Vol 24 4 PDF

TABLE OF CONTENTS
TESOL QUARTERLY Volume 24, Number 4 ❑ Winter 1990
A Journal for Teachers of English to Speakers of Other Languages

and of Standard English as a Second Dialect
Editor
SANDRA SILBERSTEIN, University of Washington
Guest Editor
BERNARD SPOLSKY, Bar-Ilan University
Review Editor
HEIDI RIGGENBACH, University of Washington
Brief Reports and Summaries Editor
GAIL WEINSTEIN-SHR, University of Massachusetts at Amherst/
Temple University
Research Issues Editor
GRAHAM CROOKES, University of Hawaii at Manoa
Assistant Editor
CAROLYN MARGON, University of Washington
Editorial Assistant
MAUREEN P. PHILLIPS, University of Washington
Editorial Advisory Board
Roberta G. Abraham Miriam Eisenstein
Iowa State University New York University
Margie S. Berns Liz Hamp-Lyons
Purdue University University of Colorado at Denver
Joan Eisterhold Carson Mary McCroarty
Georgia State University Northern Arizona University
Ruth Larimer Cathcart Thomas Ricento
Monterey Institute of International Studies Japan Center for Michigan Universities/
Graham Crookes Central Michigan University
University of Hawaii at Manoa May Shih
Catherine Doughty San Francisco State University
The University of Sydney James W. Tollefson
Patricia A. Dunkel University of Washington
The Pennsylvania State University Vivian Zamel
University of Massachusetts at Boston
Credits
Advertising arranged by Helen Kornblum, TESOL Central Office, Alexandria, Virginia
Typesetting, printing, and binding by Pantagraph Printing, Bloomington, Illinois
Design by Chuck Thayer Advertising, San Francisco, California
VOLUMES MENU
TESOL QUARTERLY
CONTENTS To print, select PDF page
nos. in parentheses
ARTICLES
Introduction to a Colloquium: The Scope and Form
(10-17)
of a Theory of Second Language Learning 609
Bernard Spolsky
“Conscious” versus “Unconscious” Learning 617 (18-35)
Barry McLaughlin
The Competence of Processing: Classifying Theories
of Second Language Acquisition 635 (36-49)
Ellen Bialystok
The Least a Second Language Acquisition Theory
Needs to Explain 649 (50-67)
Michael H. Long
Extending the Scope of the Acculturation/Pidginization
Model to Include Cognition 667 (68-85)
John H. Schumann
Learning Without Rules: PDP and a Resolution of (86-97)
the Adult Language Learning Paradox 685
M. E. Sokolik
The Need for an Integrated Theory: Connecting Modules 697 (98-117)
Evelyn Hatch, Yasuhiro Shirai, and Cheryl Fantuzzi
REVIEWS
Recent Publications on Classroom Research 717
Observation in the Language Classroom
Dick Allwright
Second Language Classrooms: Research on Teaching and Learning
Craig Chaudron
The Classroom and the Language Learner:
Ethnography and Second-Language Classroom Research
Leo van Lier
Reviewed by Courtney B. Cazden
Writing Across Languages: Analysis of L2 Text
Ulla Connor and Robert B. Kaplan
Reviewed by Melvin R. Andrade, Jr.
BOOK NOTICES 729
The Literature-Enriched Curriculum
Donna Brinton, Guest Editor
Literacy Through Literature, Terry D. Johnson and Daphne R. Louis
(Christine Holten)
The Inward Ear: Poetry in the Language Classroom, Alan Maley
and Alan Duff (Foong Ha Yap)
Volume 24, Number 4 ❑ Winter 1990
Longman American Structural Readers (Annette C. Sheehan)

Forestville Tales: InternationaI Folk Stories, Aaron Berman (Fang-Lian Liao)
Outsiders: American Short Stories for Students of ESL, Jean S. Mullen
(Lyn M. Repath)
At the Door: Selected Literature for ESL Students, Sandra McKay
and Dorothy Petitt (Martha Dunn)
Changes: Readings for ESL Writers, Jean Winthrow, Gay Brooks,
and Martha Clark Cummings (Frank Beyer)
Literary Portraits: An Anthology of Modern American Prose and Poetry for
Students of English, J. Donnie Snyder (Melinda Kodimer)
BRIEF REPORTS AND SUMMARIES
DIF in Native Language and Gender Groups
in an ESL Placement Test 741
Antony John Kunnan
The Effect of Syntax, Speed, and Pauses on Listening Comprehension 746
Eileen K. Blau
Evaluations of Essay Prompts by Nonnative Speakers of English 753
Malcolm Hayward
THE FORUM
Two Commentaries on Brian K. Lynch's
“A Context-Adaptive Model of Program Evaluation” 759
A Reader Reacts . . .
Mark W. Meinke
A Context-Restrictive Model for Program Evaluation?
John W. Swales
The Author Responds . . .
Brian K. Lynch
A Logical Difficulty of the Parameter Setting Model
Yoshinori Sasaki
Research Issues
The Use of Multiple t Tests in Language Research
James Dean Brown
Multiple t Tests: Some Practical Considerations
Andrew F. Seigel
Information for Contributors 777

Editorial Policy
General Information for Authors
Publications Received 781
Publications Available from the TESOL Central Office 785
Cumulative Index for the TESOL Quarterly,
Volumes 23 and 24, 1989-1990 791
TESOL Membership Application 823
TESOL QUARTERLY 603

TESOL QUARTERLY
Editor's Note
■ On behalf of the TESOL Quarterly, I am pleased to congratulate the

winners of the 1990-91 Fred W. Malkemas Prize. Alastair Pennycook’s
article “The Concept of Method, Interested Knowledge, and the Politics of
Language Teaching” (TESOL Quarterly, Vol. 23, No. 4, December 1989)
shares the award with James Lydon’s “From Gattegno to Freire: Musings
on a Decade Gone By” (NYS TESOL Selected Conference Proceedings,
1989). The annual prize is conferred by the American Language Institute
of New York University’s School of Continuing Education for articles in
English that make a contribution to our knowledge of teaching and
classroom practice.
This issue of the Quarterly is the first to be devoted to a single topic. I
thank Bernard Spolsky (Bar-I1an University) who served as guest editor for
a series of articles representing a diversity of current perspectives on
second language acquisition theory. Thanks also to Carolyn Margon for
serving as Assistant Editor and to Maureen Phillips and Linda Stolfi for
work on the cumulative index, which appears in this issue.
This issue premiers a new subsection of The Forum, Research Issues,
edited by Graham Crookes. In each instance, an aspect of qualitative or
quantitative research will be addressed, frequently from somewhat
different perspectives. Although contributions will typically be solicited,
readers are encouraged to submit topic suggestions and/or make known
their availability as contributors by contacting Graham Crookes at the
address provided in the Information for Contributors section of the
Quarterly.
605
In this Issue
Bernard Spolsky’s interpretive essay examines the scope and form of a

theory of second language learning, documenting the variety of current
approaches to modeling the learning process. His discussion serves as an
introduction to six full-length articles, each reflecting a different
perspective on this topic.
Also in this issue:
● Reviews: In a review of recent publications on classroom research,
Courtney Cazden reviews Dick Allwright’s Observation in the Lan-
guage Classroom, Craig Chaudron’s Second Language Classroom, and
Leo van Lier’s The Classroom and the Language Learner; Melvin
Andrade reviews Ulla Connor and Robert B. Kaplan’s Writing Across
Languages: Analysis of L2 Text.
● Book Notices: The literature-enriched curriculum is the focus of this
issue’s Book Notices section, for which Donna Brinton has been the
guest editor.
• Brief Reports and Summaries: Antony Kunnan identifies potential
sources for differential item functioning among four native language
groups and the two gender groups; Eileen Blau investigates the effect
of syntax, speed, and pauses on listening comprehension; and Malcolm
Hayward reports on nonnative speakers’ evaluations of specific essay
prompts, largely in terms of perceived ease of answering the question.
● The Forum: Brian Lynch responds to comments by Mark Meinke and
John Swales on his recent TESOL Quarterly article, “A Context-
Adaptive Model of Program Evaluation”; in the new subsection
“Research Issues,” James Dean Brown and Andrew Siegel comment on
the use of multiple t tests; Yoshinori Sasaki identifies a logical difficulty
of Universal Grammar theory’s parameter setting model.
Sandra Silberstein
606 TESOL QUARTERLY

TESOL QUARTERLY, Vol. 24, No. 4, Winter 1990
Introduction to a Colloquium:
The Scope and Form of a Theory
of Second Language Learning
BERNARD SPOLSKY
Bar-Ilan University
This paper introduces a colloquium on the scope and form of a

theory of second language learning. It argues for the value of a
general theory, considers the relation of theory to practice, and
argues that the papers that follow—by McLaughlin; Bialystok;
Long; Schumann; Sokolik; and Hatch, Shirai, and Fantuzzi;—point
to a field where new and competing paradigms are being
explored.
The history of language teaching is sometimes written as though

it follows a simple progression: Every few years, a new theory
appears to drive out the old one and furnish a new method. Analysis
shows that this view is flawed. First, new theories do not generally
succeed in replacing their predecessors but continue to coexist with
them uncomfortably; second, theories have not usually been
realized in new methods but have more mundanely provided
ideological underpinnings, intellectual backing, or advertising
slogans for some newly discovered teaching panacea; and finally,
teaching practice in the foreign language classroom has not usually
been derived from the new methods but continued as a loosely
eclectic amalgam of old habits with new garnishes. Recent study of
the reality of language teaching has in fact shown that it is often
economic or political factors rather than theoretical ones that have
determined school policy and classroom practice (Phillipson, 1990;
Richards, 1984).
Nor do the theories offer a firm and unchanging basis. For a short
time in the 1950s and early 1960s the harmony of the pact between
structural linguists and Skinnerian psychologists provided justifica-
tion for the Audiolingual Method. By the 1970s the hegemony of
transformational grammar and cognitive psychology empowered
much of second language acquisition research. More recently we
609
seem to have entered a period of uncertainty, of challenges stronger
than the proposed answers.
One of the big disappointments in the study of language has, for
some people, been the failure of linguists and psychologists to form
the genuine interdisciplinary team that, it was once hoped, would
solve the outstanding problems. George Miller (1990), who has long
been at the forefront of these attempts at interdisciplinary commu-
nication and cooperation, has recently proposed a convincing expla-
nation of the breakdown: the fact that linguists and psychologists
have different theories of explanation. For a linguist, the proposal of
a simpler set of rules is itself an explanation; for the psychologist,
such a theory is only the starting point, requiring explanation in
terms of cause and effect, or stimulus and response. This topic is
discussed in detail in McLaughlin’s paper in this issue of the TESOL
Quarterly.
If linguists and psychologists have difficulty in agreeing on a
theory of language and language learning, how much more difficult
things become in the related field of second language learning,
where we have a traditional concern to consider not just the
explanatory power of the theory but also its relevance to second
language pedagogy (Spolsky, 1968). It is no wonder then that there
has been suspicion of the usefulness of theory, a suspicion no doubt
compounded by the harm done by some of the application of
theory to second language teaching.
There is much more than a theory of second language learning
involved in understanding language teaching. Some years ago,
Spolsky (1968) suggested a wider model in which learning theory
too must play a role; Strevens (1988) proposed the notion of the
teacher informed (but not controlled) by knowledge from a
number of disciplines; and the significance of political and
economic factors is now being convincingly demonstrated
(Pennycook, 1989; Phillipson, 1990; Richards, 1984; Spolsky, 1990;
Strevens, 1971). The simple notion that linguistic theory was
something to be applied to language teaching has done much harm:
As an anonymous linguist pointed out many years ago, “The most
pernicious thing linguists ever did was to sell the idea of pattern
practice drill to high school language teachers” (cited in Spolsky,
1974, p. 2021). The “century-old obsession” (Stern, 1985, p. 249)
with the search for the perfect method has provided scope not just
for the inventors of the new panaceas, such as Total Physical
Response, Community Language Learning and Suggestopedia (see
Richards & Rodgers, 1986), but has also encouraged premature and
inappropriate attempts to translate theoretical proposals like the
Monitor Model or proposed taxonomies of pragmatic functions like
610 TESOL QUARTERLY

the notional/function syllabus into all-encompassing and indepen-
dent methods. As the paper in this collection by Hatch, Shirai, and
Fantuzzi points out, we have not yet had much of the opposite but
necessary process of including knowledge from language teaching
in language learning theory.
The argument that a theory of second language learning must not
be saddled with the extra burden of providing a basis for language
teaching methods has been presented on a number of occasions; in
a recent paper in this journal (Spolsky, 1988), I summarized my
position as follows:
If one looks at the complexity of the circumstances under which second

languages are learned, or fail to be learned, one immediately sees that a
theory must not just be equally complex but must be able to account for
the successes and failures of the many different methods that have been
and are used throughout the language teaching world. (p. 378)
Understanding the conditions under which learning occurs does not

directly say how to reproduce those conditions; indeed, the general
theory as I understand it shows rather the complex trade-offs that
make any teaching approach a less than perfect provision of
learning conditions (i.e., individual differences make it clear that no
one method will meet the needs of any group).
The full-length articles in this issue of the TESOL Quarterly,
modified versions of papers read at a colloquium at the 24th Annual
TESOL Convention (1990), are in full agreement that a theory of
second language learning is not required to be the basis of a method
of second language teaching. But this leaves us with the task of
saying what it should do. What is the task of a theory, and what is
its scope?
The papers that follow present various answers to this question.
For McLaughlin, the first task of a theory is to provide a set of
hypotheses about the acquisition of second language grammars that
are capable of being investigated and falsified. Bialystok, who
recognizes the distinctions between micro- and macrolevel theories
and among those that derive their scope from applied linguistics or
sociolinguistics or psycholinguistics, specifically chooses to focus on
the question of a psycholinguistic theory of second language
acquisition. Long, too, recognizes the problem of scope; like
Bialystok, he concedes that it can be discipline-oriented and may
vary; his summary of accepted findings shows that he sees the value
of a broad scope. Schumann looks for a cognitive model that will
explain observations about pidginization in the grammar of
interlanguages. Sokolik focuses on the learning of grammars but
INTRODUCTION TO A COLLOQUIUM 611

shows the value of trying to deal with child-adult differences in the
same model. Hatch, Shirai, and Fantuzzi make the major break-
through of going beyond grammar (linguistic competence) to oral
and written ability (communicative competence). All of these take
us beyond the limitation of answering single questions about a
single type of second language learning.
These papers demonstrate that over the last few years work in
second language acquisition has been moving beyond its earlier
limited scope. As Ellis (1985) pointed out in his excellent survey of
the field, much of the work until recently had been limited to a
small number of items of phonology and grammar. Following the
lead of first language acquisition research, which has tended to have
its agenda set by the generative concern to establish linguistic
universals, it looked for evidence of how these same structures were
acquired by second language learners. As Ellis noted, one necessary
effect of this concentration on what used to be the microlevel was
that second language acquisition models find little place for the
larger factors such as attitude and aptitude. The macrolevel studies
like those of Gardner (1985) include such factors, and so seem to
have much more to say to the language teacher.
But, as Hatch, Shirai, and Fantuzzi point out, studies in recent
years have moved beyond the limitations of interlanguage research,
to include such topics as conversation, rhetoric, and lexicon. The
field is much broader than that surveyed by Ellis (1985), but there
is still no general theory. There is a great advantage in a general
theory, one with the widest possible scope. The generality of the
theory makes it possible to consider within one model (and so to
attempt to understand and describe the relevant differences that
exist between) second and foreign language learning, learning for
general and specific purposes, formal and informal learning,
developing knowledge and skills, to mention just a few of the ways
theories are sometimes specialized. Saying that a general theory
must cover both second and foreign language learning, or both
formal and informal learning, or both child and adult learning, does
not say that they are the same; it simply sets as a task of that theory
to show how they are different, rather than ignoring this issue. Thus,
I would expect a general theory to include anything that can
reasonably be considered relevant different learners, different
goals, different kinds and amounts of exposure. If these variations
make a difference, a general theory must show how that happens.
The contributors to this issue are, I suspect, not in disagreement
with this general aim and definition of scope, although clearly they
would take the position that certain areas need preliminary
attention or concentration before the most general questions can be
612 TESOL QUARTERLY

tackled. On the more difficult question of the nature and form of
such a theory, there is naturally less agreement.
But all seem to agree that we have reached a point in developing
a theory of second language learning where the fundamental
questions of the form are open, the older set of views is under
question, and something in the nature of a paradigm shift is in
effect.
The papers in this issue show this readiness for change clearly.
McLaughlin shows how vagueness in using the distinction between
“conscious” and “unconscious” learning has handicapped the field.
He considers the contrast in theories based on rules and those based
on connectionist principles, and points out the need to integrate the
representational and the processing perspective. Bialystok takes up
this issue in more detail, makes clear the confusions that have
existed in second language learning between competence and
processing models, and shows how they may well be complemen-
tary views of the same facts. Long considers various kinds of
explanation and various mechanisms that may be proposed as
accounting for the observed facts of second language learning.
Schumann summarizes a number of cognitive models and considers
how they might account for pidginization: cognitive theory, the
experiential approach, the knowledge and control dimensions, the
active control of thought model, the connectionist lexical theory
memory model, and the cognitive interactionist model; he con-
cludes that a satisfactory model will have to incorporate cognitive
processing and brain architecture. Sokolik describes parallel
distributed processing (PDP) models, connectionist models based
in the brain analogy that do not assume rules, and shows how one
such model might account for child/adult differences in second lan-
guage learning. Hatch, Shirai, and Fantuzzi argue for the likelihood
of overlapping modules, developed on an artificial intelligence or
connectionist model, that will account for various aspects of second
language learning and build a total but fuzzy theory.
There is, it will be noted, a healthy and unusually polite accep-
tance of the possibility of pluralism in the answers proposed, a will-
ingness to concede that different models might be needed for dif-
ferent aspects of the problem, an acceptance that different points of
view may lead to different theories. I must say that I find this en-
thusiastic ambivalence congenial. Those who were trained in struc-
tural linguistics will recall the heat regularly engendered by the de-
bates between what Householder (1952) strikingly characterized as
the God’s truth or hocus-pocus approaches to grammars. While the
former is attractive intellectually, its notion of a single correct
answer to all problems is somewhat presumptuous and certainly

disfavors “elephantine” (remember the blind men and the ele-
phant?). Perhaps avoiding the pejorative associations of the term
hocus-pocus, we might still accept that academic wizardry involves
exploring various techniques and models that will cast more light on
the phenomena we are studying.
There is nothing wrong in a degree of tentativeness, ambivalence,
or cautious eclecticism in the failure to plump all our savings on a
single proposed model. At the macrolevel, we are learning a great
deal from multifactor causal models; the preference theory
(Spolsky, 1989) offers, I believe, a particularly useful way to state
the assumptions to be explored in these causal models, but other
models are useful. At the microlevel, it is time to start moving
beyond the productive lessons of the purest competence models
into the complex and messy world of process models. Here,
preference models (translated perhaps into expert systems) offer
one possible route; PDP models offer another, if only to those with
the courage of long detailed grinding work. With such detailed
exploration, we would be ready to start doing more than
speculating about combining both levels. Hatch, Shirai, and
Fantuzzi go furthest in this argument for a multimodular solution.
This collection of papers is intended to contribute to a productive
reconsideration of the state of second language learning theory.
Individually, each paper makes important contributions: McLaugh-
lin and Bialystok help clear away the confusions engendered by the
increasingly fuzzy use of such crucial terms as unconscious and
conscious, competence and performance, and processing; L o n g
identifies important findings that need to be included in theory;
Schumann and Sokolik show the potential of neurophysiological
and connectionist models; Hatch, Shirai, and Fantuzzi explore the
neuropsychological basis for second language acquisition. Taken
together, these papers provide a view of the current state of the
field, pointing to the directions in which it is likely to be moving in
the coming years, and capturing what might well turn out to be a
moment of paradigm shift in our conception of the scope and form
of a theory of second language learning.
614 TESOL QUARTERLY

THE AUTHOR
Bernard Spolsky, Professor in the Department of English at Bar-Ilan University,
Ramat-Gan, Israel, is a former President of TESOL. His book, Conditions for
Second Language Learning: Introduction to a General Theory (Oxford University
Press, 1989) has been awarded the 1990 Book Prize by the British Association for
Applied Linguistics and the tenth Kenneth W. Mildenberger Prize by the Modern
Language Association of America.
REFERENCES
Ellis, R. (1985). Understanding second language acquisition. Oxford:
Oxford University Press.
Gardner, R. C. (1985). Social psychology and second language learning:
The role of attitudes and motivation. London: Edward Arnold.
Householder, F. W. J. (1952). [Review of Methods in structural linguistics].
International Journal of American Linguistics, 18, 260-268.
Miller, G. A. (1990). Linguists, psychologists, and the cognitive sciences
[Discussion note]. Language, 66 (2), 317-322.
Pennycook, A. (1989). The concept of method, interested knowledge, and
the politics of language teaching. TESOL Quarterly, 23 (4), 589-618.
Phillipson, R. (1990). English language teaching and imperialism.
[Doctoral dissertation, University of Amsterdam]. Tronninge, Denmark:
Transcultura.
Richards, J. C. (1984). The secret life of methods. TESOL Quarterly,
18 (1), 7-23.
Richards, J. C., & Rodgers, T. S. (1986). Approaches and methods in
language teaching: A description and analysis. Cambridge: Cambridge
University Press.
Spolsky, B. (1968). Linguistics and language pedagogy—Applications or
implications? In J. E. Alatis (Ed.), Twentieth annual round table
meetings on languages and linguistics (pp. 143-155). Washington, DC:
Georgetown University Press.
Spolsky, B. (1974). Linguistics and education: An overview. In T. A.
Sebeok, A. S. Abramson, D. Hymes, H. Rubenstein, E. Stankiewicz, &
B. Spolsky (Eds.), Current trends in linguistics: Linguistics and adjacent
arts and sciences. (Vol. 13, pp. 2021-2026). The Hague: Mouton.
Spolsky, B. (1988). Bridging the gap A general theory of second language
learning. TESOL Quarterly, 22 (3), 377-396.
Spolsky, B. (1989). Conditions for second language learning: Introduction
to a general theory. Oxford: Oxford University Press.
Spolsky, B. (1990, April 19). Educational linguistics: Definitions, progress,
problems [Keynote address]. 9th World Conference of Applied
Linguistics, Thessaloniki, Greece.
Stern, H. H. (1985). [Review of Methods that work: A smorgasbord of
ideas for language teachers]. Studies in Second Language Acquisition, 7,
249-251.

Strevens, P. (1971). Where has all the money gone? The need for cost-
effectiveness studies in the teaching of foreign language. In G. Perrin &
J. L. M. Trim (Eds.), Applications of linguistics: Selected papers of the
second international congress of applied linguistics, Cambridge 1969
(pp. 389-406). Cambridge: Cambridge university press.
Strevens, P. (1988). Learning English better through more effective
teaching: Six postulates for a model of language learning/teaching.
World Englishes, 7 (l), 51-63.
616 TESOL QUARTERLY

“Conscious” versus “Unconscious”

Learning
BARRY McLAUGHLIN
University of California, Santa Cruz
This article examines the concept of consciousness in second lan-

guage research. After defining theoretical assumptions and
reviewing a number of controversies in the psycholinguistic and
second language literature, I argue that although the terms
conscious and unconscious have a place in our prescientific
vocabulary, they have acquired too much surplus meaning and
should be abandoned in favor of clearly defined empirical
concepts. Lacking an adequate theory of mind that allows us to
decide that particular mental states or operations are “conscious”
or “unconscious,” one cannot falsify claims regarding conscious-
ness in second language learning.
In this article, I discuss the theoretical ramifications of a persistent

debate that divides researchers in first language acquisition, in
second language acquisition, and in psycholinguistics generally.
This is the debate between those who would assign a greater role to
unconscious processes in language acquisition and those who see
little or no role for unconscious processes. The debate takes many
forms, some of which will be outlined in this paper.
THREE ASSUMPTIONS
I begin by making clear a number of assumptions I hold about
theory. The first assumption is that research is inseparable from
theory. By theory is meant a way of interpreting, criticizing, and
unifying established generalizations. These generalizations arise
from hypotheses that are supported in research and can be said to
constitute “facts” or “laws,” according to whether they are particu-
lar or general in content. A theory is a system of facts and laws that
have been altered, reinterpreted, and given meaning. We impose
this meaning on all research—there is no “phenomenal language”
distinct from “theoretical language.” We cannot not interpret. In our
617
research we are inevitably testing hypotheses about a theory. The
theory may be vaguely or not at all described, but the researcher is
never unencumbered by expectancies and beliefs about the
phenomenon in question.
Second, I would argue that there is no one scientific method. The
traditional research cycle proceeds from description to correlation
to experimentation. We describe a phenomenon, see with what it
covaries, and then try to determine why it happens as it does. We
have been taught to move from descriptive to inferential statistics,
with the assumption that the degrees of control possible in experi-
mentation will lead to true understanding. But laboratory control
has its perils, including the possible trivialization of the phenom-
enon under investigation and the lack of motivation on the part of
the subject. Experimental research has its value, but I would argue
for a catholicity of outlook when it comes to method—single-case
studies can be more valuable than carefully controlled laboratory
research or large-scale multivariate analyses in furthering our
understanding of the phenomena we are trying to understand.
Finally, I make the assumption that there is no single scientific
truth. It follows from what I have been saying about method that
we are going to make the most progress if we take an open, multi-
method strategy. Data triangulation is achieved most often through
the work of different researchers with varied interests getting a fix
on the same phenomenon from different angles. I side with those
philosophers of science who argue that truth can never be known
directly and in its totality. All knowledge is mediated by the symbol
systems used by scientists. Just as there are multiple perspectives,
multiple accounts of what is seen, so there are multiple truths. The
symbol system or metaphor used by a particular scientific approach
may help us see more clearly, but it does not constitute ultimate
truth. Nor does the combination of all partial representations of
truth add up to ultimate truth. Ultimate truth is only approximated
by the shadows cast by the metaphors of our theories.
However, one needs to be careful to avoid theoretical solipsism.
There must be criteria for evaluating the adequacy of different
theoretical points of view. Kaplan (1964) proposed that theories can
be evaluated by norms of correspondence, norms of coherence, and
pragmatic norms. Norms of correspondence relate to the fit
between the theory and external reality and to the correspondence
of the theory to the facts that the theory is to explain. A good theory
must have definitional adequacy and explanatory power. It must
have coherence in the sense that it must fit with the body of
knowledge that is already established and must be consistent with
related acceptable theories. An acceptable theory must also be
618 TESOL QUARTERLY

practical in the sense that it is heuristically rich, in that it guides and
stimulates the ongoing process of scientific inquiry. It should raise
new questions.
To Kaplan’s norms many would add a fourth criterion—falsifia-
bility. A theory can fit the data well, be consistent with related
theory and elegant in its formulation, and be heuristically rich—
without being falsifiable. That is, it can meet all of Kaplan’s criteria
and still be of little value. Theories that are self-contained cannot be
tested. Such a theory may survive, though at best it will survive as
an impervious fortress, perhaps invincible but in splendid isolation.
The results of research probe but do not prove a theory. A theory
may repeatedly survive such probing—but it may be always
displaced by a new probe. In practice, this means that a theory is
either disconfirmed or escapes being disconfirmed. But it is never
confirmed. This is the logic of statistical inference as well: The null
hypothesis is never accepted—it can only be rejected or fail to be
rejected.
To summarize (for a more extended discussion, see McLaughlin,
1987), my position concerning theory is that we cannot do without
it; we will achieve progress as we come to illuminate our under-
standing by taking different perspectives and by utilizing diverse
methods. Good theories fit the data well, are consistent with related
formulations, and are heuristically rich. Perhaps most important,
they are capable of disconfirmation.
THREE DEBATES
The Krashen/McLaughlin Debate
I turn now to three debates that bear on the distinction between
conscious and unconscious processes. The first is a debate that took
place a number of years ago between Stephen Krashen and myself.
One of our differences of opinion related to a subjective experience
that is quite common in learning a new language. I gave the
example from my own experience with German:
When I “feel” that something is wrong with Ich habe nicht das Kind
gesehen, I also know that there is a rule about the placement of
negatives. Similarly, while I have to have recourse to the rule to be sure
that Ich habe es ihm gegeben is correct, I also have a feel that Ich habe
ihm es gegeben is wrong. At least in my own introspection, it is unclear
whether I am working on the basis of “rule” or “feel.” (McLaughlin,
1978, pp. 317-318)
Krashen, of course, would explain this by saying that in the first
case I had unconsciously (or subconsciously) “acquired” the rule
CONSCIOUS VERSUS UNCONSCIOUS LEARNING 619

and in the second I had consciously “learned” it. In fact, Krashen
said as much in his reply to my critique of the Monitor Model. He
argued (Krashen, 1979), “McLaughlin’s observations about himself,
on the other hand, seem clear to me if not to him” (p. 152). But this
was exactly my point—that two or more people can disagree about
whether one and the same experience is based on rule or feel. How
are such differences of opinion to be resolved? This is an important
question for a theory such as Krashen’s, where the ultimate test of
whether a process involves conscious learning or unconscious
acquisition is whether the learner is working on the basis of “rule” or
“feel” (McLaughlin, 1978).
In his theory, Krashen argued that the starting point for all lan-
guage performance is acquisition, thought of as an unconscious
process similar to that by which children acquire their native lan-
guages. Utterances are initiated by the acquired system, with con-
scious learning available only as a monitor to alter the output of the
acquired system. In contrast to this approach, I have argued
for a distinction between controlled and automatic processing
(McLaughlin, Rossman, & McLeod, 1983). Controlled processes
require active attention but are not always available to conscious
perception. They regulate the flow of information between short-
and long-term memory systems. Automatic processes are associated
with long-term memory and take substantial time to develop and
become established. They do not require attention, are usually not
available to consciousness, but can become the focus of attention in
certain circumstances. This information-processing model assumes
that, in the initial stages of learning, controlled processes are adop-
ted and used to perform accurately and are in effect the “stepping
stones” for the development of subsequent automatic processes.
Like Shiffrin and Schneider (1977), from whom these concepts
were adopted, I have argued that the distinction between controlled
processing and automatic processing is not based on conscious
versus unconscious awareness. Shiffrin and Schneider were careful
not to associate their distinction of controlled and automatic pro-
cessing with a distinction between conscious and unconscious pro-
cessing: Both controlled and automatic processes can in principle be
either conscious or not. Because most automatic processes occur
with great speed, their constituent elements are usually, but not
necessarily, hidden from conscious perception, Some controlled
processes also occur with great speed, so that they may not be
available to conscious experience. Shiffrin and Schneider (1977)
called these controlled processes “veiled” (p. 138). Other controlled
processes, those they referred to as “accessible” (p. 138), are easily
perceived by the learner.
620 TESOL QUARTERLY

Thus the distinction between controlled and automatic process-
ing allows us to avoid recourse to such notions as “conscious” or “un-
conscious” or a sense that one is operating by “rule” or “feel.”
Instead, the distinction between controlled and automatic process-
ing relates to the degree to which the skills in question have been
routinized and established in long-term memory. This is empirically
testable. For example, in second language research, investigators
have used such measures as reaction time, semantic facilitation
effects, and the Stroop task to determine the degree of automaticity
of lexical retrieval (see McLaughlin, Rossman, & McLeod, 1983, for
further examples).
Certainly John Schumann (1983) was correct in his contention
that Stephen Krashen and I are employing different metaphor
systems. He went on, however, to maintain that both views
can coexist as two different paintings of the language learning
experience—as reality symbolized in two different ways. Viewers can
choose between the two on an aesthetic basis, favoring the painting
which they find to be phenomenologically true to their experience.
(p. 56)
This may be true on an intuitive, prescientific level, but if we are to
progress theoretically, there must be norms for deciding between
competing theories. As was mentioned earlier, one of these norms is
falsifiability. The information-processing or cognitive account
provides means of empirically testing its predictions. It is difficult
to see how Krashen’s claims about conscious learning and uncon-
scious acquisition can be falsified.
The Reber/Dulany Debate

In a number of papers, Reber and his associates (e.g., Reber,
1976; Reber & Allen, 1978) have argued for what they have called
implicit learning. In their experiments, Reber and his colleagues
examined the process whereby complex abstract knowledge of a
structured stimulus domain is acquired implicitly, held tacitly, and
used unconsciously to make accurate decisions about the well-
formedness of novel items. In this research, subjects were exposed
to finite-state grammars made up of letter strings; subjects were
found to be significantly accurate when they subsequently had an
opportunity to judge the grammaticality of novel grammatical and
non grammatical strings.
The subjects in these experiments were encouraged to maintain a
running commentary on the reasons and justifications for their
judgments. This introspective evidence convinced Reber and his

colleagues that “learning occurs in the absence of explicit code-
breaking strategies; our subjects cannot tell us very much about
what they know” (Reber & Allen, 1978, p. 204).
This conclusion has been challenged by Dulany, Carlson, and
Dewey (1984), who questioned the degree to which the knowledge
of subjects in these experiments is properly characterized as
abstract and the degree to which it is truly unconsciously held.
Using the same artificial grammar that Reber and his associates
used, Dulany et al. replicated the finding that subjects acquire
sufficient knowledge of the rules of grammar by simple observation
of exemplary strings to be able to judge the grammaticality of novel
strings.
In addition, however, Dulany et al. required their subjects to
identify the aspect of the string that defines its grammaticality by
underlining the part of the item that made it correct or crossing out
the part of an item that violated the rule. Using this procedure,
Dulany et al. found that subjects developed personal and idiosyn-
cratic sets of rules and that these sets of rules correlated with the
finite-state grammar in the sense that both sets of rules resulted in
the same grammatical classifications. The subjects’ idiosyncratic
rules were of imperfect validity and of limited scope but were
accessible to consciousness. Dulany et al. (1984) concluded with this
comment:
We often intuitively judge the grammaticality of a sentence or the
legality of a move or the propriety of an act without conscious access to
the formal syntax of the domain. But let us turn the tables somewhat. It
is an interesting possibility that each of those intuitions is one of a set of
informal rules of limited scope and perhaps imperfect validity. The
intuitions seem quite conscious. We know something that seems right or
wrong, even when we don’t think of or know the proper rule from a
formal system. With intuition reclaimed for consciousness, we would not
disagree with Allen and Reber (1980, p. 178) that “decisions about the
well-formedness of test strings are made largely on an intuitive basis.”
(p. 554)
In their reply to the Dulany et al. (1984) article, Reber, Allen, and
Regan (1985) stated that they have no issue with the notion of
“correlated grammars” (p. 533). They regarded it as not improbable
that subjects emerge from the learning phase of their experiments
with sets of rules that, to a degree, make the same classifications as
finite state grammars. However, they did not agree that the sets of
rules that represent these grammars appear in consciousness. Reber
et al. argued that the procedure employed by Dulany et al. —
identifying the aspect of the string that defines its grammaticality—
is not a recall technique, but more akin to a recognition technique.
622 TESOL QUARTERLY

Such a technique allows for the existence of knowledge to be
confirmed or disconfirmed but does not allow for any statement
about the explicitness of this knowledge.
In their reply to the Reber et al. (1985) reply, Dulany, Carlson, and
Dewey (1985) rejected the methodological critique of Reber et al.
They argued that by requiring subjects to report on the grammatical
status of a string and to report which letters of the string implied that
status, they were obtaining a report of each rule as close as possible
to the moment of judgment. It is not reasonable, they argued, to insist
that only memory for the rules as measured by a recall test taps the
subject’s explicit and conscious knowledge. Dulany et al. concluded
that explicit and implicit learning are distinguished not by one being
conscious and the other unconscious, but by the different forms of
nonconscious operations that yield information in consciousness from
other information in consciousness.
This may not resolve the debate. Perhaps Reber et al. (1985) came
closest to the truth when they defined their differences with Dulany
et al. (1984) in these terms:
So, to cast the problem into its simplest form, (a) some mental “stuff” is
unconscious and some mental “stuff” is conscious, but (b) no one has an
adequate theory about how to categorize the actions of mind so that
their primary causal roots are put in the proper category. Dulany et al.
are arguing for a cognitive paradigm that emphasizes the role of con-
sciousness as an explanation of human behavior; their prejudice is to cast
phenomena into the conscious control category. We are pushing for a
perspective that grants unconscious processes a larger explanatory role;
our prejudice is to place phenomena into the unconscious control box.
(p. 22)
In other words, it comes down to personal prejudice whether one
stresses conscious or unconscious processes. If this is true, it
becomes very difficult to falsify theories that emphasize one or the
other of these processes.
The McClelland, Rumelhart/Pinker, Prince Debate

The issue in this exchange concerned the question of whether
knowledge of language consists of mentally-represented rules.
McClelland, Rumelhart, and their colleagues (McClelland,
Rumelhart, and the PDP Research group, 1986) took the con-
nectionist view that information is encoded in a large number of
simple but highly interconnected units or nodes that receive excita-
tory or inhibitory activity along their arcs (Anderson, 1983;
McClelland et al., 1986). According to McClelland et al., mentally

represented rules are merely convenient approximate fictions; the
real causal processes in language acquisition and use are the transfer
of activation levels among units and the modification of the weights
of their connections.
For the connectionists, learning takes place through the strength-
ening and weakening of interconnections in response to examples
encountered in the input. There is no need for learners to appreciate
the significance of specific syntactic forms or to make an inductive
leap to abstract rules. Instead, learning consists of a network of units
that enable the learner to produce rule-like behavior, but the rules
themselves exist only as association strengths distributed across the
entire network.
McClelland, Rumelhart, and the PDP Research group (1986) have
described a connectionist model of the acquisition of the past tense in
English that they claim successfully maps many stems onto their past
tense forms, both regular (walk/walked) and irregular (go/went),
and that also replicates some of the errors and sequences of language
development in young children. The McClelland/Rumelhart model
does this without explicit rules; instead, the model consists of a set of
neuron-style units that stand for trigrams of phonetic features of the
past form, and an array of connections between the sets of units
whose strengths are modified during learning.
Pinker and Prince (1988) rejected these claims, arguing that the
model explains both too little and too much: It does not learn very
many rules and among those it does learn are rules found in no
human language. Furthermore, Pinker and Prince maintained that
the connectionist model does not account very well for the facts of
morphological overgeneralization in child learners nor does it
provide a satisfactory account of the manner in which children
master the past tense in English. In the eyes of its critics, the
connectionist approach is incompatible with contemporary linguis-
tic theory, with many of the facts of acquisition, and with much
experimental psycholinguistics (see also Fodor & Pylyshyn, 1988).
Nonetheless, the connectionist approach represents an important
theoretical contribution because it attempts to provide an explicit
and testable account of how linguistic knowledge is organized in the
mind. Moreover, the connectionist approach rejects the notion of
unconscious rules that control linguistic behavior. Instead, the
model posits that linguistic behavior is the result of a network of
examples exerting their respective associative weights. These
associative weights are built up over time via parallel processing.
Instead of relying, then, on “unconscious” knowledge of an
elaborate, formal grammar of rules, one can think of linguistic
behavior as the product of a rather heterogeneous compilation of
associated memories.
624 TESOL QUARTERLY

IS THE CONSCIOUS-UNCONSCIOUS DISTINCTION
USEFUL FOR SECOND LANGUAGE THEORY?
From the preceding discussion it should be clear what the
author’s answer is to the question of whether the conscious-uncon-
scious distinction is useful for second language theory. However, a
strong case has been made recently for the usefulness of this
distinction. This argument will be reviewed briefly before the
counterargument is presented.
On the Utility of the Conscious-Unconscious Distinction

In his excellent discussion of the role of consciousness in second
language learning, Richard Schmidt (1990) took the position that
both conscious and unconscious processes are involved in second
language learning and that an important goal in research should be
to assess the available evidence for the relative contribution of each.
To do this, he argued, we need to have a clear understanding of
what is meant by “conscious” and “unconscious.” Schmidt noted
that there are various meanings for consciousness in the literature
and that the question of whether second language learning is to be
thought of as conscious or unconscious is not a single question, but
is a reflection of a number of different contrasts. These include the
question of whether the learner is aware of learning something (“un-
conscious” in this case would mean that the learner is not aware of
learning anything), whether the learner’s awareness is at the level of
noticing (“unconscious” here would mean that the learner has not
noticed anything), whether the learner has understanding and
insight (“unconscious” would mean that something has been learned
without conscious understanding). Other contrasts relate to inten-
tion, at one level whether the learner intends to notice or pays atten-
tion (“unconscious” meaning that this is not the case), and at a more
global level, whether there is a deliberate plan involving study and
other intentional learning strategies (“unconscious” meaning that
there is no such planning). Finally, there is the question of articulate
report, whether learners are able to say what they appear to know
(“unconscious” in this sense would mean that knowledge is not
accessible to introspective report).
In discussing these contrasts, Schmidt pointed out that we have
grown used to accounts that talk about child or adult language
learners as though they were sophisticated grammarians. Learners
are said to start with certain assumptions about language, to scan the
input for certain crucial data, then to add, delete, and reorganize
rules in an attempt to confirm or reject hypotheses—all part of an
attempt to construct a theory of the language they are learning.

However, although the cognitive operations described in this
process are those that typify conscious reasoning of the highest
order—testing hypotheses through inductive and deductive reason-
ing, drawing inferences, and making abstract generalizations—
theorists do not typically conclude that learners are consciously
engaged in the enterprise of theory construction when learning lan-
guages (but see McLaughlin, 1986, for a discussion of Soviet
theory). The usual solution has been to assume that this reasoning is
going on unconsciously, and that unconscious thought processes are
similar to conscious thought processes except that they are not
accessible to introspection and awareness.
The problem with this account of language learning is that it is in
apparent contradiction to a large body of evidence that suggests
that there is no learning without awareness. Reviewers of the
psychological literature (Brewer, 1974; Dawson & Schell, 1987;
Ericsson & Simon, 1984) are consistent in their conclusion—adult
humans do not learn without awareness. Experiments claiming to
show that learning without awareness is possible are rejected,
usually because they employ insensitive measures of awareness.
This is essentially the argument made by Dulany et al. (1984) in their
critique of Reber’s work.
The literature in experimental psychology indicates that memory
requires attention and awareness. There have been a number of
studies that have shown that unattended stimuli can influence
behavior. For example, subliminally presented words that the
subject does not “see” can prime semantic associates (Marcel, 1983).
That is, unattended information can make its way into short-term or
working memory. However, the effects of such subliminal process-
ing are generally small and involve stimuli that are already well
known: native language words, letters of the alphabet, etc. There is
no evidence that unattended information becomes part of the long-
term store. As Schmidt (1990) noted, familiar stimuli may be
processed subliminally, because habitual encodings are automatic
and do not depend on limited capacity, but it is not possible to learn
a second language or anything else through subliminal perception
(see Pratkanis & Greenwald, 1988).
Schmidt (1990) maintained that there is insufficient evidence to
conclude that second language learning occurs without awareness.
Nor did he agree that understanding is epiphenomenal to learning
or that most second language learning is unconscious. Instead, he
argued that the evidence suggests that understanding and learning
are in most cases tightly linked—most of the time when we want to
learn something we make ourselves conscious of it, and the more
there is to learn, the greater the need for sustained conscious in-
volvement.
626 TESOL QUARTERLY

Schmidt argued strongly for the thesis that nothing in the target
language input becomes intake for language learning other than
what learners consciously notice. He accepted the possibility of
incidental learning—learning without consciously trying to learn—
especially when task demands focus attention on relevant features
of the input. However, he rejected the possibility for adults of
incidental learning in the sense of picking up target language forms
from input when they do not carry information critical to the task at
hand. For Schmidt, paying attention to language form is hypothe-
sized to be facilitative in all cases and may be necessary for adult
acquisition of redundant grammatical features.
The Case for Abandoning the Conscious-Unconscious Distinction

Schmidt’s (1990) discussion of the role of consciousness is an
important contribution to second language theory. Schmidt is
correct in maintaining that the role of unconscious second language
learning has been exaggerated, and that our research should focus
on what learners notice and what they think as they learn second
languages. However, Schmidt also clearly articulated the diffi-
culties involved in using the concepts conscious and unconscious.
He has shown that there are multiple meanings for both terms and
that they are frequently ill defined. Nonetheless, Schmidt took the
position that the notion of consciousness is respectable because of
an impressive body of research that deals with its role in cognition
and learning and because it ties together many related concepts,
including attention, short-term memory, controlled versus auto-
matic processing, and serial versus parallel processing.
My own bias, however, is to avoid use of the terms conscious and
unconscious in second language theory. I believe that these terms
are too laden with surplus meaning and too difficult to define em-
pirically to be useful theoretically. Hence my critique of Krashen’s
distinction between learning and acquisition—a distinction that
assumes that it is possible to differentiate what is conscious from
what is unconscious. Here I am in agreement with Terence Odlin
(1986):
One does not need to attempt dubious distinctions between types of
knowledge about language in terms of the notoriously slippery notion of
“consciousness,” or to speculate about the supposed differences in the
knowledge about language that people may “acquire” instead of “learn.”
(p. 138)
This is not to deny the importance of the subjective experience
that theorists attempt to describe using the terms conscious and un-
conscious. However, it seems to me that such subjective experience

is best left to the psychoanalyst. My own preference is to short-
circuit the use of these terms and to define specifically the particular
contrasts that one is interested in. Then the task becomes one of
empirically defining the operations that distinguish the terms being
contrasted. Figure 1 lists the contrasts distinguished by Schmidt
(1990); which were discussed earlier, as well as other contrasts from
the literature of cognitive psychology. My point is that we should
avoid using the umbrella concepts conscious and unconscious when
making such contrasts, and should instead specify precisely what
contrast we are making.
Although all of the contrasts listed in Figure 1 have not been
examined in detail, some have. For example, as we have seen, there
FIGURE 1
Contrasts Underlying Various Uses of the Terms Conscious and Unconscious
has been considerable research on the question of learning without

awareness, and the limitations of such learning have been fairly well
defined. Similarly, there has been considerable discussion of the
concepts of intentional and incidental learning. In fact, there is still
general consensus in the cognitive literature on what I wrote many
years ago:
The operational procedures by which INT [intentional] and INC
[incidental] learning are distinguished typically involve the comparison
of performance under instructions to learn the relevant material and no
instructions to learn this material. Data from research show quantitative
differences between the instructions and no-instructions groups, but all
that can be concluded on the basis of such data is that learning is more
628 TESOL QUARTERLY

difficult under disadvantageous (no-instructions) conditions. The no-
instructions procedure does not preclude the possibility of sporadic self-
instructions, and consequently intergroup differences are attributable
only to the functional relations of a number of parameters of efficient
and inefficient INT learning. There is no justification for the implication
that two types of learning-defined by different operational anteced-
ents—are being investigated in this research. (McLaughlin, 1965, p. 373)
Another contrast that has received a good deal of attention is
that between explicit and implicit knowledge. This was the
contrast at issue in the Reber/Dulany debate, which is a good
example of why the terms conscious and unconscious are not
helpful and why it is more useful to specify the contrast in
question and the empirical operations that define that contrast.
The debate went nowhere on the issue of “conscious” versus “un-
conscious” learning because there was no adequate theory of the
architecture of the mind to which both parties could agree.
However, I believe that the methodological refinement intro-
duced by Dulany et al. (1984) constitutes a significant advance for
this line of research because it allows judgments regarding the
existence of knowledge and the explicitness of such knowledge
(pace Reber et al.). Furthermore, Dulany et al. have made the
important point that the grammar that is explicit in the mind of the
learner may correlate but not be identical with the formal
grammar of the target language (see also Odlin, 1986).
Related to the issue of explicit versus implicit knowledge is the
question of whether the subject has the ability to report accurately
on knowledge or on the intent to use particular strategies in
learning. A number of second language researchers (Faerch &
Kasper, 1983; McGroarty, 1988; O’Malley & Chamot, 1989; Oxford,
1986) have attempted to specify strategies that good language
learners use. A major problem with research on self-reported
strategies is that it relies on retrospection. Subjects’ reports may
derive more from what they think they should have been doing than
from what they actually were doing. Attempts to obtain more
directly on-line data by interrupting the task at intervals and asking
subjects what they were doing threaten the ecological validity of the
task. Nonetheless, more accurate introspective data may be possible
using methods such as the yoked-subjects technique, where subjects
are told to tell another potential subject just what they are doing
(Nayak, Hansen, Krueger, & McLaughlin, in press), or by using
subjects trained in introspection (Ericsson & Simon, 1984).
As was noted earlier, the distinction used in the information-
processing literature between controlled and automatic processing
does not make any assumptions about conscious or unconscious

processing. Earlier formulation (e.g., Posner & Snyder, 1975) iden-
tified control with consciousness and automaticity with unconscious
activation. However, Shiffrin and Schneider (1977) were careful not
to associate the distinction between controlled and automatic
processing with the distinction between conscious and unconscious
processing.
Nonetheless, the controlled-automatic distinction is not unproble-
matic. As Schneider, Dumais, and Shiffrin (1984) noted, there are
various ways of characterizing these operations that have been
proposed by different authors. Each of these characterizations,
however, represents an attempt to define empirically the nature of
the process involved. Similarly, other contrasts in the information-
processing literature, such as that between focal and peripheral
attention, and between short- and long-term memory, have been
defined and reconceptualized as empirical evidence required.
Thus, Baddeley (1986) and many recent researchers use the term
working memory as a construct replacing such older constructs as
focal attention and short-term memory. The reason why new
concepts and terms are adopted is because old ones have too much
baggage and do not fit the empirical facts.
To summarize my argument: I believe that the terms conscious
and unconscious have too much surplus meaning and should be
abandoned in favor of clearly defined empirical concepts. In this
way, it is possible to be more precise about specific cognitive states
and operations. This is no longer the case for the terms conscious
and unconscious. It is no longer possible to get an empirical handle
on these terms, and hence they have little theoretical utility.
CONCLUSION
In the beginning of this paper, I argued that the role of theory is
to illuminate our understanding. I do not believe that theories that
depend on concepts such as “conscious” learning and “unconscious”
acquisition provide much insight. Progress in second language
research will come as the constructs used in our theories are clearly
specified and testable. If we cannot falsify our theories, we will get
nowhere.
As I noted above, the terms conscious and unconscious are used
to refer to important subjective experiences. We all have the sense
that we have learned something “unconsciously.” My argument,
however, is that although the terms conscious and unconscious have
a place in our prescientific vocabulary, they should be excluded
from theory. As Reber et al. (1985) noted, no one has developed an
adequate theory of mind that allows us to decide that particular
630 TESOL QUARTERLY

mental states or operations are “conscious” or “unconscious.” This is
a serious impediment to falsifiability.
What one finds in the literature of experimental psychology is that
researchers have looked at various manifestations of the conscious-
unconscious contrast: learning with and without awareness, inten-
tional and incidental learning, explicit and implicit knowledge. For
the most part, this research leads to the conclusion that these con-
trasts represent continua, and performance is determined by in-
structions and the nature of the experimental task. The contrasts
that appear in the recent cognitive psychology literature-con-
trolled versus automatic processing, serial versus parallel process-
ing—are neutral with respect to the conscious-unconscious distinc-
tion.
I believe that second language theory should follow a similar
strategy. The point is that theories need to be tied to empirical
operations and empirical data. The constructs that we use to do this
need to be as unambiguous as possible—they cannot mean all things
to all people. They can, however, come from many different
metaphor systems. I am not arguing for the parochial view that the
only way to truth is through cognitive psychology. As I stated
above, theoretical progress depends on taking a catholicity of
perspectives.
I agree with Carroll (1989), who has argued for the importance of
integrating the representations of classical linguist theory with the
processing perspective of cognitive, information-processing theo-
ries. Indeed, representations and processes are inseparable because
an adequate theory requires explicit assumptions about both. Nei-
ther Universal Grammar theory nor present information-process-
ing models used in second language research have attained a very
satisfactory level of explicitness. As Carroll notes, for example, one
of the most researched concepts in second language acquisition,
markedness, is not formally defined in this research and remains
very much an intuitive notion. Similarly, cognitive approaches have
been vague about the specific way in which linguistic representa-
tions are organized in the mind. One hopes that the next
development in second language theory will be the integration of
more explicitly defined representational and processing perspec-
tives.

