Kaufman 1988). Borrowing transfer refers to the influence a second
language has on a previously acquired language (which is typically one’s
native language). For example, borrowing transfer has been amply doc¬
umented in the case of Young People’s Dyirbal (Schmidt 1985). Dyirbal,
an Aboriginal language spoken in northeastern Australia, has undergone
considerable attrition (i.e., change from the traditional norms) as a result
of the younger Aborigines’ exposure to English. While not all of the
attrition is directly due to English influence, many characteristics of
Young People’s Dyirbal show clear evidence of transfer. Other cases of
attrition with concomitant borrowing transfer include a variety of Swiss
Romansh influenced by German (Weinreich 1953/1968), some varieties
of Greek influenced by Turkish (Dawkins 1916), and American varieties
of Norwegian influenced by English (Haugen 1953).
Substratum transfer is the type of cross-linguistic influence investigated
in most studies of second language acquisition; such transfer involves
the influence of a source language (typically, the native language of a
learner) on the acquisition of a target language, the “second” language
regardless of how many languages the learner already knows.6 Aside
from studies of such transfer in the performance of second language
students, there are a number of language contact studies showing the
importance of native language influence, as in the case of the influence

6 Not all cases of cross-linguistic influence can be neatly classified as borrowing or

substratum transfer. In the case of Ethiopia, for example, the evidence suggests a
possible occurrence of both kinds of transfer (Thomason and Kaufman 1988). In
cases of young children learning two languages simultaneously, cross-linguistic
influence can occur (e.g., Taeschner 1983), but such influence differs considerably
from typical instances of borrowing or substratum transfer (Sections 8.2, 8.3).
 Earlier thinking on transfer 13

of Irish on the English spoken in Ireland (Henry 1957; Filppula 1986),

the influence of Quechua on the Spanish spoken in Peru and Ecuador
(Lujan, Minaya, and Sankoff 1984; Muysken 1984), and the influence
of Arabic and Turkish on the Dutch spoken by immigrant workers in
Holland (Jansen, Lalleman, and Muysken 1981).7 Since the focus of this
book is on native language influences, the term transfer will hereafter
serve as an abbreviation for substratum transfer, except where there is
an explicit comparison made between borrowing transfer and substra¬
tum transfer.
While borrowing transfer and substratum transfer are similar in some
ways, their results are often very different. Thomason and Kaufman
(1988) argue that such differences reflect differences in social as well as
linguistic factors. Borrowing transfer normally begins at the lexical level,
since the attrition of the language absorbing the foreign vocabulary
normally begins with the onset of strong cultural influences from speak¬
ers of another language. The group exerting the influence is often, though
not always, a speech community with larger numbers, greater prestige,
and more political power. In such cases, words associated with the
government, the legal system, the schools, the technology, and the com¬
mercial products of the dominant majority are among the first to make
their way into the minority language, but massive lexical borrowing may
also supplant much of the vocabulary of everyday living. In the case of
Young People’s Dyirbal, for example, Schmidt found that younger Ab¬
origines were not always aware of the traditional words to describe
activities such as cooking and would therefore employ English words
instead/ When borrowing transfer comes to have such a major effect
on lexical semantics, there is often a great deal of cross-linguistic syn¬
tactic influence as well. As Schmidt indicates, Dyirbal syntax now shows
many influences from English. However, the phonetics and phonology
of the native language are less likely to be affected by borrowing transfer.
According to Schmidt, the pronunciation of Young People’s Dyirbal
does not depart drastically from that of older speakers.
Substratum transfer, on the other hand, will normally show a different
weighting of cross-linguistic influences. Thomason (1981) suggests that
the effects of substratum transfer will be more evident in pronunciation
(and also in syntax) than in the lexicon. If pronunciation is the most
difficult aspect of a second language to master, as some believe, the
influence of native language phonetics and phonology will be more per¬
vasive than the influence of other language subsystems (cf. Section 8.2).

7 In subsequent chapters (especially 6 and 8), there is more extended discussion of

these cases.
8 Such behavior may seem to be the same as code-switching, but many studies of
switching indicate that such mixing does not result from a lack of knowledge of
either language in the switch (Section 8.3).
14 Language transfer

The effects of native language influence may survive the historical ac¬
quisition situation, with both pronunciation and syntax in the newly
acquired language often providing an enduring testimony of earlier trans¬
fer, as is evident in the case of Hiberno-English, the English dialect
spoken in many parts of Ireland. Only a small number of people now
speak Irish as their native language, while nearly everyone speaks En¬
glish. However, many monolingual English speakers in Ireland use a
pronunciation and grammar that reflect a “substratum” of earlier cen¬
turies of widespread Irish-English bilingualism. Some of the clearest
examples of this type of transfer involve verb phrases that have parallels
in Irish, such as He’s after telling a lie (“He’s told a lie”) (cf. Wright
1898; Henry 1957; Bammesberger 1983; Bliss 1984).9

The social dimensions of transfer research

Since transfer occurs in a wide variety of social contexts, a thorough
understanding of cross-linguistic influence depends very much on a thor¬
ough understanding of those contexts. Accordingly, historical research
on transfer may serve as a useful cross-check on research in which social
factors are not investigated. Because many second language investiga¬
tions are detailed case studies or experimental studies involving relatively
small numbers of individuals, they cannot always suggest the importance
of demographic influences or other social factors whereas many histor¬
ical studies can. In the case of Ireland, for example, the widespread
adoption of English in the nineteenth century appears to have resulted
in much more transfer than njight be evident in the English of the
relatively small numbers of foreign students at American universities,
where so many second language studies are conducted (cf. Section 8.3).
Studies of both historical and contemporary cases of wide-scale language
contact are also useful for assessing the effects of formal instruction
on transfer. Some have claimed that transfer will be significant in ac¬
quisition affected by formal instruction but will be less so in naturalistic
second language acquisition (e.g., Dulay, Burt, and Krashen 1982). That
claim is supported by some evidence, but it is probably an oversimpli¬
fication, as there are studies of language contact in Ecuador, West Africa,
and elsewhere indicating that formal education may constrain transfer
(Section 8.3). While transfer is primarily a psychological phenomenon,
its potential effect on acquisition may be large or small depending on
the complex variations of the social settings in which acquisition takes
place.
9 In discussions of Hiberno-English there has been little disagreement that transfer is
responsible for constructions such as He’s after telling a lie. However, some other
constructions may be related to nonstandard usages in British dialects going back to
the seventeenth century (Harris 1984).
 Earlier thinking on transfer 15

2.2 Transfer as a controversy in

language teaching
The growth of contrastive studies
Despite the conflicting views on the significance of language contact in
historical linguistics, the notion of language transfer remained uncon-
troversial among language teachers well into the twentieth century. As
early as the schools of the ancient world, teachers were writing down
contrastive observations about the languages students knew and the
languages they wished to learn (Kelly 1969). And as recently as the
schools influenced by figures such as Sweet (1899/1972), Jespersen
(1912), Palmer (1917), Fries (1945), and other proponents of new (or
seemingly new) methods of language teaching, there was a widespread
acceptance of the idea that native language influences could greatly affect
second language acquisition. Challenges to assumptions about the im¬
portance of transfer did not have much impact on the history of language
teaching until the late 1960s.
The challenges that arose in that period were largely in reaction to
two claims that American scholars had made about transfer in the pre¬
ceding twenty or so years. The first of those claims was that the existence
of cross-linguistic differences made second language acquisition ex¬
tremely different from first language acquisition. In the foreword to
Linguistics Across Cultures, a highly influential manual on contrastive
analysis by Lado (1957), Fries stated:
Learning a second language .. . constitutes a very different task from learning
the first language. The basic problems arise not out of any essential difficulty
in the features of the new language themselves but primarily out of the
special “set” created by the first language habits.
Along with many linguists of his time, Fries subscribed to the behaviorist
analysis of linguistic competence as a series of habits (Bloomfield 1933).
Native language influence was thus the influence of old habits, some
potentially helpful, some potentially harmful (cf. Section 3.1).
The second claim that came to be challenged was that the difficulties
of second language acquisition could be determined through contrastive
analyses:
We assume that the student who comes in contact with a foreign language
will find some features of it quite easy and others extremely difficult. Those
elements that are similar to his native language will be simple for him, and
those elements that are different will be difficult. The teacher who has made
a comparison of the foreign language with the native language of the
students will know better what the real learning problems are and can better
provide for teaching them. (Lado 1957:2)
16 Language transfer

While other claims of Fries and Lado have also been disputed, the po¬
sitions just quoted would serve as major catalysts for research chal¬
lenging the significance of transfer.
Although Lado and others are sometimes accused of having only been
concerned with narrow structural analyses of language, the title of Lin¬
guistics Across Cultures indicates otherwise. Moreover, Lado stated that
“the fundamental assumption” of his book was that
individuals tend to transfer the forms and meanings, and the distribution of
forms and meanings of their native language and culture [emphasis added] to
the foreign language and culture - both productively when attempting to
speak the language and to act in the culture, and receptively when attempting
to grasp and understand the language and the culture as practiced by the
natives. (Lado 1957:2)
The importance of cultural as well as linguistic factors in acquisition
was clear to Lado, and he devoted a chapter of his book to suggesting
how to compare cultures. Despite this emphasis on culture as well as
on language, most contrastive analyses in the 1950s and 1960s concen¬
trated on pronunciation and grammar. However, a discussion by Kaplan
(1966) of cultural differences and second language writing proved to be
highly influential in subsequent research on second language discourse.
Kaplan claimed:
The foreign-student paper is out of focus because the foreign student is
employing a rhetoric and a sequence of thought which violate the
expectations of the native reader. (1966:4)
The rhetoric and “sequence of thought” reflected, according to Kaplan,
the discourse and thought patterns of the student’s own culture, and
were analogous to the transfer alleged to be so important in second
language pronunciation and grammar. Though Lado’s analysis only in¬
directly touched on discourse, Kaplan proposed that contrastive studies
were possible beyond the sentence level, and his arguments encouraged
the study of what is now frequently termed contrastive rhetoric (cf.
Sections 4.2, 4.3).
Fries, Lado, and others saw as extremely important the development
of materials specifically designed for different groups of students (cf.
Fries 1945; Lado 1957). According to Fries (1949:97), “The problems
of the Chinese student are very different from those of the Spanish
speaker,” and the materials for teaching English that he had helped
develop some years before reflected that belief. The pedagogical practices
advocated by Fries and others encouraged an expansion of contrastive
studies for pedagogical purposes. Many books, articles, and graduate
theses from 1950 to 1970 reflect the growth of such study (cf. Sajavaara
and Lehtonen 1981; Dechert, Briiggemeir, and Futterer 1984). However,
another area, empirical studies of transfer, was slower in developing.
 Earlier thinking on transfer 17

Although Lado (1957) and others did recognize the need for such re¬
search, it is clear that he believed that the language contact research of
Haugen (1953) and Weinreich (1953/1968) provided enough of an em¬
pirical demonstration of the importance of transfer to warrant the de¬
velopment of contrastive analyses for language teaching. More research
on cross-linguistic influences did eventually appear, though much of it
called into question earlier thinking on transfer.

Problems with contrastive analysis

The claims made by Lado and Fries about the predictive power of con¬
trastive analysis and about the relation between first and second language
acquisition faced serious challenges by the 1970s. The predictive validity
of many contrastive analyses seemed questionable: empirical research
was beginning to show that learning difficulties do not always arise from
cross-linguistic differences and that difficulties which do arise are not
always predicted by contrastive analyses. Moreover, the theoretical sig¬
nificance of transfer seemed dubious to a number of researchers struck
by the similarities between first and second language acquisition.10 Some
problems of contrastive analysis will be discussed in this section, and
the question of first and second language acquisition will be considered
in the next section.
Some differences between languages do not always lead to significant
learning difficulties. For example, two verbs in Spanish correspond to
—
different senses of the English verb know conocer and saber. While
this lexical difference poses many problems for English speakers learning
Spanish, Spanish speakers learning English seem to have little difficulty
in associating two lexical senses with one form (cf. Stockwell, Bowen,
and Martin 1965; Lee 1968). Thus the difference between Spanish and
English is not in itself enough to allow for accurate predictions of dif¬
ficulty (cf. Section 3.2).
An even more serious challenge to the validity of contrastive analyses
is the occurrence of errors that do not appear to be due to native language
influence. For example, a contrastive analysis of Spanish and English
would not predict that Spanish speakers would omit forms of the verb
be, since Spanish has similar grammatical structures." Whatever other

10 Criticism of contrastive analysis arose for other reasons in addition to those

discussed here, as seen, for example, in the analysis of Lado’s claims about testing
by Upshur (1962). However, the most serious challenges, those that most
influenced arguments against transfer up to the present, have been related to
problems of prediction and similarities between first and second language
acquisition.
1 1 Spanish has not one but two verbs equivalent to be: ser and estar. While this
contrast causes considerable problems for English speakers learning Spanish, it
 18 Language transfer

errors might arise, one would not predict that Spanish speakers would
say That very simple instead of That’s very simple (Peck 1978). Yet a
number of studies, including Peck’s, have documented the omission of
forms of be in the speech of Spanish-speaking learners of English (e.g.,
Butterworth and Hatch 1978; Peck 1978; Schumann 1978; Shapira
1978).
The errors that Spanish-English contrastive analyses fail to predict are
not just ones involving be forms. Other kinds of errors occur despite
clear similarities between English and Spanish in such areas as verb
tenses, word order, and prepositional usage (cf. Butterworth and Hatch
1978; Schumann 1978; Andersen 1979). Moreover, in looking at studies
of child bilingualism that had been conducted earlier in the twentieth
century as well as at more recent ones (e.g., Ronjat 1913; Ravem 1968),
scholars noticed that a number of these studies showed only minimal
evidence of transfer or any other kind of language mixing (cf. Section
8.2).
( Further questioning of the worth of contrastive analysis came from
classifications of learners’ errors in studies that became known generi¬
cally as error analyses (e.g., Duskova 1969; Richards 1971). Some errors
seem to arise not from language transfer but from other sources such
as transfer of training, that is, the influences that arise from the way a
student is taught (cf. Seiinker 1972; Stenson 1974; Felix 1981). While
some influences from teaching are no doubt beneficial, others can induce
errors that might not otherwise occur. For example, Felix notes that
question-and-answer drills can produce errors such as the following:
Teacher: Am I your teacher?
Student: Yes, I am your teacher.
As Felix observes, one would not predict such an error from a com¬
parison of pronouns in the student’s native language, which was German,
and pronouns in English.
Other errors documented in error analysis research seem to arise spon¬
taneously. For example, overgeneralizations such as This program of
the Kissinger (said by a Spanish speaker) often appear to be due to the
inappropriate application of a target language rule, here an overexten¬
sion of an article to a proper noun (Schumann 1978). Since Spanish,
like English, does not use articles with personal names such as Kissinger,
transfer is not a viable explanation. Similarly, errors such as omitting
articles, copulas, and other forms often seem to involve simplification
rather than transfer, as Schuchardt had suspected several decades earlier
(Section 2.1). For example, the omission by a Spanish speaker of an

does not seem to lead to any special problems for Spanish speakers learning
English (cf. Stockwell, Bowen, and Martin 1965).
 Earlier thinking on transfer 19

article in the phrase Picture is very dark (Schumann 1978) is not at¬
tributable to Spanish since a Spanish translation of the sentence would
require an article (see Section 3.2).
While error analysis research has done much to show the complexity
of acquisition behaviors, it is not without its own problems (Schachter
and Celce-Murcia 1977; Long and Sato 1984). One of the major chal¬
lenges for error analysts is deciding what category to assign a particular
error to. For example, omitting an article in English may quite arguably
be a case of simplification with a Spanish speaker but a case of transfer
with a Korean speaker (see Section 3.2). Moreover, there is reason to
believe that processes such as transfer and simplification interact (see
Section 3.3). Aside from such problems, the error analyses of the 1960s
and 1970s often found some evidence of native language influence, even
while opinions varied about the importance of such influence.
Despite the evidence of cross-linguistic influence in some error anal¬
yses, the credibility of contrastive analysis had been seriously damaged
by the 1970s. Some scholars (e.g., Lee 1968; Wardhaugh 1970) sug¬
gested that contrastive analysis had no predictive power and that con¬
trastive studies could only be useful after the fact. In other words, a
comparison of the native and target languages would be useful for ex¬
plaining why certain errors arise, but in the absence of actual data about
learners’ errors little if anything could be reliably predicted. Other re¬
searchers (e.g., Schachter 1974) nevertheless offered empirical arguments
for the predictive ability of certain kinds of contrastive analysis (Section
6.2). How much a contrastive analysis can or should predict has re¬
mained a controversial question up to the present (Section 3.3).

Universal processes in acquisition?

Aside from casting doubt on the value of contrastive analysis, the em¬
pirical studies of the 1960s and 1970s contributed to skepticism about
transfer in other important ways. The research showed not only the
similarity of some errors made by learners of many different language
backgrounds, but also the similarity of some errors in both first and
second language acquisition, which led many to wonder how different
the two processes really were.
The error analyses of the 1960s and 1970s showed that some types
of errors are common in the emerging second language of speakers of
virtually any native language. For example, the omission of is in cases
such as That very simple is an error made not only by Spanish speakers
but also by speakers of Chinese, Japanese, and other languages (Huang
and Hatch 1978; Itoh and Hatch 1978). In some cases the omission of
is would be predicted from a contrastive analysis, as in the case of the
stereotypical Russian’s monologue given in Chapter 1. Likewise, the
 20 Language transfer

omission of the verb in That very simple by a Chinese speaker might be

predicted from structural facts about the use of copula forms in Chinese.
Nevertheless, the success of this contrastive prediction seems unimpres¬
sive in light of the omission of English copula forms by speakers of
Spanish, Japanese, and other languages that do have present-tense copu¬
las. Moreover, the transfer explanation for such errors seems question¬
able in light of the fact that the omission of is and other copula forms
also occurs in the speech of children learning English as their native
language: for example, That a kitchen (Brown 1973).
For many researchers, such errors are nothing more or less than in¬
dicators of developmental processes found in both first and second lan¬
guage acquisition, and accordingly such errors are often termed
developmental errors. The stages of development, many have argued,
are evident in the relative accuracy that learners show in using particular
structures. For example, Dulay and Burt (1974) observed similar ac¬
curacy orders for several English structures in the speech of two groups
of bilingual children, one group speaking Spanish as their native language
and the other group speaking Chinese. Part of the accuracy order is
given below:
Most accurate contractible copula (’s)12
regular past tense (-ed)
Least accurate possessive (’s)

Thus, errors such as That very simple were less frequent than tense
errors such as I play a new game last night, which in turn were less
frequent than omissions of the possessive marker, as in John hook.
x! ' The methodology of studies of accuracy order rests largely on two
assumptions: (1) that learners have reached a particular developmental
stage if they make very few or no errors with particular types of struc¬
tures; and (2) that the more frequent particular errors are, the further
learners are from attaining a particular stage. If one accepts these as¬
sumptions, the succession of structures that are mastered constitutes a
developmental sequence, that is, a succession of phases of learning to
master new structures. The most important evidence for any particular
sequence comes from longitudinal studies, in which learners’ progress
with various structures is charted over some interval of time (e.g., Hakuta
1976). Though not as persuasive, evidence for a sequence can also come

12 In spoken English the copula may be contracted most of the time. In some cases,
however, the full form of the copula is required. For example, when the copula
follows a word that ends with a strident sound such as /S/ (as in The glass is
fragile), one cannot omit the vowel of the copula. In order to separate difficulties
due to grammar problems from difficulties due to pronunciation problems,
researchers have generally distinguished contractible and uncontractible copulas.
 Earlier thinking on transfer 21

from cross-sectional studies, in which individuals’ relative performances

are measured at a single interval. 13
In the debate on language transfer, the notion of developmental se¬
quences has been a key concept, especially since the appearance of the
first studies of accuracy orders by Dulay and Burt (e.g., Dulay and Burt
1973; Dulay and Burt 1974). Their cross-sectional studies have often
been viewed as strong evidence, for the existence of developmental se¬
quences in second language acquisition, since rather similar results have
been evident in studies of other groups of learners, including adults and
speakers of other languages besides Spanish and Chinese (e.g., Bailey,
Madden, and Krashen 1974; Fathman 1975). Moreover, there are some
similarities in these results with those in child language research (cf.
Brown 1973; de Villiers and de Villiers 1973; Van Patten 1984).^ Al¬
though the accuracy orders shown by children are not the same as those
shown by second language learners, many researchers have seen enough
similarities to begin paying serious attention to the idea that all language
acquisition, first and second, proceeds largely in terms of asevryLfi xecf-
JevefcrprfTental sequenCesru~7Vside from the particular structures studied
''by' Dufay and Burt, developmental similarities between child language
and second language acquisition have been documented in other areas,
such as the evolution of negation patterns (Section 6.3). Moreover, re¬
search on the relative clause patterns of adult second language learners
indicates developmental similarities regardless of the native language of
the learners (Section 6.2).
Largely from the evidence for developmental sequences, Dulay, Burt,
Krashen, and others have argued that transfer plays only a minimal role
in the acquisition of grammar. The study of Chinese and Spanish speak¬
ers by Dulay and Burt (1974) illustrates the basic thrust of such argu¬
ments. Despite the fact that the copula is often absent in Chinese,
speakers of that language were nearly as successful in using the con¬
tractible copula is as Spanish speakers were. And despite the fact that
Chinese and Spanish are very different languages, the accuracy orders
for all structures studied were comparable for both groups. In other
words, the native language appears to have little influence on whether
one target language structure will be easier than another.

13 From a methodological point of view, longitudinal studies are superior to cross-

sectional studies. From a practical point of view, however, cross-sectional studies
are usually much more feasible. In studying most kinds of human behavior over
long stretches of time, researchers often find it difficult to keep track of the
individuals under study.
14 It is not entirely clear how good the evidence is for developmental sequences in the
cross-sectional studies of Dulay, Burt, and others. While problems with the
methods and data interpretation of Dulay and Burt are beyond the scope of the
present discussion, such problems have been discussed by many researchers (e.g.,
Rosansky 1976; Andersen 1977; Hatch 1983; Huebner 1983).
 22 Language transfer

If native language influence does not determine the accuracy orders

or the developmental sequences presumably represented by those orders,
some other influence (or influences) must account for the relative ease
or difficulty of a target language structure. Dulay and Burt saw in the
results of their Spanish and Chinese study evidence
that universal cognitive mechanisms are the basis for the child’s organization
of a target language and that it is the L2 system [the target language] rather
than the LI system [the native language] that guides the acquisition process.
(1974:52)
Krashen (1981) and others have argued that this position is just as
applicable to adult second language acquisition, at least in those cases
in which formal instruction plays little, if any, role. Dulay, Burt, Krashen,
and others claim that second language acquisition is essentially no dif¬
ferent from child language acquisition. While they acknowledge that
there frequently are differences in the success achieved in second lan¬
guage acquisition in comparison with first language acquisition, they see
such differences as related to motivation, anxiety about making errors,
the learner’s environment, and other factors. The claim made about the
fundamental identity of first and second language acquisition is thus
exactly the opposite of the one made by Fries in the 1950s.
The skepticism about transfer did not result simply from the growth
of empirical research, for many scholars, transfer was a behaviorist
— concept that had been appropriated by structural linguists such as Fries,
and in the 1960s and 1970s both behaviorism and structuralism were
becoming unattractive, as major theoretical shifts occurred in linguistics
and psychology (cf. Section 3.1). Where structuralist analyses of gram¬
mar had prevailed during much of the twentieth century, transforma¬
tional syntax now became the dominant form of grammatical analysis
(cf. Fries 1952; Chomsky 1957; Robins 1979). Where the vast differ¬
ences between some languages had made many linguists hesitant to see
any common patterns, the similarities of languages now became clearer
to scholars more and more interested in the notion of language universals
(cf. Section 3.4). Where an earlier generation of scholars thought little
about biological predispositions to learn language, a new generation
looked back to'the old philosophical problem of innate ideas. Where
the behaviorist doctrine of habit formation had once dominated psy¬
cholinguistic investigations (including investigations of language acqui¬
sition), cognitive psychology stressed the creative capacities of human
thought and language. For many scholars, transfer was too much the
theoretical cr-eature of dubious psychology and dubious linguistics.

Some doubts about the case against transfer

The empirical failures of contrastive analysis and the important simi¬
larities between first and second language acquisition did much to bring
 Earlier thinking on transfer 23

the notion of transfer into disrepute. Nevertheless, the most skeptical

positions taken on transfer and contrastive analysis are highly ques¬
tionable on both theoretical and empirical grounds. Since much of this
book presents evidence that rebuts the most skeptical positions, the
shortcomings of those positions will be discussed only briefly in this
section.
There are many theoretical difficulties with the arguments that min¬
imize the importance of transfer. One problem with many of these ar¬
guments is their focus on errors (e.g., Whitman and Jackson 1972).
While errors no doubt provide important evidence for the strength or
weakness of particular native language influences, they are far from being
the only evidence (cf. Section 3.3). Another problem lies in an assumption
frequently made in the consideration of language universals: namely,
that if universal developmental sequences play a major role in acquisi¬
tion, transfer cannot play much of a role. In fact, however, there are
reasons to believe that cross-linguistic influences work in tandem with
the psychological factors governing developmental sequences (Sections
6.2, 6.3). Still another difficulty is the assumption sometimes still made
that theories of transfer are inextricably linked to theories of habit for¬
mation. Yet by no means is there any necessary connection between
such theories (Section 3.1).
One further theoretical shortcoming found in many of the most skep¬
tical positions on transfer also warrants mention: their overemphasis on
morphology and syntax. Even though the motivations have been some¬
what different, there is a clear similarity between the positions on lan¬
guage mixing taken by Muller, Whitney, and others in the nineteenth
century and the positions on transfer taken by Dulay, Burt, and others
in the twentieth. As in the nineteenth century, grammar often continues
to be seen as the “blood and soul” of language (cf. Section 2.1). In
contrast to grammatical transfer, transfer affecting second language pro¬
—
nunciation has been less controversial even though empirical studies
have identified many of the same problems in contrastive analyses of
phonetics and phonology that have been evident in analyses of mor¬
phology and syntax (e.g., Johansson 1973). Too often in polemics on
transfer the evidence from research not only on pronunciation but also
on discourse and vocabulary is either taken for granted or simply ig¬
nored. Obviously the question of grammatical transfer is important and
warrants much of the research that has been undertaken. Nevertheless,
the studies surveyed in Chapters 4 through 7 indicate that: ( 1) transfer
can occur in all linguistic subsystems, including morphology and syntax;
and~J2Eother influences besides transfer affecf aTTsubsystems.
The theoretical shortcomings evident in the more skeptical positions
on transfer are more than matched by empirical shortcomings. In vir¬
tually none of the most extreme polemics on transfer has there been a
careful look at some of the most relevant language contact evidence.
24 Language transfer

Dulay and Burt (1974/1983), for example, equate the bilingual situations
in Weinreich’s study with those in Haugen’s research (which concerned
borrowing transfer), and claim that Weinreich’s research has little to do
with contrastive analysis (cf. Section 2.1). Haugen’s study, it is true,
focuses on borrowing transfer and thus is only indirectly relevant to
transfer in second language acquisition. However, Weinreich’s study
cites a great deal of research not only on borrowing but also on sub¬
stratum transfer (e.g., Schuchardt 1884/1971; Marckwardt 1946; Harris
1948). While Weinreich’s use of the term interference to refer to both
kinds of transfer (and also to code-switching) was no doubt confusing,
many of the cases discussed in his book involve native language influence
on the second language of bilinguals having varying degrees of linguistic
proficiency.
Along with the relatively early studies cited by Weinreich, much of
the empirical research in the 1970s and 1980s has led to new and ever
more persuasive evidence for the importance of transfer in ail subsystems.
A rather large number of studies comparing the grammar, vocabulary,
and so forth of learners with different native languages indicate acqui¬
sition differences attributable to cross-linguistic influence (e.g., Ringbom
and Palmberg 1976; Schachter and Rutherford 1979; Jansen, Lalleman,
and Muysken 1981; Ard and Homburg 1983; Andrews 1984; Appel
1984; White 1985; Schumann 1986; Singler 1988). Furthermore, there
are studies indicating the influence of a second language on the acqui¬
sition of a third (Sections 3.3, 8.3). And with the growth of transfer
research, there have appeared more studies that give some idea of how
transfer interacts with many other factors in acquisition (cf. Chapter 8).
From the nineteenth century on, the standards of evidence for transfer
have been rising, and the empirical support for the importance of cross-
linguistic influences on grammar, vocabulary, pronunciation, etc., is now
quite strong. Before a detailed look is taken at empirical work on specific
subsystems, however, it is worthwhile to consider some basic problems
in doing research on transfer. Chapter 3 addresses several of those
problems.

Further reading
The role of language contact in language change is discussed at some length
by Bynon (1977) in a very readable introduction to historical linguistics.
Surveys by Appel and Muysken (1987) and by Miihlhausler (1986) offer
more detailed discussions of many of the same issues. Though somewhat
dated, an article by Sridhar (1981) is a very thoughtful survey of problems
related to error analysis and transfer. Hakuta (1986) offers useful insights
not only on transfer but also on many other issues important in the study of
bilingualism.
 3 Some fundamental problems
in the study of transfer

There are many theoretical and practical problems that attend the study
of transfer. It is beyond the scope of this chapter to review them all, but
problems in four areas have an especially important bearing on the
discussion in subsequent chapters: definition, comparison, prediction,
and generalization.

3.1 Problems of definition

The terminology used to study language reflects - and sometimes creates —
vexing problems, and in the terminology of second language research, the
term transfer is as problematic as any. The issue of cross-linguistic influ¬
ence is controversial with or without the term, but the long-standing use
of transfer has itself led to differences of opinion. Some scholars have ad¬
vocated abandoning the term or using it only in highly restricted ways
(e.g., Corder 1983; Kellerman and Sharwood Smith 1 986), yet many oth¬
ers continue to use it without restriction. In this section, a definition of the
term transfer will be presented, along with a critique of that definition.
However, before any observations are made about what transfer is (or at
least seems to be), some observations about what transfer is not are
v appropriate.
Transfer isnot simply a consequence of habit formation. A discussion
of contrastive analysis and behaviorism by Carroll (1968) makes clear
that the behaviorist notion of transfer is quite different from the notion of
nativeTanguage influence (cf. Section 2.2). For one thing, the behaviorist
notion of transfer often Implies the extinction of earlier habits, whereas
the acquisition of a second language need not (and normally does not)
lead to any replacement of the learner’s primary language. 1 This and other
considerations suggest that behaviorism may never have been relevant to
the study of transfer (cf. Hakuta 1986). Even though behaviorism has
1 The acquisition of a second language can lead to borrowing transfer and attrition,
and to that extent the behaviorist conception might apply to second language
behavior (cf. Section 2.1). However, in many acquisition contexts the effects of such
attrition are extremely negligible.

25
26 Language transfer

contributed extremely little to the study of transfer since the 1 970s, much
of the dislike of the term transfer comes from its traditional association
with behaviorism. Yet as Kellerman (1984) has observed, there is less and
less danger of people associating transfer with habit formation, which has
in many ways been superseded by concepts from cognitive psychology (cf.
Section 2.2). In fact, behaviorism is now so widely discredited in the field
of psycholinguistics that some leading textbooks in that field give virtually
no attention to behaviorist analyses (e.g., Clark and Clark 1977; Foss and
Hakes 1978). It is worth noting that over a hundred years ago Whitney
(1881) used the term transfer to refer to cross-linguistic influences long
before any linguists thought of linking it to the notion of habit formation.
—
In all likelihood, no amount of scholarly protestation will keep that term
from being used far into the future.
Transfer isjwt^imply interference or without any behaviorist
connotations, the notion of interference does seem applicable in the
description of some aspects of second language performance, such as
phonetic inaccuracies that resemble sounds in the learner’s native lan¬
guage (Section 7.2). Not surprisingly, then, the term interference con¬
tinues to be widely used. Nevertheless, much of the influence of the^
native language (or of some other previously Teamed language) can be
very helpful, especially when the differences between two languages are
relatively few. F<^ example, the number of Spanish-English cognates
(e.g., publico anapublic) is far greater than the number of Arabic-English
cognates.2 Accordingly, native speakers of Spanish have a tremendous
advantage over native speakers of Arabic in the acquisition of English
vocabulary (Section 5.2). The term interference implies no more than
what another term, negative transfer, does, but there is an advantage in
using the latter term since it can be contrasted with positive transfer,
which is the facilitating influence of cognate vocabulary or any other
similarities between the native and target languages.
Transfer is not simply a falling back on the native language. In an
elaboration of an analysis originally proposed in the 1960s, Krashen
(1983:148) claims that:
Transfer . . . can still be regarded as padding, or the result of falling back on
.old knowledge, the LI rule, when new knowledge . . . is lacking. Its cause may
simply be having to talk before “ready,” before the necessary rule has been
acquired.
Tn addition to this claim, which might seem plausible, is another: “Use
of an LI rule... is not ‘real’ progress. It may be merely a production
strategy that cannot help acquisition” (Krashen 1983:148). There are

2 From a historical point of view, the similarity of public and publico is not a cognate
relation in the same sense that mother and madre are cognates. From the learner’s
point of view, however, any lexical similarities may be considered cognates.
 Some problems in the study of transfer 'll

several problems with analyzing transfer as merely a falling back. First,

it ignores the head start that speakers of some languages have in coming
to a new language (cf. Singleton 1987). For example, the similarities. in-
vocabulary, writing systems, and other aspects of English and Spanish
reduce the amount that may be utterly new in English for Spanish speak7 /
ers in comparison with, say, Arabic speakers (cf. Sections 3.3, 5.2, 7A).[
Second, .Krashen’s statements imply that native language influence is
always manifested in some transparent “TT rule.” In fact, however,
native language influences can interact with other influences so that
^sometimes there is no neat correspondence between learners’jnative lan¬
guage patterns and their attempts to use the target languag^cf. Section'
"3.3). Third, Krashen’s claim that transfer may be a mere /‘production
strategy” fails to recognize that crdss-linguisric influences can be bene¬
ficial in listening or reading comprehension/Fourth, Krashen’s analysis
cannot account for the long-term results of language contact in some
settings. In the case of Ireland, for example, learners of English seem to
have fallen back frequently on knowledge of Irish, but such falling back
was never entirely eradicated nor did it halt the wide-scale adoption of
—
English. While the Hiberno-English that bilinguals a’nd later, mono-
—
linguals spoke was often a nonstandard variety, the bilinguals of Ire¬
land had indeed acquired Ehglish (Section 8.3).
Transferors not always native language influence-. Throughout this
book, the phrase native language influence is used as a synonym for
transfer. Such usage, however, is only a convenient fiction. When in¬
dividuals know two languages, knowledge of both may affect their ac¬
quisition of a third (Section 3.3). Most probably, knowledge of three or
more languages can lead to three or more different kinds of source
language influence, although pinning down the exact influences in mul¬
tilingual situations is often hard (Section 8.3). Since knowledge of a
single native language is the most typical basis for substratum transfer,
such cases will have priority in this book.
Although the four conceptions of transfer just discussed do not ade¬
quately characterize the phenomenon, a working definition of transfer
is nevertheless feasible. The following definition of substratum transfer
applies to all subsequent discussions in this book:
Transfer is the influence resulting from similarities and differences between
the target language and any other language that has been previously (and -
'perhaps imperfectly) acquired.
This, it should be stressed, is only a working definition, since there are
problematic terms within the definition. While the word influence seems
appropriate, it is somewhat vague. Just how does the influence work?
Without question, the influence arises from a learner’s conscious or
unconscious judgment that something in the native language (most typ-
28 Language transfer

ically) and something in the target language are similar, if not actually
the same. However, the conditions that trigger judgments of similarity
or identity remain incompletely understood. The term acquired also
remains only partially understood. Various models of second language
acquisition have been proposed, but the time seems distant when scholars
will agree on a definitive model (cf. Ellis 1985).
A fully adequate definition of transfer seems unattainable without
adequate definitions of many other terms, such as strategy, process, and
simplification. Such definitions may presuppose an account of bilin¬
gualism that accurately characterizes relations between transfer, over¬
generalization, simplification, and other second language behaviors. An
adequate account of bilingualism would in turn have to include an
accurate neurological model of language since, presumably, the influence
of one language on another has something to do with the storage of
two knowledge systems within the same brain (Albert and Obler 1979).
Thus, one might plausibly argue that a fully adequate definition of trans¬
fer presupposes a fully adequate definition of language.

3.2 Problems of comparison

The study of transfer depends greatly on the systematic comparisons of
languages provided by contrastive analyses. While it may seem obvious
that any investigation of transfer presupposes such comparisons, Thoma¬
son and Kaufman (1988) cite instances of historical linguists attempting to
study language contact situations without knowing much about the lan¬
guages involved. The essential criteria for sound contrastive analyses are
easy enough to state, but the development of comparisons based on those
criteria has proved to be difficult. As a result, there is considerable varia¬
tion in the quality of crossdinguistic comparisons. Yet no matter how good
a contrastive analysis is, tnore than \ust structural comparisons are neces¬
sary for a thorough understanding of transfer,’since native language influ¬
ence interacts with nonstructural factors. And aside from native— target
language comparisons, a second type of comparison is useful, and often
necessary, to establish the occurrence of transfer: a comparison of the per¬
formance of two or more groups of learners with different native
languages.

Descriptive and theoretical adequacy

An ideal contrastive analysis would provide much of the same infor¬
mation that ideal grammars of the native and target language would.’

3 The use of the term grammar here is in the more general sense - it describes not
just the morphology and syntax of a language but also its phonology, lexicon, etc.
 Some problems in the study of transfer 29

By the criteria discussed by Chomsky (1965), an ideal grammar would

be both descriptively and theoretically adequate, and the same criteria
apply to a contrastive analysis. Accurate and thorough descriptions are
obviously important. Yet just as important are descriptions that embody
sound theoretical principles which enable contrastive analysts to predict
better what will be easy or difficult to learn.
While many contrastive analyses provide useful and sometimes highly
perceptive information about languages they compare, none comes close
to meeting in full the criteria of descriptive and theoretical adequacy.
The descriptive shortcomings of contrastive accounts are quite evident;
even much-studied languages such as English have not been thoroughly
described, and many languages have yet to be studied in any detail.
Consequently, no theoretically adequate grammar exists, since descrip¬
tive adequacy is, as Chomsky has observed, a precondition for theoretical
adequacy. Moreover, the range of unsolved theoretical problems is im¬
mense and has led to very different opinions about the foundations of
grammatical analysis (e.g., Givon 1979; Chomsky 1981; Comrie 1981;
Gazdar et al. 1985; Lakoff 1987).
Despite this lack of consensus, certain theoretical approaches have
found much favor among contrastive analysts and students of second
language acquisition. Chomskyan approaches have long been especially
popular (e.g., Stockwell, Bowen, and Martin 1965; Di Pietro 1971;
James 1980; Sharwood Smith 1986). Scholars have often called both
for rigorous analyses and for a set of analytical principles relevant to a
general understanding of human language, and in the opinion of many
scholars the generative analyses of Chomsky and others speak to such
concerns. Yet whatever a contrastive analysis gains through the use of
Chomskyan formalisms, it also inherits their theoretical and empirical
liabilities. Aside from the analyses by Lakoff, Givon, and the others cited
above, there have been numerous other discussions of linguistic theory
that challenge the assumptions and methods of Chomskyan approaches
(e.g., Labov 1972; Bolinger 1975; Coulmas 1981a). In this book there
will not be a great deal said about Chomskyan analyses, as most of the
research described in it provides important insights about transfer with
little or no reliance on Chomskyan theories and formalisms.4

4 Some researchers (e.g., Flynn 1984) do formulate their analysis within a

Chomskyan framework, but rarely does that framework seem to be the only way to
explain the results the researchers obtained. It might be mentioned that other
varieties of formal linguistics, such as Generalized Phrase Structure Grammar, have
received little or no attention from second language researchers. Whatever
advantages any particular formalist approach may have, the methods of
investigation employed often show many of the same limitations of Chomskyan
approaches, including an overreliance on introspection (cf. Labov 1975; Chaudron
1983; Trudgill 1984).
30 Language transfer

Some problems in contrastive descriptions

Whether contrastive analysts use a Chomskyan or some other approach,
they encounter formidable problems in trying to formulate sound de¬
scriptions. One of the most fundamental problems is idealization, which
is the characterization of the most important aspects of a language with
the elimination of unneeded details. Idealization of linguistic data is
unavoidable since there are many minute variations in the speech of
individuals who consider themselves to be speakers of the same language.
The more idiosyncratic variations in a speech community clearly have
no place in a contrastive description. However, too much idealization
amounts to distortion. For example, contrastive descriptions of Arabic
and English do not always specify the regional varieties in question, and
while such generalized comparisons are often appropriate, there do exist
important differences in, for instance, the pronunciation of Arabic speak¬
ers in Iraq and Egypt. As Broselow (1983) has pointed out, differences
such as those between Iraqi and Egyptian Arabic can result in differences
in learners’ pronunciation of English. Social variation can matter as much
as regional variation in contrastive descriptions (Wolfram 1978; Gonzo
and Saltarelli 1983). For example, Classical Arabic enjoys greater social
prestige than does Egyptian or any of the other regional varieties of
Arabic, and this sociolinguistic fact seems to influence formal and in¬
formal speech not only in the Arabic but also in the English of Egyptians
(Section 8.3).
As noted earlier, criticisms of contrastive analysis have frequently
attacked the Lado-Fries assumption that linguistic difference is equiv¬
alent to linguistic difficulty (Section 2.2). More refined approaches to
contrastive analysis have attempted to deal with this problem by iden¬
tifying the types of differences that will lead to difficulties and the types
that will not (e.g., Stockwell, Bowen, and Martin 1965). For example,
one difference that frequently, if not always, leads to difficulty is that
in which a structure in one language has not one but two (or more)
counterparts in another language. As mentioned previously (Section 2.2),
Spanish, for instance, has two verbs, conocer and saber, that correspond
to different senses of the English verb know, and this lexical discrepancy
constitutes a real area of difficulty that English speakers encounter with
Spanish vocabulary, even though this lexical difference does not seem
to be problematic for Spanish speakers learning English. A fully devel¬
oped contrastive analysis would include an accurate hierarchy of diffi¬
culty, that is, a definitive statement about which contrasts are most and
least likely to cause problems. That hierarchy would have to account
for, among other things, those cases in which similarities between lan¬
guages prove to be more troublesome than some differences.' For ex-

5 A claim is frequently made that similarities between languages are a greater source
 Some problems in the study of transfer 31

ample, the formal resemblance between English embarrassed and

Spanish embarazado (which means “pregnant”) can lead an embarrassed
Englishman to make the embarrassing statement Estoy muy embarazado
(“I am very pregnant”) (Section 5.2).
Another challenge for any contrastive description is the interaction of
linguistic subsystems. Psycholinguistic research has demonstrated a
strong interdependence among discourse, syntax, phonology, and other
subsystems in the comprehension and production of language (e.g., San¬
ford and Garrod 1981; Bock 1982). For example, the production of
discourse has important effects on the pitch contours, pauses, and other
suprasegmental aspects of phonological structure (Section 7.2). Simi¬
larly, the treatment of topics in discourse often has important effects on
word order and other syntactic structures (Section 6.1). Some systematic
analyses of language have taken these interactions into account, but
relatively few contrastive analyses have explored in much detail the
interdependence of various subsystems (e.g., Schachter and Rutherford
1979).

Structural and nonstructural factors

In this book, the term structure is roughly synonymous with the term
tagmeme, a unity of form and function (Pike 1954). In other words,
nothing is a structure unless it has both a form (some definite pattern)
and a function (some definite use). Such a notion of structure has been
implicit in much contrastive research, but with the increased interest in
discourse from the 1970s onward the boundaries of contrastive analysis
no longer seem as clear-cut as they once did.6 There do exist units of
discourse structure, such as paragraphs and narratives, that clearly in¬
volve form-function relations analogous to those in phonology, syntax,
and other subsystems (Section 4.2). Nevertheless, discourse involves
much more than what a purely structural analysis covers. For example,
politeness is an aspect of discourse that is very important for any cross-
linguistic comparison, but it is also one that involves many nonlinguistic
factors. Differences between, say, apologies in English and apologies in
Japanese may reflect differences in cultural beliefs and values (Section
4.1). In such areas, the boundaries between anthropology, philosophy,

of difficulty than differences (e.g., Pica 1984). While some similarities doubtless can
occasion great difficulty, it is an oversimplification to deem similarities to be the
. greater problem. Such claims generally rely on data from error analyses (Section
2.2), but do not account for other relevant evidence. If the claim that similarities
cause more difficulties were fully true, some very improbable events would be
normal; for example, students literate in Chinese would have a big advantage over
students literate in Spanish in learning the English alphabet (cf. Section 7.4).
6 The notion of structure here includes categories (e.g., past tense), rules (e.g., number
agreement), relations (e.g., word order), and vocabulary items.
32 Language transfer

and linguistics become very hazy. In attempting to deal with the question
of just what a contrastive analysis of discourse should contrast, some
researchers have proposed intricate systems encompassing a myriad of
variables (e.g., Hartmann 1980). However, practical tests of the merit
of one system over another would be extremely complicated.
Another problematic relation between structural and nonstructural
factors is language distance, or the degree of similarity between two
languages. As discussed in Chapter 1, intuition suggests that some lan¬
guages are more closely related than others. For example, English seems
more closely related to French than to Eskimo, and Spanish seems more
closely related to French than to English. While resemblances such as
those just cited often arise from various historical relationships (Section
2.1), there are clear resemblances between languages whose historical
relationships are not certain (e.g., Korean and Japanese), and even be¬
tween languages having no known historical relationship (e.g., Japanese
and Quechua). Objective measures of the distance between languages
can be established through careful comparisons of structural similarities,
which would show, for example, that the patterns of noun phrases in
Spanish are more like those of French than like those of English. While
the cumulative similarities between languages might be quantified in an
objective manner, the subjective judgments of language distance by learn¬
ers can matter considerably. For example, although English and Dutch
have many cognates, Dutch students of English appear to be frequently
skeptical about the possible use of certain cognate forms (Section 8.3).
Other nonstructural factors, which are discussed in Chapters 4 and 8,
also suggest that there is more to transfer than what a good contrastive
analysis will indicate.

Comparison of performances
While a contrastive analysis is a necessary condition to establish the
likelihood of transfer, it is not always a sufficient condition. By com¬
paring the performances of speakers of at least two different native
languages, researchers can better determine any effects of negative
transfer. Such comparisons are frequently necessary since transfer in¬
teracts with other factors, and explanations based only on contras¬
tive analyses are sometimes misleading. For example, there are
problems in attributing only to Persian language influence the errors
that Persian speakers make by using resumptive pronouns, as seen in
the last word in the sentence I know the man that John gave the
book to him. A contrastive analysis of relative clauses in Persian and
English does suggest that the use of resumptive pronouns would be
a Persian speaker’s error; Persian relative clauses often have resump¬
tive pronouns. Simply relying on a contrastive description would be
 Some problems in the study of transfer 33

misleading, however, since such errors are also made by speakers of

languages not having resumptive pronouns in equivalent relative
clauses (e.g., most Romance languages). Nevertheless, comparisons of
speakers of languages such as Persian, on the one hand, and speak¬
ers of languages such as French, on the other, indicate differing pro¬
pensities to use such pronouns in the target language; in other
words, the presence of resumptive pronouns in the native language
can lead to a greater use of them in the target language (Section
6.2).
In some cases the need for a formal comparison of performances is
not very great. Certain spelling errors made by ESL students reflect
characteristic pronunciation problems of speakers of particular native
languages. For example, the spelling of playing as blaying is more likely
to be the spelling error of an Arabic speaker than of a Spanish speaker
(cf. Sections 3.3, 7.4). Moreover, while grammatical errors such as re¬
sumptive pronouns and the omission of the copula occur in the English
of speakers of many different languages, other errors are characteristic
of speakers of very few languages. For example, the anomalous verb
form found in Hiberno-English sentences such as He’s after telling a lie
(“He’s told a lie”) is rarely if ever cited in error analyses of the English
of speakers of other languages. While formal comparisons of such spell¬
ing and verb errors are possible, the distinctiveness of the errors amounts
to an implicit comparison. In cases where an error is rather common
among speakers of several different native languages, explicit compar¬
isons are preferable.
While explicit comparisons are often desirable in determining negative
transfer, they are indispensable in determining positive transfer. Research
on articles clearly illustrates this need. Like the omission of the copula,
the omission of articles is common even among speakers of languages
having articles. For example, the following error comes from a native
speaker of Spanish: Picture is very dark (Schumann 1978). A contrastive
analysis of Spanish and English would not predict this error since there
is a word-for-word grammatical correspondence between Spanish and
English:
El cuadro es muy oscuro.
The picture is very dark.
Because of examples such as Picture is very dark, many linguists be¬
lieve that the native language helps learners very little in their acqui-

7 The context of this statement is not given, so it is not dear that cuadro is the best
translation of picture. However, this uncertainty does not affect the grammatical
analysis of article use in the two languages. It should be noted that another (and
highly probable) Spanish translation involves a different word order in which the
subject appears in sentence-final position: Es muy oscuro el cuadro (cf. Section 6.1).
34 Language transfer

sition of English articles. Spanish-speaking learners, it might be, ar¬

gued, learn correct use of articles (e.g., The picture is very dark) not
as a result of any positive transfer from Spanish but simply as the
result of sufficient exposure to the definite article in English. Several
comparative studies, however, suggest that this argument is dubious.
Three studies of student performances on written tests of ESL indi¬
cate that speakers of languages having articles tend to use them
more accurately than do speakers of languages not having articles
(Oller and Redding 1971; Kempf 1975; Ringbom 1976). Moreover,
speaking tests of learners’ abilities to use articles produced results
similar to those of the written tests (e.g., Dulay and Burt 1974;
Fathman 1977), as did comparisons of learners’ speech in naturalis¬
tic contexts (Zobl 1982; Gilbert 1983). The similarity of results in
several studies thus suggests that some positive transfer occurs even
when the contrastive prediction is stated in a very crude way (e.g.,
“Having articles in the native languages will make articles in the tar¬
get language easier to acquire”). It is likely that more refined com¬
parisons of learner performances would show even better evidence of
positive transfer. Errors of article omission in the English of Spanish
speakers appear to correspond closely - though not entirely
areas of contrast between English and Spanish: for example, where
— to

indefinite articles are used in English but not in Spanish (Andersen

1977).
The existence of differences in acquisition patterns seen in comparative
studies may not in all cases be due to cross-linguistic influence alone. It
is conceivable, for example, that some of the differences reflect transfer
of training, such as schooling that encourages learners to translate from
their native language (cf. Krashen 1983). Such an explanation, however,
cannot explain a number of findings in the comparative studies. For one
thing, some of the studies are of children who probably have not had
formal training in translation (e.g., Zobl 1982), and some other studies
are of adults having little or no schooling (e.g., Schumann 1986). More¬
—
over, other research suggests that schooling may decrease not increase
— the likelihood of negative transfer (Section 8.3). Even if schooling did
invariably encourage all forms of cross-linguistic influence, transfer-of-
training explanations could not account for results such as those seen
in work on the perception of tones (Section 7.2) and on alterations of
structures to forms that do not closely resemble structures in either the
native or target language (Section 3.3).
Regardless of how much or how little any training in translation or
other second language behaviors may encourage substratum transfer,
there is strong evidence that different acquisition patterns are associated
with different native language backgrounds. Not only articles but several
other areas of second language performance have been the focus of
 Some problems in the study of transfer 35

comparative research - over three dozen empirical studies which provide

explicit comparisons of learners’ behaviors and which show effects of
transfer are cited in this book/

3.3 Problems of prediction

Forecasts and explanations
The literature on contrastive analysis frequently refers to predictions
that are determined by cross-linguistic comparisons. In reality, however,
the “predictions” of learners’ behavior are often derived after the fact:
What counts as a prediction is frequently based on data about learner
performances already known to a linguist who has interpreted the data
record with the help of cross-linguistic comparisons (cf. Wardhaugh
1970). Such a record does have the predictive value that other kinds of
knowledge of past events have. A record of temperatures in the month
of July in Texas for several years is likely to be a good predictor of
average temperatures in Texas next July. Similarly, a record of errors
in French made by English-speaking students in previous years can serve
as a predictor of errors that English-speaking students will make in a
French course next year. However, such predictions are clearly different
froin the kind made before the actual performance of learners is
investigated.
One could argue, as Wardhaugh and others have, that the ultimate
test of a contrastive analysis is one in which the predictions are based
only on comparisons of the linguistic systems. Successful contrastive
analyses developed in such a way could make it easier to discover general
principles for making sound predictions about transfer in any language
contact situation that might arise in the future. Yet while such predictions
may constitute the ultimate test, the explanatory power of any cross-
linguistic comparison is another important test. In other words, a good
contrastive analysis should make it easier to explain why transfer will
or will not occur in any given instance. Without a clear understanding
of the conditions that occasion transfer, there is little hope of developing
highly sophisticated contrastive analyses that make predictions of the
kind discussed by Wardhaugh. Thus, while good predictions may be the
ultimate goal, good explanations are a crucial part of achieving that
goal. In view of the descriptive and theoretical problems of comparison
already discussed (Section 3.2), it is not surprising that contrastive anal-

8 Other problems involving comparison, such as the notion of translation

equivalence, are discussed in Section 3.4. Sridhar (1981) discusses in some detail
still other problems of comparison.
36 Language transfer

yses have yet to succeed completely as either explanations or forecasts.

Part of the challenge in developing better explanations and predictions
is to understand better the many possible outcomes of cross-linguistic
similarities and differences.

A classification of outcomes
The following classification offers some idea of the varied effects that
cross-linguistic similarities and differences can produce:
I. Positive transfer
II. Negative transfer
A. Underproduction
B. Overproduction
C. Production errors
D. Misinterpretation
III. Differing lengths of acquisition
POSITIVE TRANSFER
The effects of positive transfer are only determinable through compar¬
isons of the success of groups with different native languages (Section
3.2). Such comparisons often show that cross-linguistic similarities can
produce positive transfer in several ways. Similarities between native
language and target language vocal^lary can reduce the time needed to
develop good reading comprehension, as discussed later/ Similarities
between vowel syyterps can make the identification of vowel sounds
easier (SectioX 7.2)! Similarities between writing systems can give learners
a head start in reading and writing in the target language,(SectionJ7.4).
And similarities in syntactic>structures can facilitate the' acquisition of
grammar: Learners speaking a language with a syntax similar to that of
the target language tend to have less difficulty with articles, word order,
and relative clauses (Sections 3.2, 6.1, 6.2). Future research is likely to
show that cross-linguistic similarities in other areas will also promote
acquisition.
NEGATIVE TRANSFER
Since negative transfer involves divergences from norms in the target
language, it is often relatively easy to identify. Although negative transfer
tends to be equated with production errors, there are other ways in
which an individual’s second language performance may differ from the
behavior of native speakers.
Underproduction. Learners may produce very few or no examples of
a target language structure. Often the examples learners produce result
in comparatively few errors, but if the structure is more infrequent than
it is in the language of native speakers, the infrequency constitutes a
 Some problems in the study of transfer 37

divergence from target language norms. There is good evidence for one
form of underproduction related to language distance: avoidance. If
learners sense that particular structures in the target language are very
different from counterparts in the native language, they may try to avoid
using those structures. Schachter ( 1 974) found that Chinese and Japanese
students of ESL tended to use fewer relative clauses than did students
whose languages have relative clause structures more like those of En¬
glish (Section 6.2). Similarly, Kleinmann (1977) found evidence of avoid¬
ance involving other structures (Section 8.1).