THE AUTHOR
Barry McLaughlin teaches in the Program in Experimental Psychology at the
University of California, Santa Cruz, where he is Codirector of the National Center
for Research on Cultural Diversity and Second Language Learning.
REFERENCES
Anderson, J. (1983). Syllable simplification in the speech of second lan-
guage learners. Interlanguage Studies Bulletin, 7, 4-36.
Baddeley, A. (1986). Working memory. Oxford: Clarendon Press.
Bialystok, E. (1979). Explicit and implicit judgments of L2 grammaticality.
Language Learning, 29, 81-103.
Brewer, W. (1974). There is no convincing evidence for operant or classical
conditioning in adult humans. In W. Weimer & D. Palermo (Eds.),
Cognition and the symbolic processes (pp. 123-143). Hillsdale, NJ:
Lawrence Erlbaum.
Carroll, S. (1989). Second language acquisition and the computational
paradigm. Language Learning, 39, 535-594.
Chomsky, N. (1986). Knowledge of language: Its nature, origin, and use.
New York: Praeger.
Dawson, M., & Schell, A. (1987). Human autonomic and skeletal classical
conditioning: The role of conscious cognitive factors. In G. Davey (Ed. ),
Cognitive processes and Pavlovian conditioning in humans (pp. 89-111).
Chichester, England: Wiley.
Dulany, D. E., Carlson, R. A., & Dewey, G. I. (1984). A case of syntactical
learning and judgment: How conscious and how abstract? Journal of
Experimental Psychology: General, 113, 541-555.
Dulany, D. E., Carlson, R. A., & Dewey, G. I. (1985). On consciousness in
syntactic learning and judgment: A reply to Reber, Allen, and Regan.
Journal of Experimental Psychology: General, 114, 25-32.
Ericsson, K., & Simon, H. (1984). Protocol analysis: Verbal reports as data.
Cambridge, MA: MIT Press.
Faerch, C., & Kasper, G. (Eds.). (1983). Strategies in interlanguage com-
munication. London Longman.
Fodor, J. A., & Pylyshyn, Z. W. (1988). Connectionism and cognitive
architecture: A critical analysis. In S. Pinker & J. Mehler (Eds.), Connec-
tions and symbols (pp. 3-71). Cambridge, MA: MIT Press.
Krashen, S. (1979). A response to McLaughlin, “The monitor model: Some
methodological considerations.” Language Learning, 29, 151-167.
Kaplan, A. (1964). The conduct of inquiry: Methodology for the behavioral
sciences. San Francisco: Chandler.
Marcel, A. (1983). Conscious and unconscious perception: An approach to
the relations between phenomenal experience and cognitive processes.
Cognitive Psychology, 15, 238-300.
632 TESOL QUARTERLY

McClelland, J. L., Rumelhart, D. E., & the PDP Research Group. (1986).
Parallel distributed processing: Explorations in the microstructure of
cognition: Vol. 2. Psychological and biological models. Cambridge, MA:
Bradford/MIT Press.
McGroarty, M. E. (1989). The “good learner” of English in two settings
(Tech. Rep. No. 12). Los Angeles: University of California, Center for
Language Education and Research.
McLaughlin, B. (1965). “Intentional” and “incidental” learning under
conditions of intention to learn and motivation. Psychological Bulletin,
63, 359-376.
McLaughlin, B. (1978). The monitor model: Some methodological
considerations. Language Learning, 28, 309-332.
McLaughlin, B. (1986). Multilingual education: Theory east and west. In
B. Spolsky (Ed.), Language and education in multilingual settings
(pp. 35-52). Clevedon, Avon, England: Multilingual Matters.
McLaughlin, B. (1987). Theories of second-language learning. London:
Edward Arnold.
McLaughlin, B., Rossman, T., & McLeod, B. (1983). Second language
learning: An information-processing perspective. Language Learning,
33, 135-158.
McLeod, B., & McLaughlin, B. (1986). Restructuring or automaticity?
Reading in a second language. Language Learning, 36, 109-123.
Nation, R., & McLaughlin, B. (1986). Experts and novices: An informa-
tion-processing approach to the “good language learner” problem.
Applied Psycholinguistics, 7, 41-.56.
Nayak, N., Hansen, N., Krueger, N., & McLaughlin, B. (1990). Lan-
guage-learning strategies in monolingual and multilingual subjects. Lan-
guage Learning, 40, 221-244.
Odlin, T. (1986). On the nature and use of explicit knowledge. IRAL, 24,
123-144.
O’Malley, J. M., & Chamot, A. U. (1989). Learning strategies in second
language acquisition. New York: Cambridge University Press.
Oxford, R. L. (1986). Second language learning strategies: Current
research and implications for practice. University of California, Los
Angeles: Center for Language Education and Research.
Pinker, S., & Prince, A. (1988). On language and connectionism: Analysis
of a parallel distributed processing model of language acquisition.
Cognition, 28, 73-193.
Posner, M. I., & Snyder, C. R. R. (1975). Attention and cognitive control.
In R. L. Solso (Ed.), Information processing and cognition: The Loyola
symposium (pp. 145-175). Hillsdale, NJ: Lawrence Erlbaum.
Pratkanis, A. R., & Greenwald, A. G. (1988). Recent perspectives on
unconscious processing: Still no marketing application. Psychology and
Marketing, 5, 337-353.
Reber, A. (1976). Implicit learning of synthetic languages: The role of
instructional set. Journal of Experimental Psychology: Human Learning
and Memory, 2, 88-94.

Reber, A. S., & Allen, R. (1978). Analogic and abstraction strategies in
syntactic grammar learning A functionalist interpretation. Cognition, 6,
189-221.
Reber, A. S., Allen, R., & Regan, S. (1985). Syntactical learning and
judgment, still unconscious and still abstract: Comment on Dulany,
Carlson, and Dewey. Journal of Experimental Psychology: General, 114,
17-24.
Schmidt, R. (1990). The role of consciousness in second language learning.
Applied Linguistics, 11, 129-158.
Schneider, W., Dumais, S., & Shiffrin, R. (1984). Automatic and controlled
processing and attention. In R. Parasuraman & D. R. Davies (Eds.),
Varieties of attention (pp. 45-79). London: Academic Press.
Schneider, W., & Shiffrin, R. M. (1977). Controlled and automatic human
information processing: I. Detection, search, and attention. Psychologi-
cal Review, 84, 1-66.
Schumann, J. (1983). Art and science in second language acquisition
research. Language Learning, 33, 49-76.
Sharwood Smith, M. (1981). Consciousness-raising and the second lan-
guage learner. Applied Linguistics, 2, 159-168.
Shiffrin, R. M., & Schneider, W. (1977). Controlled and automatic human
information processing: II. Perceptual learning, automatic attending,
and a general theory. Psychological Review, 84, 127-190.
White, P. (1982). Beliefs about conscious experience. In G. Underwood
(Ed.), Aspects of consciousness (pp. 156-187). London: Academic Press.
634 TESOL QUARTERLY

The Competence of Processing:

Classifying Theories of
Second Language Acquisition
ELLEN BIALYSTOK
York University
Theories of second language acquisition are frequently divided

into two groups: competence theories and processing theories.
These categories seem to be derived from the Chomskyan
distinction between competence and performance in language
use. This paper evaluates this distinction by examining Chomsky’s
definitions for competence and performance, and setting out
criteria for competence theories. This analysis leads to the need for
a reclassification of theories in these terms.
THE COMPETENCE OF PROCESSING:

CLASSIFYING THEORIES OF SECOND LANGUAGE ACQUISITION
Not all theories can be expected to do everything. With a few
notable exceptions, such as Newton’s three laws of physics, most
theories are fairly limited descriptions of specific problems. And
even Newton, as Einstein showed, was not exactly correct. Because
theories are inherently limited descriptions, the problem of saying
what a given theory is a theory of is crucial in the evaluation,
interpretation, and application of that theory.
Determining the correct limitations and classification for theories
is especially critical when the problem being addressed is in its early
stages of explanation, when there is a fuzziness about the scope of
the problem and its relation to other problems, and when existing
theories dealing with related problems are still imprecise. Because it
is a relatively young field, these conditions describe the problem
domain of second language acquisition. Hence, it is important to be
able to classify existing theories to understand what contribution
each makes to the field. Only with precise concepts of what infor-
mation is provided by different types of theories and how these will
eventually fit together can broader interpretations develop.
Several parameters have been used to classify theories of second
635
language acquisition. McLaughlin (1987) categorized theories ac-
cording to values on two dimensions: form and content. Regarding
form, he distinguished between theories that are primarily inductive
and those that are primarily deductive. Regarding content, he
distinguished between microtheories and macrotheories according
to their scope. Ellis (1986) implicitly used content to distinguish
among theories and identified the major theories in the field as re-
flecting different perspectives on the field, such as the acculturation
model, accommodation theory, discourse theory, and the like. Per-
haps a more principled way of describing this dimension is by the
linguistic subdiscipline upon which the theory is based. Hence the
categories would be sociolinguistic theories, psycholinguistic
theories, applied linguistic theories, and so on. This system, in fact,
is rather effective for distinguishing among theories as it clarifies the
ways in which the theories are different while legitimizing each as
a valid description of second language acquisition. What is less clear
is how the theories are to be integrated into a more complete
analysis of the problem.
COMPETENCE THEORIES AND PROCESSING THEORIES

The focus of this article is an analysis of the form of psycholin-
guistic theories of second language acquisition. The distinction at
issue is one that places a wedge between what have been called
competence theories on the one hand and processing theories on the
other (e.g., Spolsky, 1989). I shall argue that this distinction, as it has
usually been interpreted, has not been a productive means of
classifying these theories. Some theories that are usually considered
to be processing models for second language acquisition are easily
subsumed by the class of competence models. Clarification of the
difference between these types of explanations is required for them
to provide an effective means of classifying theories.
An evaluation of the metatheoretical distinction between compe-
tence theories and processing theories is important because of the
implications of this classification for theory building. First, the
designation sets out a particular course of investigation for research:
Each type of theory leads in a substantially different direction.
Commitment to one or the other type of theory has important
consequences for the type of questions pursued, research con-
ducted, and evidence accumulated. Clearly, differences of this sort
are important in shaping the field.
Second, the distinction cannot help but carry with it a value
judgment concerning the relative merit of each type of theory. This
value judgment traces its lineage to Chomsky who, in setting out the
636 TESOL QUARTERLY

possibility for such a distinction, awarded only competence his
attention. The implication of this favoritism is an imbalance in
research attention to all phases of producing a coherent description
for second language acquisition. In order to redress this imbalance,
it must be clear what aspects of the problem have been neglected.
Ultimately, both directions need to be defined in terms that make
them tenable avenues of study.
If competence theories and processing theories are to constitute
different categories, there must bean objective means for assigning
descriptions to these two classes. Yet, the criteria for distinguishing
between competence theories and processing theories have been
poorly formulated. In this article, two criteria for distinguishing
competence theories will be discussed. These are the definition of
competence and the role of idealization in theorizing. These criteria
will then be applied to standard processing theories which, it will be
argued, fit easily into the domain of competence theories. Finally, a
reappraisal of the relation between competence and processing will
be presented.
The usual way of distinguishing between competence and pro-
cessing models is to consider competence models as descriptions of
knowledge of rules for linguistic structure and processing models as
descriptions of methods of storage and means of accessing those
rules. Competence models, that is, provide descriptions of the set of
rules and the utterances that are produced by those rules without re-
course to any of the intervening mental operations. Processing
models tend to be descriptions of language use, in particular for
cases in which imperfect or nonstandard mastery is demonstrated.
Interlanguage is a case in point: It is marked by deviance and pro-
duced by the gap between competence and performance. Under
this view, any theory of interlanguage falls within the domain of a
processing model.
This interpretation of knowledge excludes two important aspects
of language acquisition and use. First, to merit study in these terms,
descriptions of linguistic knowledge, or linguistic competence must
form a complete and static system. Under this restriction,
descriptions of linguistic knowledge possessed by learners cannot
be the basis for competence theories since the knowledge in this
case is quite obviously incomplete. Second, evidence for variability
is excluded because it is interpreted as providing an imperfect ap-
proximation to the knowledge system. A complete and stable sys-
tem could not manifest the diversity known to characterize real lan-
guage use. Hence, descriptions of variability in language use are
relegated to the domain of processing theories. According to
this definition of knowledge, competence theories provide more
CLASSIFYING THEORIES OF SECOND LANGUAGE ACQUISITION 637

significant formulations with greater breadth and impact than do
processing theories.
The exclusion of incomplete knowledge and variability in
language use from the jurisdiction of competence theories,
however, does not necessarily follow from a stricter interpretation
of competence. First, there is an important difference between
linguistic knowledge that is static and linguistic knowledge that is
stable. The claim (from the processing point of view) is not that
interlanguage is static, but that it is stable. At a given point in time,
language learners have structured representations of their knowl-
edge of the language and these representations are used systemati-
cally to produce utterances. Moreover, this representational struc-
ture is not idiosyncratic but general across groups of language learn-
ers with similar backgrounds and experiences, and systematic
growth and development in this structure can be observed. These
premises form the basis for a competence theory at the centre of
which is some construct that may be described as interlanguage.
Second, variability need not undermine the attempt to produce
descriptions of learner’s competence. As I have argued elsewhere
(Bialystok & Sharwood Smith, 1985), variability takes two forms:
synchronic and diachronic. Synchronic variability is the alternation
in forms that a learner will employ at a single point in time. These
changes are typically motivated by social and contextual features of
the situation in which the language is being used. Such changes may
be explained by processing theories or, indeed, by production
models. Diachronic variability is the gradual change over time in
the way in which a learner represents linguistic forms. The learner’s
knowledge and conception of the linguistic system evolves,
presumably moving closer to some standard set of norms. This vari-
ability also manifests itself in changes in the learner’s performance,
but it is change that directly reveals competence. An explanation for
diachronic variability, while superficially part of a processing
model, must ultimately be formulatal in a competence theory.
In sum, the criteria for determining whether a model of second
language acquisition is providing an explanation at the level of a
competence theory or a processing theory are vague and conten-
tious. Minimally, competence theories are overly restricted by ex-
cluding incomplete knowledge systems and variability. Conse-
quently, traditional distinctions in the literature in which theories
are classified in this way are likely to be incorrect. In order to clari-
fy the way in which competence theories and processing theories
relate to each other and the type of description that each provides,
it is necessary to examine more closely the meaning of these terms.
638 TESOL QUARTERLY

COMPETENCE AND PERFORMANCE
The tradition of distinguishing between perfect knowledge of a
system and its flawed use in actual production can be traced of
course to Chomsky (1965). Perhaps his most well-known contribu-
tion is his distinction between competence and performance.
“Linguistic theory,” he states, “is concerned primarily with an ideal
speaker-listener, in a completely homogeneous speech-community,
who knows its language perfectly and is unaffected by memory
limitations, distractions, etc.” (p. 3). This definition of the proper
domain for linguistic theory led Chomsky to elevate competence to
the level of legitimate study while relegating performance to the
routine description of idiosyncratic use.
The legacy of this distinction has been enormous. One of its
applications has been as a means of classifying theories, producing
descriptions of language and language performance that are
considered to be either competence theories or performance theo-
ries. This division appears on the surface to be clear and applicable
to a variety of language studies: One can imagine, for example, that
studies of the nature of the linguistic system and its internal struc-
ture belong to the domain of competence descriptions, while
studies of the retrieval of that system during on-line performance in
real contexts are subsumed by the category of performance studies.
This distinction is extended to determine territorial boundaries
between disciplines; linguistics is in charge of setting out the
structure of the linguistic system and psychology is in charge of
determining how that system is used by the imperfect machinery
with which we are all burdened. Not only is this classification
ineffective, but it is very likely not what Chomsky intended. Its
impact on models of second language acquisition and use illustrates
its failure as a classification system and at the same time provides an
opportunity for the reconceptualization of the distinction.
First, let us attempt to reconstruct Chomsky’s intent in establish-
ing the competence-performance distinction. Chomsky is une-
quivocal in separating knowledge from ability, reserving compe-
tence only for knowledge. But Chomsky’s definition of knowledge
is more consistent with psychological than with formal linguistic
usage. Knowledge is considered in terms of mental structures, or
representations. It is not a description of the properties of an
external system that exists apart from a “knower.” He provides the
following explanation for what he intends by knowledge of
language:
To know a language, I am assuming, is to be in a certain mental state . . .
I assume further that to be in such a mental state is to have a certain

mental structure consisting of a system of rules and principles that
generate and relate mental representations of various types. (Chomsky,
1980, p. 48)
He goes on, however, to caution against excessive focus on the
“knower” in which the use of this system, observed through
performance data, becomes the basis for the correct description of
language competence:
Alternatively, one might attempt to characterize knowledge of language
. . . as a capacity or ability to do something. . . in which case one might
be led (misled, I think) to conclude that behavior provides a criterion for
the possession of knowledge. In contrast, if such knowledge is
characterized in terms of mental state and structure, behavior simply
provides evidence for possession of knowledge, as might facts of an
entirely different order—electrical activity of the brain for example.
(Chomsky, 1980, p. 48)
Chomsky’s use of the term knowledge in this way was unusual,
especially for a linguist. Even so, had he simply labelled this as
linguistic knowledge, or grammatical knowledge, the misinterpreta-
tions might have been minimized. Because he described this
knowledge as “competence,” the tendency has been to attribute to
the concept complete and intact faculties equipped with the
necessary knowledge to understand language that exist apart from
other mental systems. The term competence, that is, appears to
have the effect of placing the linguistic system outside of the
context and constraints of the mind. Chomsky is aware of this
confusion and explains his choice of term as follows:
The term competence entered the technical literature in an effort to
avoid entanglement with the slew of problems relating to knowledge,
but it is misleading in that it suggests ability-an association that I would
like to sever. (Chomsky, 1980, p. 159)
Chomsky, that is, intended competence to describe the linguistic
representations possessed by the ideal speaker/hearer of his or her
native language. He did not mean to refer to an abstract description
of linguistic structure, but rather to a psychological description of
mental structure:
By grammatical competence, I mean the cognitive state that encom-
passes all those aspects of form and meaning and their relation, includ-
ing underlying structures that enter into that relation. This is “the lan-
guage faculty,” (Chomsky, 1980, p. 59)
Solutions to the problem of how to describe the use of the linguis-
tic system by speakers and hearers takes two forms in Chomsky’s
model. The first, and trivial, solution is to relegate language use to
640 TESOL QUARTERLY

performance and then dismiss it as uninteresting. Such performance
models must be directly informed by competence models and take
account of “the structure of memory, our mode of organizing
experience, and so on” (Chomsky, 1980, p. 225).
The second and more serious attempt is through his construct of
pragmatic competence:
Pragmatic competence underlies the ability to use such knowledge along
with the conceptual system to achieve certain ends or purposes. It might
be that pragmatic competence is characterized by a certain system of
constitutive rules represented in the mind, as has been suggested in a
number of studies. (Chomsky, 1980, p. 59)
In this way, Chomsky defines both knowledge and the interesting
description of the use of knowledge in terms of mental structures
consisting of rules and representations. Both grammatical compe-
tence and pragmatic competence are part of the computational
system of language. More importantly, both are competence
theories.
Thus, Chomsky accepts, at least in principle, the importance of
some forms of production data and the potential relevance of
processing models, at least to the extent that they are descriptions of
pragmatic competence. He cautions as well that production data
are not privileged, that other kinds of evidence are equally impor-
tant. Nonetheless, their contribution is clear:
If we accept—as I do—Lenneberg’s contention that the rules of gram-
mar enter into the processing mechanisms, then evidence concerning
production, recognition, recall, and language use in general can be
expected (in principle) to have bearing on the investigation of rules of
grammar, on what is sometimes called “grammatical competence” or
“knowledge of language.” (Chomsky, 1980, p. 200)
Fodor (1981) goes (even further, stating that because the classifica-
tion is not principled, there are no such things as competence data
or performance data. Data do not come labelled for their
application. The basis of the distinction, rather, is a difference
between kinds of explanations.
Finally, although Chomsky speaks about performance as the
clear opposition to competence, it is less clear that the functioning
of the system he envisages treats them categorically as oppositional
terms. Rather, he implies that there is an intricate interdependency
between the two constructs. He argues that hard-line distinctions
between competence and performance characterize the division
between linguistics and psychology, each preferring to focus on a
different side of this division:

Thus, linguistics is taken to be the field that relies on informant
judgments, elicited material, whatever limited use can be made of an
actual corpus, and so on, to try to determine the nature of grammar and
universal grammar. Its concern is competence . . . Psychology, in
contrast, is concerned with performance, not competence; its concern is
the processes of production, interpretation, and the like, which make use
of the knowledge attained, and the processes by which transition takes
place from the initial to the final state, that is, language acquisition. To
me, this distinction has always seemed quite senseless. (Chomsky, 1980,
pp. 201-202)
In fact, he criticizes Kintsch et al. (1974) for making a firm and
dogmatic distinction of this type.
To summarize, Chomsky’s definition for competence contains
two features. First, competence is defined in terms of mental
structures—it is a psychological, or psycholinguistic conception of
the rules and representations (that later become principles and
parameters [Chomsky, 1986]) that exist in the minds of speakers
and hearers, enabling them to use a language. This “mentalistic”
aspect of competence is frequently overlooked, and the descrip-
tions of linguistic competence that result tend to be formal
descriptions of linguistic structure without attention to how those
structures might be represented in the minds of speakers.
The second feature of a competence description is that it involves
an idealization of the way a system functions under optimal
conditions. This idealization may not be an accurate description of
actual performance where mitigating factors interfere with this
optimal behavior, but the abstraction is necessary to create a
coherent model. Let us now examine this point in more detail.
IDEALIZATION AND THEORY BUILDING

Chomsky’s charge that the appropriate focus of study is the “ideal
speaker/hearer” led not only to the distinction between competence
and performance, but to the development of a theory that may be
described as an “idealization.” The theory, that is, was not intended
to model actual knowledge of real speakers (the actual way in
which real speakers use language), but the nature and structure of
the system in its ideal state. This attempt at idealization is
characteristic of competence theories. To what extent is such
idealization justified? Chomsky (1980) put the question this way:
So we are left with what must be the crucial question: does the
idealization so falisfy the real world that it will lead to no significant
insight into the nature of the language faculty, or does it, on the contrary,
642 TESOL QUARTERLY

open the possibility for discovering fundamental properties of the
language faculty? In short, is the idealization legitimate? (p. 25)
Macnamara (1990) argues that theories need to be idealizations in
order to have productive consequences. He defines idealizations as
the “limits towards which performance can tend but which are not
fixed by a physically observable reality” (p. 15).
Macnamara uses Chomsky’s distinction between linguistic
competence and performance as an important illustration of the use
of idealization in theorizing. The grammars in our heads that are
described in competence models may or may not be revealed in
actual performance because of production limitations. But,
continues MacNamara, this should in no way prevent processing
descriptions from being idealizations as well. Real progress, he
argues, is made only when one has postulated an ideal system and
attempted to understand it. He asks: “Is there any understanding the
mind in its logical aspects without access to the mind’s logical
ideas?” (Macnamara, 1990, p. 19).
Macnamara goes on to urge theorists to approach the problem of
description as an idealization and suggests the type of theory that
would result in a number of domains. None of these theories
addresses problems of functioning but rather considers competen-
cies in terms of the structure of knowledge and the logical possibil-
ities that follow. Competence theories by their very nature must be
idealizations.
EXPLANATIONS OF PROCESSING
If these two criteria for competence theories, namely, being
based on mental structures and constructed as idealizations, are
rigorously applied to theories in the field, then the classification into
processing theories and competence theories is drastically re-
ordered. To illustrate the problem, consider information-processing
theories of language development. These theories generally have
the following structure: Knowledge is represented in some
structured format that is usually made quite explicit. Performance
involves the accessing of that knowledge. Access is constrained by
extrinsic factors such as task demands, social context, experience,
and the like, although not determined by them. Performance, then,
is only an indirect (and undoubtedly) imperfect reflection of the
knowledge structures that were involved. But the description of
these knowledge structures is the central feature of information-
processing theories. Development (and to some extent, learning) is
generally explained in terms of the increasing richness of the

knowledge structures and improved cognitive operations for
accessing that knowledge.
On the surface, the mechanism outlined in these models would
suggest that they are properly classified as processing theories.
Consider, however, the two criteria described above for a compe-
tence theory: knowledge as mental structures and idealization.
The central claim of these models is the nature of the mental
representation of language and the changes that occur in the form of
that mental representation. The principal idea is that structural
changes in how information is represented are responsible for
advances in proficiency. While it is true that such models are silent
with regard to the specific content of linguistic structures, the claims
for linguistic representation are claims for the structure of the
system and not for its use. Questions such as how information is
accessed, how sentences are strung together in real time, and how
interpretations take place are generally not addressed. Rough
categories of types of processes are frequently proposed, such as
controlled and automatic processes, but these descriptions serve
only to classify performance outcomes into binary classes and not to
explain them. The models aim primarily to describe the mental
representation for linguistic knowledge in terms that support the
concept of representational change and fit with the broad
dimensions that distinguish between different types of perfor-
mance.
Second, the models tend to be idealizations of the structure of
linguistic representation and the dynamics of its transformational
properties. The instantiation of this sort of model in the mind of any
real language learner would undoubtedly differ from this abstract
description. The combination of systematic individual differences
and random contextual variation determines the specific features
that characterize a learner’s representation of linguistic knowledge.
Presumably, all such specific descriptions are instances of this gen-
eral idealization, varying only in detail. Moreover, actual perfor-
mance in any specific situation would almost certainly underesti-
mate the learner’s level of development of competence.
An example of a processing model that is in fact a competence
theory by these criteria is that of Jackendoff (1987). The problem he
addresses is how to describe the ontological status of consciousness
if we accept the assumption of mind being a computational
(computer-like) system, the standard metaphor in information-
processing psychology. In his solution to this problem, he describes
in abundant detail how such a system could function and solve the
problems of consciousness that he describes. Again, the terminology
is process-oriented and the mechanism seems to be a process; yet,
644 TESOL QUARTERLY

by the criteria described above, it is nonetheless a competence
theory. The theory is an idealization of how knowledge stored in
certain kinds of mental structures may have the potential for
supporting particular performances.
Although these information-processing models are classified as
competence models on the basis of criteria derived from Chomsky’s
work, there is no need to adopt the specific content of Chomsky’s
theory. The criteria of mentalistic definitions for knowledge and
idealization of operation apply irrespective of the specific content
of the representation. One need not accept claims of the centrality
of universal grammar or of a language faculty conceived as a
distinct cognitive entity. Indeed, there is evidence to reject both
these claims (see, for example, Jackendoff, 1987), yet both are
fundamental to Chomsky’s linguistic theory.
What, then, would constitute a processing theory for second
language learning? One example may ironically be some work that
Spolsky (1989) describes as a competence model. In accounting for
the variability of language use in different contexts, Tarone (1988)
argues that learners represent their knowledge along a continuum of
styles. The use of these styles is determined by social and contextual
factors. Accessing linguistic knowledge proceeds as a function of
the pragmatic context, and changes in this context determine the
style that is used. What changes along the continuum is the
likelihood that a particular rule will be used correctly. At some
points on this continuum, the rule may contain the instruction that it
be used in 100% of obligatory contexts, while at other points, the
instruction will be to use in only 50% of the obligatory contexts.
Since the knowledge of language is represented in a single
continuum, and it is the access to this continuum that determines the
likelihood of using the correct rule, the explanation is one of a
processing model. No specifications for competence or the nature
of the representational structures are offered.
A more far-reaching example of a processing theory for second
language learning is the use of parallel distributed processing
(PDP), or connectionist models. These models, developed by
Rumelhart and McClelland (1986) and applied to second language
acquisition by Sokolik (in this issue of the TESOL Quarterly)
provide a description of learning and cognition in the absence of the
need to represent a rule system. Competence, that is, is eliminated
from the system. Learning and performance are described proce-
durally in terms of connecting weights between nodes in complex
neural networks. The system is quintessentially a processing model.
Processing theories, according to this interpretation, have two
general features. First, they are neutral regarding the nature or

structure of the mental representations that underlie performance.
Not only will different kinds of formats for representing knowledge
be consistent with a given processing theory, but a specific format
does not need to be explicated for the processing theory to be
coherent. Competence theories require more detail on the nature of
representation. Second, processing theories are descriptions that
apply over a limited and specific point in time. They include,
therefore, on-line processing descriptions in real time but exclude
processes that occur over long or indefinite periods of time. For this
reason, they also exclude accounts of development. By these
criteria, theories of development are almost inevitably competence
theories.
When the contrast between competence models and processing
models is construed in these terms, the distinction becomes an
important one. First, it provides a means for assessing the relative
contribution of standard information-processing theories (see
Jackendoff, 1987) and PDP theories (see Rumelhart & McClelland,
1986). Since these models stand on different sides of this compe-
tence-processing distinction, the theories endeavour to do different
things. It is at least conceivable that at some point they could be
found to coexist, not as competing descriptions, but as complemen-
tary explanations for different problems.
Second, this differentiation between competence models and
processing models provides a set of conceptually distinct alterna-
tives for approaching psycholinguistic issues in second language
acquisition. Both alternatives are means for examining the mental
processes involved in second language use, but each entails
different theoretical traditions and employs different methodolo-
gies.
In classifications based on content or discipline, the theories are
also distinct in this way, but they stand as descriptions of different
aspects, or different appendages, of a larger problem: Psycholin-
guistic theories, sociolinguistic theories, applied linguistic theories—
all address individual parts of the larger problem, but each is
described in isolation from the others. In contrast, competence
theories and processing theories as defined here represent
conceptually distinct but intimately related approaches to a
problem. It is possible, then, to know what information is provided
by each to the larger issues of second language acquisition and to
anticipate how the two kinds of theories will eventually fit together.
This is the ideal situation for dichotomies: The two sets of options
are significantly different from each other yet related in ways that
are possible (at some point) to specify. It is possible, that is, for both
to be correct descriptions of a different problem or of a problem
646 TESOL QUARTERLY

differently construed. An analysis of content-based divisions reveal-
ing the connections that unite these theories as well as the differ-
ences that separate them would advance the field even further by
suggesting how the various pieces must eventually fit together. The
division between conceptually distinct but potentially reconcilable
alternatives is ultimately the sort of choice that metatheoretical
distinctions should offer.
ACKNOWLEDGMENT
This paper was prepared with the support of Grant #A2559 from the Natural
Sciences and Engineering Council (NSERC), Canada.
THE AUTHOR
Ellen Bialystok is Professor of Psychology at York University in Toronto, Canada.
Her research is in language acquisition, bilingualism, and cognitive development.
REFERENCES
Bialystok, E., & Sharwood Smith, M. (1985). Interlanguage is not a state of
mind: An evaluation of the construct for second language acquisition.
Chomsky, N. (1965). Aspects of a theory the syntax. Cambridge, MA: MIT
Press.
Chomsky, N. (1980). Rules and representations. New York: Columbia
University Press.
Chomsky, N. (1986). Knowledge of language: Its nature, origin and use.
New York: Praeger.
Fodor, J. A. (1981). Introduction: Some notes on what linguistics is about.
In N. Block (Ed.), Readings in the philosophy of psychology (Vol. 2,
pp. 197-207). Cambridge, MA: Harvard University Press.
Jackendoff, R. (1987). Consciousness and the computational mind.
Kitsch, W., Crothers, E. J., Glass, G., Keenan, J. M., McKoon, G., &
Monk, D. (1974). The representation of meaning in memory. New York:
Wiley.
Macnamara, J. (1990). Ideals and psychology. Canadian Psychology, 31,
14-25.

Edward Arnold.
Rumelhart, D. E., & McClelland, J. L. (1986). Parallel distributed
processing. Cambridge, MA: MIT Press.
Spolsky, B. (1989). Conditions for second language learning. Oxford:
Tarone, E. (1988). Variation in interlanguage. London: Edward Arnold.
648 TESOL QUARTERLY

The Least a Second Language

Acquisition Theory Needs
to Explain
MICHAEL H. LONG
University of Hawaii at Manoa
Theories of second language acquisition (SLA) are attempts to

explain well-attested empirical findings about relationships
between process and product in interlanguage development and
universals, and variance in learners and learning environments. An
important component of such theories will be one or more
mechanisms to account for interlanguage change. While theories
differ in scope and so often relate only to partial descriptions, they
must account for major accepted findings within their domains if
they are to be credible. Identification of “accepted findings,”
therefore, is an important part of theory construction and
evaluation. Such findings will be the least an SLA theory needs to
explain. Sample accepted findings on learners, environments, and
interlanguages are proposed along with some implications for
current SLA theories.
SECOND LANGUAGE ACQUISITION

SOME STRUCTURAL CHARACTERISTICS
Second language acquisition (SLA) is a relatively new, interdisci-
plinary field of inquiry. While several important studies appeared
much earlier (see Hatch, 1978a, for review), most empirical
research has been conducted since 1980 by researchers drawing
heavily (although some would say not heavily enough) upon theory,
research findings, and research methods in a variety of fields,
including education, psychology, linguistics, anthropology, foreign
languages, ESL, and applied linguistics. Data-based SLA research is
presented at a variety of conferences, most of which were originally
designed with focuses other than reporting SLA research results,l
1 Examples include TESOL, the International Association of Applied Linguistics (AILA), and
the American Educational Research Association (AERA). The Second Language Research
Forum (SLRF) is the only regular international conference (first held in 1977) devoted
exclusively to SLA research findings although two new organizations, the European Second
Language Association (EUROSLA) and Second Language Research Forum for the Pacific
(PacSLRF), will begin holding regular annual SLA conferences in l991 and 1992,
respectively.
649
and is published in a wide range of journals, only three of which
(Language Learning, Studies in Second Language Acquisition and
Second Language Research) are primarily devoted to it. Important
results often go unnoticed because they appear in obscure regional
publications or remain buried between the covers of master’s and
doctoral theses. There is very little funding available, and virtually
none at all in the U. S., where many SLA researchers work.
SLA’s brief history means that few issues have yet been
investigated exhaustively. The dearth of funding causes what tends
to be labor-intensive work to be conducted cross-sectionally (that is,
using data collected on different subjects at one point in time only,
instead of on the same subjects over time, or longitudinally) and on
small samples. The diversity of disciplines represented in the field
often results in skepticism about findings when research methods
from other traditions are used—from controlled laboratory
experiments, through work using interview data, grammaticality
judgments and other kinds of introspection, to case studies and
ethnographies. Also, SLA research with origins in one source
discipline, for example, theoretical linguistics, often seems
irrelevant to that inspired by developments in another, for example,
social psychology. To illustrate, it is difficult to relate research
findings on access to Universal Grammar in adult SLA motivated by
Chomsky’s ideas to the results of studies of nonnative speech
accommodation to an interlocutor motivated by Giles’ Accommo-
dation Theory. Finally, the shortage of specialist SLA conferences
and the fragmented publication of research findings makes it
difficult to review the literature to assess what is known about a
given topic. What is “the literature” on SLA?
DESCRIPTION AND EXPLANATION

IN THEORY CONSTRUCTION
It is often difficult to determine just what is known, or thought to
be known about second language acquisition. It becomes very
dangerous to claim that X is an established fact or that Y has
attained the status of a generalization or perhaps even of a law when
there is disagreement over what constitutes legitimate data and
when researchers and textbook writers are not reading or respecting
the same literature. Yet the identification of at least some uncontro-
versial results is a prerequisite for developing and evaluating
theories in any field. A synthesis of well-attested empirical findings
about process and product in interlanguage development related to
universals and variance in learners and learning environments is
essential for a valid description of SLA. The description delineates
650 TESOL QUARTERLY

the scope of the problem to be solved; it becomes part of the data
for which a theory needs to account and against which it may be
testable. The description specifies what is acquired; the theory
explains how.
Several qualifications are in order, however. First, not all forms of
theory attempt to explain how. Axiomatic and causal-process forms,
for example, do. The set-of-laws form does not. It consists of a com-
pilation of repeatedly observed patterns, but does not necessarily
seek to explain them (Reynolds, 1971; Long, 1985).
Second, what counts as an explanation varies from one discipline
and scientific subcomrnunity to another, and over time (Bunge,
1985; Cummins, 1983; Trusted, 1979). For some, such as behavior-
ists in several fields, biochemists, and many psychologists, explana-
tion means the empirically verified ability to predict future events
(either that they will occur or when they will occur); for others, such
as some ethnographers and anthropologists, it can mean post hoc
understanding of a single past event. For some, a purported
explanation must be empirically testable; for others, it need not, and
for still others, e.g., theorists in some branches of contemporary
physics, a theory cannot be tested due to the current unavailability
of technology required to conduct such a test.
Finally, the work undertaken to produce a description really does
more than provide a mere collation of the data to be explained. A
description has the beginning of explanation embedded within it;
explaining and theorizing are not separate activities, as is often
thought (Pronko, 1988). Description and explanation are better
viewed as two overlapping circles or as two points on a continuum
(Figure 1). What researchers select for observation is seldom
arbitrary, but a reflection of their own or others’ biases about what
FIGURE 1
Description and Explanation in Theory Construction
THE LEAST AN SLA THEORY NEEDS TO EXPLAIN 651

is likely to be important, or “worth studying.” That is, the choice is
already an implicit theoretical claim. In addition, with the possible
exception of certain kinds of constitutive ethnography (e. g., Mehan,
1978), what is incorporated into the resulting description is both far
less and far more than what was observed.
What emerges from a study is far less than what was observed
because researchers eliminate what they consider to be irrelevant
detail and draw attention to recurrent patterns. Thus, in the
following examples, it is safe to assume that subjects exhibited some
deviation from the norms, i.e., that there was variation in the data,
but this information has not been retained.
1. The frequency of no V constructions declined as that of don’t

V constructions increased.
2. Subjects’ suppliance of plural s was more target-like on the
picture description task than in the narrative.
3. Whether or not learners exhibited adverb-fronting on the
pretest predicted their control of particle separation after
instruction.
Descriptive statements like 1-3 (those in area A of Figure 1) are

ostensibly observations. They record that learning or some kind of
interlanguage change occurred or failed to occur and are neutral as
to how or why. Their very inclusion in a final report, however,
inevitably reflects the investigator’s initial assessment of their
potential significance for explanation. To the extent that messy
detail has been eliminated, they are already low-level generaliza-
tions about the data collected from particular samples although not
yet generalizations about behavior across studies or to populations.
On the other hand, what emerges from a study is far more than
what was observed because many descriptive statements about
patterns do take the form of generalizations and/or link two or
more variables in a way that implies a potential causal relationship.
To a greater or lesser degree, that is, they are abstractions from the
data, and abstraction is an essential step in constructing theories of
every kind. In the following examples, patterns are being claimed to
exist across samples (or across performance by the same subjects on
different task types) and/or across studies and in populations.
4. Accuracy was greater on tasks performed after planning than

on tasks performed with no planning.
5. After equivalent periods of exposure, child starters score
higher on proficiency tests than learners who begin as adults.
652 TESOL QUARTERLY

6. Constructions (such as topicalization) which involve move-
ment of an element from final to initial position are learned
before constructions (such as particle separation) which
require both disruption of a string and movement of an
internal element to a salient (initial or final) position.
Consciously or not, descriptive statements like 4-6 (those in area B

of Figure 1) already suggest at least low-level potential explanations
for the findings they record.
While descriptions are the basis for theory construction, theories
need not account for all the facts in every description to be viable.
Theories rarely purport to address every kind or aspect of SLA;
they vary greatly in scope. A particular theory may deal with
naturalistic, instructed, or mixed learning, with children or adults,
with specific language skills and modalities (oral or written,
comprehension or production), with a specific cognitive capacity or
resource (such as memory, attention, or aptitude), a specific psy-
cholinguistic process (such as transfer, restructuring, or stabiliza-
tion), a specific linguistic system (such as phonology, syntax, or
pragmatics), a specific subsystem (such as syllable structure, tense-
aspect-modality, relative clauses, or politeness), and so on. The vari-
ance in scope makes it legitimate for theories to relate to different
partial descriptions, or to selected findings in the field.
7. SLA is just one aspect of acculturation and the degree to which

a learner acculturates to the TL [target language] group will
control the degree to which he [sic] acquires the second
language. (Schumann, 1978, p. 34)
8. There are two independent ways of developing ability in a
second language. “Acquisition” is a subconscious process
identical in all important ways to the process children utilize in
acquiring their first language, while “learning” is a conscious
process that results in “knowing about” language. (Krashen,
1985, p. 1)
9. Second language learning, like any other complex cognitive
skill, involves the gradual integration of subskills as controlled
processes initially predominate and then become automatic.
(McLaughlin, 1987, p. 139)
In fact, as these statements (area C of Figure 1) show, in addition to

ignoring vast bodies of SLA research findings, theoretical claims of
greater or lesser scope, that is, explanations, may not refer explicitly
to SLA research findings at all.

MECHANISMS
A theory that referred to every accepted finding about SLA
would still not necessarily provide an explanation of the SLA
process, even in the unlikely event that all the findings were correct.
For explanatory power, a theorist needs to propose one or more
mechanisms to account for change. In the present context,
mechanisms are devices that specify how cognitive functions
operate on input to move a grammar at Time 1 to its new
representation at Time 2. The output of these mechanisms is
observable in learner data, in this case, an interlanguage sample.
Mechanisms in theories of first language acquisition are dis-
cussed, among others, by Atkinson (1982), McShane (1987),
MacWhinney (1987), and Jensen and Kiel (1988). Behaviorist
theories, they note, have relied upon data-driven mechanisms of
association, differentiation, and generalization. Innatist theories
typically employ some form of hypothesis testing constrained by
innate knowledge, for example, Chomsky’s language acquisition
device. Flavell (1972) provides a taxonomy of mechanisms
governing possible relationships between stages in a developmental
sequence, including addition, substitution, modification (either by
differentiation or generalization), inclusion, and mediation.
Mechanisms in the SLA literature to date are rather vaguely
defined and poorly supported. A partial list includes some borrow-
ings from first language research, plus restrictive and elaborative
simplification (Meisel, Clahsen, & Pienemann, 1981; Meisel, 1983),
conformative simplification (Stauble, 1984), regularization (Long,
1982), nativization and denativization (Andersen, 1983), and
hypothesis testing constrained by innate knowledge of language
universals, either syntactic (White, 1985; Flynn & O’Neil, 1989) or
semantic (Adamson, 1988; Bickerton, 1984).
In addition to specifying mechanisms driving development from
one stage to the next, Atkinson (1982) suggests that an adequate
explanation will also identify why the stages in a developmental
sequence have to occur in exactly the order they do, and cannot
occur in some other order (see also Johnston, 1985). Few proposals
of this type have yet been made in SLA theory, but one interesting
example is Rutherford’s attempt to explain various morphological
and syntactic accuracy and acquisition orders in terms of marked-
ness (Rutherford, 1982).
Another such SL example is that originally proposed by Meisel,
Clahsen, and Pienemann (1981) and Clahsen (1987) to account for
German as a second language word order, and later extended to a
654 TESOL QUARTERLY

variety of morphological and syntactic constructions in ESL
(Pienemann & Johnston, 1987). Meisel et al. and Clahsen claim that
the surface structures observed at each of five stages in the
development of German as a second language word order reflect
the cumulative shedding of three underlying processing strategies:
(a) the canonical order strategy, (b) the initialization-finalization
strategy, and (c) the subordinate clause strategy.
The strategy combinations are hierarchically related such that
each new one entails and adds to the sophistication of the previous
one, thereby gradually allowing psycholinguistically more complex
structures to be processed. At stage X + 1 (initialization-finaliza-
tion), for example, the learner can move elements from one salient
position to another (string-initial to string-final position or vice
versa), but only if this does not disturb the canonical word order.
Thus, an ESL learner can produce utterances like In Vietnam, I am
teacher (adverb-fronting). At stage X + 2 (disruption and move-
ment into a salient position), the learner is no longer constrained by
the canonical order strategy, and can move string-internal elements
to salient (initial or final) position, too, producing utterances like
Have you job? (yes/no inversion) and You take your coat off
(particle separation). Whatever the merits of this particular analysis,
it is an attempt to move beyond presentation of observed series of
structures as developmental sequences with no attempt to explain
why they occur in the order they do (for a critique of the model, see
White, 1989a).
SOME ACCEPTED SLA FINDINGS

Whether or not they refer to empirical findings or specify
explanatory mechanisms, theories have to account for well-attested
facts in their domain if they are to be taken seriously. A proposed
explanation for something that either ignored or could not account
for one or more of its most salient characteristics and/or for some
other well-established facts would lack credibility. For example, a
theory that birds can fly because they eat flying insects would
immediately be rejected because many flying birds do not eat
insects, many animals that eat flying insects cannot fly, and because
such an explanation would ignore a salient characteristic of all flying
birds, that is, wings. Theorists need to survey such known facts
when evaluating theories in a field and when constructing their own
theories. (See Spolsky, 1989, for a recent attempt at SLA theory
construction in this vein.)
There is an underlying assumption here that should be made

explicit, namely, that SLA research is at least partly a rational
endeavor. It is this quality that makes well-attested findings
important, not merely the fact that they are well attested.
It should be noted, however, that the rational nature of science in
general is not universally acknowledged. Proponents of an extreme
relativist, “anything goes” view of the growth of knowledge (e.g.,
Feyerabend, 1975) argue that voodoo, magic, or religious beliefs are
held no less certainly than are scientific beliefs, and that scientific
beliefs are different, not superior. As Gould (1981) and others have
shown, it is sometimes the case that the apparent superiority of
science over other belief systems arises simply from scientists as a
community deciding to value their work higher than others’. Then,
either they are in a position of power or represent elite interests with
sufficient power to impose their views on others.
The strength with which a belief is held and the value systems
underlying it are not the only relevant criteria for judging an
epistemology. As Newton-Smith (1982) has argued, scientists are
different precisely because they recognize the threat to the validity
of their work potentially posed by a lack of objectivity, unreliability
of observation and measurement, and the inevitable contribution of
the social construction of truth by closed communities. They,
uniquely, attempt to deal with the threat. They do this through
efforts to disengage researchers from their findings by establishing
minimum criteria that they must meet when making a claim. There
must be evidence for the claim, the evidence must have been
obtained using certain procedures, must be replicable by third
parties, and so on. The experiment is a key illustration of work
carried out within this approach, as is the importance of replication
and prediction. While scientists may never be able to know if they
have found the truth, increasing accuracy in their predictions is one
indication that they are getting closer to it.
Generalizations are only possible when empirical observations
are repeatedly supported. Generalizations are crucial to, and must
be accounted for by, credible theories. It is not just the existence of
consensus that makes generalizations important for theory construc-
tion, but the rationality of the scientific process that underlies them.
(For insightful discussion of these and related issues, see Beretta, in
press.)
Unfortunately, certain structural characteristics of SLA research
make known facts rather easier to determine with respect to how
birds fly than why some people can learn a second language and
others do not. Notwithstanding these difficulties, the following are
a few examples (by no means intended as a comprehensive list) of
what I would claim are well-established findings about learners,
environments, and interlanguages, along with some of the
656 TESOL QUARTERLY

challenges they pose to current SLA theories, Space limitations
preclude surveys of supporting literature. For each generalization,
references are provided to recent reviews and/or to key studies of
the phenomenon concerned, (The very existence of reviews, of
course, attests to the familiarity of many of the results.)
Learners
Wide variation in learners’ abilities (e.g., intelligence), states (e.g.,
motivation) and traits (e.g., extroversion) has relatively little effect
on most aspects of (first or second) language acquisition by young
children. Child language development is strikingly regular in both
course, rate, and ultimate attainment, and success is the norm
(Slobin, 1982). On the other hand, individual differences do affect
adult first (e.g., American Sign Language) or second language
acquisition. SLA processes and sequences are again fairly regular,
but learning rate and ultimate SL attainment are highly variable and
failure is common (Ellis, 1985; Newport, 1984).
Differences in learners’ starting age (Krashen, Long, & Scarcella,
1979; Scovel, 1988,), aptitude, attitude, and motivation (Skehan, 1989;
Spolsky, 1989), for example, are systematically related to variance in
rate of progress and ultimate attainment. The role of affective factors
appears to be indirect and subordinate to more powerful develop-
mental and maturational factors, perhaps influencing such matters as
the amount of contact with the L2, or time on task (Schumann, 1986).
The most positive attitudes to target language speakers and the
strongest motivation, for example, cannot overcome psycholinguistic
constraints on learnability at a particular stage of development
(Clahsen, 1987; Meisel, Clahsen, & Pienemann, 1981; Pienemann &
Johnson, 1987; Schmidt, in press) or maturational constraints on what
older starters can achieve (Long, 1990). Both L1 and L2 development
appear to depend on the same universal cognitive abilities (e.g., the
capacity for implicit and inductive learning) and to be subject to the
same cognitive constraints (e. g., limited human memory, attentional
resources, and information-processing capacity) (McLaughlin, 1987;
Schmidt, 1990; in press).
Environments
Variation in the linguistic environment has surprisingly little
effect on first language acquisition by children, where a high degree
of success is achieved even under conditions of quite severe linguis-
tic deprivation (Gleitman, 1986). The effect on adult language
learning of differences in the amount and kind of input available is

much greater and varies among different groups of learners, in part
as a function of L1/L2 relationships (Larsen-Freeman & Long,
1991). Both children and adults need the language they encounter to
be comprehensible for it to become potential intake (Krashen,
1985). Comprehensibility is not dependent on linguistic “simplifica-
tion” from the source (speaker/writer), which is often absent, but
may result from interfactional or elaborative modifications, which
are frequently the product of negotiation for meaning between the
source and the learners themselves (Hatch, 1978b; Long, 1983;
Parker & Chaudron, 1987). Exposure to comprehensible input is
necessary but not sufficient (White, 1987). Both children and adults
can learn from positive evidence alone, as evidenced by successful
untutored development in the absence of negative input, such as
overt error correction (Bley-Vroman, 1986), but a focus on form
(which overt error correction can sometimes induce in the learner),
along with any other behaviors or tasks that make certain L2
features salient, improves rate and ultimate SL attainment
(Doughty, in press; Long, 1988). Attention to form is necessary for
mastery of certain types of L1/L2 contrasts, for example, where the
way the L2 encodes a grammatical relationship is more marked than
the equivalent L1 structure (Eckman, 1981; Schachter, 1989), and
where the L1 allows two options (such as placement of frequency
adverbs before or after the direct object), only one of which is
grammatical in the L2, but both of which are communicatively
successful. This situation preempts negative input on the ungram-
matical item via repair sequences (White, 1989b). Noticing, brought
about by feedback, task structure, or other means, is necessary for
input to become intake, and negative evidence must be recognised
as such for it to be effective (Schmidt, in press). Much of a language
is not learned unconsciously.
Interlanguages
Interlanguages, the psycholinguistic SL equivalent of idiolects,
exhibit systematicity and variability at any time in their develop-
ment (Huebner, 1985; Selinker, 1969). The systematicity manifests
itself in many ways, including the regular suppliance and nonsuppli-
ance of both targetlike and nontargetlike features in certain linguis-
tic contexts and in the persistence of the same errors for often quite
lengthy periods (Sato, 1990; Schmidt, 1981). Interlanguages, that is,
are, or at least appear to be, rule-governed. Much of the variability
they also reveal turns out to be systematically related to such factors
as task, task requirements (e.g., attention to form and planning),
interlocutor, and linguistic context (Crookes, 1989; Hulstijn, 1989;
658 TESOL QUARTERLY

Kasper, 1988; Preston, 1989; Tarone, 1988). However, some of it
does appear to be random, or free, as when a learner produces no
put and don’t put or I born and I was born within moments of one
another under seemingly identical conditions (Ellis, 1985). Change
over time also follows predictable paths. With some differences for
first language background—L1 transfer being constrained by such
factors as L1/L2 markedness relationships and perceived
transferability (Eckman, 1985; Kellerman, 1984; Odlin, 1989; Zobl,
1982) —learners of different ages, with and without instruction, in
foreign and second language settings (Lightbown, 1983; Pica, 1983),
follow similar developmental sequences for such items as English
negation (Schumann, 1979), English and Swedish relative clauses
(Hyltenstam, 1984; Pavesi, 1986), German word order (Meisel,
Clahsen, & Pienemann, 1981), and a variety of other morphological
and syntactic constructions (Johnston, 1985). Progress is not linear;
backsliding is common, giving rise to so-called U-shaped behavior
observed in first and second language acquisition (Huebner, 1983;
Kellerman, 1985). Development is for the most part gradual and in-
cremental, but some sudden changes in performance suggest occa-
sional fundamental restructuring of the underlying grammar
(McLaughlin, 1990).
SAMPLE IMPLICATIONS FOR CURRENT SLA THEORIES

If the above can be considered a sample of “accepted findings” in
the sense indicated earlier, and so some of the facts in need of
explanation and/or constituting part of that explanation, a number
of implications follow for any theories that purport to be
comprehensive accounts of SLA. The following eight are offered by
way of illustration:
1. Common patterns in development in different kinds of learners
under diverse (conditions of exposure means that a theory that
says nothing about universals in language and cognition is
incomplete or, if considered complete, inadequate.
2. Systematic differences in the problems posed learners of
different L1 backgrounds by certain kinds of L1/L2 configura-
tions and by other qualitative features of the input, such as the
salience of certain linguistic features or lack thereof, means that
a theory that says nothing about environmental factors is
incomplete or, if considered complete, inadequate.
3. Differences in rate of acquisition and the level of proficiency
achievable by children and adults under comparable conditions
of exposure requires that viable theories specify either different
THE LEAST AN SLA ‘THEORY NEEDS TO EXPLAIN 659

mechanisms driving development in learners of different starting
ages or differential access to the same mechanisms.
4. The subordination of affective factors to linguistic and cognitive
factors means that a theory that purports to explain development
solely in terms of affective factors can, at most, be an account of
facilitating conditions, not an explanatory theory of acquisition
itself.
5. The need for awareness of and/or attention to language form for
the learning of some aspects of a SL means that a theory that
holds all language learning to be unconscious is inadequate.
6. The impossibility of learning some L2 items from positive evi-
dence alone means that a theory that holds that nativelike
mastery of a SL can result simply from exposure to comprehen-
sible samples of that language is inadequate.
7. Interlanguage systematicity, including adherence to regular
developmental sequences and systematic production of
nontargetlike forms never modeled in the input indicates a strong
cognitive contribution on the learner’s part and means that
environmentalist theories of SLA are inadequate.
8. The gradualist, often U-shaped course of much interlanguage
development renders inadequate a theory that assumes sudden,
categorical acquisition of grammatical knowledge triggered by
recognition of linguistic features of the input. A theory that
assumes that change is a product of the steady accumulation of
generalizations based upon the learner’s perception of the
frequencies of forms in the input is also incomplete.
CONCLUSION
It is perfectly reasonable for particular theories to discount or
ignore certain supposed empirical findings in the field because they
lie outside a theorist’s domain of interest or because a theorist’s
assumptions preclude the findings being correct and/or from
holding explanatory relevance. Nevertheless, a theory must account
for at least some of the major accepted findings within its scope if
it is to be useful. The same descriptions of findings to which a
theory is accountable may often simultaneously serve as the
beginning of an explanation for them, but an adequate SLA theory
also needs to specify one or more mechanisms to explain
interlanguage change.
Accepted research findings show that SLA is a multidimensional
phenomenon, with many (although by no means all potential)
660 TESOL QUARTERLY

learner and environmental variables determining variation in
developmental processes and product. Consequently, theories that
attempt to explain acquisition by recourse to a single factor (for
example, motivation, comprehensible input, or the workings of an
innate language acquisition device) or to a single type of factor (for
example, affect variables) lack face validity.
An explanatory theory of SLA that hopes to be viable will have to
be interactionist. That is to say, it will need to do two things. First,
it will need to recognize the role of both learner variables and
environmental variables in language development: Which aspects
of SLA are universal (presumably as a result of all learners
possessing common cognitive abilities and constraints), and which
aspects vary systematically as a function, for example, of age,
aptitude, and attention, or of the kind of input different learners
encounter? Second, it will need to specify which of those learner
and environmental variables exert a constant influence and which
ones interact, when and how. Which variables’ effect on develop-
ment is mediated by which others, and when? For example, is in-
creasing age of first exposure to the L2 always negatively related to
ultimate attainment, or only up to a particular age after which it
makes no difference? Are affective factors important in second
language settings, where, for example, positive attitudes may lead
learners to seek out and obtain more input or different kinds of
input, but only relevant for L2 literacy skills in foreign language
settings, where increased contact with native speakers may be
unavailable even when attitudes are positive? Is increased attention
to form useful only for adults or only for certain classes of linguistic
items? Does the structure of a learner’s L1 facilitate acquisition of
some features in a particular L2, impede acquisition of certain
others, and have no effect on still others?
As can be seen from the above examples, interactionist theories
are more powerful than unidimensional or single factor solutions in
their ability to account for the same set of data because they invoke
two or more (often many more) variables, types of variables, and
relationships among variables. This is always undesirable from a
theory construction perspective, where parsimony is valued.
However, the increase in power is clearly justified by the intriguing
combination of universals and variability in adult language learning,
which is the least an SLA theory needs to explain.

ACKNOWLEDGMENTS
This is a revised version of a paper presented at the 24th Annual TESOL
Convention in San Francisco, March 1990. I thank Graham Crookes, Alan Beretta,
and TESOL Quarterly editor Sandra Silberstein for helpful comments on earlier
versions.
THE AUTHOR
Michael H. Long is Professor of ESL and Chair of the PhD Program in Second
Language Acquisition at the University of Hawaii at Manoa. He is a member of the
Editorial Board of Studies in Second Language Acquisition and coeditor of the
Cambridge Applied Linguistics Series.
REFERENCES
Adamson, H. (1988). Variation theory and second language acquisition.
Washington, DC: Georgetown University Press.
Andersen, R. W. (1983). Introduction: A language acquisition interpreta-
tion of pidginization and creolization. In R. W. Andersen (Ed.),
Pidginization and creolization as language acquisition (pp. 1-56).
Rowley, MA: Newbury House.
Atkinson, M. (1982). Explanations in the study of child language
development. Cambridge: Cambridge University Press.
Beretta, A. (in press). Theory construction in SLA: complementarily or
opposition? Studies in Second Language Acquisition.
Bickerton, D. (1984). The language bioprogram hypothesis and second
language acquisition. In W. E. Rutherford (Ed.), Language universals
and second language acquisition (pp. 141-161). Amsterdam: John
Benjamins.
Bley-Vroman, R. (1986). Hypothesis testing in second language acquisi-
tion. Language Learning 36, 353-376.
Bunge, M. (1985). Types of psychological explanation. In J. McGaugh
(Ed.), Contemporary psychology: Biological processes and theoretical
issues (pp. 489-501). Amsterdam: Elsevier Science Publishers.
Clahsen, H. (1987). Connecting theories of language processing and
(second) language acquisition. In C. Pfaff (Ed.), First and second
language acquisition processes (pp. 103-116). Cambridge, MA:
Newbury House.
Crookes, G. (1989). Planning and interlanguage variation. Studies in
Second Language Acquisition, 11 (4), 367-387.
Cummins, R. (1983). Psychological explanation. Cambridge, MA: MIT
Press.
Doughty, C. (in press). Second language instruction does make a
difference: Evidence from an empirical study of SL relativization.
Studies in Second Language Acquisition.
662 TESOL QUARTERLY

Eckman, F. (1981). On the naturalness of interlanguage phonological rules.
Language Learning, 31, 195-216.
Eckman, F. (1985). The markedness differential hypothesis: Theory and
applications. In B. Wheatley, A. Hastings, F. Eckman, L. Bell, G. Krukar,
& R. Rutkowski (Eds.), Current approaches to second language
acquisition: Proceedings of the 1984 University of Wisconsin-Milwaukee
Linguistics Symposium (pp. 3-21). Bloomington, IN: Indiana University
Linguistics Club.
Feyerabend, P. (1975). Against method. London: Verse.
Flavell, J. (1972). An analysis of cognitive-developmental sequences.
Genetic Psychology Monographs, 86, 279-350.
Flynn, S., & O’Neil, W. (Eds.). (1989). Linguistic theory in second
language acquisition. Dordrecht, The Netherlands: Reidel.
Gleitman, L. R. (1986). Biological programming for language learning. In
S. L. Friedman, K. A. Klivington, & R. W. Peterson (Eds.), The brain,
cognition, and education (pp. 119-149). New York: Academic Press.
Gould, S. J. (1981). The mismeasure of man. New York: Norton.
Hatch, E. M. (1978a). Introduction. In E. M. Hatch (Ed.), Second language
acquisition: A book of readings (pp. 1-18). Rowley, MA: Newbury
House.
Hatch, E. M. (1978b). Discourse analysis and second language acquisition.
In E. M. Hatch (Ed.), Second language acquisition: A book of readings
(pp. 402-435). Rowley, MA: Newbury House.
Huebner, T. (1983). Linguistic system and linguistic change in an
interlanguage. Studies in Second Language Acquisition, 6, 33-53.
Huebner, T. (1985). System and variability in interlanguage syntax.
Language Learning 35, 2, 141-163.
Hulstijn, J. H. (1989). A cognitive view on interlanguage variability. In
M. Eisenstein (Ed.), The dynamic interlanguage (pp. 17-31). New York:
Plenum Press.
Hyltenstam, K. (1984). The use of topological markedness conditions as
predictors in second language acquisition: The case of pronominal
copies in relative clauses. In R. W. Andersen (Ed.), Second language: A
cross-linguistic perspective (pp. 39-58). Rowley, MA: Newbury House.
Jensen, K. A., & Kiel, E. (1988). Innateness and language acquisition: Two
perspectives. Pluridicta, 10, Odense University, Denmark.
Johnston, M. (1985). Syntactic and morphological progressions in learner
English [Research report]. Canberra, Australia: Department of Immi-
gration and Ethnic Affairs.
Kasper, G. (1988). Variation in interlanguage speech act realization.
University of Hawai’i Working Papers in ESL, 7 (2), 117-142.
Kellerman, E. (1984). The empirical evidence for the influence of the L1 in
interlanguage. In A. Davies, C. Criper, & A. P. R. Howatt (Eds.),
Interlanguage (pp. 98-122). Edinburgh: Edinburgh University Press.

Kellerman, E. (1985). If at first you do succeed . . . . In S. Gass &
C. Madden (Eds.), Input in second language acquisition (pp. 345-353).
Rowley, MA: Newbury House.
Krashen, S. D. (1985). The input hypothesis: Issues and implications. New
York: Longman.
Krashen, S. D., Long, M. H., & Scarcella, R. (1979). Age, rate, and eventual
attainment in second language acquisition. TESOL Quarterly, 13 (4),
573-582.
Larsen-Freeman, D., & Long, M. H. (1991). An introduction to second
language acquisition research. London: Longman.
Lightbown, P. M. (1983). Exploring relationships between developmental
and instructional sequences in L2 acquisition. In H. W. Seliger &
M. H. Long (Eds.), Classroom oriented research in second language
acquisition (pp. 217-243). Rowley, MA: Newbury House.
Long, M. H. (1982, August). Foreigner talk and early interlanguage: A
cross-linguistic study. Paper presented at the 2nd European-North
American Workshop on Cross-Linguistic Second Language Acquisition
Research, Gorde, West Germany.
Long, M. H. (1983). Linguistic and conversational adjustments to non-
native speakers. Studies in Second Language Acquisition, 5 (2), 177-193.
Long, M. H. (1985). Input and second language acquisition theory. In
S. Gass & C. Madden (Eds.), Input and second language acquisition
(pp. 377-393). Rowley, MA: Newbury House.
Long, M. H. (1988). Instructed interlanguage development. In L. M. Beebe
(Ed.), Second language acquisition: Multiple perspectives (pp. 115-141).
Cambridge, MA: Newbury House.
Long, M. H. (1990). Maturational constraints on language development.
Studies in Second Language Acquisition, 12 (3), 251-285.
McLaughlin, B. (1987). Theories of second language learning. London:
Edward Arnold.
McLaughlin, B. (1990). Restructuring. Applied Linguistics, 11 (2), 1-16.
McShane, J. (1987). Do we need a metatheory of language development?
Language and Communication 7 (2), 111-121.
MacWhinney, B. (Ed.) (1987). Mechanisms of language acquisition.
Hillsdale, NJ: Lawrence Erlbaum.
Mehan, H. (1978). Structuring school structure. Harvard Educational
Review, 48, 32-64.
Meisel, J. (1983). Strategies of second language acquisition: More than one
kind of simplification. In R. W. Andersen (Ed.), Pidginization and
creolization as language acquisition (pp. 120-157). Rowley, MA: Newbury
House.
Meisel, J., Clahsen, H., & Pienemann, M. (1981). On determining
developmental stages in natural second language acquisition. Studies in
Second Language Acquisition, 3 (l), 109-135.
Newport, E. (1984). Constraints on learning: Studies in the acquisition of
American Sign Language. Papers and Reports on Child Language
Development, 23, 1-22.
664 TESOL QUARTERLY

Newton-Smith, W. (1982). Relativism and the possibility of interpretation. In
M. Hollis & S. Lukes (Eds.), Rationality and relativism (pp. 106-122).
Oxford: Basil Blackwell.
Odlin, T. (1989). Language transfer. Cross-linguistic influence in language
learning. Cambridge: Cambridge University Press.
Parker, K., & Chaudron, C. (1987). The effects of linguistic simplifications
and elaborative modifications on L2 comprehension. University of
Hawai’i Working Papers in ESL, 6 (2), 107-133.
Pavesi, M. (1986). Markedness, discoursal modes, and relative clause
formation in a formal and an informal context. Studies in Second
Language Acquisition, 8 (l), 38-55.
Pica, T. (1983). Adult acquisition of English as a second language under
different conditions of exposure. Language Learning, 33 (4), 465-497.
Pienemann, M. (1984). Psychological constraints on the teachability of
languages. Studies in Second Language Acquisition, 6 (2), 186-214.
Pienemann, M., & Johnston, M. (1987). Factors influencing the development
of language proficiency. In D. Nunan (Ed.), Applying second language
acquisition research (pp. 45-141). Adelaide, Australia: National Curricu-
lum Resource Centre.
Preston, D. (1989). Sociolinguistics and the learning and teaching of foreign
and second languages. In D. Preston, Sociolinguistics and second language
acquisition (pp. 239-272). Oxford: Basil Blackwell.
Pronko, N. H. (1988). Explanation or description? In N. H. Pronko, From AI
to zeitgeist (pp. 55-6). New York: Greenwood Press.
Reynolds, P. (1971). A primer in theory construction. Indianapolis: Bobbs-
Merrill.
Rutherford, W. E. (1982). Markedness in second language acquisition.
Language Learning, 32 (l), 85-108,
Sato, C. J. (1990). The syntax of conversation in interlanguage development.
Tubingen, Germany: Gunter Narr.
Schachter, J. (1989). On the issue of completeness in second language
acquisition. Unpublished manuscript. University of Southern California,
Department of Linguistics, Los Angeles.
Schmidt, R. W. (1981). Interaction, acculturation and the acquisition of
communicative competence. University of Hawai’i Working Papers in
Linguistics, 13 (3), 29-77.
Schmidt, R. W. (1990). The role of consciousness in second language
learning. Applied Linguistics, 11 (2), 17-46.
Schmidt, R. W. (in press). Consciousness, learning, and interlanguage
pragmatics. In G. Kasper and S. Blum-Kalka (Eds.), Research in second
language pragmatics. Oxford: Oxford University Press.
Schumann, J. H. (1978). The acculturation model for second language
acquisition. In R. Gingras (Ed.), Second language acquisition and foreign
language teaching (pp. 27-50). Arlington, VA: Center for Applied
Linguistics.

Schumann, J. H. (1979). The acquisition of English negation by speakers of
Spanish: A review of the literature. In R. W. Andersen (Ed.), T h e
acquisition and use of Spanish and English as first and second languages
(pp. 3-32). Washington, DC: TESOL.
Schumann, J. H. (1986). Research on the acculturation model for second
language acquisition. Journal of Multilingual and Multicultural
Development, 7 (5), 379-392.
Scovel, T. (1988). A time to speak. A psycholinguistic inquiry into the critical
period for human speech. Cambridge, MA: Newbury House.
Selinker, L. (1969). Language transfer. General Linguistics, 9, 67-92.
Skehan, P. (1989). Individual differences in second language learning.
London: Edward Arnold.
Slobin, D. I. (1982). Universal and particular in the acquisition of language.
In E. Wanner and L. Gleitman (Eds.), Language acquisition: State of the
art (pp. 128-170). Cambridge: Cambridge University Press.
Spolsky, B. (1989). Conditions for second language learning. Oxford: Oxford
University Press.
Stauble, A-M. (1984). A comparison of the Spanish-English and Japanese-
English interlanguage continuum. In R. W. Andersen (Ed.), Second
languages: A cross-linguistic perspective (pp. 323-353). Rowley, MA:
Newbury House.
Tarone, E. (1988). Variation in interlanguage. London: Edward Arnold.
Trusted, J. (1979). The logic of scientific inference. London: Macmillan.
White, L. (1985). Universal grammar as a source of explanation in second
language acquisition. In B. Wheatley, A. Hastings, F. Eckman, L. Bell,
G. Krukar, and R. Rutkowsky (Eds.), Current approaches to second
language acquisition. Bloomington: Indiana University Press.
White, L. (1987). Against comprehensible input: The input hypothesis and
the development of L2 competence. Applied Linguistics, 8, 95-110.
White, L. (1989a, February). Processing strategies: Are they sufficient to
explain adult second language acquisition? Paper presented at the Ninth
Second Language Research Forum, University of California, Los Angeles.
White, L. (1989b, October). The principle of adjacency in second language
acquisition: Do L2 learners observe the subset principle? Paper presented
at the Boston Conference on Child Language Development, Boston.
Zobl, H. (1982). A direction for contrastive analysis: The comparative study
of developmental sequences. TESOL Quarterly, 16 (2), 169-183.
666 TESOL QUARTERLY

Extending the Scope of the

Acculturation/Pidginization
Model to Include Cognition
JOHN H. SCHUMANN
University of California, Los Angeles
This paper examines five cognitive models for second language

acquisition (SLA) and assesses how each of them might account
for the pidginized interlanguage found in the early stages of SLA.
The conclusion drawn from this analysis is that all the models offer
some account for how a pidginized interlanguage might be
formed, but none provides an adequate explanation of the
cognition that would be involved in pidginization. Based on this
assessment, the paper discusses criteria for evaluating cognitive
models for SLA. In addition, an argument is made that abstract
characterizations of learner behavior (Principles, rules, strategies,
etc.) must not be confused with the cognitive mechanisms that
cause the behavior. Finally, it is suggested that the structure and
function of the brain might be a good source of ideas about the
cognitive architecture and mechanisms involved in SLA and that a
complete general theory of second language learning will
ultimately have to contain an account of cognitive processing.
In the summer of 1987 after reading Barry McLaughlin’s (1987)

work on cognitive theory, I thought it might be interesting and
timely to try to extend the scope of the Acculturation/Pidginization
Model (Schumann, 1976, 1978, 1986; McLaughlin, 1987) to include a
cognitive component. I wanted to find a cognitive mechanism that
could account for early fossilization, particularly fossilization as
seen in pidginization. The assumption of the model was that if social
and psychological distance were high, a learner would not get the
requisite interaction and input to continue acquiring a second
language. Some cognitive mechanism would then operate on the
deficient input to produce the fossilized system. My procedure was
to examine cognitive models that had been proposed in the second
language acquisition (SLA) literature to see if such a cognitive
mechanism could be found.
667
McLAUGHLIN’S COGNITIVE THEORY
The first model examined was that proposed by McLaughlin
(1987) who applied Schiffrin and Schneider’s (1977) notions of
controlled and automatic processes to second language acquisition.
Controlled processing is seen as cognition under the attentional
control of the subject. This kind of processing capacity requires
time for activation and is generally limited to one or two items or
sequences at a time.
Automatic processes are learned responses that are activated
whenever the appropriate stimuli are present. They require
relatively little processing capacity and can be carried out in
parallel. They result from controlled processes that have been
extensively practiced, and once they are established, they are
difficult to alter or suppress.
In what ways might controlled and automatic processes result in
a pidginized interlanguage? One possible scenario is that the
deficient input picked up by the learner is maintained largely under
controlled processes and does not become automatic. This would
explain the labored, agrammatical speech characteristic of many
pidginized learners. Another scenario is that through extensive use,
the learner’s agrammatical speech becomes automatic, resulting in
the fluent agrammatical speech characteristic of other pidginized
speakers.
THE EXPERIENTIAL APPROACH

Hatch and Hawkins’ (1985) “experiential approach” to SLA
contains a cognitive component, which is based on ideas about
scripts proposed by Schank and Abelson (1977) and ideas about
language production proposed in Kempen and Hoenkamp’s (1987)
Incremental Procedural Grammar (IPG). Hatch and Hawkins
present a model of internal, integrated mental systems consisting of
a conceptualize, a formulator, and an articulator.
The conceptualize contains scripts, which are knowledge
structures built up through experience with sequences of events.
The events (such as going to the grocery store, registering at a
university, ordering a meal in a restaurant, depositing money in a
bank) become routinized and their mental representations allow the
speaker to operate with facility when situations involving them are
encountered. The scripts contained in the conceptualize have
language attached in the form of memorized routines or formulas
(for example, "Can I help you?"; “Are you kidding?”; “No way”; “I
see”; “I don’t know”). The conceptualize retrieves, activates and
668 TESOL QUARTERLY

sorts scripts; as the process continues, it sends fragments of the
scripts (and attendant pieces of language) to the formulator, which
puts the fragments together, selects lexicon and provides this
material with pragmatic and syntactic organization. The conceptual
fragments that are sent to the formulator are handled piecemeal,
with separate systems providing parallel processing of units such as
noun phrases, prepositional phrases, verb phrases. These formu-
lated structures are then passed on to the articulator where they are
queued in the order received and are supplied with appropriate
morphophonemics markers and phonological form.
How might the cognitive component in Hatch and Hawkins’
experiential model operate in the construction of a pidginized
interlanguage? If the learner is an adult immigrant, we can assume
that the learner’s conceptualize contains a great many scripts
acquired in experiences with the native language. However, he or
she will not only have to build new L2 scripts but also learn to fit
new knowledge into existing scripts. If learners are socially and
psychologically distant from the target language, then they may not
have sufficient interaction and input for the conceptualize to make
the appropriate additions and changes in their knowledge struc-
tures. Therefore, the formulator will receive deviant conceptual
fragments from the conceptualize. The formulator itself may not
be sufficiently developed to appropriately assemble the conceptual
fragments, choose lexicon, and provide pragmatic and syntactic
organization. In this case, it would likely fall back on L1
formulations or on some possible universal process such as “focus
last” (Perdue, in press).
Finally, we can imagine that the articulator would be insuffi-
ciently equipped to supply appropriate morphophonemics markers
and phonological form. This would result in a lack and/or misuse of
morphology and a foreign accent. From this whole process, then,
could very well emerge the nonsyntactic, morphologically skeleton-
ized form of speech characteristic of pidginized interlanguage.
BIALYSTOK AND RYAN’S KNOWLEDGE

AND CONTROL DIMENSIONS
Bialystok and Ryan (1985) have proposed a model of language
processing consisting of two dimensions: knowledge and control.
The knowledge dimension refers to the degree to which the mental
representations of the learner’s language knowledge are analyzed.
The assumption is that the learner’s knowledge is structured with
rules of grammar, semantics, phonology, and discourse, and that the
learner has varying degrees of awareness of that structure. This
EXTENDING THE ACCULTURATION/PIDGINIZATION MODEL 669

awareness is not necessarily conscious in the sense that the learner
can articulate the rules although, at the highest levels on the analysis
dimension, this may be possible. The dimension of analyzed
knowledge is simply the learner’s awareness of the elements
(lexicon, morphology, syntax, pragmatics) that comprise the mental
representations of language knowledge. Nonanalyzed knowledge,
in its extreme, consists of memorized patterns and routines, the
elements of which are not part of the learner’s awareness. The
analysis dimension, however, is not dichotomous; rather, it
constitutes a continuum consisting of degrees of analysis of
language information.
The control dimension refers to the degree of automaticity in the
selection and coordination of language information. Essentially
automaticity constitutes the fluency with which learners can
retrieve the language information contained in their knowledge
structures. Higher levels of automaticity enable learners to display a
more skilled performance. Once again the automaticity dimension
constitutes a continuum extending from low control to high control.
Confusion may develop when attempting to associate Bialystok and
Ryan’s (1985) notion of “control” (p. 221) with McLaughlin’s (1987)
notion of “controlled processing” (p. 135). McLaughlin’s “automatic
processing” (p. 134) should be associated with Bialystok and Ryan’s
“high control” (p. 218), and McLaughlin’s (nonautomatic) “con-
trolled processing” (p, 135) should be associated with Bialystok and
Ryan’s (nonautomatic) “low control” (p. 218). Finally, the analysis
and control dimensions are independent of each other, and it is
possible for a learner to have a high degree of analyzed knowledge
with low control and vice versa.
How then might analysis of knowledge and degree of control as
cognitive processing dimensions explain pidginized varieties of
interlanguage? We can assume that insufficient interaction with and
input from target language speakers caused by low acculturation
would result in a knowledge base that might contain representations
of linguistic structure that are not correct by target language
standards (Bialystok & Sharwood Smith, 1985). In addition, the
quantity of this knowledge base might be relatively small,
consisting of a limited lexicon, morphology, and syntactic structure.
However, this knowledge representation may be highly analyzed,
thus permitting flexibility and creativity, or it may be largely
unanalyzed and thus limited to relatively rigid and stereotypical
representations of language structure. Control over either type of
representation may be high or low. These two dimensions allow for
speakers of pidginized interlanguage varieties that are
670 TESOL QUARTERLY

1. halting (i.e., low control) or fluent (i.e., high control) expressions
of memorized patterns and routines (i.e., unanalyzed); or
2. halting or fluent expressions of a knowledge base that is
agrammatical with respect to the target language but flexible,
creative, and perhaps systematic (i.e., analyzed) with respect to
its own norms.
ACT* (ACTIVE CONTROL OF THOUGHT)

O'Malley, Chamot, and Walker (1987) have suggested that J. R.
Anderson’s (1983) ACT* (read ACT star, the asterisk is meant to
indicate that this is considered the last version of the model)
framework offers a perspective from which cognition in second
language might be viewed. Anderson distinguishes between
declarative knowledge, which consists of memory for facts and
events, and procedural knowledge, which consists of skills. All
knowledge enters the system in the declarative form and is
represented in memory as temporal strings (such as phone numbers
and lists), special images, or abstract propositions. The information
in declarative memory is organized into associative networks
consisting of nodes with links between them. An element in the
network becomes active when it encodes an environmental stimulus
or an internal goal (an intention). By a process called spreading
activation, activation moves from these active nodes to other nodes
in the network. The whole network is viewed as long-term memory,
and the portion of it that is active constitutes working memory.
Procedural knowledge consists of productions that are condition/
action pairs. For example:
IF the goal is to describe a topic
and the listener does not know the topic
THEN set as subgoals
1. To generate the indefinite article, a
2. To describe the topic
(based on Anderson, 1983, p. 262)
If the data patterns in the condition clause match the active
information in working memory, then the production applies and
the action clause specifies what is to be done.
There are three stages of procedural learning in the ACT* system:
a declarative stage, a compilation stage, and a tuning stage. In the
declarative stage, interpretive procedures guide behavior. For
example, if the information is in the form of instructions, the
procedure might be to follow the instructions step by step. If the

information is less direct, it can be used by general problem-solving
procedures or analogy-forming procedures (i.e., modeling pro-
cedures for a new behavior on those of already learned behaviors).
In the compilation stage, the heavy burden on working-memory
capacity is eased by two processes: composition and proceduraliza-
tion. Composition collapses a sequence of productions into a single
production. For example, this process might combine the condition
clauses and action clauses of two productions into one production:
Pl: IF the goal is to spell a word and letter 1 is the first letter
of the word, THEN say letter 1.
P2: IF the goal is to spell a word, and letter 1 has just been
spoken, and letter 2 is after letter 1 in the word, THEN
say letter 2.
P1/P2: If the goal is to spell a word, and letter 1 is the first letter
of the word, and letter 2 is after letter 1, THEN say letter
1 and then letter 2.
Proceduralization eliminates the need both to retrieve information
from long-term memory and to hold it in working memory. Instead
it creates a knowledge-specific procedure. For example,
P* If the goal is to spell sleep THEN say S–L–E–E–P.
The asterisk indicates that the production that follows is the product
of the composition and proceduralization of the preceding
productions. With the original productions P1 and P2, any word
could be spelled. P* only spells sleep. Anderson points out that
proceduralization produces phenomena associated with automati-
zation (Shiffrin and Schneider, 1977).
Compiled tasks can still be refined by three tuning mechanisms:
(a) generalization, which broadens the applications of production
rules, (b) discrimination, which narrows them, and (c) strengthen-
ing, which reinforces successful productions and weakens less
successful ones.
How then might Anderson’s model account for pidginized
interlanguage? First, we can imagine that the reduced interaction
and input resulting from low acculturation would produce a good
measure of inaccurate and incomplete declarative knowledge of the
target language. In addition, it would produce knowledge about
which the learner is uncertain. Interpretive procedures operating on
this knowledge would result in a labored and halting pidginized
variety. However, if the defective declarative knowledge were then
to undergo knowledge compilation through extensive communica-
tive demands, then the inadequacies in syntax, morphology, and
672 TESOL QUARTERLY

lexicon could become fossilized (Cushing, 1989). This situation
would result in the fluent agrammatical speech that characterizes
certain pidginized varieties.
CONNECTIONIST LEXICAL MEMORY

Another model of cognition in second language acquisition is
Gasser’s (1988) connectionist model of sentence generation. This
framework, called the Connectionist Lexical Memory (CLM),
reflects recent thinking about cognition that is inspired by the neural
networks that make up the brain. These models, when they are
learning models and are implemented on a computer, attempt to be
brain-like in the sense that they contain highly interconnected
networks that acquire knowledge through experience, and the
networks are not programmed with rules that already represent the
knowledge. Gasser’s model is not a learning model, but it falls
within the connectionist framework because it contains no rules as
such; instead, the system’s knowledge is maintained in a network of
processing units and weighted connections.
Gasser conceives of the mental structure of cognition as a
network of schemas, There are three types of schemas for sentence
production: those for concepts, those for generalized utterances
(GUs), and those for generalized illocutions (GIs). Figure 1
represents a concept schema. The head node of the schema is on the
left (DEPOSIT-IN-BANK); to the right are role nodes (object,
recipient), and to the far right are role values (MONEY, BANK),
The DEPOSIT-IN-BANK schema is an instance of a more general
concept schema, TRANSFER-OF-CONTROL.
Linguistic knowledge is represented by GUs and GIs. Figure 2
depicts part of a GU for the *DEPOSIT-IN-BANK schema. The
asterisk is used to distinguish the GU names from concept names.
The GU has roles for linguistic information such as verb and direct
object. In Figure 2, we can conceive of the GU as being embedded
in the concept schema for DEPOSIT-IN-BANK and connected to
the general schema, TRANSFER-OF-CONTROL.
GIs are planning schemas that represent the plan a speaker has to
achieve a goal. They are specializations of a more general INTEND
schema. Figure 3 shows a GU that represents a plan by a speaker to
get someone to undertake something for him/her. It specifies that
actor is also the hearer, and that the allocution will consist of a yes-
no question of the “could you” form where you is the subject and
could is a modal.
Figure 4 shows the three types of schemas embedded in a large
network that would be involved in the generation of the utterance,

FIGURE 1
Representation of a Concept *
FIGURE 4
Partial Semantic Net for “Could You” Request
From The Relevance of Connectionism to Theories of Second Language Acquisition (p. 72) by
C. Fantuzzi, 1988. Unpublished master’s thesis. Los Angeles University of California. Copyright
1988 by C. Fantuzzi. Adapted by permission.
“Could you deposit this money for me?” In the CLM model the
generation process begins when a set of nodes representing the
speaker’s intention fires. Activation spreads from the firing node to
nodes connected to it. Each node in the network has an activation
level that varies from –1 to +1. When a node reaches its threshold
activation, it fires and sends activation along its connections. After
a node has fired, it enters a refractory period during which it cannot
be affected by further inputs. This prevents infinite feedback loops
from neighboring nodes.
Schemas are selected via winner-take-all (WTA) networks. That
is, when a node fires it sends out inhibition to all competing nodes,
preventing them from also firing. For example, if the TRANSFER-
OF-CONTROL node fired, activation would be sent to several GUs
such as *DEPOSIT-IN-BANK, *WITHDRAW-FROM-BANK,
*GIVE, and *LEND. However, if the information in the network
caused *DEPOSIT-IN-BANK to fire, that GU would be selected
and *WITHDRAW-FROM-BANK, *GIVE, and *LEND would be
inhibited. In this sense, *DEPOSIT-IN-BANK is the “winner” that
“takes all” (the activation).
In the CLM model, another important aspect of utterance
generation is role binding. For example, in Figure 2, the semantic
object role (money) of the *DEPOSIT-IN-BANK schema has to be
bound to the direct object role of the schema. This binding is
accomplished by having both role nodes fire more or less
simultaneously. Sequencing is also crucial in this cognitive model
where processing is parallel. In the CLM model, sequencing is
accomplished by having each constituent or phrase bounded by a
start node and an end node.
Now we want to ask how the CLM model could account for
pidginization in second language acquisition. We might assume that
an adult second language learner would have all the necessary
conceptual schemas, perhaps lacking only those that are specific to
the target language culture. However, because of social and
psychological distance from the target language group and the
concomitant reduced input and intake, we could expect the learner
to have deficient GUs and GIs. The former might lack appropriate
linguistic information. For example, constituent roles might not
have target language morpheme values or might maintain incorrect
ones. This would lead to either the skeletonized interlanguage or the
nontarget morphological systems that characterize pidginized
varieties.
Deficient GIs would result in planning schemas that would not
contain accurate information on what sorts of utterances satisfy
what sorts of goals. The learner may have the appropriate goal, that
676 TESOL QUARTERLY

is, may know what mental state or behavior of the hearer the learner
wants to achieve, but may not have the appropriate plan, that is, the
general type of utterance that can accomplish this goal. So from the
perspective of the CLM, it would be deficiencies in GUs and GIs
that would cause pidginization in early second language acquisition.
THE COGNITIVE INTERACTIONIST MODEL

Roger Andersen (1988) has been developing a perspective that he
calls the Cognitive Interactionist Model for second language
acquisition. The major feature of this approach is its cognitive
operating principles in the style of Slobin (1985). Andersen’s
operating principles, however, are not merely translations of
Slobin’s formulations to second language acquisition; rather, they
derive from careful observation of the linguistic behavior of
learners acquiring English and Spanish as second languages. In
Andersen’s view these cognitive operating principles “govern the
path the learner takes in developing an increasingly more native-like
and more efficient and successful competence in a second
language” (p. 117). Andersen has developed seven principles, and to
illustrate his view of cognition I will discuss four of them.
The One-to-One Principle states that “an interlanguage system
should be constructed in such a way that an intended underlying
meaning is expressed with one clear invariant surface form (or
construction)” (p. 117). Anderson offers several pieces of evidence
for this principle.
First, learners of German, in the early stages, tend to maintain a
subject-verb-object word order even though native German under
certain conditions requires placement of the verb before the
subject, placement of nonfinite verb forms in clause final position,
and placement of finite verb forms in final position in subordinate
clauses.
Second, in the acquisition of negation, learners of English,
French, German, and Swedish tend initially to place the negator
before whatever entity is to be negated. Only later do they learn the
more complicated rules for negation formation that exist in each of
these languages.
Finally, in the acquisition of target language temporal systems,
second language learners tend initially to restrict past tense
morphology to punctual verbs. Only later is the morphology
extended to verbs that are durative.
The second principle, the Multifunctionality Principle states:
(a) Where there is clear evidence in the input that more than one form

marks the meaning conveyed by only one form in the interlanguage, try
to discover the distribution and additional meaning (if any) of the new
form. (b) Where there is evidence in the input that an interlanguage
form conveys only one of the meanings that the same form has in the
input, try to discover the additional meanings of the form in the input.
(Andersen, 1988, p. 119)
Andersen believes that one aspect of second language development
is the movement from the One-to-One Principle to the principle of
multifunctionality. This development can be illustrated by the
acquisition of English negation. Initially learners place no before
whatever they want to negate. This frees them to notice that forms
such as don’t and not also express negation in English, and learners
extend their negation to these forms as well. Andersen also reports
that in the untutored acquisition of Spanish, the past preterit form is
initially restricted to punctual verbs. Later, learners extend the use
of this form to verbs that are both punctual and durative and finally
to completely nonpunctual verbs.
Another operating principle is the Transfer to Somewhere
Principle:
A grammatical form or structure will occur consistently and to a
significant extent in the interlanguage as a result of transfer if and only
if (1) natural acquisition principles are consistent with the L1 structure or
(2) there already exists within the L2 input the potential for (mis-)
generalization from the input to produce the same form or structure.
Furthermore, in such transfer, preference is given in the resulting
interlanguage to free, invariant, functionally simple morphemes which
are congruent with the L1 and L2 (or there is congruence between the L2
and natural acquisition processes and [to] morphemes [which] occur
frequently in the L1 and/or the L2. (Andersen, 1988, pp. 125-126)
This principle can be illustrated in the placement of object
pronouns by English learners of French and French learners of
English. Both languages place full object noun phrases after the
verb, but in French, object pronouns are placed before the verb.
When English speakers learn French they initially tend to place
object pronouns after the verb. Their model for this is the identical
French and English post verbal placement of full object noun
phrases. French learners of English, however, do not place object
pronouns preverbally because there is no model in the English input
to promote such transfer.
The final principle to be discussed is the Relexification Principle:
When you cannot perceive the structural pattern used by the language
you are trying to acquire, use your language structure with lexical items
from the second language. (Andersen, 1988, pp. 126-127)
678 TESOL QUARTERLY

This principle tends to apply most strongly in environments where
the target language input is highly restricted, but is rarely if ever
categorical. It would appear to be a default principle to which
learners resort when they have insufficient guidance from the target
language input.
As mentioned earlier, Andersen has developed seven principles of
which four have been discussed here. Andersen will add to the
model as evidence for new principles is accrued from interlanguage
studies. (Slobin now has over forty such principles for first language
acquisition.) The Cognitive Interactionist Model is still developing;
nevertheless, even among the four principles that have been
examined in this paper, we can see how a pidginized interlanguage
might be explained within Andersen’s framework.
The Relexification Principle would apply to the degree that target
language input was restricted by the learner’s lack of interaction
with speakers of the target language. That is, we can expect that the
nonacculturating learner will, in many cases, be unable (or even
unwilling) to perceive the structural patterns of the target language
and, therefore, will follow the structures of the native language,
lexifying these with acquired target language words.
One would also expect that the pidginizing learner would main-
tain the One-to-One Principle and might never (or only minimally)
move on to multifunctionality. The Transfer to Somewhere Princi-
ple and the Relexification Principle would interact, with relexifica-
tion perhaps dominating, and with Transfer to Somewhere reinforc-
ing that tendency in situations where natural developmental princi-
ples coincide with the native language structure and where some of
the limited target language intake is (misperceived as matching the
native language. For example, a Spanish speaker who has fossilized
a pidgin variety where the negative particle no precedes all negated
elements (e.g., no go, no here, no tall, no went) would provide
evidence for Andersen’s principles as follows:
1. The One-to-One Principle because the interlanguage contains
only one invariant form for negation;
2. The Transfer to Somewhere Principle because (a) placing the
negative particle before the negated entity appears to be a
natural acquisition principle, and (b) the English negative no
matches the Spanish negator and therefore can lead to the
misgeneralization that English negation is like Spanish
negation where no precedes all negated items;
3. The Relexification Principle because placing an English word
after no results in the Spanish structural pattern with target
language words.