Overproduction. Overproduction is sometimes simply a consequence

of underproduction. For example, in an effort to avoid relative clauses,
Japanese students may violate norms of written prose in English by
writing too many simple sentences. Overproduction can also arise for
other reasons, however. For example, the use of apologies appears to
be more frequent in American English than in Hebrew, and English
speakers learning Hebrew appear to follow the norms of their native
language in making apologies (Section 4.1).

Production errors. In speech and writing there are three types of errors
especially likely to arise from similarities and differences in the native
and target languages: (1) substitutions, (2) caiques, and (3) alterations
of structures. Substitutions involve a use of native language forms in the
target language. For example, Ringbom (1986) noted the following use
of the Swedish word bort (“away”) in an English sentence written by a
native speaker of Swedish: Now I live home with my parents. But some¬
times I must go bort.
Caiques are errors that reflect very closely a native language structure.
For example, Fantini (1985) notes the following sentence spoken by a
Spanish-English bilingual child:
Vamos rapido a poner el fuego afuera.
Let’s quickly put the fire out.
The child made a literal translation of the English expression put the
fire out, which normally translates into Spanish as extinguir el fuego.
Aside from such idiomatic expressions, certain word-order errors can
also be evidence of caiques. For example, an error made by a Spanish¬
speaking ESL student shows the same word order as the translation
equivalent in Spanish: the porch of Carmen, as opposed to the more
natural English phrase, Carmen's porch (Section 6.1).
Substitutions and caiques are frequently the types of errors to which
writers on bilingualism refer when they discuss transfer errors, and these
types might suggest that transfer always involves an obvious correspon¬
dence between the native and target languages. Krashen (1983), for
38 Language transfer

example, characterizes transfer as a falling back on some “LI rule”

(Section 3.1). While this assumption holds in the case of errors due to
caiques and substitutions, it cannot explain some of the most important
cases of cross-linguistic influences which involve alterations of structures,
as seen, for instance, infhypercorrections. Sometimes hypercorrections
are overreactions to a particular influence from the native language. For
example, Arabic speakers occasionally make ESL spelling errors that
involve substitutions of the letter b for the letter p, as in blaying (Section
7.4); however, Arabic speakers also use p inappropriately (e.g., in hapit)
in mistaken attempts to avoid b/p substitutions (Ibrahim 1978). Other
alterations resemble hypercorrections insofar as they do not reflect any
direct influence from the native language. For example, Schachter and
Rutherford (1979) discuss cases of Chinese and Japanese students in¬
appropriately using English syntactic structures involving forms such as
There is and It is, as in: There were many new patriots in my country
gathered together and established a new country. Schachter and Ruth¬
erford argue that although Chinese and Japanese do not have syntactic
structures comparable to There is and It is, sentences such as the one
just cited reflect discourse influences from those languages. Still other
cases of alterations may lead to phonological errors. For example, Thom¬
ason (1981) discusses errors observed in the Serbo-Croatian stress pat¬
terns of Hungarians; the stress rule reflects influence from Hungarian,
but the stress rule applied by Hungarian learners does not exist either
in Serbo-Croatian or in Hungarian.9

Misinterpretation. Native language structures can influence the inter¬

pretation of target language messages, and sometimes that influence leads
to learners inferring something very different from what speakers of the
target language would infer. Variant interpretations can arise from mis¬
perceptions of target language sounds that become categorized in terms
of native language phonology (Section 7.2). Misinterpretations may also
occur when native and target language word-order patterns differ (Sec¬
tion 6.1) and when cultural assumptions differ (Section 4.2).
DIFFERING LENGTHS OF ACQUISITION
The distinction between positive and negative transfer is useful, but the
study of these types of transfer tends to focus on specific details and not
on the cumulative effects of cross-linguistic similarities and differences
on the acquisition process. One way of assessing such effects is to look
at the length of time needed to achieve a high degree of mastery of a
language. As noted in Chapter 1, native speakers of one language often

9 Caiques are the only type of error invariably reflecting native language influence;
substitutions and alterations may arise from sources besides the native language.
 Some problems in the study of transfer 39

believe that they will find certain other languages especially hard to
learn. While little research has been carried out to demonstrate the
validity of that belief, some relevant evidence does exist, such as the
lengths of language courses offered to members of the U.S. diplomatic
corps. The following list shows the maximum lengths of intensive lan¬
guage courses at the Foreign Service Institute (1985) of the U.S. State
Department:
Language Number of Weeks
Afrikaans 24
Amharic 44
Arabic 44
Bengali 44
Bulgarian 44
Burmese 44
Chinese 44
Czech 44
Danish 24
Dari 44
Dutch 24
Finnish 44
French 20
German 20
Greek 44
Hebrew 44
Hindi 44
Hungarian 44
Indonesian 32
Italian 20
Japanese 44
Korean 44
Lao 44
Malay 32
Norwegian 24
Filipino 44
Polish 44
Portuguese 24
Rumanian 24
Russian 44
Serbo-Croatian 44
Spanish 20
Swahili 24
Swedish 24
Thai 44
Turkish 44
Urdu 44
In all of the FSI language courses listed, the aim is to develop students’
linguistic skills to a high level of proficiency that is comparable in each
of the languages. For example, students who spend twenty-four weeks
40 Language transfer

studying Swedish are expected to be as proficient as students who spend

forty-four weeks studying Finnish. In each course, students spend the
same amount of time in class each week (thirty hours), and there are
no significant differences in the language-learning aptitudes of the groups
studying any particular language.10 Accordingly, the most straightfor¬
ward explanation for the varying FSI course lengths is that the languages
themselves are of varying difficulty for students who (usually) are native
speakers of English. The notion of language distance is clearly relevant
to differences seen in the list (cf. Section 3.2). The least difficult lan¬
guages, as determined by course lengths, are mainly Germanic and Ro¬
mance languages, which are similar to English in many respects." The
most difficult languages are diverse, but in general they share fewer
structural similarities with English.
Such differences in course lengths have occasionally been noted,
but in only a few discussions of transfer (e.g., James 1971; Ringbom
1987) has there been much attention given to the issue of time. This
issue is, however, quite important, since it indicates that there may be
a fundamental difference between first and second language acquisi¬
tion. Linguists by and large agree that children acquire their native
languages in approximately equal periods of time. That is, both
Turkish-speaking children and Italian-speaking children, for example,
master the essentials of linguistic structure in about the first five years
of childhood, even though not all the particulars of each language
prove to be equally easy (Slobin 1982). On the other hand, the FSI
figures suggest that it will take an English-speaking adult over twice
as long to become highly skilled in Turkish as to become highly
skilled in Italian.12
There is, however, a counterargument to claims about the importance-

10 The lengths of shorter FSI courses are not cited since they often reflect
administrative and not pedagogical considerations. In the case of Arabic and some
other languages, certain regional varieties are taught in separate courses but these
show no divergences in the overall pattern of course length. I would like to
express my gratitude to Willow Shlanta and Hedy St. Denis of the FSI for this
information and for help in interpreting it.
11 Actually, the course-length data give only a conservative estimate of the difference
in difficulty of various languages. Oxford and Rhodes (1988) discuss estimates
that suggest that it will take ninety-two, not forty-four weeks, to reach a rather
high level of proficiency in Arabic, Chinese, and Japanese comparable to what
learners studying French, Spanish, or Swedish can achieve in twenty-four weeks
(cf. Ringbom 1987).
12 The greater amount of time needed by English speakers to learn Turkish may well
be due in part to greater cultural differences (cf. Sections 4.1, 4.2). However, it is
highly instructive that FSI course lengths differ between languages spoken in areas
where the cultures are very similar (e.g., Hungarian versus Rumanian, and Finnish
versus Swedish). Thus, linguistic structure seems to play more of a role in course
length than culture does.
 Some problems in the study of transfer 41

of transfer that are based on course-length data. As Littlewood (1973)

has noted, one might argue that “correct” methods of teaching could
eliminate the differences in course lengths that seem to be necessary.
Yet, while the correct-methods argument is taken seriously by some
researchers (e.g., McLaughlin 1978), the logic of the argument leads to
predictions that are, to say the least, implausible. As noted earlier, the
difference between Spanish and Portuguese is close enough to be con¬
sidered a dialect difference (Section 2.1). Nevertheless, the correct-meth¬
ods argument would lead one to predict, for example, that some teaching
method would enable Spanish-speaking students to learn Japanese
as quickly as they can Portuguese. Such a method would be truly
miraculous.
Aside from course-length data, another kind of evidence that sug¬
gests a strong relation between language distance and length of acqui¬
sition comes from a study of reading comprehension of Dutch by
English-speaking university students (Singleton and Little 1984). In
that study, none of the students had had any instruction in Dutch,
but one subgroup had already had some instruction in German, and
that subgroup showed a better understanding of the Dutch text that
they were given to read. The other subgroup in the study did succeed
in understanding some of the text (largely because of a modest num¬
ber of Dutch-English cognates), but they did not have the advantage
of the subgroup that had studied German, a language highly similar
to Dutch.1’ These results clearly support the belief that several years
of study of one foreign language can greatly reduce the time needed
to acquire a similar language.

Transfer and simplification J

The taxonomy of transfer effects in the previous section is not compre¬
hensive, largely because the relation between transfer and other processes
in second language acquisition remains only partially understood^ Dis¬
tinguishing between transfer and simplification is often quite difficult.
There are cases in which cross-linguistic influence is not a very compelling
—
explanation for example, the omission of the English copula by Spanish
speakers (Section 2.2). On the other hand, the omission of pronouns
seen in the following statements of a Spanish speaker may or may not
reflect transfer:
In Saturday no like, no time, watch TV. . .
On Saturday I don’t like [to go to the movies], I don’t have any time. I watch
TV...(Giv6n 1984b: 124)

13 Some of those in the non-German subgroup had studied other foreign languages,
but such language study helped them little.
42 Language transfer

Spanish frequently allows pronoun omission, and a study of learners’

grammaticality judgments by White (1985) indicates that Spanish speak¬
ers are more tolerant of pronoun omissions than are speakers of French,
which, like English, does not usually allow pronoun omission. Yet pro¬
noun omissions can result from other influences, such as interactions
between discourse and syntax (Section 6. 1 ). It remains unclear how much
of a role native language influence has in the omission of pronouns,
prepositions, articles, and other structures in particular acquisition con¬
texts. Such uncertainty is reflected in the conflicting estimates of native
language influence seen in error analyses (Section 2.2); as Ellis (1985)
has suggested, such estimates often seem to reflect theoretical biases on
the part of the researchers. However, comparative studies such as those
of article usage offer hope of further understanding. When cross-
linguistic differences have some real influence, there should be evidence
in the form of different performances on the part of groups of learners
speaking different languages. Comparative research in fact indicates that
transfer and simplification may sometimes converge in some errors, such
as the omission of articles (Section 3.2).

Individual outcomes
Up to this point, the discussion of predictions has not distinguished
between the performance of groups and the performance of individuals.
Since contrastive analyses compare linguistic systems, they are more
relevant to collective than to individual behavior. Contrastive analysts
have a hard enough time in predicting, for example, the range of possible
ESL errors of Spanish speakers as a group. The difficulty of predicting
the specific errors that any particular Spanish speaker will make is ob¬
viously far greater (cf. Lee 1968). Individuals vary in many ways, in¬
cluding in their experience and aptitude for learning languages, and such
variation can definitely affect transfer (Section 8.1).
The effects of individual variation make any contrastive prediction
subject to probabilities. Contrastive analysts have often been criticized
for hedging their predictions with words such as probably and tend to
(cf. Dickerson 1974). However, such hedges reflect a strong sense of
realism about human behavior. The physical sciences have yet to develop
to the point where meteorologists can flawlessly predict tomorrow’s
weather or where geologists can long foresee any occurrence of an earth¬
quake. It should thus come as no surprise that in the study of complex
individuals who speak complex languages, predictions are statements of
probabilities. Progress, such as it may come, will result from refining
contrastive predictions so that they more frequently tally with actual
outcomes.
 Some problems in the study of transfer 43

3.4 Problems of generalization

The discovery of valid generalizations about transfer depends very much
on the discovery of valid generalizations about the nature of language,
that is, about language universals. Along with structures found in all
languages or in most languages, structures that distinguish certain types
of languages from others are the focus of many universalist investigations
and are often important for the understanding of cross-linguistic
influences.

Language universals
Two of the most distinctive approaches to the study of universals are
associated with the linguists Noam Chomsky and Joseph Greenberg. As
Comrie (1981) observes, the Chomskyan approach favors the intensive
analysis of one language as part of an effort to identify abstract principles
of a Universal Grammar, whereas the Greenbergian approach favors
cross-linguistic comparisons. Much, though by no means all, of the work
in the Chomskyan approach has focused on various characteristics of
the syntax of standard written English. In contrast, Greenbergian anal¬
yses have generally focused on the cross-linguistic variations seen in
particular structures such as word order. Some researchers (e.g., Haw¬
kins 1983) have attempted to combine both approaches, but many in¬
vestigators of universals have favored one approach over the other.
In some respects, the Chomskyan approach is the more ambitious of
the two since it advances many more claims about language structure,
language acquisition, and linguistic theory y One key hypothesis is that
Universal Grammar is a biological inheritance which simply requires
activation in child language acquisition. Just as there seems to be a
biological “program” that guides infants in their efforts to walk, there
seems to exist, according to Chomskyan views, a program that guides
them in their efforts to talk. With its basis in Universal Grammar, the
language program is generally successful: Under normal conditions chil¬
dren will inevitably learn to talk, just as under normal conditions they
will inevitably learn to walk.14 The interest of some Chomskyan linguists
in child language acquisition has intensified in recent years, and many
linguists who differ in their adherence to Chomskyan views have never¬
theless agreed about the importance of Universal Grammar in language
acquisition (cf. Wexler and Culicover 1980; Bickerton 1981). Still more

14 Examples of abnormal conditions include such cases as where children are severely
retarded or deprived of opportunities to speak with others. Even in cases where
their speech organs are hopelessly damaged, children can acquire a sign language
as highly structured as any spoken language (de Villiers and de Villiers 1978).
44 Language transfer

recently, Chomskyan analyses have led to empirical investigations in

second language acquisition that aim to determine the extent to which
Universal Grammar is still “available” to guide the progress of adults
learning a second language (e.g., Flynn 1984; Eubank 1986).
The Greenbergian approach involves relatively few theoretical as¬
sumptions, but the cross-linguistic regularities identified in that approach
have provided the basis for much research on grammatical theory and
language acquisition. Greenberg’s own work on word order exemplifies
the usefulness of cross-linguistic surveys (Greenberg 1966). The basic
word order of English is one in which grammatical subjects precede
verbs (or verb phrases), which in turn precede objects, and thus the
abbreviation SVO characterizes the canonical order of constituents in
an English clause (e.g., John bought the car). While the SVO order is
quite common in the world’s languages, Greenberg found two other
orders also to be common:
Order Examples Cross-linguistic frequency
VSO Irish, Classical Arabic Somewhat common
SVO English, Russian Very common
SOV Persian, Japanese Very common
VOS Malagasy Rare
OVS Hixkaryana Very rare
OSV Apurina Very rare15
At first glance there would seem to be little “universal” in such findings.
However, the fact that the first three word-order types (VSO, SVO,
SOV) account for the vast majority of languages in Greenberg’s survey
is itself highly significant. It appears that subjects tend to precede objects
in most languages. That tendency does not constitute a universal in the
strictest sense of the word, but it is too consistent a regularity to be the
result of chance; there seems to be a strong preference for having subjects
appear early in sentences (Keenan 1978). Greenberg found, moreover,
that other word-order patterns were often predictable from basic word
order. For example, one can predict with considerable accuracy that an
SOV language will use postpositions instead of prepositions, as in Jap¬
—
anese (e.g., tookyoo ni “Tokyo to”), and one can predict with even
more accuracy that a VSO language will use prepositions (e.g., in Irish
—
sa bhaile “at home”). Linguists frequently put such predictions in the
form of implicational statements. Thus one would state, for example,
that if a language is VSO it will also have prepositions (cf. Hawkins
1983).

15 Scholars have often been skeptical about the existence of OSV languages, but there
is growing evidence that some languages in the Amazon region rely primarily on
OSV (Derbyshire 1986). A discussion of the notion of basic word order appears in
Section 6.1.
 Some problems in the study of transfer 45

The great interest in universals during the last thirty years or so has
undoubtedly led to a greater appreciation of both the complexity and
the unity of human languages. Nevertheless, that interest has not resulted
in the discovery of many facts about linguistic structure that are universal
in the strictest sense of the term, that is, facts which hold for all human
languages. Linguists usually do concede the truth of some universalist
statements, such as the claim that all languages have vowels. However,
the number of uncontroversial claims is small, and what agreement there
is in the study of universals has generally been about implicational uni¬
versals, which provide useful information for the study of language types.

Linguistic typologies
Languages may be classified in many different ways. Typology, the study
of such classifications, benefits work in many fields, including historical
linguistics, grammatical theory, and contrastive analysis. A very rudi¬
mentary example of a typological comparison of English, Classical Ar¬
abic, and Thai is given below:
Inflectional Basic Resumptive Lexical
Language morphology word order pronouns? tones?
English Simple SVO No No
Arabic Complex VSO Yes No
Thai Negligible SVO No Yes
The comparison indicates various structural similarities and differences.
Among the ways that Arabic, for example, differs from English and Thai
is in its intricate system of inflections to express gender, number, and
other categories. On the other hand, Thai is different from both English
and Arabic in that it uses tones to distinguish meanings of words (Section
7.2).
Typological analyses contribute to the study of transfer in three ways.
First^they provide a basis for estimating language distance. While the
example given cites only four structural characteristics, it does suggest
that the language distance between Thai and English may be smaller
than the distance between Thai and Arabic. Second, typological analyses
encourage the study of transfer in terms of systemic influences. For
example, research indicates that speakers of Japanese sometimes have
difficulty both with word order and relative clause structure in English
(Sections 6.1, 6.2). Since there is a fairly strong implicational relation
between word order and relativization, the difficulties that Japanese
learners have in these two areas may be related. Ejnal1y,\typological
analyses allow for a clearer understanding of relations Eetween transfer
and developmental sequences. For example, Greenbergian research on
negation suggests that certain patterns of negation occur much more
46 Language transfer

often than others. These typologically common patterns are also frequent
in first language acquisition both as errors and as correct forms, and in
second language acquisition these negation patterns may sometimes re¬
flect native language influences, sometimes developmental factors, and
sometimes perhaps both transfer and developmental factors (Section
6.3).
Typologically common features give clues to universal preferences in
linguistic structure. For example, the five-vowel system of Spanish is
extremely common, whereas the eleven-vowel system of Vietnamese is
rare (Maddieson 1984). The reasons for widespread preference for the
“Spanish” system probably reflect both physical and psychological fac¬
tors. The five vowels in Spanish (/i/, /e/, /u/, /o/, /a/) are among the easier
ones that the human vocal tract can produce, and their acoustic dis¬
tinctiveness makes perceptual confusions between them unlikely. A wide
range of physical and psychological factors must be involved in typo¬
logically common (or universal) characteristics of linguistic structure (cf.
Gass and Ard 1984). Accordingly, the hierarchy of difficulty in a refined
contrastive analysis would have to take into account the significance of
such factors (Section 3.2).

Universalist assumptions
Even though their focus is often on particular kinds of human languages,
typological analyses and contrastive analyses often involve universalist
assumptions. One of the most important is the assumption that there
are categories applicable to the analysis of all languages. For example,
Greenberg’s classification of languages in terms of basic word order
assumes that categories such as Subject are universal. Many researchers
accept that assumption, but there is no consensus about a set of necessary
and sufficient conditions to define the category Subject, ^ome have ar¬
gued against viewing linguistic categories as sets of necessary and suf¬
ficient conditions, and attempt to define categories such as Subject and
Object as highly correlated bundles of syntactic, semantic, and discourse
properties (e.g., Lakoff 1972; Keenan 1976; Comrie 1981; Bates and
MacWhinney 1982). Whatever the merits of such proposals, it is clear
that until the general nature of linguistic categories is well understood,
all analyses assuming the universality of particular categories are highly
tentative (cf. Lakoff 1987).
Another crucial universalist assumption in typological and contrastive
analyses is that there are certain meanings that are equivalent in the
discourse and semantic systems of all human languages. Without some
notion of translation equivalence, there can be no useful cross-linguistic
comparison of structures. In constructing a typology of syntactic ne¬
gation, for example, linguists assume that negation is a logical construct
 Some problems in the study of transfer 47

found in all languages (cf. Sections 5.1, 6.3). Yet how much translation
equivalence exists between the discourse and semantic systems of two
languages is problematic. Areas in discourse such as politeness expres¬
sions show considerable variation, and not all areas of semantic structure
are necessarily uniform (Sections 4.1, 5.1, 5.2). The most extreme forms
of linguistic relativism have ascribed highly distinct worldviews to mem¬
bers of different cultures and have minimized similarities not only in
meanings in different languages but also in speakers’ perceptions of the
world (Section 4.1). Yet despite many important differences in beliefs
and attitudes in different cultures, such differences probably affect lan¬
guage less than what extreme relativist claims suggest. In areas such as
color vocabulary, a long-cherished topic in relativist analyses, there are
reasons to believe that semantic universals exist (Berlin and Kay 1969;
Rosch 1973; Mervis and Roth 1981).
There is one assumption of universalist analyses that is found in many
other kinds of research - that a reasonably good sample of behavior
allows for reasonable inferences about all such behavior. However,
“large” samples in Greenbergian analyses usually cover only a small
fraction of the known languages of the present and past, and they ob¬
viously cannot say much about undiscovered languages or about the
many languages that once existed but that were never written down.
Similarly, studies in second language acquisition sample only a small
number of the language contact situations in the world, and caution is
certainly advisable in claims about how universal the results of such
studies are. As the discussion of discourse in the next chapter indicates,
there is good reason to be cautious in assessing the importance of either
universals or transfer.

Further reading
Texts by Clark and Clark (1977) and Foss and Hakes (1978) are somewhat
dated introductions to the psychology of language, but are still worth
consulting. While these texts offer almost no discussion of second language
research, a survey by Hatch (1983) provides a useful look at the implications
of psycholinguistic studies for second language acquisition. James (1980)
raises a number of important issues related to the development of contrastive
analyses. One of the best discussions of typologies and universals is a text by
Comrie (1981).
4 Discourse

Of all the areas of contrastive analysis, cross-linguistic comparisons of

discourse are probably the most challenging. As noted in Chapter 3,
discourse analysis involves a wide array of nonstructural as well as
structural characteristics, and the boundaries between contrastive dis¬
course and other disciplines such as cultural anthropology are not clear¬
cut.1 Moreover, models of discourse are necessarily complex and are
thus quite difficult to test (Section 3.2). Yet despite the lack of compre¬
hensive cross-linguistic descriptions, there has been progress in the study
of contrastive discourse. Researchers now have detailed information
about specific cross-linguistic contrasts in requests, apologies, mono¬
logues, and other forms of discourse. Such information points to some
probable cases of discourse transfer.
While there are many difficulties attending the study of discourse
transfer, such study is undeniably important. As Richards (1980) has
noted, when learners violate norms of conversation in the target lan¬
guage, the violations are potentially much more serious than syntactic
or pronunciation errors since such violations can affect what is often
termed “the presentation of self.” Two areas of discourse in which
effects on the presentation of self can be especially dangerous are _po-
liteness and coherence. While politeness is probably a universal no¬
tion, the expression of politeness in different societies varies consider¬
ably. Similarly, the notion of coherence is applicable to conversations
and monologues in every society, but the relations between sentences,
phrases, and other units can vary a great deal in the discourse pat¬
terns of different languages. If native language patterns influence
learners in inappropriate ways, the language that a learner uses may
seem impolite or incoherent. Cross-linguistic differences in discourse
may affect comprehension as well as production. A learner may inter¬
pret conversations and monologues in the target language in terms of

1 Many of the topics discussed in this chapter are sometimes considered to fall within
the realm of pragmatics, which some scholars (e.g., Levinson 1983) see as a field
distinct from discourse analysis. However, there is no consensus about terminology
in such matters, and in the interests of terminological simplicity, only the term
discourse will be used.

48
 Discourse 49

native language norms, and may mistakenly believe that native speak¬
ers are being rude in situations where they are actually behaving ap¬
propriately according to the norms of their speech community. A
learner may also have difficulty in seeing the coherence of target lan¬
guage discourse and fail to grasp the points a speaker or writer is
trying to make. Since much of the research on contrastive discourse
has dealt either with politeness or coherence, this chapter focuses on
research in those areas.

4.1 Politeness
Types of politeness
One of the basic challenges in the study of politeness is understand¬
ing the differences of interpretation that different cultures make of
certain kinds of behavior. What counts as an apology in one culture
may be seen as an expression of thanks in another, and what consti¬
tutes a proper request in one culture may seem very rude in another.
Brown and Levinson (1978) have provided a useful framework for
understanding how politeness may be interpreted by different cultures
in different ways.2 According to their analysis, all people have a
strong interest in preserving face, which has two aspects: (1) positive
face, the self-image and self-respect that a person has; and (2) nega¬
tive face, the claim to privacy, freedom of action, and other elements
of personal autonomy. Since individual and social needs often lead to
actions that threaten the positive or negative face of other people, it
is important for individuals performing such actions to minimize the
sense of threat created by an action. Brown and Levinson describe a
large number of options individuals can use to minimize the impact
of a face-threatening action, but most options are instances of either
positive or negative politeness, which are strategies that minimize
threats to positive or negative face.
Which politeness strategy may be chosen in any given discourse con¬
text depends on many situational and cultural factors. Brown and Lev¬
inson assert that in some societies speakers will generally opt for
strategies aimed at positive politeness, whereas in other societies they
will opt for ones aimed at negative politeness. Comparisons of the norms
of politeness in two communities are therefore possible in terms of how

2 Brown and Levinson (1978) criticize the characterization of behaviors described in

speech-act theory as “norms” or “rules,” and suggest that an approach emphasizing
“strategies” can better characterize the dynamism of speech acts. Yet, while their
objections to the static characterization of discourse in some analyses are
warranted, the “strategies” they propose resemble rules.
50 Language transfer

much the members of those communities use positive or negative po¬

liteness in their speech acts, which are the requests, apologies, and other
purposive uses of language (cf. Levinson 1983). Such comparisons sug¬
gest that there can be important divergences in the speech acts of two
communities even when the sociolinguistic norms of both communities
are generally similar. For example, the norms of linguistic politeness in
France and the United States are the same in many situations, but there
is at least one context in which the norms differ considerably - using
the telephone. According to Godard (1977), telephone calls in France
are seen as impositions more often than they are in the United States,
and thus the etiquette of making calls in France more frequently requires
callers to make an apologetic statement (which is a form of negative
politeness) at the beginning of the call. Consequently, phone calls be¬
tween French and American individuals who are bilingual but unfamiliar
with the differences in telephone etiquette may give rise, as Godard notes,
to perceptions of bizarre or rude behavior. A somewhat similar problem
resulting from differing norms is described by Gumperz (1982) in an
analysis of the communicative difficulties between a speaker of British
English and a Pakistani who had nativelike control of English syntax
and vocabulary but whose conversational strategies differed from the
norms of British English.
Whether the norms of polite speech in two languages vary a great
deal or only a little, there do seem to be norms that are shared and that
are related to the universals of politeness suggested by Brown and Lev¬
inson. In their analysis, negative politeness is most likely to be used to
redress actions that are serious threats to face, whereas positive politeness
is more likely to be used in less threatening situations. Grammatical
mood is one area of linguistic structure where the positive-negative dis¬
tinction appears very useful: the evidence presented by Brown and Lev¬
inson shows that questions are correlated with negative politeness and
statements with positive politeness. In one sense, then, grammatical
moods can be viewed in terms of a politeness scale: interrogative mood
is somewhat “more polite” than indicative mood, since the former
can do more to diminish threats to face; by the same token, imperative
mood is the “least polite,” since imperative forms often seem to be face¬
threatening acts. This politeness scale has empirical support from both
first and second language acquisition research indicating that learners
are sensitive to the correlation between politeness and grammatical mood
(Bates 1976; Rintell 1979; Walters 1979a, b; Carrell and Konneker
1981; Obilade 1984). For example, Carrell and Konneker found that
ESL students who were native speakers of Spanish, Arabic, Persian,
Japanese, or other languages consistently ranked sentences such as Could
you give me a pack of Marlboros?, I want a pack of Marlboros, and
Give me a pack of Marlboros on a scale of politeness: the first sentence
 Discourse 51

was consistently deemed the most polite and the third the least polite.’
Such similarities in judgments among students of very different back¬
grounds indicate that learners can sometimes successfully use their in¬
tuitions about what is “naturally polite” as an aid in distinguishing
between target language expressions which have the same referential
meanings but different social meanings. Yet such intuitions can be per¬
ilous. Much of what seems “naturally polite” in one society will not
necessarily be so in another, as research on requests and apologies
indicates.

Requests
Some contrastive studies indicate that speakers of different languages
prefer different levels of directness in their requests. In a detailed em¬
pirical study, Kasper (1981) showed that native speaker norms in
German and English differ; German usage allows for more directness
in requests than does British English usage.j^or example, German
speakers show a strong preference for modal forms suggesting a sense
of obligation, as in Du solltest das Fenster zumachen (“You should
'

close the window”), whereas English speakers prefer modal forms

with a weaker force, as in Can you close the window?. Moreover, it
appears that German speakers more often prefer declarative state¬
ments in contrast to English speakers, who more often prefer interro¬
gative statements to make requests, as in the preceding example. In
terms of the relation between grammatical mood and politeness de¬
scribed previously, the preferred request strategies in English seem
“politer.”
Nevertheless, it would be mistaken to use any scale, no matter how
carefully constructed, as a means of determining how polite speakers of
one language are in comparison with speakers of another language. Using
an approach common in tagmemic analyses of language and culture
(Pike 1954), House and Kasper (1981) contend that a universal scale of
politeness must always be interpreted in language-specific terms. In other
words, even though a universal scale will determine the range of points
on the politeness scale in every language, individual speech communities
determine the specific points on the scale that mark deference, neutrality,

3 Thomas (1983) claims that because correlations between mood and politeness are
only “probabilistic,” they contribute little to the understanding of problems in the
acquisition of discourse. She also notes that there are many counterexamples
involving polite statements in the imperative form and impolite ones in the
interrogative. Such objections, however, cannot account either for cross-linguistic
correlations between mood and politeness, or for a similar correlation between
utterance length and politeness, or for the heuristic value that such correlations may
have in acquisition.
52 Language transfer

imposition, and so forth. Thus, while the sentences Du soiltest das Fens¬
ter zumachen and Kannst du das Fenster zumachen? have formally equiv¬
alent counterparts in English, You should close the window and Can
you close the window? , they are not, in social terms, translation equiv¬
alents. On a German politeness scale, the two German sentences do not
have such different politeness values as do the two English sentences on
an English politeness scale.
The notion of language-specific politeness scales is not altogether novel
- good translators intuitively make use of the notion, as do good lan¬
guage teachers. The notion does, however, warrant greater attention in
classrooms. The speech act study of Kasper (1981) indicates that speak¬
ers of German often produce requests in ESL that are too direct, and
that finding has obvious implications for the teaching of English in
German-speaking countries. In other acquisition contexts, however, the
problems of directness can be just the opposite of those described by
Kasper. Blum-Kulka (1982) found that native speakers of English often
make requests in Hebrew that are, by the standards of native speakers
of Hebrew, not direct enough; apparently the English politeness scale
induces learners of Hebrew to use structural correlates of rnodals such
as can in requests, whereas politeness in Hebrew does not require such
forms.
The distinction between positive and negative politeness proposed by
Brown and Levinson suggests a possible generalization about the dif¬
ferences in politeness scales. English speakers more often seem to prefer
negative politeness in their requests: Respect for the autonomy of people
who have the power, but not necessarily the desire, to grant favors seems
to explain the frequent use of rnodals such as could, would, and can.
In contrast to English speakers, German and Hebrew speakers appear
to make more use of positive request strategies, in which social bonds
between the speaker and hearer are assumed to be strong. Thomas (1983)
gives an interesting example of such an assumption among speakers of
Russian, who may use extremely direct requests even with strangers in
cases where the imposition is considered slight: for example, Daite si-
garetu! (“Give [me] a cigarette!”). This example suggests that speakers
of a language such as English might have difficulty in learning to make
requests that seem rude in their native language. Moreover, the example
points to a clear danger that speakers of Russian might have in making
requests in English. The problem for Russian speakers is especially great
since some of the polite equivalents of Daite sigaretu in English are
syntactically more complex (e.g., Excuse me, you wouldn’t happen to
have a cigarette, would you?). Research in both first and second language
acquisition suggests that learning the full range of polite formulas often
poses grammatical problems that less capable learners avoid through
 Discourse 53

—
the use and often misuse
1976; Scarcella 1979).
— of simpler forms, such as Please (Bates

Apologies
Apologies also show considerable cross-linguistic variation and pose
problems for second language learners. Research indicates that the prob¬
lems may arise from two kinds of cross-linguistic differences: differences
in the frequency of use of apologetic formulas, and differences in the
relations between apologies and other speech acts. Important differences
in the frequency of apologies are evident in comparisons of the verbal
behavior of speakers of Hebrew, Russian, and English (Cohen and Olsh-
tain 1981; Olshtain 1983). The comparisons, which involved various
role-playing tasks, showed the following tendency:
English > Russian > Hebrew
That is, English speakers used apologetic formulas the most, and Hebrew
speakers used them the least. The evidence furthermore suggests that
native speakers of English generally use apologies when using Hebrew
more often than native speakers of Hebrew do, and that native speakers
of Hebrew generally use apologies when using English less often than
native speakers of English do. These tendencies are not necessarily seen
in all contexts, however. Native speakers of Hebrew, for example, may
be more likely to make apologies that imply personal responsibility in
auto accidents.4
Differences in the relations between apologies and other speech acts
can lead to inappropriate uses of apologetic formulas. Borkin and Rein¬
hart (1978) claim that Excuse me and I'm sorry are often used inap¬
propriately by Thai and Japanese ESL students because of imperfect
matches between those forms and analogous forms in the students’ native
languages. Borkin and Reinhart cite as an example a Japanese student
responding “I’m sorry” to an American saying, “I have so much home¬
work to do!” The relations between apologies and expressions of grat¬
itude seem to occasion particular difficulty. The underlying similarities
between these speech acts are analyzed by Coulmas (1981b), who shows,
for example, that in English, French, and German, expressions of both
thanks and apology from speaker A can elicit identical responses to both
types of speech act from speaker B:

4 The lesser willingness of English speakers to take responsibility for auto accidents
may have a legal explanation. If the English speakers in the role-playing situations
were of American or British origin, they may have been thinking of the legalistic
advice that American and British insurance companies routinely give to clients to
not admit any fault at the scene of an accident.
 54 Language transfer
A: Thank you so much. A: Excuse me, please.
B: That’s all right. B: That’s all right.

A: Merci Monsieur. A: Excusez-moi.

B: De rien. B: De rien.

A: Danke schon. A: Verzeihung.

B: Bitte. B: Bitte.
In Japan, the relations between thanks and apologies are more overt,
and expressions of gratitude are often formulated with terms that in
other languages only express apologies. Westerners may find strange the
double function of phrases expressing either thanks or apologies, such
as O-jama itashimashita (“I have intruded on you”), but in the analysis
of Coulmas it has a straightforward explanation:
In Japan, the smallest favor makes the receiver a debtor. Social relations can
be regarded, to a large extent, as forming a reticulum of mutual responsi¬
bilities and debts. Not every favor can be repaid, and if circumstances do not
allow proper repayment, Japanese tend to apologize. They acknowledge the
burden of the debt and their own internal discomfort about it. (1981b:88)
. Other speech acts besides thanks and apologies probably also show
underlying similarities, similarities that might have a great deal to say
about the relation between discourse universals and language-specific
forms in speech acts. Since cross-linguistic research on speech acts is, as
Levinson (1983) has noted, still in its infancy, it would be premature to
speculate on such relations. However, there is sufficient evidence to
indicate that the differences between speech acts in a target language
will create problems for speakers of a language in which the differences
between the same speech acts are less evident. It does seem to be true
that Japanese speakers have transfer-based difficulties in predicting the
use of I have intruded on you. Thank you, and I’m sorry (Coulmas
1981b; Loveday 1982a). An important implication of the work on Jap¬
anese speakers is that future research on discourse transfer should con¬
trast not only individual speech acts in two languages but entire speech
act systems.

Other speech acts

^Requests and [apologies are not the only types of speech acts that can
cause difficulties in learning to be polite in a foreign language. While
greetings are a likely language universal (Ferguson 1981), the rules gov¬
erning the use of them can vary considerably. In parts of the West Indies,
for example, greetings are expected in almost every social encounter,
whereas in the United States they are used less often (Reisman 1974).
Moreover, some languages (e.g., English) show variable patterns of greet-
 Discourse 55

ing, while others (e.g., Arabic) show fixed patterns (Applegate 1975).
Proverbs are another likely discourse universal (Taylor 1962), but their
role in polite speech varies considerably in different cultures. In many
countries, including much of the Middle East and Africa, proverbs and
other formulaic utterances are frequently employed as aids in arguing,
in complimenting, in expressing condolences, and so forth (Tannen and
Oztek 1981; Wolfson 1981). While proverbs (as well as related phe¬
nomena, such as commercial and political slogans) are common in En¬
glish-speaking countries, there are stylistic constraints on their use both
in speech and in writing (as seen, for example, in advice offered by
Hornby 1974).
Formulaic statements in one language do not always have close trans¬
lation equivalents in other languages. Some of these expressions are quite
simple, as the formula bon appetit said by French speakers at the be¬
ginning of a meal, but others are quite complex, such as the chanting
ceremonies of Cuna Indians (Sherzer 1974). These language-specific
speech acts pose additional challenges for second language learners.
Unlike requests and apologies, which may in some sense be already
“known” to learners even when details of the patterning are not clear,
language-specific speech acts require learners to become familiar with
very new patterns of culture. It is safe to say that the more dissimilar
two cultures are, the more learners will need to make use of speech acts
that appear in one speech community but not in the other.
Along with speech acts, the rules governing turn-taking and other
procedural aspects of conversation show considerable cross-linguistic
variation. Sacks, Schegloff, and Jefferson (1974) and other analysts of
social interaction have noted complex rules related to who can take a
turn at what point, who can “hold the floor,” and so forth. While there
may be a universal element in some of the rules they have identified,
much about turn-taking involves culture-specific rules that can cause
problems for learners. For example, German sociolinguists have ob¬
served that Turkish workers in Germany often have difficulty in partic¬
ipating in conversations where turns are exchanged rapidly .^According
to Barkowski, Harnisch, and Krumm (1976), the norm in ratal Turkey
is for individuals to take extended turns without having to deal with
questions, comments, or other interruptions. Similarly, Applegate ( 1 975)
has noted different levels of tolerance of interruptions between Scan¬
dinavians, Americans, and people from the Middle East.

Conversational style
Many linguists (e.g., Tannen 1981) consider the totality of discourse
devices that signal the imprint of a specific culture on an individual’s
56 Language transfer

speech to be the domain of style.5 From one language to the next,

conversational style can vary along many dimensions, some of which
are not related to politeness. However, one of the most important stylistic
dimensions is very much related to politeness: formality. Structural dis¬
tinctions signaling formal and informal speech are found in many, and
perhaps in all, societies. However, there are considerable differences in
when formal speech is necessary and what aspects of language are used
to convey formality. For example, most dialects of English do not make
use of pronoun distinctions for purposes of formality, but such distinc¬
tions ate' common in many languages.6 For instance, the distinction in
French between tu (“you” singular) and vous (“you” plural) can signal
a difference in number, but often it only signals a difference in formality:
Vous is the expected form to use when addressing one individual in
formal situations. The acquisition of pronouns in languages such as
French is thus a special challenge for speakers of a language such as
English. Successful acquisition requires coming to grips early with so-
ciolinguistic norms in the target language. In French, as in other lan¬
guages, pronouns constitute basic vocabulary; moreover, French
sociolinguistic norms are often coded not only in the pronoun system
but also in the verb inflection system, as in the different verb endings
for tu penses (“you think,” singular) and vous pensez (“you think,”
plural). In the latter case, again, the phrase can be used in formal sit¬
uations to refer to one individual. With other target languages (e.g.,
Persian, Thai, and Japanese) the challenges are even greater because
formality distinctions extend to first- or third-person pronouns. More¬
over, in such languages there is often a complex system of honerifics
involving other types besides pronouns (Brown and Gilman 1 960; Hinds
1975; Beeman 1976; Richards and Sukwiwat 1983).
Along with purely linguistic elements such as pronouns, many para-
linguistic elements can also serve to mark a conversational style: inton¬
ation and related characteristics such as loudness and speech rate, ges¬
tures, facial expressions, physical posture, and the like. The importance
of these elements is as great as the complications they cause for con¬
trastive analysis. The problem of style would be far simpler if para-

5 The question of style is far more complex than the discussion in this chapter may
indicate. Some discourse analyses distinguish registers (e.g., the baby talk that adults
use with children) from styles. Moreover, the role of norms is a basic issue that
would be covered in a more detailed analysis of style. Some scholars have viewed
style largely in terms of norms, whereas others have seen it as the systematic
departure from norms. This and a host of other questions would have to be
considered in order to arrive at a fully adequate theory of style. For further
discussion, see Sebeok (1960), Crystal and Davy (1969), Kinneavy (1971), and
Corbett (1971).
6 Some British dialects still use thou forms to signal familiarity, and the use of ya'll
serves a somewhat similar function in parts of the United States.
 Discourse 57

linguistic elements were always language-specific; however, there are

probably some universals of paralanguage. A cross-cultural experiment
by Ekman (1972) suggests, for example, that some facial expressions
(e.g., smiles) can serve as signals of the same emotional state in any
society. Thus, Ekman found that Americans could accurately infer the
emotional states of Japanese photographed with various facial expres¬
sions, New Guineans could infer the same from pictures of Americans,
and so forth. However, there also are culturally specific uses of facial
expressions. For example, Loveday (1982b) has observed that in north¬
ern Germany adults talking in the “baby talk” register with infants
pucker and protrude their lips. The mix of universal and specific elements
is also found in voice quality. Shouting, for instance, has a natural
advantage everywhere in that it can attract the attention of a distant
individual. However, what people consider to be an acceptable level of
volume in conversation is culturally variable. Thus, Americans speaking
at “normal” volume by American standards might be considered rather
quiet in some parts of Africa and rather loud in some parts of the Far
East (cf. Applegate 1975).
Similar problems involving universal and culture-specific norms are
seen in the uses of silence. For example, Japanese and English speakers
appear to have very different interpretations, as Loveday observes:

Japanese performing in English often do not realize how much distress is

caused by remaining silent for long periods. A hesitancy to speak out and
verbalize one’s thoughts and feelings may be interpreted in the L2 setting as
coldness, hostility, unconcern and even wiliness. Of course, well-meaning
attempts to make the Japanese partner “speak up” often tend to cause silent
frustration and resentment, since, from the Japanese viewpoint, the Westerner
is the culprit who should rather be taught how to shut up. (1982a:8)

Differing community standards on the role of talk and silence are also
known to affect other language contact situations, such as those in¬
volving speakers of English and speakers of some American Indian lan¬
guages (Scollon and Scollon 1981).
, In second language acquisition, much of the difficulty in becoming a
competent speaker (and listener) is likely to come from the simultaneous
existence of universal and specific elements in spoken interactions. The
difficulty may be compounded by beliefs on the part of learners that
their requests, their greetings, their facial expressions, their volume, and
so on, are not arbitrary in the way that words in their native language
— —
are. That is, learners may suspect not altogether mistakenly that the
rules guiding their interactions are natural and therefore universal. In¬
formation that sorts out the seemingly natural from the truly natural in
paralanguage thus has an important place in foreign language instruction
in oral skills.
58 Language transfer

4.2 Coherence
Differences in norms oF-politeness are only one of the ways in which
cross-linguistic variation can lead to misunderstandings* Differences re¬
lated to expectations about coherence in discourse may create special
problems for learners in their reading or listening comprehension efforts.
Alternatively, those differences may lead members of a speech com¬
munity to consider the speech or writing of non-native speakers inco¬
herent. It is not yet clear just how often such differences actually result
in negative transfer (Section 4.3). What is clear is the potential that
cross-linguistic variations in discourse have for creating misun¬
derstandings.
The notion of coherence is closely related to the notions of logicality
and relevance, with an absence of either one seriously jeopardizing the
coherence of a discourse. Sometimes only an apparent absence of either
can create an impression of incoherence. That is, conversations or mono¬
logues may seem incoherent if they appear to lack sufficiently logical
relations between ideas, or more technically, propositions (Section 5.1).
Conversations or monologues may also seem incoherent if there appears
to be too little relation between a focus of information in a discourse
(i.e., a topic) and other information, in other words, if too much appears
to be “off-topic” (cf. Section 6.1).
Incoherence is sometimes very real. For example, the speech of dis¬
turbed people can be quite illogical or full of irrelevancies. On the other
hand, the incoherence that readers or listeners may perceive sometimes
has little to do with logicality or relevance. In some cases a particular
audience may simply lack sufficient knowledge of the topic to make
sense out of a discourse. The language in technical reports in various
fields may seem incoherent to those unfamiliar with the subject matter,
whether or not the discourse is really incoherent. Similarly, discourse
that presupposes some familiarity with another culture may seem in¬
coherent when listeners or readers lack sufficient knowledge of the cul¬
ture. In other cases, audiences may not have problems with the content
of the discourse but with the presentation of information. For audiences
unfamiliar with certain patterns of organization, the information pre¬
sented through those patterns may prove difficult or even impossible to
understand.

Narratives
Like linguistic expressions of politeness, narratives probably constitute
a discourse universal. Appearing in most, and perhaps all, societies,
narratives allow for many cross-linguistic comparisons of discourse
 Discourse 59

form. Some types of narratives recur in an extremely wide variety of

cultures. Myths and folktales have long been a focus of study by an¬
thropologists and literary theorists on account of pervasive character¬
istics both in content and form. Stories about the creation of the world,
a flood, a return of a god or hero, and so forth have appeared in a
remarkable number of communities. Moreover, traditional narratives
are known to have formulaic characteristics that recur in the oral lit¬
erature of many communities (Levi-Strauss 1955; Dundes 1964; Propp
1968). On the basis of some of these formulaic characteristics, scholars
have developed story grammars and other methods of analysis to un¬
derstand narratives and related forms of discourse (e.g., Thorndyke
1977; Mandler et al. 1980). In story grammars, relations between set¬
tings, themes, plots, episodes, and characters are made explicit through
a series of rules similar to those in generative descriptions of syntax.
While such rules do not entirely account for the coherence seen in nar¬
ratives, they do suggest the role that linguistic organization plays in
signaling coherence.
The range of possible stories is far greater than what is found in the
narratives analyzed in terms of story grammars, and it seems unlikely
that that or any other analytic system will ever account for the entire
range. However, many of the narratives characterized in story grammars
are quite common in storytelling traditions and have a number of re¬
current properties. Among such properties are chronological and causal
order, as well as a sense of narrative tension. That is, the events in such
stories show a succession of events in time, with many of the events
being causally related to each other, and a problem introduced early in
the story is often not resolved until the end of the story, with the audience
being kept in suspense during most of the narrative. In many cultures
children are introduced to such stories at an early age, and adult listeners
(or readers) are quite able to recognize and expect patterns of chronol¬
ogy, causality, and narrative tension.
These patterns certainly recur in storytelling in many lands. However,
it is not clear that such recurrences are universal. Kintsch and Greene
(1978) have claimed that there are culturally specific patterns of nar¬
ratives and that cultural differences in narrative form have consequences
for language comprehension. This claim has some experimental support,
including the results of studies showing that English-speaking under¬
graduates had more success in remembering European than Amerindian
folktales (Kintsch and Greene 1978). Others, however, have criticized
the conclusions that Kintsch and Greene drew from such evidence and

7 Carrell (1982) and others have observed that coherence in discourse (including
narratives) comes not from the formal structure of discourse but from the coherence
—
of content that is, from the thought processes underlying any discourse form.
60 Language transfer

have argued that cultural differences in narrative are far less than what
they seem to be (Mandler et al. 1980).
However culturally specific or universal certain patterns of narrative
may be, it is undeniable that stories from other cultures can seem rather
strange. The following story from the Kathlamet of the Pacific Northwest
is a classic example of just how different stories can appear:
One night two young men from Egulac went down to the river to hunt seals,
and while they were there it became foggy and calm. Then they heard war¬
cries, and they thought: “Maybe this is a war-party.” They escaped to the
shore, and hid behind a log. Now canoes came up, and they heard the noise
of paddles, and saw one canoe coming up to them. There were five men in
the canoe, and they said:
“What do you think? We wish to take you along. We are going up the
river to make war on the people.” One of the young men said: “I have no
arrows.”
“Arrows are in the canoe,” they said.
“1 will not go along. I might be killed. My relatives do not know where 1
have gone. But you,” he said, turning to the other, “may go with them.”
So one of the young men went, but the other returned home.
And the warriors went on up the river to a town on the other side of
Kalama. The people came down to the water, and they began to fight, and
many were killed. But presently the young man heard one of the warriors
say: “Quick, let us go home: that Indian has been hit.” Now he thought:
“Oh, they are ghosts.” He did not feel sick, but they said he had been shot.
So the canoes went back to Egulac, and the young man went ashore to his
house, and made a fire. And he told everybody and said: “Behold I
accompanied the ghosts, and we went to fight. Many of our fellows were
killed, and many of those who attacked us were killed. They said I was hit,
and 1 did not feel sick.”
He told it all, and then he became quiet. When the sun rose he fell down.
Something black came out of his mouth. His face became contorted. The
people jumped up and cried.
He was dead.
(Bartlett 1954:65)

While the story is coherent, the supernatural events in it cause compre¬

hension problems for readers who do not note some important details.
In a famous psychological experiment, Bartlett (1954) showed that Brit¬
ish students had considerable difficulty in remembering much of the
story, while students had considerably less difficulty in recalling details
of an African story with a pattern rather similar to that of many English
stories. Although the Kathlamet ghost story does show formulaic char¬
acteristics, the formula of the story was apparently hard for English
speakers to recognize (Dundes 1964).
Lack of familiarity with a discourse pattern is not the only possible
source of comprehension difficulties. When listeners or readers are not
familiar with another culture, they may not succeed in correctly inter¬
preting the content of a discourse. In a study of reading comprehension,
 Discourse 61

Steffensen, Joag-Dev, and Anderson (1979) prepared two passages in

English, one describing a wedding in the United States and the other a
wedding in India. In both cases the passages were constructed to resemble
a narrative letter written by someone who had attended one of the
weddings and who was a member of the same culture: for example, the
Indian wedding was described by an Indian writer for an Indian reader.
Test results indicated that subjects needed less time to read the narrative
describing the type of wedding more familiar to them and that they were
able to recall the gist of the familiar type of wedding better than the gist
of the unfamiliar type. An especially significant finding is that subjects’
recollections showed fewer distortions of facts in the text of the familiar
wedding. For example, Indians readily understood such statements
as “After two days of marriage she was taken to Nagpur. Her father-
in-law accompanied her.” The Indian readers understood this to mean
that the bride was going to Nagpur to live with her in-laws, as they
knew that the norm for marriages in India is for the bride to live with
the groom’s family. In contrast, a number of the American readers in¬
terpreted that statement to mean that the bride was going to Nagpur
for a honeymoon.
Culturally specific knowledge can affect not only the comprehension
but also the production of discourse, as shown in a study by Win¬
field and Barnes-Felfeli (1982). Two groups of ESL students, Spanish¬
speaking and non-Spanish-speaking, were asked to read paragraphs of
about two hundred words and then write summaries of the paragraphs.
One of the two paragraphs was about the Spanish classic Don Quixote,
while the other was about the Japanese Noh theater. The Spanish-speak¬
ing students were able to write significantly longer descriptions of the
Quixote paragraph than were their non-Spanish-speaking counterparts,
although in some other respects there were no clear-cut differences be¬
tween the two groups. Since most of the Spanish-speaking students were
already familiar with the Quixote story but not with Noh theater, cul¬
tural knowledge apparently contributed to their relative fluency. Aside
from fluency, more subtle effects of cultural knowledge appear in dis¬
course production. For example, Greek and American students produced
rather different narrations about a short film that they had seen (Tannen
1984). The Greek students tended to provide more details about possible
social or psychological characteristics of the individuals seen in the film,
whereas the Americans tended to provide more details about actions
performed by the individuals and about the film-making technique.

Indirection in discourse
Research on narratives provides some evidence of cross-linguistic dif¬
ferences in discourse, but there is even stronger evidence from recent
studies of Japanese and Korean. These studies suggest that cross-
 62 Language transfer

linguistic differences can affect not only comprehension and memory

but also value judgments of writing. The evidence comes from prose
that shows a great deal of indirection.
In an effort to categorize the ESL writing of students from various
countries, Kaplan (1966) compared the writing done in various countries
in terms of different types of lines (cf. Section 2.2). In Kaplan’s analysis,
writing in English resembles a straight line since it supposedly is, direct
'and “to the point.” In contrast, writing in Russian, for example, resem¬
bles, according to Kaplan, a zigzag, and writing in Oriental languages
“a widening gyre.”8 Kaplan provided examples in support of his im¬
pressionistic linear characterizations, although there is still much un¬
certainty about the nature of cross-linguistic variation in discourse
(Section 4.3). Despite that uncertainty, however, Kaplan’s analysis has
had much intuitive appeal among scholars who have encountered par¬
agraphs such as the following translation from Korean:

- Foreigners who reside in Korea as well as those who study the Korean
language in foreign countries are, despite their deep interest, ignorant of the
basis on which the Korean alphabet, Hangul, was formulated. The Korean
alphabet, composed of combinations of lines and curves, at first seems more
difficult than Japanese kana for those who use the Roman alphabet, and as
the combination of vowels and consonants multiplies, it appears more
difficult to memorize all the combinations. This seemingly complicated
combination of vowels and consonants can, on the contrary, be mastered
with no more effort than is needed to learn the Roman alphabet or Japanese
kana, for one must merely memorize two dozen vowels (Eggington
1987:154).

The first sentence seems to set the stage for a discussion of the history
of the Korean alphabet. However, the focus of information in the par¬
agraph then shifts without any warning to a description of the seeming
difficulty of Hangul. Such a sudden change in topic is not unusual in
Korean prose, according to Eggington (1987), who sees abrupt shifts as
a normal feature of Korean style. Similarly, Hinds (1983, 1984) has cited
several examples of Japanese news stories written in a form known as
ki-shoo-ten-ketsu. According to Hinds, the ki-shoo-ten-ketsu form has
its origins in Chinese poetry and constitutes a norm of Japanese style.