If a pidginized French speaker, whose native language has
negation in the form of ne verb pas, were to use the same form, only
1 and 2a above would apply.
DISCUSSION
The preceding analysis demonstrates that each of these models
provides us with a cognitive perspective on how a pidginized
interlanguage might be formed. At issue is which model should be
chosen to bring cognitive processing within the scope of the
Acculturation/Pidginization Model. What is needed are criteria for
evaluating the various cognitive perspectives. One possible criterion
might be the type of cognitive mechanism and architecture that the
particular model proposes.
Long, in this issue of the TESOL Quarterly, states that a theory of
SLA must include mechanisms, devices that specify “how cognitive
functions operate on input to move a grammar at Time 1 to its new
representation at Time 2, the output [of which] is observable in
learner data, in this case, an interlanguage sample.” But in addition
to mechanisms, a cognitive system also must have an architecture,
that is, an overall structure in which the cognitive operations take
place. Metaphorically we might view architecture as a blueprint or
design of the “machine” or structure in which the cognitive
mechanisms operate.
Studies of second language acquisition, however, are generally
studies of behavior. Researchers look at behavior in interlanguage
development, and on that basis, they speculate about what internal
mechanisms may have caused the behavior. But as the neurobiolo-
gists, Changeux and Dehaene (1989) have pointed out, “reconstruct-
ing architecture on the basis of external observation alone is a
complex matter, if not an ill-defined task, with no unique solution”
(p. 98). A further difficulty is ensuring the characterizations of
external behavior are not confused with the internal mechanisms
that cause that behavior.
To illustrate the distinction between descriptions of behavior and
models of cognition, I will briefly examine two of the cognitive
models discussed above, Roger Andersen’s Cognitive Interactionist
Model and Michael Gasser’s Connectionist Lexical Memory.
Andersen’s model does not specify any particular mechanism or
architecture. It consists of a series of operating principles that are
essentially abstract summaries of learner behavior derived from
careful observation of learner interlanguage. These summaries of
behavior are then thought of as mental processes. In the One-to-One
680 TESOL QUARTERLY

Principle, the observation is made that at some stage of interlan-
guage development, learners use a single invariant form to express
a particular meaning. For example, learners of English when
expressing negation always place the negative particle no before the
entity to be negated, or they may only use past tense morphology on
punctual verbs. Because this behavior is consistent and widespread,
the behavior itself is characterized as a general principle and is seen
as a cognitive process. I suspect this is the best we can do on the
basis of external observation alone. Limiting ourselves to external
behavior means that the mechanisms we postulate must bear a close
relation to observable behavior, but in fact, the actual cognitive
mechanisms may be very different.
The same point can be made for the mechanisms referred to in
Long’s paper. Long considers mechanisms to be such things as
generalization, differentiation, simplification and regularization.
(Long himself is clearly not happy with these designations. In his
paper in this collection, he says the mechanisms “in the SLA
literature to date tend to be rather vaguely defined and poorly
supported.”) Are these indeed mechanisms? Are they not instead
legitimate abstract labels or characterizations of behavior? If they
are mechanisms, what does the machine look like? What are these
things mechanisms of?
Gasser’s model provides a very detailed architecture for utterance
production in a second language. The mind, as depicted in this
model, consists of networks of highly interconnected schemas.
There are schemas for concepts, utterances, and illocutions. They
are related to each other hierarchically such that schemas lower on
the hierarchy inherit the characteristics of those higher up. For
example, the DEPOSIT-IN-BANK concept inherits the characteris-
tics of the more general TRANSFER-OF-CONTROL schema. The
utterance generation process is implemented by spread of activa-
tion, which causes the nodes to fire. Schemas are selected through
winner-take-all procedures. Role binding is accomplished by
simultaneous firing of nodes, and sequencing is maintained by
bounding linguistic units with start and end nodes. Note that none
of these mechanisms are isomorphic with behavior. In the model,
they cause the behavior.
What about the brain? Do we believe that we can understand the
cognitive mechanisms of SLA without understanding their neurobio-
logical bases? All the models discussed in this paper are essentially
“brainless.” Gasser’s Connectionist Lexical Memory is neurally
inspired; it consists of a network of processing units analogous to the
brain’s neuronal networks, but this constitutes very minimal neuro-
biological inspiration. As Churchland and Sejnowski (1988) point out,

it would be convenient if we could understand the nature of cognition
without understanding the nature of the brain itself. Unfortunately, it is
difficult if not impossible to theorize effectively on these matters in the
absence of neurobiological constraints. The primary reason is that
computational space is consummately vast, and there are many
conceivable solutions to the problem of how a cognitive operation could
be accomplished. Neurobiological data provide essential constraints on
computational theories, and they consequently provide an efficient
means for narrowing the search space. (pp. 744-745)
If we accept this view, we have very little basis for adopting any
of the models presented in this paper except as useful conceptual
frameworks or depictions that allow us to think about cognition in
SLA in different ways. This, of course, is no small contribution. An
appropriate philosophy for exploring cognition in SLA might
involve the notion that all cognitive models are metaphorical
depictions that are valued as vehicles for thinking about the
question of how cognitive processes operate in SLA, but that we do
not expect or demand that they provide ultimate answers to those
questions. We would then hope that future cognitive models of SLA
would incorporate neurobiological considerations and augment our
current metaphors of the mind with metaphors based on the brain.
In concluding, I would relate this discussion specifically to the
issue of the scope of a theory of second language learning. I think
that we cannot have a general theory or a complete theory of
second language learning without including cognitive processing
within its scope. At the same time, I have to agree with Spolsky
(1989):
A[n] . . . important question is whether or not a theory of second
language learning needs to be a processing model, proposing a working
model of exactly how language learning takes place. I think the answer
is, not yet . . . for the challenge remains to specify exactly how such a
model could work in the human brain as we know it. And . . . it is too
early to do this with any feeling of certainty. (pp. 4-5)
It would appear that models of cognition in second language
learning may have to be worked on independently, or they may
have to be developed as components of less general models of
second language learning. For example, Roger Andersen’s Cogni-
tive Interactionist approach, while not offering a detailed mental
architecture, may be a very useful way to think about cognition
within the framework of his Nativization Model for second
language acquisition. And for the Acculturation/Pidginization
perspective, I have proposed a neurocognitive component
(Schumann, 1990) in which the amygdaloid complex of the brain,
682 TESOL QUARTERLY

which is important in both emotion and memory, can be seen as a
mediator between affect and cognition in acculturation and second
language acquisition.
It may be the case that cognitive processing cannot, at this time,
be satisfactorally incorporated into a complete general theory of
second language learning. But, the paradox is that without such a
cognitive component, the theory can be neither general nor
complete.
THE AUTHOR
John H. Schumann is the Chair of the Department of TESL and Applied
Linguistics at the University of California, Los Angeles. His research includes the
study of the social and psychological factors in second language acquisition, the
acculturation process, and pidginization. His most recent work has been in the area
of the neurobiology of emotion and cognition in second language acquisition.
REFERENCES
Andersen, R. A. (1988). Models, processes, principles, and strategies:
second language acquisition in and out of the classroom. IDEAL, 3, 111-
138,
Anderson, J. R. (1983). The architecture of cognition. Cambridge: Harvard
University Press.
Bialystok, E., & Ryan, E. B. (1985). A metacognitive framework for the
development of first and second language skills. In D. L. Forrest-
Pressley, G. E. MacKinnon, & T. Gary Waller (Eds.), Metacognition,
cognition and human performance (pp. 207-249). Orlando, FL: Aca-
demic Press.
Bialystok, E., & Sharwood Smith, M. (1985). Interlanguage is not a state of
mind: An evaluation of the construct for second-language acquisition.
Changeux, J-P, & Dehaene, S. (1989). Neuronal models of cognitive
functions. Cognition, 33, 63-109.
Churchland, P. S., & Sejnowski, T. J. (1988). Perspectives on cognitive
neuroscience. Science, 242, 741-745.
Cushing, S. T. (1989). ACT theory: A description and implications for
second language acquisition. Unpublished manuscript, University of
California, Los Angeles, Department of TESL and Applied Linguistics.
Fantuzzi, C. (1989). The Relevance of Connectionism to Theories of
Second Language Acquisition. Unpublished master’s thesis, University
of California, Los Angeles.

Gasser, M. E. (1988). A connectionist model of sentence generation in a
first and second language. Unpublished doctoral dissertation, University
of California, Los Angeles.
Hatch, E., & Hawkins, B. (1985). Second-language acquisition: An
experiential approach. In S. Rosenberg (Ed.), Advances in applied
psycholinguistics (Vol. 2, pp. 241-283). New York: Cambridge
University Press.
Kempen, G., & Hoenkamp, E. (1987). An incremental procedural
grammar for sentence formulation. Cognitive Science, 11, 201-258.
McLaughlin, B. (1987). Theories of second-language learning. Baltimore:
Edward Arnold.
O’Malley, J. M., Chamot, A. U., & Walker, C. (1987). Some applications of
cognitive theory to second language acquisition. Studies in Second
Language Acquisition, 9, 287-306.
Perdue, C. (in press). The European Science Foundation project on adult
language acquisition. In C. A. Ferguson & T. Huebner (Eds.), Second
language acquisition: Contributions and challenges to linguistic theory.
Cambridge, MA: The MIT Press.
Schank, R., & Ableson, R. (1977). Scripts, plans, goals, and understanding.
Hillsdale, NJ: Lawrence Erlbaum.
Schneider, W., & Schiffrin, R. M. (1977). Controlled and automatic
processing: I. Detection, search, and attention. Psychological Review,
84, 1-64.
Schumann, J. H. (1976). Second language acquisition: The pidginization
hypothesis. Language Learning, 26, 135-144.
Schumann, J. H. (1978). The acculturation model for second language
acquisition. In R. C. Gingras (Ed.), Second language acquisition and
foreign language learning (pp. 27-50). Washington, DC: Center for
Applied Linguistics.
Schumann, J. H. (1986). Research on the acculturation model for second
language acquisition. Journal of Multilingual and Multicultural Develop-
ment, 7, 379-392.
Schumann, J. H. (1990). The role of the amygdala in mediating affect and
cognition in second language acquisition. In Georgetown University
Round Table on Language and Linguistics 1990 (pp. 169-176).
Washington, DC: Georgetown University Press.
Schiffrin, R. M., & Schneider, W. (1977). Controlled and automatic
processing: II. Perceptual learning, automatic, attending, and a general
theory. Psychological Review, 84, 127-190.
Slobin, D. I. (1985). Crosslinguistic evidence for the language making
capacity. In D. I. Slobin (Ed.), The Crosslinguistic study of language
acquisition (pp. 1157-1256). Hillsdale, NJ: Lawrence Erlbaum.
Spolsky, B. (1989). Conditions for second language learning. New York:
684 TESOL QUARTERLY

Learning Without Rules: PDP

and a Resolution of the Adult
Language Learning Paradox
M. E. SOKOLIK
Texas A & M University
Parallel distributed processing (PDP) or connectionist models are

systems that “learn without rules,” thus providing an alternative to
more traditional symbolic grammars. In this paper, a common
problem in adult second language learning—the disparity between
child and adult second language learning—here called the Adult
Language Learning Paradox (ALL Paradox) is examined within
this framework. The alternative analysis examined here offers
some insight into this problem by taking into consideration the
influence of neurological constraints on language learning and the
parallel processing that is intrinsic to brain function.
Rules seem to be an omnipresent, if not omnipotent, factor in the

study of human behavior in general and language in particular. As
researchers, we often see our task as one of discovering rules, that is,
distilling behavior into appropriate propositional generalizations.
As teachers, we frequently present rules in order to impart
condensed knowledge to our students. And as students, we try to
use rules to make our task of learning an easier, more efficient one.
Given this situation, how exactly does one approach the task of
learning without using rules? Before attempting to answer these
questions, it might be helpful to begin with the converse; that is,
what are the specific proposals for learning with rules?
We begin with a definition of the term rule. A rule is a
formalization of an operation that manipulates symbols. Symbols
take the form of mental representations, for example, a word or an
idea which are manipulated by computational algorithms. Rule-
based processing typically is done in a step-by-step or serial fashion,
for example, “If Monday, pick up dry-cleaning.” This type of
system may take the form of a generative grammar, an expert
system, and so on.
685
The goal inherent in rule writing is to capture relevant
generalizations in the most parsimonious form possible; the result
may be extremely abstract symbolic grammars. As noted by
MacWhinney, Leinbach, Taraban, and McDonald (1989), in the
case of the early formulations of English phonological rules, these
rules were
highly symbolic, having numerous subconditions, alternative environ-
ments, and variables. Moreover, the forms upon which those rules
operated were highly abstract, often with no direct relation to any
actually occurring phonological form. (p. 255)
Generative and other symbolic grammars have also embraced this
type of representation, with null elements and empty nodes making
up an important part of the system.
It has been successfully argued, however, that these sorts of
symbolic grammars cannot be acquired by children through mere
inductive learning or hypothesis testing. Thus, in order to acquire
the rule-based system of language, that is, to learn to use rules, at
least in the generative sense, one must arrive into this world with an
innate ability to manipulate symbols with specific types of
operations. Exposure to the data (in the case of language, the
spoken word) “triggers” the appropriate rule, or parameter setting,
or whatever the current instantiation of the theory requires.
Precisely how this triggering operates in the brain is not specified.
At one level, there is nothing wrong with the idea of innate
knowledge and rule formation. In fact, it would be surprising to
learn that there is no innate component to human language acquisi-
tion. However, at the behavioral level, there is no direct evidence
that these types of symbols are indeed manipulated by rule-like
operations in the brains of language learners (Ervin, 1964; Hockett,
1968; Slobin, 1971). Furthermore, at the physiological, or neuronal
level, there are certain difficulties in trying to account for the
evidence using rule-like structures.
First, from what we know about the structure and operation of
the neurons and synapses in the brain, we know that certain
cognitive tasks—for example, answering a simple true/false
question—take only about a half of a second. If the individual steps
needed to complete this task were done using a serial, rule-based
system, that is, step one followed by step two, and so on, the
process would likely take much longer (Churchland & Sejnowski,
1989).
Second, it is also known that the brain is a highly interconnected,
parallel system, where neurons A, B, and C are probably all
mutually interconnected. One consequence of this design is that if
686 TESOL QUARTERLY

one area of the system fails, the result may only be a slight decrease
in efficacy of the entire system. In a serial system, where neuron A
would be connected to B, which in turn would be connected to C,
if one step fails (for example, the connection between B and C), the
entire system is likely to be disrupted.
Thus, if we compare the parallel and serial models to the human
nervous system, which we know is a somewhat plastic and
redundant structure that can in many cases compensate for failures
in part of the system (for example, the recovery of function after a
brain injury), the parallel model is obviously the better match. Thus,
we shall now turn to an examination of the framework known as
Parallel Distributed Processing (PDP) and consider how it might
affect our understanding of second language learning.
PARALLEL DISTRIBUTED PROCESSING

Although often touted as a new model of cognition and learning,
the ideas underlying PDP models are hardly new. The roots of
connectionist ideas run deep in the fields of psychology and
neurology and have been around for decades. What is new are the
ongoing attempts to apply the principles of PDP to varied fields of
interest.
When we talk about PDP or connectionist models, we are really
speaking of a group rather than one specific model. However, the
basic principle of all models in the group is the same: Connectionist
or parallel models are systems that behave intelligently without the
explicit manipulation of symbol systems, that is, the learning of
rules.
Within the field of language acquisition, models have been
developed to accomplish such tasks as to produce English past tense
verbs (Rumelhart & McClelland, 1987) and appropriately case/
gender/number-marked German articles (MacWhinney et al.,
1989), to recognize spoken language (Elman & McClelland, 1984)
and the gender of French nouns (Sokolik & Smith, 1989), and to
read aloud from a printed text (Sejnowski & Rosenberg, 1987). This
article examines the phenomenon of second language acquisition
(SLA) in light of this “new” theory of learning and investigates the
application of the principles of PDP to a familiar problem found in
SLA.
The Basics of the Model

The architecture of PDP models is inspired in part by the
structure of the brain (Clark, 1989). These models have also been
PDP AND THE ADULT LANGUAGE LEARNING PARADOX 687

demonstrated to exhibit some of the robust pattern-recognition
capabilities characteristic of human cognitive processes (Rumelhart
& McClelland, 1987). That is, at least in some respects, they share
structural and functional characteristics with the human brain.
In a PDP network, information or knowledge is coded by a
specific pattern of activity distributed over a set of low-level
featural nodes or units (analogous to neurons in the brain). A
processing module might comprise two or more layers of such units.
Any two layers can be linked by interconnections of individual units
(analogous to synapses) across layers, and activity may be shared
between units via these interconnections. The strength or weight of
connectivity between specific pairs of units varies, and thus the
degree to which activity is shared between pairs of units also varies.
Cognitive transformations in these types of architectures do not
rely upon explicit rule systems. Instead, the transformation a given
module performs is implemented as a mapping of activity from one
layer of units to another; the form this transformation takes is
determined by the pattern of weights that link the units at one level
to those at the next. For example, the activity of a unit in one layer
will have a relatively large positive influence on the activity of a unit
in the next layer if those two units are interconnected by a relatively
large positive weight, or conversely, a small or negative (inhibitory)
influence if the weight is small or negative.
The ability of a given processing module to accomplish anything
useful is determined by the pattern of its interconnection weights.
The critical developmental issue is how to determine an appropriate
set of weights for a particular desired transformation. One option is
to specify the weights a priori, if an appropriate set of weights to
accomplish the transformation is known in advance. However, a
more naturalistic approach is to allow the model to evolve a
functional set of weights in response to its interaction with a
stimulus environment. Thus, in many connectionist models the
weights are modifiable over time in a manner that reflects the
processing history of the system. That is, the strengths of
connectivity between units in different layers can be construed as
learned associations. When equipped with a suitable modification
function, such networks possess the ability to learn complex
mappings between representations. That is, given an arbitrary set of
pairs of input and output patterns, such networks can learn to
produce an appropriate output pattern when presented with a
particular input pattern. This type of pattern recognition is
performed automatically without any “look-up” procedure or
explicit application of if/then-type rules.
One common PDP architecture is often referred to as a “pattern
688 TESOL QUARTERLY

associator.” Pattern associators are among the most basic types of
PDP architectures, where an input layer (IL) is connected by a
single layer of modifiable connections to a set of output layers (OL).
PDP architectures of this kind have been studied extensively as
learning paradigms for both linguistic and nonlinguistic domains
(Anderson, 1983; Kohonen, 1984; Hinton & Anderson, 1981;
McClelland, Rumelhart & the PDP Research Group, 1986).
Several properties of pattern association networks are consistent
with those minimally required for models of human learning
mechanisms. For instance, they can learn to abstract features
common to a set of inputs; they can generalize these features to the
recognition of novel patterns that are similar to those in the studied
set; and their disruption in the face of damage or noise is graded
rather than complete—that is, when connections are damaged, the
system continues to function, albeit in a degraded fashion.
PDP AND SECOND LANGUAGE ACQUISITION

We can concretely illustrate the model outlined above by
applying it to a recurring problem in second language acquisition
studies: How can we explain the apparent difference between child
and adult second language acquisition? (For another discussion of
PDP and SLA, see Gasser, 1990.)
Learning Rate and Nerve Growth Factor

In any type of learning, an adjustment must be made between the
actual output of the learner and the desired output or target. The
following example shows how a simple learning procedure,
expressed as a mathematical formula, uses the difference between a
desired or “target” activation and the obtained activation to guide
learning in the system. The equation shows that the new weight
(state of knowledge) equals the prior weight (former state of
knowledge) plus a fraction of the difference between the actual and
desired activation level (the amount learned from the error) on the
output node. This formula is stated as
1. w ij (t) = w ij (t – 1) + re i a j
where wti (t) refers to the new weight (new output) between OL u i
(u= unit) and IL u j , w ij (t – 1) refers to the prior weight (old
output) for this connection, r represents a learning rate parameter
(that is, the speed with which the system learns—this learning rate

is generally expressed as a decimal fraction between 0 and 1), and
e i represents the difference (that is, the difference between the
target and the actual output), or “error” for OL u, given by
In this formula, t i is the target activation on OL u i and a i is the

obtained activation computed as in equation 1.
In the model, the correct set of weights is approached gradually
as the training procedure continues through several sweeps
(commonly termed epochs) through some list of inputs, such as
English past tense verbs, German nouns, etc. Each time the model
receives new bits of data, it adjusts what it knows based on that
information.
As described above, the model adjusts its knowledge at a rate that
is determined by the variable r, or the learning rate parameter. This
portion of the model is particularly important to the argument
which follows. In the example presented below, the size of the
learning rate variable in humans is considered to be linked to the
availability of some nerve growth factor (NGF) in the brain
(Purves, 1986; Fine & Rubin, 1988). In other words, the rate at
which learning in humans takes place is, in part, dependent upon
the amount of this substance present in the system. Thus, a higher
level of NGF would translate into a higher speed (or value of r) of
learning.
Research in NGF is fairly new, and so this interpretation must be
accepted with certain caveats. However, this parallel is drawn
based on the fact that children, who presumably have higher levels
of NGF, generally undergo a larger recovery of brain function after
injury than do adults (Lenneberg, 1967). This implies diminished
capacity for neural growth in adults. The relationship of the nerve
growth factor to child and adult second language acquisition is
shown in the following example.
Child Versus Adult SLA

It is a broadly accepted generalization that children acquire
second languages more readily than do adults. There is, of course,
variation implicit in this generalization, but most would concur with
the basic claim. This generalization, however, poses a further
problem for the concept of rule-driven language systems. We will
call this the Adult Language Learner Paradox (ALL Paradox),
which is stated below:
Given the superior ability of adult humans in comparison to children to
generalize and extract rules from their experience (Inhelder and Piaget,
690 TESOL QUARTERLY

1958), and given that human language is rule-governed, it would follow
that adults should be better able to learn a second language than are
children. However, our observations of actual second language learning
contradict this inference.
We will see how a PDP model contributes to resolving this paradox.
The general details of the model have already been discussed. We
now turn to a specific example showing how this model might be
applied to a problem in second language learning.
To illustrate this problem, we take the following hypothetical
example: A three-year-old English speaking child and 25-year-old
English-speaking parent are in France. They are both attempting to
learn Feature Xf of French. Recalling the earlier formula, we will
say, arbitrarily, that each learner begins with a starting weight of .2
for this feature. Complete knowledge of Feature Xf would be
represented by a connection weight of 1 between an IL unit and an
OL unit. We will posit .25 as the learning rate value associated with
the availability of NGF to the 3-year-old, and .05 as that available to
the 25-year-old adult.
In Figure 1, the results of this procedure are calculated and
plotted.
In this figure, the x-axis represents the number of epochs, or the
amount of time, that it takes to complete the task. The first line
illustrates the hypothetical 3-year-old, with a higher level of nerve
growth factor, and thus a higher learning rate. The second line
represents the learning of the 25-year-old, with a lower level of
NGF available. The effect of the difference in learning rate is that
the 3-year-old reaches a point of near unity (.95), that is, has
acquired Feature Xf after 10 epochs or exposures to the data while
the adult takes 55 epochs to reach the same point.
It is important to note that this model does not take into account
all of the psychological and sociological factors that may influence
learning rates. It merely shows how physiological differences in the
brains of children and adults could affect the rate of time with
which new information is assimilated.
To summarize briefly, based on the decreased availability of
NGF and/or the interference of hormones after puberty, specific
features of language learning in adults may be slower than in a
child. This is reflected in the age-dependent learning rate: the
observed disparity between adult and child learning. This interpre-
tation is echoed in the following two quotations:
The high plasticity frequently observed in the early stages of a learning
process may reflect, at least in part, an expected reduction in the
modifiability of neuronal response characteristics as their synaptic
connections are strengthened as a result of experience. (Munro, 1986,
p. 471)

692 TESOL QUARTERLY
In accordance with what is known about SLA across different ages, it
seems plausible that, within relative time limitations, the earlier one
begins acquiring a second language, the greater the chance that this
second language will become involved in plasticity of a formational/
organizational nature. (Jacobs, 1988, p. 324)
Similarly, the degree to which connections are alterable, even in
adults, may be genetically determined to some degree. This could
account for the variation that is observed between “good” and
“poor” language learners of all ages.
In summary, PDP models can offer an alternative explanation for
one recurrent theme in explaining the facts of second language
learning.
NEW THEORIES FOR OLD

One criticism of the models of development in PDP is that the
computer models start with some form of tabula rasa, which would
seem to disallow any innate linguistic abilities as a driving force in
language acquisition. One example that is traditionally cited as
proof of some innate set of rules is the absence of linguistic
operations such as “reverse the order of words in a declarative
sentence to form a question.” This type of algorithm is not
automatically ruled out by a connectionist system. In fact, it could
be readily learned by such a model with the appropriate set of
input. However, it is notable that the “form questions by word
order reversal” algorithm could be learned easily by humans as well
(as a sort of language play), indicating that although it is not a
feature of any known natural language, it is not unlearnable. There
is a lot of evidence of linguistic play by chidren in particular, where
this type of algorithm is used (for example, Pig Latin, or what I
called as a child “Bubble Talk,” where speech is produced by
inserting the letter B before the vowel in each syllable—thus, “How
are you?” becomes “H-Bow B-are y-B-ou?)
In fact, there is nothing inherent to PDP that requires the initial
state to be a system of zero weights or randomly defined ones.
There is certainly an initial state, a predefine, prenatally and/or
genetically determined set of weights or connections in the human
by the time of birth (Jacobs, 1988). The problem posed to all
cognitive science is, of course, what that initial state is (Fodor,
1983), and what contribution it makes to the end result (Rumelhart
& McClelland, 1986).
Just as the connectionist theory does not disallow some
preformed connections, it is likely that there are some connections
that cannot be altered. For example, one would not want the

connections responsible for important bodily functions—heartbeat,
respiration, and so on—to be easily altered. Thus, it seems possible
that there is also a genetic predisposition, or set of preprogrammed
connections, toward certain types of linguistic structures, and these
connections may not be alterable.
Finally, there is the problem that arises concerning the term
learning itself. In the PDP model, learning is simply the setting and
adjustment of weights within- the network structure. However, in
many areas of SLA research, the notion of some mechanism to
evaluate learning has been proposed such as Krashen’s monitor and
Bialystok’s concept of control. That is, a second system is posited
that involves awareness of the learning process. In the PDP
literature, this system has been termed “deliberate conscious
control” (DCC) (Norman, 1986, p. 543). It works in tandem with the
parallel system but is described as a serial system that is “not very
effective for some kinds of skills” (Norman, 1986, p. 544). This
introspective process may contribute to the changing of developing
weights, but the individual does not have any direct access to the
reasons a particular state in the system was reached. Further
specification of the DCC system is necessary for a more complete
and satisfactory model of learning.
CONCLUSION
In this paper, I have attempted to present a model that is able to
learn without the use of explicit rules. One motivation for using this
framework is the observation that in second language, in particular,
rules do not seem to be consistently helpful in explaining learning.
In fact, there is something of a paradox in the system—adults,
typically better at rule learning than children, usually learn second
languages less proficiently than do prepubescent children. In this
modal, the paradox is explained by the greater availability of nerve
growth factors in children, which is hypothesized to affect language
learning rate.
Using as a model the subsymbolic paradigm, we can retain a
concept of innateness as the initial state of the neurons and their
connections in the system. The learning of a first (or second) lan-
guage can be interpreted as the changing, strengthening, and
reorganization of some of those connections as a result of exposure
to the data, perhaps guided by some degree of self-evaluation and
conscious learning. At the symbolic level, the generalizations that
are formed as a result of such connections might reflect what we
know as a grammar or syntax. Both levels are important to the
understanding of cognitive processes, and it is to the detriment of
694 TESOL QUARTERLY

our general knowledge that we ignore either of these levels when
attempting an explanation of what it is to learn or know a language.
ACKNOWLEDGMENTS
I would like to thank Michael Sharwood Smith for his careful reading, comments,
and criticisms of this paper.
THE AUTHOR
Maggi Sokolik teaches linguistics and applied linguistics at Texas A & M
University. She has also taught ESL courses at MIT, Harvard, and UCLA, where
she received her PhD in Applied Linguistics, focusing on issues of second language
acquisition.
REFERENCES
Anderson, J. R. (1983). The architecture of cognition. Cambridge, MA:
Harvard University Press.
Churchland, P. S., & Sejnowski, T. J. (1989). Neural representation and
neural computation. In L. Nadel, L. A. Cooper, R. M. Harnish, & P.
Culicover (Eds.), Neural connections, mental computation (pp. 15-48).
Cambridge, MA: Bradford.
Clark, A. (1989). Microcognition. Cambridge, MA: Bradford.
Elman, J. L., & McClelland, J. L. (1984). The interactive activation model
of speech perception. In N. Lass (Ed.), Language and speech
(pp. 337-374). New York: Academic Press.
Ervin, S. (1964). Imitation and structural change in children’s language. In
E. Lenneberg (Ed,), New directions in the study of language.
Fine, R. E., & Rubin, J. B. (1988). Specific trophic factor—receptor
interactions: Key selective elements in brain development and
“regeneration.” Journal of the American Geriatric Society, 36, 457-466.
Fodor, J. A. (1983). Modularity of mind: An essay on faculty psychology.
Gasser, M. (1990). Connectionism and universals of second language ac-
quisition. Studies in Second Language Acquisition, 12, 179-199.
Hinton, G. E., & Anderson, J. A. (Eds.). (1981). Parallel models of
associative memory. Hillsdale, NJ: Lawrence Erlbaum.
Hockett, C. (1968). The state of the art. The Hague: Mouton.
Inhelder, D., & Piaget, J. (1958). The growth of logical thinking from
childhood to adolescence (A. Parsons & S. Milgram, Trans.). London:
Routledge & Kegan Paul.
Jacobs, B. (1988). Neurobiological differentiation of primary and
secondary language acquisition. Studies in Second Language Acquisi-
tion, 10, 303-337.

Kohonen, T. (1984). Self-organization and associative memory. Berlin:
Springer-Verlag.
Lenneberg, E. H. (1967). Biological foundations of language. New York:
Wiley.
MacWhinney, B., Leinbach, J., Taraban, R., & McDonald, J. (1989). Lan-
guage learning: Cues or rules? Journal of Memory and Language, 28,
255-277.
McClelland, J. L. (1988). Connectionist models and psychological
evidence. Journal of Memory and Language, 27, 107-123.
McClelland, J. L., Rumelhart, D. E., & The PDP Research Group (Eds.).
(1986). Parallel distributed processing: Explorations in the micro-
structure of cognition: Volume 2. Psychological and Biological Models.
Cambridge, MA: Bradford.
Munro, P. W. (1986). State-dependent factors influencing neural plasticity:
a partial account of the critical period. In J. L. McClelland, D. E.
Rumelhart, & The PDP Research Group (Eds.), Parallel distributed
processing: Explorations in the microstructure of cognition: Vol. 2.
Psychological and biological models (pp. 471-502). Cambridge, MA:
Bradford.
Norman, D. (1986). Reflections on cognition and parallel distributed
processing. In J. McClelland, D. Rumelhart, & the PDP Research Group
(Eds.), Parallel distributed processing: Explorations in the microstruc-
ture of cognition: Vol. 2. Psychological and biological models,
(pp. 531-546). Cambridge, MA: MIT press.
Purves, D. (1986). The trophic theory of neural connections. Trends in
Neuroscience, 9, 486-489.
Rumelhart, D. E., & McClelland, J. L. (1987). Learning the past tenses of
English verbs. Implicit rules or parallel distributed processing. In B.
MacWhinney (Ed.), Mechanisms of language acquisition (pp. 249-308).
Hillsdale, NJ: Lawrence Earlbaum.
Sejnowski,T. J., & Rosenberg, C. R. (1987). Parallel networks that learn to
pronounce English text. Complex systems, 1, 145-168.
Slobin, D. L (1971). On the learning of morphological rules. In D. I. Slobin
(Ed.), The ontogenesis of grammar. New York: Academic Press.
Sokolik, M. E., & Smith, M. E. (1989, February). French gender
recognition: A network model and implications for second language ac-
quisition. Paper presented at the Ninth Second Language Research
Forum, Los Angeles, CA.
696 TESOL QUARTERLY
—
The Need for an Integrated Theory:

Connecting Modules
EVELYN HATCH, YASUHIRO SHIRAI, and CHERYL FANTUZZI
While second language acquisition research has been broad in

scope, the approach has been largely modular. Findings and the
theories used to account for findings have addressed only small,
isolated parts of the total language learning picture. We argue that
it is time to begin integrating modules in our research, to develop
and test integrated models for second language data, and to
develop integrated explanatory theories to explain these data. The
result should be theories that cover a much broader scope, that are
neurally plausible, and that allow us to integrate theories of
teaching with theories of learning.
As we think about our experiences as teachers and researchers

and examine our beliefs about the learning process, it becomes clear
that we have no all-encompassing theory of language acquisition
that matches what we have learned from experience. Rather, we
find a great deal of research on small parts of the total picture
without an integrated theory to guide our work. This is as true of a
theory of teaching as it is a theory of second language acquisition
(SLA). We agree with Spolsky’s (1989) statement that “the issue of
the area of coverage of theory is not fully appreciated” (p. 9).
There is no theory that integrates all areas of language acquisition.
There is no theory of SLA that includes acquisition of rhythm and
intonation (perhaps one of the first things learned in primary
language acquisition and the last in formal SLA). Nor is there a
theory that includes the acquisition of vocabulary (not just gender
or verb class categories but semantics in general). Where is the
theory that recognizes the fundamental process of everyday
metaphor? Or the theory that includes rhetorical organization of
expository text or the organization of conversational interaction?
Where is the theory that includes conflicting processing models such
as top-down versus bottom-up models and allows for multidirec-
tionality in processing? What theory makes explicit how L1 transfer
697
works in L2 processing for all areas of language? Where is the
theory that includes not only all of these but also intentional versus
incidental learning by students and considers the effects of formal
and informal instruction? Since foreign and second language
acquisition includes formal and informal instruction, the effect of
teaching—and the different types of programs that promote
particular types of teaching—should be made explicit in any theory
of SLA.
While we agree that the limited scope of theory is not well
recognized, the scope of research on SLA is amazingly broad. What
holds this research together is the fact that it all deals with the
acquisition of languages other than the native language. Of course
to be feasible, research can only cover some small area of SLA.
Thus, interlanguage research has looked for sequence or order in
the acquisition process: sequence in acquisition of question forms,
sequence in acquisition of certain English vowels, sequence in
acquisition of inflectional morphology, and so forth. Cognitivists
studying interlanguage have looked at sequence in the acquisition of
linguistic markers of spatial and temporal relations. Conversational
research has looked at pieces of the conversation system such as
openings and closings, length of gap between turns in the turn-
taking systems, noticings, brackets for asides, repairs, and the
functions of repair dysfluencies. Cross-cultural research h a s
compared speech act structures (and sometimes full event scripts)
of ESL and EFL students from different L1 backgrounds:
compliments, complaints, requests, refusals, and so on. Rhetoric
research has centered around narrative, procedural, and argumenta-
tive texts and the influence of L1 rhetorical organization on
comprehension and production of L2 text. Studies of how second
language learners establish reference, use deictic markers, and
promote coherence with cohesive ties have also been carried out.
There has been some interest in using rhetorical structure theory to
investigate the organization of the written text of second language
learners. Contextual analysis has also been used to contrast written
and oral language use of such learners. Lexical research has looked
at the learner’s flexible use of items via type-token ratios;
investigations have been carried out on the acquisition of core
semantics; and some few studies have begun to look at the
importance of conceptual metaphor in acquisition of second
languages.
Attempts to account for these parts of the total SLA picture have
centered on naturalness claims related to markedness theory,
parameter setting related to GB (government and binding) theory,
acculturation as related to learner and social group characteristics or
698 TESOL QUARTERLY

to accommodation theory, psychological factors—individual
factors such as motivation, anxiety, and cognitive style. While it has
been unfashionable to make explicit claims for the effect of
teaching, there is some research that has seriously considered the
role of explicit and implicit instruction on the learning process. (For
further explication of the scope of SLA research, see Hatch, 1989.)
The scope of SLA research might be represented in a flow chart
which appears on pages 700-701.
As you can see, if there is any “theory” that research in each of
these areas shares, it is that variation in sequence or in patterns of
use can be accounted for by some combination of L1 or topological
interference and transfer, universals or natural difficulty, and
individual learner or social group factors. In general, the research
has an “internal” focus on the learner, emphasizing the role of
innateness and universal orders to the neglect of the “external”
world where teaching and learning take place. Communication
studies, while not strongly influenced by innateness claims, still
focus on the learner rather than on the total interaction (i.e., the
evolving discourse produced by learners and other speakers or
readers). Surprisingly little importance is attached to teaching.
FEASIBILITY OF MODULAR AND INTEGRATED RESEARCH

Because each researcher must limit the scope of his or her
research, the questions asked and answers sought are almost always
about one separate subsystem of the total language picture.
Individual modules are investigated and findings are discussed as
though these modules were autonomous. However, we have always
been aware that modules “leak” or overlap. For instance, we know
the errors in third-person singular present tense may be either a
representation problem in morphology or a phonology problem
(the difficulty of consonant clusters) or both. A modular approach
to phonology would simply ignore the overlap. Pennington (1990)
shows that in Japanese-English interlanguage, L2 phonology and
discourse levels are interlocking systems and must therefore be
investigated as an integrated whole, rather than separately. In
syntax, we have found that temporality (verb tense and aspect)
cannot be successfully studied apart from rhetorical structure (since
tense and aspect in narratives, the most common data source for
investigations of verb tense, relate as much to discourse functions as
to syntactic functions). Definite and indefinite articles and pronouns
cannot be accurately studied apart from the discourse framework in
which they occur (since these forms depend on the role of the noun
THE NEED FOR AN INTEGRATED THEORY 699