8 The “straight line” pattern may not be characteristic of all English discourse, as the
following remarks by a Scandinavian linguist (Dahl 1979a: 199) on John Lyons’s
(1977) Semantics suggest. Dahl criticizes Lyons’s “tendency to write in the rambling
essay style not uncommon among British scholars ... he may get off the main line of
discussion to elaborate on some minor point at length, after which the reader will
have lost the thread.” Moreover, correspondence in other languages sometimes
shows a “linear” pattern even in cases where influence from English may be
minimal. The Vai people of West Africa, for example, are known to use a very
direct type of paragraph structure in their letters (Scribner and Cole 1981).
 Discourse 63

Texts in this form have a four-part pattern of development, as seen in

the following translation of an article in a Japanese newspaper:
(ki) This columnist first learned to drive and obtained a driving license in
New York City. At the time, what the driving instructor most naggingly
stressed was “harmony.” He said that the knack of driving lay first in
harmony, second in harmony, no third and fourth and fifth in harmony.
(shoo) Ignoring the question of how to shift gears, he lectured, while on
the road, on the importance of maintaining the minimum necessary distance
between cars. There were times when this writer became sick and tired
because he kept harping on the matter so much. It may be questionable
whether American drivers actually place importance on “harmony,” but at
least that aged instructor kept insisting on it all the time.
(ten) The most frightening thing in the accident in the Nihonzaka Tunnel
of the Tomai Expressway on July 11 was that there were about 170 vehicles
within the tunnel and most of them burned. Why were there so many as 170
vehicles inside the tunnel?
In order to run at a speed of 80 kilometers per hour within the tunnel,
vehicles must keep a distance of 80 meters between each other. If the vehicles
had been running at 80-meter intervals, the total of vehicles on the two lanes
from the entrance to the site of the accident about 1 .6 kilometers away
should have been 40 at the most. Since the expressway was crowded that
day, the speed may have been less than 80 kilometers per hour. Still, 170
vehicles are just too many.
First, there was disregard of the proper distance between vehicles. On
expressways, there are cases of vehicles running at 100 kilometers an hour
with only 10 or 20 meters between them. Even if a driver tries to maintain
the proper distance between vehicles, other vehicles cut into the space in
front of that driver, immediately destroying harmony. Drivers are aware of
the danger of a collision and pile-up but keep on driving, comforting
themselves with the thought, “It will be all right.” The piling up of such
disharmony is dangerous.
There was also the fact that warnings were ignored. Immediately after the
accident occurred, the panel at the tunnel entrance lit up with the warning,
“Fire Outbreak, Entry Banned.” But it appears that a considerable number of
cars entered the tunnel after the warning had been posted. Did they speed
into hell, unable to apply brakes suddenly because the distance between
vehicles was too small?
(ketsu) The preventive measures taken by the Japan Highway Public
Corporation were grossly inadequate. Experts should be well aware of what
a lack of water for firefighting means in emergencies. They knew but closed
their eyes to the fact. The psychology of “It will be all right” on the part of
the drivers and of the corporation caused this major accident.
(Hinds 1983:188-89)
While the first two parts of this rhetorical format will not likely seem
strange to an English-speaking audience, the third part (the ten) seems
to be an abrupt shift away from the topic originally introduced. In fact,
the rhetorical norm is for the ten to introduce a subtopic that is only
indirectly related to the ki and the shoo. Hinds’s research indicates that
64 Language transfer

Japanese readers are quite accustomed to reading such articles. In com¬

parison with an American group that read translated passages such as
this one, a group of Japanese readers more accurately recalled infor¬
mation in news articles in the ki-shoo-ten-ketsu form.
Taking a somewhat different experimental approach, Eggington found
that Korean readers were often able to recall more information when it
was presented in “nonlinear” forms like the ki-shoo-ten-ketsu than when
it was presented in the “linear” form that Kaplan deemed to be char¬
acteristic of writing in English. The indirection seen in some Japanese
and Korean prose thus seems to produce no adverse effects on compre¬
hension among readers used to such forms. There is, furthermore, evi¬
dence that readers in the Far East consider indirection to be quite
acceptable: results of a survey by Hinds (1983) indicate that Japanese
readers are more likely than American readers to consider articles in ki-
shoo-ten-ketsu form as well written. The results of such investigations
thus suggest that a passage may be more readable or less readable de¬
pending on readers’ expectations, which are partially shaped by language
and culture.9 '

4.3 Discourse transfer and other factors

Most of the research discussed up to this point has shown considerable
evidence of cross-linguistic differences in discourse, but less evidence of
cross-linguistic influences. There is strong evidence of discourse transfer
in a small number of studies. However, just as other factors besides
transfer can affect the acquisition of syntactic competence in a second
language, other factors can affect the acquisition of discourse
competence.

Cross-linguistic differences and influences

The earlier discussion of the cross-linguistic dimensions of politeness
and coherence cited many discourse contrasts evident in human lan¬
guages. While there is still only fragmentary information available about

9 Some other interesting evidence for cross-linguistic differences in discourse comes

from the literature being written in several former colonies of Western powers.
Several authors have perceived language-specific discourse differences and have
attempted to convey them in poems, short stories, and novels. For example, in Raja
Rao’s novel Kanthapura (1963), which was written in English, there are several
stylistic devices used to convey a sense not only of the oral tradition in a village in
south India but also of the syntax of the language spoken there (cf. Kachru 1969;
Richards 1979; Platt, Weber, and Ho Mian Lian 1984).
 Discourse 65

such contrasts, the available evidence suggests that English and Japanese,
for example, show a number of differences, such as the following:
English Japanese
Less overlap in expressions of gratitude and apology More overlap
Simple pronoun system Complex honorific
system
More avoidance of silences Less avoidance
Less use of “nonlinear” patterns of organization More use
Observations by Loveday (1982b) and others suggest that a fully de¬
veloped contrastive analysis of discourse in Japanese and English would
show even more differences than these, but the full range of differences
will probably remain unknown for a long time to come. For languages
less investigated than English and Japanese, even preliminary contrastive
studies of discourse are often lacking.
Along with the differences, however, there probably are typological
similaritiesJn discourse between languages just as there are typological
similarities in phonology and syntax (Section 3.4). While such resem¬
blances are not well understood, typological research is certainly feasible
for some aspects of discourse, such as indirection. Yamuna Kachru
(1983) has argued that indirection is highly characteristic of discourse
in Hindi and other Indian languages (cf. Pandharipande 1983). It is
natural to wonder if the kinds of indirection found in Hindi closely
resemble the kinds found in Korean and Japanese.10
While the differences and similarities of discourse warrant further
study, their effects on second language acquisition remain only partially
understood. There is some evidence that negative transfer occurs in
certain situations. Some of the best evidence comes from an investigation
by Bartek (1983) of repetition in the ESL writing of American Indian
students. In Amerindian languages such as Navajo and Apache, repe¬
tition is frequently used for emphasis, and Bartek’s study indicates that
Navajo and Apache students writing in English use repetition so much
that their papers often seem verbose in the eyes of teachers conscious
of the norms of English style. The students are most likely to use rep¬
etition when they have strong feeljngs toward a topic, as seen in the
following excerpt from an English composition written by a student
describing a raid the Navajos suffered at the hands of a marauder:
Carson invaded the Canyon De Chelley to destroyed Navajo crops and
livestocks and capture or kill all the Navajos, so they burned all their crops

10 Hinds (1987) has proposed still another typology, one involving the variations in
“responsibility” that writers and readers may have in different societies. The
Japanese, he claims, more frequently put the burden of responsibility for
understanding a text on the reader rather than on the writer.
66 Language transfer

and bring all their livestock, all their livestock. Finally when the Navajos
found they have destroyed all their crops and livestock they shoot down all
their livestock .. . (Bartek 1 983:299)
Since livestock and crops were the heart of the Navajo economy, the
repetitions doubtlessly serve to emphasize the scale of the calamity. Such
reiteration is not characteristic of all ESL writers, although some evidence
suggests that Arabic discourse may encourage Arab students to repeat
words and phrases in English (Koch 1983). Since the use of repetition
seems to be characteristic of only some groups and since there is evidence
of the use of repetition in the students’ native languages, the case for
discourse transfer in Bartek’s study is quite strong.
Some of the evidence for discourse transfer also involves syntactic
transfer - such an overlap of evidence is not surprising in view of the
important influence of discourse on syntactic structure (Section 6.1).
English makes considerable use of cleft sentences for purposes of em¬
phasis, as in the following examples: It’s the transmission that isn't
working and It was the copilot who saw the oncoming plane. In cleft
sentences there is often a contrast between an entity referred to in
the sentence and one known from the discourse context; thus, the
first sentence could serve to contrast a broken-down automobile trans¬
mission with a smoothly running carburetor, and the second sentence
equid serve to contrast an attentive copilot with an inattentive pilot. In
certain other languages the use of cleft sentences is more pervasive. Irish,
for example, makes use of a wide range of cleft patterns, and the language
contact situation in Ireland has led to a use of cleft sentences unlike
those found in other English-speaking regions: It’s flat it was (“It was
flat”) and It must be working for her he was (“He must have been
working for her”) (Henry 1957). As Henry and others have observed,
such Hiberno-English sentences reflect cleft patterns in Irish (cf. Stenson
1981; Bliss 1984). The case for transfer is also supported by the fact
that cleft sentences seem to be especially common in counties of Ireland
where Irish is still spoken (Filppula 1986).
The use of repetition by American Indians writing in English and the
use of cleft sentences by speakers of Hiberno-English are cases of transfer
in which the comparisons of performance of different native language
groups are implicit. That is, since only certain groups of non-native
speakers of English produce certain patterns, the divergences from target
language norms are not seen in all language acquisition and they there¬
fore suggest the influence of particular native langhages (Section 3.2).
In studies that make explicit comparisons of target language speech or
writing by speakers of different languages, there is also some credible
evidence for discourse transfer. For example, the study of apologies in
Hebrew by Olshtain (1983) compared native speakers of Hebrew with
immigrants from the United States and the Soviet Union (Section 4.1).
 Discourse 67

The English and Russian groups differed in their use of apologies not
only from the native speaker group but also from each other.
Despite such evidence for cross-linguistic influences in discourse, the
extent of discourse transfer is not clear. Some studies of contrastive
discourse have found little or no evidence for transfer (e.g., Connor and
McCagg 1983, 1987; Scarcella 1984; Mohan and Lo 1 985). The dearth
of evidence in such studies suggests that the discourse influences from
the native language may be weaker than, for example, phonological
influences, although the relative likelihood, of transfer in phonology and
other subsystems is still very much an open question. In any case, the
enormous complexity of linguistic factors related to politeness, coher¬
ence, and so forth suggests that a learner’s native language may well be
only one of a host of influences on second language discourse.

Other possible influences

In second language discourse (and also in syntax, phonology, etc.), the
relative importance of factors besides transfer is not clear. Nevertheless,
there is evidence that other factors can be quite important.
OVERGENERALIZATION

Just as morphological errors may result from false generalizations (e.g.,

We sleeped a long time), inappropriate discourse may result from gen¬
eralizations carried too far. In her detailed study of speech acts of Ger¬
man ESL students, Kasper (1981) noted many instances such as the
following:
Y: Could you show me the drawings?
X: Yeah."
In a role-playing task, a student (X) made the response Yeah to a request
from an older woman (Y) whom he had never met before. While Yeah
might be polite enough as a response to a request from a friend of the
same age, it could be seen by an older person as extremely curt. Since
the norms of politeness in German and English are not very different in
this regard, transfer is not as good an explanation for X’s response as
overgeneralization is.
A subclass of false generalizations probably results from transfer of
training, specifically from foreign language instruction. Kasper cites a
case that seems to reflect an inappropriate use of a polite formula: Would
you like to drink a glass of wine with met, which was used in a context
in which Can I get you another glass of wine? would have been more

I I The capitalization and punctuation of this example differ in minor ways from the
system used by Kasper.
68 Language transfer

appropriate. The overuse of formulas such as Would you like .. . prob¬

ably results from ESL instruction that does not make clear the limits on
the use of polite formulas.

DEVELOPMENTAL INFLUENCES
In some respects the mastery of writing in a second language resembles
the development of writing in one’s native language. For example, chil¬
dren writing in their native language and adults writing in a second
language often show a preference for discourse constructed with very
simple syntactic structures (Homburg 1984; Hillocks 1986). This fact
complicates the study of transfer, since some discourse characteristics
that appear to be due to native language influence may instead reflect
little more than a normal development of writing abilities. For example,
writing in Persian shows considerably more use of coordinating con¬
junctions than English writing does (Dehghanpisheh 1978). It therefore
is tempting to attribute to transfer any overuse of and or other coor¬
dinating conjunctions by Persian-speaking ESL students. However, the
frequent use of simple and coordinate sentences is a common develop¬
mental characteristic among both native and non-native speakers of
English (cf. Shaughnessy 1977; Larsen-Freeman 1978). Accordingly,
transfer is not necessarily the best explanation for this aspect of Persian
speakers’ writing, just as transfer is not always the best explanation for
certain syntactic errors.12

LITERACY
Many of the problems that second language writers face may be due
primarily to inexperience in reading or writing in any language. For
example, the problems that Japanese ESL students have in writing classes
may more reflect a lack of skill in composing in Japanese than an influ¬
ence of the ki-sboo-ten-ketsu form (Section 4.2). In fact there is evidence
that native-language literacy skills affect a number of aspects of second
language performance, including writing (Section 8.1).

12 The use of coordinate clauses might seem to be an issue primarily related to

syntax and not to discourse. However, the developmental preference for
coordinate clauses over subordinate clauses may result from a need of inexpe¬
rienced writers to concentrate on the global demands of discourse construction
instead of on syntactic elaboration. There is empirical support for the idea that
differences in discourse tasks can affect linguistic performance. Working with
English-speaking children, Levin, Silverman, and Ford (1967) found that when
children were asked to explain simple physical processes (in English), they paused
more often and spoke more slowly than they did when asked to describe objects.
That finding is consonant with some ESL writing teachers’ intuitions that it is
more difficult for students to write accounts of processes than descriptions of
objects.
 Discourse 69

A word about positive transfer

Up to this point, most of the discussion of cross-linguistic influences in
discourse has focused on negative transfer. While the likelihood of posi¬
tive transfer also being important is high, most research on contrastive
discourse has not addressed this issue. Such transfer is virtually impos¬
sible to document in the absence of studies comparing the performance
of speakers of different language backgrounds (Section 3.2), and so far
there have been few comparative studies in the area of discourse.
One hindrance to doing comparative research is the dearth of infor¬
mation about discourse similarities (or differences) in various languages.
There is a clear need for several kinds of information, including: (1)
discourse universals such as those hypothesized by Brown and Levinson
about politeness (Section 4.1); (2) stylistic variation in different lan¬
guages; (3) the teaching of rhetoric for written communication; (4) the
teaching of rhetoric for spoken communication (e.g., Albert 1964); and
(5) the influence of English and other Western languages on spoken and
written discourse in Asia, Africa, and elsewhere. On that last point, it
seems likely that the impact of Western science, education, and jour¬
nalism on many cultures can explain some of the similarities seen in the
non-narrative prose of many speech communities. For example, Yamuna
Kachru (1983) has shown that some scientific writing in Hindi follows
the (supposedly) “linear” pattern of English prose. Such influences con¬
stitute a kind of borrowing transfer, and in some languages the effects
of such transfer are evident in the wide range of styles used in writing
(e.g., in Persian). As more information is becoming available about the
teaching of writing in various countries (e.g., Purves 1986), future studies
of contrastive discourse should be better able to take into account such
factors as borrowing transfer and stylistic variation.

4.4 Summary and conclusion

Although the evidence is still fragmentary, enough exists to indicate that
transfer involving discourse can often occur in second language acqui¬
sition. The probability of future research showing more instances of such
transfer seems high since much research has already identified consid¬
erable cross-linguistic differences in discourse. Some of those differences
can cause misunderstandings and may also lead to second language
speech or writing that differs greatly from the discourse norms of the
target language. Misunderstandings related to politeness and coherence
are especially dangerous, and thus discourse transfer should be a matter
of special concern for teachers. As with phonology, syntax, and other
70 Language transfer

subsystems, however, transfer is not the only source of divergences be¬

tween native speakers’ and learners’ use of target language discourse.
At present, it would be hard to quantify the importance of cross-linguistic
influences on discourse. Nevertheless, it appears that discourse transfer
has considerable potential to interact with other subsystems, including
syntax, as seen in the case of Hiberno-English sentences such as It must
be working for her he was. Other interactions involving transfer also
seem likely; as research described in the next chapter indicates, semantic
transfer often requires the study of several kinds of interaction.

Further reading
Two highly useful surveys of problems related to discourse are texts by
Levinson (1983) and Brown and Yule (1983). Many of the problems
especially important to second language discourse are discussed by Loveday
(1982b). Although their views on discourse transfer are rather different,
articles by Hinds (1983) and by Mohan and Lo (1985) provide much-needed
critical assessments of some of the work on contrastive rhetoric.
5 Semantics

The study of discourse transfer and the study of semantic transfer overlap
a great deal. Since discourse normally consists of sequences of statements,
discourse analysis is closely related to propositional semantics, or the
study of meaning in statements (cf. Section 4.2). Moreover, since state¬
ments normally consist of sequences of words, discourse analysis is also
related to lexical semantics, which is the study of meaning in words.'
Accordingly, if discourse transfer occurs, as evidence discussed in Chap¬
ter 4 suggests, semantic transfer is also probable. The discussion of
semantic transfer in this chapter is divided into two sections: first, a look
at cross-linguistic differences evident in propositional semantics, and
then a look at lexical transfer. /

5.1 Propositional semantics

Semantic universals and linguistic relativism
A fundamental issue in the study of semantic transfer is the relation
between language and thought. Expressions such as “learning to think
in French” reflect a common belief that learning a particular language
requires adopting a worldview which, to some extent, is unique to that
language. One implication of this belief is that if learners do not “think
in French,” for example, they must still be using their native language
as a reference point for cognitive activities. It is still very much an open
question just how closely language and thought are related (cf. Whorf

I Scholars frequently point out that discourse is more than the sum of its structural
parts. That is, while words are components of statements, and statements are
components of discourse, discourse also involves aspects of meaning not truly
amenable to a componential analysis. Another caution seems advisable here: The
term statement should not be construed in the narrow sense of an assertion or a
description, but in a broader sense to include speakers’ communications about their
attitudes and feelings. Sentences are the most typical statements, but words and
phrases can also be statements, as in the case of the response Yes to the question Is
Paris the capital of France?.

71
 72 Language transfer

1956; Foss and Hakes 1978; Lakoff 1987). Accordingly, it is also an

open question just how much native-language semantic structure can
influence performance in a second language. Some research, however,
does suggest that cross-linguistic differences in structure sometimes re¬
flect differences in thinking, and a look at such differences is therefore
appropriate.
Whatever differences there may be in the thought patterns of people
in different speech communities, there are indications of universals in
cognition. Propositional reasoning is a case in point. Although not all
cultures have developed an explicit study of logic and although cultures
in which logic is studied have not all developed exactly the same type
of logical analysis, human reasoning processes show considerable sim¬
ilarities in empirical cross-cultural research. For example, an investiga¬
tion by Hamill (1978) of reasoning in conversations in Mende (a West
African language) and in English showed there was little cross-cultural
variation in either successes or failures in reasoning. In conversations in
both languages, speakers showed an awareness of such basic logical
rules as contradiction: that is, if a proposition “X” is true, a negation
/ of the same proposition (i.e., “not X”) must be false. Thus, the statements
Paris is the capital of France and Paris is not the capital of France cannot
both be true. In Hamill’s study, English speakers and Mende speakers
seemed to make use of the same basic rules in both valid and fallacious
interpretations of chains of statements (cf. Hutchins 1981).
Despite such evidence for the universality of basic reasoning processes,
cross-cultural differences in cognition do seem to exist. One important
source of differences is probably the form of education that children
receive (Scribner and Cole 1981). Another probable source of differences
— and one more directly relevant to the problem of contrastive analysis
— is linguistic variation. Many scholars have claimed that differences in
thought processes are somehow reflected in differences in language. A
passage from an essay by Whorf (1956:212—13) states some of the
strongest claims:
the background linguistic system (in other words, the grammar) of each
language is not merely a reproducing instrument for voicing ideas but rather
is itself the shaper of ideas, the program and guide for the individual’s mental
activity, for his analysis of impressions, for his synthesis of his mental stock
in trade. Formulation of ideas is not an independent process, strictly rational
in the old sense, but is part of a particular grammar, and differs, from
slightly to greatly, between different grammars. We dissect nature along lines
laid down by our native language. The categories and types that we isolate
from the world of phenomena we do not find there because they stare every
observer in the face; on the contrary, the world is presented in a
kaleidoscopic flux of impressions which has to be organized by our minds -
and this means largely by the linguistic system in our minds.
 Semantics 73

Several of the statements in this passage are consonant with the so-called
< strong relativist position, which claims that language can determine cog¬
nitive processing, as exemplified in the statement “We dissect nature
along lines laid down by our native language.” Yet, as many scholars
have observed, this determinist position is beset with insurmountable
problems (cf. Section 3.4). For example, if individual languages (and
thought patterns) were as radically different as the determinist position
claims, the acquisition of a second language might sometimes be im¬
possible. That is, there might be some language whose speakers were
utterly incapable of learning even a rudimentary amount of English
because of radically different structures in their native language; or by
the same token, there might be some language that speakers of English
would find utterly impossible to learn. While the results of second lan¬
guage acquisition often fall short of what is desired, there are no known
cases of an absolute “acquisition barrier” between speakers of different
languages. This and many other considerations suggest that the differ¬
ences seen between languages and cognitive processes are far less vast
than what determinist arguments claim (cf. Penn 1972; Rosch 1974;
Foss and Hakes 1978; Bloom 1981).
Although the “strong” position is discredited, the “weak” relativist
position is plausible; as the last sentence in the passage of Whorf quoted
—
earlier suggests, language may have an important but not absolute —
influence on cognition.2 One structural characteristic found in many
languages that might influence cognition is grammatical gender, as in
the case of French, which classifies every noun as either masculine or
feminine. For example, mot (“word”) is masculine and langue (“lan¬
guage”) is feminine. Gender classification may encourage mental asso¬
ciations that differ according to the languages that individuals speak (cf.
Guiora and Acton 1979; Clarke et al. 1981). In certain cases, cultural
traditions may encourage or discourage certain types of thinking, and
those cultural patterns may be reinforced by the structural characteristics
of a particular language. A contrastive study of English and Chinese
(Bloom 1981) illustrates one possible instance of interaction between
language and culture, but also illustrates exceedingly well the difficulties
of achieving conclusive evidence that supports the relativist position.
Like speakers of other languages, Chinese speakers are perfectlyx:apable
of talking about unreal states of affairs. However, the syntactic structure
of Chinese does not explicitly encode some semantic differences asso-

2 Much of the controversy surrounding linguistic relativism has arisen from

ambiguities in statements by Whorf and others, with the result that discussions of
language and cognition do not always distinguish the “strong” and the “weak”
relativist positions (cf. Penn 1972).
74 Language transfer

dated with unreality; in contrast, the English verb system explicitly codes
differences, such as those seen in the sentences If you burned your finger,
it would hurt and If you had burned your finger, it would have hurt.
The syntax of Mandarin allows for one type of sentence construction
to apply to both types of conditions described in the two English sen¬
tences. Thus, there is no special syntactic device in Chinese to signal the
difference between a hypothetical state of affairs (in the first sentence)
and a counterfactual state of affairs (in the second) in which an event
that did not take place (e.g., a finger burned) is talked about as an
imaginary event in the past. The absence of such an overt distinction is,
according to Bloom, consonant with certain Chinese intellectual
traditions.
However plausible relativist analyses, such as that of the Chinese case,
may appear, they can be difficult to validate. Bloom has provided in¬
teresting but controversial evidence in support of his Whorfian inter¬
pretation of counterfactuals in Chinese. Some of his strongest evidence
comes from the results of a reading test given to Chinese monolinguals,
English monolinguals, and Chinese-English bilinguals. The passage used
on the reading test contained many counterfactual statements and the
state of affairs described in the passage was explicitly unreal: the effect

— —
that a certain philosopher might have had but actually did not have
upon the development of science. In answers to questions designed to
establish how well readers understood that the philosopher had actually
had no effect on the development of science, the English monolinguals
proved the most successful and the Chinese monolinguals the least
successful.
Bloom argues that the performance of the bilingual group constitutes
further evidence for his Whorfian analysis since the performance of that
group fell in between those of the monolingual groups; by virtue of
having studied English, this group had become more sensitive to coun¬
terfactual states of affairs in abstract discourse. In effect, such study may
have resulted in a form of borrowing transfer. Bloom’s research has met
with vigorously argued rebuttals from other researchers who have
claimed, among other things, that Bloom’s reading test was poorly trans¬
lated into Chinese, and who have provided counterevidence from other
reading tests (cf. Au 1983; Bloom 1984; Liu 1985). Bloom’s claims are
likely to be controversial for a long time to come, but some other evidence
besides the reading test argues for a limited Whorfian position and also
suggests a more straightforward methodology for investigating some
issues in linguistic relativism.’ Bloom searched for examples of coun-

3 Birdsong and Odlin (1983) present a detailed discussion of several aspects of

Bloom’s analysis. In that article and in this chapter, the use of counterfactual differs
somewhat from Bloom’s terminology.
 Semantics 75

terfactual arguments in a Chinese newspaper and found only one, which

happened to be in a translation of a speech of an American statesman.
While Bloom described this survey as “informal,” a controlled textual
analysis comparing, for example, hundreds of editorials in English with
hundreds of editorials in Chinese might establish - or disconfirm - that
counterfactual arguments are indeed rarer in Chinese discourse. Such
comparative studies might conceivably also work for other aspects of
the Whorfian question.
While the results of Bloom’s investigation are not conclusive, they do
suggest intriguing implications for the study of transfer. Celce-Murcia
and Larsen-Freeman (1983) note that the syntax of English hypothetical
and counterfactual statements is an area of exceptional difficulty for
learners. Some of the difficulty is no doubt related to the intrinsic com¬
plexity of the English system, but some of the difficulty may be related
to differences between languages such as English and Chinese. If bor¬
rowing transfer can occur, as suggested earlier, substratum transfer is
also conceivable.

Semantic case
One of the most important concepts in propositional semantics is the
notion of semantic case, or semantic roles. Many relationships between
syntactically different sentences can be specified with a small number of
cases, as in the correspondence between active and passive voice in the
following sentences: Bob stole the tomatoes and The tomatoes were
stolen by Bob. While the syntactic structures of the two sentences are
obviously different, the sentences share an identical proposition.4 Se¬
mantic case analysis allows the semantic relation between the two sen¬
tences to be specified in some detail. In both sentences the noun Bob
has the same semantic case of agent and the noun tomatoes has the same
case of patient (i.e., an entity that is affected through an action). The
analysis of propositions in terms of semantic case is useful not only for
an understanding of the semantic system of a language but also for cross-
linguistic comparisons of morphology and syntax (cf. Fillmore 1968;
Chafe 1970; Lyons 1977). For example, the possessive constructions of
English and Spanish differ somewhat in their morphosyntactic charac¬
teristics. Thus, the Spanish phrase los heroes de la nation can translate
into English either as the heroes of the nation or the nation’s heroes. A
contrastive description of Spanish and English grammar in this area
would posit a morphosyntactic but not a semantic difference; that is,

4 The traditional criterion determining that identity is that the two sentences must
share the same truth conditions; in other words, one sentence must be true if the
other is also true, or false if the other is also false.
76 Language transfer

while genitive (i.e., possessive) constructions are common to both lan¬

guages, the morphosyntactic manifestations are only partially the same
— English makes use of both a possessive inflection (as in nation's') and
a prepositional construction (as in of the nation), whereas Spanish only
allows a prepositional construction (e.g., de la nation). This contrast is
significant for the analysis of some of the difficulties that Spanish speak¬
ers encounter with possessive constructions in English (Section 6. 1).5
Since there is a fairly high degree of similarity in the Spanish and
English structures just described, the notion of semantic case might not
seem especially original or useful. However, case analyses are just as
applicable in comparisons of two languages that are typologically very
different. For example, Russian differs from Spanish and English in that
it relies greatly on morphological case, which (in Russian) is primarily
a system of noun endings that signal semantic case.6 The fact that Russian
relies so much on word endings to signal semantic case is relevant to
the considerable flexibility seen in Russian word order (cf. 6.1). A prop¬
ositional analysis employing the notion of semantic case allows for a
clear and consistent comparison of very different uses of morphology
and word order in Russian and English. In general, the notion of se¬
mantic, as opposed to morphological, case is essential to any contrastive
analysis since, as was stated earlier, only when sentences in two languages
have comparable meanings can aspects of their morphological and syn¬
tactic structure be compared (Section 3.4). By referring to a (presumably)
universal inventory of semantic cases as well as to other characteristics
of propositional structure, contrastive analysts can be explicit about how
sentences in two languages have comparable meanings.
The notion of case is also useful in the study of interactions between
propositional and lexical semantics. For example, in Irish there is no
word closely corresponding to the English verb have. The closest trans¬
lation of / have money is as follows:
Ta airgead agam.
Is money at-me.
Airgead is the syntactic subject of the sentence and agam is an example
of the “prepositional pronouns” that are common in Irish. While the
semantic case relations of Ta airgead agam and 1 have money are iden-

5 In many theoretical analyses of semantic case, the genitive construction is not

treated as a basic case but instead as a derivative of some other case (cf. Fillmore
1968; Lyons 1977). Aside from a further consideration of basic case relations, a
more detailed look at Spanish and English possessive constructions would have to
include a description of the restrictions on the use of prepositional constructions:
for example, the constraint that makes Tom's book more often acceptable than the
book of Tom.
6 Often the relation between semantic case and morphological case is rather weak (cf.
Jespersen 1929; Fillmore 1968).
 Semantics 77

tical, the syntactic forms of the two sentences are very different due to
Irish lacking a verb like have. This lexical difference between Irish and
English has had grammatical consequences. Hiberno-English often em¬
ploys prepositional constructions where other varieties of English do
not, and it therefore makes more extensive use of the objective mor¬
phological case found in the pronoun forms me, him, her, us, and them.
Thus, the standard English sentence He had a bad heart has a unique
counterpart in Hiberno-English: The heart was bad on him (Henry
1957). Such uses of prepositions in Hiberno-English are highly system¬
atic. As in other varieties of English, the most common antonym of on
in Hiberno-English is off, which can sometimes be used to express the
idea of not having something. Thus, a counterpart of the standard En¬
glish You won't have money worries is The money will be off you (Henry
1957).

5.2 Lexical semantics

Transfer and cognate vocabulary
Many language teachers and linguists have believed that similarities and
dissimilarities in word forms, along with similarities and dissimilarities
in word meanings, play a major role in how quickly a particular foreign
language may be learned by speakers of another language; intuition
suggests that the similarity between, say, French justifier and English
justify will give English speakers studying French (and French speakers
studying English) a head start in the acquisition of vocabulary. Com¬
menting on the comparative difficulties that speakers of European lan¬
guages encounter in learning to read European and Oriental languages,
Sweet observed:
Mastering the vocabulary of most European languages means simply learning
to recognize a number of old friends under slight disguises, and making a
certain effort to learn a residue of irrecognizable words, which, however,
offer less difficulty than they otherwise would through being imbedded in a
context of familiar words. The higher vocabulary of science, art, and abstract
thought hardly requires to be learnt at all; for it so consists either of Latin

7 Lass (1986) has noted some dubious claims about Irish-language influence in
Hiberno-English, including claims about prepositional constructions. Despite the
skepticism shown by Lass, there is strong evidence pointing to Irish influence in
certain prepositional constructions. First, some sentences involving such
constructions do appear to be unique to dialects in Ireland (e.g., The heart was bad
on him). Second, the wide range of uses of certain Hiberno-English phrases, such as
in it, suggests Irish influence even though such phrases are found in other English
dialects. Third, some prepositional constructions, such as with a while, occur as
part of phrases that are caiques of Irish idioms (Section 8.3).
78 Language transfer
and Greek terms common to most European languages or of translations of
them.
It is very different with a remote disconnected language such as Arabic or
Chinese. The abstract vocabulary of Arabic shows Greek influence, although
this affords very little practical help; but the terminology of Chinese
philosophy and science is independent of Western influence, so that every
extension of the vocabulary requires a special effort of memory and
reasoning. The task of mastering such languages is literally an endless one.
Enough Arabic grammar for reading purposes is soon acquired, the
—
construction being always perfectly simple at least in ordinary prose, but
the student may read one class of texts for years, and then, when he proceeds
to another branch of the literature, he may find that he can hardly
understand a word, this being almost entirely the result of the unfamiliarity
of the new vocabulary required. (Sweet 1899/1972:64—65)
Sweet’s observation has been expressed in somewhat different ways by
many other scholars, but only recently has there been much intensive
study of the effects of lexical similarity. In a comparison of the success
of Finnish- and Swedish-speaking students on an ESL test, Sjoholm
(1976) found that the former group did not do as well as the latter on
vocabulary questions, probably because Finnish does not share nearly
as much cognate vocabulary with English as Swedish does/ In a some¬
what similar study in the United States, Ard and Homburg (1983) com¬
pared the performances of ESL students speaking two different native
languages, Arabic and Spanish. Here again, the speakers of the language
having more lexical similarities with English (in this case, Spanish) were
considerably more successful on vocabulary questions. While both the
Finnish and American studies controlled for such factors as the linguistic
proficiency of the test takers, the American study went a step further.
In looking closely at the performance of Spanish speakers, Ard and
Homburg determined that those students did especially well with the
words on test items that had spelling identical or at least similar to that
found in Spanish forms (e.g., exiled and exilado). The benefits of rec¬
ognizing cognates may not be the only advantage that Spanish speakers
have in learning English; as Ard and Homburg observe, another likely
advantage is that Spanish speakers will have more time to concentrate
on unfamiliar vocabulary.
Despite the advantages of a large lexicon common to two languages,
there are nevertheless pitfalls in the form of “faux amis,” the “false
friends” notorious to many language teachers. For example, the forms
of French prevenir and English prevent seem to be as reliable signals of
a cognate relation as the forms of justifier and justify. Yet while the
latter pair is a true instance of a cognate relation, the former is not:

8 As suggested in Chapter 3, the use of the term cognate in second language

acquisition research is broader than the use of the term by historical linguists.
 Semantics 79

Prevenir means “to warn,” and thus the pair prevenir and prevent is a
pitfail for English learners of French and French learners of English
(Holmes 1977).
False friends come in other guises as well. One of the most common
problems in second language acquisition is when there is only a partial
semantic identity of cognates. Thus, the translation of English succeed
into Spanish as suceder will be acceptable in some contexts but not in
others: For example, while Truman sucedio a Roosevelt (“Truman suc¬
ceeded Roosevelt”) is acceptable Spanish, Sucedio en su trabajo (“He
succeeded in his work”) is not (Anthony 1952—3). Lexical transfer can
also occur when there is no morphological similarity between words
that appear to be semantically equivalent, as is seen in the following
error made by a Finnish student: He bit himself in the language. In
Finnish, a single form, kieli, is used for both “tongue” and “language”
(Ringbom 1986).
While a pair of cognates may be semantically similar, there are often
grammatical restrictions found in one language but not in another, and
such restrictions can occasion difficulty. For example, Adjemian (1983)
notes a cognate problem in the English of a French-speaking student:
At sixty-five years old they must retire themselves because this is a rule
of society. While the form retire reflects a true French-English cognate,
the French lexical item has a grammatical restriction that the ESL student
—
applied erroneously to the English form the use of the reflexive pronoun
is necessary in French (the infinitive verb form is se retirerf whereas the
use of the English reflexive themselves is not grammatical in the context
of the sentence. Transfer can also occur when the word forms are not
similar but the meanings are. For instance, the Swahili auxiliary weza
is roughly equivalent to English can. However, the Swahili auxiliary
often suggests the moral capacity for doing something. Hocking (1973)
claims a likely transfer error for a Swahili speaker trying to suggest a
person’s capacity for cruelty would be to say He’s a very cruel man
he can heat his children with a hoe. It is an open question whether this
—
type of transfer is as likely as in cases where there is an overt formal
similarity of lexical forms in two languages. Whatever the relative ad¬
vantages or disadvantages that cognate forms occasion, more and more
research on contrastive lexical semantics shows that recognition of cog¬
nates is often a problem. Learners may not always note the formal
similarities that mark a cognate relation, and they may not always believe
that there is a real cognate relationship (cf. Sections 8.1, 8.3).
Yet, even with the problems attendant in lexical similarities, there can
be little doubt that learners will find one language far easier to learn
than another if the one language shows many lexical similarities with
their native language and the other does not. Much of the advantage in
lexical similarity is likely to be evident in reading comprehension. The
80 Language transfer

Ard and Homburg study suggests, for example, that Spanish-speaking

students can be exposed to written texts with great lexical variety much
earlier than Arabic speakers can.

Lexical universals and acquisition

Even while transfer can play a major role in the acquisition of vocab¬
ulary, other factors are also at work. Although relatively little systematic
investigation has been undertaken, there appear to be important simi¬
larities between first and second language learners in their patterns of
lexical acquisition. For example, Strick (1980) compared judgments
made by native speakers of Persian about the lexical similarity of words
in their own language and their judgments about words in English. He
found that the judgments Persian speakers made about the similarity of
English words differed from their intuitions about Persian words. To a
considerable extent, their judgments about English words resembled
semantic intuitions of English-speaking children (Clark 1973). For ex¬
ample, the Persian speakers tended to classify the four terms Mr., Mrs.,
John, and Mary more frequently on the basis of a perceptible attribute
(i.e., male or female sex) than on the basis of social affiliation (i.e.,
formal terms of address in contrast with personal names). Such results
suggest that there may be a universal core of semantic information
accessible to all learners and which may aid in the acquisition of new
vocabulary. Determining the exact nature of that core is difficult, how¬
ever. Even while there seems to be a strong similarity between the se¬
mantic classifications of first and second language learners, there are
limits to the similarity.9 In children the evolution of lexical competence
is accompanied by major steps in cognitive development (Nelson 1974).
Although the notion of continuing cognitive development in adults is
plausible, there is little reason to believe that the acquisition of lexical
semantics in a second language is a close recapitulation of all earlier
developmental processes.
Whether they are communicative errors or simply innovations, the
uses of words in a second language often show influences besides those
from the native language, and some influences probably have a universal
basis. Although no classification has proven wholly satisfactory, there
seems to be a rather wide range of semantic innovations. Overextension
of word meanings is as likely as any other kind of overgeneralization.
Bamgbose (1982) notes that while many lexical innovations in Nigerian
9 Even in Strick’s study there is evidence of nonuniversal influences in the form of
transfer in some similarity judgments made by the Persian speakers. And in a study
using methods similar to those of Strick, Ijaz (1986) found that cross-linguistic
influences affected learner judgments of similarities between pairs of English
prepositions.
 Semantics 81

English have resulted from transfer from West African languages, some
have not. For example, in Nigerian English My father has traveled is
equivalent to My father is away. The Nigerian use of travel reflects a
new semantic range that has not resulted from transfer, according to
Bamgbose, but from a natural process of semantic extension, a process
found just as commonly among monolinguals (cf. Clark and Clark 1979;
Clark 1982; Lakoff 1987).'° When learners are not familiar with the
customary uses of a word, they may have recourse to approximation,
which is similar to overextension in that it is not a direct reflection of
either the native or target languages. For instance, Blum and Levenston
(1978) note that English speakers learning Hebrew may consider the
phrase asaf kesef (literally, “gather money”) an acceptable construction
to express the idea of saving money, even though the phrase does not
correspond directly to any phrasing either in Hebrew or in English. Some
approximations clearly result from metaphoric coinages that function
in lieu of the normally accepted target language form. For example,
Bartelt (1982) notes the description by an Apache ESL student of rotting
food as dead food, and Varadi (1983) notes instances of Hungarian ESL
students describing a balloon as an air ball or gas ball. Overextensions
and approximations as well as other types of semantic innovations are
also found in first language acquisition (Clark 1973; Clark and Clark
1977; de Villiers and de Villiers 1978). Accordingly, some of the errors
seen in lexical semantics are among the best evidence of the universality
of processes at work in all language acquisition contexts.
Such universality, however, coexists with language-specific nuances
in the lexicon. For example, while many languages have a word roughly
equivalent to the English word family, the kinship concept signaled by
the form can vary from complex extended families to nuclear families.
Indeed, the form family signals different meanings in different varieties
of English, such as Nigerian and American English (Platt, Weber, and
Ho 1984). One cross-linguistic study of universals and language-specific
nuances in the lexicon (Osgood, May, and Miron 1975) supports a
limited form of linguistic relativism. While Osgood, May, and Miron
found quite similar connotations associated with certain words (e.g.,
girl and its translation equivalents) in various languages, the emotional
significance of some words can vary considerably (e.g., policeman and
its translation equivalents). Such variation appears to have important

10 The use of the term overextension is not meant to suggest that lexical innovations
such as that seen in the case of travel are “errors” in the Nigerian context.
However, it is clear that in varieties of English found in Nigeria, Ireland, India,
and elsewhere, lexical innovations and other types of innovations have resulted
from such natural processes in second language acquisition as transfer and
overgeneralization.
82 Language transfer

effects on synonymy judgments of second language learners (e.g., Blum

and Levenston 1980).

Lexicon and morphology

As the “dictionary” component in the structural description of a lan¬
guage, the lexicon contains not only information about the meaning of
words but also morphological and syntactic information. In the case of
cognate forms, occurrences of lexical transfer are generally cases of both
morphological and semantic transfer. The studies of Sjoholm (1976) and
Ard and Homburg (1983) cited earlier suggest that morphological trans¬
fer involving independent words is as likely as any other kind of transfer.
However, in speech and writing there may be very strong constraints
working against the transfer of bound morphemes, which are prefixes,
suffixes, and any other forms that are meaningful yet incapable of stand¬
ing alone. For example, it may be the case that Spanish speakers do not
have any special advantage in using the English plural suffix seen in
tops, kites, and cakes even though the form is almost identical with the
Spanish form (cf. Dulay, Burt, and Krashen 1982). Yet while there is
skepticism about such transfer, some studies show instances of plural-
ization rules in one language being used in another, as in the phrase of
a Spanish-English bilingual child too manys cars, which reflects a number
agreement rule in Spanish whereby adjectives must agree in number with
the nouns they modify (Fantini 1985).
A number of scholars have claimed that the transfer of bound gram¬
matical morphemes is rare or nonexistent (e.g., Whitney 1881; Krashen
1978). Nevertheless, Weinreich (1953/1968) found a number of cases
where such transfer seems to have occurred. While noting the apparent
rarity of these cases, he suggested a formal criterion regarding the use
of native-language bound morphemes in another language:
it stands very much to reason that the transfer of morphemes is facilitated
between highly congruent structures; for a highly bound morpheme is so
dependent on its grammatical function . . .that it is useless in an alien system
unless there is a ready function for it. (Weinreich 1953/1968:33)
Thus, for example, it seems unlikely that an Arabic speaker trying to
learn Thai would attempt using the complex rules of Arabic word for¬
mation in Thai, which has virtually no inflectional morphology (cf.
Section 3.4). Weinreich’s notion of “highly congruent structures,” as
well as a related notion of “crucial similarity,” may well be essential to
any adequate account of language transfer in syntax as well as in mor¬
phology (cf. Section 6.3). However, just how much congruence there
has to be remains problematic (cf. Andersen 1983b).
The existence of general lexical similarities is probably a major influ-
 Semantics 83

ence on how much transfer of bound morphemes will take place. In the
case of two languages with many lexical similarities, such as Spanish
and Italian, the transfer of bound morphemes in speech appears to be
quite possible, as indicated in research by Meo Zilio (1959, 1964) on
both borrowing and substratum transfer. For example, in the Spanish
of Uruguay, where many Italian immigrants settled, the adjective nub-
ladeli (“rather cloudy”) reflects a fusion of Spanish nublado (“cloudy”)
and an Italian suffix -eli. Even in very dissimilar languages, however,
there do appear instances of morphological transfer. The Greek spoken
in some parts of Turkey earlier in this century frequently showed bor¬
rowing transfer in the form of Turkish suffixes on Greek nouns and
verbs (Dawkins 1916). And morphology in a variety of the Chinese
language Hui has apparently been affected by substratum transfer; in
the dialect studied by Li (1984) there are morphological case suffixes
unlike those found in other varieties of Chinese but very much like those
of some Mongolian and Turkic languages spoken in the same region of
western China.
Whatever the constraints may be on the transfer of bound morphemes
in production, there are probably fewer constraints on transfer in com¬
prehension processes. In other words, the similarity of bound morphemes
in two languages may facilitate reading and listening in the same way
the similarity of free morphemes does. For example, the similarity of
suffixes in English and Spanish, such as -ous and -oso in scandalous and
escandaloso, is likely to help readers identify words as cognates.

5.3 Summary and conclusion

There is little question that lexical similarities in two languages can
greatly influence comprehension and production in a second language.
Cognates can provide not only semantic but also morphological and
syntactic information, and while some of the information may be mis¬
leading, some can facilitate acquisition. What is less clear is the impor¬
tance of linguistic relativism. In all probability, anything that can be said
in one language can also be said in any other, although it might be easier
to express a particular thought in one language (cf. Hockett 1954). A
moderate Whorfian position has some empirical support (e.g., some of
the findings of Osgood, May, and Miron 1975), but there remains con¬
siderable uncertainty about how much influence semantic structures in
one language can have on production and comprehension in another
language. Whatever the importance of relativism, the need for the study
of universals is evident in some cross-linguistic research on cognition
and in studies of the development of word meaning in first and second
language acquisition. It seems clear that second language research on
84 Language transfer

universals and relativism cannot make great progress without a clearer

understanding of the problem of meaning not only in discourse and
semantics but also in interactions between syntax and other subsystems.
In the next chapter, there is considerable attention given to interactions
between syntax, discourse, and semantics.

Further reading
Lyons (1977) provides a very detailed discussion of the vast literature on
propositional and lexical semantics. Good historical reviews of the problem
of linguistic relativism are found in studies by Penn (1972) and Bloom
(1981). Ard and Homburg (1983) discuss not only lexical transfer but also
some methodological problems in determining such transfer.
6 Syntax

The notion of syntactic transfer has long been controversial, and em¬
pirical studies of second language syntax have fueled much of the debate
(Sections 2.1, 2.2). Despite the apparent absence of cross-linguistic in¬
fluence in some studies, however, there is considerable evidence for
positive transfer involving articles and other syntactic structures (Section
3.2). There is also evidence for negative transfer in cases such as the
Hiberno-English verb phrase seen in He’s after telling a lie (Sections 2.1,
3.2). And a great deal of evidence has also been found for syntactic
transfer (both positive and negative) in studies of word order, relative
clauses, and negation. An extended look at research in those three areas
is appropriate for several reasons. First, the number of studies in those
areas is rather large. Second, several studies have involved target lan¬
guages other than English. Third, such work is related in important ways
to work in other areas of linguistics, such as discourse analysis and
syntactic typology. Finally, many of the studies indicate that transfer
interacts with other factors in acquisition.

6.1 Word^order—
Word order has been one of the most intensively studied syntactic prop¬
erties in linguistics, and in second language acquisition research there
are now numerous studies of learners’ word-order patterns. The study
of second language word order has been useful not only for a better
understanding of transfer but also for an understanding of discourse,
syntactic typology, and other factors affecting second language
acquisition.

Word-order rigidity
As discussed earlier, the vast majority of human languages have either
VSO, SVO, or SOV as their basic word order (Section 3.4). Yet while
most languages can be compared in terms of these three patterns, a

85
86 Language transfer

contrastive analysis based only on such comparisons could be mislead¬

ing. In the case of Russian and English, for example, both languages
have SVO as their basic order but they vary considerably in terms of
rigidity. For instance, each of the following Russian sentences cited by
Thompson (1978) has a different word order, whereas most of the
English translations have SVO order:
Kolja kupil masinu Kolya bought the car (neutral)
Kolya bought car
S V O
Kolja masinu Tupil Kolya BOUGHT the car
S O V
Kupil Kolja masinu Kolya did buy the car
V S O
Kupil masinu Kolya KOLYA bought the car
V O S
Masinu Kolja kupil The car, Kolya BOUGHT it
O S V
Masinu kupil Kolja The car, it was Kolya who
O V S bought it
The flexibility of Russian word order is readily explained. The form
masinu signals the syntactic role of direct object in contrast with other
forms such as masina, which can signal the role of subject.1 Such a
reliance on bound morphology (as in the use of the morphological case
suffix -m) is often seen in languages having flexible word orders.
The relative rigidity of word orders is, according to Thompson (1978)
and others, just as important a typological property as basic word order.
VSO, SVO, and SOV languages can thus be subcategorized according
to the rigidity of their word order, as illustrated below:

Language Word order Rigidity

Irish VSO Rigid
Biblical Hebrew VSO Flexible

Twssian
— SVQ=
SVO Flexible
Persian SOV Rigid
Turkish SOV Flexible

I The flexibility of word order in languages such as Russian raises the question
whether all languages have a “basic” order such as SVO or SOV. Some linguists are
more skeptical than others on this question (cf. Hawkins 1983; Givon 1984a).
Certain analyses suggest that it may be possible to develop criteria for “basicness”
even when word order is quite flexible (e.g., Keenan 1976, 1978). Even if a
consensus, ever develops on such criteria, there are still other reasons to be cautious
in using the concept of basic word order in a contrastive analysis. For example, a
 Syntax 87

Languages designated as rigid can vary; Irish may be more rigid than
English (Filppula 1986). Similarly, languages designated as flexible can
vary; Russian is probably more flexible than Spanish (Thompson 1978).
Nevertheless, the classification of languages as either rigid or flexible
allows for a more detailed characterization of syntactic contrasts be¬
tween languages.
Rigidity appears to be a transferable property. Speakers of a flexible
language, for example, may use several word orders in English even
though English word order is quite rigid. Evidence of such transfer
appears in a study by Granfors and Palmberg (1976), which lists nu¬
merous errors in English word order in a guided composition task per¬
formed by native speakers of Finnish, a flexible SVO language. One
example of negative transfer that resulted is This weekend got F. any
fish (“This weekend F. caught no fish”). Granfors and Palmberg attribute
such errors to the flexibility of Finnish word order; in the same study,
native speakers of Swedish, a more rigid SVO language, made far fewer
errors. Similarly, Trevise (1986) notes cases that reflect the relatively
flexible word order of French: for example, I think it's very good the
analysis between the behavior of animals and the person.
Aside from studies of production, studies of comprehension also point
to the importance of rigidity (Bates and MacWhinney 1981; Gilsan
1985; Gass 1986). For example, the results of Gilsan’s study indicate
that English speakers learning Spanish experience comprehension dif¬
ficulties related to the relatively flexible word order of Spanish. Some¬
times, however, rigid word order may be quite helpful in the
comprehension and production of a second language by younger learn¬
ers. There is evidence that some children in the early stages of acquisition
are likely to prefer rigid word order regardless of whether word order
is rigid in either the native or target language. For example, a study by
Pienemann (1981) of Italian children learning German shows a frequent
use of a rigid SVO order, and many of the children’s sentences do not
seem to be based on word-order patterns of either German or Italian
(cf. Section 8.2)?

single language can employ word-order patterns that seem diametrically opposed.
German, for instance, often has SVO in main clauses but SOV in subordinate
clauses, and thus scholars have often debated whether the basic order in German is
SVO or SOV (cf. Comrie 1981; Zobl 1986).
2 While reliance on rigid word order appears to be a common developmental
characteristic, it is not found universally in first language acquisition. Slobin (1982)
has shown that the morphological typology of the language being acquired can
affect children’s reliance on word order. Such typological factors might also affect
second language acquisition (cf. Zobl 1983).
88 Language transfer

Discourse factors in word order

With the exception of Pienemann’s study, most of the results discussed
in the preceding section are compatible with the predictions of a con¬
trastive analysis that takes word-order rigidity into account. However,
some of the production and comprehension difficulties that learners
experience with word order in a second language have little to do with
native language influence. Pienemann’s study and others indicate that a
rigid word order is advantageous because it simplifies (or at least seems
to simplify) language processing routines (cf. Clahsen 1984). On the
other hand, flexible word order also has advantages, since variations in
word order can signal special discourse conditions. Flexible word order
in learners’ use of a second language may at times reflect discourse
constraints in the native language, but at other times it may reflect
discourse signals found in all languages.
The expression “free word order” is sometimes used to characterize
languages having a flexible order, but discourse considerations suggest
that that expression is misleading. Surveys of Ute, Turkish, and Serbo-
Croatian, all languages with a very flexible order, show that some of
the six possible permutations (i.e., SVO, SOV, VSO, VOS, OSV, OVS)
are much more frequent than others (Slobin 1982; Givon 1984a). Far
from being “free” or random, word order in flexible languages seems
to reflect constraints imposed by the discourse needs of speakers and
listeners. The constraints are complex and some seem to be language¬
specific, but some of the most important constraints are evident in the
limited number of patterns used to signal information about topics in
discourse (Section 4.2). Definitions of the term topic are problematic,
partly because topics and grammatical subjects are often distinct, as the
discussion later indicates (cf. Li and Thompson 1976; Brown and Yule
1983). Nevertheless, a necessary condition for something to qualify as
a topic in conversation or writing is that it be a focus of information.
Once a focus has been established, speakers (or writers) typically elab¬
orate on the information already presented, with any elaboration func¬
tioning as a comment.
A conventional way to present information is through the use of a
topic-comment sequential order. In English syntax, this order frequently
— —
though not always coincides with a subject-predicate order. Such
ordering seems to play an important role in first and second language
acquisition. Establishing the topic in a stretch of discourse can be difficult
for learners, and the use of a topic-comment order is one way of min¬
imizing the difficulty. As Givon (1 984b:1 28) observes:
If one has difficulty establishing a new topic, or if one suspects that the
hearer is likely to experience such difficulty, the most sensible strategy is first
 Syntax 89

to make sure that the topic is firmly established [emphasis in the original]
and only then to come up with the new information.
There is detailed evidence for the heavy reliance of some learners on
topic-comment patterning in the early stages of acquisition. In a highly
detailed study of the speech of a Hniong refugee in Hawaii named Ge,
Huebner (1983) presents evidence that topic and comment are basic
categories in Ge’s early speech and that topic-comment order is likely
when both categories are overtly signaled, as in the following sentences:
mii wok
As for me, I walked
hos, ai reis
As for horses, I raced (them)3
In both sentences, the first word can be identified as the topic and all
subsequent words as the comment. It is possible that Ge was influenced
by word-order patterns in Hmong, but other research suggests that a
learner’s use of topic-comment order need not result from native lan¬
guage influence. The following example from a Spanish immigrant in
Germany (Klein 1986) shows a patterning just as rigid as Ge’s:

——
—
Heute vier Schule neu meine Dorf; ich klein Kind eine Schule vielleicht
hundert Kind; heute vielleicht ein Chef o Meister zwanzig oder
—
fiinfundzwanzig Kind; ich Kind vielleicht hundert Kind

—
Today four school new my village; 1 little child — one school perhaps
hundred child; today perhaps one boss or master — twenty or twenty-five
—
child; 1 child perhaps hundred child

The comparison that this speaker attempted to make is that: (1) there
are now four schools in his village whereas when he was a young child,
there was only one school with about one hundred children; (2) while
the student-teacher ratio today is twenty or twenty-five to one, when
the speaker was a child the ratio was about a hundred to one. This
comparison is analyzable in terms of a topic-comment analysis:

Topic Comment
today four schools
in the past one school
today twenty children
in the past one hundred children

3 Huebner notes that such dear cases of overtly expressed topics and comments are
somewhat rare because Ge seemed to know that some topics were easily identifiable
and therefore did not need to be expressed.
 90 Language transfer

The word order of Spanish does not require that information be pre¬
sented in such a rigid sequence, and Givon’s explanation quoted earlier
may best account for the speaker’s performance (cf. Schumann 1 984).4
Although Huebner’s study and others have shown the importance of
the topic-comment pattern, an explicit indication of the topic is not as
frequent in the speech of learners as one might expect. Once topics have
been clearly established, they are predictable for listeners and therefore
are not incessantly signaled, even by speakers with very little proficiency.
Technically known as zero anaphora, the omission of a form signaling
a predictable topic (here, the speaker) is seen in the following example
from a Spanish speaker responding to a question about going to the
movies:
In Saturday no like, no time. Watch TV . . .
On Saturday I don’t like [to go to the movies], I don’t have any time. I watch
TV... (Givon 1984b)

Instead of being omitted, topics may follow comments, as seen in another

example from the same speaker:
. . . (me come-back Mexico in 1974), is come my family ...
... (1 came back from Mexico in 1974), they came too, my family did . . .
(Givon 1984b)

The appearance of family after come instead of before it is known

technically as right dislocation, as opposed to left dislocation, which is
seen in the following example from a Korean speaker:
. . . diploma my son high school get . . .
. . . diploma, my son got a high school diploma . . . (Givon 1 984b)
z Learners probably do not use these different word-order patterns ran¬
domly. In a study of the discourse of second language learners with three
different native language backgrounds, Givon (1984b) has argued that
the striking similarities in learners’ word-order patterns reflect universal
principles of signaling topic continuity. The basic principle is summa¬
rized thus by Givon (1984b: 126):
Of topics that are fully expressed as an independent word or pronoun, those
that are most continuous/predictable will display COMMENT-TOPIC (VS,
VO) word order; while those that are less continuous/predictable will display
TOPIC-COMMENT (SV, OV) word order.