FIGURE 1
The Scope of SLA Research
phrase within a particular discourse). (For more detailed examples
of module overlap, see the Appendix.)
To make our research feasible, we try to limit our investigations
to one area, but in doing so, we may advance explanations that are
faulty. The overlap among modules is admittedly great, but our
need to substantiate modules as autonomous is also great. Our fear
is that “confusion will reign” if we attempt to approach the whole
picture at once, rather than to separate the acquisition of language
into smaller, approachable facets.
We have apparently despaired of creating a theory that could
address all areas of language learning in an integrated fashion, a
theory that would allow for seemingly endless interaction among
components we have always described and perceived as separate.
Separate hypotheses about the acquisition of L2 phonology, syntax,
semantics, conversational structure, event scripts, and rhetorical
organization can be enlightening, but some way of bringing all of
these pieces together and allowing for large numbers of
interconnections is needed. We are beginning to see that it may be
possible to build models that overlap (are messy and fuzzy) in the
extreme and also become separate modules on demand.
Our ideas of what is feasible change with time. Philosophers of
science are fond of saying that the search for theory is influenced
not only by the work of great leaders but also by the technology
available for use in searching for answers. Such leaders may use a
variety of research methods to empirically test their assumptions. In
addition to testing their claims with empirical data, they may use
available technology to model learning processes. Technology may
advance theory because it is useful in solving problems (e.g.,
computers can search through countless texts with great accuracy),
because it offers a new and helpful metaphor (e.g., we talk about
mental processing in terms of computer processing), and because it
allows one to build models of learning and test them for replicabil-
ity.
Early computers with binary features gave us a way of describing
syntax as a formal series of sequential rules plus transformations of
the output of such rules. Early implementations of Chomsky’s
transformational generative theory used this approach. Such
attempts showed us that it was possible to model an autonomous
system of rules for portions of English syntax. Because of the limits
of computers at that time, the model was strictly sequential—that is,
if the system failed, the sentence was not produced and the
“machine” went back to the beginning to create a new sentence. It
couldn’t repair in midstream and, thus, didn’t model human
behavior in any strict sense. Further, the model only described
702 TESOL QUARTERLY

sentences and was limited in what it could tell us about the relations
among sentences aside from derivational history. The model was
limited by the technology to formal derivational, serial rules, and
“trees.” The model also reflected computer technology in the sense
that such rules could not be “taught” or “induced” by the computer
but had to be fed into it in some way (i.e., a language acquisition
device programmed into human biology). With improved technol-
ogy many of these problems in modeling language now appear to
be approaching solution, allowing us to reconsider our notions of
learning and teaching.
A SIMPLE METAPHOR
Rather than jump immediately to a more advanced computer
metaphor of windows, stacks, and so forth, let’s begin with a
metaphor from a simpler technology—a metaphor that might make
clear the type of messy models that are beginning to emerge. Think
of having stacks of overhead transparencies or slides available for
use in any order you wish. Since the light can shine through several
transparencies at once, you can overlap as many as you want in
whatever way you wish. Imagine you wanted to make an oral
complaint on returning some merchandise to a store. You can use a
“conversation system” transparency (openings and closings, turn-
taking rules, dysfluencies for face-giving, aside structures, and so
forth), a service-encounter script transparency with the complaint
section highlighted (actors, roles, props, acts and evaluation with a
sequence of statements of complaint, self-justification, other-
justification, negotiation of remedies, and so forth), an accompany-
ing lexical sheet with appropriate strength clines (vocabulary to
negotiate the remedy from threats, to suggestions, to hints),
intonation and stress sheets (again, with appropriate clines for
intensity), and syntactic sheets for any nonformulaic utterances not
already incorporated by ISA, HAS, DO or TRANSfer links in the
script transparencies. The light of the overhead projector (OHP)
shines through them all and some mechanism (sometimes called a
“Central Processing Unit” [CPU]) can move them about in relation
to each other; one may be highlighted more than another at one
instant and another a moment later. That is, while each may indeed
turn out to be an individual modular sheet, they are all intercon-
nected in multiple ways and at ever-changing strengths. What kinds
of models can be devised that will make sense in such an ever-
shifting system?
The picture becomes even more complicated when we realize
that the OHP metaphor can’t work because language learners do

not just talk to themselves. How will the CPU of one OHP work
when constant demands are being made for it to attend to the
ongoing interaction. Let’s switch metaphors for a moment and think
of the interaction as a basketball game where our OHP is expected
to understand the rules of interaction: accepting the ball, passing it
to others, recognizing the unfolding game plan, assisting others or
receiving assistance from others to bring about a successful play,
holding onto the ball, and so forth. In a second language, we may
not have the right uniforms, may not know the rules of the game, or
the patterns may be somewhat different in our league, and the close
interpersonal interaction of the team may not soon be offered to us.
With supercomputers, we can envision parallel processing of
components much as in our single OHP metaphor. With networked
computers, it is possible to see how the demands of more than one-
party communication might be modeled. With parallel processing,
a nonsequential model—where connections are activated and work
simultaneously, with the possibility of drawing from each other on
demand—becomes a possible language processing model. The
“parts” need not be seen as fixed separate modules or transparencies
but as large numbers of connections that happen to get turned on or
activated together so that they form patterns connected by links.
MODELING LEARNING
Much of this discussion on the possibilities for computer model-
ing comes out of AI (artificial intelligence) research on natural
languages. The modeling has not all been at the level of phonology
and syntax. Successful models have been built for various
communication tasks (airline information, medical information,
catalogue ordering, bank information calls, and, in our applied lin-
guistics AI class, cooking scripts). Successful AI models of L2 lexi-
cal errors within scripts have been constructed (see Gasser, 1988;
Zernik, 1986, 1987). In cognitive science, where there is a concern
that models match to some extent what neuroscientist claim about
ways in which the brain works, they speak of models that actiwzte
and link and inherit information in scripts: As actors are activated,
roles may be linked to actors, and props may be inherited by the
roles (these become the givens or background information for the
interaction). Stereotyped utterances and syntactic links for
background information may also be activated (John IS A painter,
painter HAS A brush, painter DO: provide a service—activate
service encounter superscript; paint walls, cleanup, evaluate job,
receive payment for job, and so forth). Some AI models also include
syntactic parsers within them. For example, a model designed to
704 TESOL QUARTERLY

comprehend news reports would have a rhetorical processor, but it
might also have a syntactic parser within it. In some cases an
auditory/oral component is added so that processing for compre-
hension and production can be modeled interactively; it may be-
come possible to add stress and intonation information to such
models.
The problem for AI models has been that the information in
models (as was the case in early transformational generative (TG)
models of language acquisition) must be given to the computer—
written in as a part of the program that is then tested on new
material. That is, these are not models of acquisition but are flexible
templates that allow the computer to process and produce language
in relation to the templates. The models are improvements over
simple early TG rules, though, for they do not simply produce
isolated sentences. They cover a much broader scope of language.
Still, there has not been a way to teach the computer to form the
model on its own (i.e., to become an intelligent computer regarding
language). The search for, and perfecting of, a model that learns the
material rather than simply (actually, not simply at all!) providing
an “innate” program is the goal of much connectionist research.
That goal is important to all of us who believe that language
learning must include an important role for teaching as influencing
internal processing.
Connectionists have shown it is possible to simulate learning of
small parts of language. In connectionist experiments, for example,
the computer is able, given input as an instructional program, to
sense patterns in the data it is given. (We have not included a
description of how connectionist experiments work here. For the
details of connectionist learning, see Gasser, 1988, 1990; Sokolik &
Smith, 1988; and Fantuzzi, 1990, as well as Sokolik’s paper in this
issue of the TESOL Quarterly.) The outcome of these experiments
shows that the computer is able to acquire small parts of the
language (e.g., parts of phonology, morphology, noun gender) with
a fair degree of accuracy and, while learning, produce errors some
of which are similar to those of human learners.
Following Rumelhart and McClelland (1986), Plunkett and
Marchman (1989) ran a connectionist experiment to see what kind
of input would get the best learning of English past tense forms and
also produce learner-like errors in generalization. This is similar to a
teacher asking questions such as whether it would be best to present
all regular verbs first or start with irregular forms, how many items
should be presented at a time, and so forth. If a teacher could alter
the input in various ways, which input would produce the best
outcome? By controlling input, we might get the machine to

accurately produce irregular past forms such as eat-ate, sing-sang,
write-wrote, and so forth. If we then flooded the model with lots of
regular -ed past tense forms, we would expect the output pattern to
shift to eat-eated or perhaps eat-ated. Once we shift input to include
irregular verbs again, would the computer produce accurate eat-ate
output once again? That is, would it give us a U-shaped curve of
ate-eated-ate similar to the U-shaped curve posed for children
acquiring English as a first language? If all types of past tense forms
were presented together, what would happen? If the computer
produced errors such as “eated,” then the connectionist experiment
might be a viable learning model. And this, of course, is what these
two studies strive to do.
A second type of human-like behavior is shown in Gasser’s (1990)
connectionist learning experiment, which investigated the effect of
L1 on L2 learning. Gasser “taught” simple forms of an L1 in a
connectionist experiment. He then introduced forms in an L2 that
was either quite similar or quite different from the L1. Some of the
findings match those of human learning. That is, the L2 was never
as well learned as the Ll, but learning was somewhat easier initially
than had been the case with L2. The L2 patterns were easier to learn
and there was less interference if the words and the word order of
the two languages were similar. As more and more L2 patterns were
learned, this initial difference in favor of similar words and patterns
disappeared. Most interesting, the effect of different word order
was greater when the words in the two languages were similar. This
finding led Gasser to suggest that such experiments could be used to
further test Kellerman’s (1979) notion of perceived language
distance influencing transfer.
At first glance, such connectionist experiments are pleasing par-
ticularly since they allow us to consider once again the importance
of teaching—formal and informal—on early L2 learning. The
models assign teaching an important role even though the particular
type of input used to teach the computer would not appeal to most
language teachers. The research also allows us to take seriously
questions about scope and sequence in teaching and its effect on
inferencing and generalization in learning. The experiments will
also allow us to test the importance of a sequenced instructional
program and of L1 transfer in a variety of ways. (For further
discussion of connectionism and language transfer and interference,
see Shirai, in press.)
But, again, AI and connectionist models do not yet give us a
model of integrated language on which to build theory. That is, it is
possible to program computers with flexible templates of scripts,
conversation components, rhetorical forms, a syntactic parser, and
706 TESOL QUARTERLY

so forth and to test these templates on new data to see if they work
(i.e., whether the computer, given new data, can answer questions
based on this knowledge). It is possible to get the computer to
“learn” patterns with teaching from input examples. So far, these
learning experiments are very limited in scope; of necessity, they
deal with only some small part of a module (and, as we have
argued, findings from modular research may need to be
reconsidered in terms of an approach that integrates modules).
There needs to be much more work to bring together the two
extremes: the broad areas such as AI models of scripts and the
narrow areas of research in connectionist experiments (such as past
tense, gender, and so forth).
PROCESS RESEARCH
The language acquisition research that connectionists believe
illustrates their view of learning is that concerned with processes.
These studies focus on some small part of language where findings
are used to illustrate such processes as matching or generalization or
reorganization or where writers speak of activation and patterns or
connections (rather than of rules, rule-governed behavior, o r
parameters and markedness). Other favorites are those which
combine explanations that cross linguistic modules. For example,
Karmiloff-Smith’s (1985) process model predicts generalization and
U-shaped curves that are interesting not because the notion is new
but because the particular curves she uses to illustrate her process
model combine discourse and syntax; that is, process models can
cross “transparencies.” In language development, the modules need
not be treated as autonomous subsystems.
In Karmiloff-Smiths process model, phases similar to those
producing the U-shaped curve discussed earlier for irregular past
tense verbs are illustrated. Initially, the learner’s language may form
a fairly close match to that shown in the input. Young children
produce correct irregular verbs such as sang, went, ate as they
match the input they hear. This is similar to connectionist
experiments where computers are easily able to accurately match
patterns in the input. In the first phase, the learner attempts to
match the input. The learner is sensitive to positive and negative
feedback and adjusts to match the desired output. Once success is
obtained so that the learner receives only positive feedback, a
second phase of reorganization takes place to obtain better control
of the internal representation, a representation that fits better with
other parts of the system being learned. This second phase is
internally driven rather than input- or data-driven. Even though the

input doesn’t change, the learner’s output may change as the learner
reorganizes Phase 1 into something that fits better in the total
system. The child learner, for example, produces singed, goed,
eated. In the third phase, there is an integration of input—the
externally driven impetus to learning provided by various types of
teaching—and internally driven processes. Thus, the child learner
once again produces the correct irregular past tense forms sang,
went, and so forth.
Karmiloff-Smith has also given other less well-known examples of
acquisition in terms of phases. The phases are not to be confused
with stages; phases refer to the types of mental processes that
learners use as they experience the interconnectedness of language.
For example, initially learners may simply match the input in
producing pronouns. Their stories at this stage will show evolving
accuracy for pronouns in terms of grammatical number and person.
1. The little girl is walking in a path. The little girl buys a ice cream. She
eats it. She falls down and she drops it. She’s crying. (Hatch &
Hawkins, 1990)
While such matching may be successful as a Phase 1 process, it
does not necessarily fit well with what the learner must simultane-
ously learn about subject selection in discourse. Learners may then
use pronouns in a way that is inaccurate for the discourse. For exam-
ple, they may use pronouns for whatever person they have selected
to fill the subject slot in their narratives:
2. The little girl is walking. She wanted an ice cream and she
((unstressed)) sold her an ice cream and she fell and she dropped a
ice cream and she cried. (Hatch & Hawkins, 1990)
These examples (data from a child telling a story from a set of
pictures) show an incomplete blend where the storyteller puts the
actor in subject slot and then uses pronoun reference. A more
harmonious fit can only be obtained by placing the hero of the story
in subject slot and then noting that if someone else appears in that
slot, this information must be highlighted by giving full NP
identification of the person. That is, if a secondary protagonist
appears, a full NP or a stressed pronoun must be used.
3. The little girl is walking. She sees a ice cream wagon and she goes to
buy a ice cream. The lady gives her her ice cream. Oh, she doesn’t
watch out. She felled on a . . . on a? rock? and drop her ice cream.
And then she cries. (Hatch & Hawkins, 1990)
Process models such as that briefly illustrated above are by no
means new, but they are important because they show us how
708 TESOL QUARTERLY

modules overlap and give us clues to where instruction might be
most helpful to the learner. The learner’s output also shows us the
type of knowledge teachers need to stimulate further internal
reorganization. These models can highlight the importance of
teaching in an integrated theory of language learning.
These models are not new, but they may be more precise
explications to show that
[the learner] is neither a pure associative machine nor a sovereign
constructor of concepts. Rather his speech is based on the continuing
interaction of external impressions with internal systems which usually
function unconsciously, it is thus the result of a constant “convergence.”
The detailed investigations pertaining to the development of speech and
thought should determine the relative participation of both forces and
also show how they accommodate each other. (Stern & Stern, 1907)
Process models are a step towards building theories of how inter-
nal and external processes might influence and/or accommodate
each other. They have not specified how teaching might influence
either internally or externally driven phases. That task is the job of
both SLA researchers and learning and instruction specialists.
We would like to end this section on process models on this rather
positive note. However, we must admit that when we have at-
tempted to verify and validate U-shaped learning curves similar to
those discussed above, we have run into difficulty. For example, in
a pilot study to test Karmiloff-Smith's phase model, Hatch and
Hawkins (1990) analyzed sets of 64 stories from children in the
process of becoming bilingual. (The Fabian example in the
Appendix is drawn from this data base.) It was impossible to
categorize the narratives, as Karmiloff-Smith had done, into the
three phases she outlined, i.e., how children use the following not
only as linguistic forms but also according to their discourse
functions in stories: (a) definite and indefinite articles, (b) stressed
and unstressed pronouns, and (c) full noun phrases. The analysis of
the narratives revealed that an individual narrative might have
evidence of Phase 3 along with instances of Phase 1 and Phase 2
structures within it. Why should this be so? There are many possible
explanations other than those involving bilingualism, but we’d like
to point out one as an illustration of how interconnected the system
may be.
Using a broader notion of connectionism, is it possible that in
natural language use (our Ss were telling their own stories rather
than retelling a picture story), too many connections are activated.
To put it another way, many modules and interconnections within
and between modules are activated and used. At one moment, the

child focuses on the hero of the story in one way and, a moment
later, in another. The “transparencies” shift with incredible speed as
the child alters goals, changes heroes, shifts temporal and spatial
frames, moves from narration to description and back again. Such
data makes it impossible for us to see the small, neat processes that
may be operating in one small part of the total messy picture. We
can look at this as either a positive or negative outcome. If the scope
of the model (and the theory to explain the model) is narrow, it
makes it possible for us to see processes that would otherwise be
buried in the data (narrowed scope allows us to see trees in the
forest). On the other hand, the narrow scope may result in
misleading “insights’’—insights that work only in the experimental
setting and perhaps not even consistently there (not only can’t we
see the forest because of the trees but we may be describing
something other than forests or trees).
Natural data that integrates all areas of language use calls into
question the feasibility of research and accuracy of research-based
explanations. This situation is not too different from the problem all
teachers face. If we try, for example, to teach articles in terms of
their forms, we can limit the context and then look at how well our
learners use articles in this context. We feel pleased that learners
acquire the forms. However, when we compare the article use of
native speakers to that of our students, we realize that our heuristic
presentation of the article system no longer makes sense. Many
interacting components (including discourse function) are involved
as we select articles. To make our teaching task feasible, we have
narrowed the focus and trust that, however students first learn
articles, they can successfully reorganize them later as we offer
them more and more connections. This is often very difficult for
both teachers and researchers, for we are not sure that we really see
clear evidence of phase-like processes in the data our students
produce. We are afraid that students are not really learning.
Similarly, in our SLA research, we are afraid that we may not really
have data that would support a phase model.
We have spent so much time on this problem of feasibility
because SLA (and most language) research has had as its focus an
attempt to discover the “underlying representation” for some small
piece of language. To do this, we have had to ensure that this is also
the learner’s focus (i.e., the light of the OHP is shining on this
particular part of the total communication picture). We do this by
working out an artificial procedure that will force the learner to
produce the forms of interest in the context we select. The findings
then, as in all controlled research, may not generalize to other
situations. Chomsky and others solved this problem for linguists by
710 TESOL QUARTERLY

saying that the “underlying representation” is linguistic knowledge
that is imperfectly produced in performance. Such a view is more
compatible with the early technology than with the fully
distributed, parallel processing models we now know are possible.
If we believe that language is made up of connections that are
multiply distributed, exceedingly overlapping in nature, then it is
this messy, fuzzy, overlapping picture that is real rather than the
descriptions that we use to bring better order to the picture.
THE EXPERIENTIAL FRAMEWORK REVISITED

In proposing an “experiential framework” for second language
acquisition (Hatch, Flashner, & Hunt, 1986; Hatch & Hawkins,
1987), we tried to demonstrate that a theory of SLA must include a
much broader scope of research than that centered on two
modules—syntax and phonology. We showed that conversational
structure, scripts, and lexicon are not only important but also that
they interact with other language components as well. Any theory
should cover at least all the areas shown in Figure 1 and the
interconnections among these areas. In the framework, we also
emphasized the importance of learning experiences (whether in
formal or informal learning situations). We tried to show how
internal mechanisms might structure or restructure the data
obtained from experience. To do this, we used some of the early
parallel processing models such as that proposed by Kempen and
Hoenkamp (1981). Since then, connectionism and other AI models
have presented us with even better alternatives to use in visualizing
this process—a view of what a theory might look like that integrated
all areas of language learning.
Given this approach, it is tempting to say that partial theories are
misleading (even though they are supported rather than falsified by
models that test them). If the partial theories fail or are only
apparent in controlled experiments or in controlled classroom
environments, are they useless? How can they be right when they
fail once the scope is enlarged? Of course, this negative view is not
warranted. We simply need to be cautious about our claims keeping
in mind the metaphor that the light, the CPU, may only be directed
at one tiny place in the total system and that once the transparencies
are shifted around, we can again investigate what happens as the
focus shifts. Connectionism has given life to such a metaphor
because it has emphasized the many, many connections and
interconnections to be made in learning. It’s disheartening to realize
that we can’t look at all these connections and interconnections at
once any more than we can teach all of language at once. All we can
THE NEED FOR AN lNTEGRATED THEORY 711

do is tease apart some small portion of connections and examine
them in detail without forgetting that whatever system is discovered
to operate for this small part of language may, in fact, be quite
different once we integrate it into the total language picture.
Where does all this leave us? In his 1990 paper, Gasser says that
“Once we are willing to accept the possibility of an adult system in
which redundancy is rampant, concepts are fluid, metaphor is a
fundamental process, and exceptions are the rule, our picture of the
learner and our research strategy change dramatically” (p. 196).
Gasser says that a great deal of rethinking will be needed but that
the outcome may be a more elegant model of acquisition, one that
“allows it to be integrated into the rest of mind, and perhaps even
the brain” (p. 197). We would like to see it integrated not only in this
way but also in a way that allows us to collapse models of teaching
and learning. We have already seen that AI researchers can model
much of discourse structure (rhetorical structure, script structure,
conversational structure, and so forth). The computer can also be
used to test our ideas about teaching. The amazing computer-
assisted programs (such as those produced at Brigham Young
University for Spanish and other languages) show how useful and
innovative computers can be in teaching adults foreign languages.
Such programs and others like it could be used to see if it is possible
(as Horwitz, 1989, suggests) to lessen anxiety and fear of failure, and
to change learners’ beliefs about the learnability of second and
foreign languages. In addition, they give us the opportunity to
model various types of teaching and to test these with real learners.
Given our present technology, all kinds of experiments could be
carried out to test our hypotheses about both teaching and learning,
about how learning is generalized, and about how findings change
once the effect of another “module” of language or communication
is changed and new parts added. If we can combine a neurally
plausible theory of learning with a pedagogically plausible theory
of teaching, producing an integrated theory that is broad in scope,
we could not only get back “on track” (as the theme of the 1990
TESOL conference suggests) but also advance our understanding
of second language acquisition in exciting new ways.
712 TESOL QUARTERLY

THE AUTHORS
Evelyn Hatch is Professor Emeritus in the Department of TESL and Applied
Linguistics at the University of California, Los Angeles. Her most recent book is
The Research Manual: Design and Statistics for Applied Linguistics (1991,
Cambridge University Press) coauthored with Anne Lazaraton. She is now
working on her next book, The Discourse Manual, to be published by Newbury
House.
Yasuhiro Shirai is a doctoral student in the Department of TESL and Applied

Linguistics at the University of California, Los Angeles. His research interests
include L1/L2 lexical and semantic development, and cognitive models of
language processing/acquisition. He is currently working on his dissertation on the
L1/L2 acquisition of tense/aspect.
Cheryl Fantuzzi is a doctoral student in the Department of TESL and Applied

Linguistics at the University of California, Los Angeles, specializing in second
language acquisition. Her research interests include cognitive science and
developmental psycholinguistics.
REFERENCES
Clark, A. (1989). Microcognition. Cambridge, MA: MIT Press.
Fantuzzi, D. (1990). Connectionism and second language acquisition.
[Qualifying paper]. University of California, Los Angeles, Department
of TESL and Applied Linguistics.
Gasser, M. (1988). A connectionist model of sentence generation in a first
and second language (Tech. Rep. No. UCLA-AI-88-13). Los Angeles:
University of California, Computer Science Department.
Gasser, M. (1990). Connectionism and universals of second language
acquisition. Studies in Second Language Acquisition, 12, 179-199.
Hatch, E. (1989, August). The scope of SLA research. Paper presented at
the Educational Testing Service Invitational Symposium on Language
Acquisition and Language Assessment, Princeton, NJ.
Hatch, E., Flashner, V., & Hunt, L. (1986). The experience model and
language teaching. In R. Day (Ed.), Talking to learn (pp. 5-22). Rowley,
MA: Newbury House.
Hatch, E., & Hawkins, B. (1987). Second language acquisition: An
experiential approach. In S. Rosenthal (Ed.), Advances in applied
psycholinguistics: Vol. 2. Reading, writing and language learning
(pp. 241-283). Cambridge: Cambridge University Press.
Hatch, E., & Hawkins, B. (1990). [Oral and written narratives: At-risk LEP
children]. Unpublished raw data.
Horwitz, E. K. (1989). Recent research on second language learners:
Beliefs and anxiety. In D. Koike & A. Simoes (Eds.), Negotiating for
meaning: Papers on foreign language teaching and evaluation. Austin,
TX: University of Texas at Austin, Department of Foreign Language
Education Studies.

Karmiloff-Smith, A. (1985). Language and cognitive processes from a
developmental perspective. Language and cognitive processes, 1 (l),
60-85.
Kellerman, E. (1978). Giving learners a break: Native speaker institutions
as a source of predictions about transferability. Working papers in
bilingualism, 15, 59-92.
Kellerman, E. (1979). Transfer and non-transfer: Where we are now.
Studies in second language acquisition, 1, 37-57.
Kempen, G., & Hoenkamp, E. (1981). A procedural grammar for sentence
production (Internal Report No. 81). Nijmegen, The Netherlands:
Katholieke Universiteit, Vakgroep psychologische functieleer,
Psychologisch laboratorium.
Pennington, M. (1990). The context of L2 phonology. In H. Burmeister &
P. L. Rounds (Eds.), Variability in second language acquisition:
Proceedings of the tenth meeting of the Second Language Research
Forum (Vol. 2, pp. 541-564). Eugene, OR: University of Oregon,
Department of Linguistics.
Plunkett, K., & Marchman, V. (1989). Pattern association in a back
propagation network: implications for child language acquisition (Tech.
Rep. No. 8902). La Jolla, CA: University of California, San Diego Center
for Research in Language.
Rumelhart, D., & McClelland, J. (1986). On learning the past tenses of
English verbs. In J. McClelland, D. Rumelhart, & the PDP Research
Group (Eds.), Parallel distributed processing: Explorations in the
microstructure of cognition: Vol. 1, Foundations. Cambridge, MA: MIT
Press.
Shirai, Y. (in press). Conditions on transfer: A connection approach. Issues
in Applied Linguistics.
Sokolik, M., & Smith, M. (1988). French gender recognition: A network
model and implications for SLA. Paper presented at the Eighth Second
Language Research Forum, Los Angeles.
Spolsky, B. (1989, August). Language assessment in the light of a general
theory of second language learning. Paper presented at the Educational
Testing Service Invitational Symposium on Language Acquisition and
Language Assessment, Princeton, NJ.
Stern, C., & Stern, W. (1907). Die Kindersprache. Leipzig: Barth.
Zernik, U. (1987). Failure driven acquisition of figurative phrases by
second language speakers. Unpublished manuscript. University of
California, Los Angeles, Artificial Intelligence Department.
Zernik, U., & Dyer, M. (1986). Disambiguation and language acquisition
through the phrasal lexicon (Tech. Rep. No. 86-2). Los Angeles:
University of California, Computer Science Department.
714 TESOL QUARTERLY

APPENDIX
Illustrations of Module Overlap
PHONOLOGY AND MORPHOSYNTAX OVERLAP
Depending on the researcher’s focus, the following “errors” (of third-grade children) could
be cited as morphosyntax data, pronunciation data, or orthographic data. Clearly, these
language modules should not be investigated in isolation from each other.
1. Lins is like cat. There father and the mother protections there line cub. The father is very
bik. I like the lins. (Hatch & Hawkins, 1990)
2. The bats are in the buk The babys they clings to it’s mother fur as she sails through the
air I like the bat They are creepy Furit is a very nurasheen food it has vitamen and many
good thing. Almost evry one in the world has taysted furit. I have a little dog. We putted
her in a warm box. When morning came she jumpet in the bed. We were serpris and
alarm. Now evry morning she come and wakes me up. (Hatch & Hawkins, 1990)
SYNTAX AND GENRE OVERLAP
In most narratives, the hero (primary protagonist) is placed in subject slot and carries out
actions directed at other participants or objects. Once the primary protagonist has been
established, a pronoun is used to refer to the protagonist. If a secondary participant is placed
in subject slot, that participant is given high focus via full Noun Phrase identification or with
stressed pronouns. In the following excerpt, note that the child repairs the first utterance so
that he, rather than Fabian, is the primary protagonist. To appear in subject slot, the
secondary protagonist is mentioned by name (“and Fabian got mad at me”) rather than
deictically with an unstressed pronoun. From this point on, the storyteller seems to have a
conflict of goals: Should he present himself as the Macho Kid, keeping himself in subject slot
and emphasizing himself as the strong and admirable aggressor, or should he become the
recipient, the Picked On Kid, to garner sympathy from his audience? The selection of
pronouns, stressed pronouns, and full names shows this dilemma. Pronouns should not be
described without reference to narrative structure and the goal of the storyteller in the speech
event of storytelling.
3. This story’s about when Fabian got-. . . I-when I got mad at Fabian and Fabian got
mad at me. H-he got mad at me because me and Joe were just passing the ball to each
other and we didn’t pass it to him. . . in basketball. And . . . just me and Joe were
working on our shots, an’ . after the bell rang . He got mad at me an’ he ws gonna
hit me an’ . . . I just hit him back an’ Mr. L came . . . (Hatch & Hawkins, 1990)
SYNTAX AND GENRE OVERLAP
We expect that simple tense (past or present) will be used for the main story line. Continuous
aspect is frequently used to show actions taking place in parallel with the main story line.
Perfective aspect is often used to show previous actions which form tbe background for the
ongoing actions in the main story line. In this excerpt from the same narrative, notice that the
dilemma continues as the speaker tries to make Fabian the major actor and aggressor. He uses
weakened statives (“just sat”) and weakened serial acts (“starting hitting,” “kept on hitting”)
for his own actions; simple active verbs are used for Fabian, the aggressor (“he threw me”);
and they are both to-be-pitied recipients of passive constructions (“got suspended”) in the
end. Tense cannot be described without reference to narrative structure and the goal of the
storyteller.
4. Teacher: He was gonna hit you and you hit him before he hit you?
MA: Yeah . . . the-then he-he uses-he just-got me and pushed me. Then Mr. L came an’
. . .-before Mr. L came . . . I pushed him back, and Mr. L just told us to
stop it an’ Then in the classroom when we were walking back uhmmm,
Fabian-Fabian just . . . I wa-just sat in my seat an’ Fabian Fabian got me an’
. . . he threw me against the chalkboard an’ then I just got backup an’ next. .
I started-I started hitting him, an’ . . .Mrs. D. told us to stop it, but Fa-Fabian

didn’t let me go, so I kept on hitting him, too. An’ . . uhmmmm . . . we got
suspended for f-f-for a week, I think. St-We had to stay up in the office. (Hatch&
Hawkins, 1990)
GENRE AND SCRIPT OR EVENT OVERLAP
In terms of rhetorical structure, the following text appears to be a short description followed
by a narrative (telling what “I” did). The note, however, was attached to a gift of seeds. The
gift-giving script or speech event leads the reader to reinterpret the rhetorical structure of the
text as procedural (advice-giving) discourse. Nevertheless, it does not contain the usual
markers of procedural text such as neutral actor, neutral time, high use of passives and
medals. Nor does it use an opening framework such as “To grow sweetpeas,” or a “good
luck” closing. Rhetorical structure cannot be described without reference to the script
module.
Evelyn,
These are second generation giants. Their parents got to be about 8 feet tall. I dug a ditch
about 3“ deep and 5’ long and buried them about 1/2” deep in the center. Then, as they
grew I shoveled the dirt back into the ditch. So the roots would not be exposed. I may have
used a small amount of rose food fertilizer too.
Best wishes,
Grant
PHONOLOGY, SYNTAX AND GENRE OVERLAP
In English, new information is generally placed toward the end of sentences. The accent
pattern of English is also phrase final. Old information often appears in sentence-initial
position and is unstressed (see Example 3 above for an illustration of the placement of
unstressed pronouns in subject slot as “old information” once the major protagonist of a
narrative has been identified). In addition, content words receive stress within phrases and
these typically occur following any function words or modifiers. In Japanese, on the other
hand, this accent prominence is given to phrase-initial positions. If a Japanese student of
English, telling a story similar to that of MA above, identified a primary protagonist with a
high pitched, stressed pronoun, would this “error” be one of discourse, syntax, or phonology?
Again, it is impossible, as Pennington (1990) points out, to study these independently. Any
explanation of the phenomena must consider all three language modules.
716 TESOL QUARTERLY

REVIEWS
The TESOL Quarterly welcomes evaluative reviews of publications relevant to
TESOL professionals. In addition to textbooks and reference materials, these
include computer and video software, testing instruments, and other forms of
nonprint materials.
Edited by HEIDI RIGGENBACH

University of Washington
Recent Publications on Classroom Research

Observation in the Language Classroom.
Dick Allwright. London and New York: Longman, 1988.
Pp. xvi + 288.
Second Language Classrooms: Research on Teaching and Learning.
Craig Chaudron. Cambridge Cambridge University Press, 1988,
Pp. xvii + 221.
The Classroom and the Language Learner: Ethnography and Second-
Language Classroom Research. Leo van Lier. London and New York:
Longman, 1988. Pp. xviii + 262.
These three excellent books on classroom research, written

primarily for researchers, are very different: Chaudron’s is the most
comprehensive review. Allwright’s is a useful introductory text
containing long excerpts from original reports, van Lier’s is entirely
about what I would call discourse analysis.
Chaudron’s book, Second Language Classrooms: Research on
Teaching and Learning, is a critical review of the methodology and
findings of research on talk in second/foreign language classrooms
“in order to determine the degree to which specific classroom
processes or behaviors are sources of positive effects on second lan-
guage learning” (p. 2).
Chapter 2 presents issues of methodology, with an overview of
four research traditions—psychometric, interaction analysis,
discourse analysis, and ethnography—that have gained currency in
roughly this chronological order. Chaudron’s overview documents
how each subsequent development has called our attention to new
aspects of classroom talk. The psychometric studies of the 1950s
and 1960s evaluated instructional programs—e.g., comparing
proficiency test scores in grammar/translation with audiolingual
717
classrooms. Such studies ignored variation within a program
category and offered no information about what features of a
“successful’ program contributed to its effect.
Research attention shifted in the 1970s from nominal program to
actual classroom processes, with more detailed analyses of teacher
and student talk. The first such attempts, called interaction analysis,
were designed for on-the-spot coding; later analytical schemes,
termed discourse analysis, work from transcripts and incorporate
more complex sociolinguistic constructs in the coding categories.
Finally, ethnographic studies call attention to significant contextual
variables that help to explain why classroom processes are as they
are in any particular situation. They also call attention to partici-
pants’ interpretation of the function of talk as a component of valid
coding, but they rarely provide quantitative information on student
learning. It is a credit to Chaudron’s breadth as a researcher that his
discussions of alternative methods and issues in research decisions
are relatively unbiased by his personal research training and
experience.
Chapters 3 and 4 discuss analyses of teacher talk and student talk,
respectively. Teacher talk in second language classrooms has at
least three broad functions: First, it constitutes input (of special
importance in foreign language situations); second, it sometimes
shifts from focus on meaning to focus on form—that is, from com-
munication to metalanguage, from use to mention; and third, it
structures the context for learner language use. In this chapter, 32 of
the 39 pages are devoted to the first function, and Chaudron
summarizes the ways in which teacher talk is adapted to the needs
of L2 learners in slower rate, more exaggerated pronunciation,
simplified vocabulary and syntax, and more self-repetitions. Only
two pages are devoted to research on the second function, which
Chaudron refers to in the index as “metatalk’ (p. 218). The third
function, structuring student participation, is discussed in chapter 5.
Chapter 4, “Learner Behavior,” includes a short discussion of
covert learner strategies and learner talk. Chaudron organizes the
analysis of research on learner talk around three plausible
hypotheses: that learners develop their L2 by “producing the target
language more frequently, more correctly, and in a wider variety of
circumstances . . . by generating input from others . . . [and] by
engaging in communicative tasks that require negotiation of
meaning” (p. 90). In general, research serves to qualify each of these
hypotheses. Especially interesting is Chaudron’s discussion of the
small number of studies (all but one published since 1985) of
interaction among NNS learners in small group tasks. The
hypothesis that privileges correct language input might question
718 TESOL QUARTERLY

small group talk out of earshot of the correcting teacher, while
the hypothesis that privileges the importance of negotiation of mean-
ing might consider such small group work the most beneficial of
all.
Chapter 5, on teacher-student interaction, includes discussion of
teacher questions and teacher feedback. These are the functions of
the first and third slots in the three-part structure of classroom
interaction (teacher initiation, student response, and teacher
evaluation) that is so common in both L1 and L2 classrooms. Both
questions and feedback have special significance in L2 classrooms:
Teacher questions (whether display questions to which teachers
know the answer or referential questions to which they do not)
serve as invitations to student language use, and feedback serves as
comment (by peers as well as teacher) on meaning or form.
Chapter 6 revisits these variables as well as the issue of explicit
attention to language form (“teachers as instructors rather than
simply as conversant” [p. 164]) in search of evidence of their effect
on student language learning. Unfortunately, research results do not
yet provide clear recommendations to teachers, and Chaudron ends
the book by revisiting some of the methodological issues raised at
the beginning of the book.
Allwright’s book, Observations in the Language Classroom,

complements Chaudron’s in three ways. First, it follows roughly the
same sequence of research categories: from program/product to
process/product and from interaction analysis to observations that
start with hypotheses derived from second language acquisition
research. However, the book makes more explicit the chronological
nature of this journey during the past 25 years. Second, whereas
Chaudron refers to many studies (and provides excellent compari-
sons of them in charts in every chapter, e.g., of the 26 studies in the
chapter on teacher talk alone), Allwright combines the functions of
author and editor by providing extended excerpts from 16 studies
that exemplify both trends and some of the best work along the way
(although none more recent than 1980), Third, Allwright includes
more discussion of the relationship of observational research to
teacher training—not by discovering generalizable knowledge that
can be transmitted to teachers, but by suggesting ways in which
teachers can monitor processes in their own classrooms. At the end,
Allwright seems less convinced than Chaudron of the value of a
continued search for generalizable statements of causal relation-
ships between classroom process variables and L2 learning.
Allwright seems, in short, to have less “faith in the observable”
(p. 248).
REVIEWS 719
Significant nonobservables in the classroom are what is happen-
ing in the minds of students and teachers. In the words of two L1
process-product researchers (Gage & Needels, 1989) in responding
to criticisms of that work, future research should be based on “more
complete conceptualizations . . . [which] should include (a) the
student cognitive processes mediating between process and
product variables and (b) the teachers’ thought processes associated
with process variables” (p. 294).
The map Chaudron presents (p. 3) to show how the process vari-
ables that will be his focus relate to other sets of variables (presage,
context, and product) is taken from a 1974 source, before cognitive
mediating variables had gained the significance they have now. He
devotes seven pages in the chapter on learner behavior to learner
strategies, including one chart summarizing seven studies and another
one giving definitions of 26 strategies in three categories: meta-
cognitive (such as self-monitoring), cognitive (such as repetition) and
socioaffective (such as question for clarification). But he does not
return to the importance of such mediating variables in his final
chapter, and he nowhere suggests that teacher thinking (in contrast to
presage variables such as whether the teacher is a native or nonnative
speaker) may be important to consider.
Allwright’s final chapter, on the other hand, suggests the value of
combining observation with more “mentalistic” research proce-
dures that take advantage of the fact that learners in L2 classrooms
are older, often much older, than L1 language learners and can
report on at least some of their language learning strategies in
diaries and more immediate, elicited self-reports. Allwright
mentions research on teacher cognition as “a new area for enquiry”
(p. 209), and cites two studies, one by Chaudron, of teachers’
responses (and presumably their thinking about responding) to
student errors. Allwright’s desire “to encourage teachers to see
themselves as researchers, and to bring teaching and research close
together” (p. 251) suggests a more active role for teachers in the
research process.
Van Lier’s The Classroom and the Language Learner: Ethnog-

raphy and Second-Language Classroom Research makes a strong
case for the importance—for teaching practice as well as for
theoretical understanding—of a more contextualized analysis of lan-
guage use in L2 classrooms. In his words:
The teaching profession is ill-served by pedagogical recommendations
which isolate specific observable phenomena, such as types of questions
asked, time lapses between answer and evaluation, and so on, without
720 TESOL QUARTERLY

showing how such phenomena flow naturally from the kinds of activities
that are conducted. It makes no sense to recommend a change in those
isolated aspects without addressing the underlying organizational
characteristics of the activities themselves. (p. 215)
The term context is inherently problematic: How much is enough
for any one researcher’s analysis or any one teacher’s reflection?
The author acknowledges the existence of both “local” and
“nonlocal” aspects of context (which he calls “forces” [p. 179]), and
Candlin affirms the importance of both in his series editor’s preface.
This book is devoted to the local aspects that can be captured by
methods of discourse analysis, for which van Lier adopts Erickson’s
term microethnography. (Surprisingly, especially given the label
ethnography in the book’s subtitle, van Lier tells us nothing about
the speakers of the talk he analyzes except that the classrooms were
in either the U.K. or the U.S. )
Chapters 1-4 present a detailed argument for the importance of
such work, followed by three chapters on van Lier’s own microeth-
nographic research: on turn-taking, activities and their participant
structures, and repairs.
Chapter 5, on turn-taking is the longest in the book (55 pages)
attesting to the importance to van Lier of student participation and
initiative. In the author’s words, “I have not aimed to prove that
active participation . . . is beneficial, but rather to show how
interaction can be analysed. . . . The importance of participation
and initiative has been assumed on the basis of recommendations
from communicative teaching theory, and evidence from first-lan-
guage studies” (p. 144). The result is a detailed four-level category
scheme which van Lier recommends for further research that could
not only relate participation to learning but also show how such
participation happens.
In chapter 6 van Lier analyzes five kinds of activities—which he
calls conversation, telling, elicitation, ritual, and group work—in
terms of the degree of prespecification by the teacher, topic and its
limitations, social rules and constraints, and participant structure.
For example, whether a teacher question is intended, or interpreted,
as a “test question” or an authentic request for student knowledge or
opinion will depend largely on the matrix activity (which Levinson
[1979], in a particularly useful article, calls “the matrix language
game”). If teachers want to shift the proportion of such questions,
they will have to think in terms of a shift in the larger activity or
game.
Chapter 7 discusses “The Organization of Repair in Second-Lan-
guage Classrooms.” This title signifies the important contemporary
REVIEWS 721
shift away from behaviorist notions of student error and teacher
correction toward interfactional processes of adjustment between
speakers and hearers, and constructivist notions of learner
hypotheses and self-repair. (Several recent obituaries for Pit Corder
have paid tribute to his key role in this reconceptualization of
learner error.) As with turn-taking, van Lier offers a detailed
category scheme for classroom discourse analysis. Both schemes
should be useful for second language classroom researchers but are
probably too detailed for classroom teachers to use (though
Candlin’s preface suggests they area primary audience).
For van Lier as for other microethnographers, “context” refers to

the immediate context of activity and discourse within it. This is one
important meaning, but not the only one. Schiffrin (1987) opens her
article on “Discovering the Context of an Utterance” with one
categorization of how aspects of context are divided among
branches of linguistics:
Pragmatics focuses on how language is influenced by frames of mutual
knowledge, discourse analysis on the influence of textual structures and
conversational patterns, sociolinguistics on the influence of social
situation and speaker/hearer identities, and the ethnography of commu-
nication on the influence of cultural constructs. (p. 11)
Without devaluing the discourse analysis presented by van Lier,
readers should realize that this is not all that ethnography can
contribute to second language research and teaching. (See Watson-
Gegeo, 1988, for a comprehensive review.) Not discussed in any of
these three books are two less local but potentially important
contexts: the cultural expectations and norms of behavior brought
into the classroom by the participants and the institutional context in
which they are embedded.
First, the participants’ cultural background: Precisely because
communicative language teaching depends on student participa-
tion, differences in background experience and expectations
between teacher and students can be critically important, even
more so than in native language classrooms (Cazden, 1986, 1987, for
researchers; 1988, for teachers). Ron and Suzanne Wong Scollon
describe one dramatic example. Drawing on their previous research
on Native American/Canadian patterns of interaction (Scollon &
Scollon, 1981), their own 20-year Anglo/Chinese marriage, and their
recent experience teaching English for two years in Taiwan and
Korea, they offer a rich analysis:
The Confucian emphasis on benevolence and respect between teacher
and student makes it difficult to use conversational English teaching
722 TESOL QUARTERLY

methods which are based on Western assumptions about classroom
relations between teachers and students. If an English teacher does not
first take this cultural difference into account, the direct application of
conversational methods can produce frustration in the teacher and a
sense in the students that the teacher is not behaving appropriately. . . .
A successful language program for Chinese students (and we think for
other Confucian Asians as well) first needs to establish a cross-cultural
understanding of the differences between Chinese face-based
interpersonal relationships and the Western conversational model of
negotiated interpersonal relationships. . . . [Then] conversational
methods of language teaching can most successfully be applied by
strategies which produce longer teacher interturn pauses and a richer
teacher language environment. (Scollon & Scollon, 1990, p. 2)
Without such a cultural dimension, even the kind of microethno-
graphic analysis advocated by van Lier could help only to describe
a problem, but would fall short of the kind of interpretation
necessary for devising a solution.
Second, the institutional context: Ulichny (1989) analyzes a non-
credit reading course for beginning ESL students at an urban state
university through classroom observations and interviews with the
teacher throughout the year. She shows how the marginalized
academic status of these students, this course, and the teacher
herself affects how the teacher thinks about, and acts in, her
classroom.
In its broadest interpretation, process-product research is a useful
label for all attempts—as researchers of someone else’s classroom or
our own—to determine which experiences have what kind of effect
on the participants that lasts and is transferable beyond the
momentary time and place; there is no inherent reason, only the
contingencies of history and structure of academic work, why such
research should be limited to work within one academic discipline
(predominantly psychology) and one way of analysing “process.”
All three of these books contribute to opening up the meanings of
the term process and ways to study it, and Chaudron’s Chapter 6
begins to do the same for the term product.
Admittedly, no one researcher—unless working as a member of
an interdisciplinary team—can take into account all, or even several,
of the nested levels of context that Schiffrin lists. However, as
Chaudron suggests (pp. 23, 46), a minimal ethnographic perspective
is essential for all research to ensure the validity of the researcher’s
analytic categories and then to ensure the appropriateness of
variations on classroom discourse that are essential for the varied
situations of L2 teaching, particularly “communicative language
teaching” around the world (Savignon, 1990).
REVIEWS 723
REFERENCES
Cazden, C. B. (1986). Classroom discourse. In M. E. Wittrock (Ed.),
Handbook of research on teaching (3rd ed., pp. 432-463), New York:
Macmillan.
Cazden, C. B. (1987). Language in the classroom. In R. B. Kaplan et al.
(Eds.), Annual review of applied linguistics (Vol. 7, pp. 18-33). New
York: Cambridge University Press.
Cazden, C. B. (1988). Classroom discourse. Portsmouth, NH: Heinemann.
Gage, N. L., & Needels, M. C. (1989). Process-product research on
teaching: A review of criticisms. The elementary school journal, 89,
253-300.
Levinson, S. C. (1979). Activity types and language. Linguistics, 17, 365-99.
Savignon, S. J. (1990, March). Communicative language teaching:
Definitions and directions. Plenary address, Georgetown University
Round Table on Language and Linguistics, Washington, DC.
Scollon, R., & Scollon, S. B. K. (1981). Narrative, literacy, and face in
interethnic communication. Norwood, NJ: Ablex.
Scollon, R., & Scollon, S. W. (1990). Some cultural aspects of teaching
English to Asian adults. Unpublished manuscript, Haines, AK.
Schiffrin, D. (1987). Discovering the context of an utterance. Linguistics,
25, 11-32.
Ulichny. P. D. (1989). Exploring a teacher’s practice: Collaborative
research in an adult ESL reading class. Unpublished doctoral
dissertation, Harvard Graduate School of Education, Cambridge.
Watson-Gegeo, K. A. (1988). Ethnography in ESL: Defining the essentials.
TESOL Quarterly, 22 (3), 575-592.
COURTNEY B. CAZDEN
Harvard University
Writing Across Languages: Analysis of L2 Text. Ulla Connor and

Robert B. Kaplan (Eds.). Reading, MA: Addison-Wesley, 1987.
Pp. v + 202.
This volume of commissioned papers is a useful and informative

addition to the growing literature on second language writing
research. The editors hope that the collection will make an
important contribution to two related aspects of language study.
First, they are concerned with “the characteristics of writing as a
mode of language use (not just an alternative medium) different
from speech and worthy of particular attention” (p. iii). Second,
they wish to emphasize “that texts have schematic structures which
are culturally variable” (p. iii-iv).
724 TESOL QUARTERLY

There are 10 papers in this collection. The first two focus on
theory and background. The next five are concerned with
expository and argumentative text. The final three are “inter-lan-
guage” (cross-cultural comparison) studies. Each paper is followed
by several questions for study and discussion, making this collection
useful as a textbook for graduate courses in TESOL or applied lin-
guistics.
Kaplan’s paper, “Cultural Thought Patterns Revisited,” which
opens the first section, provides a state-of-the-art discussion of
contrastive rhetoric. He argues that although various rhetorical
modes are possible in any language, each language has certain clear
preferences. He points out the differences between oral and written
discourse and briefly discusses the topic of universals in writing.
Kaplan’s paper is thought provoking and establishes a framework
for the papers that follow. In “Text Linguistics for the Applier: An
Orientation,” Enkvist discusses four types of models operating in
text linguistics: sentence-based, predication-based, cognitive, and
interfactional. He also discusses the difference between structural
and process models of linguistic analysis. This theoretical paper
introduces some of the basic terminology linguists use to discuss
written text.
The next five papers focus on models of exposition and argu-
ment. In “Text as Interaction: Some Implications of Text Analysis
and Reading Research for ESL Composition,” Carrell reviews
several empirical findings directly relevant to ESL composition
instruction. She advises “that teaching ESL writers about top-level
rhetorical, organizational structures of expository text, teaching
them how to choose the appropriate plan to accomplish specific
communication goals, and teaching them how to signal a text’s
organization through appropriate linguistic devices should all
function to make writing more effective” (p. 54). What Carrell has
done, as she herself notes, is “stated the obvious” (p. 55) for those
who view reading and writing as complementary processes.
Connor’s “Argumentative Patterns in Student Essays: Cross-Cul-
tural Differences” presents a system for describing and evaluating
argumentative writing. She examined essays written by high school
students in England, Finland, Germany, and the United States for
(a) argumentative text structure (situation + problem + solution +
evaluation), (b) succession of speech acts, and (c) direct appeal to
the audience. Although there was little variation in average holistic
scores, Connor did find some cross-cultural variation in discourse
style.
In the following paper, “A Contrastive Study of English Exposi-
tory Prose Paraphrases,” Connor and McCagg argue that language
REVIEWS 72.5
proficiency and writing fluency—not cultural factors—may account
for native/nonnative variation in writing. In this study, native
speakers of English, Japanese, and Spanish were asked to recall in
writing a previously read passage. The researchers found few
differences between the nonnative groups. The nonnative writers,
however, did differ from the native writers in the amount of detail
and support they provided for their generalizations.
Lantamatti’s “Observation on the Development of the Topic of
Simplified Discourse” is somewhat lengthy and technical. It
investigates factors that contribute to text simplification other than
sentence length, syntactic complexity, and vocabulary type.
Although the paper raises some important points, it appears out of
place in this collection, seeming more suited to a book on reading
comprehension.
Another technical article, Grabe’s “Contrastive Rhetoric and
Text-Type Research,” attempts to define expository prose through
the quantification and statistical analysis of linguistic variables. This
study raises three questions: Do distinct text-types in fact exist? If
they do exist, can they be defined objectively? How can text-types
be compared across languages? In a computer analysis of texts from
academic journals, college textbooks, and popular journals, Grabe
found that expository writing differs from other kinds of writing on
such dimensions as immediacy of context, interactive-informational
orientation, and logical versus situational context.
The third group of papers on inter-language studies begins with
Hinds’s “Readers versus Writer Responsibility: A New Typology.”
Hinds postulates that the degree of involvement expected of
readers will depend on their cultural background: “What this means
is that in some languages such as English, the person primarily
responsible for effective communication is the speaker, while in
other languages such as Japanese, the person primarily responsible
for effective communication is the listener” (p. 143). He argues that
a breakdown in communication in English is attributed to the
speaker (or writer) not having made the message clear; in some
other languages it is considered the fault of the listener (or reader)
for not exerting enough effort to understand.
In “Written Academic Discourse in Korean: Implications for
Effective Communication” Eggington, like Hinds, asserts there is a
“contract” between reader and writer; that is, readers expect writers
to provide information in certain predictable orders. He describes
traditional Korean academic discourse and shows how the linear
style of English academic prose has begun to influence some
Korean writers. In a recall study, Eggington found that Korean
students understood passages written in Korean better if the
726 TESOL QUARTERLY

passages followed a traditional rhetorical pattern than if they
followed a pattern common in English.
The final paper, Ostler’s “English in Parallels: A Comparison of
English and Arabic Prose,” tries to explain why Arabic-speaking
students, who seem to have mastered most English grammatical
structures, continue to produce foreign-sounding essays. Through
statistical and discourse analysis, she shows that the English prose
style of Arabic-speaking students is quantitatively and conceptually
different from that of native writers of English.
The editors have aimed in this volume to make available recent
research in contrastive rhetoric that has relevance for the classroom
(the “primary concern is to provide practical applications rather
than theoretical positions,” [p. 2]). Despite this claim of practicality,
however, readers for the most part will have to draw their own
conclusions about how to translate this research into practice.
Instructors looking for step-by-step guidance will need to look
elsewhere. The critical reader will also note problems in the design
of several of the studies, for example, equating recall protocols with
original essay writing or comparing mismatched samples (student
writing with professional writing). Regardless of technical flaws in
some of the research, the case for contrastive rhetoric is, on the
whole, well grounded. This collection is a much needed contribu-
tion to the area of second language writing. Readers with an interest
in this area will find this book informative and worthwhile reading.
MELVIN R. ANDRADE, JR.

Nanzan University, Nagoya, ]apan
REVIEWS 727
BOOK NOTICES
The TESOL Quarterly welcomes short evaluative reviews of print and nonprint
publications relevant to TESOL professionals. Book notices may not exceed 500
words and must contain some discussion of the significance of the work in the
context of current theory and practice in TESOL.
The Literature-Enriched Curriculum

Recognition of the importance of context and of authentic linguistic and
cultural input has caused many teachers to reassess the role of literature in
ESL/EFL teaching and to recognize it as a rich and enabling resource for
their students. Accordingly, this issue’s book notices focus on the topic of
literature and present those interested in enriching their curriculum
through literature with a look at several teacher resource texts as well as
some of the more recent student texts that are literature-based. A broad
definition of literature has been adopted; folk tales and student-generated
literature are included with works of great literature written by famous
authors. Similarly, anthologies relevant for different audiences and which
have been compiled with different teaching purposes in mind have been
included.
Not among the books reviewed below are several texts that have been
reviewed in previous issues of the TESOL Quarterly. Since these texts also
present a rich resource for teachers, a brief annotated bibliography
follows. Those wishing more information are directed to the specific issue
in which the book was reviewed.
Literature in the Language Classroom: A Resource Book of Ideas and

Activities. Joanne Collie and Stephen Slater. Cambridge: Cambridge
University Press, 1987. Pp. vi + 266.
This teacher resource text proposes a host of activities that assist ESL/EFL
students to enter into literature, guide them through it, and transport them
beyond it to the realm of ideas. [Reviewed 24 (2), 305-306]
Voices: New Writers for New Readers. Patty Richards (Ed.). Surrey,
British Columbia, Canada: Lower Mainland Society for Literacy and
Employment, 1990.
Geared for the adult literacy market, this quarterly journal includes student
works (writings, art, photos), thus validating the use of student-generated
literature to enhance students’ literacy skills. [Reviewed 23 (2), 333-334]
Don’t Be Afraid, Gringo: A Honduran Woman Speaks From the Heart:

The Story of Elvis Alvarado. Medea Benjamin (Trans. and Ed.). San
Francisco: The Institute for Food and Development Policy, 1987.
729
Based on the oral history of a Honduran peasant woman, this text provides
a culturally sensitive vehicle for promoting adult literacy skills. [Reviewed
23 (2), 324-325]
The House on Mango Street. Sandra Cisneros. Houston Arte Publico

Press, 1985. Pp. 103.
This collection of poems and short stories, which centers around the life of
a young Chicana, offers new readers material that mirrors their own
immigrant experience. [Reviewed 23 (2), 326-327]
Tell Me About It: Reading and Language Activities Around Multi-

Cultural Issues Based on an Oral History Approach. Azi Ellowitch. Phila-
delphia: LaSalle University, Urban Studies and Community Services
Center, 1986. Pp. vi + 150.
This Freirian approach to literacy development employs the genre of the
interview, documenting the lives of representatives of various cultural
groups. [Reviewed 23 (2), 327-328]
Need I Say More: A Literacy Magazine of Adult Student Writings.

Boston: Publishing for Literacy Project.
Like its Canadian counterpart Voices, this journal allows the authentic
voices of new writers to be employed in the venture of adult literacy
teaching. [Reviewed 23 (2), 331-332]
DONNA BRINTON, Guest Editor

Literacy Through Literature. Terry D. Johnson and Daphne R. Louis.

Portsmouth, NH: Heinemann, 1987. Pp. vi + 160.
More and more school districts around the country have adopted a
literature-based curriculum for teaching literacy in the elementary grades.
Many teachers, while glad to see “real books” rather than basal readers as
the basis of the school’s literacy program, find themselves searching for
effective ways to promote literacy skills through literature. For teachers
seeking such practices, Literacy Through Literature presents a rich
compilation of practical and creative techniques for reading and talking
about stories and poems with children. While designed for English
speakers, the techniques, with some modification, are equally applicable
to second language (L2) students.
Chapters 2 and 3 contain teaching strategies for both initial and
developing literacy. Many of the practices outlined in these chapters can
be used effectively with children because, in keeping with sheltered
730 TESOL QUARTERLY

instruction, they introduce difficult concepts and foster critical thinking
skills through concrete approaches to comprehension and active engage-
ment with the characters, plot, and issues of the literature being read.
While the techniques deliberately integrate oral and written language, each
activity type focuses on one of the following learning modalities: (a) text-
based activities such as journals, story ladders, and story or character
report cards require the integration of reading and writing to enhance
comprehension; (b) visually/graphically-based activities such as literacy
posters or plot graphs provide visual reinforcement as a comprehension
tool; and (c) orally-based activities such as readers’ theater and
dramatization allow students to relive orally selected portions of the
literary work. The authors further insist on modelling to guide students
through each new routine, a teaching approach especially beneficial to L2
students who need repeated encounters with language, concepts, and
procedures.
Several limitations in classroom practices must be noted if this resource
book is to be successfully adapted by L2 literacy teachers. Regrettably
absent from the list of stories and poems are selections from other lands
and by ethnic writers. However, the authors do ask users of the book to
experiment: “[the classroom practices] will not be truly yours until you
implement the routines with your words and with books that you and your
children have chosen” (p. 154). L2 teachers will want to use these
templates with literature that draws on the diverse backgrounds of their
students. Modifications also need to be made in the rhymes used for
beginning literacy activities. Poems like “Hickory Dickory Dock” are
unfamiliar to our students, containing unknown words, difficult sound
patterns, and culturally irrelevant ideas. These same activities would be
better undertaken with pattern books and rhymes in more contemporary
language (e.g., Shel Silverstein poems) or even with big books produced
by upper elementary students.
Literacy Through Literature is an indispensable book for every literacy
teacher’s professional library. Its creative techniques help teachers
rediscover the joy of literature and teach children to love words.
CHRISTINE HOLTEN
The Inward Ear: Poetry in the Language Classroom. Alan Maley and Alan
Duff. Cambridge: Cambridge University Press, 1989. Pp. vi+ 186.
The Inward Ear arrives at a time when a growing number of language

teachers feel the need to find out more about a contemporary literature-
based approach to language instruction. As a resource book filled with
creative ideas about the use of poetry in the language classroom, The
Inward Ear proves handy and timely.
In terms of content and organization, the text begins with a brief but
BOOK NOTICES 731

persuasive introduction advocating a place for poetry within the ESL/EFL
program, followed by eight chapters filled with interactive activities
which, if carefully selected, are also challenging and thought-provoking.
Chapters 1-3 together form a cohesive unit: The focus of the first chapter
is schema-activating tasks that facilitate access to the poem; the second
chapter covers activities that help students grapple with both the language
and the ideas of the poem, and the third chapter gives assignments that
prompt student reactions to the poem. At the heart of The Inward Ear,
Chapter 4 pays homage to the sound patterns of poetry, devoting itself to
read-aloud activities. Next, Chapters 5-8 form another cohesive unit by
virtue of their common focus on poetry-anchored writing activities.
A noteworthy pedagogical feature of Maley and Duff’s activities is that
they promote the integration of language skills—calling, for example, for
paired and/or small group discussions to help increase student interaction
and utilizing peer input to generate progressively more polished student
responses, both orally and in writing. In the process, students practice all
four language skills in the context of authentic and communicative linguis-
tic exchanges.
An important assumption in The Inward Ear is that “difficulty is a
relative concept” (p. 15, italics added). Thus, Maley and Duff grant that
some poems are less accessible than others—due, for example, to their lin-
guistic complexity, their culture-specific opacity, or even to students’
naivete about some of life’s experiences. The authors nevertheless defend
each teacher’s right to select poetic texts that some might judge “too
difficult,” noting that too often in the past our notions of a poem’s
difficulty have been biased by the presence/absence (as well as by the
frequency and type) of unknown words found in a poem. Since schema-
activating and other comprehension facilitating activities are better
understood and (hopefully) more widely practiced these days, the authors
unabashedly leave the task of poem-selection in the hands of the language
teachers themselves, and quite rightly so.
The authors note that The Inward Ear should interest language teachers
working with students at “the secondary school level upwards” (p. 1). I
would agree with this wide audience designation since most of the text’s
activities can readily be adjusted to satisfy different ability levels. Because
the ESL/EFL students’ language experience is greatly enriched by being
exposed to the wonder and power of poetry in languages other than their
own, one would hope to see this invaluable handbook added to each
language teacher’s reference library.
FOONG HA YAP
732 TESOL QUARTERLY

Longman American Structural Readers: Horizontal Readers. Various
authors. New York: Longman.
Sixteen of the Longman American Structural Readers are horizontally

shaped texts containing high-interest simplified English fiction and
nonfiction. The top half of each page contains a series of colorful,
artistically designed pictures that provide a pictorial representation of the
storyline—thus making the story itself comprehensible. The pictures are
modern and invite potential readers to engage in a rewarding experience.
Designed for teens and young adults, the readers include biographies as
well as romance and adventure.
Of course, students can begin reading authentic materials even faster and
increase their comprehension level when provided with guided activities
before, during, and after the reading. The student workbooks that
accompany the readers provide a great variety of guided exercises. Some-
times, however, students may enjoy being able to read alone without being
involved in guided activities. Use of the workbooks is therefore optional,
with teachers being able to choose those exercises that best suit their
objectives and/or the needs of their students. Outside of structured class
time, the readers can also be used to augment a program’s free-reading
component. In such a situation, the workbooks may or may not be used.
Current language teaching pedagogy reflects ambivalence towards
simplified literature. On the one hand, this form of literature is heralded as
a means for beginning ESL students to read extensively while building
confidence in their reading skills and gaining appreciation for works of
literary value. On the other hand, it can be cursed for distorting the
authentic source material. Even more seriously, simplified literary texts are
sometimes accused of producing prose that lacks the coherence and
natural redundancy of the original work; that is, they fall short of their
aim—to provide a comprehensible prose base.
In examining a series such as The Longman Structural American
Readers, one needs to be mindful of arguments against simplified
literature. Nonetheless, the question here is not so much whether
simplified literature is intrinsically good, but rather whether it provides
students with a means to learn English and whether it motivates them to
continue to do so. I would contend that these readers do achieve their
purpose of providing comprehensible, interesting input, and are highly
motivating to beginning-level ESL students. I would argue, further, that
students who envision themselves as readers of English will persevere, will
continue to read, because they have joined the “literacy club” (Smith,
1988).
In sum, reading in a second language is a process that requires nurturing,
with students continually being encouraged to immerse themselves in the
richness of (simplified or original) literary text. We must trust that the
process will evolve as the appetite for literacy grows through successful
experiences. With this graded series of readers, we can trust that it will
evolve to include more complicated texts.
BOOK NOTICES 733

REFERENCE
Smith, F. (1988). Joining the literacy club. Portsmouth, NH: Heinemann,
ANNETTE C. SHEEHAN
California State University, Los Angeles
Forestville Tales: International Folk Stories. Aaron Berman. New York:

Collier Macmillan, 1977. Pp. 71.
In the introduction to Forestville Tales, the author states that his book is
a supplementary reader for young and young-at-heart ESL learners. No
mention is made of the students’ English proficiency level or their first lan-
guage background. However, since the stories are in simplified English, it
can be safely assumed that this text is appropriate for lower-level ESL
students who are still developing their English language skills.
The presentation of the eight international folk stories in Forestville
Tales conforms to procedures advocated in current reading research. As
explained by Berman in his introduction, each literary selection is prefaced
by prereading activities and culminates in a variety of reading and post-
reading activities. More specifically, each story is introduced by a
vocabulary list and is followed by four exercise types selected from the
following: True-False, Vocabulary Practice, Fill Ins, Write Your Own
Sentence, Circle Exercise, Sentence Combining, and Verb Practice. Each
unit is self-contained, allowing the teacher to choose among readings; each
selection is short enough to be covered in one or two lessons.
Although some might balk at the use of simplified English in the ESL
classroom, Forestville Tales is a textbook I would recommend, especially
for young ESL learners. The use of simplified English in these stories does
not detract from the ideas. Inasmuch as these stories are entertaining and
originate from various cultures, the stories in this book relate to issues with
which students can identify. These tales also fulfill an important need of
limited English readers. They provide students a successful experience of
having read and understood written English text. Accordingly, the use of
simplified English in these stories does not detract from their usefulness.
Nevertheless, there is still room for improvement. According to the
author, “Forestville Tales is designed to develop reading and speaking
skills” (p. 7). However, the exercises accompanying each story promote
speaking and comprehension rather than reading skills. Outside of the
suggestion that the students read silently as the teacher reads the story
aloud, there is little reading practice. I would also have preferred to see
some activities related to each story’s theme, either within each story unit
or as suggested activities in the introductory section.
The text’s soft cover with its illustrations from the stories connotes a
book designed for pleasure reading. The typeface is large and distinctive
(e.g., boldface titles) and the illustrations are very clearly drawn. In spite
734 TESOL QUARTERLY

of its minor drawbacks, this book would enliven any ESL class for young
learners. I would. not hesitate to recommend it for this context.
FANG-LIAN LIAO
Outsiders: American Short Stories for Students of ESL. Jean S. Mullen.

Englewood Cliffs, NJ: Prentice Hall, 1984. Pp. xxi+ 223.
In our field today, there are two widely divergent approaches to the use
of literature in the classroom: Literature is seen either as a vehicle for the
detailed study of usage (i.e., structure and form) or as a springboard for
language use (i.e., true communication). While Mullen’s anthology could
be adapted to teach language use in a communicatively oriented
classroom, the activities and format of this book lean heavily toward the
former approach, that is, toward investigating how U.S. authors
manipulate language and structure to express their thoughts.
The first chapter presents a detailed description of how to study
literature, including a thorough explanation of the vocabulary aids that
accompany each selection. As this text is intended to introduce “nonnative
English users in . . . universities” (p. xix) to the study and analysis of
literature and aims to improve students’ writing skills, the emphasis the
author has chosen to place on structure and form is warranted.
The remaining eleven chapters all follow the same format: Each short
story selection includes an extensive glossary (including informal speech
and standard vocabulary) followed by comprehension/discussion
questions, vocabulary exercises, and a variety of language usage exercises.
The selections are lightly based on the theme of “outsiders” in U.S. culture;
yet this is not a collection of immigrant experiences, but rather stories of
any type of outsider, including anyone not in the cultural mainstream of
society. Several of the stories, such as “The Test” and “I Want to Work,”
are ideal for inter- and intracultural discussion.
While the stories themselves are provocative and could lead to
stimulating discussions on cultural idiosyncrasies, there are few exercises to
exploit these possibilities. In addition, there is no provision made to
introduce students to the readings through “schema activating” or other
such activities. Even when using this text with advanced students—who
may have no overt difficulty reading the text—teachers will need to plan
additional activities to assist students in comprehending subtle details and
plot twists. An additional disappointment are the illustrations, which are
cartoon-like in nature and detract from the presentation rather than
enhance comprehension.
There is a teacher’s manual accompanying the text, which in addition to
providing answers to the exercises also provides an occasional suggestion
for activity expansion. I highly recommend that teachers read the sections
entitled “Suggestions and Ideas” and augment these suggestions with their
own repertoire of activities.
BOOK NOTICES 735

Overall, this anthology has potential for use in advanced classes,
especially at the university and university preparatory levels. While there
are many shortcomings in the prereading activities, a teacher with a
moderate investment of time could overcome these and put the existing
writing and language focus sections to good use.
LYN M. REPATH
At the Door: Selected Literature for ESL Students. Sandra McKay and
Dorothy Petitt. Englewood Cliffs, NJ: Prentice Hall, 1984. Pp. xi+ 142.
At the Door is a collection of literature for high-intermediate to

advanced ESL students. According to the authors, the selections were
chosen for their relevance to the previous experiences of language
minority students and for their linguistic accessibility.
The first selection, a poem by Adrienne Rich, sets the book’s thematic
tone:
The door itself
Makes no promises.
It is only a door.
(p. 2)
Like much of the text, this poem deals with the problems encountered by
immigrants who have passed through “the door” from their native country
and culture into a strange and often discriminatory environment.
The book’s first topic concerns adjusting to a new culture. Thus in “West
Side Story,” Anita and Rosalia argue the pros and cons of life in the U.S.
and Puerto Rico. In “A Measure of Freedom,” Jade Snow Wong shocks her
family by doggedly setting her own course to pursue a higher education.
And in “Sweet Promised Land,” Robert Laxalt relates how his father—a
sheepherder from the Pyrenees of France—faced a clash of cultures when
dining at a sophisticated New York City restaurant.
Discrimination and racism is a second theme explored. In two Saroyan
selections, a young Filipino is driven to homicide by a chauvinistic, “real
American” and Saroyan tells of the discrimination he himself suffered in
school because of his outspoken Armenian ways. Maya Angelou
remembers seeing family after family disappear to the Japanese-American
internment camps of World War II while Jeanne Wakatsuki Houston
writes of her mother’s anguish when selling the family’s belongings before
their forced exodus to Manzanar.
Finally, several selections show immigrants who rise above their
conditions. In “Doria Ramirez,” a young migrant worker refuses to accept
her lot and becomes a labor organizer. And in Clark’s “Medicine Man’s
Daughter,” a Native American leaves the reservation to attend private
boarding school and achieves an uneasy adjustment.
736 TESOL QUARTERLY

The text’s activities go far beyond the usual postreading comprehension
questions and include penetrating questions that promote reflection and
incite students to explore ideas in the context of their prior experience. In
order to increase their command of English, students also learn new
vocabulary via context clues, make judgments on the effectiveness of
metaphorical language, and focus on such aspects as sentence structure
and idiom. Many selections are followed by suggestions for composition
topics that, although based on the content of the reading, require the
students to go beyond the reading to express their own ideas.
Aside from these extensive and varied activities, what impresses me
most is that virtually all the literary pieces are culturally relevant, linguis-
tically accessible, and of high literary quality. My only regret is that the
text does not include more selections showing people successfully passing
through “the door” and achieving their goals. This is perhaps an unrealistic
expectation given the extreme hardships faced by many of our students,
but I believe that as ESL teachers, we have a responsibility to inspire hope
in our students; otherwise, we cannot expect them to give us their best
efforts.
MARTHA DUNN
Manual Arts High School
Changes: Readings for ESL Writers. Jean Winthrow, Gay Brookes, and
Martha Clark Cummings. New York: St. Martins Press, 1990. Pp. 225.
Changes is a literature-based rhetoric focusing on reading and writing

skills for adults at intermediate to advanced levels. It is attractive, easy to
use, and provides teachers with numerous activities for initiating student
writing. The readings are varied—both by the ethnicity of their authors
and by their complexity of thought and language—and range from such
canonized professionals as Robert Frost to student writers.
Grounded in current theory, Changes emphasizes student self-
expression and encourages individual, personalized responses to the
readings. Thus, for example, throughout the text it is assumed that students
are maintaining a log or journal for both pre- and postreading activities.
Clearly, much beneficial student interaction based on collaborative
learning could result from these activities.
Notwithstanding these attributes, there are several features which
detract from the text’s attractiveness. Some of these are minor, such as the
choice of Changes as a title. In fact, some of the readings seem much more
closely related to this concept than do others. As well, the book is visually
disappointing, failing to take advantage of numerous opportunities to
provide visual stimulation as an invitation into the selections.
Coexisting with these more minor problems are a few more serious ones,
such as the text’s lack of coherence and the varying difficulty levels of the
reading selections. In Chapter 1, for example, the eclecticism of the
BOOK NOTICES 737

selections is a bit overwhelming, with Frost’s “Stopping by Woods”
immediately preceding two student narratives dealing with emigrating
from China to the U.S. The connection between these readings seems
tenuous at best, and the language problems are significantly different.
A more fundamental problem is exemplified by the use of Jamaica
Kincaid's “A Walk to the Jetty,” the last chapter of her lovely novel Annie
John. The beginning of the chapter, dealing with Annie’s cool, awkward
breakfast with her parents before embarking for England, has been
omitted. This type of anthologizing is considered acceptable by many. But
it raises the question of why students shouldn’t read the entire chapter.
Finally, the Kincaid selection also evidences the text’s lack of attention to
vocabulary. One postreading activity asks the student to deal with simile
and metaphor, concepts introduced only in footnotes. Such sophisticated
concepts need greater attention than provided here.
The short, abbreviated readings evidence that Changes is not so much a
reading and writing text as a writing text with readings. It is deficient in
attention to vocabulary and so culturally eclectic that it is sometimes
confusing. It is also apparatus heavy, with somewhat redundant pre- and
postreading activities presented in great detail throughout. Perhaps its
greatest advantage is its teacher friendliness. For this reason, Changes is
probably a good text to hand an instructor on Monday when class begins
on Tuesday. It is theoretically sound, flexible, and readable, and would
serve well in a variety of classroom settings.
FRANK BEYER
Pierce College
Literary Portraits: An Anthology of Modern American Prose and Poetry

for Students of English. J. Donnie Snyder. New York: Macmillan, 1989.
Pp. xiii + 217.
Written for students at the high-intermediate level, Literary Portraits

presents 24 essays and 3 poems, ranging in length from one page to 14
pages. The texts are unsimplified and define only those words and
expressions not to be found in an English-language dictionary (i.e.,
“sardine transportation: tightly-crowded conditions in public transporta-
tion” [p. 41]). The readings are arranged into three sections, each
containing one poem and eight essays. Part I, “Observing the Melting Pot,”
deals with ethnic diversity in U.S. culture (e.g., “At Ellis Island,” “Halfway
to Dick and Jane: A Puerto Rican Pilgrimage,” “Jackie Robinson: Each
Game Was a Crusade”). Part II, “American Introspection,” deals with
everyday cultural aspects of life in the United States (e.g., “Rituals at
McDonald’s,” “The Invisible Poor,“ “Portrait of an American: Ralph
Nader”). Part III, “Changing Roles and Perspectives,” deals with current
social issues (e. g., “You Are What You Say,” “Those Good Ole Boys,” “You
Haven’t Come a Long Way, Baby: Women in Television Commercials”).
738 TESOL QUARTERLY

Each reading selection is prefaced by a short biographical paragraph
about the author, a brief background paragraph to prepare the reader for
the story’s subject, and two or three prereading considerations. Following
each essay are a list of Discussion Points, a variety of “Composition
Topics,” and several research ideas for interested students.
This book is excellent in its presentation of authentic works written in a
highly literate style. Students are exposed to an impressive range of
vocabulary and sentence structure—features that are often edited out of
ESL reading anthologies. The selections are thought provoking as well; the
prereading considerations lead students to look at their beliefs and ideas
about a subject before reading an essay dealing with that subject. The
Discussion Points do not simply ask the student to parrot information, but
rather to comprehend the underlying subtleties in each essay. The
Composition Topics provide students a wide range of choices such as
putting themselves in the place of someone in the reading, comparing well-
known people in the U.S. to people from the students’ own countries, or
applying the new knowledge from the reading to a present-day issue.
This text helps international students access U.S. society and culture in a
realistic manner. It lends itself to speaking, reading, and writing activities
in both academic and nonacademic settings and brings together a valuable
collection of readings for any instructor at the high-intermediate to
advanced levels.
MELINDA KODIMER
BOOK NOTICES 739

BRIEF REPORTS AND SUMMARIES
The TESOL Quarterly invites readers to submit short reports and updates on their
work. These summaries may address any areas of interest to Quarterly readers.
Authors’ addresses are printed with these reports to enable interested readers to
contact the authors for more details.
Edited by GAIL WEINSTEIN-SHR

University of Massachusetts at Amherst/
Temple University
DIF in Native Language and Gender Groups in

an ESL Placement Test
ANTONY JOHN KUNNAN
■ The relative cultural fairness of educational tests has been a fertile field
for researchers. The primary concern of such studies has been tests or test
items that behave differently for different native-language, cultural,
ethnic, and/or gender groups. These studies are popularly known as test
bias studies but the term test bias is misleading and inaccurate. The term
assumes that the researcher is examining tests or test items that are biased
while what we know from test-taker responses is only that there are
differences in performances by different individuals or groups that could
be due to several reasons, only one of which is bias. A more accurate term,
therefore, that has been used recently in testing literature, is differential
item functioning (DIF), referring to the way items function differently for
individuals or groups of test takers who have similar abilities.
Among the few studies on DIF in the field of second/foreign language
testing reported recently are the following: Chen and Henning (1985)
examined the Winter 1985 version of the ESL placement examination
(ESLPE) at the University of California, Los Angeles (UCLA) to
determine the nature, direction, and extent of bias present for members of
the Spanish and the Chinese native language groups; Zeidner (1986, 1987)
investigated the English Language Aptitude Test (ELAT) used routinely
for student selection and placement in Israel; Spurling (1987) studied the
“fair use” of the Marin Community College English admissions and
placement test; Hale (1988) reported on the interaction of student major-
field and text content in the reading comprehension section of the Test of
English as a Foreign Language (TOEFL); and Angoff (1989) tested the
hypothesis that items of the TOEFL that contain references to people,
places, regions, etc., of the U.S. tend to favor test takers who have spent
some time living in the United States. Kunnan and Sasaki (1989) extended
the Chen and Henning (1985) study by including five native language
741
groups in their examination of the Fall 19S7 version of the ESLPE. Their
study, like that of Chen and Henning (1985), identified several vocabulary
and grammar items that favored certain native language groups. Other
studies that have investigated performance across native language groups
include Alderman and Holland (1981), and Ohman, Stricker, and Barrows
(1988).
METHOD
The present study is concerned with the identification of differential item
functioning among four native language groups and the two gender groups
in the Fall 1987 version of UCLA’s ESLPE. (This version is different from
the Winter 1985 form used in the Chen and Henning (1985) study in terms of
the actual items though the composition of the testis the same.)
The sample for the study was 844 nonnative-speaking entering students
at UCLA. These students were from 72 countries, with 61 language
backgrounds, pursuing 76 academic specializations. The native language
groups analyzed were the four largest ones: Chinese (262); Spanish (81);
Korean (76); and Japanese (59), for a total of 478. In terms of gender, the
sample was distributed as follows: male, 478 and female, 347 (for a total of
825; 19 test takers did not identify their sex). In its greater variety of native
languages and the inclusion of gender, this study extends the work by
Chen and Henning (1985), and Kunnan and Sasaki (1989).
The test instrument, the ESLPE, consists of 150 items in five 30-item
subtests and one 20-minute composition. The five subtests in this test are
listening comprehension, reading comprehension, grammar, vocabulary,
and writing error detection. The data for the study came from the 150
multiple-choice items that were dichotomously scored for our analyses.
RESULTS
The one-parameter Rasch model from Item Response Theory (IRT)
group, which calibrates item difficulty estimates, was used for this study.1
The analyses did not take into account the possibility of guessing. Item
difficulty estimates were determined for all items using the BICAL
unconditional maximum likelihood estimation procedure for the total
sample. An SAS least-squares regression procedure using the item
difficulty estimates was used to plot all items against all combinations of
native language and gender groups. Items that were placed outside the 95%
confidence interval of the regression plot were considered to be items that
displayed DIF. These items were then examined in detail so that sources of
DIF could be hypothesized.
1 Similar
use of Rasch analysis in identifying DIF was made earlier by Chen and Henning
(1985), and Madsen and Larson (1986). Besides, the unidimensionality assumption
underlying the Rasch model application appeared to be satisfied for earlier, though similar,
data by Henning, Hudson, and Turner (1985) and for the present data by Davidson (1988).
For more details about IRT principles, assumptions, modeling, and applications, see
Henning et al. (1985), for a less technical elaboration, and Muthen and Lehman (1985) and
Hambelton and Swaminathan (1985), for technical expositions.
742 TESOL QUARTERLY

Results of the analysis indicated the following: Thirteen items displayed
DIF in the native language group analysis, and twenty-three items
displayed DIF in the gender group analysis.2
Native Language Group Analysis

The three items that favored the Japanese group and the three items that
favored the Chinese group were all grammar items but different ones.
These grammar items tested the appropriate use of grammar points, such
as the definite article, a preposition, or verb tense. These items may have
been easy for these two language groups because of their instructional
background and/or familiarity with discrete-point grammar testing in
their home countries. Thus, the source for this type of DIF can be
hypothesized to be instructional background both in terms of test content
(grammar-based teaching and testing) and test method (multiple-choice
format).
The four items that favored the Spanish group were all vocabulary
items. The words that were tested were hypothetical, implication, elabo-
rate, and alcoholics. All these words have Spanish cognates making the
items potentially easy for this group. This finding is similar to that of the
Chen and Henning (1985) study, where the words that favored the Spanish
group were approximate, animated, maintain, and obstruct. Thus, the
source for this type of DIF may be hypothesized to be cognates, a test-
content facet based on test-taker native language.
The potential sources of DIF for two grammar items, one that favored
the Japanese group and one that favored the Chinese, and one vocabulary
item that favored the Korean group could not be identified.
Gender Analysis
The 20 items that favored the male group came from all sections of the
test: 7 in listening, 4 in reading, 3 in grammar, 4 in vocabulary, and 2 in
writing error detection. The 11 listening and reading items were based on
the listening and reading passages respectively and, therefore, the content
of the passages could be a source of the DIF. These items were based on
passages from business, culture/anthropology, and aerospace engineering
passages. These subject areas seem to favor the male group. In addition,
one of the vocabulary items tested, simulates, may be used frequently in
engineering/science classes. Out of 478 male students who took part in this
study, 72% indicated that they were engineering/science majors compared
with 24% of the 347 female students. The potential source of this type of
DIF may have been test-taker major field, a test-content facet.
The potential source for DIF for 3 grammar, 3 vocabulary, and 2 writing
2 Thedescriptive results showed that the test is internally consistent across all sections for all
the native language groups and gender groups (KR-20: .95) and that the mean scores for the
Korean and the Chinese groups and the male gender group was slightly higher than the
other groups.
BRIEF REPORTS AND SUMMARIES 743

error detection items could not be hypothesized. In addition, the source
for DIF for 3 items that favored the female group, one each from listening,
vocabulary and writing error detection, could not be hypothesized.
In summary, in both analyses, out of the 150 items in the ESLPE, 36
items (24%) with no overlap between native language and gender groups
were identified as displaying DIF. Potential sources of DIF for 22 of these
items (61%) were hypothesized, leaving 14 items (39%) for which there were
no hypotheses.
DISCUSSION
This study identified three potential sources of DIF for both native lan-
guage and gender groups: instruction, native language, and major field.
But identifying potential sources of DIF is only the first step in this kind of
analysis. The next step, determining what to do with items that display
DIF and how to compensate test takers for such “bias” remains to be dealt
with. Two procedures can be used here: (a) sources causing DIF can be
examined, hypothesized and causally related, if possible, and (b) items
displaying DIF can be improved or discarded. The second procedure
should only be resorted to by test writers and administrators as a short-
term plan: to remedy a test in use, or when there are not enough resources
for a detailed study. The first procedure, which will have a more lasting
effect, is recommended otherwise.
Let us use the first procedure for this study and examine the instructional
background effect. This effect may be due to certain language teaching
methodologies. For example, language learners who are taught through a
teaching methodology that focuses on grammar and multiple-choice,
discrete-point testing may find such items (as on the ESLPE) easier to
answer (Farhady, 1979). This sort of practice effect can be reduced if a
broad range of test content and formats is presented so as not to favor test
takers from any one instructional background.
The problem of the major-field effect is easier to handle. For example,
it is possible to bring about a balance in content on the basis of a test-taker
background questionnaire administered to native language and gender
groups that reveals reading interest, academic background, career interest,
and so forth so that content that is favorable to one group does not
dominate the test. Another way of balancing content would be to invite
test writers from native language and gender groups who are represented
in the test to balance cultural and gender-related material. (See Hale, 1988,
for other suggestions.) Through these means, DIF based on content can be
reduced considerably in a test.
Finally, let us examine the native language source for DIF, the cognate
effect. Any language test can face this problem as specific language family
ties come into play: English is more closely related to Romance and
Germanic languages than, for example, Chinese or Dravidian languages.
Therefore, students with a native language that is closely related to English
will tend to be favored on ESL tests. Many researchers argue that this is a
744 TESOL QUARTERLY

fact of language learning and, therefore, the question of discarding DIF
items that have cognates (a linguistic effect) should not arise. Even so,
many issues remain unresolved: How many questions in a 150-item test
should rely on cognates? How many in each section (e.g., vocabulary,
grammar)? Should the cognates reflect written language or oral language?
And so forth.
CONCLUSION
This study shows that a placement test cannot only be examined for
items that display DIF by using an Item Response Theory, but also the
identification of potential sources for these DIF items can be attempted
and short- and long-term remedial measures to reduce DIF can then be
proposed. A more extensive use of test-taker characteristics, not used in
this study, will strengthen this approach. Methodologically, DIF studies
could also be undertaken using the Mantel-Haenszel approach (Holland &
Thayer, 1988) or Structural Equation Modeling (Muthen, 1989). Finally, a
practical need for this type of study is clear: If a substantial number of
items in a test display DIF, test scores could be unreliable and invalid for
all groups of test takers, not just the affected groups; most tests still make
placement decisions based on norm-referenced rather than criterion-
referenced statistics. An example of the usefulness of this kind of study is
the revision process the ESLPE has been undergoing for the past year,
partly as a result of this study.
REFERENCES
Alderman, D. L., & Holland, P. W. (1981). Item performance across native lan-
guage groups on the TOEFL (TOEFL Research Report No. 9). Princeton, NJ:
Educational Testing Service.
Angoff, W. H. (1989). Context bias in TOEFL (TOEFL Research Report No. 29).
Princeton, NJ: Educational Testing Service.
Chen, Z., & Henning, G. (1985). Linguistic and cultural bias in language
proficiency tests. Language Testing, 2 (2), 155-163.
Davidson, F. (1988). An exploratory modeling survey of the trait structures of some
existing language test data sets. Unpublished doctoral dissertation. University of
California, Los Angeles.
Farhady, H. (1979). The disjunctive fallacy between discrete-point and integrative
tests. TESOL Quarterly, 13 (3), 347-357.
Hale, G. A. (1988). The interaction of student major-field group and text content in
TOEFL reading comprehension (TOEFL Research Report No. 25). Princeton,
NJ: Educational Testing Service.
Hambelton, R. K., & Swaminathan, H. (1985). Item response theory: Principles
and applications. Boston: Kluwer-Nijhoff.
Henning, G., Hudson, T., & Turner, J. (1985). Item response theory and the
unidimensionality for language tests. Language Testing, 2 (2), 141-154.
Holland, P. W., & Thayer, D. T. (1988). Differential item performance and the
Mantel-Haenszel procedure. In H. Wainer & H. I. Braun (Eds.), Test validity
(pp. 129-146). Hillsdale, NJ Lawrence Erlbaum.

Kunnan, A. J., & Sasaki, M. (1989, March). Item bias in UCLA’s ESL Placement
examination. Paper read at the Ninth Second Language Research Forum, Los
Angeles.
Madsen, H. S., & Larson, J. W. (1986). Computerized Rasch analysis of item bias
in ESL tests. In C. W. Stansfield (Ed.), Technology and language testing
(pp. 47-67). Washington, DC: TESOL.
Muthen, B. (1989). Latent variable modeling in heterogeneous populations.
Psychometrika, 54, 557-585.
Muthen, B., & Lehman, J. (1985). Multiple group IRT modeling: Applications to
item bias analysis. Journal of Educational Statistics, 10 (2), 133-142.
Oltman, P. K., Stricker, L. J., & Barrows, T. (1988). Native language, English
proficiency, and the structure of the TOEFL (TOEFL Research Report No. 27).
Princeton, NJ: Educational Testing Service.
Spurling, S. (1987). The fair use of an English language admissions test. The
Modern Language Journal, 71 (4), 410-421.
Zeidner, M. (1986). Are English language aptitude tests biased towards culturally
different minority groups? Some Israeli findings. Language Testing, 3 (l), 80-98.
Zeidner, M. (1987). A comparison of ethnic, sex and age bias in the predictive
validity of English language aptitude tests: Some Israeli data. Language Testing,
4 (l), 55-71.
Author’s Address: Department of TESL and Applied Linguistics, 3300 Rolfe Hall,
University of California, Los Angeles, 405 Hilgard Avenue, Los
Angeles, CA 90024-1531
The Effect of Syntax, Speed, and Pauses

on Listening Comprehension
EILEEN K. BLAU
University of Puerto Rico
Whether or not we fully accept Krashen’s theory of second language

acquisition (SLA), few would deny that comprehensible input (CI), in
conjunction with other factors, is an essential ingredient for SLA.
Acquisition is fueled by exposure to input that is somehow rendered
comprehensible either by the opportunity for negotiation of meaning via
interaction or through the aid of characteristics of the input itself (Long,
1983; Snow as cited in Bohannon & Warren-Leubecker, 1985). In one-way
input, i.e., lectures or a listening lab, the characteristics of the input itself
are important. Long (1985) found that nonnative speakers (NNS)
comprehended a foreigner-talk (FT) version of a lecture significantly
better than an unmodified version. The FT version included rephrasing
and restatements, was syntactically slightly less complex, and was
delivered at a somewhat slower rate. However, it is not possible to
determine which of the modifications (or combination of modifications) is
responsible for the positive effect.
Along with rephrasing, restatements, and simplification, the notion that
slowing down the flow of speech is one of the characteristics of input that
746 TESOL QUARTERLY

facilitates comprehension for NNSs has always been intuitively attractive.
However, studies that manipulate the velocity of speech, either mechani-
cally with variable-speed tape recorders or naturally, have yielded varied
results. Smith (1980) and Dwyer (1987) found that reducing the rate of
speech did not significantly aid language learners, while Flaherty (1979)
and Kelch (1985) did find support for reduced speed. Unfortunately, these
researchers used different rates as the norm, making it difficult to compare
findings and draw conclusions.
The studies reported below represent an attempt to determine which of
several alterations to the input directed to second language learners affect
comprehensibility. The first study manipulated both speed and syntax,
and the second study examined the introduction of pauses into the input.
STUDY 1
Study 1 was conducted with samples from two populations: 72 first-year
English institute students at a university in Poland and 100 students taking
a required English as a second language course at a university in Puerto
Rico. The three versions of the 18 brief passages used in Blau (1982) to
examine the effect of syntax on reading comprehension were recorded at
normal speed (approximately 170 wpm) and then slowed to 85% of normal
(approximately 145 wpm) on a VSC Soundpacer tape recorder. Within
each sample, six existing classes were randomly assigned (by drawing
numbers from a hat) to one of the six treatments: Version 1 (simple
sentences) at slow and normal speeds, Version 2 (complex sentences with
clues to underlying structure left intact) at slow and normal speeds, and
Version 3 (complex sentences without optional surface clues to underlying
structure) at slow and normal speeds. All groups heard the same passages
in a language lab and answered the same written multiple-choice compre-
hension questions in English immediately after hearing each passage.
There were a total of 24 comprehension questions. Comprehension was
measured as the percent of correct responses out of the total of 24
questions.
Results of Study 1
A regression approach to analysis of covariance (ANCOVA) was used
with English institute entrance exam scores as the covariate for the Polish
sample and English as a Second Language Achievement Test (ESLAT)
scores as the covariate for the Puerto Rican sample. ANCOVA was chosen
in order to control for varying levels of proficiency among subjects while
testing for the effect of speed and syntax on listening comprehension. By
using ANCOVA to eliminate the variation due to level of proficiency, the
effects of speed and sentence structure can be seen more clearly. ABSTAT
(IBM Personal Computer) was used to analyze the data for the Polish
sample, and Statview (Macintosh Apple Computer) was used to analyze
the data for the Puerto Rican sample. Mean scores on the comprehension

questions for the two samples are presented in Table 1. In answer to the
research question on speed and syntax, there appear to be no differences
across treatments except that within the Puerto Rican sample, the group
that heard Version l/slow had higher comprehension scores than the other
groups.
TABLE 1
Group Means of Comprehension Scores (in %):
Study 1
Although the overall regressions were significant (F [4,67] = 6.34,

p <.0002 for the Polish sample and F [4,95] = 8.275, p <.0001 for the
Puerto Rican sample), there was no significant effect for speed
(F [1,69] = 0.673, p >.05 for the Polish sample and F [1,97] = 0.022,
p >.05 for the Puerto Rican sample) or for sentence structure
(F [2,68] = 1.071, p >.05 for the Polish sample and F [2,96] = 1.839,
p >.05 for the Puerto Rican sample). The only significant variable in
each case was the covariate which measures level of proficiency
(F [1,67] = 23.898, p <.0001 for the Polish sample and (F [1,95] = 24.275,
p <.0001 for the Puerto Rican sample).
Reduced velocity, although usually yielding slightly higher comprehen-
sion scores, was not a significant aid in these studies. Sentence structure,
which made more of a difference in the reading comprehension study
(Blau, 1982) conducted with a similar sample of Puerto Rican students,
seems to be a less salient modification when the input is aural rather than
written.
STUDY 2
Because of the insignificant effect on the comprehension of mechani-
cally slowed aural input and sentence structure modifications obtained in
Study 1, it seemed logical to investigate an additional input modification-
pausing. Chamot (1977), Crow (1984), Flaherty (1979), Hatch (1983),
Rivers (1980), and Stevick (1976) support the notion of allowing additional
748 TESOL QUARTERLY

processing time without disturbing the natural features of the normal flow
of speech through the use of pauses. But the actual effect of this modifica-
tion to aural input has not, to my knowledge, been empirically tested.
The second study compared the effect of speed and pauses. Because the
passages used in Study 1 were deemed too brief to measure the effect of
pauses, three longer monologues from Listening in and Speaking Out
(Bode, Whitley, & James, 1981) were recorded three times: at normal
speed (which this time turned out to be approximately 200 wpm), slowed
on the VSC Soundpacer to approximately 185 wpm, and with 3-second
pauses inserted at selected sentence, clause, and phrase boundaries which,
in effect, slowed the rate of delivery to approximately 150 wpm. Pauses
were inserted, on the average, every 23 words.
Samples of 36 Polish and 70 Puerto Rican students were drawn from the
same two populations as in the previous study. Three existing classes
within each sample were randomly assigned to each of the three treat-
ments. All students heard the three texts in a language lab. Comprehension
of each monologue was measured by wh- questions immediately after
hearing the monologue. The questions for the Puerto Rican sample were in
Spanish and were more specific and therefore more numerous than the
questions for the Polish sample. However, in both cases the questions
allowed for display of as much recall as possible. Subjects were also asked
to indicate what percentage of the monologues they thought they
understood.
Results and Discussion of Study 2

Data were analyzed using a regression approach to analysis of
covariance with the same covariates as in Study 1. Mean scores on the com-
prehension questions are presented in Table 2, which shows higher com-
prehension on the version with pauses. The overall regression for the
TABLE 2
Group Means of Comprehension Scores (in %):
Study 2
Slow Pauses Normal

Polish Sample 55 71 59
Puerto Rican Sample 55 69 58
Polish sample, analyzed using ABSTAT (IBM-PC), was significant

(F [3,32] = 3.12, p < .05), and the version with pauses did in fact yield
significantly higher comprehension than both other versions
(t [32] = 1.776, p < .05). The overall regression for the Puerto Rican

sample, analyzed using Statview (Macintosh Apple Computer), was also
significant (F [3,66] = 14.347, p < .0001). As with the Polish sample,
pauses yielded significantly higher comprehension than the other versions
(t [66] = 1.966, p < .05). The covariate was significant for the Puerto Rican
sample (F [1,66] = 36.192, p < .0001) but was not significant for the Polish
sample.
The significantly higher comprehension scores for the version with
pauses parallels the students’ assessment of the percentage they thought
they understood (See Table 3). Though the differences are more dramatic
for the Polish sample, the results indicate that the version with pauses is
consistently the most comprehensible, whereas the mechanically slowed
version is the least comprehensible.
TABLE 3
Self-Assessment (in %):
Study 2
Slow Pauses Normal

Polish Sample 43 84 65
Puerto Rican Sample 54 62 58
The problem of the lack of a standard “normal rate” has already been
noted. In this study the normal rate was actually relatively rapid (200
wpm). In spite of this, the effect of mechanically reducing velocity yielded
a negative effect both in comprehension and self-assessment of quantity
understood, making the case against mechanical reduction of speed even
stronger.
The graphs in Figures 1 and 2 allow viewing of the results across
proficiency levels. Unfortunately, the measures of proficiency for each
sample (on the x-axis) are different, making it impossible to compare
formally the proficiency levels of the two groups. Furthermore, the Polish
students studied are specializing in English and are among the best English
students in a country where English is a foreign language. The Puerto
Rican students, on the other hand, are not English majors and are enrolled
in an ESL course as a requirement. The role of English in Puerto Rico, a
U.S. commonwealth where students take ESL throughout their schooling,
is quite different from its role in Poland.
Despite the different settings for the two samples used in this study, the
different measures of proficiency used, and the different systems for
measuring comprehension, some interesting preliminary findings may be
noted. Comprehension of the version with pauses is generally higher than
that of the normal version. For the Puerto Rican sample the benefit of
pauses increases as proficiency increases within the range of this sample
while the contrary seems to be the case for the Polish sample. It is likely
750 TESOL QUARTERLY

FIGURE 1
Polish Sample
FIGURE 2
Puerto Rican Sample

that a certain threshold of proficiency is necessary to be able to take
advantage of the extra processing time provided by the pauses. If the
higher-level Polish students are more advanced than the higher-level
Puerto Rican students, this would indicate that beyond a certain level, one
can comprehend normal input better; a NNS at this level no longer needs
the pauses and might even find them bothersome.
At the lowest levels of proficiency the effect of mechanical slowing is
positive for the Puerto Rican sample, whereas for the Polish sample,
reduced velocity yields slightly higher results than the normal version but
lower than the version with pauses. For both samples, the higher the
proficiency level, the more negative the effect of slowing.
At most levels of proficiency, then, pauses seem to aid auditory compre-
hension more than either mechanical slowing or “normal” rates of delivery.
At the lowest levels, however, slowing may be of some assistance, whereas
at the very highest levels neither slowing nor pausing modification is
necessary. These findings suggest that, as Scarcella and Perkins (1987)
have noted, input features might have different effects at different stages
of SLA.
PEDAGOGICAL IMPLICATIONS
The results of these studies offer several suggestions for language
teaching professionals as providers of comprehensible aural input, either
prerecorded or live. First, we should not be overly concerned with
sentence structure. Second, we should not be overly concerned with speed
of speech. Even using a relatively rapid rate of delivery as the norm,
mechanical slowing of speech did not enhance comprehension except at
the lowest levels of L2 proficiency. Flaherty (1979) states that “too slow an
input rate can impair comprehension by prolonging the time a pattern
must be held in the short-term memory and allowing more time for
memory traces to fade” (p. 275). Pauses at constituent boundaries, on the
other hand, do enhance the comprehensibility of aural input significantly.
The use of pauses is a modification to input that is easy to implement
and has the potential to significantly help NNSs receive the CI they need
to successfully acquire a second language. This finding thus contributes to
the gradually emerging picture of what constitutes comprehensible aural
input.
REFERENCES
Blau, E. K. (1982). The effect of syntax on readability for ESL students in Puerto
Rico. TESOL Quarterly, 16 (3), 517-528.
Bode, S., Whitley, C., & James, G. (1981). Listening in and speaking out. New
York: Longman.
Bohannon, J. N., & Warren-Leubecker, A. (1985). Theoretical approaches to lan-
guage acquisition. In J. Berko Gleason (Ed.), The development of language.
Columbus, OH: Charles E. Merrill.
752 TESOL QUARTERLY