4 Klein uses a different set of terms, which may in fact be preferable to topic and
comment. However, his evidence is very similar to that of Huebner, who does
employ topic and comment, hi the interest of terminological consistency, only
Huebner’s classification is used in this chapter.
 Syntax 91

The signaling system thus reflects speakers’ judgments about how con¬
tinuous and predictable a topic is.5 Aside from word order, other syn¬
tactic devices play a role in the system, part of which is given in the
following scale:
Most continuous topic
Zero anaphora
Unstressed pronoun
Right dislocation
Neutral order
Left dislocation6
Least continuous topic
This analysis has been consistently applied not only to some aspects of
learners’ syntax but also to syntactic characteristics of a wide variety of
languages, including Japanese, Ute, Amharic, Spanish, and English (Gi¬
ven 1983a). In light of such cross-linguistic regularities, the word-order
errors of speakers of Finnish and French noted earlier might not reflect
transfer so much as universal principles of topic continuity. Several re¬
searchers have in fact argued that what appears to be word-order transfer
is often a reflex of discourse constraints (e.g., Muysken 1984). There
are, however, problems with such arguments as an explanation for all
cases of anomalous word order. First, many of the errors are remarkably
like flexible patterns in the native languages of some learners (cf. An¬
dersen 1984). Second, comparative evidence cited by Granfors and Palm¬
berg (1976) suggests that native speakers of a language with rigid word
order (Swedish) make fewer errors than do speakers of a language with
flexible word order (Finnish). Finally, negative transfer accounts better
for many errors made in cases where the basic word orders of two
languages differ, as evidence in the next section suggests.

Transfer in basic word-order patterns

Among the most controversial topics concerning the acquisition of sec¬
ond language syntax is the extent of word-order qransfer, especially
regarding the two basic word orders, SVO and SOV. Despite reported

5 The Spanish speaker’s clause cited earlier, is come my family, might seem to
contradict Givon’s generalization about more predictable topics having a VS order;
such a clause could well be the first time that speaker has mentioned his family. The
generalization about VS word order is still tenable, however, if such a clause
appears early in a discourse “paragraph” (Givon 1983b). It should be noted that
the word order in is come my family does not rule out the possibility of transfer
interacting with discourse universals, since the VS order here has a close translation
equivalent in Spanish.
6 In the interest of brevity, some details (e.g., the role of definiteness in noun phrases)
have been omitted from the discussion of Givon’s analysis.
92 Language transfer

instances of basic word-order transfer, a number of skeptics have con¬

tended that such transfer does not take place, and others have argued
that if it does take place, it is a negligible phenomenon. Arguments
against word-order transfer have generally rested on universalist inter¬
pretations (cf. Section 3.4). One is that what appears to be transfer may
be an artifact of discourse manipulations related to topic continuity and
the like (e.g., Muysken 1984). Another universalist argument is that
some constraint of Universal Grammar blocks the use of basic word¬
order transfer (e.g., Zobl 1986). While both arguments are universalist,
there is a crucial difference between them: The discourse-based argument
suggests that in the early stages of acquisition, learners’ word-order
patterns are “asyntactic” and reflect universal principles of discourse
organization. On the other hand, the Universal Grammar argument
suggests that some innate, principle of syntactic organization is available
in second as well as in first language acquisition. Clearly.) whatever the
facts are with regard to word-order transfer, they are relevant to fun¬
damental questions in the study of second language acquisition.
The universalist arguments are relatively easy to test when they involve
predictions directly the opposite of those of a contrastive analysis. Ruth¬
erford, for example, has claimed that “Japanese learners of English do
not at any time produce writing in which the verb is wrongly placed
sentence-finally” (1983:367). Since the basic order of Japanese is SOV
and the basic order of English SVO, transfer might take place. Yet in a
word-order study by Rutherford, virtually no evidence of such transfer
emerged, a finding supported by other studies (cf. Zobl 1986). Never¬
theless, universalist arguments cannot be true in any absolute sense for
the simple reason that some studies do show clear evidence of basic
word-order transfer. For instance, a highly detailed study of Japanese
speakers of Pidgin English in Hawaii contains examples of full-fledged
SOV patterns, as, for example, mi: cu: stoa gecc (me two store get =
“1 got/acquired two stores”) as well as several examples of clauses in
which objects appear before verbs (i.e., OV), as in hawai kam (“I came
to Hawaii”) (Nagara 1972).
With regard to OV patterns, the claim has sometimes been made that
they do not reflect native language influence but instead discourse strat¬
egies. For example, native speakers of English sometimes use OSV order
for purposes of contrast. The clause The soup we ordered, for instance,
can be used contrastively in sentences such as The soup we ordered, the
salad we did not. If the subject were omitted, this pattern would be The
soup ordered. Evidence in the preceding section suggests that zero an¬
aphora is common in the early stages of second language acquisition.
Thus, it is possible to argue that even when a Japanese speaker uses OV
patterns, a discourse strategy instead of transfer is at work. Such ar¬
guments have not, however, been substantiated with studies of second
 Syntax 93

language discourse like those described in the previous section. More

important, such claims do not square with comparative data on Ha¬
waiian Pidgin English. Bickerton and Givon (1976) found that speakers
of Philippine languages such as Ilocano and Tagalog, which are VSO,
produced a large number of VSX sentences, and they found that speakers
of Japanese produced a large number of SXV sentences. Two additional
facts strongly suggest that word-order transfer is the best explanation
of these findings. First, the immigrants from the Philippines produced
almost no SXV sentences, and the immigrants from Japan produced
almost no VSX sentences. Second, the least proficient speakers in both
groups tended to produce the largest number of variant patterns (whether
SXV or VSX). Still other comparative evidence is seen in Givon’s (1984b)
study of topic continuity in second language acquisition. The Korean-
English text in that study shows more instances of OV than VO phrases,
whereas the Philippine-English text shows many VO but almost no OV
phrases. Since Korean is, like Japanese, an SOV language, the data in
Givon’s study parallel those of the Bickerton and Givon study comparing
Japanese and Filipinos.
Studies of Andean Spanish also provide evidence of word-order trans¬
fer. Many of the Indians of Peru and Ecuador speak both Spanish, an
SVO language, and Quechua, an SOV language, and studies of both
child and adult bilinguals indicate that Quechua word order influences
the local varieties of Spanish.” A study of the Spanish spoken by Peruvian
children shows that five-year-olds used as many OV patterns as they did
VO ones, and while seven- and nine-year-olds used more VO patterns,
OV patterns were still quite common (Lujan, Minaya, and Sankoff
1984). The decrease in OV patterns appears to continue as speakers
grow older and speak a less distinctly regional variety of Spanish. Other
research indicates that adults who use Quechua more also tend to use
OV patterns more (Muysken 1984). As in the case of Hiberno-English,
the speech of bilinguals (past or present) has probably influenced the
speech of monolinguals; Ecuadorean Spanish speakers who do not know
any Quechua also frequently use OV patterns (cf. Section 8.3).7 9

7 As with some other scholars whose work is discussed in this section, Bickerton and
Givon contrast the patterns of the pidgin speakers in terms of a difference between
SVX, VSX, and SXV instead of SVO, VSO, and SOV. The use of X allows
grammarians to account for other patterns besides ones involving a direct object.
For example, the indirect question seen in / asked what she was doing can be
described as X in an SVX pattern.
8 Non-Andean varieties of Spanish do use SOV patterns with pronoun objects, but
there is little doubt that the basic word order of Spanish is SVO.
9 Muysken (1984) attempts to show that word-order transfer is only indirectly
involved in the basic SOV patterns of Andean Spanish. Space does not permit a
fully detailed analysis of Muysken’s claims, but the most important one he makes
should be noted (cf. Odlin 1987). Muysken claims that the reason that objects
94 Language transfer

Studies of Dutch and German offer particularly intriguing examples

of where word-order transfer can lead to different acquisition patterns.1"
Both languages employ SOV in subordinate clauses and SVO in main
clauses, although other main-clause word orders are possible under spe¬
cial circumstances. Examples of Dutch word orders cited by Jansen,
Lalleman, and Muysken (1981) appear below:
ik zag gisteren een beer Independent clause, SVO
I saw yesterday a bear
gisteren zag ik een beer Independent clause, VSO"
yesterday saw I a bear
omdat ik gisteren een beer zag Subordinate clause, SOV
because I yesterday a bear saw
A study of the speech of immigrant workers by Jansen, Lalleman, and
Muysken indicates that speakers of Moroccan Arabic, a language that
mainly uses SVO, tend to identify the basic word order of Dutch as
SVO, whereas speakers of Turkish, an SOV language, tend to identify
the basic order as SOV. Consistent with such identifications, the Dutch
word-order patterns of the Arabic speakers showed a strong preference
for SVO and those of the Turkish speakers for SOV (cf. Appel 1984).
The transfer-based preferences of the workers for SVO or SOV were
especially evident among the less proficient speakers of Dutch; as in the
study of Bickerton and Givon, there seemed to be an inverse relation
between transfer and proficiency (cf. Section 8.1).
Research on Italian and Spanish workers in Germany also provides
strong evidence of transfer of basic word-order patterns (Meisel, Clah-
sen, and Pienemann 1981). The SVO order of Italian and Spanish appears
to have influenced some workers’ use of SVO instead of SOV order in
German subordinate clauses (cf. LoCoco 1975). These findings are es¬
pecially significant in light of research on the acquisition of German
(Clahsen 1982). Children acquiring German as their native language

occur before verbs so often in Andean Spanish is not a result of transfer but a
result of “stylistic” (i.e., discourse) strategies. His claims are not, however,
supported by any actual analysis of discourse. Moreover, if one accepts Muysken’s
discourse-based explanation, one has to account for why Peruvian five-year-old
bilinguals use OV patterns so much more than do seven-year-olds, who in turn use
them much more than do nine-year-olds (cf. Lujan, Minaya, and Sankoff 1984). In
fact, Zobl (1983) provides evidence suggesting that four- and five-year-old children
seldom use word orders influenced by discourse factors (Section 8.2).
10 Zobl (1986) has seen the German and Dutch cases as examples of word-order

— —
transfer. NevertheiesSn’is analysis cannot account for and indeed it is
contradicted by the cases of Hawaiian Pidgin English and Andean Spanish.
Moreover, there are other cases that contradict the analyses of Zobl and
Rutherford (cf. Odlin 1987).
1 1 The presence of a sentence-initial adverb is one condition under which VSO order
is employed in Dutch and German.
 Syntax 95

rarely seem to use SVO in subordinate clauses, and thus the evidence
suggests that the development of basic word order can and sometimes
does proceed differently in first and second language acquisition.
Evidence from the acquisition of English, Spanish, Dutch, and German
thus strongly suggests that basic word order is one kind of syntactic
pattern susceptible to native language influence. Accordingly, univer-
salist arguments cannot fully account for the acquisition of basic word
order. Yet a contrastive analysis may sometimes overpredict word-order
problems. The transfer of basic word order does not always occur in
situations where, for example, the native language is SOV and the target
language is SVO. Although there has yet to appear any satisfactory
explanation for cases in which transfer does not occur, several factors
are probably involved. Target language patterns can lead to overgener¬
alizations (Section 2.2). For instance, a study by Snow (1981) shows
that English speakers may overgeneralize the SOV order of Dutch sub¬
ordinate clauses and thus produce them in main clauses. Aside from
tendencies of learners to overgeneralize target language patterns, there
seem to be other cases in which cross-linguistic influences are highly
improbable, such as when a native language pattern is not really “basic.”
For example, French and Spanish, which are clearly SVO languages,
nevertheless have a rigid SOV order when pronouns instead of nouns
signal objects; however, Zobl (1980) has noted that virtually no cases
of transfer into English of such SOV patterns (e.g., I them see} have
appeared in the second language acquisition literature.12 In addition to
such structural factors, there are probably other factors that inhibit
transfer, such as the linguistic awareness of the individual (Odlin 1987).
Accordingly, even while a contrastive analyst may justifiably predict
basic word-order transfer in some acquisition contexts, an array of struc¬
tural and nonstructural factors may affect the prediction - in some cases
basic word-order transfer may not take place, and in other cases such
transfer may be seen in only the earliest stages of acquisition.

Word order within the clause

Other constituents besides S, V, and O are also subject to word-order
rules. In noun phrases (NPs) in English, for example, articles normally
precede adjectives and nouns in the noun phrase (e.g., the fierce lion},
and languages generally have rules specifying the occurrence of elements
within noun phrases, verb phrases, and other constituents. Even lan¬
guages with a somewhat flexible order can show considerable rigidity

12 Transfer may, however, take place in the opposite direction, as when an English
speaker learning French uses SVO instead of SOV patterns regardless of whether
the object is a noun or a pronoun (cf. Zobl 1980; Andersen 1983b).
96 Language transfer

in some areas of their syntax (Hawkins 1983). Since the rules governing
the position of adjectives, adverbs, and other word classes vary consid¬
erably from one language to the next, it is natural to expect to find cases
of word-order transfer in constituents within clauses, and indeed such
cases exist. Yet, as with basic word order, learners may encounter dif¬
ficulty for reasons besides native language influence.
Some of the clearest evidence of within-clause transfer comes from a
study by Seiinker (1969) of Hebrew speakers learning ESL. Using data
obtained from interviews with Israeli students, Seiinker found a strong
tendency for speakers to follow Hebrew instead of English norms for
the placement of adverbial elements, as seen in the following error: /
like very much movies. Seiinker did not attribute all adverbial errors to
transfer, but native language influence did account well for the vast
majority of the errors. Other research on the English of non-native
speakers also shows strong evidence of transfer. In a study of possessive
constructions in ESL papers written by Spanish-speaking students, An¬
dersen (1979) found frequent examples of NPs that were word-for-word
translations from Spanish, as in the porch of Carmen from el balcon de
Carmen. Nevertheless, Andersen also found many NPs that could not
be explained in terms of native language influence but that could be
explained in terms of overgeneralization: for example, the United State
President, which does not conform to Spanish word order. In addition
to these types of errors, another type described in Andersen’s study
suggests that some errors may result from an interaction of transfer and
syntactic overgeneralization. For example, the flute’s lessons reflects not
only the English word order (i.e., flute lessons) but also the Spanish use
of an overt signal of a genitive construction (cf. Section 5.1). That is,
the use of the inflection ’s seems to serve as a semantic counterpart of
the Spanish preposition de in the translation equivalent lecciones de
flauta.
In studies of the acquisition of other languages besides English, there
is also considerable evidence for the importance of cross-linguistic as
well as other types of influences on the production of within-clause word
order (e.g., Snow 1981; Veronique 1984; Lujan, Minaya, and Sankoff
1984). One area of special interest is the implicational statements for¬
mulated by Greenberg (1966), Hawkins (1983), and others (Section 3.4).
Some researchers have tried to determine, for example, if there is a
developmental relationship between the acquisition of rules governing
the sequence of nouns and adjectives (NA or AN) and the acquisition
of rules governing basic word order/ (VSO, SVO, etc.) when the word
orders of the native and target languages differ (cf. Lujan, Minaya, and
Sankoff 1984; Zobl 1986). One tendency, for example, might be for
the acquisition of SVO order in English not to take place until learners
have mastered the preposition-noun order, which contrasts with the
 Syntax 97

noun-postposition order seen in languages such as Japanese and Turkish

(3.4). Yet while the theories upon which such studies are based are indeed
interesting, empirical observations do not always support them (e.g.,
Muysken 1984). In the discussion of negation in Section 6.3, there will
be further consideration of some of the issues related to developmental
sequences and word order.

6.2 Relative clauses

Branching direction in relative and adverbial clauses
Although there is no necessary connection between the position of words
and phrases in a clause and the position of clauses in a sentence, some
research suggests important relations. For example, Kuno (1974) has
investigated the tendency in SOV languages for relative clauses to precede
the noun they modify as opposed to the tendency in VSO (and most
SVO) languages for relative clauses to follow the noun, as seen in the
following examples from English (SVO) and Japanese (SOV):
The cheese that the rat ate was rotten
Nezumi ga tabeta cheese wa kusatte ital!
rat ate cheese rotten was

These examples only begin to show how different English and Japanese
syntax can be. If a relative clause modifies rat in the above English
sentence, as in the rat that the cat chased, the result will be a very complex
sentence with center-embedding, as seen in The cheese that the rat that
the cat chased ate was rotten. The SVO order of English appears to
constrain the use of relative clauses following the subject of the main
clause, since a center-embedding within a center-embedding may lead
14
to extreme comprehension difficulties.13 In contrast, Japanese syntax
does not lead to center-embedding with this type of relative clause.

13 The particles ga and wa in Kuno’s example mark the syntactic categories of topic
and subject (cf. Li and Thompson 1976).
14 Kuno’s analysis makes strong claims about the relation between language structure
and human perceptual capacities. Such claims are suspect, however, since there
exist languages in which the degree of permissible center-embedding appears to be
greater than that found in English and Japanese (cf. Hagege 1976; Gazdar and
Pullum 1985). Nevertheless, it is not at all clear that such languages are numerous.
If there are few of them, the problem may be similar to what typologists find in
other domains. For example, a language such as Vietnamese may have an unusual
eleven-vowel system, but there are many more languages that have five-vowel
systems (Section 3.4). Even though the significance of some statistical patterns is
often hard to interpret, it would be short-sighted to claim that such patterns have
no significance just because there are counterexamples to the dominant trend.
98 Language transfer

According to Kuno, the Japanese translation equivalent of the above

English sentence is quite comprehensible:
Neko ga oikaketa nezumi ga tabeta cheese wa kusatte ita
cat chased rat ate cheese rotten was

Such differences in comprehensibility suggest that the branching direc¬

tion favored in a language is an influence on the types of syntactic
complexity likely to occur in a language. English primarily relies on a
Right Branching Direction (RBD), as is seen in the above examples in
which the relative clauses appear to the right of the head noun (cheese).
In contrast, Japanese primarily relies on a Left Branching Direction
(LBD), as is seen in the relative clause nezumi ga tabeta cheese wa, in
which the modifying clause appears to the left of the head noun.
The above examples suggest that the SOV typology of Japanese more
easily accommodates syntactic complexity, but there are many instances
in which SOV does not allow much complexity. When a noun is in
direct-object position in English, multiple embeddings of relative clauses
are possible:
John read the letter that Mary wrote to the boy that Jane was in love with.
The LBD of Japanese, on the other hand, leads to a translation with a
highly complex center-embedding:
John ga Mary ga Jane ga aisite iru syoonen ni kaita tegami o yonda
loving is boy to wrote letter read
There is evidence that when two languages show a difference in prin¬
cipal branching directions, the acquisition of complex syntax will be
more difficult than when both languages show the same branching di¬
rection. In a study of Japanese-speaking and Spanish-speaking ESL stu¬
dents, Flynn (1984) found that Japanese speakers had more difficulty in
repeating sentences with adverbial clauses, such as The boss informed
the owner when the worker entered the office. Flynn attributed the
greater success of the Spanish speakers to the fact that Spanish, like
English, is an RBD language. According to her analysis, Spanish-speaking
students have an advantage in repeating sentences such as the one above
since the adverbial clause follows the main clause, thus conforming to
an RBD pattern found in Spanish. 15 Flynn and Espinal (1985) have

15 The results of Flynn’s study do not seem attributable to differences in listening

comprehension or in overall proficiency, as there were controls for these factors in
her experimental procedure. However, there may be other problems in her design.
For example, there is an ambiguity in the sentence The boss informed the owner
when the worker entered the office - the subordinate clause may be either an
indirect question or an adverbial clause. Such ambiguity might affect
comprehension. Moreover, Eubank (1988) was not able to replicate some of
Flynn’s results on branching direction effects.
 Syntax 99

provided further evidence that when an LBD is predominant in the native

language, the acquisition of RBD patterns in English is more difficult. 16
Aside from difficulties encountered on repetition tests, another man¬
ifestation of negative transfer is evident in one of the earliest empirical
studies of relative clauses in second language research. Schachter (1974)
provided evidence that differences in branching directions favored in
relative clause patterns will occasion underproduction (Section 3.3).
Consistent with the predictions of a contrastive analysis, the speakers
of Japanese and Chinese (LBD languages) in Schachter’s study seem to
have often avoided using relative clauses in written compositions; in
contrast, equally proficient students who spoke Arabic and Persian
(which are, like English, RBD languages) used many more such clauses
(cf. Schachter and Hart 1979). Although the Arabic and Persian speakers
produced a greater number of errors in the relative clauses that they
used, the similarity of patterns in the native and target languages ap¬
parently led them to attempt writing more sentences with relative con¬
structions (cf. Section 8.1).

Relativized positions, transfer, and universals

The order of clauses in_a_sgntence is not the only factor affecting the
acquisition of complex syntax. In the case of relative clauses, another
crucial factor is the gramtnaticaLm/e^of nouns andoronouns^Keenan
(1985) defines restrictive irelativization as a construct'usualTy having a
“domain noun” and invariably having a modifying clause. For example,
the sentence The musician who played at the concert is from China has
a relative construction consisting of a domain noun (in this case, mu¬
sician) and a clause modifying the noun (in this case, who played at the
concert).'7 One of the most common patterns in relative clauses is to
have within the clause a pronoun with the same reference as the domain
noun (e.g., who and musician in the preceding example). However, one
important difference between the domain noun in main clauses and the
pronoun in relative clauses is that they may take different grammatical
roles. While who functions as subject (SU) within the subordinate clause
and musician as subject within the main clause in the preceding example,
this identity of grammatical roles is not inevitable, as five of the following
examples show:

16 Flynn and others who have worked on branching direction problems do not
attribute to transfer all difficulties encountered by learners. Nevertheless, they do
see native language influence as an important factor in many difficulties.
17 While domain nouns in a relative clause are not always the same as what
grammarians often refer to as head nouns, this identification is valid for English
and for the other languages to be discussed in this section (cf. Keenan 1985).
100 Language transfer
SU The musician who played at the concert is from China.
DO The musician whom we met at the concert is from China.
IO The musician to whom we sent the message is from China.
OPREP The musician from whom we got the message is from China.
GEN The musician whose son played at the concert is from China.
OCOMP The musician who George is taller than is from China.
In each of these sentences musician is the subject of the main clause,
but in each sentence (except for the first) the relative pronoun takes a
different grammatical role: direct object (DO), indirect object (IO), prep¬
ositional object (OPREP), genitive (GEN), and object of comparison
(OCOMP).
Some languages mark DO, IO, and other constituents with a re¬
sumptive instead of a relative pronoun, as in the following example from
Persian (Keenan 1985:146):
Man zan-i-ro ke John be u sib-e-zamini dad misnasam
I woman that John to her potato gave know
The translation of this sentence in English, I know the woman that John
gave the potato to, uses a relative, but not a resumptive, pronoun, and
a translation that would more closely represent the IO constituent in
Persian (that is, be u) would be anomalous: I know the woman that
John gave the potato to her.'*1 The designations DO, IO, and so forth
are useful even in cases where no pronoun appears in the subordinate
clause, as in The musician we met at the concert is from China. The
relative clause in that sentence is considered to have a DO position since,
if there were a relative pronoun marking that position, it would be
functioning as a direct object within the clause. Thus, whether a con¬
stituent is marked by a relative pronoun, a resumptive pronoun, or by
no form, the designations of DO, IO, and so on are used to describe
relativized positions.
Not all languages have syntactic equivalents of some of the above
types of relative clauses. A cross-linguistic survey of relativization pat¬
terns by Keenan and Comrie (1977) suggests that there is an implica-
tional relationship among relativizable positions:
SU>DO>IO>OPREP>GEN>OCOMP
For example, if OPREP is a relativized position in a given language, IO,
DO, and SU will also be relativized positions. The converse is not nec¬
essarily true, however; if SU, for instance, is a relativized position, DO,
IO, and the rest may or may not be relativized. Keenan and Comrie’s

18 Persian ke is usually translated as that, but it is more of a subordinating

conjunction than a pronoun (Lazard 1957). The transliteration of the Persian
words presented in this chapter departs in minor ways from the transliteration
given by Keenan.
 Syntax 101

survey of fifty languages indicates considerable variation in what posi¬

tions may be relativized in any language.19 While English allows rela¬
tivization on all six of the above positions, Tagalog allows it only on
SU. Thus, although The musician who played at the concert is from
China could be translated easily into Tagalog, a more circuitous trans¬
lation would be needed for The musician whom we met at the concert
is from China.
Keenan and Comrie see the implicational sequence as related to com¬
prehensibility; thus, structures incorporating SU, DO, and IO may be
more easily understood than ones incorporating OPREP, GEN, and
OCOMP (cf. Keenan and Bimson 1975; Fox 1987). Support for the
Keenan-Comrie interpretation is seen in the fact that the implicational
relationship appears to be a good predictor of how frequently the six
relativized positions will appear in the discourse of a given language. In
a count of relative clause types in two English novels, Keenan (1975)
found that SU clauses were the most frequent, DO the next most fre¬
quent, and so forth. Still further support for the Keenan-Comrie inter¬
pretation is seen in the fact that the implicational relationship also seems
to be a good predictor of which relativized positions will make use of
resumptive pronouns, as in the Persian example cited above.20 The po¬
sitions on the right-hand end of the implicational sequence, OPREP,
GEN, and OCOMP, are far more likely to have such pronouns than are
those on the left-hand end, SU, DO, and IO. Keenan and Comrie suggest
that the use of resumptive pronouns aids in comprehension by making
the structure of the relative clause more transparent.
Characteristics of relative clause structure, such as pronoun retention,
pose an interesting problem for second language acquisition researchers.
On the one hand, the implicational hierarchy posited by Keenan and
Comrie may somehow reflect language universals. On the other hand,
there is considerable cross-linguistic variation in relative clause struc¬
tures, and such variation may occasion language transfer. In fact, re¬
search on the acquisition of English and Swedish suggests that transfer
does indeed play a role in the erroneous use of resumptive pronouns.
From the results of a sentence-combining task given to ESL students
with nine different native languages, Gass (1979, 1983) determined that
the native language was likely to influence the frequency of resumptive

19 Comrie and Keenan (1979) have provided an updated account, but one not
substantially different from what is summarized here.
20 More often than not, positions such as GEN and OCOMP require resumptive
pronouns if they can be relativized. In fact, Keenan and Comrie have identified an
implicational relation that seems to govern the use of resumptive pronouns. Singler
(1988) has noted that cases do exist that do not conform to such an implicational
relation; however, he believes such exceptions do not constitute major problems
for the Keenan-Comrie analysis (cf. Joseph 1983b).
102 Language transfer

pronouns in DO, IO, and OPREP positions. Thus, sentences like I know
the woman that John gave the potato to her were more often produced
by speakers of languages such as Persian (i.e., languages in which the
IO position must have a resumptive pronoun). The influence of transfer
also appears in some of the results of a grammaticality judgment test
that Gass gave to the same subjects: speakers of languages such as Persian
were more likely to accept DO and IO sentences like I know the woman
that John gave the potato to her as grammatical in comparison with
speakers of languages that do not make use of pronoun retention.
Further evidence of transfer comes from a recent study by Singler
(1988) of pidginized forms of Liberian English. For example, speakers
of Vai rarely used resumptive pronouns in SU position, while speakers
of a language called Dan used them rather frequently. As Singler ob¬
serves, such tendencies reflect the fact that Vai does not allow resumptive
pronouns in SU position, whereas Dan requires such pronouns in the
same position. Moreover, Singler’s evidence suggests that some highly
unusual patterns of resumptive pronoun use in certain other West Af¬
rican languages are mirrored in relativization patterns in some varieties
of Liberian English. As in other cases to be discussed, the social factors
involved in the acquisition of these varieties of Liberian English seem to
have greatly influenced patterns of transfer (Section 8.3).
The findings on English relativization have been corroborated by re¬
search on the acquisition of Swedish, a language that, like English, does
not use resumptive pronouns.21 Using a picture-description test, Hylten-
stam (1984) elicited relative clauses from speakers of four different lan¬
guages. Speakers of Greek and Persian, languages that allow pronominal
retention, produced many more instances of resumptive pronouns than
did speakers of Finnish and Spanish, languages that do not allow pro¬
nominal retention. Another result similar to Gass’s findings was that the
resumptive pronouns used by Greek and Persian speakers were fre¬
quently in DO, IO, and OPREP positions, whereas such pronouns were
less frequently used in the same positions by speakers of Finnish and
Spanish.
Drawing on evidence from grammaticality judgment tests, some re¬
searchers (loup and Kruse 1977; Tarallo and Myhill 1983) have offered
universalist arguments that the use of resumptive pronouns does not
indicate transfer. For example, in the study carried out by Tarallo and
Myhill, native speakers of English studying German, Portuguese, and
other languages that do not use resumptive pronouns often considered
as acceptable ungrammatical sentences that had such pronouns. In a

21 Resumptive pronouns do occur in relative clauses in some nonstandard varieties of

English (Jespersen 1954). However, the influence of such nonstandard varieties on
the English of students in Gass’s study is unlikely.
 Syntax 103

similar study of grammaticality judgments of French sentences, native

speakers of English frequently accepted sentences with resumptive pro¬
nouns even though French, like English, does not use such pronouns
(Birdsong, Johnson, and McMinn 1984). Since neither native nor target
language facts account well for the performance of English speakers
in these studies, a universalist explanation appears highly plausible
especially since there is also some support for this explanation in the
—
research of Gass and Hyltenstam. In the latter two studies, learner be¬
haviors closely matched the implicational patternings predicted by the
Keenan-Comrie analysis. Nevertheless, such an explanation cannot ac¬
count for all the known facts about the use of resumptive pronouns.
The English and Swedish studies indicate that speakers of languages not
using pronoun retention were less likely to employ resumptive pronouns
than were speakers of languages using pronoun retention. Thus, even
though a sentence such as / know the woman that John gave the potato
to her could come from a French speaker, it would more likely come
from a Persian speaker.
Despite its inability to explain everything, a universalist explanation
does account well for a number of facts. First, the studies of Gass and
Hyltenstam indicate that speakers of any native language are likely to
use resumptive pronouns in the GEN and OCOMP positions. Second,
the findings of Keenan (1975) on the frequency of relative clause types
are corroborated by a count of the types of relative clauses used in student
compositions (Gass 1983): SU sentences were more frequent than DO
sentences, which in turn were more frequent than IO sentences, and so
forth. Third, although there are some discrepancies in the data (especially
in Liberian English), the frequency of correct production of relative
clauses generally conforms to the Keenan-Comrie implicational se¬
quences (Gass 1979, 1983; Hyltenstam 1984). Fourth, while a contras¬
tive analysis can predict the use of avoidance strategies, as Schachter’s
research indicates, the Keenan-Comrie analysis provides insights about
what relative clause patterns students are most likely to avoid (Gass and
Ard 1984).

Other considerations
Relative clauses show other variations in structure besides those con¬
sidered so far, but the role of such variations in second language ac¬
quisition has not yet been studied in as much detail. For example, there
is some evidence supporting Kuno’s analysis of the difficulties that
embedded relative clauses may occasion. In a study of ESL students’
grammaticality judgments, loup and Kruse (1977) found that student
judgments were less accurate for sentences with embedded relative
clauses (e.g., The dish which fell on the floor broke in half) than for
104 Language transfer

ones in which the relative clause was not embedded (e.g., The little girl
is looking for the cat which ran away). It is not clear, however, that the
results of loup and Kruse’s study reflect universals at work in second
language acquisition. In an investigation of relative clauses in first lan¬
guage acquisition, Sheldon (1977) found that embedding did not affect
children’s ability to interpret relative clauses in their native languages
(English and French). More crucial were sentences having “parallel func¬
tions,” for example, The lion that pushes the horse knocks down the
cow, where both the domain noun and the relative pronoun function
as subjects within their respective clauses. Even though that sentence
has an embedded relative clause (i.e., that pushes the horse), it proved
to be of a type easier to understand than nonembedded types that did
not have parallel functions (e.g., The lion knocks down the cow that
pushes the horse).
Structures that are highly language-specific are another aspect of rel¬
ative clauses warranting further investigation. For example, English
makes use of a somewhat unusual relative clause pattern that does not
have a relative pronoun but that does have a stranded preposition mark¬
ing an IO or OPREP position in a clause (for example, / want the pencil
I write with). Birdsong, Johnson, and McMinn (1984) found that native
speakers of English do not frequently accept as grammatical the French
translation equivalents of such patterns: in this case, Je veux le crayon
f ecris avec. Moreover, research by Adjemian and Liceras (1984) suggests
that French learners of English and French learners of Spanish (a lan¬
guage with relative clause patterns more like those of French) are equally
unwilling to accept as grammatical sentences with stranded prepositions.
On the other hand, preposition stranding is transferable to some extent.
In a study of English-speaking children in a French immersion program,
Seiinker, Swain, and Dumas (1975) cite the following case of transfer:
Un chalet qu'on va aller a (“A cottage that we’re gonna go to”) (cf.
White 1987). The description of the French immersion program suggests
that the social context, a relaxed setting of learners speaking with each
other, may well have encouraged cross-linguistic influence despite pu¬
tatively universal constraints on the transferability of preposition strand¬
ing (cf. Section 8.3).

6.3 Negation
Formal dimensions of negation
The study of negation in second language acquisition has sometimes
been regarded as simply a question of word order. In some languages,
negators (i.e., forms expressing negation, such as no and nicht) may
 Syntax 105

either precede a verb phrase, as in Spanish Juan no va (“Juan is not

going”), or follow a verb phrase, as in German Ingrid kommt nicht
(“Ingrid is not coming”). As much of the following discussion will show,
the use of preverbal or postverbal negation is indeed an important clue
to relations between transfer and universals in second language acqui¬
sition. Nevertheless, fundamental differences between negation and
word order must be noted. Negation is fundamentally a semantic notion,
whereas word order is a formal arrangement (Section 5.1). As such,
word order can play a role in expressing negation, but such a role
presupposes constructions of two or more words. Since one-word ne¬
gation is possible, the need for word order is not absolute, as the fol¬
lowing example shows:
A: Are 'you hungry?
B: No.
Aside from independent words, prefixes and suffixes may also serve
as negators, as in uninterested, disinterested, nonaligned, and thought¬
less. The boundary between words and suffixes functioning as negators
is sometimes hazy. In English, negators may be either words (as in Alice
is not here) or contracted forms that have some properties of suffixes
(e.g., Alice isn’t here). The rules for negation are not always as simple
as the Spanish and German examples given above suggest. For example,
English verb-phrase negators are often neither preverbal nor postverbal,
strictly speaking, since in negative constructions the verb phrase usually
has both an auxiliary and a main verb (e.g., Alice hasn't come). And
multiple negation in French shows that there can be both preverbal and
—
postverbal negators in the same clause (e.g., Jean ne voyage pas “John
is not traveling”), as two forms, ne . . . pas, function together as negators.
Thus although the following discussion considers in detail preverbal and
postverbal patterns, a full understanding of negation in second language
acquisition involves much more.

Negation and developmental sequences

As with the other areas of syntax already reviewed in this chapter,
negation shows evidence not only of transfer but also of other influences
on acquisition. Negation, in fact, was one of the areas first described by
researchers who maintained that transfer did not offer a fully satisfactory
account of acquisition (cf. Section 2.2). For example, data from
Norwegian-speaking children (Ravem 1968) show that the predictions
of contrastive analysis are sometimes utterly wrong. Like German, Nor¬
wegian uses postverbal negation in main clauses having one-word verb
phrases, and so a contrastive analysis would naturally predict that a
Norwegian learner of English would frequently use postverbal negation
106 Language transfer

in the target language. Ravem found, however, that preverbal patterns

were predominant: for example, I not like that and I not looking for
edge. Moreover, there were almost no examples of postverbal negation.
Occurrences of preverbal negation have also been described in studies
of the English spoken by speakers of other native languages, such as
Spanish and Greek (summarized by Schumann 1979). In these cases the
occurrence of preverbal negation is generally compatible with a transfer
explanation, since Spanish and Greek (and many other languages) use
preverbal negation. Yet, since transfer cannot account for all learners’
uses of the preverbal pattern, other factors influencing acquisition might
account for any use of preverbal negation, no matter whether an indi¬
vidual’s native language uses preverbal or postverbal negation or some
other pattern.
In attempts to compare first and second language acquisition, Wode
(1981, 1983a, 1983b, 1983c) studied the use of negation in the perfor¬
mance of monolingual and bilingual children in terms of developmental
sequences. In both types of acquisition, Wode (1981, 1983a) found
evidence of a developmental sequence consisting of three major stages:
(1) one-word negation; (2) two- or more word negation; and (3) intra-
sentential negation. From his analysis emerges a pattern of evolution in
negation that is similar in both first and second language acquisition.22
One-word negation is somewhat like zero anaphora (Section 6.1) a
lone negator is often sufficient to signal negation and to maintain topic
—
continuity, as seen in the question and response given earlier of Are you
hungry? No. The formal simplicity of a single negator and the minimal
need to know any other vocabulary are undoubtedly key factors in the
early appearance of one-word negation in the acquisition either of one’s
native language or of another language. The form a learner uses need
not be identical with the target form. As Wode (1983a) has noted in
observations of children learning German as their native language, the
pronunciation of the negator may only weakly resemble an adult’s pro¬
nunciation of the form; nevertheless, the choice of negator that a child
uses often seems to be modeled after adult pronunciation.
Two-word utterances that include a negator are clearly similar to ones
having three or more words, although longer utterances certainly require
more knowledge of vocabulary and of some kind of clause structure. In
two-word negation, the negator may accompany either a verb (as in a
Spanish speaker’s use of No understand [Cancino, Rosansky, and Schu¬
mann 1978]) or some other element (e.g., No pink [Young 1974; cited

22 Despite giving one of his papers a title that suggests a four-stage model, Wode
presents a tripartite analysis in which the second part (two- or more word
negation) consists of two substages. Other analyses by Wode (e.g., Wode 1981)
show a somewhat different approach to the developmental sequence, but the
evolutionary pattern remains essentially the same.
 Syntax 107

by Schumann 1979]). As utterances grow longer and new syntactic pat¬

terns begin to emerge, the negation pattern may still be far from the
target pattern, as in No drink some milk (Wode 1981).
Intrasentential negation presupposes the development of other syn¬
tactic structures, since in this stage the negator generally occurs within
the sentence. In the sentence I not looking for edge, for example, the
negator occurs between the subject and the predicate. According to
Wode, another crucial indicator of intrasentential negation is the grow¬
ing use of other negators besides those found in the earliest stages of
acquisition. Thus, in English the use of the form no often gives way to
the use of other forms such as not and don’t. However, the development
of alternative forms may not necessarily begin in the intrasentential stage
(cf. Schumann 1979).
Wode has acknowledged that other stages are also involved in the
acquisition of negation, but his findings generally concur with those of
other researchers (cf. Schumann 1979; Stauble 1984). Much in his anal¬
ysis supports a universalist position. Neither native- nor target-language
facts can explain all of what goes on in the early stages. For example,
German-speaking children’s use of two-word constructions, such as no
finish and no cold in English, cannot be fully explained in terms of either
the structure of German or of English (Wode 1983b). And in all like¬
lihood, the acquisition of negation in any language by children or by
adults will often show an initial use of one-word negation followed in
turn by two-word negation and then by more complex patterns. The
three-stage analysis also offers insights about the use of word order in
negation. Noting the frequency of preverbal negation in the second and
third stages of acquisition, Wode (1983c) has speculated that universal
language-processing abilities may be the ultimate explanation for the
frequent appearance of preverbal negation both in language acquisition
and in the syntax of many of the languages of the world (Dahl 1979b).
Although there is a clear universalist strain in his analysis, Wode

—
—
(1983b) has also shown that important differences differences related
to transfer can arise in the development of negation by children learning
a second language and children learning their native language. While
there are attested cases of German-speaking children using English sen¬
tences such as I’m steal not the base, which shows a postverbal pattern,
there are few if any comparable examples of this done by children
learning English as their native language. Thus, a contrastive analysis
of English and German does appear to have some predictive power; for
at least one stage in the acquisition of negation, German-speaking chil¬
dren may use postverbal negation in English.
In attempting to account for the appearance of transfer, Wode has
suggested that structures in both the native and target languages must
be developed to the point where a “crucial similarity” between them
108 Language transfer

can occasion transfer (cf. Section 5.2). In the case of the bilingual children
that he studied, for example, the syntax of learners’ German and English
— —
had to be considerably developed and similar for there to be post¬
verbal negation, which did not develop until preverbal patterns had
appeared (cf. Wode 1978, 1981, 1983a; Zobl 1980). Without sufficient
development, structures in both languages might be too similar for a
distinct native language influence to be evident. Since the earliest de¬
velopmental stages in the acquisition of any language are remarkably
alike and since the forms in these stages are structurally simple, some
phenomena have several possible explanations, including transfer and
target language influence (cf. Adiv 1984).

Developmental sequences: one or many?

Research on the acquisition of negation in other contexts shows addi¬
tional evidence for the existence of developmental sequences. Hyltenstam
(1977) examined the use of negation by adult learners of Swedish, which
employs postverbal negation in main clauses and preverbal negation in
subordinate clauses. Using a written grammar test, Hyltenstam found
frequent word-order errors, such as preverbal negation in main clauses
and postverbal negation in subordinate clauses. Errors in subordinate
clauses were much more frequent; in fact, mastery of negation in sub¬
ordinate clauses generally implied mastery of negation in main clauses,
but not vice versa. In other words, a developmental sequence seems to
exist in the acquisition of Swedish - learners first master postverbal
order in main clauses and only then the preverbal order in subordinate
clauses. A common intermediate stage in this developmental continuum
was evident in the use of postverbal negation in all contexts, with such
overgeneralization only gradually giving way to the correct use of word
order in both types of clauses. Hyltenstam was able to arrive at an even
finer-grained characterization of learners’ negation patterns in terms of
the type of verb found with the negator: there was a clear tendency in
main clauses to use preverbal negation with main verbs and postverbal
negation with auxiliaries.2’ While he found virtually no evidence of
transfer in learners’ use of negation, Hyltenstam (1984) has more re-

23 Hyltenstam’s analysis has sparked some controversy. Jordens (1980) criticizes,

among other things, Hyltenstam’s assumptions about the role that auxiliaries play
in the development of negation (cf. Hyltenstam 1982; Jordens 1982). While the
claims and counterclaims go beyond the scope of this chapter, one of the questions
Jordens raises is highly significant for language acquisition studies generally: how
closely do the categories of a learner’s developing linguistic competence match
those of an adult native speaker? Since it involves problems of comparison
(Section 3.2), that question is especially significant for transfer research (cf. Bley-
Vroman 1983).
 Syntax 109

cently observed that any firm conclusions regarding transfer and nega¬
tion in Swedish were difficult to reach in light of a number of meth¬
odological problems (cf. Section 6.2).
The negation studies of Wode and Hyltenstam have provided valuable
insights about developmental sequences, but their results do not entirely
clarify certain issues. For example, while Wode’s research indicates that
the development of negation in first and second language acquisition is
not invariably the same, there remains the question of just how much
the acquisition sequences can differ. And in second language acquisition
there is also the question of just how much transfer can influence the
evolution of different sequences. A number of researchers have argued
that such influence is considerable (e.g., Schumann 1979; Zobl 1980).
Support for that position comes from a detailed comparison of the
negation patterns in the English of Spanish and Japanese speakers (Stau-
ble 1984). Stauble’s findings suggest that the paths along which the
syntax of negation develops sometimes cross but often diverge. The
Spanish and Japanese speakers least proficient in English made use of
preverbal negation, but Spanish speakers tended to employ no extremely
frequently whereas Japanese speakers used both no and not. Among the
more proficient individuals, the Japanese speakers tended to use a wider
variety of negators, although the most proficient individuals demon¬
strated relatively few differences due to transfer.
Stauble’s results strongly suggest that a lexical similarity between two
languages can increase the likelihood of not only lexical but also syntactic
transfer. The Spanish negator no is phonologically similar to the English
form no, and this similarity may explain why Spanish speakers seem
more ready to employ it than to employ such forms as not and can’t.14
Although the cross-sectional nature of Stauble’s research does not allow
for firm conclusions, it may also be the case that Spanish speakers persist
longer in using no. A study by Schachter (1986) of a Colombian named
Jorge indicates that use of no persists even while other types of negators
such as don’t and never are used more frequently. While the lexical
similarity between English no and Spanish no is one possible reason for
such persistence, discourse function also seems to be an important factor.
Schachter’s analysis suggests that some negators are especially likely to
be used when particular speech acts are negated. For example, when
Jorge used no before a prepositional phrase, the negator sometimes
signaled the nonexistence of something (e.g., No in Colombia?), but he
preferred using don’t to signal rejections (e.g., I don’t like).

24 Another likely instance of lexical and syntactic transfer interacting is found in

Hiberno-English. Henry (1957) suggests that sentences such as It’s flat it was were
especially likely to develop among Irish-English bilinguals because of the similarity
of the Irish copula is and English it’s (cf. Section 4.3).
110 Language transfer

Aside from variations in a developmental continuum that are due to

transfer, individual differences in learners may lead to somewhat dif¬
ferent continua. This is more than just a possibility, as Clahsen (1984)
has shown in a longitudinal study of Spanish and Portuguese learners
of German (cf. Wode 1981). Even though, for instance, both Spanish
and Portuguese have preverbal negation, not all the individuals studied
by Clahsen used preverbal negation in German a great deal.25 While
Clahsen maintains that his data support the notion of developmental
stages, he stresses that some learners show more of a tendency to simplify
target language patterns:
it might be expected that simplifying learners will tend to use preverbal NEG
placement, whereas learners who are more oriented toward use of the target
norm will tend to prefer postverbal negation from the beginning of the
learning process. (1984:237)
In the case of “simplifying” learners, there is likely to be more of a
resemblance between first and second language acquisition. Such dif¬
ferences in learner behavior show the importance of taking individual
variation into account in the study of transfer (Section 8.1).

6.4 Summary and conclusion

A comprehensive treatment of syntactic transfer would have to include
much more than the present survey of word order, relative clauses, and
negation. Nevertheless, some tendencies evident in those three areas
suggest that similar tendencies will also emerge when other areas of
syntax are investigated in greater detail in the future. In the acquisition
of word order, relative clauses, and negation, transfer figures as an
important factor, but it often occurs in conjunction with other acqui¬
sition processes, some of which show hints of typological and universal
influences at work. With regard to word order, an important influence
of language universals seems to be the effect that discourse has on the
arrangement of basic clause constituents. With regard to relative clauses,
typological factors such as the range of relativizable positions and the
primary branching direction that a language shows appear to have a
strong influence on the development of complex syntax, in what sub¬
ordinate clauses learners first use, and how successfully they use them.
And with regard to negation, there is also evidence of developmental
sequences, albeit with some individual differences. Aside from the likeli-

25 In a study of negation patterns used by English-speaking adults learning German,

Eubank (1986) found even less evidence for the use of preverbal negation. Eubank
attributed such a difference to the effects of formal instruction (cf. Section 8.3).
 Syntax 111

hood that one-word negation is a universally occurring first stage of

development, there appears to be a strong typological influence that
favors the use of preverbal negation at an early stage of syntactic de¬
velopment in some languages even when neither the native nor the target
language uses preverbal negation. Yet while typology and universals
seem to play a role in the acquisition of negation and the other syntactic
structures discussed, so does transfer. Whether syntactic transfer is as
important as (for example) phonological transfer is still an open ques¬
tion. However, the discussion of typologies and universals in the next
chapter indicates that the problem of transfer in second language pro¬
nunciation is similar to the problem of syntactic transfer, since language¬
specific facts are not the only data important to consider.

Further reading
A text by Givon (1984a) on syntax provides an extensive discussion of facts
about word order and negation relevant to this chapter, and an article by
Keenan (1985) provides a good typological survey of relative clause
formation. Gass (1984) surveys much of the work on second language syntax
that has led to a reconsideration of transfer by many researchers in the
1980s. A very detailed longitudinal study by Wode (1981) offers insights not
only about the development of second language syntax but also many other
aspects of child bilingualism.
7 Phonetics, phonology, and
writing systems

There is little doubt that native language phonetics and phonology are
powerful influences on second language pronunciation, and this chapter
will consider some of the more important aspects of those influences.
As with other aspects of second language performance, pronunciation
often shows other influences besides cross-linguistic ones; accordingly,
developmental and other factors will be examined, as well as transfer
involving writing systems.

7.1 General versus specific predictions

The literature on second language acquisition and language teaching is
replete with descriptions of the difficulties that learners encounter in
trying to pronounce sounds in a foreign language, and contrastive
explanations for such difficulties are quite common (Dechert, Briigge-
meir, and Futterer 1984). As the discussion in the next section shows,
— —
there is considerable but not total support for specific contrastive
predictions. Aside from specific predictions, general predictions are
also possible. For example, one might predict that a speaker of Thai
will have more difficulty in learning to pronounce English than a speak¬
er of Persian will. In fact, some research on ESL pronunciation (Su¬
ter 1976; Purcell and Suter 1980) supports that prediction; in carefully
controlled analyses of judgments of pronunciation accuracy, the pro¬
nunciation of speakers of Thai and Japanese did not receive as high
ratings as did the pronunciation of speakers of Arabic and Persian.
Suter (1976) acknowledges that such results do not in themselves ex¬
plain what features of native language pronunciation contribute most
to a high or low evaluation of pronunciation accuracy. Moreover, such
results do not entirely rule out the possibility that nonstructura) fac¬
tors (e.g., personality factors) account for some of the differences
in performance (Section 8.1). Nevertheless, detailed analyses of
several possible factors showed the native language of students to be
an especially good predictor of what evaluation their speech in En¬
glish received as opposed to such factors as length of residence in the
112
 Phonetics, phonology, and writing systems 113

United States and motivation to pronounce English correctly (Purcell

and Suter 1980).

7.2 Phonetic and phonological transfer

Phonetic differences
A cross-linguistic comparison of sounds in two languages should include
descriptions of the phonetics as well as the phonology of the native and
target languages (cf. Briere 1968). A phonetic description is necessary
since sounds in two languages often show different physical character¬
istics, including both acoustic characteristics (e.g., the pitch of a sound)
and articulatory characteristics (e.g., how widely the mouth is open in
producing a sound). Two languages frequently have sounds which may
seem identical but which in fact are acoustically different. For example,
a comparison of an American English /d/ with a Saudi Arabian Arabic
/d/ shows several differences (Flege 1980). Among the differences, the
duration of an English /d/ at the end of a word (e.g., in bad} tends to
be shorter than its Arabic counterpart. As a contrastive analysis would
predict, Saudi learners’ pronunciation of the English /d/ tends to show
a duration longer than the English norm. Nevertheless, Flege’s analysis
shows that learners are capable of modifying their production of sounds
so that their pronunciation comes closer to the target language norms.
The modifications often do not result in the attainment of target language
norms, but rather in approximations that are neither fully nativelike nor
targetlike. Thus, the “English” /d/ that some Arabic speakers produce
is longer than the target language norm but is shorter than a normal
Arabic /d/. 1 The development of such compromise forms shows the im¬
portance of learners’ unconscious judgments. While these judgments are
influenced by learners’ knowledge of forms in their native language, they
are also based on judgments about how phonetic material is structured
in the target language.2

Phonemic differences
Learners’ identification of the Arabic /d/ and the English /d/ illustrates
the importance of phonetic similarity in interlingual identifications,
1 Although the term interlanguage coined by Seiinker (1972) usually connotes more
than simply compromise forms, the example of the /d/ in Flege’s study supports
Seiinker’s contention that second languages are systematic varieties in their own
right. Trudgill (1986) has shown the usefulness of Seiinker’s analysis for the study
of compromise forms in cases of dialect contact (cf. Section 2.1).
2 In the case of subtle phonetic differences such as those that distinguish the Arabic
/d/ from the English /d/, learners’ judgments of similarities and differences are likely
to be unconscious (cf. Section 8.3).
114 Language transfer

which are the equivalence relations that learners establish between the
native and target languages. While any resemblance between sounds
creates the potential for identifications, the judgments of equivalence
that learners make are affected by much more than just the acoustic
properties of sounds in the native and target languages. The similarity
of cognate forms, for example, may induce learners to establish corre¬
spondences between sounds that are phonetically very different (cf. Sec¬
tion 5.2). For instance, the uvular /r/ of Parisian French and the retroflex
/r/ of American English have very different phonetic properties, but there
are other acoustic, as well as orthographic, cues that may induce Amer¬
ican learners of French to equate the French /r/ with the English /r/ in
cognates such as route.
Another factor that influences interlingual identifications is the set of
relations implicit in the phonemic system of a language. A study by
Scholes (1968) of the perception of vowels by native and non-native
speakers of English indicates that non-native speakers are likely to cat¬
egorize foreign language sounds largely in terms of the phonemic in¬
ventory of the native language (cf. Liberman et al. 1957). In Scholes’s
study, native speakers distinguished between the vowels /e/ and /ae/ (as
in the words rain and ran}, whereas speakers of Russian and Greek did
not. In contrast to other non-native speakers of English, speakers of
Persian, which, like English, has a phonemic contrast between lei and
leel, did distinguish between the two vowel sounds. Although such re¬
search clearly demonstrates the importance of native-language phonemic
systems, the explanation for some perceptual confusions is a bit less
straightforward. For example, Spanish has a nasal consonant /n/ pho¬
netically similar to the English /n/ and, like the English sound, a Spanish
/n/ can occur at the end of a word. One might naturally expect, then,
that Spanish speakers would never have difficulty in perceiving the En¬
glish /n/ at the end of a word (e.g., in fan}. However, a study by Marck-
wardt (1946) showed some confusion on the part of Spanish speakers
in distinguishing between /n/ and the nasal phoneme /r)/ in fang. Since
the latter nasal sound exists in Spanish, but never in a phonemic contrast
such as between fan and fang, the phonetic similarities of the Spanish
and English nasals do not always outweigh the differences in phonemic
systems for purposes of interlingual identification. In Marckwardt’s
study, such systemic differences appear to have encouraged hyper¬
correction.
While Scholes’s study indicates that major differences in phonemic
inventories can cause perceptual confusions in foreign language learning,
the phonemic inventory of the native language does not totally impede
perception of foreign language sounds.. Phonetic mimicry is one kind of
evidence that individuals can recognii^ sounds“rather different from
those in the native language. Flege and Hammond (1982) studied the
 Phonetics, phonology, and writing systems 115

“Spanish accents” imitated by native speakers of English and determined

that the sounds the English speakers produced were often approxima¬
tions of Spanish vowels and consonants. Another kind of evidence, iden¬
tification of differing accents, also indicates that individuals can
recognize foreign language sounds. For example, loup (1984) found that
native speakers of English can identify different groups of non-native
speakers simply on the basis of their pronunciation (cf. Smith and Bisazza
1982).
The phonetic sensitivity needed for such identifications is evident
among individuals learning a second language even when their native
language seems to impede certain perceptual distinctions. The distinction
between / 1/ and /r/ in English is notoriously difficult for speakers of
languages such as Japanese and Korean, which do not have that pho¬
nemic distinction. Frequent problems involving misperceptions and mis¬
pronunciations of these sounds in words such as lice and rice make it
natural to suppose that Japanese speakers, for example, may not actually
attend to the phonetic material that distinguishes these sounds. However,
research by Mann (1986) indicates that Japanese speakers can attend to
the differences. In one sense, such findings should not be surprising, since
learners often do show success in achieving pronunciations closer and
closer to target language norms (cf. Dickerson 1974; Flege 1980; Borden,
Gerber, and Milsark 1983). Nevertheless, individuals differ in their per¬
ceptual acuity, and it may be that only individuals with especially high
phonetic sensitivity will be able to overcome most of the inhibiting
influence of phonological patterns in the native language (cf. Section
8.1).

Types of segmental errors

Although cross-linguistic differences in phonetics and phonology have
important consequences for perception and comprehension, the most
salient consequences of linguistic differences are production errors which
result in pronunciation patterns that diverge from those found in the
target language. Most attempts at classification of pronunciation errors
have emphasized phonemic contrasts (e.g., Weinreich 1953/1968; Lado
1957). However, the evidence of phonetic transfer discussed earlier sug¬
gests that an adequate classificatory scheme must take into account other
factors. An error taxonomy devised by Moulton (1962a) takes into
account much, though not all, of the complexity found in second lan¬
guage pronunciation. Although many of the assumptions about pho¬
nological theory and transfer that are implicit in Moulton’s classification
have been challenged, his taxonomy still provides a valuable analysis of
the range of second language segmental errors (i.e., errors involving
vowels and consonants). Based mainly on a contrastive analysis of En-
1 16 Language transfer

glish and German, Moulton’s taxonomy recognizes four types of errors:

(1) phonemic errors; (2) phonetic errors; (3) allophonic errors; and (4)
distributional errors.
Phonemic errors can arise when the phonemic inventories of two
languages differ. For example, German has a phonemic contrast between
the voiceless velar fricative /x/ and the voiceless velar stop /k/. Thus,
German has minimal pairs such as /naxt/ (“night”) and /nakt/ (“naked”).
While the latter consonant has phonemic status in English, the former
does not (except in some dialects in the British Isles). In fact, many native
speakers of English have difficulty pronouncing /x/ and often fail to
distinguish minimal pairs such as /naxt/ and /nakt/. An analogous prob¬
lem in ESL is the /r/—/1/ distinction in English discussed earlier, which
results in frequent pronunciation confusions among speakers of Chinese,
Japanese, and Korean.
Phonetic errors in Moulton’s classification involve cases of cross-
linguistic equivalence at the phonemic but not the phonetic level. Thus,
while the German uvular /r/ and the English retroflex /r/ are correspond¬
ing consonants in cognate forms such as German rar and English rare,
their acoustic properties differ considerably. Moreover, speakers of
American English are normally not accustomed to using the uvula,
whereas speakers in many regions of Germany are.3 The /r/ sounds that
English speakers are likely to produce will thus differ considerably from
the target language consonant.
Allophonic errors can arise in cases of interlingual identifications of
phonemes in two languages. A particular sound or allophone that is a
manifestation of a native language phoneme is not always an accepted
manifestation of a corresponding target language phoneme. For example,
both English and German have a voiceless alveolar stop /t/. When the
American English /t/ occurs between vowels, however, it is not always
voiceless; thus, the sound of /1/ in words such as writer and whiter is
acoustically quite similar to the sound of /d/ in rider and wider. The
German /t/, on the other hand, remains essentially voiceless when it
occurs between vowels. Americans learning German are thus liable to
use a voiced consonant between vowels in words such as bitter
(“bitter”).4
Distributional errors sometimes resemble allophonic errors, but they
may involve combinations of sounds. For example, German has a pho-

3 As discussed earlier, the equation of transfer and “old habits” is misleading (Section
3.1). However, articulatory problems that English learners of German have with
uvulars suggest that a theory of habit formation may be applicable to certain types
of phonetic transfer.
4 Allophonic and other types of errors in Moulton’s classification might also be
described in terms of rules such as those in generative phonology (e.g., Schane
1973).
 Phonetics, phonology, and writing systems 117

neme /ts/ which is acoustically similar to the consonant cluster found

at the end of English words such as its and bits. While speakers of
English have no difficulty in pronouncing the German phoneme when
it occurs at the end of words, as in Sitz (“seat”), they do often have
difficulty in pronouncing it at the beginning of words, as in zu (/tsu/,
“to”). Thus the position of a sound within a word or syllable can affect
how easy a sound is to pronounce. When there are distributional dif¬
ferences in the sounds of two languages, transfer errors may occur.

Suprasegmental patterns
Although cross-linguistic influences on pronunciation frequently involve
segmental contrasts, the influences are also frequently evident in supra¬
segmental contrasts involving stress, tone, rhythm, and other factors.
Stress patterns are crucial in pronunciation since they affect syllables
andTfiFsegments that constitute syllables, as seen in the stress alternation
in English between certain nouns and verbs, such as between COMbine
and comBlNE. The first syllable in these two words has a different
vowel sound, with the sound varying according to the acoustic promi¬
nence of the syllable. Such interactions have important implications not
only for speech production but also for comprehension. Research re¬
viewed by Cutler (1984) indicates that stress patterns play a crucial role
in listeners’ recognition of words. When non-native speakers do not use
a stress pattern that is a norm in the target language, vowels and con¬
sonants may also vary from the target pattern, and this can result in a
total misperception by listeners. Bansal (1976) argues that errors in stress
are the most important cause of unintelligibility in Indians’ pronunci¬
ation of English, and gives examples of misidentifications by listeners.
For instance, diVisions was sometimes pronounced Divisions and was
consequently misperceived by British listeners as REgions, and talking
among themSELVES was sometimes pronounced as talking among
THEMselves and was consequently misperceived as talking among
DAMsels.
Stress errors such as those noted by Bansal do not necessarily reflect
native language influence. However, a somewhat similar study by Tiffin
(1974) found differences in the intelligibility of Nigerian English that
were related both to stress errors and to native language. According to
Tiffin, Yoruba speakers in the study tended to make more errors in¬
volving stress than Hausa speakers did, and as a result British listeners
had more difficulty in identifying what the Yoruba speakers had said.
The greater intelligibility of the Hausa speakers seems to have been due
mainly to a somewhat greater similarity between the suprasegmental
system of Hausa and that of English. Other evidence of cross-linguistic
influences related to stress is found in a study by Andrews (1984). French
118 Language transfer

speakers, for example, tended to accent syllables at the end or close to

the end of English words; that tendency accords with stress patterns in
French, and it proved to be different from the stress patterns of speakers
of other languages. Such a tendency suggests that cognate forms (e.g.,
motor and moteur} might often be unrecognizable when listeners
(whether native or non-native speakers) do not take into account dif¬
ferences in stress patterns.
One of the most important typological distinctions between languages
involves tone aijcLintonation (cf. Section 3.4). In tone languages, pitch
levels havepiionemic significance. A common example of the phonemic
status of tone in certain languages involves the Mandarin Chinese syl¬
lable ma, which, among other denotations, represents “mother” when
it is used with a high level tone, and “horse” with a low rising tone
(Bloomfield 1933). Such a system uses pitch levels quite differently from
what is encountered in most European languages.
The ease of acquisition of the phonology of tone languages may
depend very much on the typological similarity between the native
and target languages. Research by Gandour and Harshman (1978)
suggests that knowledge of one tone language (e.g., Yoruba) can aid
learners in identifying the significant suprasegmental units in another
tone language (e.g., Thai). Thus, it is likely that speakers of a lan¬
guage such as Yoruba would have less difficulty than would speakers
of English, which is not a tone language, in mastering the supraseg¬
mental system of a language such as Thai. There is evidence that
speakers of English do have considerable difficulty in learning to
identify and use tones in Mandarin (Chiang 1979; Broselow, Hurtig,
and Ringen 1987). However, it also appears to be true that speakers
of tone languages encounter difficulties in learning another closely re¬
lated tone language; Leung (1978) notes that Cantonese speakers
often have difficulties with Mandarin and that such difficulties are
due in part to the tone system of the native language.
While tone languages such as Chinese and Yoruba are found in many
parts of the world, intonational languages such as English are also quite
common. Pitch in English does not signal phonemic distinctions as it
does in Chinese, but it does convey important information about speak¬
ers’ attitudes and emotional states. As such, pitch and other supraseg¬
mental features, including stress, rhythm, and loudness, play a role in
discourse similar to that of gestures and other paralinguistic signals (cf.
Section 4.1). A study by Rintell (1984) suggests that speakers of Chinese
have special difficulty in identifying the emotional states of speakers of
English; in contrast to speakers of Spanish and Arabic, Chinese speakers’
judgments of the emotional tone of conversations often diverged from
the judgments of native speakers of English. While intonation may be
only one of the factors accounting for the difference between Chinese
 Phonetics, phonology, and writing systems 119

and other ESL students in Rintell’s study, it does seem significant that
Spanish and Arabic are, like English, intonational languages, whereas
Chinese is not.
Intonational signals have other functions besides suggesting speakers’
attitudes and emotions; they also help to structure conversation by pro¬
viding signals for openings and closings, for the managing of turns, and
for other functions (Brazil, Coulthard, and Johns 1980). Moreover, in¬
tonation often interacts with discourse and syntactic structures. There
do seem to be some universal tendencies in the functions that supraseg-
mental units will have in phrases and clauses. For example, Bolinger
(1978) observes that a rising intonation is characteristic of yes-no ques¬
tions in many languages (e.g., Are you coming?, which can have either
an affirmative or a negative reply). Nevertheless, there is considerable
cross-linguistic variation in suprasegmental systems, and the effects of
similarities and differences in systems are evident in second language
acquisition.
A similarity in the suprasegmental patterns of two languages can
^give^lMearner importarrradvantages in learning the syntax of the tar-,
get language according to a study by Keller-Cohen (1979). In a com¬
parison of the acquisition of English by children who spoke Japanese,
Finnish, and German, Keller-Cohen found that the similarity between
the question intonation patterns of English and those of German and
Japanese aided speakers of those languages in acquiring the syntax of
questions in English. In contrast, the absence of rising intonation in
yes-no questions in Finnish appears to delay the acquisition of En¬
glish question patterns.
Similarity or dissimilarity between native- and target-language into¬
nation patterns can also affect production in other ways. Adams (1979)
attributes much of the divergence of ESL speakers’ speech rhythms to
the rhythmic systems in their native languages (in her study, Vietnamese,
Cambodian, and other languages). Contrastive studies of intonation in
German, Dutch, English, and other languages also point to native lan¬
guage influence (Piirschel 1975; Willems 1982; Van Els and De Bot
1987); indeed, one of the surest clues to the specific “foreign accent”
of an individual appears to be the ensemble of characteristics of sentence
rhythm and pitch in the native language.
The effects of suprasegmental (or segmental) transfer may often be
relatively unimportant. When speaking English, a German may “sound
German” and a Korean may “sound Korean,” but they may still succeed
in communicating gracefully, fluently, and accurately in most respects
(cf. Chapter 10). Nevertheless, non-native speakers may at times risk
giving offense simply from the use of intonation patterns that signal one
emotional state in the native language and a different one in the target
language (cf. Kasper 1981; Loveday 1982b).
120 Language transfer

73 Pronunciation, language universals, and

typologies
The cross-linguistic frequency of phonemes
Although the native language has a major influence on the phonetic and
phonological patterns evident in second language acquisition, research
in the last decade or so has raised important questions about the in¬
dependence of native language influence from other factors. One factor
that many contrastive analyses have unfortunately not taken into ac¬
count is the cross-linguistic frequency of the sounds being contrasted.
Languages tend to have a_m.ix of sounds, some found in many languages
and some not so commonly found. "While there is no particular vowel
or consonant phoneme found in eVery—language of the; world, some
sounds are extremely common. For instance, in a controlled samplqjaf
317 languages, the vowels /i/, /u/, and /a/ all appeared in the phonemic
inventories of over 250 languages; similarly, the bilabial nasal _/m/ ap¬
peared in almost 300 languages, and the voiced bilabial stop / b/ in almost
200 languages (Maddieson 1984). In contrast, some sounds in Maddie-
son’s survey were somewhat rarer: for example, the German / x / appeared
in 76 languages and the German /ts/ in 46 (cf. Section 7.2). Some sounds
were rarer still. One such example is a voiceless pharyngeal fricative
/h/ in Kurdish that appeared in only 12 other languages in Maddieson’s
survey.
The significance of such cross-linguistic facts for second language
acquisition is thaT there seems to be a rough correlation between the
frequency of a sound and its difficulty for adults learning a second
langn3ge.’ FoFexample, a study by Briere (1968) suggests that American
students will have considerably more trouble in learning the rarest
sounds: /h/ as opposed to /x/, for instance. The significance of such
experimental results for contrastive analysis is great. A comparison of
Kurdish and English, for example, shows that the former language has
both /h/ and /x/, whereas the latter has neither. Any contrastive analysis
is likely to predict that both sounds will cause difficulty for English-
speaking learners of Kurdish, but the facts of cross-linguistic frequency
suggest that /h/ will cause far more difficulty.

Common phonological rules

As with the cross-linguistic frequency of particular sounds, the frequency
of phonological rules can be an important clue to what will be easy or
5 The relative ease that children may have in learning target language sounds is
discussed in the next chapter (Section 8.2).
 Phonetics, phonology, and writing systems 121

difficult in acquisition. One rule that has proved to be especially inter¬

esting involves consonant devoicing; in some languages certain conso¬
nants at the ends of words become voiceless. In German, for instance,
the final consonant of Rad (“wheel”) is pronounced the same as the
final consonant in Rat (“advice”), although the plural form of Rad,
which is Rader, has a voiced consonant (/d/). Linguists often interpret
the voiceless consonant in words such as Rad as evidence of a devoicing
rule by which a voiced consonant becomes voiceless whenever it appears
at the end of words. In the sense that similar rules appear in many other
languages, this German rule is a very “natural” rule, and this “natu¬
ralness” leads to predictions rather different from those of traditional
contrastive analyses (Eckman 1977). Even though English does not have
such a devoicing rule, English speakers have little difficulty, according
to Moulton (1962b), in learning to pronounce Rad and Rat identically
despite the fact that the plural form of Rad (i.e., Rader) has a voiced
consonant. In contrast, German speakers learning English will have,
Moulton claims, considerable difficulty in suppressing the native lan¬
guage devoicing rule, and thus will find it difficult to avoid pronouncing
words such as nod the same way as not.6 The presence of the devoicing
rule in German and its absence in English therefore do not imply that
English learners and German learners will have comparable difficulty in
acquiring the rule, on the one hand, and suppressing the rule on the
other. It is highly likely that cross-linguistic contrasts frequently entail
such asymmetrical patterns of difficulty (cf. Section 7.4).
The relative lack of difficulty of the devoicing rule for English learners
seems to be related to the cross-linguistic frequency of the rule, which
is found in many other languages. There is other evidence of the natu¬
ralness of the rule. Eckman (1981a, 1981b) has documented cases of
speakers of Cantonese and Spanish devoicing word-final stops in English
even though such a rule does not exist in either the native or target
languages. For example, Cantonese speakers may pronounce pig like
pick even though Cantonese does not have a devoicing rule like that
found in German (Eckman 1981a). Thus, in some acquisition contexts

6 Eckman develops his analysis in terms of markedness, a frequently discussed

concept in the study of linguistic typologies and universals. Scholars generally agree
that the more marked a structure is, the more unusual it is, and the more difficult it
will be to learn. Beyond that consensus, however, there is considerable uncertainty
about how markedness is best defined (cf. Moravcsik and Wirth 1986). In the
interest of terminological simplicity, the term markedness has been avoided in this
and other chapters where it is relevant (e.g., the discussion of relative clauses in
Chapter 6).
7 Hecht and Mulford (1987) claim that consonant devoicing occurs in some dialects
of English. While devoicing may occur, the importance of this rule in English is
probably negligible in comparison with its importance in German.
122 Language transfer

the devoicing rule has an existence somewhat independent of both the

native and target languages (cf. Eckman 1984).

Syllable structure
The relative independence of the devoicing rule from the structure of
the native and target languages might suggest that transfer is not a factor
in, for example, Cantonese speakers’ pronunciation of pig and pick.
Nevertheless, native-language phonological rules do appear to interact
with rules such as the devoicing rule. The pronunciation errors of Can¬
tonese speakers described by Eckman (1981a) differ considerably from
those of Japanese speakers in the same study. The Japanese speakers
never devoiced final consonants; thus, words such as pig were not pro¬
nounced like pick but instead often had a vowel added to create a second
syllable as in [pigs]. Eckman attributes such errors to syllable structure
typology - Japanese is one of many languages that allow very few con¬
sonants to occur at the end of a word. The addition of a vowel to words
such as pig therefore seems to be a consequence of a typological pref¬
erence in Japanese for open syllables, syllables that do not end in con¬
sonants, as in the consonant-vowel (CV) sequence in pa. Even though
the vowel-insertion rule posited by Eckman does not exist either in
Japanese or English, it reflects a possible influence of native language
syllable structure (cf. Anderson 1987).
According to Eckman’s analysis, the preference of Japanese speakers
for open syllables arises from the influence of the native language. There
is, however, a possibility that this preference is related to language uni¬
versals. Tarone (1980) compared the English pronunciation errors of
native speakers of Cantonese, Portuguese, and Korean and found that
while many errors could be attributed to language transfer, not all could.
Many of the errors, according to Tarone, suggest a universal preference
for open syllables of the CV type. For example, even though Korean
allows for nasal consonants such as /n/ at the end of words, speakers of
Korean did not always pronounce that consonant at the end of English
words such as then and thus produced words with a CV syllable struc¬
ture. In light of the observation of Hyman (1975) and others that CV
is the most widespread syllable type, the errors documented by Tarone
suggest that speakers of all languages may be predisposed to using CV
syllables in a second language (cf. Greenberg 1983).
If there is such a predisposition, however, it interacts with other fac¬
tors. The Cantonese speakers studied by Eckman (1981a) often produced
syllables with a final consonant (e.g., CVC), and frequent use of CVC
syllables is also evident in data obtained from studying Spanish speakers
(Eckman 1981b). Moreover, Sato (1984) has argued that Vietnamese
speakers prefer CVC syllables to CV. For example, Vietnamese speakers
 Phonetics, phonology, and writing systems 123

appear less likely to pronounce next (/nekst/) as [ne] than as [ne?], where
the final consonant is a glottal stop. Other research on Vietnamese
speakers’ pronunciation, however, suggests that some universal factors
are at work. Benson (1986) was able to use an implicational analysis of
Greenberg (1965) to predict what syllable types Vietnamese speakers
would find most difficult. Greenberg’s analysis indicates, for example,
that languages are more likely to have syllables ending in two voiceless
consonants (e.g., /-ps/ as in tops') than to have syllables ending in two
voiced consonants (e.g., /-bd/ as in rubbed). Benson’s findings very clearly
indicate that Vietnamese speakers did indeed have far more difficulty
pronouncing consonant clusters such as /-bd/.

Transfer and developmental factors

The evidence presented so far in this section has suggested that studies
in language universals and linguistic typology are relevant to research
on second language pronunciation in a number of ways, including the
cross-linguistic frequency of phonemes, the naturalness of phonological
rules, and preferences for certain syllable structure types. In all of these
areas, the evidence points to influences on second language acquisition
that are not due solely to native language patterns of pronunciation. As
in the case of syntactic patterns such as negation, the acquisition of
sound patterns appears to involve developmental factors both in first
and in second language learning (cf. Section 6.3). A very probable ex¬
ample of a developmental influence is the consonant devoicing rule in¬
vestigated by Eckman and others. As noted, the devoicing of final
consonants at the end of words can occur in the ESL pronunciation not
only of speakers of languages in which the rule operates (e.g., German)
but also of speakers of languages in which the rule does not operate
(e.g., Cantonese). Some evidence from child language acquisition also
suggests that consonant devoicing is a developmental rule. Edwards
(1979) found that children acquiring English as their native language
tend to devoice word-final consonants such as the /z/ in shoes.
The relative importance of transfer and of developmental processes
such as consonant devoicing has been the object of some investigations.
In a study of a child from Iceland in the United States, Hecht and Mulford
(1987) have provided interesting evidence that transfer and develop¬
mental factors may affect learners’ pronunciation in different ways. They
found that developmental errors were especially common with fricatives
(e.g., /v/), whereas transfer errors were more common with other types
of consonants and with vowels. Major (1986, 1987) has hypothesized
that transfer errors will be more evident in the earlier stages of acquisition
and that developmental errors will not be common until learners have
made considerable progress; his hypothesis is similar to others that have
124 Language transfer

been made about syntactic errors (Section 8.1). While such analyses are
provocative and are supported by some evidence, the relation between
transfer and developmental factors is probably as complex in phonology
as it is in other areas.

7.4 Writing systems

In contrast to other systemic aspects of language, such as phonology
and syntax, writing systems are unique. Since writing systems do not
exist in every language, cross-linguistic comparisons of such systems will
not always be possible. Furthermore, in cases where writing systems do
exist in both the native and target language, contrastive analyses of those
systems will not be applicable to the performance of learners who never
learn the target language writing system or who have not learned the
system of their own language (cf. Section 8.1). However, many acqui-
sition contexts do involve cases of learners who are literate in one lan¬
guage and who seek to become literate in another. In this section, the
focus will be on such cases since they are the only ones eligible for
contrastive analyses of writing systems and since transfer involving such
systems can play an important role in acquisition.

Contrasts in writing systems

Any consideration of transfer involving writing systems must take into
account the relation that frequently exists between pronunciation and
writing. As suggested at the beginning of the chapter, writing systems
often reflect the sound patterns in a language, and therefore a contrastive
analysis of writing systems often presupposes some familiarity with the
phonetics and phonology of the languages being compared. Much of
the negative transfer evident in misspellings has its origins not in native
language orthography but in native language pronunciation (as discussed
later). Thus, even in cases in which learners do not read or write any
language except the target language there can be native language influ¬
ences on second language literacy.
There is no question that literacy involves much more than just the
ability to encode and decode the symbols used in a writing system.
Coming to terms with the meanings intertwined in rhetorical, lexical,
and syntactic patterns surely constitutes much of the challenge of learn¬
ing to read and write in a new language. However, successful reading
and writing also presuppose a certain mastery of encoding and decoding
skills. Such skills involve not only individual symbols but also systems
of symbols. To become literate in an alphabetic system, one must come
 Phonetics, phonology, and writing systems 125

to recognize the correspondence (however rough) between letters and

phonemes. To become literate in a syllabic system such as the Vai script
used in West Africa, one must recognize correspondences between writ¬
ten symbols and syllables (Scribner and Cole 1981). And to become
literate in a so-called ideographic system such as Chinese, one must
recognize the correspondences between written symbols and morphemes
(Coulmas 1983). The most widespread orthographies function exclu¬
sively on an alphabetic, or on a syllabic, or on an ideographic principle,
although some languages, such as Japanese, make important use of more
than one principle (cf. Gelb 1963; Sampson 1985).
Pedagogical practice reflects the fact that the more similar the writing
systems of two languages are, the less time learners will need to develop
basic encoding and decoding skills. Spanish and English, for example,
are extremely similar in their writing systems, whereas Chinese and
English are not. Textbooks that introduce Spanish to English speakers
—
therefore need not and generally do not - devote much space to the
writing system of the target language, whereas introductory texts on
written Chinese for English speakers obviously must devote a great deal
of space to the ideographic system. The Spanish writing system does
show some differences from English in spelling and handwriting con¬
ventions, but what differences there are are differences between alpha¬
betic systems.8 While such differences can occasion spelling difficulties,
the similarities that arise from two languages having essentially the same
alphabet are so great as to reduce considerably the time needed to become
literate in the target language (cf. Section 3.3).
The overlap between alphabetic systems is not always as great as in
the case of Spanish and English. The Cyrillic alphabet of Russian, for
example, has some letters in common with English (e.g., the printed
uppercase letters M and T), but many of the letters will not be familiar
to an English reader even when they represent phonemes that are func¬
tionally equivalent in Russian and English. Empirical work on readers’
processing of the Cyrillic and Roman alphabets used in Yugoslavia sug¬
gests that when students learn an alphabet having some correspondences
with the one they have already mastered, they make interlingual iden¬
tifications of familiar letters and thus begin their mastery of the new
alphabet on the basis of the similarities of the two writing systems
(Lukatela et al. 1978). With some alphabets there is little or no overlap
in the writing systems of two languages except for the alphabetic prin¬
ciple itself. For example, the Persian and the English alphabet have

8 Vachek (1964) has argued that there are important reasons to distinguish between
written language and printed language. Such a distinction might have implications
for studies of transfer, but there seems to be little research on transfer and
handwriting or on related issues.
126 Language transfer

virtually no letters in common and they use opposite directional prin¬

ciples, that is, the Persian alphabet is written from right to left and the
English alphabet is written from left to right. Since the two languages
only share the alphabetic principle, there is little if any positive transfer
aiding the acquisition of English by Persian speakers or the acquisition
of Persian by English speakers, though there is probably some advantage
arising from having already learned to encode and decode written lan¬
guage (cf. Section 8.1).9

Spelling problems
A similarity in writing systems doubtlessly can reduce the amount of
time needed to learn to encode and decode written symbols in a second
language. Yet, as in other cases of linguistic similarity, there may arise
difficulties due to partial but not complete overlap in writing conven¬
tions. For example, Oller and Ziahosseiny (1970) cite instances of mis¬
spellings that are clearly due to the cognate status of words in English,
which was the target language in their study, and words in the native
language of students. Thus, speakers of Spanish, Italian, and Portuguese
all spelled comfort as confort (the cognate form in all three languages
uses n instead of m).
Even in words that are not cognates there may be spelling errors due
to the influence of spelling conventions in the native language, as in the
case of traied, the spelling of tried used by a Spanish speaker (Oller and
Ziahosseiny 1970). Another important influence besides spelling con¬
ventions can be pronunciation. For example, English makes a phonemic
distinction between /p/ and /b/, but Arabic does not, and accordingly
Ibrahim (1978) attributes the following ESL spelling errors of Jordanian
— —
students blaying, bicture, and bombous to phonological influence
from Arabic.10
As with other systemic errors, not all spelling problems can be at¬
tributed to native language influence. Although the categorizations differ,
the results of a number of investigations make clear that sources such
as overgeneralization also account for many errors (Kamratowski and
Schneider 1969; Oller and Ziahosseiny 1970; Ibrahim 1978; Bebout
1985). In the case of languages such as English, which has an orthog-

9 In a description of the reading skills of Persian-English bilinguals, Cowan and

Sarniad (1976) suggest that some of the problems evident in the bilinguals’ skills
are attributable to the very different alphabets used in the two languages.
10 While such evidence suggests that native language pronunciation can have an
effect on spelling, “spelling pronunciations” that reflect imperfections in the target
language orthography can lead to pronunciation errors. Altenberg and Vago
(1983) give some interesting examples of spelling pronunciations interacting with
native language phonology.
 Phonetics, phonology, and writing systems 127

raphy notoriously difficult for native as well as non-native speakers, the

variety of errors reflects in large measure the idiosyncrasies of the system;
thus, the errors made by ESL students are often identical to those made
by native speakers, as fought (instead of taught} and sleaping (instead
of sleeping) (Ibrahim 1978).
The difficulties of the English spelling system provide further evidence
that whatever proves difficult for native speakers can also prove difficult
for non-native speakers. This fact is crucial for understanding the asym¬
metry of linguistic systems, which any thorough contrastive analysis
must take into account (cf. Section 7.3). The differences between the
Spanish and English spelling systems, for example, are not likely to lead
to symmetrical patterns of difficulty. The letter-phoneme correspond¬
ences in Spanish are far more straightforward than the correspondences
in English are, and it seems improbable that spelling in Spanish will
prove as difficult for English speakers as spelling in English will for
Spanish speakers. In the case of Chinese, a language with a writing system
far more intricate than the English system, the consequences for acqui¬
sition are even more evident: most books in English consist of various
combinations of only twenty-six letters, whereas books in Chinese con¬
sist of combinations of hundreds and often thousands of different char¬
acters (cf. Ong 1982). Chinese speakers find the system difficult and
English speakers are likely to find it even more so.

7.5 Summary and conclusion

Native language influence is an important factor in the acquisition of
target language phonetics and phonology. The importance of transfer is
evident in studies of specific pronunciation contrasts and also in research
comparing the overall pronunciation accuracy of speakers of different
native languages. As with syntax and other language subsystems, transfer
is not the only factor affecting the ease or difficulty of reproducing target
language sounds. Typological and apparently universal factors some¬
times operate independently of transfer and sometimes operate together
with it. Thus, the study of developmental patterns is as important in
investigations of second language pronunciation as it is in investigations
of other subsystems. Similarities and differences in writing systems can
result in positive or negative transfer. Learners have tremendous advan¬
tages in learning a language with a writing system similar to that of their
native language. As with pronunciation, transfer is not the only factor
that must be considered in the study of encoding and decoding written
symbols in a second language.
This and the previous three chapters have considered the question of
transfer in specific subsystems. While the evidence in all four chapters
128 Language transfer

argues for the importance of cross-linguistic influences, the problem of

transfer cannot be understood only through reference to system-internal
factors. With regard to the probability (or improbability) of transfer in
particular contexts, who the learners are and what their environment is
can be just as important as the languages they speak. Chapter 8 therefore
considers factors usually not discussed in contrastive analyses.

Further reading
A collection edited by loup and Weinberger (1987) not only has many
important articles on second language pronunciation but also provides a
glossary. Schane (1973) provides a short, readable introduction to modern
phonology.
§ Nonstructural factors
in transfer

In the preceding chapters on semantics, syntax, and pronunciation, the

focus of discussion was on cross-linguistic differences in language struc¬
ture and the consequences that those differences have for acquisition.
Any study of transfer must naturally provide a detailed consideration
of cross-linguistic differences in structure. However, as the discussion
of discourse in Chapter 4 noted, a purely structural analysis of discourse
will not suffice to account for cross-linguistic influences in conversation
and writing. Nonstructural factors relevant to transfer involve not only
discourse hut also pronunciation, vocabulary, and grammar.
A comprehensive review of nonstructural factors in second language
acquisition is beyond the scope of this book, but mention of a couple
should make it clear that structural factors are not the only influence
on acquisition. Motivation surely plays a major role. Even though, for
example, the smaller number of structural differences between Spanish
and English in comparison with Chinese and English can give Spanish
speakers a considerable head start in their acquisition of English, a highly

— —
motivated Chinese speaker will probably learn more English and learn
it faster than will a poorly motivated Spanish speaker. Class size may
also have an important effect. Sixty Spanish-speaking students in an
English class will not likely receive as much individual attention as will
four Chinese speakers who comprise another English class, and so in
aspects of second language acquisition where individual attention is
extremely important (e.g., writing), a Chinese speaker may acquire a
greater mastery of English. Factors such as motivation and class size
have an obvious potential to influence acquisition no matter what native
or target languages are involved. Nonstructural factors may thus operate
independently of transfer. However, there are other factors which are
not structural yet which interact with transfer. Those factors are the
subject of this chapter.
Chapter 8 reviews not only the widest range of nonstructural factors
affecting transfer - those related to individual variation - but also single
factors that affect transfer: the age of the learner, and human awareness
of language, especially as it exists in social contexts. Individual variation,

129
130 Language transfer

age, and language awareness interact with each other in various ways,
and some of these interactions will be discussed.

8.1 Individual variation

Variation is one of the most important characteristics of language. Not
only the existence of different languages, but also the differences seen
within languages in dialects and speech styles provide evidence of the
staggering range of possible variation. Moreover, the potential diversity
extends to individual behavior. No two people speak exactly the same:
differences in voice quality, intonation, and vocabulary choice are among
the most common distinguishers of individual speech, and there are other
distinguishers as well. The existence of such differences poses an im¬
portant problem for the study of transfer. Contrastive analysis, the struc¬
tural basis for predictions of transfer, normally relies on comparisons
of collective, not individual, linguistic behavior (cf. Section 3.2). While
one could compare, for instance, the English of one American with the
Sindhi of one Pakistani, a contrastive analysis of English and Sindhi
would more likely consist of idealized descriptions of both languages.
Such descriptions might or might not allow for regional and social var¬
iation in English and Sindhi, but in either case they still would only be
approximations of the speech of any single American and any single
Pakistani.1
As descriptions of collective behavior, contrastive analyses may fre¬
quently give rise to inaccurate predictions of individual performances
(Sections 2.2, 3.3). Even while some kind of transfer is likely in the
second language performance of most learners, the manifestations of
transfer can vary from one learner to the next. In this section, the fol¬
lowing kinds of individual variation will be considered: (1) personality;
(2) aptitude for phonetic mimicry; (3) linguistic proficiency; and (4)
literacy.

Personality
While individuals can vary in any number of ways, personality differ¬
ences are among the most obvious distinguishers, and some of those
differences probably increase or decrease the likelihood of transfer. There
is, however, a major problem that complicates the study of personality
and transfer: the uncertain status of theories of personality (Littlewood
1983). Personality traits such as aggressiveness and friendliness figure

1 The problem of individual variation is even more complex since, among other
reasons, individual speech changes over time.
 Nonstructural factors in transfer 131

importantly in many theories of personality, but there is no certainty

about how consistent any individual’s personality traits will be in a wide
array of contexts. Despite this and other problems, there are enough
hints in the literature on second language acquisition which point to
traits that may make transfer more likely or less likely.
In her research on avoidance of relative clauses by Chinese and Jap¬
anese students, Schachter (1974) noted the probability that many learn¬
ers experience anxiety about using unfamiliar structures (cf. Section 6.2).
In another study of avoidance, Kleinmann (1977) sought empirical ver¬
ification of the relation between anxiety and cross-linguistic differences.
Some results of Kleinmann’s study suggest that students especially sus¬
ceptible to certain forms of anxiety are more likely to avoid structures
that a contrastive analysis would predict to be difficult.
Personality factors may also account for the varying degrees of success
that individuals have in approximating pronunciation patterns in the
target language. Guiora (1972:145) hypothesized that “individual dif¬
ferences in the ability to approximate native-like pronunciation should
reflect individual differences in the flexibility of psychic processes, or
more specifically, in the empathetic capacity.” Considering pronuncia¬
tion to be the most concrete manifestation of a “language ego,” Guiora
and colleagues have sought to show that the more permeable an indi¬
vidual’s “ego boundary” is (i.e., the more empathetic an individual’s
disposition is), the more attainable a nativelike pronunciation of foreign
language sounds will be. There is some experimental evidence supporting
the notion of ego permeability and empathy. Individuals given moderate
amounts of alcohol or Valium have been able to improve their pronun¬
ciation of foreign language sounds (Guiora et al. 1972; Guiora et al.
1980). By Guiora’s interpretation, such altered states of consciousness
are analogous to the empathetic state of mind believed to facilitate the
acquisition of nativelike pronunciation (cf. Taylor et al. 1969). If his
interpretation is correct, individuals more susceptible to feeling emo¬
tionally “inside” the target language speech community are more likely
to overcome their foreign accent. Accordingly, there is a probable inverse
relation between individual empathy and transfer: the less an individual
learner can feel emotionally “inside” the target language speech com¬
munity, the more pervasive the influence of native language pronunci¬
ation will be.2
Anxiety and empathy are two personality characteristics that appear
to interact with transfer, and there are probably others as well. Identi¬
fying all such characteristics, however, presupposes a thorough under-

2 As discussed earlier, not every characteristic of a foreign accent can be attributed to

native language influence (Section 7.3). Nevertheless, much evidence does point to
the importance of phonetic and phonological transfer (Sections 7.1, 7.2).
132 Language transfer

standing of relations between societies and individuals. The personality

that any learner seems to have may be as much a reflection of socio-
linguistic norms in the native-language speech community as it is of an
individual self. For example, in a study of introversion and extroversion
in Japanese students, Busch (1982) suggested that the norms for class¬
room behavior in Japanese schools often encourage the development of
introverted behavior. If this and similar observations accurately char¬
acterize the influence of native-language speech communities on second
language behavior, much of an individual’s apparent personality may
say more about a native language community than about the individual.
As Scollon and Scollon (1981) and others have suggested, community
norms may lead to unfortunate stereotypes of individual speakers, but
in many acquisition contexts two mitigating influences may come into
play. First, the norms of a native-language speech community may be
rather flexible and may thus overlap with norms of the target language
community. Second, individuals do not always wish to adhere to the
norms of their native-language speech community and may there¬
fore find it relatively easy to adapt to a new set of norms (cf. Paulston
1978).

Aptitude for phonetic mimicry

In some societies, individuals are believed to have or not have “an
ear” for foreign languages, and research on adult capacities has led
to detailed characterizations of the concept of foreign language apti¬
tude. Several abilities seem to underlie that aptitude. Among the
most important is a “phonetic coding ability,” which Carroll
(1981:105) defines as “an ability to identify distinct sounds, to form
associations between these sounds, and symbols representing them,
and to retain these associations.” A related ability, Carroll notes, is
an aptitude for phonetic mimicry. Individuals vary considerably in
their capacity to mimic sounds in a foreign language (Pike 1959),
and research by Purcell and Suter (1980) indicates that this capacity
is a significant predictor of pronunciation accuracy.
Part of the aptitude for mimicry is probably related to the capacity
for empathy discussed earlier. However, there is no logically compelling
reason to suppose that an aptitude for mimicry is entirely attributable
to empathy. Just as the capacity for rote memorization (which Carroll
considers to be another component of foreign language aptitude) prob¬
ably has little to do with individual personality, so the capacity for
mimicry may not simply reflect a personality trait (such as empathy).
No matter what their personality characteristics are, individuals with
little aptitude for mimicry are likely to show the effects of phonetic and
phonological influence from their native languages.
 Nonstructural factors in transfer 133

Proficiency
The notion of second language proficiency is somewhat controversial;
many problems attend its defiinition and measurement, and there is no
consensus among language testers about what test or combination of
tests would constitute a thoroughly adequate index of proficiency. It is
nevertheless undeniable that learners’ abilities differ vastly, with much
(though not all) of the individual variation reflecting different degrees
of second language skill. Some evidence suggests that there is a relation
between proficiency and transfer.
Taylor (1975) has argued that less proficient learners will rely more
on transfer. Much of his evidence comes from the results of a translation
test given to Spanish-speaking students at an American university. Stu¬
dents placed in less advanced classes were especially likely to produce
translations with errors reflecting Spanish-language influence: for ex¬
ample, Can the director to speak with me now?, which reflects the
probable influence of a verb phrase rule in Spanish. Students placed in
more advanced classes were more likely to produce translations with
errors reflecting overgeneralizations, such as Does Gilbert don't speak
French?, which seems to reflect confusion about English question and
negation patterns more than it does any influence from Spanish (cf.
Sections 2.2, 6.3). Taylor argues that the apparent differences in reliance
on transfer mainly reflect the differences in the knowledge base that less
advanced and more advanced learners have to work with. More ad¬
vanced learners know much more about the target language and can
more often make analogies on the basis of that information. Since less
advanced learners have less such information, they will, by Taylor’s
analysis, tend to draw more on their native language (or some other
source language) for analogies that appear relevant. As Taylor observes,
this interpretation of the effects of previous knowledge is consistent with
claims made about the general nature of human learning (Ausubel 1968).
Taylor’s analysis has a great deal of intuitive appeal, and other evi¬
dence supports his position. For example, the studies of word order
transfer discussed earlier indicate that the transfer of basic word order
is most probable among learners with little proficiency in the target
language (Section 6.1). In phonology, moreover, studies by Major (1986)
and others suggest that transfer is especially prevalent in the earlier stages
of proficiency. Nevertheless, there are reasons to be cautious about claim¬
ing a strong relation between transfer and proficiency level. Most im¬
portant, as a study in error analysis, Taylor’s research speaks only to
negative transfer. When one considers the facilitating effects of some
cross-linguistic similarities, such as cognate vocabulary, the likelihood
seems great that positive transfer will occur at the advanced stages just
as much as at the beginning stages of second language acquisition. An-
134 Language transfer

other problem with taking Taylor’s claim as unequivocally true is that

certain kinds of negative transfer cannot occur until learners have
reached a certain level of proficiency (cf. Kellerman 1984). For example,
the existence of transfer errors made with resumptive pronouns in rel¬
ative clauses presupposes that learners can form relative clauses, and
that capacity often takes a long time to develop (cf. Section 6.2). Since
there is evidence that the use of resumptive pronouns is sometimes af¬
fected by cross-linguistic influence (as in the case of Liberian English),
it is probable that transfer can occur whenever a structure seems new,
and therefore problematic, to a learner. While Taylor’s study is an im¬
portant indicator that proficiency may interact with transfer, the relation
between target language ability and native language influence is likely
to be a complex one.3

Literacy
Literacy can have a major impact on second language acquisition, and
a modest amount of evidence suggests that literacy interacts with trans¬
fer. There is no logically necessary connection between literacy in one
language and successful acquisition of another language; one can read
and write in one language without being able to read and write in another
(cf. Section 7.4). Nevertheless, several studies reviewed by Cummins
(1979) indicate that bilinguals highly literate in one language tend to
find easier the acquisition of many second language skills, including
literacy skills.4 A word of caution is necessary here, however. The com¬
parative success of literate bilinguals does not as clearly indicate the
importance of language transfer in the sense of native language influence
as it indicates the importance of transfer of training. That is, literate
3 Another problem with Taylor’s analysis is that the evidence he uses presupposes a
classification of errors into mutually exclusive categories of transfer and
overgeneralization (along with a few other categories). Even if that classification
correctly suggests the relative importance of transfer and overgeneralization and—
—
that assumes that all errors were properly classified such a taxonomy discounts
the possibility of interactions between transfer and overgeneralization (cf. Section
3.3).
4 Defining the term literacy is problematic. While virtually everyone would agree that
literacy requires certain abilities to read and write, there is some controversy over
just what those abilities are. Part of the problem of studying literacy is that different
cultures have different uses for reading and writing (cf. Scribner and Cole 1981;
Heath 1983). Moreover, there is some controversy over what the differences are - if
—
any between spoken and written language (cf. Biber 1986; Shuman 1986). Any
thorough understanding of discourse transfer will have to account for these and
other problems. Despite such problems, however, some facts about literacy relevant
to transfer are not controversial. One is that literate adults tend to have a wider
range of literacy skills (or, a “higher degree of literacy”) than children do. Another
fact is that people differ in their abilities to read and write certain types of texts. In
light of such facts, it is meaningful to discuss literacy in developmental terms (i.e.,
degrees of literacy) in relation to language transfer.
 Nonstructural factors in transfer 135

bilinguals may have an advantage not just because of their linguistic

skills but also because of problem-solving skills that they may have
acquired in the course of their education. Yet there does appear to be
a specifically linguistic factor in bilingual literacy, as Genesee (1979)
indicates in a discussion of the reading skills of Canadian children who
were literate in three languages: French, English, and Hebrew. Genesee
notes that tests of literacy in the three languages yielded higher corre¬
lations between French and English reading skills than between English
and Hebrew reading skills. As discussed earlier, differences in writing
systems can make the acquisition of a second language more difficult
(Section 7.4). Since Hebrew uses a writing system very different from
those of French and English, it is not surprising that there would be less
of a relation between reading in Hebrew and reading in English. By
virtue of the great similarity in the writing systems of French and English,
reading in the one language would be rather comparable to reading in
the other.
A high degree of literacy in one’s native language can increase the
likelihood of positive transfer in recognizing cognate vocabulary (cf.
Section 5.2). A relation between transfer and native language vocabulary
development is evident in a study by Limper (1932) of the French vo¬
cabulary known by American high school and college students having—
a better knowledge of English vocabulary, the college students were
much more successful in identifying cognate forms. Limpet’s research
thus supports the intuition that individuals unable to read a wide range
of texts in their native language will not be in a good position to take
advantage of cognate forms in a second language (cf. Stendahl 1972).
If an English speaker does not understand ambitious, for example, the
similarity to it of French ambitieux will be of little help in understanding
the French word.
Native language literacy also seems to be a factor in success in learning
to write in a second language. One particular area in which native
language literacy might make an immense difference is the ability to
compose for the needs of a specific audience. Scribner and Cole (1981)
found that literate individuals who generally had experience in writing
letters were able to compose more usable descriptions of the rules for
playing a board game than were nonliterate individuals. While Scribner
and Cole’s research does not address directly the problem of literacy in
a second language, it does suggest that individuals literate in their native
language will have a head start on some second-language writing tasks.
There is also reason to believe that individuals with more developed
native-language literacy skills will perform better in second language
writing. Linnarud (1978) asked native speakers of English to judge ESL
writing by students in Sweden, and she asked native speakers of Swedish
to judge writing by the same students in their native language. She found
136 Language transfer

a considerable consistency in both sets of native speaker judgments: the

most favorably judged essays in English and Swedish were generally
written by the same students, and the least favorably judged essays in
English were written by the students whose Swedish essays were also
judged least favorably.5 Using a rather different approach, Masny and
d’Anglejan (1985) found that the native-language reading skills of
French-speaking students in Canada correlated significantly with their
skills in ESL, including writing.

Other factors
Aside from characteristics such as literacy, other aspects of learners’
backgrounds may prove to be important factors in how much transfer
affects any individual’s acquisition of a second language. Social and ped¬
agogical factors are especially likely to interact with transfer. There is
considerable evidence that the social background of learners can and
does affect their reliance on transfer, and some of that evidence will be
discussed later in this chapter (Section 8.3). Along with social factors,
a number of pedagogical factors may have an important effect. For
example, teachers who know the native language of their students may
provide information about native-target-language contrasts that other
teachers cannot provide. Similarly, textbooks and other materials that
present analogies between the native and target languages may promote
or inhibit some kinds of transfer. Pedagogical practices such as the use
of translation tests may also encourage transfer, as Dulay, Burt, and
Krashen (1982) argue.
Factors such as schooling, literacy, personality, and so forth surely
account for much of the individual variation affecting transfer. Yet even
if all those factors were better understood, there would still be the
possibility that seemingly idiosyncratic characteristics of learners could
affect their reliance on transfer. The differences among individual learn¬
ers described by Fillmore (1979) and others make clear how wide the
range of possible acquisition behaviors is.

8.2 Transfer and age of acquisition

The population of most speech communities shows considerable vari¬
ation in the age of community members, and to that extent chronological

5 Although Linnarud’s sample size is quite small (eight students), her hypothesis
about native and second language skills receives additional support from the fact
that she edited the English papers so that native speaker judgments would be
independent of the grammatical accuracy of the writers.
 Nonstructural factors in transfer 137

age is one more manifestation of individual variation. However, the

inevitability of the aging process makes the relation between language
and age special. It is not inevitable that an adult will have an empathetic
personality, a high aptitude for mimicry, or any skill in reading or writ¬
ing. It is inevitable, however, that an adult will have experienced child¬
hood and the sequence of developmental changes that accompany
growing up. If the common belief is true that “younger is better” for
the acquisition of a second language, there may be a definite relation
between transfer and aging. Younger learners may be more receptive to
transfer, or they might be less receptive. Not all second language re¬
searchers subscribe to the “younger is better” position, but even if that
position is incorrect, the apparent relation between aging and foreign
accents warrants consideration. Moreover, there is other evidence for a
relation between child-adult differences and transfer that also merits
consideration.

Aging and foreign accents

Although non-native pronunciation shows influences besides those of
the native language, the importance of phonetic and phonological trans¬
fer has considerable empirical support (Sections 7.1, 7.2). The relation
between such transfer and the age of learners, however, is not very clear.
In the large body of literature on age and second language acquisition,
a number of studies indicate that younger children are more likely to
achieve nativelike pronunciation (e.g., Asher and Garcia 1969; Seliger,
Krashen, and Ladefoged 1975; Oyama 1976). Such studies thus support
the notion that the older the learner, the more important the influences
of native language pronunciation will be. Other studies have shown very
different results, however. Some indicate that older learners are likely
to have just as good or better pronunciation than younger (e.g., Olson
and Samuels 1973; Snow and Hoefnagel-Hohle 1977; Neufeld 1978);
moreover, age has not always proved to be a significant factor when
compared with related factors, such as the length of residence in an area
in which the target language is spoken (Purcell and Suter 1980).
To a large extent the conflicting views on the relation between age
and pronunciation reflect disagreements about the overall relation be¬
tween age and second language learning. Several studies of the acqui¬
sition of morphology and syntax support the “younger is better” position
(e.g., Krashen, Long, and Scarcella 1979), yet the evidence has not con¬
vinced all scholars (e.g., McLaughlin 1981; Hatch 1983). Some research
suggests that neither the “younger is better” nor the “older is better”
position is entirely correct. In a study of the acquisition of Thai by native
speakers of English, loup and Tansomboon (1987) found that younger
learners were better able to control pitch contrasts in Thai, which is a
138 Language transfer

tone language (Section 7.2). On the other hand, adults showed a greater
ability to master some segmental contrasts. It is entirely possible that
such differences in ability also exist in the acquisition of morphology
and syntax. Moreover, it is possible that differing abilities that have been
observed in adults and children are related to differences in their literacy
skills, their use of the target language, their manner of learning the target
language, their social background, their social attitudes, and many other
factors as well.
While the pronunciation-age issue is part of a larger debate, some as¬
pects of that issue involve questions specific to phonetics and phonology.
One explanation for the nontargetlike sounds so commonly found in
adult second language speech is an alteration in the motor control pro¬
gram governing speech organs (Scovel 1969). With the passing of years,
the argument goes, the program will change and no longer allow the vocal
tract to form sounds that learners can nevertheless perceive. Another (and
possibly related) explanation is that speakers who have learned one lan¬
guage much earlier than another will tend to make interlingual identifica¬
tions resulting in a target language sound being categorized in terms of
phonetic norms of the native language (cf. Section 7.2). Flege (1981) sug¬
gests that the simultaneous acquisition of two languages in childhood
may be the only situation in which such identifications can be avoided; by
establishing early two distinct sets of phonetic norms, young bilingual
children may have an advantage over older children and adults learning a
second language (cf. Flege 1987). Whatever the merits of these and other
explanations, it does seem that pronunciation is more likely than other as¬
pects of linguistic performance to show age-related differences. Studies of
dialect contact in the English-speaking world also suggest that there is
something unique about pronunciation. Trudgill (1986) discusses re¬
search indicating that American adults in Britain and British adults in the
United States rarely lose their accents completely. As in the case of lan¬
guage contact, the reasons for this preservation of native-dialect features
appear to be numerous: perceptual, articulatory, and affective factors are
all probably among those involved.

Other evidence
In the debate on age and second language acquisition, there has been
relatively little attention given to the source of various learner errors,
although in the case of pronunciation, native language influence is prob¬
ably the most important source. Aside from research on pronunciation,
there is evidence that age and native language also interact in the lexical
and syntactic development of second language learners.
Analyzing data from several studies of first and second language ac¬
quisition, Zobl (1983) argues that word order transfer varies with the
 Nonstructural factors in transfer 139

chronological age in which a second language is acquired. Up to about

age three, the second-language word orders used by children are quite
variable, with the variation being governed, according to Zobl, by dis¬
course considerations analogous to those that govern languages with
flexible word order (cf. Section 6.1). For example, a three-year-old child
whose native language was French produced the following sentence:
Elephant plane is this (“This is the plane elephants fly on”). The sentence
reflects, in Zobl’s analysis, an interaction between the relatively flexible
word order of French and the tendency of young children to use word
orders determined by discourse factors. From about age four up to about
age ten, a period of “syntactic conservatism” develops. In that period,
children are more inclined to stick with one word order, which may or
may not be that of the target language (cf. Harley 1984). This decrease
in word order variability is consistent with first-language acquisition
studies showing a preference for consistency in syntactic patterns in
middle childhood. In contrast to children, French-speaking adults and
adolescents do not show as much “conservatism” in their use of word
order, although they seem to avoid employing most French word-order
patterns in their English sentences (Trevise 1986).
In very young bilingual children an extreme sort of language mixing
is quite common. For example, in the speech of Danny, a two-year-old
whose mother spoke English and whose father spoke German, English
and German words and phrases were often mixed:
Mehr books (“More books”)
Guck, alle Auto on the ship (“Look, all auto on the ship”)
With the cars rauf (“With the cars on top”)
(Redlinger and Park 1980:343)
Studies by Redlinger and Park (1980), Vihman (1985), and others in¬
dicate the usual period in which such mixing occurs ranges from about
a year and a half up till about three. At the age of three, mixing begins
to taper off. Although mixing is common in child bilingualism, the extent
of mixing varies considerably. Children in some settings do very little
mixing, yet in other settings mixing may continue into adulthood (cf.
Taeschner 1983; Fantini 1985; Le Page and Tabouret-Keller 1985).
Although such language mixing and code-switching are often equated
in discussions of bilingualism, code-switching is only one possible type
of mixing. Switching is superficially similar to the language mixing of
young bilinguals, since words, phrases, and sentences of two (or more)
languages may appear in juxtaposition, as seen in the following English-
Spanish switch:
Why make Carol SENTARSE ATRAS PA’QUE (“sit in the back so”)
everybody has to move PA’QUE SE SALGA (“for her to get out”)? (Poplack
1980:589)
140 Language transfer

However, the conditions under which adult code-switching and infant

language mixing take place are often quite different. For example, young
children who mix languages may not always be aware of the existence
of two separate languages, but individuals who code-switch do have
such an awareness. Other characteristics of mixing and code-switching
will be discussed in the next section.

8.3 Transfer, linguistic awareness, and

social context

The difference between “knowing” and “knowing about” a language

has long been recognized, and in the general increase of study of second
language acquisition that difference has received considerable attention.
Opinions vary greatly about the role of linguistic awareness (i.e., “know¬
ing about”), especially as to how helpful such awareness is (cf. Krashen
1981; McLaughlin, Rossman, and McLeod 1983). Whatever the exact
nature of the role that linguistic awareness plays, such awareness is a
nonstructural factor that interacts with cross-linguistic influences.6
Linguistic awareness can be either conscious or unconscious. When
awareness is conscious, people frequently give a name to the object of
their awareness. The designation that people give may be imprecise from
the standpoint of the descriptive linguist (cf. Odlin 1986). For example,
many speakers of English mistakenly believe that the pronunciation
represented with an apostrophe in such words as nothin' is a case of
“dropping a g." Yet, while the expression “dropping a g" is misleading
(the pronunciation of nothin’ does not involve any “dropping”), the
expression does show that listeners with no background in phonetics
have some conscious awareness of the difference between two nasal
consonants (i.e., /n/ and /q/). Even if individuals do not know the phrase
“dropping a g” or show any other conscious awareness of that difference,
they may be unconsciously aware that the pronunciation of nothin’ is
not an example of the prestigious “Network Standard” spoken by
newscasters.
However aware people are of various formal units, their awareness
is frequently as much social as it is linguistic. The attention that English
speakers pay to “dropping a g” reflects the fact that in English some
uses of nasal consonants have more social prestige. In second language
acquisition the linguistic awareness that learners show frequently reflects
social considerations. Thus in much of the following discussion of the
6 The term metalinguistic awareness is frequently used to describe individuals’
awareness of language. In the interest of terminological simplicity, it is not used in
this chapter.
 Nonstructural factors in transfer 141

relation between linguistic awareness and transfer, the importance of

social context will loom large.