Chamot, A. U. (1977). Listening skills in the elementary ESL class. In H. D. Brown,
C. A. Yorio, & R. H. Crymes (Eds.), On TESOL ’77 (pp. 74-80). Washington,
DC: TESOL.
Crow, J. T. (1984). Psycholinguistics and listening comprehension for the ESL
classroom, Cross Currents, XI, 29-36.
Dwyer, E. C. (1987). The effects of speech rates on the listening comprehension of
fifth-grade students. Unpublished master’s thesis, University of Puerto Rico,
Mayaguez.
Flaherty, Sister E. (1979). Rate-controlled speech in foreign language education.
Foreign Language Annals, 12, 275-280.
Hatch, E. M. (1983). Psycholinguistics: A second language perspective. Rowley,
MA: Newbury House.
Kelch, K. (1985). Modified input as an aid to comprehension. Studies in Second
Language Acquisition, 7, 81-90.
Long, M. H. (1983). Linguistic and conversational adjustments to non-native
speakers. Studies in Second Language Acquisition, 5, 177-193.
Long, M. H. (1985). Input and second language acquisition theory. In S. Gass &
C. Madden (Eds.), Input in second language acquisition (pp. 377-393). Rowley,
MA: Newbury House.
Murphy, J. M. (1965). Examining ESL listening as an interpretive process. TESOL
Newsletter, XIX, 6, 23-24.
Rivers, W. (1980). Listening comprehension. In Kenneth Croft (Ed.), Readings on
English as a second language (2nd ed., pp. 265-277). Cambridge, MA: Winthrop.
Scarcella, R., & Perkins, L. (1987). Shifting gears: Krashen’s input hypothesis.
Studies in Second Language Acquisition, 9, 347-353.
Smith, P. D., Jr. (1980, Spring/Summer). A study of the effect of “slowed speech"
on listening comprehension of French. NALLD Journal, 9-13.
Stevick, E. W. (1976). Memory meaning and method. Rowley, MA: Newbury
House.
Author’s Address: Department of English, University of Puerto Rico, Mayaguez,

PR 00709
Evaluations of Essay Prompts by

Nonnative Speakers of English
MALCOLM HAYWARD
Indiana University of Pennsylvania
For a nonnative speaker of English, full competency in a U.S. college

or university includes the ability to take essay tests well. Such tests often
require students to select one or several prompts. How do ESL students
choose questions? Do they focus on the same features as native speakers?
This study, part of a more extended analysis of affective responses to essay
questions, finds that ESL students in an English I composition class are
influenced by the length of a prompt and to a degree by its readability.
Also, more important than the grade in choosing the prompt are such

factors as interest, ability to find much to say about the topic, and ease of
writing the essay.
Answering an essay prompt requires skills in vocabulary, syntax,
organization, and rhetoric, as well as a grasp of the material under
discussion. Most studies of essay tests concern evaluation. The few studies
of the effects of varying a prompt’s phrasing find that wording may affect
written responses. While Brossell (1983) and Brossell and Ash (1984) find
that small variations in the prompt’s structure did not result in significant
differences in the essays, Smith et al. (1985) report that larger variations
resulted in significant variations in essay quality. Kegley (1986) and
Kindilien (1986) have measured the impact of rhetorical mode in responses
to essay prompts. Related research concerns students’ affective responses
to essay prompts. In evaluating and selecting essay prompts, native Eng-
lish-speaking beginning writers are sensitive to such features of the prompt
as word length, readability level, and rhetorical structure (Hayward, 1989).
This study continues that latter line of research.
METHOD
Subjects
Subjects were 46 native speakers in two sections of English 101 and 27
nonnative speakers in three ESL-only sections of English 101 at Indiana
University of Pennsylvania, a medium-size state university. The
nationalities of the ESL students were: Bangladesh/Pakistan (5), Malaysia
(4), Hong Kong (3), Japan (2), India (2), Cyprus (2), Turkey (l), Costa
Rica (l), France (l), Mauritius (l), Nicaragua (l), Paraguay (l),
Philippines (l), Senegal (l), and Somalia (l).
Materials and Design

A survey asked first for demographic information and a self-rating of
success as a writer and essay test taker. On a 5-point Likert scale, ranging
from strongly agree (1) to strongly disagree (5), the students rated three
items:
1. I am a pretty good essay test taker.
2. My writing skills are fairly strong.
3. My grades are/were helped by essay tests.
Each student was then presented with one of two questionnaires
containing 15 essay prompts drawn from past departmental exams
(yielding 30 prompts in the total sample). Prompt length ranged from 29 to
121 words with an average of 53 words, and readability ranged, on the
Flesch Index, from a very easy 90 to a difficult 45 with a mean of 63.7
(standard on the Flesch Index). Each prompt was followed by eight
statements:
754 TESOL QUARTERLY

1. The essay will be easy to write.
2. The essay will be interesting to write.
3. I might choose this one if given a choice.
4. It is easy to see what the questioner is looking for.
5. It would be easy to organize the essay.
6. I would probably express a lot of my own, personal ideas here.
7. I would probably find a lot to say in this essay.
8. I might get a good grade on this essay.
After reading a prompt, students rated each statement on a 5-point Likert
scale, circling responses ranging from strongly agree (1) to strongly
disagree (5).
RESULTS AND DISCUSSION

Self-ratings were first analyzed and found to be strongly correlated
(measured by the Pearson product-moment correlation coefficients)
among the three items. The correlations held for native and ESL groups
(see Table 1, p <.01 in all cases). Native speakers ranked their abilities
significantly higher, as measured by a t test, than did ESL students on the
TABLE 1
Pearson Product-Moment Correlation Coefficients for Self-Ratings
Q1 Q2 Q3
Native Speakers (n= 46)
1 (Good test taker) 1.0000 .5245 .6265
2 (Strong writing skills) .5245 1.0000 .4079
3 (Essay tests help grades) .6285 .4079 1.000
ESL Students (n = 27)

1 (Good test taker) 1.0000 .7661 .5758
2 (Strong writing skills) .7661 1.0000 .4913
3 (Essay tests help grades) .5758 .4913 1.0000
items “good essay test taker” (p < .00l) and “writing skills are fairly
strong” (p < .01); no significant differences were found on the self-rated
item, “grades . . . helped by essay tests.” The ESL group also gave
consistently and significantly lower ratings (p < .05) to all of the eight
statements concerning the prompts but one, Item 3, “might choose this
one,” as measured by t tests. (Item 1, p < .001; Item 2, p < .05; Item 4,
p < .001; Item 5, p < .05; Item 6, p < .001; Item 7, p < .001; Item 8,

p <.001. An overall test was performed to check the validity of these
multiple t tests; Hotelling’s T2 test was significant, p < .05.) If these ratings
indicate the student’s sense of ability to handle an essay task, then ESL
students are far less sure of themselves than native speakers.
Next, relations among the eight statements were considered. Results
showed the expected high correlations among the eight items. To
determine why an essay might be chosen, the relations between Item 3, “I
might choose this one, ” and the other seven items were analyzed. For
native speakers, the highest correlations were with Item 2, “interesting to
write,” Item 8, “might get a good grade,” and Item 7, “find a lot to say.”
Like their native-speaking counterparts, ESL students ranked most highly
the relation between “the essay will be interesting to write” (Item 2) and “I
might choose this one” (Item 3). “Find a lot to say” (Item 7) also correlated
well with “get a good grade’’(Item 8)—but the issue of a good grade
proved less significant for ESL students than ease of writing (Item 1). For
both groups, Item 6, expressing “personal ideas,” was the item least well
correlated with the other 7 items. Finally, for ESL students there was a
strong relation between Item 1, “easy to write,” and Items 4 and 5, “easy
to see what the questioner is looking for” and “easy to organize the essay.”
The correlations for both groups are displayed in Table 2.
TABLE 2
Pearson Product-Moment Corrrelation Coefficients for Prompt Ratings
Item No.
Item No. 1 2 3 4 5 6 7 8
756 TESOL QUARTERLY

The next issue addressed was readability, or ease of text comprehension,
here measured by the Flesch Index. For native speakers, the Pearson
correlation coefficients between prompt readability and average
responses to each of the eight items found significant correlations (p <.05,
one-tailed) between readability and the items “interesting to write” and
“find a lot to say.” The others were nearly, but not quite, significant: “get
a good grade” (p = .056), “easy to write” (p = .077), and “might choose”
(p= .072).
If native speakers respond to levels of readability, ESL students might
be expected to show even greater sensitivity; yet that was not the case.
Although the correlation approached significance on several items for ESL
students, no single item reached the .05 level of significance. The closest
items were “interesting to write” (p = .087), “might choose” (p = .081),
and “easy to see what the questioner is looking for” (p = .068). Thus
readability was not as strong a factor for ESL students as for native
speakers.
While readability was less important than expected for ESL students,
prompt length was another matter. The length of the prompt proved quite
important. For native English speakers the prompt length was correlated
(p< .05) with positive ratings for expressing “personal ideas,” finding a
“lot to say,” and getting a “good grade” on a prompt. For ESL students,
significant, positive correlations were found between the length of the
prompt and ratings for “easy to see what the questioner is looking for,”
“easy to organize,” and getting a “good grade.” The correlations for all
other factors with the exception of finding a “lot to say” were not
significant at level .05, but had p values <.10. ESL students indicate a
strong preference for longer essay prompts, and their interest in prompt
length seems related to understanding the question–knowing what the
questioner is asking and how to organize the answer.
CONCLUSION
Despite certain limitations, the fact that the comparisons to native Eng-
lish speakers revealed a number of similarities in responses suggests the
survey does provide a useful measure of students’ reactions. The most
significant limitation is that the survey does not consider the knowledge of
the texts upon which the prompts are based. Given the importance that
“interest” plays in guiding the choice of a prompt, the way the prompt
addresses both real world knowledge and knowledge of the text itself may
determine choice more than wording or length.
In conclusion, ESL writers regard interest in the prompt as important,
though with prompts of equal interest, the choice will be the one that the
student believes offers the greatest scope for doing a lot of writing.
Longer, more readable prompts may give ESL students a better sense of
what the question is asking; the length of the prompt also suggests an
organization and builds student confidence. Essay questions should be
built with these constraints in mind. The question of what makes a

question “interesting,” however, is an important one and remains to be
explored.
ACKNOWLEDGMENT
A portion of this material was delivered in a presentation at the Conference on
College Composition and Communication in Chicago, March 1990.
REFERENCES
Brossell, G. (1983). Rhetorical specification in essay examination topics.College
English, 45, 165-173.
Brossell, G., & Ash, B. H. (1984). An experiment with the wording of essay topics.
College Composition and Communication, 35, 423-425.
Hayward, M. (1989). Choosing an essay test question: It’s more than what you
know. Teaching English in the Two-Year College, 16, 174-178.
Kegley, P. H. (1986). The effect of mode discourse on student writing
performance: Implications for policy. Educational Evaluation and Policy
Analysis, 8, 147-154.
Kindilien, G. A. (1986). Re-examining essay questions. Teaching English in the
Two-Year College, 13, 102-103.
Smith, W. L., Hull, G. A., Land, R. E., Jr., Moore, M. T., Ball, C., Dunham, D. D.,
Hickey, L. S., & Ruzich, C. W. (1985). Some effects of varying the structure of
a topic on college students’ writing. Written Communication, 2, 73-89.
Author’s Address: Department of English, Indiana University of Pennsylvania, 110

Leonard Hall, Indiana, PA 15705-1094
758 TESOL QUARTERLY

THE FORUM
The TESOL Quarterly invites commentary on current trends or practices in the
TESOL profession. It also welcomes responses or rebuttals to any articles or
remarks published here in The Forum or elsewhere in the Quarterly.
Two Commentaries on Brian K. Lynch's

“A Context-Adaptive Model of Program
Evaluation”
A Reader Reacts. . .
MARK W. MEINKE
The University of Illinois at Chicago
In his recent article in the TESOL Quarterly (Vol. 24, No. 1),
Brian Lynch correctly identifies the need for “a generalized
evaluation model that addresses the full range of concerns of
language teaching programs.” Fortunately, this wheel has already
been invented.
Program evaluation is neither an arcane nor an obscure activity,
nor is it a highly specialized concern of any one field. Rather, it is a
practical effort, grounded in a variety of theories, to measure
whether a business campaign, a training program, or an ESL course
has gotten where it was supposed to go— and whether getting there
was worth the effort in the first place.
Business people and Human Resource Development (HRD)
trainers and developers have been planning and carrying out
evaluations for decades. Educators of all stripes have been doing
the same. Brian Lynch’s efforts, and TESOL program evaluations,
need not occur in isolation from all these other efforts.
Indeed large areas of Lynch’s model are reminiscent of planning
and implementation strategies developed by the Management by
Objectives (MOB) school of management. Line managers and
supervisors would recognize the goal identification and resource
analysis stages implicit in Lynch’s first three levels. The “one size fits
all” model, taken as a whole, offers a close parallel to the work of
HRD specialists and leading trainers and developers such as Dugan
Laird (1978) and Odiorne and Rummler (1988).
759
A review of Brinkerhoff’s 1987 classic, Achieving Results from
Training, shows how much effort Lynch might have saved had the
field of TESOL paid as much attention to collateral developments
in other fields as it has to linguistics, psychology, and philosophy.
Since Kirkpatrick first broached his hierarchy of evaluation stages
(1989), trainers in the U.S. and elsewhere have been examining and
refining their evaluative methods and instruments. They have dealt
with issues of qualitative versus quantitative measures, with
measuring how well training programs measure up to internal
standards rather than to organizational benefits, and with the whole
host of issues confronting TESOL program evaluators.
Additionally, business discipline has long since concentrated
trainers’ minds on issues of ultimate benefit and of utility (for which
read “cost benefit”). Of necessity, HRD professionals and trainers
have invented means of measuring learning and learning benefit in
dollar terms. While not all apply to TESOL or should even be
adapted to this field, HRD training and development models of
evaluation are certainly germane.
It is long past time that TESOL professionals begin to see
themselves within the “wider context of training and development
and begin applying useful disciplines of the business world to their
activities.
Program evaluation is not virgin territory; it is a well-travelled
and well-developed land encompassing our field and all the other
varieties of human training and development. We can save
ourselves effort and widen our pool of knowledge by drawing upon
the advances others have already made.
REFERENCES
Brinkerhoff, R. O. (1987). Achieving results from training: How to
evaluate human resource development to strengthen programs and
increase impact. San Francisco: Jossey-Bass.
Kirkpatrick, D. L. (1989, September-October). Evaluating a training
program for supervisors and foremen. The Personnel Administrator,
14 (5), 29-38.
Laird, D. (1978). Approaches to training and development, Reading, MA:
Addison-Wesley.
Odiorne, G. S., & Rummler, G. A. (1988). Training and development: A
guide for professionals. Chicago: Commerce Clearing House.
760 TESOL QUARTERLY

A Context-Restrictive Model for Program Evaluation?
JOHN M. SWALES
The University of Michigan
Brian Lynch offers a clearly articulated “context-adaptive” model

for language program evaluation, which he well illustrates through
the case of the Reading English for Science and Technology
(REST) Project at the University of Guadalajara, Mexico. Lynch
concludes his paper with the following observations
This iterative framework leads program evaluators through a set of
considerations that can adapt the evaluation to a variety of specific
program settings. The use of this model in future evaluations will lead to
further refinement and increased usefulness for ESL and other language-
teaching program contexts. (p. 39)
The reader will note that the confidence level underlying these
concluding remarks is quite high; in particular, Lynch claims that his
model, while needing further refinement, is sufficiently developed
to be ready for wider use. While it is clear that Lynch deserves
considerable credit for his attempt to construct an adaptable
evaluation instrument—as antidote to the craft of evaluating
language programs on a case-by-case basis—the model (at least as
presented and exemplified in Lynch's article) seems to be oddly,
and perhaps dangerously, self-limiting. I will argue, therefore, that
further refinement of the model would be advisable before further
use.
As a matter of practice, ESL program evaluation is a well-
established ongoing activity in our profession, most typically
orchestrated through curriculum committees, retreats, or self-study
task forces. While formal external evaluations of ESL programs
may be less common than with many other kinds of academic units
on U.S. campuses, they certainly occur. Elsewhere, organizations
like the British Council have well-established policies for periodic
external evaluation of ESL projects and programs. The perceived
rationale for all this expensive activity is that it provides objective,
reliable and expert judgments on particular programs as well as
offering reasoned recommendations for possible changes in
objectives, activities, and/or personnel. It is of course true that this
rationale is idealistic both in terms of its assumption of reduced bias
(Beretta, 1986) and in terms of its expectation that cogent
recommendations will actually be implemented. On the other hand,
it is equally the case that evaluators are typically chosen because
they are supposed to bring into the evaluation a sufficiently broad
THE FORUM 761

base of knowledge and experience to be able to access a particular
program in the light of comparable program settings elsewhere.
Given this institutional view of external evaluation, Lynch’s
model as exemplified through the REST case study seems curiously
internal; the project is examined entirely in terms of itself without
any insights from comparable programs in comparable settings.
The imposed limitation is particularly worrisome in the REST
context because of the wealth of actual comparable information
available. After all, for many years Latin American universities have
been operating REST-type programs, and the considerable
scholarly literature on such programs goes back as far as Ewer and
Latorre (1967). Important work in Mexico itself was carried out by
the Research and Development Unit at the National Autonomous
University of Mexico in the mid-1970s (see, for example, Mackay,
1978). Several of the issues which Lynch identifies as being relevant
to the thematic framework (p. 37) have already been discussed in
Latin American REST-type contexts, for example, team teaching by
de Escorcia (1984). Brazil has for a number of years produced a
journal, The ESPecialist, precisely targeted on unifying national
efforts to upgrade REST-type operations. And so on.
A less context-restrictive approach to evaluation would build such
relevant external data into the model. This would be advantageous
for at least three reasons. First, it would strengthen the chances of
program-fair evaluation (Beretta, 1986) by providing a knowledge-
base for comparing like with like. Second, it would establish a
platform whereby the audiences of the evaluation report could
perceive the comparative strengths and weaknesses of their own
program; and third, it would open the program to outside work so
that economies associated with not reinventing the wheel can be
(albeit belatedly) put into place.
Other restrictive aspects of the Lynch model can be dealt with
more expeditiously. Steps 2 and 3 of the model do not seem to easily
permit the emergence of possible alternative uses of instructional
resources. We might envision, for example, that the REST project
would work better with, say, third-year students rather than first-
year students. Swales (1989) offers the concept of opportunity cost
as a way of structuring more radical programming decisions: In
effect, opportunity cost asks ESL programs to reflect on what they
cannot “afford” not to do through to what they cannot “afford” to
do. Whatever the mechanism, a context-adaptive model needs to
allow for substitution by replacement.
Finally, there is the issue of proactive effects of the evaluation
process itself. Lynch in fact is obviously aware of the importance of
assessing the audience of the final report and in so doing of assessing
762 TESOL QUARTERLY

important social and political sensitivities. However, he only views
these considerations as constraints on the evaluator. Evaluations can
also operate as important opportunities for educating authorities
and for influencing the views of others (students, subject
instructors, etc.) via discussion, data-collection and other processes.
Just as needs analysis benefits from recognizing changing
institutional dynamics (Coleman, 1988), so does evaluation.
Lynch is to be congratulated on his bold attempt to construct a
generalizable model for guiding and rationalizing the case-by-case
character of ESL program evaluations. Yet Yin (1984) notes that
case studies are particularly valuable “when the boundaries
between phenomenon and context are not clearly evident” (p. 23).
My sense of Lynch's paper is that he has so far set the boundaries
between phenomena and contexts too narrowly and that the model
evaluation of the REST program at Guadalajara suffers somewhat
from this constriction.
REFERENCES
Beretta, A. (1986). Program-fair language teaching evaluation. TESOL
Quarterly, 20 (3), 431-445.
Coleman, H. (1988). Analysing language needs in large organizations.
English for Specific Purposes, 7 (3), 155-169.
de Escorcia, B. A. (1984). Team-teaching for students of economics: A
Colombian experience. In R. Williams, J. Swales, & J. Kirkman (Eds.),
Common ground: Shared interests in ESP and communication studies
(pp. 135-143). Oxford: Pergamon.
Ewer, J. R., & Latorre, G. (1987). Preparing an English course for students
of science. English Language Teaching, 21 (3), 221-229.
Lynch, B. K. (1990). A context-adaptive model for program evaluation.
TESOL Quarterly, 24 (l), 23-42.
Mackay, R. (1978, Summer), A reading course for Mexican veterinarians:
One approach to the total ESP operation. MALS Journal, 85-98.
Swales, J. (1989). Service English programme design and opportunity cost.
In R. K. Johnson (Ed.), The second language curriculum (pp. 79-90).
Cambridge: Cambridge University Press.
Yin, R. E. (1984). Case study research: Design and methods. Beverly Hills,
CA: Sage.
THE FORUM 763

The Author Responds . .
BRIAN K. LYNCH
Mark Meinke and John Swales offer some thoughtful criticisms of

my article to which I would like to respond. Both colleagues appear
to be concerned that I have reinvented the wheel. Meinke, in
particular, suggests that aspects of the context-adaptive model
presented in my article are “reminiscent” and “a close parallel” to
evaluation and training models from the disciplines of business and
human resource development. The complaint is made that I could
have saved myself a great deal of effort had I paid more attention
to developments in these disciplines. However, I believe that had I
started with models from the world of business and management
and worked towards the context of language teaching programs,
the result would have been a much less relevant model for TESOL
and applied linguistics. I would offer an argument parallel to the
one that Fanselow (1988) has made against starting with published
coding systems for classroom observation:
It is... more likely that somebody else’s categories can be made our
own [italics added] if we develop some of our own along the way. When
those doing the observing develop the categories, knowledge is being
created, rather than buckets being filled (Montessori, 1967) or
information being transferred (Freire, 1970, p. 67). (p. 121)
The fact that some of the “categories” or evaluation steps that I
arrived at bear resemblance to ones already arrived at in another
discipline does not invalidate my efforts. I was not so much
reinventing the wheel as “making it my own” and more relevant to
TESOL and applied linguistics.
A related criticism made by Meinke is that evaluation efforts by
me and other TESOL evaluators have occurred in isolation from
“educators of all stripes. ” An examination of work cited in my
article (Alkin & Wooley, 1970; Long, 1984; Beretta, 1986; Lynch,
1988; Brown, 1989) will show that TESOL and applied linguistics
has always grounded its evaluative work in developments from the
field of education. Perhaps the fact that my article focused the
discussion on developments specific to language teaching programs
gave the false impression that our work has or should have been
carried out in isolation. That was not my intention. However, I do
disagree somewhat with Meinke’s statement that program evalua-
tion is not “a highly specialized concern of any one field.” While
program evaluation obviously occurs in many disciplines, there is a
764 TESOL QUARTERLY

need to tailor it to the specific concerns of specific disciplines. The
context-adaptive model attempts to do this for TESOL and applied
linguistics.
John Swales’ concern that I may be fostering the reinvention of
the wheel stems from his view of the context-adaptive model as
being overly “internal” and “restrictive” and, therefore, incapable of
taking advantage of previously accumulated knowledge concerning
language teaching programs. Since the adaptive nature of the model
was meant to ensure flexibility, this criticism was, at first, extremely
puzzling to me. I hope that by responding to his specific criticisms
I can show that this view is perhaps the result of a narrow reading
of the model.
First of all, Swales contrasts “external evaluations,” where expert
evaluators are brought in from outside the program in order to offer
“knowledge and experience . . . in the light of comparable program
settings elsewhere.” The claim is made that the context-adaptive
model excludes such evaluators and such information, rendering it
“curiously internal.” One of the dimensions in Step 2 of the model,
the Context Inventory, is evaluation expertise. Only an extremely
narrow reading of the model would result in the impossibility of
external evaluators providing that expertise. While the model
cannot ensure that all evaluators will do an equally thorough job, it
assumes that evaluation experts will bring their knowledge and
experience (including that of other, comparable programs) to bear
during the analysis and report stages.
Swales follows this concern for the inclusion of relevant external
data with the assertion that the REST project context, with which
the model was exemplified, had somehow ignored “operating
REST-type programs” in Latin America. In fact, the REST project
was aware of and had studied the reports from the Centro de
Ensefianza de Lenguas Extranjeras at the Universidad National
Autonoma de Mexico and had corresponded with the ESP projects
in Brazil that Swales references. This comparative information was
used in the program design phase of the REST project. The issues
that developed during the program evaluation of the REST project
were compared with the experiences of these projects as well. The
fact that issues such as team-teaching had already been discussed in
other Latin American contexts, however, did not usually result in a
blueprint for REST project design or evaluation. The importance of
such issues for the purposes of evaluation had to do with their
particular expression within the REST context.
Another criticism that Swales raises is that Steps 2 and 3 of the
model do not permit “the emergence of possible alternative uses of
instructional resources.” In these first two steps, the purpose is to
THE FORUM 765

focus and frame the evaluation. There is no reason to believe that
alternatives to existing program practices will not be proposed
formally in the evaluation report or informally during data
collection and analysis. Swales suggests further that “we might
envision, for example, that the REST project would work better
with, say, third-year students rather than first-year students.” The
issue of whether first- and second-year chemical engineering
students would benefit more from the REST project than third- and
fourth-year students was, in fact, discussed both during the design
of the program as well as during the evaluation.
Finally, Swales points out that the discussion of the final step of
the model, the evaluation report, casts the notion of assessing the
evaluation audience and the associated political and social
sensitivities as essentially a constraint on the evaluator. He quite
correctly points out the potential for evaluation efforts to educate
and influence educational outcomes. This is a flaw in the discussion
of the model, though not necessarily in the model itself.
Swales begins his reaction to my article by advising that the
context-adaptive model, being “oddly, and perhaps dangerously,
self -limiting,” is not ready for further use. As I have tried to
demonstrate, his concerns result, perhaps, from too narrow a
reading of the model. There has already been one successful use of
the model in a vocational English as a second language context
(Franklin, 1991). Swales ends his reaction to my article with the
observation that I may have “set the boundaries between phenom-
ena and contexts too narrowly.” This underscores an important
tension in evaluation practice and research—the ability to general-
ize across program settings versus the ability to clearly and accur-
ately characterize and evaluate specific program settings. This is an
issue of no small import and I am grateful to Swales for raising it, as
I am grateful for his thoughtful reading and reaction to my article.
I hope that the discussions initiated by Meinke and Swales will
continue in an effort to clarify the role of evaluation in TESOL and
applied linguistics.
REFERENCES
Alkin, M. C., & Wooley, D. C. (1970). A framework for the evaluation of
TESOL programs (Working Paper No. 4). Los Angeles: University of
California, Center for the Study of Evaluation.
766 TESOL QUARTERLY

Beretta, A. (1986). Toward a methodology of ESL program evaluation.
TESOL Quarterly, 20 (1), 144-155.
Brown, J. D. (1989). Language program evaluation: A synthesis of existing
possibilities. In R. K. Johnson (Ed.), The second language curriculum
(pp. 222-241). Cambridge: Cambridge university press.
Fanselow, J. F (1988). “Let’s see”: Contrasting conversations about
teaching. TESOL Quarterly, 16 (3), 113-130.
Franklin, L. (1991). Using the context-adaptive model for program
evaluation for a vocational English as a second language course.
Unpublished master’s thesis, University of California, Los Angeles.
Long, M. H. (1984). Process and product in ESL program evaluation.
TESOl Quarterly, 18 (3), 409-425.
Lynch, B. K. (1988). Toward a context-adaptive model for the evaluation
of language teaching programs (Doctoral dissertation, University of
California, Los Angeles, 1987). Dissertation Abstracts International, 48,
2264A.
A Logical Difficulty of the

Parameter Setting Model
YOSHINORI SASAKI
University of Illinois at Urbana-Champaign
This paper seeks to prove that the parameter setting model (PSM)
of Chomsky’s (1981) Universal Grammar (UG) theory contains a
crucial internal contradiction. The parameter setting model is a
subsidiary hypothesis of UG theory regarding the internal
mechanism of language acquisition. Proponents of the PSM make a
strong claim for the psychological reality of UG: Parameters are
more than mere notational markers for describing languages.
Indeed, configuration of the setting of parameters is supposed to
result in different syntactic rules of languages. According to Flynn
(1987),
a theory of- UG specifies that there are abstract and linguistically
significant principles that underlie all natural languages, and which also
define the “initial state” of the L1 learner’s mind. These principles of UG
determine the basic grammar of the language being acquired and
comprise the essential language faculty with which all individuals are in
general uniformly and equally endowed. . . . Experience is needed to
set the values of these principles—setting of the parameter for one value
sets you one language, and setting of the parameter in another way sets
you another language. (p. 28)
THE FORUM 767

As is clearly stated in this quotation, the PSM claims that
“parameters . . . have to be fixed by experience” (Chomsky, 1981,
p. 4) and experience is required to set the switches (Hermon, 1987).
Some parameter values are supposed to be more “marked” than
others, and this markedness is influential upon language acquisition:
The child approaches the task [language acquisition] equipped with
Universal Grammar and an associated theory of markedness that serves
two functions: It imposes preference structure [italics added] on the
parameter of Universal Grammar and it permits the extension of core
grammar to a marked periphery. (Chomsky, 1981, p. 8)
On account of this “preference structure” imposed on the
parameters, “in the absence of evidence to the contrary, unmarked
options are selected” (McLaughlin, 1987, p. 97).
COROLLARIES OF THE MODEL

Corollary 1
The PSM was proposed with the intention of explaining within
the UG framework the existence and acquisition of diverse
syntactic forms (marked/unmarked) in different languages in terms
of the setting of internal “switches.” Marked forms in Chomsky’s
sense1 are associated with marked parameter values, and unmarked
forms are associated with unmarked parameter values. It is taken
for granted by the proponents of the PSM that some languages
contain marked forms in this sense: Some of the switches within the
speakers of those languages are indeed in marked positions.
Corollary 2
As experience is considered to be necessary to change the setting
of switches, parameter resetting would seldom, if ever, take place
spontaneously without a speaker’s experience of being exposed to a
to-be-learned form. In particular, it is virtually impossible to replace
an unmarked value (i.e., a preferred value) of parameter with a
marked one (i. e., a less preferred one) without supporting
“evidence” or “input.” Human beings, who are equally trapped in
the innate “preference structure” of parameters, can override the
1 Asopposed to UG linguists’ static view of markedness, quite a few researchers outside of
the Chomskyan paradigm address this issue from a dynamic perspective. Syntactic
markedness in their sense is only indirectly affected by bioneurological constraints. For
example, Slobin (1979) claims that diachronic changes of languages are driven by
interactions/conflicts between four charges [be clear; be processible; be quick and easy; be
expressive]. Such a cognitive-functional view of markedness, which to my mind is quite
logical and cogent, is beyond the scope of the present critique.
768 TESOL QUARTERLY

unmarked parameter values to start producing marked forms only
when they are externally motivated to do so.
THE CONTRADICTION
Assume that a certain language contains a marked form
(Corollary 1); someone in the past must have initiated it. The
initiator of the marked form, by definition, had to produce it in the
first place (i.e., marked parameter values replaced unmarked
values) without an experience of being exposed to it. This, however,
should be an impossible event according to the PSM, which claims
that experience is required to reset the switches (Corollary 2).
Hence antinomy.
Borrowing of a marked form from a foreign language simply
pulls back the issue chronologically without providing any solution
to the contradiction, because it leaves the origin of marked forms
entirely unsolved. Nor can effects on syntax of such factors as
phonology, morphology, semantics, and pragmatics provide a
coherent solution within the Chomskyan framework because such
effects would imply crucial interactions between different linguis-
tic/cognitive domains, and thus upset Chomsky’s (1965) assumption
of modularity.
In short, the above-mentioned two corollaries of the PSM are
mutually incompatible, and thus the proof of its internal
contradiction is completed.
CONCLUSION
The PSM results in an obvious self-contradiction when it is
seriously taken to model the internal state of language learners.
Indeed, the parameter setting model assumes an impossible mission
to account for the existence of what it predicts should never exist
(i.e., marked syntactic forms in Chomsky’s sense). As a logical
consequence, this model is untenable as a scientific theory of lan-
guage acquisition. Furthermore, the model’s logical breakdown
casts doubt upon the psychological reality of Chomsky’s universal
grammar.
ACKNOWLEDGMENTS
I would like to express my gratitude for comments from Elizabeth Platt,
Gabriella Hermon, and an anonymous TESOL Quarterly reviewer in
response to earlier versions of this paper, Platt’s detailed suggestions
concerning the interpretation of UG theory were particularly beneficial. I
THE FORUM 769

also appreciate encouraging comments from Stephen Leary, Robert Yates,
and Yukiko Abe Hatasa, and stylistic advice from Hajime Takagi. I also
received useful information from Toshiyuki Doi. All remaining errors are
mine.
REFERENCES
Chomsky, N. (1965). Aspects of the theory of syntax. Cambridge, MA:
MIT Press.
Chomsky, N. (1981). Lectures on government and binding. Dordrecht,
The Netherlands: Foris.
Flynn, S. (1987). Contrast and constructions in a parameter-setting model
of L2 acquisition. Language Learning, 37 (1), 19-62.
Hermon, G. (1987, March). Government and binding theory: Implications
for L1 and L2 acquisition. Paper presented at the Second Language
Acquisition and Teacher Education Seminar, Champaign, IL.
Edward Arnold.
Slobin, D. (1979). Psycholinguistics. (2nd ed.). Glenview, IL: Scott,
Foresman.
Research Issues
The TESOL Quarterly publishes brief commentaries on aspects of qualitative and

quantitative research. For this issue, we asked two researchers to address the fol-
lowing question: Under what circumstances, if any, can the use of multiple t tests
be justified?
Edited by GRAHAM CROOKES

The Use of Multiple t Tests in Language Research

JAMES DEAN BROWN
The position taken by this author is based on the “conservative”

philosophy that, in applying statistics, great care and caution must always
be practiced in order to minimize the possibility of publishing “significant”
results that may have occurred by chance alone. In the view of this writer,
a serious problem has existed for years in language research with regard to
the use of the inferential statistic known as the t test. Interpretation of a
single t test comparing two means can be clear and unambiguous. Either
there is a significant difference between the two means in question (at the
770 TESOL QUARTERLY

prespecified alpha level), or there is not. However, interpretation of the
results of multiple t tests (i.e., simple t tests used to compare two or more
pairs of means) presents a variety of potential problems (see Brown, 1988,
p. 170; Dayton, 1970, pp. 37-49; Kirk, 1968, pp. 69-98; Shavelson, 1981,
pp. 447-448). The purpose of this essay is to describe briefly some of the
issues associated with multiple t tests and to address the question of when,
if ever, they can appropriately be used.
The central problem with multiple t tests is that there is a certain
probability that one or more significant differences will be found by
chance alone. As the number of comparisons increases in a study, there is
an increasingly high probability that one or more spuriously significant
differences will be found. The severity of this problem depends on
whether the means are independent (i.e., based on different groups of
subjects) or nonindependent (i.e., based on sets of observations of the
same group of subjects). For independent means, the probability of one or
more t tests being spuriously significant (i.e., the probability y of committing
probability of a Type I error for one comparison is 5%, for six comparisons
it is 26%, for ten comparisons it is 40%, for fifteen comparisons it is 54%, for
twenty comparisons it is 64%, etc. For nonindependent means, the
probabilities are even higher: for six comparisons the probability has been
estimated to be approximately 40%, for ten comparisons about 60%, for
twenty comparisons 90%, etc. (Cochran & Cox, 1957). Whether the means
are independent or nonindependent, the general problem is that any one
(or more) of the observed significant differences in the study may have
occurred by chance alone. Since it is not possible to determine which of
the significant differences might be spurious, interpretation of the results
becomes difficult, if not impossible.
Unfortunately, researchers in our field (at least as far back as 1962 when
Anisfeld, Bogo, & Lambert reported a total of 223 t tests, and as recently
as 1989 when Carrell, Pharis, & Liberto conducted 12 t tests) often use
multiple t tests and are apparently unaware of the problem that they are
creating. Typically, they discuss multiple t tests as though the results are
directly interpretable and clear-cut. Trusting the researcher, most readers
accept the analyses at face value, and potentially fallacious results enter the
permanent body of knowledge that makes up the language teaching field.
Fortunately, it is not necessary to live with this state of affairs. For years,
procedures have been available for making clearly interpretable multiple
comparisons among means. Within univariate statistics, analysis of
variance procedures can be followed by any number of possible
a posteriori multiple comparison tests; Duncan’s new multiple range test,
the Newman-Keuls test, Dunnett’s test, Tukey’s HSD test, and Scheffe’s
test (see Kirk, 1968, pp. 87-97; Jaccard, Becker, & Wood, 1984, for more on
these methods). Other a priori strategies (including Dunn’s multiple
comparison procedure and the Bonferroni t) allow the researcher to make
all planned comparisons within an experiment. There are also a variety of
THE FORUM 771

multivariate analyses that may prove useful if two or more dependent
variables are involved (Stevens, 1986).
A related and important question that has never been addressed in
language studies is the question of which conceptual unit is most
appropriate for the rates of error. The following approaches have been
suggested: per comparison, per hypothesis, per experiment and experi-
mentwise, as well as per family and familywise (see Kirk, 1968, pp. 82-85;
Ryan, 1959; Petrinovich & Hardyck, 1969; Wilson, 1962, for further dis-
cussion). To my knowledge, the only conceptual unit for rates of error
that has ever been considered when using multiple t tests in our field is the
per comparison rate. This is unfortunate because per hypothesis or larger
error rates could be used in interpreting the probabilities involved in the
use of multiple t tests.
Given all of the above, there are two situations wherein multiple t tests
would make sense to me. The first would be in making a priori orthogonal
comparisons. The second would be in making multiple mean comparisons
in a study wherein the Type I error rates are carefully controlled. Such a
study would necessarily provide clear discussion of the conceptual units
for rates of error (per hypothesis or larger) that were used and would
report all of the associated probabilities. (For other types of multiple
comparisons other than these two rather circumscribed uses for multiple
t tests, I would strongly advocate the procedures listed in the fourth
paragraph of this essay.) Any other indiscriminate use of multiple t tests
seems to me to be totally unjustified because it represents an unnecessary
and needlessly confusing use of inferential statistics.
THE AUTHOR
James Dean Brown is Associate Professor and Director of the English Language
Institute at the University of Hawaii at Manoa. He has published articles on
language testing and curriculum development and a book on critically reading
statistical studies, Understanding Research in Second Language Learning
(Cambridge University Press, 1988).
REFERENCES
Anisfeld, E., Bogo, N., & Lambert, W. E. (1962). Evaluational reactions to
accented English speech. Journal of Abnormal and Social Psychology, 65, 223-
231.
Brown, J. D. (1988). Understanding research in second language learning: A
teacher’s guide to statistics and research design. Cambridge: Cambridge
University Press.
Cochran, W. G., & Cox, G. M. (1957). Experimental designs. New York: John
Wiley.
Carrell, P. L., Pharis, B. C., & Liberto, J. C. (1989). Metacognitive strategy training
for ESL reading. TESOL Quarterly, 23, 647-678.
Dayton, C. M. (1970). The design of educational experiments. New York:
McGraw-Hill.
772 TESOL QUARTERLY

Jaccard, J., Becker, M. A., & Wood, G. (1984). Pairwise multiple comparison
procedures: A review. Psychological Bulletin, 96, 589-596.
Kirk, R. E. (1968). Experimental design: Procedures for the behavioral sciences.
Belmont, CA: Brooks/Cole.
Petrinovich, L. F., & Hardyck, C. D. (1969). Error rates for multiple comparison
methods: Some evidence concerning the frequency of erroneous conclusions.
Psychological Bulletin, 71, 43-54.
Ryan, T. A. (1959). Multiple comparisons in psychological research. Psychological
Bulletin, 56, 26-47.
Shavelson, R. J. (1981). Statistical reasoning for the behavioral sciences. Boston,
MA: Allyn & Bacon.
Stevens, J. (1986). Applied multivariate statistics for the social sciences. Hillsdale,
NJ: Lawrence Erlbaum.
Wilson, W. (1962). A note on the inconsistency inherent in the necessity to perform
multiple comparisons. Psychological Bulletin, 59, 296-300.
Multiple t Tests: Some Practical Considerations

ANDREW F. SIEGEL
The t test is used to decide whether the average scores of two groups are
significantly different or if the difference could be due merely to random
coincidence. Statistical t tests lose their effectiveness when too many are
used at once. However, this does not always create a problem. Since
interpretation of the results can be much more difficult when more
complex tests are used, there are some advantages to the use of multiple
t tests. Some statisticians even believe that the use of multiple t tests should
be encouraged, at least for some research purposes. (See, for example,
Saville, 1990.) I will review the role of statistical hypothesis testing in
research, discuss the problem with multiple t tests, indicate when they are
(and are not) a problem, and suggest a compromise solution.
The purpose of statistical inference in general (and the t test in
particular) is to generalize from the particular cases studied (e.g., the 58
students you tested) to the much larger population that these cases
represent (e.g., all students who will take similar classes in the future).
When you find a statistically significant result (declared by statistical
tradition whenever the p value is less than .05) you have effectively ruled
out random chance as a reasonable explanation. (While you haven’t
completely ruled out random chance, you have controlled the Type I error
of wrongly declaring significance so that it happens only about 5% of the
time.) This gives you a license to explain the result as being more than pure
coincidence. Journals often discourage you from explaining results that are
not significant because, in effect, there is really nothing to explain.
Difficulties can arise as soon as you have more than one t test. The basic
problem is this: Finding one or more significant results (out of multiple
t tests) no longer eliminates random coincidence as a reasonable
THE FORUM 773

possibility. Instead of wrongly declaring significance at the low,
acceptable 5% rate, these errors may occur at an unacceptably high rate. In
effect, the “license to explain” has then been obtained under false
pretences.
When you do many t tests and find only marginal significance (i.e., the
significant p value is only slightly less than .05) you have a problem: This
could indeed happen due to random chance alone. You have not really
earned a license to explain the supposedly significant difference. Here’s
why. Suppose you did 20 t tests and found that one was significant. If each
one, by random chance alone, has a 5% chance (1 in 20) of being significant
then you would expect, on average, to find 1 significant result out of these
20 tests by random chance alone.
Fortunately, there are some clear-cut cases of multiple t tests that do not
pose any problem. One case is when none of the t tests is significant. Of
course, this case is not usually very helpful to the researcher in search of a
journal publication. (There are, however, some projects in which finding
no significant t tests is interesting. Usually a statistical power study is called
for to establish that the effects being sought are really absent and were not
merely missed due to small sample size.) Another case is when the
significant t tests are very highly significant (i.e., p < .001), as shown in the
hypothetical example of Table 1, comparing 11 experimental teaching
methods to a control group, with 180 students in all. The p values here send
an unambiguous message: Those < .05 clearly indicate significant
differences from the control group; their significance has not been
compromised by the use of multiple t tests. The remaining groups (C and
E) are not significantly different from the control group. Some statisticians
might jokingly say that this passes the “interocular test” (that is, it hits you
between the eyes) and that further worrying is unnecessary and would
only reconfirm the obvious.
TABLE 1
Comparison of Average Scores
(n= 15 for each group)
774 TESOL QUARTERLY

There is a good solution to the problem in most other cases: Use a single
overall test as a gateway to further inference. If this overall test is not
significant, you must stop because there are no significant effects. If the
overall test is significant, then you have a license to examine the individual
differences. However, if used by itself this cure can be worse than the
disease due to the extra complexity of the overall test.
The appropriate overall test to use depends upon the situation. If you are
comparing average scores of several groups, the F test for the one-way
analysis of variance (ANOVA) would ordinarily be used. When predicting
one score from several other factors, you would use the multiple regression
F test (equivalent to testing the R2 value). If you have several test scores for
each student (e.g., reading, writing, etc.) for each of two groups (or before
and after some intervention with just one group) you would use Hotelling’s
T2 test. There are also overall ANOVA and MANOVA (multivariate
analysis of variance), and other tests that are appropriate for many more
complex situations.
When there is a problem with multiple t tests, a good compromise
solution is to include both kinds of tests: (a) the overall test (if significant)
ensures that there is something to explain, while (b) the individual
estimates (e.g., group means and standard deviations) and individual
t tests will help your readers understand the results.
For example, if you have scores for reading, writing, listening, and
speaking for students in two groups, you could first use Hotelling’s T 2 test
to see if there are any differences at all between the two groups, looking
simultaneously at all scores. If this single overall test is significant, then you
could use multiple t tests (one each for reading, writing, listening, and
speaking) to examine the mean values and pinpoint the differences.
THE AUTHOR
Andrew F. Siegel is Professor of Management Science and Finance, and Adjunct
Professor of Statistics, at the University of Washington. He holds a PhD in Statistics
from Stanford University. His publications on statistical theory and applications
have appeared in many journals, including ]ournal of the American Statistical
Association, Annals of Statistics, American Statistician, Journal of the Royal
Statistical Society, and Encyclopedia of Statistical Sciences, His two most recent
books are Statistics and Data Analysis: An Introduction (Wiley, 1988) and Practical
Business Statistics (Irwin, 1990).
REFERENCE
Saville, D. J. (1990). Multiple comparison procedures: The practical solution, The
American Statistician, 44, 174-180.
THE FORUM 775

INFORMATION FOR CONTRIBUTORS
EDITORIAL POLICY
The TESOL Quarterly, a professional, refereed journal, encourages submis-
sion of previously unpublished articles on topics of significance to indi-
viduals concerned with the teaching of English as a second or foreign
language and of standard English as a second dialect. As a publication which
represents a variety of cross-disciplinary interests, both theoretical and
practical, the Quarterly invites manuscripts on a wide range of topics,
especially in the following areas:
1. psychology and sociology of language 3. testing and evaluation
learning and teaching; issues in research 4. professional
and research methodology preparation
2. curriculum design and development; 5. language planning
instructional methods, materials, and 6. professional standards
techniques
Because the Qaatierly is committed to publishing manuscripts that con-
tribute to bridging theory and practice in our profession, it particularly
welcomes submissions drawing on relevant research (e.g., in anthropology,
applied and theoretical linguistics, communication, education, English
education [including reading and writing theory], psycholinguistics,
psychology, first and second language acquisition, sociolinguistics, and
sociology) and that address implications and applications of this research to
issues in our profession. The Quarterly prefers that all submissions be
written so that their content is accessible to a broad readership, including
those individually who may not have familiarity with the subject matter
addressed.
GENERAL INFORMATION FOR A U T H O R S