Multilingualism and learners’ perceptions

Through most of this book, discussions of transfer have focused on
bilingual situations in which influence from the native language has
important consequences for the acquisition of a second language (Section
3.1). This emphasis on bilingual situations reflects the preponderance of
situations described in research and perhaps also the preponderance of
actual cases of language contact. However, individuals learning a new
language may already know two, three, or even more languages/ There¬
fore no study of transfer can neglect the importance of trilingual and
other multilingual situations. One of the most interesting kinds of evi¬
dence obtainable in multilingual settings is that regarding language dis¬
tance. Studies of trilingualism indicate that the more similar linguistic
structures in two languages are, the greater the likelihood of transfer.
However, studies of language awareness indicate that the importance of
language distance depends very much on the perceptions of that distance
by learners.
A study by Vildomec (1963) of writing errors shows the difficulties
that researchers encounter in attempting to detect negative transfer in
multilingual situations. The range of possible explanations for errors is
certainly greater in multilingual than in bilingual situations. Neverthe¬
less, credible evidence of transfer is found in some multilingual situations.
Todd (1983) briefly describes an area of West Africa where speakers’
use of English is often influenced by knowledge of French as a second
language. Todd notes that the sentence At the bottom he is a naughty
somebody probably reflects influences not from any African language
but instead from French, as in the use of the phrase at the bottom, for
which there is a clear parallel in French, au fond. Some less direct evi¬
dence of transfer comes from the results of a grammaticality judgment
test of French sentences given to students in the same part of Africa:
The judgments students gave indicate that English can have a consid¬
erable influence on their acquisition of French (Ahukana, Lund, and
Gentile 1981). While those results also indicate that knowledge of Igbo,
the native language of the students, influenced some judgments, knowl-

7 Vildomec (1963) notes some claims of extraordinary multilingual ability. For

example, biographers of a Cardinal Giuseppe Mezzolini attributed to him a
knowledge of as many as seventy-eight languages and fluency in thirty of them!
While such claims are dubious, as Vildomec suggests, the importance of positive
transfer may increase in rough proportion to the number of languages an individual
learns. There are obvious problems, however, in seeking empirical verification of
such a prediction.
142 Language transfer

edge of English, which was a second language for the students, influenced
many more judgments. Thus, even though all subjects in the study knew
both Igbo and English, the greater typological similarity between French
and English seems to have been an especially strong influence on learner
judgments (cf. Singleton 1987). Further evidence of the importance of
typological similarity appears in a study of reading (Singleton and Little
1984) that indicates that knowledge of a second language can provide
considerable help in the reading of a closely related third language (Sec¬
tion 3.3).
Such research justifies the efforts of contrastive analysts to determine
language distance. However, some evidence suggests that an objective
estimation of language distance can sometimes be misleading about the
likelihood of transfer: in some cases, the subjective estimation of distance
by learners can override an objective measure. In any learner’s attempt
to acquire a new language, language distance is ultimately in the eye of
the beholder. Research indicates that when everything else is equal,
transfer will most likely result from a learner’s judgment (made con¬
sciously or unconsciously) that particular structures in a previously
— —
learned language are quite like if not the same as structures in the
target language.
Studies by Kellerman (1977, 1978) of Dutch and English vocabulary
provide interesting evidence for the importance of judgments about lan¬
guage distance. In light of the evidence on lexical transfer discussed
earlier, it is not surprising that Dutch students learning English (and
English students learning Dutch) will find the acquisition of vocabulary
somewhat easier since the two languages have many cognates, such as
hoe and how, dat and that, and hreken and break (cf. Section 5.2). Yet,

— —
while Dutch students learning English often use and sometimes misuse
such cognate forms, they show some wariness about using certain
forms (cf. Lightbown and Libben 1984). Idioms also seem to incur a
great deal of suspicion: “I have often noted the amazement on our
students’ faces when they do discover the existence of Dutchlike idioms
in English” (Kellerman 1977:102). Kellerman found empirical support
for his observations in the form of acceptability tests. Dutch students
frequently judged as unacceptable idioms that are actually parallel in
the two languages, such as to have the victory in the bag, to lay it on
thick, and dyed-in-the-wool (Kellerman 1977). In a study of various uses
of the verb break, Kellerman (1978) found a systematic preference for
“transparent” uses. Senses of breken that are closer to the “core mean¬
ing” of the verb were seen as more transferable into English than other
forms were (with the core meaning being determined through an ex¬
perimental procedure). For example, high school and university students
frequently accepted She broke his heart as a possible translation from
Dutch into English, but less often accepted Some workers have broken
 Nonstructural factors in transfer 143

the strike, even though there is nothing anomalous about either sentence
or their translation equivalents. The importance of transparency is ev¬
ident in other research. Word association studies by Osgood, May, and
Miron (1975) suggest that form-meaning relations commonly found in
the lexicons of many languages are highly transparent (cf. Section 5.2).
Moreover, certain (possibly universal) relations between syntax and se¬
mantics suggest that transparency is relevant to other domains of lin¬
guistic structure besides the lexicon (e.g., Jordens 1977; Gass and Ard
1984).

Semantic transfer and social context

As the discussion in the preceding section indicates, Kellerman believes
that there are fairly strong constraints on what is transferable. He has
claimed, for example, that “idioms are one class of language items that
are generally not transferred” (Kellerman 1 977:101-02, emphasis in the
original). Yet, while Kellerman has acknowledged that his claims are
based primarily on research on the intuitions of students in Dutch high
schools and universities, other researchers (e.g., Krashen and Terrell
1983) have interpreted his findings much more broadly. Some of these
interpretations are universalist (cf. Section 3.4). That is, some researchers
believe that Kellerman’s work has identified constraints on semantic
transfer in any second language context. In one sense, such interpreta¬
tions are not surprising given claims made by Kellerman (1978, 1983)
about semantic transparency. According to his analysis, transparent id¬
ioms are more “transferable,” and there is little question that the notion
of transparency presupposes universal form-meaning relations (cf.
Cooper and Ross 1975; Friedrich 1979; Kellerman 1983; Slobin 1985).
Although Kellerman’s work has important implications for second
language research, it cannot provide very much support for universalist
analyses. While some expressions that he cites, such as to break some¬
one’s heart and to talk behind someone's back, are undeniably good
examples of transparent expressions, often what should or should not
be considered transparent is unclear. Ease of comprehension would seem
an essential criterion in any definition of transparency, but if that is so,
there are characteristics of semantic transfer that Kellerman’s analysis
may not explain well. The following sentences from Hiberno-English
cited by Henry (1977) suggest one of the problems: Did-ya hear of anna
wan bein' on road? and He was failin’ hack wit’ a while. Both were
used by English speakers in a bilingual region of Ireland and both have
Irish translation equivalents, yet in neither case is the idiomatic sense
likely to be clear to individuals unfamiliar with Hiberno-English.8 An
8 The standard English versions of these idioms given by Henry are: “Have you heard
144 Language transfer

expression in Malaysian English based on a Malay idiom illustrates a

similar problem: to shake legs means “to be idle,” just the opposite of
the meaning of the common English idiom to shake a leg (Platt, Weber,
and Ho 1984). Are both idioms transparent? If so, one has to make the
dubious argument that an idiom can be transparent but also ambiguous.
If the Malay idiom is not transparent, the occurrence of semantic transfer
in to shake legs is a counterexample to Kellerman’s claims.
Although the existence of such idioms may not invalidate Kellerman’s
analysis, it does suggest a need to look closely at the social contexts in
which people speak second languages. Perhaps bilinguals in Ireland and
Southeast Asia countenance a greater amount of semantic transfer in
comparison with Dutch high school and college students. There is no
reason to believe that social constraints on transfer are uniform every¬
where. In fact, there is evidence suggesting that the importance of transfer
in any situation varies largely according to the social context.

Linguistic focusing
A number of distinctions of types of social contexts are relevant to
transfer, such as formal versus informal settings and academic versus
nonacademic environments. Another distinction, however, is especially
useful: focused versus unfocused contexts. Le Page and Tabouret-Keller
(1985:116) see the notion of focusing as applicable to any social context:
[B]y means of individual adjustments in response to feedback, both
“languages” and “groups” may become more highly focused in the sense that
the behavior of members of a group may become more alike.
Linguistic focusing presupposes the existence of linguistic awareness,
since focusing involves: (1) some awareness of belonging to a group; (2)
considerable awareness of linguistic and other norms that distinguish
one group from another; and (3) adherence to and enforcement of such
norms. Focusing most typically occurs in the development of standard
languages. When individuals feel a strong sense of belonging to a group,
they are frequently concerned about preserving the linguistic forms be¬
lieved to characterize the group and such concern often leads to attempts
to standardize usage. In many cases standardization involves minimizing

of anyone going matchmaking?” and “His health was declining for some time.”
One might argue that these idioms perhaps came into Hiberno-English from
another variety of English. However, in investigating these and other Hiberno-
English idioms, 1 have not found any evidence suggesting that another dialect of
English is the source of these sayings (cf. Odlin 1988). One might also argue, of
course, that the two Hiberno-English idioms cited would be understandable in some
discourse contexts. However, little or no context is necessary to understand such
undeniably transparent idioms as to break someone's heart and to talk behind
someone’s back.
 Nonstructural factors in transfer 145

the influence from other languages. For example, there have been recent
campaigns in France to discourage the use of English loanwords such
as cocktail, software, and drugstore, all of which are used in French (cf.
Hagege 1987). Governments and educational authorities in France and
some other nations advocate “pure” forms (usually words with a long
history of use in the language) to minimize borrowing transfer (cf. Section
2.1).
The standard languages promoted by governments and schools are
not, however, the only instances of focused varieties. In the Vaupes
region of South America, multilingualism in various Indian groups is
quite common, and even though schools did not exist in the 1960s and
early 1970s, anthropologists noted a concern among Indians to keep
each language as distinct as possible (Sorenson 1967; Jackson 1974).
Another type of focusing is evident in the efforts of some parents to
discourage their bilingual children from mixing languages. As discussed
above, two-year-old bilinguals frequently mix languages (Section 8.2);
however, there have been cases in which such mixing was minimal (e.g.,
Ronjat 1913; Pavlovitch 1920). What seems to have mattered in these
cases was the avoidance of language mixing by parents (e.g., where
Ronjat spoke to his child exclusively in French while his wife spoke
exclusively in German). Even when parents do not make such efforts,
however, children are likely to reduce the amount of mixing during the
course of acquisition - provided there is some focusing in the language
of adults and older children (cf. Redlinger and Park 1980; Vihman 1985).
Although the language mixing of young children is similar to the
negative transfer seen in the performance of older learners, there are
important differences in terms of the amount of available linguistic
awareness. Older learners will more often be aware of the existence of
two (or more) distinct languages, typically their native language and the
target language. With such an awareness, older learners may try to use
only what they believe to be structures of the target language (e.g.,
Fantini 1985). Negative transfer can occur, however, on either a con¬
scious or unconscious plane. When older learners cannot think of any
other way to express what they want to say, they may have conscious
recourse to forms in their native language (or some other source lan¬
guage), or they may inadvertently resort to native language forms with¬
out any awareness that transfer is occurring. The case cited earlier of
the Swedish-speaking student who used the Swedish word bort (“away”)
in the sentence But sometimes I must go bort could involve either con¬
scious or unconscious transfer (Section 3.3). With very young children,
on the other hand, there may be little awareness that two distinct lan¬
guages are spoken in their environment; if only one linguistic code seems
to exist, nothing can transfer (Taeschner 1983). As their awareness of
the objective existence of two languages grows, children tend to develop
146 Language transfer

a sense of focusing comparable to the focusing practiced by adults.9 As

linguistic development continues, then, a largely unconscious mixture
of languages will give way to acquisition patterns liable to show the
conscious and unconscious forms of negative transfer found with older
learners.
As suggested earlier, code-switching is distinct from the language mix¬
ing of young bilinguals (Section 8.2). It is also different from transfer.
Fluent adult bilinguals frequently code-switch, and numerous studies
concur on some basic characteristics of such switching (e.g., Huerta
1978; Poplack 1980; Gumperz 1982; Bentahila and Davies 1983). First,
such switching is often intentional and can serve a variety of purposes.
For example, two Mexican-Americans speaking in English can remind
each other of their common ethnic background by occasionally switching
into Spanish. Second, code-switching is similar to conscious negative
transfer, but is more augmentative than compensatory. That is, fluent
bilinguals usually know how to say in either language what they want
to convey, and in such cases switching is not a falling-back on the native
language. Third, unconscious transfer seems to be rare in the switches
of highly proficient bilinguals. There may be some mutual influence of
the sound systems, but even the pronunciation patterns tend to conform
to the norms of each of the languages being switched. Fourth, the most
fluent bilinguals seem to have the greatest repertory of switching skills.
For example, Poplack (1980) found a wider variety of syntactic struc¬
tures in the switching patterns of fluent bilinguals. Finally, even though
switching might seem to be completely the opposite of focusing, the
language mixture created in switching can result in a focused variety.
In the Middle Ages and Renaissance, educated people developed a taste
for “macaronic” verse, in which Latin and vernacular languages (e.g.,
French and Italian) were mixed, and such verse became a recognized
genre for many years (Lazzerini 1982). A more complex case of mixing
and focusing is seen in Yiddish, which incorporated elements of German,
Hebrew, and other languages and which eventually developed into a
focused variety acquiring considerable literary prestige (Weinreich
1980).
Focusing seems to discourage most kinds of language mixing, includ¬
ing negative transfer. Unlike positive transfer, which stems from simi¬
larities between languages, any manifestation of negative transfer may
be stigmatized. Classrooms and other environments promoting the use
of focused language will normally encourage learners to develop an
awareness of the differences between their native language and the target
language. Sometimes specific materials and teaching strategies are de-

9 Vihman (1985) contends that the developing awareness of two languages by

bilingual toddlers parallels their growing self-awareness.
 Nonstructural factors in transfer 147

veloped to promote such awareness; for example, lists of faux amis

(“false friends” - erroneous French-English cognates) warn Francophone
ESL students not to consider French vocabulary too similar to English
vocabulary (cf. Section 5.2). Overt instruction may not always prevent
negative transfer, and when the instruction is poorly conceived, it may
even encourage such transfer. However, there is reason to believe that
instruction can often diminish negative transfer. In a detailed analysis
of the English used in Ghana, Sey (1973:6) claims that educated speakers
are
aware, in varying degrees depending on their standard of education of the
dangers of possible interference, and are therefore on guard against this. The
assertion ... that in West Africa there is the belief that one language may be
translated into another word for word is a truism which is applicable only to
those second language learners who lack sufficiently well informed tuition.
There is support for Sey’s contention in a study by Singler (1988) of
Liberian English, in which cross-linguistic influences appear to have an
important effect on relative clause patterns (cf. Section 6.2). Singler was
somewhat skeptical of the importance of educational factors in explain¬
ing relativization patterns. Nevertheless, adult speakers of Mande lan¬
guages tended to use more resumptive pronouns in relative clauses if
they had less than ten years of school, whereas adults who had ten or
more years made much less use of such pronouns.
In unfocused contexts, the constraints on negative transfer may be
weak, as there is generally less concern about heeding target language
norms and less linguistic awareness on the part of learners who often
have little schooling and minimal literacy skills. Detailed evidence for
this possibility appears in a study by Muysken (1984) of Ecuadorean
Spanish. Individuals whose Spanish showed the greatest influence from
Quechua tended to be peasants or unskilled laborers with little or no
schooling; these individuals were at a considerable social distance from
middle-class Spanish speakers and had little economic incentive to speak
standard Spanish.

Transfer and social prestige

In formal situations, highly focused language is quite common. Socio¬
linguists have amply documented how varying language activities affect
the formality of pronunciation (e.g., Labov 1972). For example, the
reading of lists typically induces more self-conscious pronunciation than
other types of reading do: the forms used in list reading often indicate
the pronunciation that people believe is “good,” whether or not they
normally use such forms. Such contextual variations in linguistic form
are as evident in second language acquisition as in other situations. In
148 Language transfer

a study of the pronunciation of certain English consonants by Japanese

speakers, Dickerson (1974) found that the reading of lists led to pro¬
nunciations considerably closer to target language norms than did the
reading of dialogues. The consonants studied were, significantly, among
those that a contrastive analysis of Japanese and English would predict
to be difficult.
Most of the discussion of focusing so far has been with respect to
sociolinguistic norms that define prestige in the target language. There
is evidence, however, that norms in the native language can also affect
transfer. Schmidt (1987) describes Egyptian students’ varying pronun¬
ciations of the English consonants /d/ and /e/ (as in then and thin). While
classical Arabic consonants /9/ and lei are used by educated Egyptians,
these consonants are often pronounced the same as the phonemes /z/
and /s/ in informal contexts. Schmidt found that the variation in Arabic
usage substantially influenced the pronunciation of English /d/ and lei.
Less educated Egyptians, who do not frequently use classical Arabic /d/
and lei, tended to mispronounce English /d/ and lei more frequently.
Moreover, the variations seen in the amount of transfer were predictable
in terms of the formality of the language tasks, which included the
reading of lists and passages, as in the Japanese English research of
Dickerson.
While the evidence in this section generally suggests that negative
transfer will be more frequent in unfocused situations, there are impor¬
tant exceptions to that tendency. If a particular form has great prestige
in the native language, speakers may use it more in focused second-
language situations, no matter whether the form is acceptable in the
target language. Beebe (1980) studied the variations in the pronunciation
of American English /r/ produced by speakers of Thai, a language in
which certain pronunciations of /r/ have considerable social prestige.
Transfer from Thai was more evident when the subjects read lists than
when they engaged in conversation: that is, their pronunciations of /r/
tended to have more Thailike variants, which were quite different from
the English variants, in a more formal context.
The motivation of the Thai speakers in Beebe’s study was probably
similar to that of the Arabic speakers in Schmidt’s: both groups were
probably unaware of cross-linguistic influences on what they considered
to be prestigious forms in English. In other situations, however, non¬
native speakers may be well aware that their speech shows native lan¬
guage influence and they may take pride in that fact. In former colonies
of Britain such as Singapore and India, where English has the status of
an official or quasi-official language, the English spoken and written
often shows the influence of local languages (Braj Kachru 1983; Platt,
Weber, and Ho 1984). The social significance of transfer seems to be
greater in these contexts than it does in countries where English has not
 Nonstructural factors in transfer 149

had a long tradition of widespread use. In a survey of a large number

of university students, Shaw (1981) reports that in comparison with
Thai students, Singaporean and Indian students were much more willing
to consider the way they spoke English as a legitimate variety (cf. Low-
enberg 1986). Similarly, the differences between Hiberno-English and
British English became a source of nationalist pride for some leading
literary figures in Ireland (Garvin 1977). What is usually considered
negative transfer can thus persist sometimes and achieve a favorable
social status.10

The importance of demographic factors

The social context of transfer involves factors besides those directly
related to linguistic awareness and focusing. The focusing or lack of —
—
it in any contact situation depends in part on demographic factors.
For example, the relative number of native and non-native speakers in
a community can greatly influence the type of language that learners
will hear. In the case of the Ecuadorean community described by Muys-
ken, one’s chances of hearing highly variable, Quechua-influenced Span¬
—
ish would be very high even when the Spanish speakers have no
command of Quechua. Similarly, Irish-English bilingualism was quite
widespread in Ireland during the nineteenth century, when many people
began to use English, and at the same time there were few monolingual
speakers of English in certain counties. Thus, in some parts of Ireland,
people learning English probably heard dialects of English showing the
effects of transfer more often than they heard any other dialect (Bliss
1977; de Freine 1977).
Within some classroom settings, transfer can be just as evident as in
Ecuador and Ireland. Seiinker, Swain, and Dumas (1975) noted frequent
instances of native language influence on the French of native speakers
of English who were in a bilingual immersion program in Canada (cf.
Section 6.2). Students were encouraged to speak frequently with each
other, and thus the French that they heard was often different from what
their teacher used." As Seiinker, Swain, and Dumas note, the effects of

10 The more favorable status can be ambiguous, however. The results of Shaw’s
survey indicate that Indians may never come to a complete acceptance of a
uniquely Indian variety of English. Similarly, Barry (1983) believes that Hiberno-
English has changed in the twentieth century, with many of the changes reflecting
the influence of the standard English spoken in Britain.
1 1 Some skeptics about transfer (e.g., Dulay, Burt, and Krashen 1982) seem to
assume that some social contexts, such as bilingual immersion programs, are
somehow “abnormal” since learners have little contact with native speakers of the
target language. Such assumptions about normal and abnormal contexts do not,
however, take into account the fact that in many language-contact situations,
monolingual native speakers may either be totally absent or a minority whose own
150 Language transfer

peer communication resembled the effects of second language acquisition

in language contact situations in which few monolingual native speakers
of the target language could serve as models for learners.

Further reading
Harley (1986) offers a detailed look at the problem of age differences in
second language acquisition. A recent paper by Schmidt (1988) surveys much
of the research on linguistic awareness in second language acquisition.

speech is little different from that of bilinguals. In western Ireland, for example,
children a century ago probably heard similar varieties of Hiberno-English from
monolingual speakers as well as from Irish-English bilinguals (cf. Odlin 1988).
9 Looking back and
looking ahead

The preceding chapters have discussed the evidence for the probability
or improbability of transfer in specific subsystems and specific acquisi¬
tion contexts. It is now appropriate to consider some of the limitations
in transfer research, to review some of the most important tendencies
seen in that research, and to discuss some of the areas in which more
study of transfer would be useful.

9.1 Some caveats

There has been considerable progress in the study of transfer during the
last hundred or so years, especially during the years since World War
II. Yet the controversies that have accompanied this progress make it
clear that the findings of transfer research must be interpreted cautiously.
Viewing transfer as the single most important reality of second language
—
acquisition is clearly risky though no more so than viewing transfer
as a negligible factor in acquisition.
In this book there has been relatively little discussion of the individuals
studied or of the methods used in the research. A brief look at the studies
cited will show considerable variation in the numbers of subjects, in the
backgrounds of the subjects, and in the empirical data, which come from
tape-recorded samples of speech, from student writing, from various
types of tests, and from other sources. Without question, every study
has limitations, and virtually every elicitation technique used in the stud¬
ies has its partisans and its critics (e.g., Tarone 1979; Kohn 1987). In
some of the studies discussed there are problems related to linguistic
descriptions, statistics, small sample sizes, reporting procedures, and so
forth, and improvements in data gathering would be highly desirable.
Yet the likelihood is high that future research will reproduce the results
in many (and perhaps most) of the studies cited. Despite the problems
evident in some research, the importance of transfer is clear in light of
the considerable agreement among conclusions drawn from different
types of studies.
As suggested in Chapter 3, a thorough understanding of transfer de-
151
152 Language transfer

pends a great deal on progress in other areas of linguistics. Opinions

vary as to how much progress there has been in the last thirty or so
years, and there is probably even less of a consensus about how much
progress is likely in another thirty years. Nevertheless, there is confidence
among a good many linguists, including many who study second lan¬
guage acquisition, about the growth of the empirical record. The ac¬
cumulation of better, more detailed information has led many to feel
that they are now asking some of the right questions.

9.2 Some conclusions

Uncertainties in research leave many questions about transfer still un¬
answered (cf. Section 9.3). However, a number of conclusions seem
warranted:
1. Transfer occurs in ALL linguistic subsystems. Much of the skep¬
ticism about transfer has been with regard to cross-linguistic influences
involving morphology and syntax as opposed to influences involving
phonetics, phonology, and lexical semantics. While it does seem to be
true that bound morphology is less susceptible to transfer, there are
known cases of it (Section 5.2). In syntax, cross-linguistic influence is
evident in a number of areas, including word order, relative clauses,
articles, and verb phrases (Sections 2.1, 3.2, 6.1, 6.2). Sometimes this
influence may interact with influences from other subsystems, including
discourse and lexicon. For example, Spanish speakers seem to rely more
than Japanese speakers on the form no, which is a cognate of a Spanish
negator (Section 6.3).
2. Transfer occurs both in informal and formal contexts. The evi¬
dence for transfer comes not only from second language investigations
in school settings but also from naturalistic studies of individuals who
have acquired what they know of a second language without any school¬
ing. Negative transfer may be less likely in focused contexts, those sit¬
uations which foster a considerable awareness of language (cf. Section
8.3). On the other hand, positive transfer is probably more likely in
focused contexts.
3. Transfer occurs among children as well as among adults. Although
there have occasionally been claims that adults are more susceptible to
transfer than children are, evidence from several studies suggests that
transfer is common, though not inevitable, in child second language
acquisition (e.g., Keller-Cohen 1979; Wode 1981; Appel 1984; Lujan,
Minaya, and Sankoff 1984; Hecht and Mulford 1987). It may well be
true that children have certain advantages that help them to minimize
negative transfer in their pronunciation. Nevertheless, the exact relation
between age and transfer is still far from clear (Section 8.2).
 Looking back and looking ahead 153

4. Language distance is a factor that affects transfer. While similarity

between languages can create special problems, such as errors involving
false cognates, similarity often confers important advantages. The greater
similarity in vocabulary between Spanish and English as opposed to that
between Arabic and English, for example, makes the acquisition of En¬
glish vocabulary relatively easy for Spanish-speaking students (Section
5.2). Moreover, language distance is most probably a major determinant
of the amount of time students will need in order to become highly
proficient in a language (Section 3.3).
5. Typological factors can affect the likelihood of transfer. Transfer
is likely to be common when the native language influence involves
typologically common patterns, such as preverbal negation or resumptive
pronouns in relative clauses (Sections 6.2, 6.3). In these cases, transfer
effects may be especially likely to persist. The use of some patterns can
also occur, however, even when the native language does not employ
such patterns.
6. Transfer can sometimes involve unusual structures. Native lan¬
guage influences can involve structures that are not typologically com¬
mon, such as preposition stranding in relative clauses (e.g., Un chalet
—
qu’on va alter a “K cottage that we’re gonna go to”) as well as other
structures (Section 6.2). Transfer of these structures, however, seems
less likely in comparison with cases involving typologically common
structures such as resumptive pronouns (Section 6.2).
7. Nonstructural factors can affect the likelihood of transfer. Some
individual differences, as in linguistic proficiency and literacy, can affect
how often cross-linguistic influences have an impact (Section 8.1). The
linguistic awareness of learners can also increase or decrease the prob¬
ability of transfer (Section 8.3). And in some settings, for example in
contemporary Ecuador, demographic and other social factors can make
transfer especially likely (Section 8.3).
While these findings suggest that transfer is a reality in second language
acquisition, the question remains as to just how important cross-
linguistic influences are. In an assessment that no doubt reflects the views
of many second language researchers, Ellis (1985:40) asserts that while
the learner’s native language is an important determinant of second
language acquisition, it
is not the only determinant, however, and may not be the most important.
But it is theoretically unsound to attempt a precise specification of its
contribution or even try to compare its contribution with that of other
factors.

Given the many uncertainties of second language research, Ellis’s caution

is doubtlessly warranted. However, certain language contact studies sug¬
gest that it is not premature to assess the potential that transfer has for
154 Language transfer

affecting acquisition. Among the relevant cases discussed in this book

are Japanese Pidgin English, Hiberno-English, and the Spanish spoken
in Peru and Ecuador. The evidence from these cases suggests that transfer
has very important effects on acquisition. Moreover, the social condi¬
tions in these examples seem especially favorable for transfer to have
its maximum possible impact (Sections 2.1, 8.3). Yet even in classrooms
or other social situations not especially conducive to cross-linguistic
influence, transfer can have distinct effects on the acquisition patterns
of distinct groups of learners (e.g., Ringbom 1976, 1987).
These social considerations do not, however, tell all of what is inter¬
esting about transfer. There are also psychological issues of considerable
importance. In one respect, transfer is not a phenomenon extremely
different from what occurs in child language acquisition, since cross-
linguistic influence involves the use of old knowledge in new situations,
and no one disputes that children must also use old knowledge in new
situations. The language of four-year-old children normally differs from
that of two-year-olds, but four-year-olds use a great deal of the pho¬
nological and lexical and other kinds of knowledge that they had at the
age of two. If one defined transfer as simply the use of old knowledge
in new situations, it would be tempting to argue that transfer is just as
characteristic of first as of second language acquisition. Yet such a wide
definition of transfer is unsatisfactory, even though similar psychological
mechanisms may well be involved in the ability of second language
learners to use knowledge of their native language and in the ability of
children to use old forms and functions in new contexts with socially
appropriate adjustments. There are fundamental differences in the
knowledge base available to first and second language learners. The
knowledge base in monolingual contexts (including child language ac¬
quisition) is much smaller than the knowledge base available in bilingual
contexts simply because bilinguals can draw on not one but two
languages.
Transfer is relevant to differences between first and second language
acquisition in still other ways (cf. Schachter, in press). For example,
transfer can give adults (and sometimes older children) tremendous ad¬
vantages in achieving a useful knowledge of another language even
though second language acquisition does not often result in the kind of
proficiency that native-speaker children may acquire. For example, a
well-prepared and highly motivated student of English literature can
readily take advantage of the considerable similarities in vocabulary,
syntax, writing systems, and so forth between English and other Ger¬
manic languages to become a competent reader of German literature in
a rather short time. Over the long run, of course, children who grow
up speaking German in a German-speaking country will be better pre¬
pared to understand the linguistic and cultural nuances of a story by
 Looking hack and looking ahead 155

Kafka or a poem by Rilke. Nevertheless, it does seem highly significant

that an adult speaker of English might learn to understand rather simple
texts in German (e.g., the fairy tales collected by the Brothers Grimm)
—
in a year or so much less time than the four or so years needed by
German-speaking children to understand the same texts.

9.3 Some areas for further research

There are many areas in which further work on transfer is desirable.
Among the most important are the following:
1. Social context. If it is true, as suggested in the previous section,
that transfer can be an enormous influence under certain sociolinguistic
conditions, those conditions warrant much more study. Thomason and
Kaufman (1988) have presented a number of cogent examples of where
demography and other social factors seem to affect borrowing and sub¬
stratum transfer just as much as structural factors do. The range of
possible social and linguistic variation is enormous, and a great deal of
research will be necessary to achieve a better understanding of rhe so¬
ciolinguistic dimensions of transfer.
2. Subsystem effects. In virtually every linguistic subsystem (dis¬
course, syntax, phonology, etc.), there is evidence not only of transfer
but also of developmental and other factors. As Wode (1981) and others
have suggested, there may well be general acquisition principles —
—
including principles governing transfer that will affect all subsystems.
Investigations of such principles may help linguists to understand such
problematic aspects of transfer as the apparent infrequency of cross-
linguistic influence on bound morphology (Section 5.2).
3. Longitudinal comparisons. These comparisons could help clarify
relations between transfer and developmental processes (cf. Section 2.2).
For example, it would be useful to have longitudinal studies comparing
speakers of languages using resumptive pronouns (e.g., Persian) and
speakers of languages not using such pronouns (e.g., Swedish). If Zobl
(1980) and others are correct in their analyses of the relation between
transfer and developmental errors, one might accurately predict that
Persian speakers will persist much longer in producing errors like I know
the man that John gave the book to him (Sections 6.2, 6.3). Moreover,
longitudinal studies could establish more definitively that a language
such as Chinese will require a longer time for English speakers to master
than will a language such as Spanish.
4. Subsystem interactions. The effects of native language discourse
on the development of target language lexicon and syntax are among
the areas well worth further investigation. For example, Loveday ( 1 982a)
has suggested that rules of politeness in Japan induce Japanese speakers
156 Language transfer

to try to avoid using negative forms in both their native language and
in English. Since avoidance of English relative clauses seems to be com¬
mon among Japanese learners (Section 6.2), avoidance of some types of
negators might also affect the development of syntax.
5. Bidirectionality. Gass (1986) has compared difficulties encoun¬
tered by Italian speakers learning English with difficulties encountered
by English speakers learning Italian. There is a need for more such work,
which might compare, for example, the difficulties encountered by En¬
glish speakers learning Russian and by Russian speakers learning English.
Such comparisons could provide a better idea of the general structural
principles that affect transfer.
6. Borrowing and substratum transfer. Thomason and Kaufman
(1988) describe important social and structural differences between these
types of transfer. Nevertheless, both involve cross-linguistic influences
and further comparative work might provide insights about the fun¬
damental nature of such influence. For example, a comparison of the
effects of Irish on English and the effects of English on Irish in regions
where the latter language is still spoken might show interesting parallels
(cf. Section 2.1).
7. Acquisition of non-European languages. Most of the research
discussed in this book concerned the acquisition of European languages,
especially English. Hopefully, more studies in the future will focus on
target languages with very different typological properties. Many ques¬
tions will remain unanswered until such research is carried out. Sohn
(1980) has claimed, for example, that Koreans find it easier than English
speakers do to learn to use particles in Japanese signaling discourse
topics. If this claim is true, it might prove highly significant for a better
general understanding of the acquisition of syntax.
8. Comprehension and production. It is possible, but not certain,
that transfer is more important as an influence on comprehension than
on production. To the extent that listening and reading comprehension
are prerequisites for fluent speaking and writing, positive transfer may
play an especially important role in the beginning stages of acquisition
of one language by speakers of another language that happens to be
rather similar.
9. Child bilingualism. There is, as suggested earlier, considerable
evidence for the existence of transfer among children learning a second
language. It appears, however, that such influence is related to the relative
awareness or lack of awareness that children have of the differences
between languages (Sections 8.2, 8.3). Further research on the relation¬
ship between transfer and linguistic development in childhood would
certainly be useful for an understanding of such issues as constraints on
transfer and the role of linguistic awareness in bilingualism.
1Q Implications for teaching

The preceding chapters touched only occasionally on the pedagogical

implications of studies of transfer. In comparison with research twenty
or thirty years ago, second language studies now tend to show more
caution about what conclusions teachers should draw from any partic¬
ular set of findings. This caution reflects a growing awareness of the
complexity of transfer and other topics in second language research.
Accordingly, many researchers would now hesitate in dispensing pre¬
scriptions for the classroom, especially since very little of the advice one
might give has been thoroughly tested. Yet despite the commendable
caution seen more and more in second language studies, some of the
research on transfer does have implications for teaching that merit
discussion.
The most important implication is that cross-linguistic influence has
considerable potential to affect the course of second language acquisition
both inside and outside the classroom (cf. Section 9.2). As Schachter (in
press) observes, there is far too much evidence for anyone who looks
closely at the empirical record to be skeptical about the significance of
transfer. It is true that much uncertainty remains about many issues
related to cross-linguistic influences, and it is undeniably true that re¬
searchers are far from able to predict with full accuracy when transfer
will occur (cf. Section 9.3). However, it is also true that skeptics are far
from able to predict when transfer will never occur.
Transfer is thus an important, though incompletely understood, factor
in acquisition. This chapter discusses what pedagogical concerns the
existence of transfer warrants and what steps teachers can take to un¬
derstand transfer as it occurs in their classrooms.

Attitudes
The research discussed in this book suggests that negative transfer is
—
quite possible - and often probable in the pronunciation, grammar,
and so forth of second language learners. Given the existence of that
influence, there remains the question of just how much of a problem
teachers ought to take such influence to be. Negative transfer should be
157
158 Language transfer

a cause for concern in light of the social significance of foreign accents

(including not only pronunciation but also grammar, vocabulary, etc.);
as noted earlier, there seems to be a widespread, though probably not
universal, belief that language mixing is a kind of linguistic corruption
(Section 2.1). To the extent that second language speech shows such
mixing in the form of a foreign accent, there is a possible negative
reaction on the part of listeners. In fact, some evidence suggests that the
more heavily accented a person’s pronunciation is, the more likely it is
that listeners will have negative reactions. In one case, Ryan, Carranza,
and Moffie (1977) presented native speakers of American English with
tape-recorded examples of English read by Spanish-English bilinguals.
The readers with the strongest accents tended to be judged the least
pleasant and the least friendly (cf. Brennan and Brennan 1981). While
this attitudinal research primarily involves language and society in the
United States, there is a strong likelihood that similar research in many
other countries would yield similar results (cf. Kalmar, Zhong, and Xiao
1988).’
Negative transfer does not always prompt negative attitudes: that is,
a foreign accent will not always provoke distrust or hostility. Even
though a prejudice against speakers with foreign accents exists in the
United States, many individuals with accents have been highly successful
in American society. Despite a noticeable German accent, Henry Kis¬
singer achieved distinction in public affairs, and partly because of a very
noticeable French accent, Charles Boyer became a beloved celebrity in
the United States and other English-speaking countries. Friendliness,
intelligence, and other positive personality traits can often charm people
into cultivating a liking for a certain accent - or at least into ignoring
its “foreignness” (cf. Orth 1982). A look at the history of nearly any
country is likely to show cases of talented foreigners who achieve success
even though their command of the local language differed considerably
from that of native speakers.
Yet the success that non-native speakers may achieve rarely if ever
comes easily. Xenophobia or a dislike of particular ethnic groups can
jeopardize a non-native speaker’s chances to gain respect, make friends,
get an education, or make a living, and a foreign accent can trigger

1 As discussed earlier, not all pronunciation problems can be attributed to native

language influence (Section 7.3), but there is little doubt that if the sound patterns
produced by an individual sound “foreign,” transfer will have been a major
determinant of that individual’s pronunciation. A similar claim might also be made
for grammar and vocabulary problems. Some research has suggested that
grammatical transfer does not contribute to the distinctiveness of a foreign accent
and that grammatical errors cannot aid listeners in identifying foreign accents (loup
1984). However, the conclusions of such research are questionable since
stereotypical representations of accents may involve syntactic features, as noted in
Chapter 1.
 Implications for teaching 159

hostile attitudes. It might go without saying that teachers should do

what they can to eliminate the prejudices in a society. In an ideal society,
one might dream, any use of language that was polite, clear, and well
thought out would achieve the same amount of respect regardless of the
accent of the speaker (cf. Giles 1973). However, the elimination of many
negative attitudes is likely to take a very long time, and in the meanwhile
students in a language classroom have a right to any instruction that
can help them to produce speech or writing that will minimize the threat
of a society’s prejudices.2

Comprehensibility
In some contexts, students may not wish or need to change the way they
speak. For example, many Indian students believe that the varieties of
English spoken in their country are as respectable as varieties of British
or American English (Section 8.3). Whether instructors in India them¬
selves have an Indian, a British, or some other accent, they would do
well to consider carefully how much they insist that students imitate
their own way of speaking (cf. Kachru 1987). If students are planning
to stay in India and use their English primarily with other Indians, there
is little sense in forcing students to adopt a pronunciation or vocabulary
not widely used in India. On the other hand, if students are planning
to study or work in areas where no one - or almost no one - is familiar
with Indian pronunciation or vocabulary choices (e.g., most of the South
in the United States), they could profit a great deal from becoming
familiar with features of non-Indian varieties of English. Comprehensible
language should be the goal no matter what dialect is learned, and the
comprehensibility of people can vary in different contexts (cf. Smith and
Rafiqzad 1979; Smith and Bisazza 1982).
The study of comprehensibility involves complex issues. A great deal
of work in psycholinguistics indicates that many factors besides struc¬
tural ones contribute to the comprehensibility of speech and writing (cf.
Clark and Clark 1977; Foss and Hakes 1978). Among the factors that
can contribute to the comprehensibility of any discourse are the context
in which the speaking (or writing) is produced, the cultural assumptions
that speakers and listeners share, and universals of language and cog¬
nition. Thus, it would be mistaken to assume that a foreign accent will
necessarily cause misunderstandings. Yet even with the contribution of
nonlinguistic factors to comprehensibility, structural factors can make

2 Even if it ever becomes an attainable goal, the total eradication of a foreign accent
in favor of, for example, an American or a British accent is a questionable
pedagogical aim. Sociolinguists have often commented on the risks of outsiders
speaking a language “too well” with native speakers (cf. Loveday 1982b).
160 Language transfer

a big difference in whether an individual is understood or not. For

example, a study by Bansal (1976) of the intelligibility of speakers of
Indian English indicates that pronunciation errors had a crucial effect
on how much British listeners were able to understand (cf. Section 7.2).
While further research is needed on the relation between errors and
comprehensibility, there seems little doubt that a large number of errors
— —
whether due to transfer or to other sources can seriously affect
comprehension.
The discussion of comprehensibility so far has focused on how well
non-native speakers may be understood. However, transfer studies are
also relevant to the question of how well non-native speakers understand
the target language. Research indicates that when two languages share
a large number of similar words, learners will have a considerable head
start in their efforts to read another language (Sections 3.3, 5.2). Con¬
versely, learners speaking a language with few lexical similarities to the
target language will be more likely to experience incomprehension or
miscomprehension. Since vocabulary plays a crucial role in all uses of
language, teachers should take special care to monitor what words in¬
dividual students may find easy or difficult in reading materials or in
spoken language in the classroom.5
The issues of language attitudes and of comprehensibility are by no
means completely separate. Some misunderstandings in cross-cultural
communication seem to involve differences in standards of linguistic
politeness (Section 4.1) and expectations about discourse coherence (Sec¬
tion 4.2). As discussed in Chapter 4, misunderstandings involving po¬
liteness and coherence are especially likely to cause bad feeling, and so
discourse differences are a matter of particular concern. Since languages
directly or indirectly encode some cultural differences, instruction point¬
ing out those differences is also appropriate, as Lado and many others
have recognized (Section 2.2). While it would be naive to think that all
cases of ethnic tensions within a nation or of conflict between nations
are due primarily to linguistic or cultural differences, a fundamental goal
of language instruction should be to minimize any misunderstandings
due to such differences.

3 This point may seem obvious to many teachers, but the following anecdote suggests
that not everyone is aware of the importance of lexical similarities. An American
engineer asked an Algerian technician, “Is it up to you whether or not your family
comes here?” The technician did not understand the question, and the engineer
repeated the question a number of times, with more and more frustration. When
the technician was asked by another person, “Does it depend on you whether or
not your family comes here?”, he understood and was able to answer the question
satisfactorily. The Algerian technician was fluent in French (as well as in Arabic),
and depend and dependre are cognate forms in English and French.
 Implications for teaching 161

Process
One criticism of contrastive analysis has been that it emphasizes product
over process: that is, comparisons of languages focus more on static
forms and functions in two languages than on the way people learn a
second language. Without question, teachers must be concerned not only
with forms and functions but also with the learning process. Although
transfer is only one aspect of that process, it is a crucial one. Any fully
adequate theory of second language acquisition must also be able to
account for the role of language universals in the acquisition process,
and the role of universals will never be fully understood until researchers
can account for the occurrence or nonoccurrence of cross-linguistic in¬
fluence in second language contexts (cf. Sections 3.1, 3.4). To some
extent, the product-process distinction reflects the uneasy alliance of
linguistics and psychology. While their comments are about the general
stance of that alliance, the words of two eminent psycholinguists suggest
well the specific value of contrastive research: “It is highly profitable to
know the product before studying the process by which it [the product]
is arrived at” (Clark and Clark 1977:8).

Materials
One of the most important pedagogical questions concerning the study
of transfer is how specific to each language group any classroom ma¬
terials should be. Fries, Lado, and other contrastive theoreticians be¬
lieved that Spanish-speaking students, for example, needed textbooks
and other materials very different from those needed by Chinese¬
speaking students. While error research and other investigations did
much to undermine the credibility of contrastive approaches, compar¬
isons between structures in the native language of students and in the
target language are still quite common in textbooks in certain countries.4
However popular such materials may be, there is little empirical support
for the effectiveness (or ineffectiveness) of contrastive approaches. Such
approaches may well be effective, as the long use of contrastive materials
suggests (Section 2.2), but their effectiveness is likely to remain unverified
for a considerable period of time largely because of the uncertainty about
the effectiveness of teaching linguistic structure. While most teachers
and researchers would agree that the teaching of structure has some
effect, no consensus exists about its exact nature. Research on the ef-

4 Among other reasons for the continuing use of contrastive presentations is the fact
that such comparisons are difficult to avoid if there is no transparent relation
between a native and target language structure.
162 Language transfer

fectiveness of contrastive materials is certainly feasible, but it would

have to be related to questions such as what methods of teaching are
most effective, what types of learners are most likely to benefit from
particular methods, and what periods in acquisition are especially fa¬
vorable to introduce particular structures to learners (cf. Pica 1984;
Pienemann 1984).
Whatever the merits of contrastive materials in some contexts, it is
clear that such materials are not always feasible. For example, when an
ESL class consists of speakers of Chinese, Persian, Spanish, Tamil, and
Yoruba, there is not likely to be any textbook that contrasts English
—
verb phrases with verb phrases in all of those languages and even if
there were, teachers could not profitably spend the class time necessary
to illuminate so many contrasts. Yet even in such classes, one type of
contrastive information is frequently available: bilingual dictionaries.
Although the comparisons are sometimes restricted to words in the native
and target languages, the most carefully prepared dictionaries often pro¬
vide some comparisons of pronunciation and grammar as well. If the
class size allows it, teachers can help individual students in using any
contrastive information that their dictionaries provide.

Information
Apart from dictionaries and other materials suitable for classroom use,
a great deal of other information about language contrasts is available
and may help teachers to see more clearly some of the problems that
their students encounter. Bibliographies of contrastive studies (e.g., Sa-
javaara and Lehtonen 1981; Dechert, Bruggemeir, and Futterer 1984)
are a useful place to begin looking, and the studies themselves commonly
have many useful references. Some of the contrastive research not only
describes the structures in contrast but also provides quantitative evi¬
dence for their relative ease or difficulty (e.g., Andersen 1977; Flynn
1984). While many second language studies are written primarily for
other researchers, some books and articles are written primarily for
teachers, as in the case of a recent collection of contrastive descriptions
(Swan and Smith 1987).
As with classroom materials, contrastive descriptions have their weak¬
nesses, and teachers who read these descriptions should use them with
a certain wariness. For example, a contrastive sketch by Coe (1987)
suggests that such errors as Do you can swim? result from differences
between the verb system of English and the systems of Spanish and
Catalan. While a contrastive explanation is possible, such an error seems
more likely to involve overgeneralization (cf. Sections 2.2, 8.1). Another
problem that teachers should be alert to is that students may speak their
native language differently from what is seen in a contrastive description.
 Implications for teaching 163

For example, if the native language of students is English, a contrastive

account of pronunciation difficulties of American students may not jibe
with what British teachers of French or German notice in their classes.
Furthermore, it may not be the case that students speak a standard or
even a near-standard variety of the native language depicted in a con¬
trastive description (Section 3.2).
One limitation of contrastive descriptions now available is their in¬
completeness. As noted earlier, even languages such as English have not
been thoroughly described, and for many languages information is quite
scarce (Section 3.2). The lack of information (or its inaccessibility to
teachers in certain parts of the world) should not be an insurmountable
problem to well-prepared teachers, however. It is always possible to
learn the language of students or at least enough about a student’s
language to develop contrastive descriptions. Anthropological linguists
often teach themselves languages that have never been described, and
the techniques they use in their fieldwork are not beyond the reach of
professional language teachers (cf. Gudschinsky 1967).
Becoming familiar with the culture as well as the language of their
students should be a goal for teachers, just as students usually have to
become somewhat familiar with the culture in which the target language
is used. Discourse and vocabulary are areas where such familiarity is no
doubt necessary, yet hopefully a teacher’s curiosity would also extend
to any culture for its own sake, to the history, the religion, the arts, the
literature, the cuisine, and all other aspects of life in any student’s native
land. Few language teachers would dispute the need to treat students as

—
—
individuals, and individual attention is difficult and at times even
impossible without an understanding of a person’s linguistic and cul¬
tural background. Even though language and culture are the common
property of many people in a society, they say a great deal about how
different people can be. As such, language and culture are extremely
important distinguishers of, for example, Greek students and Japanese
students. A strong interest in those distinguishers will, more often than
not, lead to a strong mutual respect between teachers and students.
Glossary

The numbers at the end of each gloss refer to the sections which introduce
the glossed term or which discuss it at greater length.

borrowing transfer: In bilingual contexts, the influences found in the

use of a person’s native language that are due to the subsequent
acquisition of another language. (Section 2.1 )
bound morphemes: Forms that cannot stand alone, as in the case of pre¬
fixes and suffixes in English. Thus, while buy is a free form, the suffix on
buying is not since it must always occur with a free form. (Section 5.2)
bound morphology: The system of bound morphemes in a language.
(Section 6.1)
branching direction: A serial relation between two constituents in a
sentence. In the case of restrictive relative clauses, a right branching
direction results from a relative clause following the noun modified
and a left branching direction from a clause preceding the noun mod¬
ified. (Section 6.2)
caiques: Literal translations from one language into another, including
cases where such translations are unacceptable: for example, a learn¬
er’s literal translation of put the fire out into Spanish as poner el fuego
afuera instead of as extinguir el fuego. (Section 3.3)
code-switching: A systematic interchange of words, phrases, or sentences
taken from two or more languages. (Sections 2.1, 8.2)
comment: Most typically, information that elaborates upon a previously
established topic. (Section 6.1)
contrastive analysis: Systematic comparison of two or more languages.
(Chapter 1)
creole: A more developed form of a pidgin. It is typically learned by
children growing up in a community where use of a pidgin is an
everyday occurrence. (Section 2.1)
cross-sectional studies: Studies in which an individual’s language is ob¬
served at only one interval. Such studies typically compare different
groups of individuals at one time period (cf. longitudinal studies).
(Section 2.2)
165
166 Glossary

developmental errors: Errors that are normal occurrences in the course

of learning either a first or a second language. (Section 2.2)
developmental sequence: A succession of phases in learning to master
new linguistic structures. (Sections 2.2, 6.3)
devoicing: The cessation of the vibrating motion of the vocal cords. For
example, if tab is changed to tap, there is no vibration in the pro¬
duction of the final consonant. (Section 7.3)
error analyses: Investigations that seek to determine the types and causes
of errors (and often the frequency of the various error types). (Section
2.2)
focused: Describes linguistic behavior that tends to show a great deal
of uniformity among different individuals. (Section 8.3)
Foreigner Talk: Adjustments in speech made by native speakers so that
non-native speakers can understand what is said to them. Repetition,
the use of simple words, and a slow rate of speech are among the
most frequent characteristics of Foreigner Talk. (Section 2.1)
genitive: A frequent formal marking, often signaled by morphological
case but sometimes by prepositions, that most typically indicates pos¬
session, as the s in the phrase Mary’s book. (Section 5.1)
glottal stop: A consonant produced by a rapid constriction of the region
near the vocal cords. In American English, glottal stops often follow
the first occurrence of the nasal consonant /n/ in such words as fountain
and mountain. (Section 7.3)
hypercorrection: An inappropriate use of a form due to excessive concern
over the use of another form: for example, Arabic speakers spelling
habit as hapit as a reaction to substitutions of the letter b for the letter
p. (Section 3.3)
implicational; Describes any relation in which the existence of one struc¬
ture implies the existence of another. For example, languages in which
verbs occur at the beginning of sentences usually have prepositions;
thus verb-initial word order implies the use of prepositions. (Sections
3.4, 6.2)
interlingual identification: A judgment made by learners about the iden¬
tity or similarity of structures in two languages. (Section 7.2)
language distance: The relative degree of similarity between two lan¬
guages. (Sections 3.2, 8.3)
language mixing: The merging of forms or functions of two or more
languages. Transfer is only one type of mixing; other types include
code-switching (Sections 2.1, 8.2) and the unfocused mixing found in
the speech of very young bilingual children. (Section 8.3)
left dislocation: A word order rule by which a linguistic constituent
appears to the left of its normal position in a written sentence (in
spoken language the same constituent occurs earlier than it normally
does). For example, the sentence John bought the car shows car in a
 Glossary 167

normal object position while The car, John bought it shows the same
constituent in a left dislocation. Left dislocations frequently signal
new topics or referents accorded special emphasis. (Section 6.1)
longitudinal studies: Studies that compare the language of the same
individual (or individuals) over at least two intervals of time. For
example, one might compare an ESL student’s use of articles after
two and after eight months of instruction (cf. cross-sectional studies).
(Section 2.2)
morphological case: An overt, formal marking of nouns, pronouns, or
other word classes to indicate the syntactic roles of noun phrases,
such as subject and object, or to indicate a semantic case. For example,
the ’s on Mary’s indicates the genitive case relation evident in Mary’s
book. Some languages such as Russian make very great use of mor¬
phological case, while others such as English make very little use, and
other languages such as Thai make virtually no use of it (cf. semantic
case). (Section 5.1)
negative face: Claims that a person has to privacy and autonomy. (Sec¬
tion 4.1)
negative politeness: Strategies that reduce threats to the negative face of
another person. (Section 4.1)
negative transfer: Cross-linguistic influences resulting in errors, over¬
production, underproduction, miscomprehension, and other effects
that constitute a divergence between the behavior of native and non¬
native speakers of a language. (Sections 3.1, 3.3)
negator: Any form used to signal negation. (Section 6.3)
open syllable: Any syllable that ends with a vowel instead of a consonant.
(Section 7.3)
overgeneralizations: Uses of a linguistic rule that go beyond the normal
domain of that rule. For example, the use of the bound morpheme
represented by -s on English nouns frequently signals pluralization,
but mans and mouses are overgeneralizations of the pluralization rule.
(Section 2.2)
phonemic system: The system of phonemes, that is, the smallest sound
units that can distinguish meanings of words. For example, the conso¬
nants /s/ and /z/ are part of the phonemic system of English since there
are contrasts in meaning that depend on such sounds: thus sip and zip
do not mean the same thing. Although Spanish (as well as some other
languages) makes some use of the z sound, there are no meaning con¬
trasts such as between sip and zip, and therefore, while /z/ is part of the
phonemic system of English, it is not part of the phonemic system of
Spanish. (Section 7.2)
pidgin: A new language that develops as a result of language contact be¬
tween speakers of different languages. Pidgins typically develop among
speakers who need to talk about trade, work, and so forth, but who are
168 Glossary

unable to learn the native language(s) of their interlocutors. Pidgins

usually show a great deal of structural simplification and sometimes a
great deal of substratum transfer (cf. creole). (Section 2. 1 )
positive face: A person’s self-image and self-respect. (Section 4.1)
positive politeness: Strategies that reduce threats to the positive face of
another person. (Section 4.1)
positive transfer: Any facilitating effects on acquisition due to the influ¬
ence of cross-linguistic similarities. Unlike negative transfer, positive
transfer results in a convergence of behaviors of native and non-native
speakers of a language. (Section 3.1)
postverbal negation: Any construction in which the negator follows the
verb, as in Ingrid kommt nicht (“Ingrid comes not” = Ingrid isn’t
coming), where nicht is the negator. (Section 6.3)
preverbal negation: Any construction in which the negator precedes the
verb, as in Juan no va (“Juan not goes” = Juan isn’t going), where
no is the negator. (Section 6.3)
proposition: The meaning of statements represented by sentences and
sometimes by other linguistic forms. (Sections 4.2, 5.1)
relativism: In linguistics, the belief that knowing a particular language
can induce one to adopt a particular worldview. (Sections 3.4, 5.1)
relativized position: The syntactic constituent in a relative clause that
may be marked by a relative pronoun. (Section 6.2)
resumptive pronoun: In relative clauses, nonrelative pronouns that help
to identify a referent, as in the case of him in The musician that I saw
him is from China. While resumptive pronouns are not allowed in
relative clauses in standard English, they are normal in relative clauses
in many languages. (Sections 3.2, 6.2)
right dislocation: A word order rule by which a linguistic constituent
appears to the right of its normal position in a written sentence (in
spoken language the same constituent occurs later than it normally
does). For example, the sentence John bought the car shows John in
a normal subject position while He bought the car, John did shows
the same constituent in a right dislocation. Right dislocations often
signal a speaker’s “afterthoughts,” expressed to make the identity of
a referent clear. (Section 6.1)
segmental errors: Pronunciation errors involving individual vowels or
consonants. (Section 7.2)
semantic case: Some role that might be ascribed to a person or thing
denoted by a noun phrase (and sometimes by other grammatical struc¬
tures). For example, in the sentences The sailor broke the window
with a rock and A rock broke the window, the noun rock is in in¬
strumental case even while it is a prepositional object in one sentence
and a subject in the other. (Section 5.1)
simplification: Any reduction resulting in a linguistic structure simpler
than what is considered to be the target language norm. For example,
 Glossary 169

two boy instead of two boys and I very good fellow instead of I’m a
very good fellow are cases of simplification (that may or may not be
a result of cross-linguistic differences). Some linguists consider over¬
generalization to be a type of simplification. (Sections 2.1, 3.3)
speech act: Purposive uses of language, such as requests, apologies,
promises, and so forth. (Section 4.1)
substitutions: Errors due to the substitution of one form (often a form
in the native language) for a form in the target language: for example,
a Swedish speaker’s use of bort instead of away in the sentence I must
go bort. (Section 3.3)
substratum transfer: In bilingual contexts, the influences found in the
use of a second language that are due to the native language. (Sections
2.1, 3.1)
suprasegmental: Describes such characteristics as word stress, rhythm,
and tone. (Sections 3.2, 7.2)
topic: The focus of information in a discourse. According to Givon
(1983b), important topics in a discourse tend to be frequently men¬
tioned and to persist through relatively long stretches of discourse.
Grammatical signals of topics include word order, intonation, and (in
some languages) special “topic markers” (cf. comment). (Sections 4.2,
6.1)
transfer of training: Influences on the production or comprehension of
a second language that are due to the ways learners have been taught
(or to ways learners have taught themselves). (Section 2.2)
tree model: A characterization of language change as divergences from
an ancestral language. The branches on the model represent new lan¬
guages or dialects. Such change is not due to language contact (cf.
wave model). (Section 2.1)
typology: Classification of languages according to structural or other
characteristics. Any such classification may reflect historical relations
between languages, but two languages having a common typological
characteristic are not always historically related. (Section 3.4)
unfocused: Describes linguistic behavior that tends to show relatively
little uniformity among different individuals (cf. focused). (Section 8.3)
wave model: A characterization of language change as a process due to
contact between speakers of different dialects or different languages
(cf. tree model). (Section 2.1)
zero anaphora: The absence of a form under special conditions, usually
when the referent of the form can be guessed through some previous
mention. For example, in George took the money and ran, the verb
ran has a “zero” subject: that is, neither a noun nor a pronoun such
as he appears as subject. Nevertheless, the referent associated with
the verb (i.e., George) can be inferred. (Section 6.1)
References

Adams, Corinne. 1979. English Speech Rhythm and the Foreign Learner. The
Hague: Mouton.
Adiv, Ellen. 1984. Language learning strategies: the relationship between LI
operating principles and language transfer in L2 development. In Second
Languages: A Cross- Linguistic Perspective, ed. by Roger Andersen. Row-
ley, Mass.: Newbury House.
Adjemian, Christian. 1983. The transferability of lexical properties. In Language
Transfer in Language Learning, ed. by Susan Gass and Larry Seiinker.
Rowley, Mass.: Newbury House.
Adjemian, Christian, and Juana Liceras. 1984. Accounting for adult acquisition
of relative clauses: Universal Grammar, LI, and structuring the intake. In
Universals of Second Language Acquisition, ed. by Fred Eckman, Lawrence
Bell, and Diane Nelson. Rowley, Mass.: Newbury House.
Ahukana, Joshua, Nancy Lund, and J. Richard Gentile. 1981. Inter- and intra-
lingual interference effects in learning a third language. Modern Language
Journal 65:281—87.
Albert, Ethel. 1964. “Rhetoric,” “logic,” and “poetics” in Burundi. American
Anthropologist 66(6) part 2, Special Publication: The Ethnography of Com¬
munication, ed. by John Gumperz and Dell Hymes.
Albert, Martin, and Loraine Obler. 1979. The Bilingual Brain. New York:
Academic Press.
Altenberg, Evelyn, and Robert Vago. 1983. Theoretical implications of an error
analysis of second language phonology. Language Learning 33:427—47.
Andersen, Roger. 1977. The impoverished state of cross-sectional morpheme
acquisition/accuracy methodology (Or: The leftovers are more nourishing
than the main course). In Proceedings of the Los Angeles Second Language
Research Forum. Los Angeles: Department of English, University of Cal¬
ifornia at Los Angeles.
1979. The relationship between first language transfer and second language
overgeneralization. In The Acquisition and Use of Spanish and English as
First and Second Languages, ed. by Roger Andersen. Washington, D.C.:
TESOL.
Andersen, Roger, ed. 1983a. Pidginization and Creolization as Language Ac¬
quisition. Rowley, Mass.: Newbury House.
Andersen, Roger. 1983b. Transfer to somewhere. In Language Transfer in Lan¬
guage Learning, ed. by Susan Gass and Larry Seiinker. Rowley, Mass.:
Newbury House.
1984. The one to one principle of interlanguage construction. Language
Learning 34:77—95.

171
172 References
Anderson, Janet. 1987. The markedness differential hypothesis and syllable
structure difficulty. In Interlanguage Phonology, ed. by Georgette loup and
Steven Weinberger. Rowley, Mass.: Newbury House.
Andrews, Geoffrey. 1984. English stress rules in adult second language acqui¬
sition. Unpublished Ph.D. dissertation, Boston University.
Anthony, Edward. 1952—53. The teaching of cognates. Language Learning
4:79-82.
Appel, Rene. 1984. Immigrant Children Learning Dutch. Dordrecht: Foris
Publications.
Appel, Rene, and Pieter Muysken. 1987. Language Contact and Bilingualism.
London: Edward Arnold.
Applegate, Richard. 1975. The language teacher and the rules of speaking.
TESOL Quarterly 9:271-81.
Ard, Josh, and Taco Homburg. 1983. Verification of language transfer. In
Language Transfer in Language Learning, ed. by Susan Gass and Larry
Seiinker. Rowley, Mass.: Newbury House.
Asher, James, and Ramiro Garcia. 1969. The optimal age to learn a foreign
language. Modern Language Journal 38:334—41.
Au, Terry Kit-Fong. 1983. Chinese and English counterfactuals: the Sapir- Whorf
hypothesis revisited. Cognition 15:155—87.
Ausubel, David. 1968. Educational Psychology: A Cognitive View. New York:
Holt, Rinehart, and Winston.
Bailey, Charles J. N. 1973. Variation and Linguistic Theory. Washington, D.C.:
Center for Applied Linguistics.
Bailey, Nathalie, Carolyn Madden, and Stephen Krashen. 1974. Is there a “nat¬
ural sequence” in adult second language learning? Language Learning
24:235-43.
Bamgbose, Ayo. 1982. Standard Nigerian English: issues of identification. In
The Other Tongue, ed. by Braj Kachru. Urbana: University of Illinois Press.
Bammesberger, Alfred. 1983. An Outline of Modern Irish Grammar. Heidel¬
berg: Carl Winter.
Bansal, R. K. 1976. The Intelligibility of Indian English. Hyderabad: Central
Institute of English and Foreign Languages. ERIC Report ED 177849.
Barkowski, Hans, Ulrike Harnisch, and Sigrid Krumm. 1976. Sprachhandlungs-
theorie und “Deutsch fur auslandlische Arbeiter.” Linguistische Berichte
45:42-54.
Barry, Michael. 1983. The English language in Ireland. In English as a World
Language, ed. by Richard Bailey and Manfred Gorlach. Ann Arbor: Uni¬
versity of Michigan Press.
Bartelt, H. Guillermo. 1982. Apachean English metaphors. In Essays on Native
American English, ed. by H. Guillermo Bartelt, Susan Penfield Jasper, and
Bates Hoffer. San Antonio: Trinity University.
1983. Transfer and variability of rhetorical redundancy in Apachean English
interlanguage. In Language Transfer in Language Learning, ed. by Susan
Gass and Larry Seiinker. Rowley, Mass.: Newbury House.
Bartlett, Frederic. 1954. Remembering: A Study in Experimental and Social
Psychology. Cambridge: Cambridge University Press.
Bates, Elizabeth. 1976. Language and Context: The Acquisition of Pragmatics.
New York: Academic Press.
Bates, Elizabeth, and Brian MacWhinney. 1981. Second-language acquisition
 References 173

from a functionalist perspective. In Annals of the New York Academy

of Sciences, ed. by Harris Winitz. New York: New York Academy of
Sciences.
1982. Functionalist approaches to grammar. In Language Acquisition: The
State of the Art, ed. by Eric Wanner and Lila Gleitman. Cambridge: Cam¬
bridge University Press.
Bebout, Linda. 1985. An error analysis of misspellings made by learners of
English as a first and as a second language. Journal of Psycholinguistic
Research 14:569—93.
Beebe, Leslie. 1980. Sociolinguistic variation and style shifting in second lan¬
guage acquisition. Language Learning 30:337—51.
Beeman, William. 1976. Status, style, and strategy in Iranian interaction. An¬
thropological Linguistics 18:305—22.
Benson, Bronwen. 1986. The markedness differential hypothesis: implications
for Vietnamese speakers of English. In Markedness, by Fred Eckman, Edith
Moravcsik, and Jessica Wirth. New York: Plenum Press.
Bentahila, Abdelali, and Eirlys Davies. 1983. The syntax of Arabic-French code¬
switching. Lingua 59:301—30.
Berlin, Brent, and Paul Kay. 1969. Basic Color Terms. Berkeley: University of
California Press.
Biber, Douglas. 1986. Spoken and written textual dimensions of English. Lan¬
guage 62:384-414.
Bickerton, Derek. 1981. Roots of Language. Ann Arbor: Karoma.
Bickerton, Derek, and Talmy Givon. 1976. Pidginization and syntactic change:
from SXV and VSX to SVX. In Papers from the Parasession on Diachronic
Syntax, ed. by Sanford Stever, Carol Walker, and Salikoko Mufwene. Chi¬
cago: Chicago Linguistic Society.
Birdsong, David, Catherine Johnson, and John McMinn. 1984. Universals versus
transfer revisited. Paper presented at the 9th Boston University Language
Development Conference.
Birdsong, David, and Terence Odlin. 1983. If Whorf was on the right track: a
review essay of “The Linguistic Shaping of Thought: A Study in the Impact
of Language and Thinking in China and the West” by Alfred H. Bloom.
Language Learning 33:401—12.
Bley-Vroman, Robert. 1983. The comparative fallacy in interlanguage studies:
the case of systematicity. Language Learning 33:1—17.
Bliss, Alan. 1977. The emergence of modern English dialects in Ireland. In The
English Language in Ireland, ed. by Diarmaid O’Muirithe. Dublin: Mercier
Press.
1978. Spoken English in Ireland, 1600—1740. Dublin: Dolmen Press.
1984. English in the south of Ireland. In Language in the British Isles, by
Peter Trudgill. Cambridge: Cambridge University Press.
Bloom, Alfred. 1981. The Linguistic Shaping of Thought: A Study in the Impact
of Language and Thinking in China and the West. Hillsdale, N.J.: Lawrence
Erlbaum.
—
1984. Caution the words you use may affect what you say: a response to
Au. Cognition 17:275—87.
Bloomfield, Leonard. 1933. Language. New York: Holt, Rinehart, and Winston.
Blum, Shoshana, and E. A. Levenston. 1978. Universals of lexical simplification.
Language Learning 28:399—415.
174 References
1980. Lexical simplification in second language acquisition. Studies in Second
Language Acquisition 2:43-63.
Blum-Kulka, Shoshana. 1982. Learning to say what you mean: a study of speech
act performance of learners of Hebrew as a second language. Applied
Linguistics 3:29—59.
Bock, J. Kathryn. 1982. Toward a cognitive psychology of syntax: information
processing contributions to sentence formulation. Psychological Review
89:1-47.
Bolinger, Dwight. 1975. Aspects of Language. New York: Harcourt, Brace, and
Jovanovich.
1978. Intonation across languages. In Universals of Language, ed. by Joseph
Greenberg. Stanford: Stanford University Press.
Bolinger, Dwight, and Donald Sears. 1981. Aspects of Language. New York:
Harcourt, Brace, and Jovanovich.
Borden, Gloria, Adele Gerber, and Gary Milsark. 1983. Production and per¬
ception of the /r/—/1/ contrast in Korean adults learning English. Language
Learning 33:499—526.
Borkin, Ann, and Susan Reinhart. 1978. Excuse me and I’m sorry. TESOL
Quarterly 12:57—69.
Brazil, David, Malcolm Coulthard, and Catherine Johns. 1980. Discourse In¬
tonation and Language Teaching. London: Longman.
Brennan, Eileen, and John Brennan. 1981. Accent scaling and language attitudes:
reactions to Mexican American English speech. Language and Speech
24:207-21.
Briere, Eugene. 1968. A Psycholinguistic Study of Phonological Interference.
The Hague: Mouton.
Broselow, Ellen. 1983. Nonobvious transfer: on predicting epenthesis errors.
In Language Transfer in Language Learning, ed. by Susan Gass and Larry
Seiinker. Rowley, Mass.: Newbury House.
Broselow, Ellen, Richard Hurtig, and Catherine Ringen. 1987. The perception
of second language prosody. In Interlanguage Phonology, ed. by Georgette
loup and Steven Weinberger. Rowley, Mass.: Newbury House.
Brown, Gillian, and George Yule, 1983. Discourse Analysis. Cambridge: Cam¬
bridge University Press.
Brown, Penelope, and Stephen Levinson. 1978. Universals in linguistic usage:
politeness phenomena. In Questions and Politeness, ed. by Esther N.
Goody. Cambridge Papers in Social Anthropology. Cambridge: Cambridge
University Press.
Brown, Roger. 1973. A First Language. Cambridge, Mass.: Harvard University
Press.
Brown, Roger, and Albert Gilman. 1960. The pronouns of power and solidarity.
In Style in Language, ed. by Thomas Sebeok. Cambridge, Mass.: MIT
Press.
Busch, Deborah. 1982. Introversion-extraversion and the EFL proficiency of
Japanese students. Language Learning 32:109-32.
Butterworth, Guy, and Evelyn Hatch. 1978. A Spanish-speaking adolescent’s
acquisition of English. In Second Language Acquisition: A Book of Read¬
ings, ed. by Evelyn Hatch. Rowley, Mass.: Newbury House.
Bynon, Theodora. 1977. Historical Linguistics. Cambridge: Cambridge Uni¬
versity Press.
 References 175
Callary, Robert. 1975. Phonological change and the development of an urban
dialect in Illinois. Language in Society 4:155—69.
Cancino, Herlinda, Ellen Rosansky, and John Schumann. 1978. The acquisition
of English negatives and interrogatives by native Spanish speakers. In Sec¬
ond Language Acquisition: A Book of Readings, ed. by Evelyn Hatch.
Rowley, Mass.: Newbury House.
Carrell, Patricia. 1982. Cohesion is not coherence. TESOL Quarterly 16:479—
88.
Carrell, Patricia, and Beverly Konneker. 1981. Politeness: comparing native and
nonnative judgments. Language Learning 31:17—30.
Carroll, John. 1968. Contrastive linguistics and interference theory. In Report
of the Nineteenth Annual Round Table Meeting on Linguistics and Lan¬
guage Studies, ed. by James Alatis. Monograph Series on Languages and
Linguistics, Georgetown University. Washington, D.C.: Georgetown Uni¬
versity Press.
1981. Twenty-five years of research on foreign language aptitude. In Indi¬
vidual Differences and Universals in Language Learning Aptitude, ed. by
Karl Diller. Rowley, Mass.: Newbury House.
Celce-Murcia, Marianne, and Diane Larsen-Freeman. 1983. The Grammar
Book. Rowley, Mass.: Newbury House.
Chafe, Wallace. 1970. Meaning and the Structure of Language. Chicago: Uni¬
versity of Chicago Press.
Chambers, J. K., and Peter Trudgill. 1980. Dialectology. Cambridge: Cambridge
University Press.
Chaudron, Craig. 1983. Research on metalinguistic judgments: a review of
theory, methods, and results. Language Learning 33:343—77.
Chiang, Thomas. 1979. Some interferences of English intonations with Chinese
tones. IRAL: International Review of Applied Linguistics 17:245—50.
Chomsky, Noam. 1957. Syntactic Structures. The Hague: Mouton.
1965. Aspects of the Theory of Syntax. Cambridge, Mass.: MIT Press.
1981. Lectures on Government and Binding. Dordrecht: Foris Publications.
Clahsen, Harald. 1982. Spracherwerb in der Kindheit. Tubingen: Gunter Narr.
1984. The acquisition of German word order. In Second Languages: A Cross-
Linguistic Perspective, ed. by Roger Andersen. Rowley, Mass.: Newbury
House.
Clark, Eve. 1973. What’s in a word: on the child’s acquisition of semantics. In
Cognitive Development and the Acquisition of Language, ed. by Timothy
Moore. New York: Academic Press.
1982. The young word-maker: a case study of innovation in the child’s lex¬
icon. In Language Acquisition: The State of the Art, ed. by Eric Wanner
and Lila Gleitman. Cambridge: Cambridge University Press.
Clark, Eve, and Herbert Clark. 1979. When nouns surface as verbs. Language
55:767-811.
Clark, Herbert, and Eve Clark. 1977. Psychology and Language. New York:
Harcourt, Brace, and Jovanovich.
Clarke, Mark, Ann Losoff, Margaret McCracken, and Jo Ann Still. 1981. Gender
perception in Arabic and English. Language Learning 31:159-69.
Coe, Norman. 1987. Speakers of Spanish and Catalan. In Learner English: A
Teacher’s Guide to Interference and Other Problems, ed. by Michael Swan
and Bernard Smith. Cambridge: Cambridge University Press.
176 References
Cohen, Andrew, and Elite Olshtain. 1981. Developing a measure of sociocultural
competence: the case of apology. Language Learning 31:113-34.
Comrie, Bernard. 1981. Language Universals and Linguistic Typology. Chicago:
University of Chicago Press.
1984. Why linguists need language acquirers. In Universals in Second
Language Acquisition, ed. by William Rutherford. Amsterdam: John
Benjamins.
Comrie, Bernard, and Edward Keenan. 1979. Noun phrase accessibility revis¬
ited. Language 55:649—64.
Connor, Ulla, and Peter McCagg. 1983. Cross-cultural differences and perceived
quality in written paraphrases of English expository prose. Applied Lin¬
guistics 4:259—68.
1987. A contrastive study of English expository prose. In Writing Across
Languages: Analysis of L2 Text, ed. by Ulla Connor and Robert Kaplan.
Reading, Mass.: Addison-Wesley.
Cooper, William, and John Ross. 1 975. World order. In Papers from the Parases¬
sion on Functionalism. Chicago: Chicago Linguistic Society.
Corbett, Edward. 1971. Classical Rhetoric for the Modern Student. New York:
Oxford University Press.
Corder, S. Pit. 1983. A role for the mother tongue. In Language Transfer in
Language Learning, ed. by Susan Gass and Larry Seiinker. Rowley, Mass.:
Newbury House.
Coulmas, Florian, ed. 1981a. A Festschrift for Native Speaker. The Hague:
Mouton.
Coulmas, Florian. 1981b. “Poison to your soul”: thanks and apologies con-
trastively viewed. In Conversation Routine, ed. by Florian Coulmas. The
Hague: Mouton.
1983. Writing and literacy in China. In Writing in Focus, ed. by Florian
Coulmas and Konrad Ehlich. The Hague: Mouton.
Cowan, J. Ronayne, and Zohreh Sarmad. 1976. Reading performance of bilin¬
gual children according to type of school and home language. Language
Learning 26:353—76.
Crystal, David, and Derek Davy. 1969. Investigating English Style. Blooming¬
ton: Indiana University Press.
Cummins, James. 1979. Linguistic interdependence and the educational devel¬
opment of bilingual children. Review of Educational Research 49:222—51.
Cutler, Anne. 1984. Stress and accent in language production and understand¬
ing. In Intonation, Accent, and Rhythm, ed. by Dafydd Gibbon and Helmut
Richter. Berlin: de Gruyter.
Dahl, Osten. 1979a. Review article of John Lyons’ Semantics. Language
55:199-206.
1979b. Typology of sentence negation. Linguistics 17:79—106.
Dawkins, R. M. 1916. Modern Greek in Asia Minor. Cambridge: Cambridge
University Press.
Dechert, Hans, Monika Briiggemeir, and Dietmar Futterer. 1984. Transfer and
Interference in Language: A Selected Bibliography. Amsterdam: John
Benjamins.
De Freine, Sean. 1977. The dominance of the English language in the nineteenth
century. In The English Language in Ireland, ed. by Diarmaid O’Muirithe.
Dublin: Mercier Press.
 References 177

Dehghanpisheh, Elaine. 1978. Language development in Farsi and English: im¬

plications for the second language learner. 1RAL: International Review of
Applied Linguistics 16:45—61.
Derbyshire, Desmond. 1986. Comparative survey of morphology and syntax in
Brazilian Arawakan. In Handbook of Amazonian Languages, ed. by Des¬
mond Derbyshire and Geoffrey Pullum. Berlin: Mouton de Gruyter.
De Villiers, Jill, and Peter de Villiers. 1973. A cross-sectional study of the
development of grammatical morphemes. Journal of Psycholinguistic Re¬
search 2:267-78.
1978. Language Acquisition. Cambridge, Mass.: Harvard University Press.
Dickerson, Lonna. 1974. Internal and external patterning of phonological var¬
iability in the speech of Japanese learners of English: toward a theory of
second language acquisition. Unpublished Ph.D. dissertation, University of
Illinois at Champaign-Urbana.
Di Pietro, Robert. 1971. Language Structures in Contrast. Rowley, Mass.: New¬
bury House.
Dulay, Heidi, and Marina Burt. 1973. Should we teach children syntax? Lan¬
guage Learning 23:245—57.
1974. Natural sequences in child second language acquisition. Language
Learning 24:37—53.
1974/1983. Goofing: an indicator of children’s second language strategies.
In Language Transfer in Language Learning, ed. by Susan Gass and Larry
Seiinker. Rowley, Mass.: Newbury House.
Dulay, Heidi, Marina Burt, and Stephen Krashen. 1982. Language Two. New
York: Oxford University Press.
Dundes, Alan. 1964. The Morphology of North American Indian Folktales.
Folklore Fellows Communication, 195. Helsinki: Academia Scientiarum
Fennica.
Duskova, Libuse. 1969. On sources of errors in foreign language teaching.
1RAL: International Review of Applied Linguistics 7:11—36.
Eckman, Fred. 1977. Markedness and the contrastive analysis hypothesis. Lan¬
guage Learning 27:315—30.
1981a. On predicting phonological difficulty in second language acquisition.
Studies in Second Language Acquisition 4:18-30.
1981b. On the naturalness of interlanguage phonological rules. Language
Learning 31:195—216.
1984. Universals, typologies, and interlanguages. In Universals in Second
Language Acquisition, ed. by William Rutherford. Amsterdam: John
Benjamins.
Edwards, Mary. 1979. Phonological processes in fricative acquisition. Papers
and Reports on Child Language Development 17:98—105.
Eggington, William. 1987. Written academic discourse in Korean: implications
for effective communication. In Writing Across Languages: Analysis of L2
Text, ed. by Ulla Connor and Robert Kaplan. Reading, Mass.: Addison-
Wesley.
Ekman, Paul. 1972. Universals and cultural differences in facial expressions of
emotion. In 1971 Nebraska Symposium on Motivation, ed. by James K.
Cole. Lincoln: University of Nebraska Press.
Ellis, Rod. 1985. Understanding Second Language Acquisition. Oxford: Oxford
University Press.
178 References
Emeneau, Murray, 1980. Language and Linguistic Area. Stanford: Stanford
University Press.
Eubank, Lynn. 1986. Formal models of language learning and the acquisition
of German word order and negation by primary and non-primary language
learners. Unpublished Ph.D. dissertation, University of Texas at Austin.
1988. Parameters in L2 learning: Flynn revisited. Presented at the Second
Language Research Forum, University of Hawaii at Manoa.
Fantini, Alvino. 1985. Language Acquisition of a Bilingual Child. Clevedon,
U.K.: Multilingual Matters.
Fathman, Ann. 1975. Language background, age, and the order of acquisition
of English structures. In On TESOL ’75, ed. by Marina Burt and Heidi
Dulay. Washington, D.C.: TESOL.
1977. Similarities and simplification in the interlanguage of second language
learners. In Actes du Sente colloque de linguistique appliquee de Neuchatel,
ed. by S. P. Corder and E. Roulet. Neuchatel: Faculte des Lettres.
Felix, Sascha. 1981. The effect of formal instruction on second language ac¬
quisition. Language Learning 31:87—112.
Ferguson, Charles. 1975. Toward a characterization of English Foreigner Talk.
Anthropological Linguistics 17:1—14.
1981. The structure and use of politeness formulas. In Conversation Routine,
ed. by Florian Coulmas. The Hague: Mouton.
Fillmore, Charles. 1968. The case for case. In Universals of Linguistic Theory,
ed. by Emmon Bach and Robert Harms. New York: Holt, Rinehart, and
Winston.
Fillmore, Lily Wong. 1979. Individual differences in second language acquisi¬
tion. In Individual Differences in Language Ability and Language Behavior,
ed. by Charles Fillmore, Charles Kempler, and William Wang. New York:
Academic Press.
Filppula, Marku. 1986. Some Aspects of Hiberno-English in a Functional Sen¬
tence Perspective. University of Joensuu Publications in the Humanities, 7.
Joensuu, Finland: University of Joensuu.
Flege, James. 1980. Phonetic approximation in second language acquisition.
Language Learning 30:117—34.
1981. The phonological basis of foreign accent: a hypothesis. TESOL Quar¬
terly 15:443—45.
1987. The production of “new” and “similar” phones in a foreign language:
evidence for the effect of equivalence classification. Journal of Phonetics
15:47-65.
Flege, James, and Robert Hammond. 1 982. Mimicry of non-distinctive phonetic
differences between language varieties. Studies in Second Language Ac¬
quisition 5:1—17.
Flynn, Suzanne. 1984. A universal in L2 acquisition based on a PBD typology.
In Universals of Second Language Acquisition, ed. by Fred Eckman, Law¬
rence Bell, and Diane Nelson. Rowley, Mass.: Newbury House.
Flynn, Suzanne, and I. Espinal. 1985. Head-initial/head-final parameter in adult
Chinese L2 acquisition of English. Second Language Research 1/2:93—1 17.
Foreign Service Institute. 1985. Schedule of Courses, 1985—1986. Washington,
D.C.: U.S. Department of State.
Foss, Donald, and David Hakes. 1978. Psycholinguistics: An Introduction to
the Psychology of Language. Englewood Cliffs, N.J.: Prentice-Hall.
 References 179

Fox, Barbara. 1987. The noun phrase accessibility hierarchy revisited. Language
63:856-70.
Friedrich, Paul. 1979. The symbol and its relative non-arbitrariness. In Lan¬
guage, Context, and the Imagination: Essays by Paul Friedrich, ed. by
Anwar Dil. Stanford: Stanford University Press.
Fries, Charles. 1945. Teaching and Learning English as a Foreign Language.
Ann Arbor: University of Michigan Press.
1949. The Chicago investigation. Language Learning 2:89—99.
1952. The Structure of English. New York: Harcourt, Brace, and Company.
Fromkin, Victoria, and Robert Rodman. 1983. An Introduction to Language.
New York: Holt, Rinehart, and Winston.
Gandour, Jackson, and Richard Harshman. 1978. Crosslanguage differences in
tone perception: a multidimensional scaling investigation. Language and
Speech 21:1—33.
Garvin, John. 1977. The Anglo-Irish idiom in the works of major Irish writers.
In The English Language in Ireland, ed. by Diarmaid O’Muirithe. Dublin:
Mercier Press.
Gass, Susan. 1979. Language transfer and universal grammatical relations. Lan¬
guage Learning 29:327—44.
1983. Second language acquisition and language universals. In The First Del¬
aware Symposium on Language Studies, ed. by Robert DiPietro, William
Frawley, and Albert Wedel. Newark, Del.: University of Delaware Press.
1984. A review of interlanguage syntax: language transfer and language uni¬
versals. Language Learning 34:115—32.
1986. The resolution of conflicts among competing systems: a bidirectional
perspective. Unpublished manuscript.
Gass, Susan, and Josh Ard. 1984. Second language acquisition and the ontology
of language universals. In Universals in Second Language Acquisition, ed.
by William Rutherford. Amsterdam: John Benjamins.
Gass, Susan, and Larry Seiinker, eds. 1983. Language Transfer in Language
Learning. Rowley, Mass.: Newbury House.
Gazdar, Gerald, Ewan Klein, Geoffrey Pullum, and Ivan Sag. 1985. Generalized
Phrase Structure Grammar. Cambridge, Mass.: Harvard University Press.
Gazdar, Gerald, and Geoffrey Pullum. 1985. Computationally relevant prop¬
erties of natural languages and grammars. New Generation Computing
3:273-306.
Gelb, I. J. 1963. A Study of Writing. Chicago: University of Chicago Press.
Genesee, Fred. 1979. Acquisition of reading skill in immersion programs. For¬
eign Language Annals 12:71—77.
Gilbert, Glenn. 1980. Introduction. In Pidgin and Creole Languages: Selected
Essays by Hugo Schuchardt, ed. by Glenn Gilbert. London: Cambridge
University Press.
1983. Transfer in second language acquisition. In Pidginization and Creoli-
zation as Language Acquisition, ed. by Roger Andersen. Rowley, Mass.:
Newbury House.
Giles, Howard. 1973. Communicative effectiveness as a function of accented
speech. Speech Monographs 40:330-31.
Gilsan, Eileen. 1985. The effect of word order on listening comprehension and
pattern retention: an experiment in Spanish as a foreign language. Language
Learning 35:443—72.
180 References
Givon, Talmy. 1979. On Understanding Grammar. New York: Academic Press.
Givon, Talmy, ed. 1983a. Topic Continuity in Discourse: A Quantitative Cross¬
Language Study. Amsterdam: John Benjamins.
Givon, Talmy. 1983b. Topic continuity in discourse: an introduction. In Topic
Continuity in Discourse: A Quantitative Cross-Language Study, ed. by
Talmy Givon. Amsterdam: John Benjamins.
1984a. Syntax: A Functional/Typological Introduction. Vol. 1. Amsterdam:
John Benjamins.
1984b. Universals of discourse structure and second language acquisition. In
Universals in Second Language Acquisition, ed. by William Rutherford.
Amsterdam: John Benjamins.
Godard, Daniele. 1977. Same setting, different norms: phone call beginnings in
France and the United States. Language in Society 6:209—19.
Gonzo, Susan, and Mario Saltarelli. 1983. Pidginization and linguistic change
in immigrant languages. In Pidginization and Creolization as Language
Acquisition, ed. by Roger Andersen. Rowley, Mass.: Newbury
House.
Granfors, Tom, and Rolf Palmberg. 1976. Errors made by Finns and Swedish¬
speaking Finns at a commercial college level. In Errors Made by Finns and
Swedish-Speaking Finns in the Learning of English, ed. by Hakan Ringbom
and Rolf Palmberg. Abo, Finland: Department of English, Abo Akademi.
ERIC Report ED 122628.
Greenberg, Cindy. 1983. Syllable structure in second language acquisition.
CUNY Forum 9:41-63.
Greenberg, Joseph. 1965. Some generalizations concerning initial and final con¬
sonant sequences. Linguistics 18:5—32.
1966. Some universals of grammar with particular reference to the order of
meaningful elements. In Universals of Language, ed. by Joseph Greenberg.
Cambridge, Mass.: MIT Press.
Gregg, Kevin. 1984. Krashen’s monitor and Ockham’s razor. Applied Linguis¬
tics 5:79-100.
Gudschinsky, Sarah. 1967. How to Learn an Unwritten Language. New York:
Holt, Rinehart, and Winston.
Guiora, Alexander. 1972. Construct validity and transpositional research: to¬
ward an empirical study of psychoanalytic concepts. Comprehensive Psy¬
chiatry 10:139—50.
Guiora, Alexander, and William Acton. 1979. Personality and language: a re¬
statement. Language Learning 29:193—204.
Guiora, Alexander, William Acton, Robert Erard, and Fred Strickland. 1980.
The effects of benzodiazepine (Valium) on permeability of language ego
boundaries. Language Learning 30:351—63.
Guiora, Alexander, Benjamin Beit-Hallahmi, Robert Brannon, Cecelia Dull, and
Thomas ScoveL 1972. The effects of experimentally induced changes in
ego states on pronunciation ability in a second language: an exploratory
study. Comprehensive Psychiatry 10:421—28.
Gumperz, John. 1982. Discourse Strategies. Cambridge: Cambridge University
Press.
Gumperz, John, and Robert Wilson. 1971. Convergence and creolization. In
Pidgins and Creoles, ed. by Dell Hymes. Cambridge: Cambridge University
Press.
 References 181

Hagege, Claude. 1976. Relative clauses, center-embedding, and comprehensi¬

bility. Linguistic Inquiry 7:198—201.
1987. Le franqais et les siecles. Paris: Editions Odile Jacob.
Hakuta, Kenji. 1976. A case study of a Japanese child learning English as a
second language. Language Learning 26:321—51.
1986. Mirror of Language: The Debate on Bilingualism. New York: Basic
Books.
Hamill, James. 1978. Transcultural logic: testing hypotheses in three languages.
In Discourse and Inference from Cognitive Psychology, ed. by M. D. Loflin
and J. Silverberg. The Hague: Mouton.
Harley, Birgit. 1984. Age as a factor in the acquisition of French as a second
language in an immersion setting. In Second Languages: A Cross-Linguistic
Perspective, ed. by Roger Andersen. Rowley, Mass.: Newbury House.
1986. Age in Second Language Acquisition. San Diego: College Hill Press.
Harris, Jesse. 1948. German language influences in St. Clair County, Illinois.
American Speech 23:106—10.
Harris, John. 1984. Syntactic variation and dialect divergence. Journal of Lin¬
guistics 20:303—27.
Hartmann, Reinhard. 1980. Contrastive Textology. Heidelberg: Julius Groos.
Hatch, Evelyn. 1983. Psycholinguistics: A Second Language Perspective. Row-
ley, Mass.: Newbury House.
Haugen, Einar. 1953. The Norwegian Language in America. Philadelphia: Uni¬
versity of Pennsylvania Press.
Hawkins, John. 1983. Word Order Universals. New York: Academic Press.
Heath, Shirley. 1983. Ways with Words. Cambridge: Cambridge University
Press.
Hecht, Barbara, and Randa Mulford. 1987. The acquisition of a second language
phonology: interaction of transfer and developmental factors. In Interlan¬
guage Phonology, ed. by Georgette loup and Steven Weinberger. Rowley,
Mass.: Newbury House.
Henry, P. L. 1957. An Anglo-English Dialect of North Roscommon. Dublin:
Department of English, University College.
1977. Anglo-Irish and its background. In The English Language in Ireland,
ed. by Diarmaid O’Muirithe. Dublin: Mercier Press.
Herman, Lewis, and Marguerite Herman. 1943. Foreign Dialects: A Manual
for Actors, Directors, and Writers. New York: Theatre Arts Books.
Hillocks, George. 1986. Research on Written Composition. Urbana, Ill.: Na¬
tional Conference on Research in English and ERIC Clearinghouse on
Reading and Communication Skills.
Hinds, John. 1975. Third person pronouns in Japanese. In Language in Japanese
Society, ed. Fred C. C. Ping. Tokyo: University of Tokyo Press.
1983. Contrastive rhetoric: Japanese and English. Text 3:183—95.
1984. Retention of information using a Japanese style of presentation. Studies
in Language 8:45—69.
1987. Reader versus writer responsibility: a new typology. In Writing Across
Languages: Analysis of L2 Text, ed. by Ulla Connor and Robert Kaplan.
Reading, Mass.: Addison-Wesley.
Hockett, Charles. 1954. Chinese versus English: an exploration of the Whorfian
theses. In Language in Culture, ed. by Henry Hoijer. Chicago: University
of Chicago Press.
182 References
Hocking, B.D.W. 1973. Types of interference. In Focus on the Learner, ed. by
John Oller and Jack Richards. Rowley, Mass.: Newbury House.
Holmes, Glyn. 1977. The problem of anglicized French at the university. Ca¬
nadian Modern Language Review 33:520.
Homburg, Taco. 1984. Holistic evaluation of ESL compositions: can it be val¬
idated objectively? TESOL Quarterly 18:87—107.
Hornby, A. S. 1974. Oxford Advanced Learner’s Dictionary of Current English.
Oxford: Oxford University Press.
House, Juliane, and Gabriele Kasper. 1981. Politeness markers in English and
German. In Conversation Routine, ed. by Florian Coulmas. The Hague:
Mouton.
Huang, Joseph, and Evelyn Hatch. 1978. A Chinese child’s acquisition of En¬
glish. In Second Language Acquisition: A Book of Readings, ed. by Evelyn
Hatch. Rowley, Mass.: Newbury House.
Huebner, Thom. 1983. A Longitudinal Analysis of the Acquisition of English.
Ann Arbor: Karoma.
Huerta, Ana. 1978. Code-switching among Spanish-English bilinguals: a so-
ciolinguistic perspective. Unpublished Ph.D. dissertation, University of
Texas at Austin.
Hutchins, Edwin. 1981. Reasoning in Trobriand discourse. In Language, Cul¬
ture, and Cognition, ed. by Ronald Casson. New York: Macmillan.
Hyltenstam, Kenneth. 1977. Implicational patterns in interlanguage syntax var¬
iation. Language Learning 27:383—411.
1982. On descriptive adequacy and psychological plausibility: a reply to
Jordens. Language Learning 32:167—73.
1984. The use of typological markedness conditions as predictors in second
language acquisition. In Second Languages: A Cross-Linguistic Perspective,
ed. by Roger Andersen. Rowley, Mass.: Newbury House.
Hyman, Larry. 1975. Phonology: Theory and Analysis. New York: Holt, Rine¬
hart, and Winston.
Ibrahim, Muhammad. 1978. Patterns in spelling errors. English Language
Teaching 32:207—12.
Ijaz, I. Helene. 1986. Linguistic and cognitive determinants of lexical acquisition
in a second language. Language Learning 36:401—51.
loup, Georgette. 1984. Is there a structural foreign accent? A comparison of
syntactic and phonological errors in a second language. Language Learning
34:1-17.
loup, Georgette, and Anna Kruse. 1977. Interference and structural complexity
as a predictor of second language relative clause acquisition. In Proceedings
of the Los Angeles Second Language Acquisition Research Forum, ed. by
Carol Henning. Los Angeles: Department of English, University of Cali¬
fornia at Los Angeles.
loup, Georgette, and Amara Tansomboon. 1987. The acquisition of tone: a
maturational perspective. In Interlanguage Phonology, ed. by Georgette
loup and Steven Weinberger. Rowley, Mass.: Newbury House.
loup, Georgette, and Steven Weinberger, eds. 1987. Interlanguage Phonology.
Rowley, Mass.: Newbury House.
Itoh, Harumi, and Evelyn Hatch. 1978. Second language acquisition: a case
study. In Second Language Acquisition: A Book of Readings, ed. by Evelyn
Hatch. Rowley, Mass.: Newbury House.
Jackson, Jean. 1974. Language identity of the Colombian Vaupes Indians. In
 References 183
Explorations in the Ethnography of Speaking, ed. by Richard Bauman and
Joel Sherzer. Cambridge: Cambridge University Press.
James, Carl. 1971. The exculpation of contrastive linguistics. In Papers in Con¬
trastive Linguistics, ed. by Gerhard Nickel. Cambridge: Cambridge Uni¬
versity Press.
1980. Contrastive Analysis. London: Longman.
Jansen, Bert, Josien Lalleman, and Pieter Muysken. 1981. The alternation hy¬
pothesis: acquisition of Dutch word order by Turkish and Moroccan for¬
eign workers. Language Learning 31:315—36.
Jespersen, Otto. 1912. How to Teach a Foreign Language. London: George
Allen.
1929. The Philosophy of Grammar. London: Allen and Unwin.
1954. A Modern English Grammar on Historical Principles. Part 3. London:
Allen and Unwin.
Johansson, Faith. 1973. Immigrant Swedish Phonology. Lund, Sweden:
Gleerup.
Jordens, Peter. 1977. Rules, grammatical intuitions, and strategies in foreign
language learning. Interlanguage Studies Bulletin 2:5—76.
1 980. Interlanguage research: interpretation or explanation. Language Learn¬
ing 30:195-207.
1982. How to make your facts fit: a response from Jordens. Language Learn¬
ing 32:175-81.
Joseph, Brian. 1983a. The Synchrony and Diachrony of the Balkan Infinitive.
New York: Cambridge University Press.
1983b. Relativization in Greek: another look at the accessibility hierarchy
constraints. Lingua 60:1—24.
Kachru, Braj. 1969. English in South Asia. In Current Trends in Linguistics 5,
ed. by Thomas Sebeok. The Hague: Mouton.
1983. English in South Asia. In English as a World Language, ed. by Richard
Bailey and Manfred Gorlach. Ann Arbor: University of Michigan Press.
1987. The Alchemy of English. New York: Pergamon Press.
Kachru, Yamuna. 1983. English and Hindi. In Annual Review of Applied Lin¬
guistics, 1 982, ed. by Robert Kaplan. Rowley, Mass.: Newbury House.
Kalmar, Ivan, Zhong Yong, and Xiao Hong. 1988. Language attitudes in Guang¬
zhou, China. Language in Society 16:499—508.
Kamratowski, Joachim, and Joachim Schneider. 1969. Zum Problem der en-
glischen Rechtschreibung. Englisch 4:69—74.
Kaplan, Robert. 1966. Cultural thought patterns in intercultural education.
Language Learning 16:1—20.
Kasper, Gabriele. 1981. Pragmatische Aspekte in der Interimsprache. Tubingen:
Gunter Narr.
Keenan, Edward. 1975. Variation in Universal Grammar. In Analyzing Varia¬
tion in Language, ed. by Ralph Fasold and Roger Shuy. Washington, D.C.:
Georgetown University Press.
1976. Towards a universal definition of “subject.” In Subject and Topic, ed.
by Charles Li. New York: Academic Press.
1978. The syntax of subject-final languages. In Syntactic Typology, ed. by
Winfred Lehmann. Austin: University of Texas Press.
1985. Relative clauses. In Language Typology and Syntactic Description,
Volume 2, ed. by Timothy Shopen. Cambridge: Cambridge University Press.
Keenan, Edward, and Kent Bimson. 1975. Perceptual complexity and the cross-
184 References

language distribution of relative clause and NP-question types. In Papers

from the Parasession on Functionalism, ed. by Robin Grossman, L. James
San, and Timothy Vance. Chicago: Chicago Linguistic Society.
Keenan, Edward, and Bernard Comrie. 1977. Noun phrase accessibility and
Universal Grammar. Linguistic Inquiry 8:63—99.
Keller-Cohen, Deborah. 1979. Systematicity and variation in the non-native
child’s acquisition of conversational skills. Language Learning 29:27—44.
Kellerman, Eric. 1977. Toward a characterisation of the strategy of transfer in
second language learning. Interlanguage Studies Bulletin 2:58—145.
1978. Transfer and non-transfer: where we are now. Studies in Second Lan¬
guage Acquisition 2:37—57.
1983. Now you see it, now you don’t. In Language Transfer in Language
Learning, ed. by Susan Gass and Larry Seiinker. Rowley, Mass.: Newbury
House.
1984. The empirical evidence for the influence of LI on interlanguage. In
Interlanguage, ed. by Alan Davies, C. Criper, and A.P.R. Howatt. Edin¬
burgh: Edinburgh University Press.
Kellerman, Eric, and Michael Sharwood Smith. 1986. Crosslinguistic influence
in second language acquisition: an introduction. In Crosslinguistic Influence
in Second Language Acquisition, ed. by Eric Kellerman and Michael Shar¬
wood Smith. New York: Pergamon Press.
Kelly, Louis. 1969. 25 Centuries of Language Teaching. Rowley, Mass.: New¬
bury House.
Kempf, Margaret. 1975. A study of English proficiency level and the composition
errors of incoming foreign students at the University of Cincinnati 1969-
1974. Unpublished Ph.D. dissertation, The Ohio State University.
Kinneavy, James. 1971. A Theory of Discourse. Englewood Cliffs, N.J.: Prentice-
Hall.
Kintsch, Walter, and Edith Greene. 1978. The role of culture-specific schemata
—
in the comprehension and recall of stories. Discourse Processes 1:1 13.
Klein, Wolfgang. 1986. Second Language Acquisition. Cambridge: Cambridge
University Press.
Kleinmann, Howard. 1977. Avoidance behavior in adult second language ac¬
quisition. Language Learning 27:93—107.
Koch, Barbara Johnstone. 1983. Presentation as proof: the language of Arabic
rhetoric. Anthropological Linguistics 25:47—60.
Kohn, Kurt. 1987. The analysis of transfer. In Crosslinguistic Influence in Second
Language Acquisition, ed. by Eric Kellerman and Michael Sharwood Smith.
New York: Pergamon Press.
Krashen, Stephen. 1978. Some issues relating to the Monitor Model. In On
TESOL '77, ed. by H. Douglas Brown, Carlos Yorio, and Ruth Crymes.
Washington, D.C.: TESOL.
1981. Second Language Acquisition and Second Language Learning. Oxford:
Pergamon Press.
1983. Newmark’s “Ignorance Hypothesis” and current second language ac¬
quisition theory. In Language Transfer in Language Learning, ed. by Susan
Gass and Larry Seiinker. Rowley, Mass.: Newbury House.
Krashen, Stephen, Michael Long, and Robin Scarcella. 1979. Age, rate and
eventual attainment in second language acquisition. TESOL Quarterly
13:573-82.
 References 185

I0ashen, Stephen, and Tracy Terrell. 1983. The Natural Approach. Oxford:
Pergamon/Alemany Press.
Kuno, Susumo. 1974. The position of relative clauses and conjunctions. Lin¬
guistic Inquiry 5:117—36.
Labov, William. 1972. Sociolinguistic Patterns. Philadelphia: University of
Pennsylvania Press.
1975. What Is a Linguistic Fact? Lisse, The Netherlands: Peter de Ridder.
Lado, Robert. 1957. Linguistics Across Cultures. Ann Arbor: University of
Michigan Press.
Lakoff, George. 1972. Hedges: a study in meaning criteria and the logic of
fuzzy concepts. Papers from the 8th Regional Meeting. Chicago: Chicago
Linguistic Society.
1987. Women, Fire, and Dangerous Things: What Categories Reveal about
the Mind. Chicago: University of Chicago Press.
Larsen-Freeman, Diane. 1978. An ESL index of development. TESOL Quarterly
12:439-48.
Lass, Roger. 1986. “Irish influence”: reflections on “standard” English and its
opposites, and the identification of caiques. Studia Anglica Posnaniensia
18:81-87.
Lazard, Gilbert. 1957. Grammaire du persan contemporain. Paris: Klincksieck.
Lazzerini, Lucia. 1982. Aux origines du macaronique. Revue des langues ro-
manes 86:11—33.
Lee, W. R. 1968. Thoughts on contrastive linguistics in the context of language
teaching. In Report of the Nineteenth Annual Round Table Meeting on
Linguistics and Language Studies, ed. by James Alatis. Monograph Series
on Languages and Linguistics, Georgetown University. Washington, D.C.:
Georgetown University Press.
Le Page, Robert, and Andree Tabouret-Keller. 1985. Acts of Identity: Creole-
Based Approaches to Language and Ethnicity. Cambridge: Cambridge Uni¬
versity Press.
Leslau, Wolf. 1945. The influence of Cushitic on the Semitic languages of Ethio¬
pia. Word 1:59-82.
1952. The influence of Sidamo on the Ethiopic languages of Gurage. Language
28:63-81.
Leung, K. C. 1978. The Cantonese student in the Mandarin class: some special
problems. Journal of the Chinese Language Teachers Association 13:51—
55.
Levin, Harry, Irene Silverman, and Boyce Ford. 1967. Hesitation in children’s
speech during explanation and description. Journal of Verbal Learning and
Verbal Behavior 6:560—64.
Levinson, Stephen. 1983. Pragmatics. Cambridge: Cambridge University Press.
Levi-Strauss, Claude. 1955. The structural study of myth. Journal of American
Folklore 68:428-44.
Li, Charles. 1984. From verb-medial analytic language to verb-final synthetic
language: a case of typological change. In Proceedings from the Tenth
Annual Meeting of the Berkeley Linguistics Society, ed. by Claudia Brugman
and Monica Macaulay. Berkeley: Berkeley Linguistics Society.
Li, Charles, and Sandra Thompson. 1976. Subject and topic: a new typology
of language. In Subject and Topic, ed. by Charles Li. New York: Academic
Press.
186 References
Liberman, Alvin, Katherine Harris, Howard Hoffman, and Bel ver Griffith. 1 957.
The discrimination of speech sounds within and across phoneme bound¬
aries. Journal of Experimental Psychology 54:358—68.
Lightbown, Patsy, and Gary Libben. 1984. The recognition and use of cognates
by L2 learners. In Second Languages: A Cross-Linguistic Perspective, ed.
by Roger Andersen. Rowley, Mass.: Newbury House.
Limper, Louis. 1932. Student knowledge of some French-English cognates.
French Review 6:37—49.
Linnarud, Moira. 1978. Cohesion and communication in the target language.
Interlanguage Studies Bulletin 3:23—34.
Littlewood, William. 1973. A comparison of first language acquisition and
second language learning. Praxis des Neusprachlichen Unterrichts 20:343—
48.
1983. Contrastive pragmatics and the foreign language learner’s personality.
Applied Linguistics 4:200—06.
Liu, Lisa Garbern. 1985. Reasoning counterfactually in Chinese: are there any
obstacles? Cognition 21:239—70.
LoCoco, Veronica. 1975. An analysis of Spanish and German learners’ errors.
Working Papers in Bilingualism 7:96—124.
Long, Michael, and Charlene Sato. 1984. Methodological issues in interlanguage
studies: an interactionist perspective. In Interlanguage, ed. by Alan Davies,
C. Criper, and A.P.R. Howatt. Edinburgh: Edinburgh University Press.
Loveday, Leo. 1982a. Communicative interference: a framework for contras-
tively analysing L2 communicative competence exemplified with the lin¬
guistic behaviour of Japanese performing in English. IRAL: International
Review of Applied Linguistics 20:1—16.
1982b. The Sociolinguistics of Learning and Using a Non-Native Language.
Oxford: Pergamon Press.
Lowenberg, Peter. 1986. Non-native varieties of English: nativization, norms,
and implications. Studies in Second Language Acquisition 8:1—18.
Lujan, Marta, Liliana Minaya, and David Sankoff. 1984. The universal con¬
sistency hypothesis and the prediction of word order acquisition stages in
the speech of bilingual children. Language 60:343—71.
Lukatela, G., M. D. Savic, P. Ognjenovic, and M. T. Turvey. 1978. On the
relation between processing the Roman and the Cyrillic alphabets. Lan¬
guage and Speech 21:113—73.
Lyons, John. 1977. Semantics. Cambridge: Cambridge University Press.
Maddieson, Ian. 1984. Patterns of Sounds. Cambridge: Cambridge University
Press.
Major, Roy. 1986. Paragoge and degree of foreign accent in Brazilian English.
Second Language Research 2:53—71.
1987. A model for interlanguage phonology. In Interlanguage Phonology, ed.
by Georgette loup and Steven Weinberger. Rowley, Mass.: Newbury
House.
Mandler, Jean, Sylvia Scribner, Michael Cole, and Marsha DeForest. 1980.
Cross-cultural invariance in story recall. Child Development 51:19—26.
Mann, Virginia. 1986. Distinguishing universal and language-dependent levels
of speech perception: evidence from Japanese learners’ perception of English
“I” and “r.” Cognition 24:169—96.
Marckwardt, Albert. 1946. Phonemic structure and aural perception. American
Speech 21:106-11.
 References 187

Masny, Diana, and Alison d’Anglejan. 1985. Language, cognition, and second
language grammaticality judgments. Journal of Psycholinguistic Research
14:175-97.
McLaughlin, Barry. 1978. Second Language Acquisition in Childhood. Hills¬
dale, N.J.: Lawrence Erlbaum.
1981. Differences and similarities between first- and second-language learn¬
ing. In Native Language and Foreign Language Acquisition, ed. by Harris
Winitz. Annals of the New York Academy of Sciences, Volume 379. New
York: New York Academy of Sciences.
McLaughlin, Barry, Tammi Rossman, and Beverly McLeod. 1983. Second lan¬
guage learning: an information-processing perspective. Language Learning
33:135-58.
Meillet, Antoine. 1948. Linguistique historique et linguistique generale. Paris:
Societe Linguistique de Paris.
Meisel, Jurgen. 1983. Strategies of second language acquisition: more than one
kind of simplification. In Pidginization and Creolization as Language Ac¬
quisition, ed. by Roger Andersen. Rowley, Mass.: Newbury House.
Meisel, Jurgen, Harald Clahsen, and Manfred Pienemann. 1981. On determining
developmental stages in natural second language acquisition. Studies in
Second Language Acquisition 3:109—35.
Meo Zilio, Giovanni. 1959. Una serie di morfemi italiani con funzione stilistica
nello spagnolo nell’Uruguay. Lingua Nostra 20:49—54.
1964. El “Cocoliche” Rioplatense. Bolettn de Filologi'a 16:61—119.
Mervis, Carolyn, and Emilie Roth. 1981. The internal structure of basic and
non-basic color categories. Language 57:384—405.
Mohan, Bernard, and Winnie Au-Yeung Lo. 1985. Academic writing and
Chinese students: transfer and developmental factors. TESOL Quarterly
19:515-34.
Moravcsik, Edith, and Jessica Wirth. 1986. Markedness — an overview. In
Markedness, ed. by Fred Eckman, Edith Moravcsik, and Jessica Wirth.
New York: Plenum Press.
Moulton, William. 1962a. Toward a classification of pronunciation errors.
Modern Language Journal 46:101—09.
1962b. The Sounds of English and German. Chicago: University of Chicago
Press.
Miihlhausler, Peter. 1986. Pidgin and Creole Linguistics. Oxford: Basil
Blackwell.
Muller, Max. 1861/1965. Lectures on the Science of Language. New Delhi:
Munshi Ram Manohar Lal.
Muysken, Pieter. 1984. The Spanish that Quechua speakers learn. In Second
Languages: A Cross-Linguistic Perspective, ed. by Roger Andersen. Row-
ley, Mass.: Newbury House.
Muysken, Pieter, and Norval Smith, eds. 1986. Substrata Versus Universals in
Creole Genesis. Amsterdam: John Benjamins.
Nagara, Susumu. 1972. Japanese Pidgin English in Hawaii: A Bilingual De¬
scription. Honolulu: University Press of Hawaii.
Nelson, Katherine. 1974. Concept, word and sentence: interrelations in acqui¬
sition and development. Psychological Review 81:276—85.
Neufeld, Gerald. 1978. On the acquisition of prosodic and articulatory features
in adult language learning. Canadian Modern Language Review 34:163—
74.
188 References

Obilade, Tony. 1984. Mother tongue influence on polite communication in a

second language. Language and Communication 4:295-99.
Odlin, Terence. 1986. On the nature and use of explicit knowledge. IRAL:
International Review of Applied Linguistics 24:123—44.
1987. Word order transfer and metalinguistic awareness. Presented at the
University of Illinois Conference on Second Language Acquisition and For¬
eign Language Learning, Champaign, Ill.
1988. Divil a lie: semantic transparency and language transfer. Presented at
the Second Language Research Forum, University of Hawaii at Manoa.
Oller, John, and Elcho Redding. 1971. Article usage and other language skills.
Language Learning 21/1:85—95.
Oller, John, and Seid Ziahosseiny. 1970. The contrastive analysis hypothesis
and spelling errors. Language Learning 20:183—89.
Olshtain, Elite. 1983. Sociocultural competence and language transfer: the case
of apology. In Language Transfer in Language Learning, ed. by Susan Gass
and Larry Seiinker. Rowley, Mass.: Newbury House.
Olson, Linda, and S. Jay Samuels. 1973. The relationship between age and
accuracy of foreign language pronunciation. Journal of Educational Re¬
search 66:263—67.
Ong, Walter. 1982. Orality and Literacy. New York: Methuen.
Orth, John. 1982. University undergraduate evaluational reactions to the speech
of foreign teaching assistants. Unpublished Ph.D. dissertation, University
of Texas at Austin.
Osgood, Charles, William May, and Murray Miron. 1975. Cross-cultural Uni¬
versals of Affective Meaning. Urbana, HL: University of Illinois Press.
Oxford, Rebecca, and Nancy Rhodes. 1988. U.S. foreign language instruction:
assessing needs and creating an action plan. ERIC/CLL News Bulletin 11:1,
6-8.
Oyama, Susan. 1976. A sensitive period for the acquisition of a nonnative
phonological system. Journal of Psycholinguistic Research 5:261—85.
Palmer, Harold. 1917. The Scientific Study and Teaching of Languages.
Yonkers-on-Hudson, N.Y.: World Book Company.
Pandharipande, Rajeshwari. 1983. Linguistics and written discourse in partic¬
ular contrastive studies: English and Marathi. In Annual Review of Applied
Linguistics, 1982, ed. by Robert Kaplan. Rowley, Mass.: Newbury House.
Paulston, Christina. 1978. Biculturalism: some reflections and speculations.
TESOL Quarterly 12:369—80.
Pavlovitch, Milivoie. 1920. Le langage enfantin: acquisition du serbe et du
franQais par un enfant serbe. Paris: Champion.
Peck, Sabrina. 1978. Child—child discourse in second language acquisition. In
Second Language Acquisition: A Book of Readings, ed. by Evelyn Hatch.
Rowley, Mass.: Newbury House.
Penn, Julia. 1972. Linguistic Relativity Versus Innate Ideas. The Hague:
Mouton.
Pica, Teresa. 1984. LI transfer and L2 complexity as factors in syllabus design.
TESOL Quarterly 18:689-704.
Pienemann, Manfred. 1981. Der Zweitspracherwerb auslandischer Arbeiter-
kinder. Bonn: Bouvier Verlag Herbert Grundmann.
1984. Psychological constraints on the teachability of languages. Studies in
Second Language Acquisition 6:186—214.
 References 189

Pike, Eunice. 1959. A test for predicting phonetic ability. Language Learning
9:35-41.
Pike, Kenneth. 1954. Language in Relation to a Unified Theory of the Structure
of Human Behavior. Glendale, Calif.: Summer Institute of Linguistics.
Platt, John, Heidi Weber, and Ho Mian Lian. 1984. The New Englishes. Lon¬
don: Routledge, Kegan, and Paul.
Poplack, Shana. 1980. Sometimes I’ll start a conversation in Spanish Y TER¬
MING EN ESPANOL: toward a typology of code-switching. Linguistics
18:581-616.
Propp, Vladimir. 1 968. Morphology of the Folktale. Austin: University of Texas
Press.
Purcell, Edward, and Richard Suter. 1980. Predictors of pronunciation accuracy:
a reexamination. Language Learning 30:271—87.
Piirschel, Heiner. 1975. Pause und Kadenz: Interferenzerscheinungen bei der
englischen Intonation deutscher Sprecher. Tubingen: Max Niemeyer.
Purves, Alan. 1986. Rhetorical communities, the international student, and basic
writing. Journal of Basic Writing 5:38—51.
Rao, Raja. 1963. Kanthapura. New York: New Directions.
Ravem, Roar. 1968. Language acquisition in a second language environment.
IRAL: International Review of Applied Linguistics 6:175—85.
Redlinger, Wendy, and Tschange-zin Park. 1980. Language mixing in young
bilinguals. Journal of Child Language 7:337—52.
Reisman, Karl. 1974. Contrapuntal conversations in an Antiguan village. In
Explorations in the Ethnography of Speaking, ed. by Richard Bauman and
Joel Sherzer. London: Cambridge University Press.
Richards, Jack. 1971. A noncontrastive approach to error analysis. English
Language Teaching 25:204-19.
1979. Rhetorical and communicative styles in the new varieties of English.
Language Learning 29:1—25.
1980. Conversation. TESOL Quarterly 14:413—32.
Richards, Jack, and Mayuri Sukwiwat. 1983. Cross-cultural aspects of conver¬
sational competence. Applied Linguistics 4:113—25.
Ringbom, Hakan. 1976. What differences are there between Finns and Swedish¬
speaking Finns learning English? In Errors Made by Finns and Swedish-
Speaking Finns in the Learning of English, ed. by Hakan Ringbom and
Rolf Palmberg. Abo, Finland: Department of English, Abo Akademi. ERIC
Report ED 122628.
1986. Crosslinguistic influence and the foreign language learning process. In
Crosslinguistic Influence in Second Language Acquisition, ed. by Eric Kel¬
lerman and Michael Sharwood Smith. New York: Pergamon Press.
1987. The Role of the First Language in Foreign Language Learning. Cleve-
don, U.K.: Multilingual Matters.
Ringbom, Hakan, and Rolf Palmberg, eds. 1976. Errors Made by Finns and
Swedish-Speaking Finns in the Learning of English. Abo, Finland: De¬
partment of English, Abo Akademi. ERIC Report ED 122628.
Rintell, Ellen. 1979. Getting your speech act together: the pragmatic ability of
second language learners. Working Papers in Bilingualism 17:97—106.
1 984. But how did you FEEL about that? The learner’s perception of emotion
in speech. Applied Linguistics 5:255—64.
Robins, Robert. 1979. A Short History of Linguistics. London: Longman.
190 References

Rona, Jose. 1 965. El Dialecto “Fronterizo” del Norte del Uruguay. Montevideo:
Adolfo Linardi.
Ronjat, Jules. 1913. Le developpement du langage observe chez un enfant bi-
lingue. Paris: Champion.
Rosansky, Ellen. 1 976. Methods and morphemes in second language acquisition.
Language Learning 26:409—25.
Rosch, Eleanor. 1973. On the internal structure of perceptual and semantic
categories. In Cognitive Development and the Acquisition of Language, ed.
by Timothy Moore. New York: Academic Press.
1974. Linguistic relativity. In Human Communication: Theoretical Explo¬
rations, ed. by Albert Silverstein. Hillsdale, N.J.: Lawrence Erlbaum.
Rutherford, William. 1983. Language typology and language transfer. In Lan¬
guage Transfer in Language Learning, ed. by Susan Gass and Larry Seiinker.
Rowley, Mass.: Newbury House.
Ryan, Ellen, Miguel Carranza, and Robert Moffie. 1977. Reactions toward
varying degrees of accentedness in the speech of Spanish-English bilinguals.
Language and Speech 20:267—73.
Sacks, Harvey, Emanuel Schegloff, and Gail Jefferson. 1974. A simplest system¬
atics for the organization of turn-taking in conversation. Language 50:696—
735.
Sajavaara, Kari, and Jaakko Lehtonen. 1981. A bibliography of applied con¬
trastive studies. In Contrastive Linguistics and the Language Teacher, ed.
by Jacek Fisiak. Oxford: Pergamon Press.
Sampson, Geoffrey. 1985. Writing Systems. London: Hutchinson.
Sandfeld, Kristien. 1930. Linguistique balkanique. Paris: Champion.
Sanford, A. J., and S. C. Garrod. 1981. Understanding Written Language.
Chichester, U.K.: John Wiley.
Sato, Charlene. 1984. Phonological processes in second language acquisition:
another look at interlanguage syllable structure. Language Learning 34:43—
57.
Scarcella, Robin. 1979. On speaking politely in a second language. In On
TESOL ‘79, ed. by Carlos Yorio, Kyle Perkins, and Jacquelyn Schachter.
Washington, D.C.: TESOL.
1984. How writers orient their readers in expository essays: a comparative
study of native and non-native English writers. TESOL Quarterly 18:671—
88.
Schachter, Jacquelyn. 1974. An error in error analysis. Language Learning
24:205-14.
1 986. In search of systematicity in interlanguage production. Studies in Second
Language Acquisition 8:119—33.
In Press. Second language acquisition and its relationship to Universal Gram¬
mar. To appear in Applied Linguistics.
Schachter, Jacquelyn, and Marianne Celce-Murcia. 1977. Some reservations
concerning error analysis. TESOL Quarterly 11:441—51.
Schachter, Jacquelyn, and Beverly Hart. 1979. An analysis of learner production
of English structures. Georgetown University Papers on Languages and
Linguistics 15:18—75.
Schachter, Jacquelyn, and William Rutherford. 1979. Discourse function and
language transfer. Working Papers in Bilingualism 19:1—12.
 References 191

Schane, Sanford. 1973. Generative Phonology. Englewood Cliffs, N.J.: Prentice-

HalL
Schmidt, Annette. 1985. Young People’s Dyirbal. Cambridge: Cambridge Uni¬
versity Press.
Schmidt, Richard. 1987. Sociolinguistic variation and language transfer in pho¬
nology. In Interlanguage Phonology, ed. by Georgette loup and Steven
Weinberger. Rowley, Mass.: Newbury House.
1988. The role of consciousness in second language learning. Presented at
the Second Language Research Forum, University of Hawaii at Manoa.
Scholes, Robert. 1968. Phonemic interference as a perceptual phenomenon.
Language and Speech 11:86—103.
Schuchardt, Hugo. 1883a. Kreolische Studien III. Uber das Indoportugiesische
von Diu. Sitzungberichte der kaiserlichen Akademie der Wissenschaften zu
Wien (philosophisch-historische Klasse) 103:3—17.
1883b. Kreolische Studien IV. Uber das Malaiospanische der Philippinen.
Sitzungberichte der kaiserlichen Akademie der W issenschaften zu Wien
(philosophisch-historische Klasse) 105:111—50.
1884/1971. Slawo-deutsches und Slatvo-italienisches. Munich: Wilhelm Fink.
1891/1980. Indo-English. In Pidgin and Creole Languages: Selected Essays
by Hugo Schuchardt, ed. by Glenn Gilbert. London: Cambridge University
Press.
1909/1980. The lingua franca. In Pidgin and Creole Languages: Selected
Essays by Hugo Schuchardt, ed. by Glenn Gilbert. London: Cambridge
University Press.
Schumann, John. 1978. The Pidginization Process: A Model for Second Lan¬
guage Acquisition. Rowley, Mass.: Newbury House.
1979. The acquisition of English negation by speakers of Spanish: a review
of the literature. In The Acquisition and Use of Spanish and English as
First Languages, ed. by Roger Andersen. Washington, D.C.: TESOL.
1 984. Nonsyntactic speech in the Spanish-English basilang. In Second Lan¬
guages: A Cross-Linguistic Perspective, ed. by Roger Andersen. Rowley,
Mass.: Newbury House.
1986. Locative and directional expressions in basilang speech. Language
Learning 36:277—94.
Scollon, Ron, and Suzanne Scollon. 1981. Narrative, Literacy, and Face in
Interethnic Communication. Norwood, N.J.: Ablex.
Scovel, Thomas. 1969. Foreign accent, language acquisition, and cerebral dom¬
inance. Language Learning 19:245—54.
Scribner, Sylvia, and Michael Cole. 1981. The Psychology of Literacy. Cam¬
bridge, Mass.: Harvard University Press.
Sebeok, Thomas, ed. 1960. Style in Language.Cambridge: Mass.: MIT Press.
Seliger, Herbert, Stephen Krashen, and Peter Ladefoged. 1975. Maturational
constraints in the acquisition of second languages. Language Sciences
38:20-22.
Seiinker, Larry. 1969. Language transfer. General Linguistics 9:67— 92.
1972. Interlanguage. IRAL: International Review of Applied Linguistics
10:209-31.
Seiinker, Larry, Merrill Swain, and Guy Dumas. 1975. The interlanguage hy¬
pothesis extended to children. Language Learning 25:139—52.
192 References
Sey, K. A. 1973. Ghanaian English: An Exploratory Survey. London:
Macmillan.
Shapira, Rina. 1978. The non-learning of English: case study of an adult. In
Second Language Acquisition: A Book of Readings, ed. by Evelyn Hatch.
Rowley, Mass.: Newbury House.
Sharwood Smith, Michael. 1986. The competence/control model, crosslinguistic
influence, and the creation of new grammars. In Crosslinguistic Influence
in Second Language Acquisition, ed. by Eric Kellerman and Michael Shar¬
wood Smith. New York: Pergamon Press.
Shaughnessy, Mina. 1977. Errors and Expectations. New York: Oxford Uni¬
versity Press.
Shaw, Willard. 1981. Asian student attitudes towards English. In English for
Cross-Cultural Communication, ed. by Larry Smith. New York: St. Mar¬
tin’s Press.
Sheldon, Amy. 1977. The acquisition of relative clauses in French and English:
implications for language-learning universals. In Current Themes in Lin¬
guistics, ed. by Fred Eckman. New York: John Wiley.
Sherzer, Joel. 1974. Namakke, Sunmakke, Kormakke: three types of Cuna
speech event. In Explorations in the Ethnography of Speaking, ed. by Rich¬
ard Bauman and Joel Sherzer. London: Cambridge University Press.
Shuman, Amy. 1986. Storytelling Rights: The Uses of Oral and Written Texts
by Urban Adolescents. Cambridge: Cambridge University Press.
Silvestri, Domenico. 1977. La Teoria del Sostrato: Metodi e Miraggi. Naples:
Gaetano Macchiaroli.
Singler, John. 1988. The homogeneity of the substrate as a factor in pidgin/
creole genesis. Language 64:27—51.
Singleton, David. 1987. Mother and other tongue influence on learner French.
Studies in Second Language Acquisition 9:327-45.
Singleton, David, and David Little. 1984. A first encounter with Dutch: per¬
ceived language distance and language transfer as factors in comprehension.
In Language Across Cultures, ed. by Liam MacMathuna and David Sin¬
gleton. Dublin: Irish Association for Applied Linguistics.
Sjoholm, Kaj. 1976. A comparison of the test results in grammar and vocabulary
between Finnish- and Swedish-speaking applicants for English. In Errors
Made by Finns and Swedish-Speaking Finns in the Learning of English, ed.
by Hakan Ringbom and Rolf Palmberg. Abo, Finland: Department of En¬
glish, Abo Akademi. ERIC Report ED 122628.
Slobin, Dan. 1977. Language change in childhood and history. In Language
Learning and Thought, ed. by John MacNamara. New York: Academic
Press.
1982. Universal and particular in acquisition. In Language Acquisition: The
State of the Art, ed. by Eric Wanner and Lila Gleitman. Cambridge: Cam¬
bridge University Press.
1985. The child as a linguistic icon-maker. In Iconicity in Syntax, ed. by John
Haiman. Amsterdam: John Benjamins.
Smith, Larry, and John Bisazza. 1982. The comprehensibility of three varieties
of English for college students in seven countries. Language Learning
32:259-69.
Smith, Larry, and Khalilullah Rafiqzad. 1979. English for cross-cultural com¬
munication: the question of intelligibility. TESOL Quarterly 13:371—80.
 References 193

Snow, Catherine. 1981. English speakers’ acquisition of Dutch syntax. In Annals

of the New York Academy of Sciences, ed. by Harris Winitz. New York:
New York Academy of Sciences.
Snow, Catherine, and Marian Hoefnagel-Hohle. 1977. Age differences in the
pronunciation of foreign sounds. Language and Speech 20:357—75.
Sohn, Ho-Min. 1980. Syntactic and semantic interference for Korean adult
English learners. In Bilingual Education for Asian Americans, ed. by John
Koo and Robert St. Clair. Hiroshima: Bunka Hyoron.
Sorenson, A. P. 1967. Multilingualism in the northwest Amazon. American
Anthropologist 69:670-84.
Sridhar, S. N. 1981. Contrastive analysis, error analysis, and interlanguage:
three phases of one goal. In Contrastive Analysis and the Language Teacher,
ed. by Jacek Fisiak. Oxford: Pergamon Press.
Stauble, Ann-Marie. 1984. A comparison of a Spanish-English and a Japanese-
English second language continuum. In Second Languages: A Cross-
Linguistic Perspective, ed. by Roger Andersen. Rowley, Mass.: Newbury
House.
Steffensen, Margaret, Chitra Joag-Dev, and Richard Anderson. 1979. A cross-
cultural perspective on reading comprehension. Reading Research Quar¬
terly 15:10-29.
Stendahl, Cristina. 1972. A note on the relative proficiency in Swedish and
English as shown by Swedish university students of English. Moderna Sprdk
66/2:117-23.
Stenson, Nancy. 1974. Induced errors. In New Frontiers of Second Language
Learning, ed. by John Schumann and Nancy Stenson. Rowley, Mass.: New¬
bury House.
1981. Studies in Irish Syntax. Tubingen: Gunter Narr.
Stockwell, Robert, J. Donald Bowen, and John Martin. 1965. The Grammatical
Structures of English and Spanish. Chicago: University of Chicago Press.
Strick, Gregory. 1980. A hypothesis for semantic development in a second
language. Language Learning 30:155-76.
Sullivan, James. 1980. The validity of literary dialect: evidence from the por¬
trayal of Hiberno-English features. Language in Society 9:195—219.
Suter, Richard. 1976. Predictors of pronunciation accuracy in second language
learning. Language Learning 26:233—53.
Swan, Michael, and Bernard Smith, eds. 1987. Learner English: A Teacher’s
Guide to Interference and Other Problems. Cambridge: Cambridge Uni¬
versity Press.
Sweet, Henry. 1899/1972. The Practical Study of Languages. London: Oxford
University Press.
Taeschner, Traute. 1983. The Sun Is Feminine: A Study on Language Acqui¬
sition in Bilingual Children. Berlin: Spnnger-Verlag.
Tannen, Deborah. 1981. Indirectness in discourse: ethnicity as conversational
style. Discourse Processes 4:221—38.
1984. Spoken and written narrative in English and Greek. In Coherence in
Spoken and Written Discourse, ed. by Deborah Tannen. Advances in Dis¬
course Processes, Volume 12, ed. by Roy Freedle. Norwood, N.J.: Ablex.
Tannen, Deborah, and Piyale Oztek. 1981. Health to our mouths: formulaic
expressions in Turkish and Greek. In Conversation Routine, ed. by Florian
Coulmas. The Hague: Mouton.
194 References
Tarallo, Fernando, and John Myhill. 1983. Interference and natural language
in second language acquisition. Language Learning 33:55—76.
Tarone, Elaine. 1979. Interlanguage as chameleon. Language Learning 29: 181—
91.
1980. Some influences on the syllable structure of interlanguage phonology.
IRAL: International Review of Applied Linguistics 18:139—52.
Taylor, Archer. 1962. The Proverb. Hatboro, Pa.: Folklore Associates.
Taylor, Barry. 1975. The use of overgeneralization and transfer learning strat¬
egies by elementary and intermediate students of ESL. Language Learning
25:73-107.
Taylor, Linda, Alexander Guiora, John Catford, and Harlan Lane. 1969. The
role of personality variables in second language behavior. Comprehensive
Psychiatry 10:463—74.
Thomas, Jenny. 1983. Cross-cultural pragmatic failure. Applied Linguistics
4:91-112.
Thomason, Sarah. 1981. Are there linguistic prerequisites for contact-induced
language change? Paper presented at the 10th Annual University of
Wisconsin-Milwaukee Linguistics Symposium. ERIC Report ED 205054.
Thomason, Sarah, and Terrence Kaufman. 1988. Language Contact, Creoli-
zation, and Genetic Linguistics. Berkeley: University of California Press.
Thompson, Sandra. 1978. Modern English from a typological point of view.
Linguistische Berichte 54:19—35.
Thorndyke, Perry. 1977. Cognitive structures in comprehension and memory
of narrative discourse. Cognitive Psychology 9:77—110.
Tiffin, Brian. 1974. The intelligibility of Nigerian English. Unpublished Ph.D.
dissertation, University of London.
Todd, Loreto. 1983. English in West Africa. In English as a World Language,
ed. by Richard Bailey and Manfred Gorlach. Ann Arbor: University of
Michigan Press.
Traugott, Elizabeth, and Mary Pratt. 1980. Linguistics for Students of Litera¬
ture. New York: Harcourt, Brace and Jovanovich.
Trevise, Anne. 1986. Is it transferable, topicalization? In Crosslinguistic Influ¬
ence in Second Language Acquisition, ed. by Eric Kellerman and Michael
Sharwood Smith. New York: Pergamon Press.
Trudgill, Peter. 1984. On Dialect. New York: New York University Press.
1986. Dialects in Contact. Oxford: Basil Blackwell.
Upshur, John. 1962. Language proficiency testing and the contrastive analysis
dilemma. Language Learning 12:123—27.
Vachek, Josef. 1964. Written language and printed language. In A Prague School
Reader in Linguistics, ed. by Josef Vachek. Bloomington: Indiana University
Press.
Van Els, Theo, and Kees De Bot. 1987. The role of intonation in foreign accent.
Modern Language Journal 71:147—55.
Van Patten, Bill. 1984. Processing strategies and morpheme acquisition. In Uni¬
versals of Second Language Acquisition, ed. by Fred Eckman, Lawrence
Bell, and Diane Nelson. Rowley, Mass.: Newbury House.
Varadi, Tamas. 1983. Strategies of target language learners: message adjust¬
ment. In Strategies in Interlanguage Communication, ed. by Claus Faerch
and Gabriele Kasper. London: Longman.
Veronique, Daniel. 1984. The acquisition and use of French morphosyntax by
 References 195

native speakers of Arabic dialects. In Second Languages: A Cross-Linguistic

Perspective, ed. by Roger Andersen. Rowley, Mass.: Newbury House.
Vihman, Marilyn. 1985. Language differentiation by the bilingual infant. Jour¬
nal of Child Language 12:297—324.
Vildomec, Veroboj. 1963. Multilingualism. Leyden: A. W. Sythoff.
—
Walters, Joel. 1979a. Strategies for requesting in Spanish and English structural
similarities and pragmatic differences. Language Learning 29:277—93.
1979b. The perception of politeness in English and Spanish. In On TESOL
'79, ed. by Carlos Yorio, Kyle Perkins, and Jacquelyn Schachter. Wash¬
ington, D.C.: TESOL.
Wardhaugh, Ronald. 1970. The contrastive analysis hypothesis. TESOL Quar¬
terly 4:123-30.
Weinreich, Max. 1980. A History of the Yiddish Language. Chicago: University
of Chicago Press.
Weinreich, Uriel. 1953/1968. Languages in Contact. The Hague: Mouton.
Wexler, Kenneth, and Peter Culicover. 1980. Formal Principles of Language
Acquisition. Cambridge, Mass.: MIT Press.
White, Lydia. 1985. The “pro-drop” parameter in adult second language ac¬
quisition. Language Learning 35:47—62.
1987. Markedness and second language acquisition: the question of transfer.
Studies in Second Language Acquisition 9:261—85.
Whitman, Randall, and Kenneth Jackson. 1972. The unpredictability of con¬
trastive analysis. Language Learning 22:29—42.
Whitney, William. 1881. On mixture in language. Transactions of the American
Philological Association 12:5—26.
Whorf, Benjamin. 1956. Language, Thought, and Reality. Cambridge, Mass.:
MIT Press.
Willems, Nico. 1982. English Intonation from a Dutch Point of View. Dor¬
drecht: Foris Publications.
Winfield, Fairlie, and Barnes-Felfeli, Paula. 1982. The effects of familiar and
unfamiliar cultural context on foreign language composition. Modern Lan¬
guage Journal 66:373—78.
Wode, Henning. 1978. Developmental sequences in naturalistic L2 acquisition.
In Second Language Acquisition, ed. by Evelyn Hatch. Rowley, Mass.:
Newbury House.
1981. Learning a Second Language: An Integrated View of Language Ac¬
quisition. Tubingen: Gunter Narr.
1983a. Four early stages in the development of LI negation. In Papers on
Language Acquisition, Language Learning, and Language Teaching, ed.
by Henning Wode. Heidelberg: Julius Groos.
1983b. On the systematicity of LI transfer in L2 acquisition. In Papers on
Language Acquisition, Language Learning, and Language Teaching, ed.
by Henning Wode. Heidelberg: Julius Groos.
1983c. Language acquisition, pidgins, and language typology. In Papers on
Language Acquisition, Language Learning, and Language Teaching, ed.
by Henning Wode. Heidelberg: Julius Groos.
Wolfram, Walt. 1978. Contrastive linguistics and social lectology. Language
Learning 28:1—28.
Wolfson, Nessa. 1981. Compliments in cross-cultural perspective. TESOL
Quarterly 15:117-24.
196 References

Wright, Joseph. 1898. The English Dialect Dictionary. London: Oxford Uni¬
versity Press.
Young, D. I. 1974. The acquisition of English syntax by three Spanish-speaking
children. Unpublished M.A. thesis, University of California at Los Angeles.
Zobl, Helmut. 1980. The formal and developmental selectivity of LI influence
on L2 acquisition. Language Learning 30:43—57.
1982. A direction for contrastive analysis: the comparative study of devel¬
opmental sequences. TESOL Quarterly 16:169-83.
1983. LI acquisition, age of L2 acquisition, and the learning of word order.
In Language Transfer in Language Learning, ed. by Susan Gass and Larry
Seiinker. Rowley, Mass.: Newbury House.
1984. The wave model of linguistic change and the naturalness of interlan¬
guage. Studies in Second Language Acquisition 6:160—85.
1986. Word order typology, lexical government, and the prediction of mul¬
tiple, graded effects in L2 word order. Language Learning 36:159—83.
Language index

Since references to English occur on almost every page of the text, only
specific varieties of English (e.g., Nigerian English) are indexed here.
Afrikaans, 39 Dutch, 13, 32, 39, 41, 94, 95, 1 19,
Albanian, 10 142, 143, 144
Amharic, 39, 91 Dyirbal, 12, 13
Ancient Egyptian, 6
Andean Spanish, 13, 93, 94, 154
see also Ecuadorean Spanish Ecuadorean Spanish, 13, 14, 93,
Apache, 65, 81 147, 149
Apurina, 44 see also Andean Spanish
Arabic, 13, 26, 27, 30, 33, 38, 39, Egyptian Arabic, 30, 148
40, 44, 45, 50, 55, 66, 78, 80, English, see Hawaiian Pidgin En¬
82, 94, 99, 1 12, 113, 118, 119, glish, Hiberno-English, Indian
126, 148, 153, 160 English, Nigerian English, Phil¬
see also Classical Arabic, Egyptian ippine English
Arabic, Iraqi Arabic, Moroccan Eskimo, 32
Arabic
Farsi, see Persian
Bengali, 39 Finnish, 39, 40, 78, 79, 87, 91, 102,
Biblical Hebrew, 86 119
Bulgarian, 9, 10, 39 French, 1, 6, 7, 8, 10, 32, 33, 35,
Burmese, 39 39, 40, 42, 50, 53, 54, 55, 56,
71, 73, 77, 78, 79, 87, 91, 95,
Cambodian, 119 103, 104, 105, 114, 117, 118,
Cantonese, 118, 121, 122, 123 135, 136, 139, 141, 142,
Catalan, 162 145, 146, 147, 149, 158, 160,
Chinese, 1, 7, 16, 19, 20, 21, 22, 163
31, 37, 38, 39, 40, 73, 74, 75,
78, 83, 99, 116, 118, 119, 125,
127, 129, 131, 155, 161, 162 Gaelic, see Irish
see also Cantonese, Hui, Mandarin German, 1, 12, 18, 39, 41, 51, 52,
Classical Arabic, 30, 44, 45, 148 53, 54, 55, 67, 86, 87, 89, 94,
Czech, 39 95, 102, 105, 106, 107, 108,
110, 116, 117, 119, 120, 121,
Danish, 39 123, 139, 145, 146, 154, 155,
Dari, 39 158, 163
see also Persian Germanic languages, 40, 154

197
198 Language index

Greek, 12, 39, 61, 78, 83, 102, 106, Mongolian languages, 83
114, 163 Moroccan Arabic, 94
Hausa, 1 17
Hawaiian Pidgin English, 11, 92, 93, Navajo, 7, 65, 66
94 Nigerian English, 80, 81, 117
Hebrew, 37, 39, 52, 53, 66, 81, 96, Norwegian, 12, 39, 105
135, 146
see also Biblical Hebrew
Hiberno-English, 4, 13, 14, 27, 33, Persian, 32, 33, 44, 50, 68, 69, 80,
66, 70, 76, 77, 81, 85, 93, 109, 86, 99, 100, 101, 102, 103,
143, 144, 149, 150, 154 112, 114, 125, 126, 155, 162
Hindi, 39, 65, 69 Philippine English, 93
Hixkaryana, 44 Pilipino, 39
Hmong, 89 see also Tagalog
Hui, 83 Polish, 39
Hungarian, 38, 39, 40, 81 Portuguese, 9, 11, 39, 41, 102, 110,
122, 126
Igbo, 141, 142
Ilocano, 93 Quechua, 13, 32, 93, 147, 149
Indian English, 61, 81, 117, 148,
149, 159, 160
Indonesian, 39 Romance languages, 8, 9, 10, 33,
Iraqi Arabic, 30 40
Irish, 2, 4, 13, 14, 27, 44, 66, Rumanian, 8, 9, 10, 39, 40
76, 77, 86, 87, 109, 143, 149, Russian, 1, 2, 3, 19, 39, 44, 52, 53,
156 62, 67, 76, 86, 87, 114, 125,
Italian, 39, 40, 83, 87, 94, 126, 146, 156
156
Japanese, 1 1, 19, 20, 31, 32, 37, 38, Serbo-Croatian, 38, 39, 88
39, 40, 41, 44, 45, 50, 53, 54, Sindhi, 130
56, 61, 62, 63, 64, 65, 68, 91, Slavic languages, 9
92, 93, 97, 98, 99, 109, 1 12, Spanish, 1, 3, 4, 8, 9, 10, 11, 13, 16,
115, 116, 1 19, 122, 125, 17, 18, 19, 20, 21, 22, 26, 27,
131, 132, 148, 152, 155, 156, 30, 31, 32, 33, 34, 37, 39, 40,
163 41, 42, 46, 50, 61, 75, 76, 78,
79, 80, 82, 83, 87, 89, 90, 91,
Korean, 4, 19, 39, 62, 64, 65, 90, 93, 94, 95, 96, 98, 102, 104,
93, 115, 116, 119, 122, 156 105, 106, 109, 110, 114, 115,
Kurdish, 120 118, 119, 121, 122, 125, 126,
127, 129, 133, 139, 146, 147,
Lao, 39 149, 152, 153, 154, 155, 158,
Latin, 1, 7, 8, 77, 146 161, 162
see also Andean Spanish, Ecuador¬
Malagasay, 44 ean Spanish
Malay, 39, 144 Swahili, 39, 79
Mandarin, 74, 1 1 8 Swedish, 37, 39, 40, 78, 87, 91, 101,
Mande languages, 147 102, 103, 108, 109, 135, 136,
Mende, 72 145, 155
 Language index 199

Tagalog, 93, 101 Urdu, 39

see also Pilipino Ute, 88, 91
Tamil, 162
Thai, 39, 45, 53, 56, 82, 1 12, 118, Vai, 62, 102, 125
137, 148, 149 Vietnamese, 46, 97, 119, 122, 123
Turkic languages, 83
Turkish, 12, 13, 39, 40, 55, 83, 86, Yoruba, 117, 118, 162
88, 94, 97 Young People’s Dyirbal, see Dyirbal
Author index

In all cases of work involving co-authors, only the name of the first
author of the work appears in this index.
Adams, C., 119 Biber, D., 134
Adiv, E., 108 Bickerton, D., 43, 93, 94
Adjemian, C., 79, 104 Birdsong, D., 74, 103, 104
Ahukana, J., 141 Bley-Vroman, R., 108
Albert, E., 69 Bliss, A., 2, 14, 66, 149
Albert, M„ 28 Bloom, A., 73, 74, 75, 84
Altenberg, E., 126 Bloomfield, L., 9, 15, 118
Andersen, R., 11, 18, 21, 34, 82, 91, Blum, S., 81, 82
95, 96, 162 Blum-Kulka, S., 52
Anderson, J., 122 Bock, J., 31
Andrews, G., 24, 117 Bolinger, D., 5, 29, 119
Anthony, E., 79 Borden, G., 115
Appel, R., 24, 94, 152 Borkin, A., 53
Applegate, R., 55, 57 Brazil, D., 1 19
Ard, J., 24, 78, 80, 82, 84 Brennan, E., 158
Asher, J., 137 Briere, E., 113, 120
Au, T., 74 Broselow, E., 30, 118
Ausubel, D., 133 Brown, G., 70, 88
Brown, P., 49, 50, 52, 69
Bailey, C., 9 Brown, R., 20, 21
Bailey, N., 21 Busch, D., 132
Bamgbose, A., 80, 81 Butterworth, G., 18
Bammesberger, A., 14 Bynon, T., 9, 24
Bansal, R., 117, 160
Barkowski, H., 55 Callary, R., 9
Barry, M., 149 Cancino, H., 106
Bartelt, H., 65, 66, 81 Carrell, P., 50, 59
Bartlett, F., 60 Carroll, J., 25, 132
Bates, E., 46, 50, 53, 87 Celce-Murcia, M., 75
Bebout, L., 126 Chafe, W., 75
Beebe, L., 148 Chambers, J., 9
Beeman, W., 56 Chaudron, C., 29
Benson, B., 123 Chiang, T., 118
Bentahila, A., 146 Chomsky, N., 22, 29, 43
Berlin, B., 47 Clahsen, H., 88, 94, 1 10

201
202 Author index
Clark, E., 80, 8 I Friedrich, P., 143
Clark, H., 26, 47, 81, 159, 161 Fries, C., 6, 15, 16, 17, 22, 30, 161
Clarke, M., 73 Fromkin, V., 5
Coe, N„ 162
Cohen, A., 53 Gandour, J., 118
Comrie, B., 4, 29, 43, 46, 47, 86, Garvin, J., 149
101 Gass, S., 5, 46, 87, 101, 102, 103,
Connor, U., 67 111, 143, 156
Cooper, W., 143 Gazdar, G., 29, 97
Corbett, E., 56 Gelb, I., 125
Corder, S., 25 Genesee, F., 135
Coulmas, F„ 29, 53, 54, 125 Gilbert, G., 11, 34
Cowan, J., 126 Giles, H., 159
Crystal, D., 56 Gilsan, E., 87
Cummins, J., 134 Givon, T., 29, 41, 86, 88, 90, 91,
Cutler, A., 117 93, 111
Godard, D., 50
Dahl, 6., 62, 107 Gonzo, S., 30
Dawkins, R., 12, 83 Granfors, T., 87, 91
Dechert, H., 16, 112, 162 Greenberg, C., 122
De Freine, S., 149 Greenberg, J., 43, 44, 46, 96, 123
Dehghanpisheh, E., 68 Gregg, K., 3
Derbyshire, D., 44 Gudschinsky, S., 163
De Villiers, J., 21, 43, 81 Guiora, A., 73, 131
Dickerson, L., 42 Gumperz, J., 10, 50, 146
Di Pietro, R., 29
Dulay, H., 14, 20, 21, 22, 23, 24, Hagege, C., 97, 145
34, 82, 136, 149 Hakuta, K., 20, 24, 25
Dundes, A., 59, 60 Hamill, J., 72
Duskova, L., 18 Harley, B., 139, 150
Harris, J(esse), 24
Eckman, F., 121, 122, 123 Harris, J(ohn), 14
Edwards, M., 123 Hartmann, R., 32
Eggington, W., 62, 64 Hatch, E., 21, 47, 137
Ekman, P., 57 Haugen, E., 12, 17, 24
Ellis, R., 5, 28, 42, 153 Hawkins, J., 43, 44, 86, 96
Emeneau, M., 10 Heath, S., 134
Eubank, L„ 44, 98, 110 Hecht, B., 121, 123, 152
Henry, P., 13, 14, 66, 77, 109, 143
Fantini, A., 37, 82, 139, 145 Herman, L., 2
Fathman, A., 21, 34 Hillocks, G., 68
Felix, S., 1 8 Hinds, J., 56, 62, 63, 64, 65, 70
Ferguson, C., 11, 54 Hockett, C., 83
Fillmore, C., 75, 76 Hocking, B., 79
Fillmore, L., 136 Holmes, G., 79
Filppula, M., 66, 87 Homburg, T., 68
Flege, J., 113, 114, 115, 138 Hornby, A., 55
Flynn, S., 29, 44, 98, 99, 162 House, J., 5 1
Foss, D., 26, 47, 72, 73, 159 Huang, J., 19
Fox, B., 101 Huebner, T., 21, 89, 90
 Author index 203

Huerta, A., 146 Lee, W., 17, 19, 42

Hutchins, E., 72 Le Page, R., 7, 139, 144
Hyltenstam, K., 102, 103, 108, 109 Leslau, W., 10
Hyman, L., 1 22 Leung, K., 1 1 8
Levin, H., 68
Ibrahim, M., 38, 126, 127 Levinson, S., 48, 50, 54, 70
Ijaz, I., 80 Levi-Strauss, C., 59
loup, G., 102, 103, 104, 115, 128, Li, C., 83, 88, 97
137, 158 Liberman, A., 114
Itoh, H., 19 Lightbown, P., 142
Limpet, L., 135
Jackson, J., 7, 145 Linnarud, M., 135, 136
James, C., 29, 40, 47 Littlewood, W., 41, 130
Jansen, B., 13, 24, 94 Liu, L., 74
Jespersen, O., 15, 76, 102 LoCoco, V., 94
Johansson, F., 23 Long, M., 19
Jordens, P., 108, 143 Loveday, L., 54, 57, 65, 70, 119,
Joseph, B., 9, 10, 101 155, 159
Lowenburg, P., 149
Kachru, B., 64, 148, 159 Lujan, M., 13, 93, 94, 96, 152
Kachru, Y., 65, 69 Lukatela, G., 125
Kalmar, I., 158 Lyons, J., 62, 75, 76, 84
Kamratowksi, J., 126
Kaplan, R., 16, 62, 64 Maddieson, L, 46, 120
Kasper, G., 51, 52, 67, 119 Major, R„ 123, 133
Keenan, E., 44, 46, 86, 99, 100, Mandler, J., 59, 60
101, 103, 111 Mann, V., 115
Keller-Cohen, D., 119, 152 Marckwardt, A., 24, 114
Kellerman, E., 25, 26, 134, 142, Masny, D., 136
143, 144 McLaughlin, B., 41, 137, 140
Kelly, L., 15 Meillet, A., 8
Kempf, M., 34 Meisel, J., 11, 94
Kinneavy, J., 56 Meo Zilio, G., 83
Kintsch, W., 59 Mervis, C., 47
Klein, W., 89, 90 Mohan, B., 67, 70
Kleinmann, H., 37, 131 Moravcsik, E., 121
Koch, B., 66 Moulton, W., 115, 116, 121
Kohn, K., 151 Muhlhausler, P., 10, 11, 24
Krashen, S., 3, 22, 26, 27, 34, 37, Muller, M., 8, 23
82, 137, 140, 143 Muysken, P., 11, 13, 91, 92, 93, 94,
Kuno, S., 97, 98, 103 97, 147, 149

Labov, W., 29, 147 Nagara, S., 11, 92

Lado, R., 6, 15, 16, 17, 29, 115, Nelson, K., 80
160, 161 Neufeld, G., 137
Lakoff, G., 29, 46, 72, 81
Larsen-Freeman, D., 68 Obilade, T., 50
Lass, R., 77 Odlin, T., 3, 93, 94, 95, 140, 144,
Lazard, G., 100 150
Lazzerini, L., 146 Oller, J., 34, 126
204 Author index
Olshtain, E., 53, 66 Schmidt, A., 12, 13
Olson, L., 137 Schmidt, R., 148, 150
Ong, W., 127 Scholes, R., 114
Orth, J., 158 Schuchardt, H., 10, 11, 18, 24
Osgood, C., 81, 83, 143 Schumann, J., 18, 19, 24, 33, 34, 90,
Oxford, R., 40 106, 107, 109
Oyama, S., 137 Scollon, R., 57, 132
Scovel, T., 138
Palmer, H., 15 Scribner, S., 62, 72, 125, 134, 135
Pandharipande, R., 65 Sebeok, T., 56
Paulston, C., 132 Seliger, H., 137
Pavlovitch, M., 145 Seiinker, L., 18, 96, 104, 113, 149
Peck, S., 1 8 Sey, K., 147
Penn, J., 73, 84 Shapira, R., 18
Pica, T., 31, 162 Sharwood Smith, M., 29
Pienemann, M., 87, 88, 162 Shaughnessy, M., 68
Pike, E., 132 Shaw, W., 149
Pike, K., 31, 51 Sheldon, A., 104
Platt, J., 64, 81, 144, 148 Sherzer, J., 55
Poplack, S., 139, 146 Shuman, A., 134
Propp, V., 59 Silvestri, D., 7
Purcell, E., 112, 113, 132, 137 Singler, J., 10, 12, 24, 101, 102, 147
Piirschel, H., 1 19 Singleton, D., 41, 142
Purves, A., 69 Sjoholm, K., 78, 82
Slobin, D., 11, 40, 87, 88, 143
Rao, R., 64 Smith, L., 115, 159
Ravem, R., 18, 105, 106 Snow, C., 95, 96, 137
Redlinger, W., 139, 145 Sohn, H., 156
Reisman, K., 54 Sorenson, A., 7, 145
Richards, J., 18, 48, 56, 64 Sridhar, S., 24, 35
Ringbom, H., 5, 24, 34, 37, 79, 154 Stauble, A., 109
Rintell, E., 50, 118, 119 Steffensen, M., 61
Robins, R., 8, 22 Stendahl, C., 135
Rona, J., 9 Stenson, N., 18, 66
Ronjat, J., 18, 145 Stockwell, R., 17, 18, 29, 30
Rosansky, E., 21 Strick, G., 80
Rosch, E., 47, 73 Sullivan, J., 2
Rutherford, W., 92 Suter, R„ 112
Ryan, E., 158 Swan, M., 162
Sweet, H., 15, 77, 78
Sacks, H., 55
Sajavaara, K., 16, 162
Sampson, G., 125 Taeschner, T., 12, 139, 145
Sandfeld, K., 9 Tannen, D., 55, 61
Sanford, A., 31 Tarallo, F., 102
Sato, C., 122 Tarone, E., 122, 151
Scarcella, R., 53, 67 Taylor, A., 55
Schachter, J., 19, 24, 31, 37, 38, 99, Taylor, B„ 133, 134
109, 131, 154, 157 Taylor, L., 131
Schane, S., 116, 128 Thomas, J., 51, 52
 Author index 205

Thomason, S., 7, 12, 13, 28, 38, Walters, J., 50

155, 156 Wardhaugh, R„ 19, 35
Thompson, S., 86, 87 Weinreich, M., 146
Thorndyke, P., 59 Weinreich, U., 12, 17, 24, 82, 115
Tiffin, B„ 117 Wexler, K., 43
Todd, L., 10, 141 White, L., 24, 42, 104
Trevise, A., 87, 139 Whitman, R., 23
Trudgill, P., 9, 113, 138 Whitney, W., 8, 9, 23, 26, 82
Whorf, B., 71, 72, 73
Willems, N., 119
Upshur, J., 17 Winfield, F., 61
Wode, H., 106, 107, 108, 109, 110,
111, 152, 155
Vachek, J., 125 Wolfram, W., 30
Van Els, T., 119 Wolfson, N., 55
Van Patten, B., 21 Wright, J., 14
Varadi, T., 81 Young, D., 106
Veronique, D., 96
Vihman, M., 139, 145, 146 Zobl, H., 9, 34, 86, 87, 92, 93, 94,
Vildomec, V., 141 95, 96, 108, 109, 138, 139, 155
Subject index

acquisition/learning distinction, 3 categories, 31, 45, 46, 89, 97, 108

adverbial clauses, 98 center-embedding, 97, 98
adverbs, 94, 96 child language, see under language
age and transfer, 136—40, 145, 146, acquisition
150, 152 cleft sentences, 66
see also language acquisition code-switching, 7, 13, 24, 139, 140,
alphabets, 31, 124, 125, 126 146
anxiety, 22, 131 cognate vocabulary, 26, 41, 77, 78,
apologies, 31, 37, 53, 54 83, 114, 116, 118, 126, 133,
aptitude, foreign language, 40, 42, 135, 142, 152, 153, 160
132 false cognates, 31, 79, 147
articles, 2, 8, 9, 10, 18, 19, 33, 34, cohesion, see topic continuity
36, 42, 85, 152 comment, 88, 89, 90
asymmetrical patterns, 17, 30, 121, see also topic
127 comprehensibility, 97, 101, 117,
attitudes, 7, 47, 71, 118, 119, 157, 159, 160
158, 160 contrastive analysis, 2, 3, 15, 16, 17,
attrition, see borrowing transfer 18, 19, 28-34, 35-42, 47, 129,
avoidance, 37, 99, 103, 131, 156 141, 142, 153
awareness of language, 1, 3, 4, 7, definition of, 2
95, 129, 140-50, 152, 153, see also language distance; predic¬
156, 160 tions; typology
contrastive rhetoric, 16, 31, 48—70
behaviorism, 15, 23, 25, 26 copula, 2, 3, 17, 18, 19, 20, 21, 33,
see also habit formation 41, 109
bibliographies, 16, 112, 162 creole, 10, 11, 12
bidirectionality, 156 see also pidgin
see also asymmetrical patterns cross-sectional studies, 21, 109
borrowing transfer, 12, 13, 24, 69, developmental processes, 68, 80, 81,
74, 75, 83, 145, 156 83, 87, 95, 108, 111, 123, 124,
see also substratum transfer 127
bound morphemes, 82, 83 and errors, 20, 155
bound morphology, 86, 152, 155 and sequences, 20, 21, 22, 23, 45,
see also morphological case 96, 105, 106, 107, 108, 109,
branching direction, 97, 98, 99 110
devoicing, 121, 122, 123
caiques, 37, 38, 77, 141, 142, 143, dialects, 4, 14, 56, 116, 121, 130,
144 138, 149, 159

207
208 Subject index

dialects (cont.) 136-40, 145, 146, 149, 150,

differences between, 30, 41, 77 152, 156
and languages 9, 1 13 by monolingual children, 3, 11,
40, 43, 68, 80, 87, 94, 104
error analyses, 18, 19, 42, 133, 134 language classifications, see historical
linguistics; typology
false cognates, see under cognate language distance, 1, 32, 37, 40, 41,
vocabulary 45, 141, 142, 153
focusing, 144, 145, 146, 147, 148, language mixing, 6, 7, 8, 9, 139,
149, 152 145, 146, 158
Foreigner Talk, 11 left dislocation, see under word
order
Generalized Phrase Structure Gram¬ linguistic relativism, 46, 47, 71, 72,
mar, 29 73, 74, 75, 81, 83, 84
genetic classification, see tree model literacy, 31, 68, 124, 125, 126, 134,
genitive, 75, 76, 96, 100, 101, 103 135, 136, 147, 154, 155
glottal stop, 123 literature, 59, 64, 146, 149, 154,
grammatical gender, 45, 73 155, 163
greetings, 54, 55 logic, 72
longitudinal studies, 20, 21, 1 10,
habit formation, 15, 22, 23, 25, 26, 111, 155
116
see also behaviorism
handwriting, 125 markedness, 121
historical linguistics, 4, 6—14, 24, 26, memorization, 132
28, 32 metalinguistic awareness, see aware¬
hypercorrection, 38, 114 ness of language
mimicry, 1, 2, 114, 115, 132
idealization, 30, 130
misinterpretation, 36, 38, 49, 57, 60,
idioms, 33, 37, 141, 142, 143, 144 61, 72
implicational relations, 44, 45, 96, morphological case, 76, 83, 86
100, 101, 103, 123
multilingualism, 7, 27, 41, 141, 142,
individual differences, 5, 42, 1 10, 146
115, 130-36
interference, 12, 24, 26
interlanguage, 113
interlingual identification, 113, 114, narratives, 31, 58, 59, 60, 61
116, 125, 138 nationalism, 149
see also substratum transfer, defi¬ negation, 21, 46, 47, 72, 85, 97,
nition of 104-10, 123, 133
intonation, 56, 118, 119, 130 negative face, see under politeness
introspection, 29 negative politeness, see under
politeness
language acquisition, 10, 15, 17, 19, negative transfer, 36, 37, 38
21, 22, 44, 59, 106, 123, 134, definition of, 26
154, 155 determination of, 28, 32, 33
by bilingual children, 12, 18, 20, negators, 104, 105, 152, 156
34, 37, 82, 87, 94, 104, 105, see also negation
106, 107, 108, 111, 119, 120, neurolinguistics, 28
 Subject index 209

overgeneralization, 18, 28, 67, 68, pronouns, 18, 41, 42, 76, 77, 79,
80, 81, 95, 96, 108, 126, 127, 91, 93, 95
133, 134, 162 and politeness, 56, 65
overproduction, 36, 37 relative, 99, 100, 104
resumptive, 32, 33, 45, 100, 101,
102, 103, 134, 147, 153, 155
paragraphs, 31, 62, 63, 64, 91 proposition, 58, 71, 72, 75
paralanguage, 6, 56, 57, 118 proverbs, 55
pedagogical materials, 4, 16, 136, see also idioms
146, 147, 160, 161, 162
perceptual errors, 38, 46, 114, 115, questions, 18, 50, 51, 71, 93, 98,
1 18 119, 133
personality, 130, 131, 132, 136, see also requests
137, 158
phonemes relative clauses, 21, 32, 33, 37, 45,
cross-linguistic frequency of, 120 97-104, 110, 111, 121, 131,
and written language, 124, 125, 134, 147, 152, 153, 155, 156
126 and relativized positions, 100,
phonemic systems, 113, 114, 115, 101, 102, 103, 104
116 relativism, see linguistic relativism
pidgin, 10, 11, 12, 92, 93, 102 relativized positions, see under rela¬
see also creole tive clauses
politeness, 31, 47, 48, 49-57, 58, requests, 48, 49, 50, 51, 52, 53, 54,
64, 65, 67, 68, 155, 156, 160 57, 67
negative, 49 resumptive pronouns, see under
positive, 49 pronouns
positive face, see under politeness right dislocation, see under word
positive politeness, see under order
politeness
positive transfer, 36 sample sizes, 14, 47, 151
definition of, 26 Sapir-Whorf Hypothesis, see linguis¬
determination of, 28, 33, 34 tic relativism
postpositions, 44, 97 schooling, 14, 22, 34, 67, 110, 132,
postverbal negation, see negators 144, 146, 147, 152
pragmatics, 48 segmental errors, types of, 115, 116,
predictions, 17, 18, 19, 20, 33, 34, 117
35-42, 44, 92, 95, 101, 103, semantic case, 75, 76, 77, 96
105, 106, 112, 120, 130, 131, silences, 57, 65
148, 157 simplification, 11, 18, 19, 28, 41, 42
prepositions, 18, 42, 75, 76, 77, 80, speech acts, 50, 109
100, 109, 153 see also apologies; greetings; prov¬
stranded, 104, 153 erbs; questions; requests; tele¬
and word order, 44, 96 phone calls
prestige, 13, 30, 140, 146, 147, 148, speech areas, 9, 10
149 speech rhythms, 117, 118, 119
preverbal negation, see negators spelling, 33, 38, 126, 127
product/process distinction, 161 see also writing systems
proficiency, 24, 39, 40, 93, 94, 109, Sprachbund, see speech areas
130, 133, 134, 146, 153, 154 standard languages, 140, 144, 145
210 Subject index
standard languages (cont.) 97, 98, 110, 118, 121, 122,
see also focusing; idealization; 123, 127, 142, 153, 156
prestige
stereotypes, 1, 2, 132, 158
underproduction, 36, 37, 99
stress patterns, 38, 117, 118
see also avoidance
structuralism, 22 Universal Grammar, 43, 44, 92
structure, definition of, 31
see also universals
style, 55, 56, 62, 65, 69 universals, 4, 11, 19, 20, 21, 22, 43—
subordinate clauses, 68, 86, 94, 95, 47, 48, 49, 50, 51, 54, 55, 57,
98, 108, 110 58, 60, 69, 72, 76, 80, 81, 82,
see also adverbial clauses; avoid¬ 83, 90, 91, 92, 95, 97-104,
ance; negators; relative clauses 107, 110, 111, 119, 122, 123,
substitutions, 37, 38 127, 143, 159, 161
substratum transfer variation, 14, 155
and borrowing transfer, 12, 13, 14 cross-linguistic, 8, 43, 44, 47
definition of, 27 individual, 5, 42, 110, 129-36
suprasegmental patterns, 117, 118, regional, 30, 149
119 social, 30, 147, 148
see also intonation; speech stylistic, 69
rhythms; stress patterns see also typology
syllable structure, 117, 122, 123
wave model, 9, 10
word associations, 17, 73, 81, 143
telephone calls, 50 word order, 10, 1 1, 18, 33, 36, 38,
testing, 17 76, 85-97, 98, 103, 105, 110,
textbooks, see pedagogical materials 111, 1 12, 152
tone languages, 34, 45, 118, 137, basic, 44, 45, 86, 87, 88, 91, 92,
138 93, 94, 95, 138, 139
topic, 31, 58, 62, 63, 64, 65, 66, 88, and discourse, 31, 88, 89, 90, 91,
89, 90, 97 138, 139
topic continuity, 89, 90, 91, 106 right and left dislocation, 90, 9 1
transfer, definition of, 1, 25—28 see also negation
see also borrowing transfer; sub¬ writing systems, 27, 36, 124-27,
stratum transfer 128, 135, 154
transfer of training, 18, 34, 67, 134 see also alphabets; literacy;
tree model, 8, 9 spelling
turn taking, 55, 119
typology, 45, 46, 47, 65, 85, 86, 87, zero anaphora, 90, 91, 92, 106
 Language Transfer
Cross-linguistic influence
in language learning
This book reconsiders a question that many language teachers
and educational researchers have discussed: how' much influ¬
ence a learner’s native language can have in making the
acquisition of a new language easy or difficult. Transfer has
long been a controversial issue, but many recent studies sup¬
port the view that cross-linguistic influences can have an im¬
portant impact on second language acquisition.

Odlin analyzes and interprets research, showing many ways

in which similarities and differences between languages can
influence the acquisition of grammar, vocabulary, and pro¬
nunciation. He provides a detailed look at work on other
areas important for the study of transfer, including discourse,
individual variation, and sodolinguistic factors. Language
teachers, applied linguists, and educational researchers will
find this volume highly accessible and extremely valuable to
their work.

in the fields of second language

" There is nothing like this hook
acquisition and language teaching methodology, and there is
a real need for it - as a useful beginning text for students of
second language acquisition and as background knowledge to
help inform and guide classroom practice of language teacher
trainees."
Jacquelyn Schachter,
University of Southern California
Terence Odlin is Assistant Professor in the Department of
English at The Ohio State University, where he teaches
courses in linguistics and in English as a second language.

ISBN 0-S21-371La-k

780521 371681