1. The TESOL Quarterly invites submissions in five categories:
Full-length articles. Manuscripts should generally be no longer than 20
double-spaced pages. Submit three copies plus three copies of an infor-
mative abstract of not more than 200 words. To facilitate the blind
review process, authors’ names should appear only on a cover sheet, not
on the title page; do not use running heads. Manuscripts should be
submitted to the Editor of the TESOL Quarterly:
Sandra Silberstein
Department of English, GN-30
Seattle, WA 98195 U.S.A.
Reviews. The TESOL Quarterly invites reviews of textbooks, scholarly
works related to the profession, tests, other instructional materials (such as
computer software, videotaped materials, and other nonprint materials),
777
and other journals concerned with issues relevant to our profession.
Comparative reviews, which include a discussion of more than one
publication, and review articles, which discuss materials in greater
depth than in a typical review, are particularly welcome. Reviews
should generally be no longer than five double-spaced pages, although
comparative reviews or review articles may be somewhat longer. Until
further notice, submit two copies of reviews to the Review Editor of the
TESOL Quarterly:
Heidi Riggenbach
Department of English, GN-30
Seattle, WA 98195 U.S.A.
Book Notices. The TESOL Quarterly also welcomes short evaluative
reviews. Book notices should provide a descriptive and evaluative
summary of a recent publication (see preceding section for appropriate
types of publications) and a brief discussion of the significance of the
work in the context of current theory and practice in the relevant area(s)
of TESOL. Submissions should range between 350 and 500 words; any
submission that exceeds 500 words will be returned. Submit two copies
of book notices to Heidi Riggenbach, Review Editor, at the address
given above.
Brief Reports and Summaries. The TESOL Quarterly also invites short
descriptions of completed work or work in progress on any aspect of
theory and practice in our profession. Reports of work in the areas of
curriculum and materials development, methodology, teaching, testing,
teacher preparation, and administration are encouraged, as are reports
of research projects of a pilot nature or that focus on topics of
specialized interest. In all cases, the discussion of issues should be
supported by empirical evidence, collected through either formal or
informal investigation, and should be grounded with references to
similar or related work. Manuscripts should summarize key concepts
and results in a manner that will make the research accessible to our
diverse readership. Although all reports and summaries submitted to
this section will be considered, preference will be given to manuscripts
of three to seven double-spaced pages (including references and notes).
Longer articles do not appear in this sectwn and should be submitted to
the Editor of the TESOL Quarterly for review. Send two copies o f
reports and summaries to the Editor of the Brief Reports and
Summaries section:
Gail Weinstein-Shr
Reading and Writing Program
School of Education
Furcolo Hall
University of Massachusetts
Amherst, MA 01003-~96
The Forum. The TESOL Quarterly welcomes comments and reactions
from readers regarding specific aspects or practices of our profession.
778 TESOL QUARTERLY
Responses to published articles and reviews are also welcome. Con-
tributions to The Forum should generally be no longer than five double-
spaced pages. Submit two copies to the Editor of the TESOL Quarterly
at the address given above.
Brief discussions of Research Issues in qualitative and quantitative
research are also published in The Forum. Although these contributions
are typically solicited, readers may send topic suggestions and/or make
known their availability as contributors by writing directly to the Editor
of Research Issues:
Graham Crookes
Department of English as a Second Language
IJniversity of Hawaii at Manoa
1890 East-West Road
Honolulu, HI 96822 U.S.A.
2. All submissions to the Quarterly should conform to the requirements
of the Publication Manual of the American Psychological Association
(Third Edition), which can be obtained from the Order Department,
American Psychological Association, 1200 Seventeenth Street, NW,
Washington, DC 20036. The Publication Manual is also available in
many libraries and bookstores.
3. All submissions to the TESOL Quarterly should be accompanied by
a cover letter which includes a full mailing address and both a
daytime and an evening telephone number. Where available, include
an electronic mail address and fax number.
4. Authors o,f full-length articles should include two copies of a very
brief biographical statement (in sentence form, maximum 50 words),
plus any special notations or acknowledgments that they would like
to have included. Double spacing should be used throughout.
5. The TESOL Quarterly provides 25 free reprints of published full-
length art icles and 10 reprints of material published in the Reviews,
Brief Reports and Summaries, and The Forum sections.
6. Manuscripts submitted to the TESOL Quarterly cannot be returned
to authors. Authors should be sure to keep a copy for themselves.
7. It is understood that manuscripts submitted to the TESOL Quarterly
have not been previously published and are not under consideration
for publication elsewhere.
8. It is the responsibility of the author(s) of a manuscript submitted to
the TESOL Quarterly to indicate to the Editor the existence of any
work already published (or under consideration for publication
elsewhere) by the author(s) that is similar in content to that of the
manuscript.
9. The Editor of the TESOL Quarterly reserves the right to make
editorial changes in any manuscript accepted for publication to
enhance clarity or style. The author will be consulted only if the
editing has been substantial.
INFORMATION FOR CONTRIBUTORS 779

10. The views expressed by contributors to the TESOL Quarterly do
not necessarily reflect those of the Editor, the Editorial Advisory
Board, or TESOL. Material published in the Quarterly should not
be construed to have the endorsement of TESOL.
780 TESOL QUARTERLY

PUBLICATIONS RECEIVED
Publishers are invited to send copies of their new materials to the TESOL Quarterly
Review Editor, Heidi Riggenbach, University of Washington, at the address listed
in the Information for Contributors section. Packages should be labeled REVIEW
COPIES.
TESOL Quarterly readers are invited to contribute review articles and evaluative
or comparative reviews for consideration for publication in the Review or Book
Notice sections of the Quarterly. These should be sent to the TESOL Quarterly
Review Editor, Heidi Riggenbach, University of Washington, at the address listed
in the Information for Contributors section.
TESOL gratefully acknowledges receipt of the following publications.
Alexander, L. C. (1990). Longrnan De Devitiis, G., Mariani, L,, & O’Mal-

English grammar practice (Self- ley, K. (1989). English grammar for
study Edition with Key). Harlow, communication exercises. Harlow,
England: Longman. England: Longman.
Bathe, E. (1989). Culture clush (2nd Eastman, C. M. (1990). Aspects of lan-
cd.). Yarmouth, ME: Intercultural guage and culture (2nd cd.). Novato,
Press. CA: Chandler & Sharp.
Baldauf, R., Jr., & Luke, A. (Eds.). Guglielmino, L. M. (1991). Adult ESL
(1990). Language pkmning and edu- instruction: A sourcebook. Glen-
cation in Australasia and the South view, IL: Scott, Foresman.
Pacific. Clevedon, Avon, England: Barrington, P. (Ed.). (1990). This land
Multilingual Matters. is your land: Preparation for amnesty
Begin, S., Bianco, P., Kirkwold, L., & legalization and citizenship. Glen-
Mackie, A. (1990). Suspicious minds: view, IL: Scott, Foresman.
Listening strategies and life skills. Hartmann, P. (1989). Clues to culture:
White Plains, NY: Longman. A cross-cultural reading/writing
Blum, L. (1990). Tuning in to spoken book. New York: Random House.
messages: Basic listening strategies. Heaton, J. B. (1990). Classroom test-
White Plains, NY: Longman. ing. Harlow, England: Longman.
Bohlman, C., & Porter, C. (1990). The Huizenga, J., & Thomas-Ruzic, M.
Uncle Sam activity book: Language (1990). Writing workout: A program
development to teach U.S. history for new students of English. Glen-
and government, Glenview, IL: view, IL: Scott, Foresman/Little,
Scott, Foresman. Brown.
Boyd, F., & Quinn, D. (1990). Stories Jacobson, R., & Faltis, C. (Eds.).
from Lake Wobegon: Advanced (1990). Language distribution issues
listening and conversation skills. in bilingual schooling. Clevedon,
White Plains, NY: Longman. Avon, England: Multilingual Mat-
Butler, P., & Harris, A. (1990). English ters.
as a second language bibliography. Jong, J. H. A. L. de, & Stevenson, D. K.
Vancouver, Canada: Vancouver (Eds.). (1990). Individualizing the
Community College. assessment of language abilities.
Byram, M., & Leman, J. (Eds.). (1990). Clevedon, Avon, England: Multilin-
Bicultural and trilingual education: gual Matters.
The Foyer Model in Brussels. Cleve- Kirchner, C. (1990). Write on cue:
don, Avon, England: Multilingual Beginning ESL writing exercises.
Matters. Glenview, IL: Scott, Foresman.
781
Kirchner, C. (1990). Write on cue: Literacy Volunteers of New York City
Beginning ESL writing exercises (Ed.). (1990). Selected from The lost
(Instructor’s Guide). Glenview, IL: _ (Writers’ Voices M. H, Clark).
Scott, Foresman. New York: Editor.
Kroll, B. (Ed.). (1990). S e c o n d Zan- Literacy Volunteers of New York City
guage writing: Research insights for (Ed.). (1990). Speaking from the
the classroom. Cambridge: Cam- heart: An anthology of writing by
bridge University Press. new writers (New Writers’ Voices).
Ladd, J. (1990). Subject India: A New York: Editor.
semester abroad. Yarmouth, ME: Literacy Volunteers of New York City
Intercultural Press. (Ed.). (1990). Speaking out on home
Lai, A. K. T. (1989). The art in the use and family: An anthology of writing
of “and” and “and/or” or “or’’-How by new writers (New Writers’ Voi-
it works. Hong Kong: Professional- ces). New York: Editor.
man ship. Literacy Volunteers of New York City
Law, B., & Eckes, M. (1990). T h e (Ed.). (1990). When dreams come
more-than-just-surviving handbook: true (New Writers’ Voices: C. Miles).
ESL for every classroom teacher. New York: Editor.
Winnipeg, Canada: Peguis.
Mair, C. (1990). Infinitival comple-
Literacy Volunteers of New York City ment clauses in English: A study of
(Ed.). (1990). Can’t wait for summer syntax in discourse. Cambridge:
( N e w Writers’ Voices T. Sanser- Cambridge University Press.
vino). New York: Editor.
Maple, R., & Ong, M. F. (1991). New
Literacy Volunteers of New York City
wave 3. Harlow, England: Longman.
(Ed.). (1990). From my imagination:
An anthology of poetry and short Monica, A., & Yalden, J. (1984). Eng-
stories by new writers (New Writers’ lish and French as second languages
Voices). New York: Editor. in Canadian teacher-education insti-
Literacy Volunteers of New York City tutions. Welland, Canada: The Cana-
(Ed.). (1990). Selected from China dian Modern Language Review.
men & The woman warrior m Nelson, V. (1991). Listening to commu-
ers’ Voices: M. H. Kingston). New nicate in English: Intermediate level
York: Editor. (Text, Instructor’s Manual, Answer
Literacy Volunteers of New York City Key, & Cassettes). Lincolnwood, IL:
(Ed.). (1990). Selected from @ National Textbook.
miner’s daughter (Writers’ Voices: L. Ochsner, R. S. (1990). Physical elo-
Lynn with G. Vecsey). New York: quence and the biology of writing.
Editor. Albany, NY: State University of
Literacy Volunteers of New York City New York.
(Ed.). (1990). Selected from contem- Porter, E. A. (1990). Foreign teachers
porary American plays: An anthol- in China: Old problems for a new
ogy (Writers’ Voices). New York: generation, 1979-1989. Westport,
Editor. CT: Greenwood Press.
Literacy Volunteers of New York City
Price, P. J. (1990). Competency in
(Ed.). (1990). Selected from. % English: A life skills approach. New
steps: The Autobiography York: McGraw-Hill.
reem Abdul-]abbar) (Writers’ Voi-
ces: K. Abdul-Jabbar & P. Knobler). Quirk, R., & Stein, G. (1990). English in
New York: Editor. use. Harlow, England: Longman.
Literacy Volunteers of New York City Radley, P., & Millerchip, C. (1990).
(Ed.). (1990). Selected from J= Mode 3 (Text, Workbook, Teacher’s
(Writers’ Voices: P. Benchley). New Book, & Cassettes). London: Collins
York: Editor. ELT.
782 TESOL QUARTERLY

Ramsey, G., & LoCastro, V. (1990), Scott, Foresman. (1990). Posters for
Talking topics. Harlow, England: English for a changing world: w
Longman. on cue & Cuing in with pictures.
Richards, J. C., IHull, J., & Proctor, S. Glenview, IL: Author.
Interchange: Student’s Book 2. Eng-
Tribble, C., & Jones, G. (1990). Con-
lish for international communication
(with Cassette). Cambridge: Cam- cordances in the classroom. Harlow,
England: Longman.
bridge University Press.
Robertshaw, S. D., Hamblen, R. E., & VanPatten, B., & Lee, J. F. (1990).
Feldman, R. (1990). Reading first: Second language acquisition/f oreign
Building reading competence (2nd language learning. Cievedon, Avon,
cd.). New York: Maxwell Macmil- England: Multilingual Matters.
lan.
Rucinski, C. J. (1990). Cuing in: Activ- Vaughan, A., & Heyen, N. (1990).
ities on blackline masters for begin-
Ready for business. Harlow, Eng-
land: Longman.
ning writers of English. Glenview,
IL: Scott, Foresman. Wilkinson, A., Davies, A., & Berrill, D.
Rucinski, C. J. (1990). Cuing in with (1990). Spoken English illuminated.
pictures. Glenview, IL: Scott, Fores- Buckingham, England: Open Uni-
man. versity.
PUBLICATIONS RECEIVED 783

PUBLICATIONS AVAILABLE
FROM THE TESOL CENTRAL OFFICE
Publications on a wide range of topics relevant to TESOL professionals are avail-

able from the TESOL Central Office. Many of these publications are offered at
reduced rates to TESOL members. An annotated list of publications can be obtained
by contacting TESOL Publications Department, Suite 300, 1600 Cameron Street,
Alexandria, Virginia 22314 U.S.A. A TESOL Order Form is printed in every issue of
the TESOL Quarterly and can be used to order any of these materials. Please note
that prices inside parentheses are special discount prices for TESOL members only.
Prices are subject to change without notice.
REFERENCE GUIDES Bilingual, ESOL and Foreip Language

Z)irectory of ProfesskmaJ Preparation Pro- Teacher Preparation: Models, Practices,
grams in TESOL in the United States, Issues. John F. Fanselow and Richard L.
1989-1991. Describes 245 college and Light, editors. An essential reader for
university programs in TESOL leading to professional preparation. 1978. $10.00
doctorate, master’s, or bachelor’s degrees, ($8.00). ISBN O-939791-03-X
and certificates. Overview of state certifi-
cation requirements. 1989.$20.00 ($15.00).
ISBN 0-939791 -35-8
1987 TESOL Membership Directory. A
comprehensive resource of who’s who ‘1’HEMATICS
in TESOL. Both alphabetical and geo-
graphical listings of more than 11,000 Coherence in Writing: Research and
institutional, commercial, and individual Pedagogical Perspectives. UBa Connor
members. $5.00. and Ann M. Johns, editors. Interpreta-
A World of Books: An Annotated Reading tions of coherence for the ESL/EFL
List for ESL/EFL Students. Dorothy S. teacher, teacher trainer, and researcher
Brown. More than 75 entries, including in L1 and L2. The processes of making
titles on various subjects and levels to assist and understanding coherence and the
the intermediate through advanced student manifestations of coherence in written
in selecting reading materials. 1987.$7.50 products. Discussion questions and
($6.00). ISBN 0-93979132-3 activities. 1990. $14.95 ($10.95). ISBN
O-939791-34-X
Selected Articles From the TESOL
Newsletter: lW-19$3. John F. Haskell, Students and Teachers Writing To-
editor. More than 100 articles on gether: Perspectives on Journal Writing.
methodology, language and culture, Joy Kreeft Peyton, editor. Examines
linguistics and grammar, language as- journal writing as a powerful stimulus
sessment, composition, ESP, reading, for curriculum and teacher change; as a
and vocabulary. 1986. $16.50 ($15,00). forum for discussing the meaning and
ISBN 0-939791 -29-3 form of literary texts and the journal
Annotated Bibliography of ESL Mate- texts themselves; and as a bridge to other
riafs. Christine Aronis. Gathered from 76 kinds of academic writing. Applications
publishers, the 636 ESL titles listed were for ESL, bilingual education, composi-
designed to teach American English. tion, writing across the curriculum,
Provides ESL teachers with an accurate writing research, literacy development,
and comprehensive list of titles available. and deaf education. 1990.$11.95 ($9.95).
1983. Sale $5.00. ISBN 0-318- 18069-3 ISBN 0-939791 -36-6
785
Research in Reading in English as a CLASSROOM PRACTICES
Second Language. Joanne Devine, Patri-
cia L. Carrell, and David E. Eskey, Classroom Practices in Adult ESL.
editors. Views of reading as an interac- Donna Ilyin and Thomas Tragardh,
tive process involving reader and text editors. Articles on topics such as
variables, of the interplay of reading classroom organization and manage-
skills and general language proficiency, ment, evaluation, materials, cultural
and of the differences h first and second considerations, literacy and reading,
language reading. 1987.$16.50 ($15.00). communicative competence, and more.
ISBN 0-939791 -30-7 1978.$8.00 ($5.00).
Children and ESL: Integrating Perspec- Classroom Practices in ESL and Bilin-
tives. Pat Rigg and D. Scott Enright, gual Education. Muriel Saville-Troike,
editors. ESL teachers as language advo- editor. A collection of articles designed
cates, ESL children’s writing, reading in to acquaint classroom teachers with
ESL, and more. 1986.$12.50 ($10.00). successful methods and materials used in
ISBN O-939791-M-2 the ESOL and bilingual education
contexts. 1979.$5.00 ($4.00).
Current Perspectives on Pronunciation:
Practices Anchored in Theory. Joan Mor-
ley, editor, The pronunciation component
of oral language how it operates to
transmit meaning, how it can be learned, ON TESOL SERIES
how teachers facilitate learning. 1987.
$14.00 ($12.00). ISBN 0-939791 -28-5 Selected papers from TESOL Annual
Conventions contained in the following
Ending Remediation: Linking ESL and
volumes:
Content in Higher Education. Sarah
Benesch, editor. How higher education On TESOL ’84. Penny Larson, Elliot L.
is bridging the gap between ESL and Judd, and Dorothy S. Messerschmitt,
content courses through collaboration editors. 1984 in Houston.
across the curriculum. Eight faculty
members describe innovative linkage On TESOL ’83. Jean Handscombe,
programs. 1988. $12.00 ($10.00). ISBN Richard A. Orem, and Barry P. Taylor,
0-939791 -33-1 editors. 1983 in Toronto.
On TESOL ’82. Mark A. Clarke and
Jean Handscombe, editors. 1982 in
TESTING Honolulu.
Reviews of English Language Profi- On TESOL ’81. Mary E. Hines and
ciency Tests. J. Charles Alderson, Karl J. William Rutherford, editors. 1981 in
Krahnke, and Charles W. Stansfield, Detroit.
editors. Descriptive and evaluative
information on the major ESL/EFL tests On TESOL “76. John F. Fanselow and
used worldwide, including more than 40 Ruth H. Crymes, editors. 1976 in New
reviews. 1987. $16.50 ($15.00). ISBN York City.
0-939791 -31-5
On TESOL ’74. Ruth Crymes and
William E. Norris, editors. 1974 in
Technology and Language Testing.
Charles W. Stansfield, editor. New Denver.
developments in measurement theory Individual volumes available for $9.95
and test administration methodology. ($7.95).
Papers from the 1985 Language Testing
Research Colloquium. $12.50 ($9.00). Complete set of all 6 volumes of the On
ISBN 0-939791 -08-0 TESOL Series – $45.95 (35.95).
786 TESOL QUARTERLY

TESOL PAPERS TESOL CLASSICS ON SALE
Statement of Core Standards for Lan- The Acquisition and Use of Spanish and
guage and Professional Preparation English as First and Second Languages.
Programs. 1985. Roger W. Andersen, editor. Colloquium
papers from TESOL 1978 in Mexico
Guidelines for the Certification and City. Reg. $6.50 Sale $3.00.
Preparation of Teachers of English to
Speakers of Other Languages in the
United States. 1976.
All papers: single copies, $1.00 each; The Construct Validation of Tests of
multiple copies in groups of 10 for $5.00. Communicative Competence. Adrian S.
Palmer, Peter J. M. Groot, and George
New!
Resolutions Adopted hy the TESOL A. Trosper, editors. Papers from the
1979 Testing Colloquium in Boston. Reg.
Membership at Legislative Assemblies,
1971 to Present. 8%” x 11”, 3-hole $6.50 Sale $3.00.
punched with covers, shrink-wrapped.
$4.00 ($2.50) .
TESOL NEWSLETTER The Human Factors fi ESL. James E.
SUPPLEMENTS Alatis and Ruth Crymes, editors. Collec-
TESOL Newsletter: 21st Anniversary tion of papers by ESOL experts: Finoc-
Issue. More than 20 articles, including an chiaro on the teacher, Paulston on
ESOL bare-bones bibliography, state of communicative competence, Tucker on
the art, early history of TESOL, facts innovative approaches, Burt on error
and faces in TESOL, and more. April, analysis, and more. 1977. Reg. $5.00 Sale
1987.$3.00. $3.00.
Computer-Assisted Language Learning.

14 articles covering various aspects of
CALL. February, 1986.$2,50.
A Handbook of Bihhgual Education
Writing and Composition. 9 articles on (Rev. Ed.). Muriel R. Saville and Rudolph
ESL writing and composition, plus tips C. Troike. Specific suggestions for-. de-
for writing teachers and a brief list of sign, instruction, and evaluation of bilin-
books and journals. February, 1984. gual programs. 1978. Sale $3.00.
$2.50.
Dialogue Journal Writing with Nonnative English

~1* Speakers: A Handbook for Teachers, Joy Kreeft Peyton
@ and Leslee Reed. This comprehensive handbook offers
practical advice in an easy-to-use format free of jargon. It
is recommended for teachers of nonnative English-
speaking students in mainstream, bilingual, or ESL
programs, from kindergarten through high school. It also
has direct application to native English-speaking, gifted
and talented, learning disabled, and special education
students. 1990.$9.95 ($6.95). ISBN 0-939791 -37-4
PUBLICATIONS AVAILABLE 787

CuMcJIA’mm I N D E X
T~OL Quarterly, Volumes 23-24 (1989-1990)
Author Index
Abraham, Roberta G., and Vann, Auerbach, Elsa, Review of Getting
Roberta J., Strategies of Unsuccess- There: Producing Photostories With
ful Language Learners, Vol. 24, No, Immigrant Women (Deborah
2> pp. 177-198. Barndt, Ferne Cristall, and dian
Aghbar, Ali A., Review of Classroom marine), Vol. 23, No. 2, pp. 323-324.
Interaction (Ann Malamah-Thomas),
Vol. 23, No. 1, pp. 141-142. Auerbach, Elsa, Review of Tell Me
About It: Reading and Language
Alford, Randall L., and Strother, Judith
Activities Around Multi-Cultural
B., Attitudes of Native and Nonna-
Issues Based on an Oral History
tive Speakers Towards Selected
Approach (Azi Ellowitch), Vol. 23,
Accents of U.S. English, Vol. 23, No.
No. 2, pp. 327-328.
3, pp. 479-495.
Andrade, Melvin R., Jr., Review of Auerbach, Elsa, Review of Voices:
Writing Across Languuges: Analysis New Writers for New Readers, Vol.
of L2 Text (Ulla Connor and Robert 23, No. 2, pp. 333-334.
B. Kaplan,’ Eds.), Vol. 24, No. 4,
pp. 724-727.
Bamford, Julian, ~omments on “Crea-
Ashby, Eda, Review of Crazy Idioms:
tive Automization: Principles for
A Conversational Idiom Book (Nina
Promoting Fluency Within a Com-
Weinstein), Vol. 24, No. 3, pp. 511-
municative Framework”: A Reader
512.
Reacts, Vol. 2.3, No. 2, pp. 363-366.
Auerbach, Elsa (Ed,), Book Notices on
Nontraditional Materials for Adult Barnes, Vincent G., Review of Dou-
ESL, Vol. 23, No. 2, pp. 321-335. blespeak (William Lutz), Vol. 24,
Auerbach, Elsa, Comments on Alien No. 3, pp. 517-518.
Winds: The Reeducation of Ameri-
ca’s Indochinese Refugees and the Barrett, Mary T., The Secondary ESL
Book Review: The Reviewer Re- Reading Course: Rationale and Im-
sponds, Vol. 24, No. 3, pp. 541-542. plementation, Vol. 23, No. 1, pp.
152-156.
Auerbach, Elsa, Review of A l i e n
Winds: The Reeducation of Ameri- Benson, Malcolm J., The Academic
ca’s Indochinese Refugees (James Listening Task: A Case Study, Vol.
W. Tollefson), Vol. 24, No. 1, pp. 85- ~, No. 3> pp. 421-445.
91.
Auerbach, Elsa, Review of English at Benson, Morton, Differences Between
Work: A Tool Kit for Teachers American English and British Eng-
(Deborah Barndt, Coordinator), lish: A Challenge to TESOL, Vol. 23,
Vol. 2.3, No. 2, pp. 322-323. No. 2, pp. 351-355.
791
Bensoussan, Marsha, Review of New- Brisk, Maria, Review of Writing in a
bury House TOEFL Preparation Kit: Bilimzual Proszram: Habia u n a Vez
Preparing for the TOEFL (Daniel B. (Car~le Edel~y), Vol. 2.3, No. 3, pp.
Kennedy, Dorry Mann Kenyon, and 536-537.
Steven J. Matthiesen) and Newbury Brown, James Dean, Improving ESL
House TOEFL Preparation Kit: Placement Tests Using Two Per-
Preparing for the Test of Written spectives, Vol. 23, No. 1, pp. 65-83.
English (Liz Hamp-Lyons), Vol. 24, Brown, James Dean, Research Issues:
No. 3, pp. 501-506. The Use of Multirde t Tests in Lan-
Beretta, Alan, Attention to Form or guage Research, ~ol. 24, No. 4, pp.
Meaning? Error Treatment in the 770-773.
Bangalore Project, Vol. 23, No. 2, Brown, James Dean, Review of Test-
pp. 283-303. ing Spoken Language: A Handbook
Beretta, Alan, Comments on “Attention of Oral Testing Techniques (Nit
to Form or Meaning? Error Treat- Underhill), Vol. 23, No. 1, pp. 144-
ment in the Bangalore Project”: The 145.
Author Responds, Vol. 24, No. 1, p. Burnaby, Barbara, and Sun, Yilin,
115. Chinese Teachers’ Views of Western
Berliner, David, Dunkel, Patricia, and Language Teaching Context In-
Mishra, Shitala, Effects of Note forms Paradigms, Vol. 2.3, No. 2, pp.
Taking, Memory, and Language 219-238.
Proficiency on Lecture Learning for
Byrd, Patricia, and Canseco, Grace,
Native and Nonnative Speakers of
Writing Required in Graduate
English, Vol. 2.3, No. 3, pp. 543-549.
Courses in Business Administration,
Bermudez, Andrea B., and Prater, Vol. 2.3, No. 2, pp. 305-316.
Doris L., Using Brainstorming and
Clustering with LEP Writers to Canseco, Grace, and Byrd, Patricia,
Develop Elaboration Skills, Vol. 2A, Writing Required in Graduate
No. 3, pp. 52.3-528. Courses in Business Administration,
Beyer, Frank, Review of Changes: Vol. 23, No. 2, pp. 305-316.
Readings for ESL Writers (Jean
Carrell, Patricia L., Carson, Joan
Winthrow, Gay Brooks, and Martha
Eisterhold, Silberstein, Sandra,
Clark Cummings), Vol. 24, No. 4,
Kroll, Barbara, and Kuehn, Phyllis
pp. 737-738.
A., Reading-Writing Relationships in
Bialystok, Ellen, The Competence of
First and Second Language, Vol. 24,
Processing: Classifying Theories of No. 2, pp. 245-266.
Second Language Acquisition, Vol.
24, No. b, pp. 635-648. Carrell, Patricia L., Pharis, Becky G.,
Blau, Eileen K., The Effect of Syntax, and Liberto, Joseph C., Metacogni-
Speed, and Pauses on Listening tive Strategy Training for ESL Read-
Comprehension, Vol. 24, No. 4, pp. ing, Vol. 23, No. 4, pp. 647-678.
746-753. Carson, Joan Eisterhold, Carrell, Patri-
Brinton, Donna (Ed.). Book Notices on cia L., Silberstein, Sandra, Kroll,
the Literature-Enriched Curriculum, Barbara, and Kuehn, Phyllis A.,
Vol. 24, No. 4, pp. 729-739. Reading-Writing Relationships in
Brinton, Donna, and Holten, Christine, First and Second Language, Vol. 24,
What Novice Teachers Focus On: No. 2, pp. 245-266.
The Practicum in TESL, Vol. 23, No. Cathcart, Ruth Larimer, Authentic
2, pp. 343-350. Discourse and the Survival English
Brinton, Donna M., Review of The Curriculum, Vol. 23, No. 1, pp. 105-
Teaching of Pronunciotwn: An Intro- 136.
duction for Teachers of English as a Cazden, Courtney B., Review of Re-
Second Language (Peter Avery and cent Publications on Classroom
Susan Ehrlich, Eds.), Vol. 2.3, No. 1, Research, Vol. 24, No. 4, pp. 717-
pp. 135-136. 724.
792 TESOL QUARTERLY

Chan, Michelle, Review of Longrnan Dunkel, Patricia A., Review of Simula-
Bibliography of Composition and tion-Gaming in the Late 1980s: Pro-
Rhetoric: 1984-1985 (Erika Linde- ceedings of the International Simula-
mann, Ed.), Vol. 23, No. 1, pp. 138- tion and Gaming Association’s I?th
140. International Conference (David
Chan, Michelle, Review of Longrmm Crookall, Cathy Greenblat, Alan
Bibliography of Composition and Coote, Jan Klabbers, and D.R. Wat-
Rhetoric: 1986 (Erika Lindemann, son, Eds.), Vol. 23, No. 1, pp. 136-
Ed.), Vol. ~, No. 1, pp. 138-140. 137.
Chapelle, Carol, The Discourse of Dunkel, Patricia, Mishra, Shitala, and
Computer-Assisted Language Learn- Berliner, David, Effects of Note
ing: Toward a Context for Descrip- Taking, Memory, and Language
tive Research, Vol. 24, No. 2, pp. Proficiency on Lecture Learning for
199-222. Native and Nonnative Speakers of
Collier, Virginia P., How Long? A English, Vol. 23, No. 3, pp. 543-549.
Synthesis of Research on Academic Dunn, Martha, Review of At the Door:
Achievement in a Second Language, Selected Literature for ESL Students
vol. ~, No. 3, pp. 509-531. (Sandra McKay and Dorothy Petitt),
Conrad, Susan, Goldstein, Lynn, Stu- vol. %, No. 4, pp. 736-737.
dent Input and Negotiation of Mean- Dunning, Mark, Review of Academic
ing in ESL Writing Conferences, Writing: Techniques and Tasks, Vol.
vol. 24, No. 3, pp. 443-460. 24, No. 1, pp. 91-95.
Cook, Cynthia, Review of Southeast
Asians: A New Beginning in Lowell Eastman, Carol M., What is the Role of
(James Higgins and Joan Ross), Vol. Language Planning in Post-Apart-
~, No. 2, pp. 326-329. heid South Africa?, Vol. 24, No. 1,
pp. 9-21.
Davies, Alan, Is International English
an Interlanguage?, Vol. 23, No. 3, Fantuzzi, Cheryl, Hatch, Evelyn, and
pp. 447-467. Yasuhiro, Shirai, The Need for an
Demel, Marjorie C., The Relationship Integrated Theory: Connecting
Modules, Vol. 24, No. 4, pp. 697-716.
Between Overall Reading Compre-
Ferguson, K. Scott, Review of A
hension and Comprehension of
Linguistic Study of American Punc-
Coreferential Ties for Second Lan-
tuation (Charles F. Meyer), Vol. 24,
guage Readers of English, Vol. 24,
No. 3, pi. 515-516.
No. 2, pp. 267-292. Fox, Wanda S., Review of The Foreign
Dollv. Martha R., Adult ESL Students’
Teaching Assistant’s Manual (Patri-
M%agement of Dialogue Journal cia Byrd, Janet C. Constantinides,
Conversation, Vol. 24, No. 2, pp. and Martha C. Pennington), Vol. 24,
317-320. No. 3, pp. 508-509.
Dubois, Betty Lou, Comments on Freedman. Sarah Warshauer, and
“Toward a Pedagogy of Possibility McKay, Sandra Lee, Language Mi-
in the Teaching of English Intern- nority Education in Great Britain: A
ationally: People’s English in South Challenge to Current U.S. Policy,
Africa”: A Reader Reacts, Vol. 24, vol. $?/i, No. 3, pp. 365-405.
No. 1, pp. 103-104. Freeman, David, Comments on “Some
Duff, Barbara, Comments on “Do We Mvths You Mav Have Heard About
Need to Teach Spelling?”: A Reader Fi’rst Language Acquisition”: A
Reacts, Vol. 23, No. 1, pp. 163-164. Reader Reacts. Vol. 23, No. 1, pp.
Dungey, Joan, Review of America: 157-159. ‘
The Early years; A m e r i c a : After Freeman, Donald, Teacher Training,
Independence; Teacher’s Guide Development, and Decision Making:
(Language Development Through A Model of Teaching and Related
Content) (Anna Uhl Chamot), Vol. Strategies for Language Teacher
.23, No. 1, pp. 133-134. Education, Vol. 2.3, No. 1, pp. 27-45.
CUMULATIVE INDEX 793

Gatbonton, Elizabeth, and Segalowitz, Hatch, Evelyn, Yasuhiro, Shirai, and
Norman, Comments on “Creative Fantuzzi, Cheryl, The Need for an
Automization: Principles for Pro- Integrated Theory: Connecting
moting Fluency Within a Communi- Modules, Vol. 24, No. 4, pp. 697-716.
cative Framework”: The Authors Hayes, Elisabeth, Hispanic Adults and
Respond, Vol. 23, No. 2, pp. 363-366. ESL Programs: Barriers to Participa-
Gathercole, Virginia C., Comments on tion, Vol. 23, No. 1, pp. 47-63.
“Some Myths You May Have Heard Hayward, Malcolm, Evaluations of
About First Language Acquisition”: Essay Prompts by Nonnative Speak-
The Author Responds, Vol. 23, No. ers of English, Vol. 24, No. 4,
1, pp. 160-162. pp. 753-756.
Genesee, Fred, Snow, Marguerite Ann,
and Met, Myriam, A Conceptual
Hoffman, Paul, and Yule, George,
Framework for the Integration of
Predicting Success for International
Language and Content in Second/
Teaching Assistants in a U.S. Univer-
Foreign Language Instruction, Vol.
sity, Vol. 24, No. 2, pp. 227-243.
~, No. 2, pp. 201-217. Holland, C. L., Kuehn, Phyllis A., and
Gilzow, Douglas F., and Ranard, Stanwyck, Douglas J., Attitudes
Donald A., Comments on Alien Toward “Cheating” Behaviors in the
Winds: The Reeducation of Ameri- ESL Classroom, Vol. 24, No. 2, pp.
ca’s Indochinese Refugees and the 313-317.
Book Review: Two Readers React, Holton, Christine, Review of Literacy
Vol. %, No. 3, pp. 529-541. Through Literature (Terry D. John-
Goldstein, Lynn, and Conrad, Susan, son and Daphne R. Louis), Vol. 24,
Student Input and Negotiation of No. 4, pp. 730-731.
Meaning in ESL Writing Conferen-
ces, Vol. 24, No. 3, pp. 443-460. Isogawa, Noriko, Review of L a n -
Gomez-Sanford, Charo, Review of guages and Children—Making the
Women’s Kit, Vol. 23, No. 2, pp. 334- Match (Helena Anderson Curtain
335. and Carol Ann Pesola), Vol. 24, No.
Granelli, Teresa, Review of Corwersa- 3, pp. 509-510.
tion Gambits: Real English Conver-
sational Practices (Eric Keller and Johnson, Valdon L., Review of T h e
Sylvia T. Warner), Vol. 24, No. 3, Alphabet Effect: The Impact of the
pp. 510-511. Phonetic Alphabet on the Develop-
Greenberg, Cindy, Review of Pro- ment of Western Civilization (Rob-
nouncing American English (Ger- ert K. Logan), Vol. 23, No. 1, pp.
trude F. Orion), Vol. 23, No. 3, pp. 140-141.
534-535. Jonz, Jon, Another Turn in the Conver-
sation: What Does Cloze Measure?
Hall, Chris, Managing the Complexity
Vol. 24, No. 1, pp. 61-63.
of Revising Across Languages, Vol.
24, No. 1, pp. 43-60. Judd, Elliot L., Review of English
Hamp-Lyons, Liz, Review of Recent Firsthand: Expanding Communica-
Publications on Statistics, Language tive Language Skills (Marc Hel-
Testing, and Quantitative Research gesen, Steven Brown, and Thomas
Methods I, Vol. 23, No. 1, pp. 127- Mandeville), Vol. 23, No. 1, pp. 134-
132. 135.
Hamp-Lyons, Liz, Review of Recent Judd, Elliot L., Review of Words at
Publications on Statistics, Language Work: Vocabulary Building Through
Testing, and Quantitative Research Reading (Betty Sobel and Susan
Methods II, Vol. 24, No. 2, pp. 293- Bookman), Vol. 24, No. 3, pp. 512-
300. 514.
794 TESOL QUARTERLY

Kaplan, Janet, Review of Spaghetti Lai, Su-Kuei, Review of E S L f o r
Forever: A Low-Intermediate Action: Problem Posing at Work
Reader in English (Jean W. Bodman (Elsa Roberts Auerbach and Nina
and Judith B. McKoy), Vol. 23, No. Wallerstein), Vol. 23, No. 2, pp. 321-
2, pp. 325-326. 322.
Kaye, Alan S., Review of Phonetic Leki, Ilona, Review of Gramrnur Prac-
Symbol Guide (Geoffrey K. Pullum tice Activities: A Practical Guide for
and William A. Laduslaw), Vol. 23, Teachers (Penny Ur), Vol. 23, No. 4,
No. 1, pp. 143-144. pp. 666-689.
Kennedy, Laura, Yule, George, and Leki, Bona, Review of Literature in the
Wetzel, Susa~, Listening Perception Language Classroom: A Resource
Accuracy of ESL Learners as a Book of Ideas and Activities (Joanne
Variable Function of Speaker Ll, Collie and Stephen Slater), Vol. 24,
VOI. ~, No. 3, pp. 519-523. No. 2, pp. 305-306.
Kessler, Carolyn, Review of Linguistic Liao, Fang-Lian, Review of Forestville
and Cultural Influences on Learning Tales: International Folk Stories
Mathematics (Rodney R. Cocking (Aaron Berman), Vol. 24, No. 4,
and Jose P. Mestre, Eds. ), Vol. 23, pp. 734-735.
No. 2, pp. 317-320. Liberto, Joseph C., Carrell, Patricia L.,
Kim, Haeyeon, Review of English and Pharis, Becky G., Metacognitive
Structure in Focus, Book Two (Polly Strategy Training for ESL Reading,
Davis), Vol. 23, No. 4, pp. 691-692. vol. ~, No. 4, pp. 647-678.
Kim, Hyun Sook, Review of English Long, Michael H., The Least a Second
Grammar: An Outline (Rodney Hud- Language Acquisition Tlseory Needs
dleston), Vol. 23, No. 4, p. 693. to Explain, Vol. 24, No. 4, pp. 649-
666.
Kodimer, Melinda, Review of Literary
Lynch, Brian K., Comments on “A
Portraits: An Anthology of Modern
Context-Adaptive Model for Pro-
American Prose and Poetry for
gram Evaluation”: The Author Re-
Students of English (J. Donnie
sponds, Vol. 24, No. 4, pp. 764-767.
Snyder), Vol. 24, No. 4, pp. 736-739.
Lynch, Brian K., A Context-Adaptive
Kroll, Barbara, Carson, Joan Eister-
Model for Program Evaluation, Vol.
hold, Carrell, Patricia, Silberstein,
24, No. 1, pp. 2.3-42.
Sandra, and Kuehn, Phyllis A., Read-
ing-Writing Relationships in First Master, Peter, Review of Techniques
and Second Language, Vol. 24, No. and Resources in Teaching Gram-
2, pp. 245-266. mar (Marianne Celce-Murcia and
Kuehn, Phyllis, Stanwyck, Douglas J., Sharon Hines), Vol. 24, No. 3, pp.
and Holland, C. L., Attitudes To- 514-515.
ward “Cheating” Behaviors in the Master, Peter, Teaching English Arti-
ESL Classroom, Vol. 24, No. 2, pp. cles as a Binary System, Vol. 24, No.
313-317. 3, pp. 461-478.
Kuehn, Phyllis A., Carson, Joan Eister- Matthews, Candace, Review of All
hold, Carrell, Patricia, Silberstein, Sides of the Issue: Activities for
Sandra, and Kroll, Barbara, Reading- Cooperative Jigsaw Groups (Eliza-
Writing Relationships in First and beth Coehlo, Lise Winer, and Judy
Second Language, Vol. 24, No. 2, Winn-Bell Olsen), Vol. 23, No. 3, pp.
P P. ~5-966. 533-534.
Kunnan, Antony John, DIF in Native McGrail, Loren, Review of Need I Say
Language and Gender Groups in an More: A Literary Magazine of Adult
ESL Placement Examination, Vol. Student Writings, Vol. 23, No. 2, pp.
~, No. 4, pp. 741-746. 331-332.

McGrail, Loren, Review of The Story Nash, Andrea, Review of The House
of Ana, La Historia de Ana (Ely on Mango Street (Sandra Cisneros),
Patricia Martinez Vasquez), Vol. 23, Vol. ~, No. 2, pp. 326-327.
No. 2, pp. 332-333. Nash, Andrea, Review of In Print:
McKay, Sandra Lee, and Freedman, Beginning Literacy Through Cultu-
Sarah Warshauer, Language Minor- ral Awareness (Lynellyn D. Long
ity Education in Great Britain: A and Janet Spiegel-Podnecky), Vol.
Challenge to Current U.S. Policy, 23, No. 2, pp. 329-330.
Vol. ~, No. 3, pp. 385-405. Nelson, Gayle, and Schmid, Thomas,
McLaughlin, Barry, “Conscious” versus ESL Reading Schema Theory and
“Unconscious” Learning, Vol. 24, Standardized Tests, Vol. 23, No. 3,
No. 4, pp. 617-634. pp. 539-543.
Meinke, Mark W., Comments on “A Nunan, David, Toward a Collabora-
Context-Adaptive Model of Pro- tive Approach to Curriculum Devel-
gram Evaluation”: A Reader Reacts opment: A Case Study, Vol. 23, No.
to Lynch, Vol. 24, No. 4, pp. 759-760. 1, pp. 9-25.
Meskill, Carla, Review of Computers
in the Language Classroom (Robert Olsen, Roger E. W-B, A Survey of
M. Hertz), Vol. 24, No. 2, pp. 308- Limited English Proficient Student
309. Enrollments and Identification
Meskill, Carla, Review of Language, Criteria, Vol. 23, No. 3, pp. 469-488.
Learners and Computers (John Hig- Oster, Judith, Seeing With Different
gins), Vol. 24, No. 2, pp. 307-308. Eyes: Another View of Literature in
Messerschmitt, Dorothy S., Review of the ESL Class, Vol. 23, No. 1, pp. 85-
Making It Happen: Interaction in the 103.
Second Language Chssroom (Patri-
cia A. Richard-Amato), Vol. 24, No. Pallen, Fay, Review of Whole Lan-
3, pp. 507-508. guage Strategies for ESL Students
Met, Myriam, Snow, Marguerite Ann, (Gail Heald-Taylor), Vol. 24, No. 2,
and Genesee, Fred, A Conceptual pp. 306-307.
Framework for the Integration of Peirce, Bronwyn Norton, Comments
Language and Content in Second/ on “Toward a Pedagogy of Possibil-
Foreign Language Instruction, Vol. ity in the Teaching of English Inter-
23, No. 2, pp. 201-217. nationally: People’s English in South
Milk, Robert D., Preparing ESL and Africa”: The Author Responds, Vol.
Bilingual Teachers for Changing 24, No. 1> pp. 105-112.
Roles: Immersion for Teachers of
Peirce, Bronwyn Norton, Toward a
LEP Children. Vol. 24, No. 3, PP.
Pedagogy of Possibility in the
407-426. ‘ --
Teaching of English Internationally:
Mishra, Shitala, Dunkel, Patricia, and
People’s English in South Africa,
Berliner, David, Effects of Note
Vol. 23, No. 3, pp. 401-420.
Taking, Memory, and Language
Proficiency on Lecture Learning for Pennington, Martha C., and Young,
Native and Nonnative Speakers of Aileen L., Approaches to Faculty
English, Vol. 23, No. 3, pp. 543-549. Evaluation for ESL, Vol. 23, No. 4,
pp. 619-646.
Nash, Andrea, Review of Don’t Be Pennington, Martha C., Comments on
Afraid, Gringo: A Honduran Woman “Approaches to Faculty Evaluation
Speaks From the Heart: The Story of for ESL: Response to Pennycook:
Elvis Alvarado (Medea Benjamin, The Political Economy of Informa-
Trans. and Ed.), Vol. 23, No. 2, pp. tion in TESOL, Vol. 24, No. 3, pp.
324-325. 559-568.
796 TESOL QUARTERLY

Pennycook, Alastair, The Concept of Rhum, Madeline, Review of T h e
Method, Interested Knowledge, and Mango Tree: Stories Told and Re-
the Politics of Language Teaching, told by Children in the Cambridge
Vol. ~, No. 4, pp. 589-618. Public Schools, Vol. ~, No. 2, pp.
Pennycook, Alastair, Commentary on 330-331.
“Approaches to Faculty Evaluation Rigg, Pat, Review of The Learner-
for ESL”: A Reader Reacts, Vol. 24, Centred Curriculum: A Study in
No. 3, pp. 555-558. Second Language Teaching, Vol. .24,
Pharis, Becky G., Carrell, Patricia L.,
No. 1, pp. 95-98.
and Liberto, Joseph C., Metacogni-
tive Strategy Training for ESL Read- Rugger, Mildred S., Review of Talking
ing, Vol. 23, No. 4, pp. 647-678. about Writing: A Guide for Tutor
Polak, Jeanne, and Krashen, Stephen, and Teacher Conferences (B. L.
comments on “Do We Need to Clark), Vol. 23, No, 3, pp. 535-536.
Teach Spelling?”: The Authors Re-
spond, Vol. 23, No. 1, pp. 164-167. Santos, Terry, Replication in Applied
Prabhu, N. S., Comments on “Attention Linguistics Research, Vol. 23, No. 4,
to Form or Meaning? Error Treat- pp. 699-702.
ment in the Bangalore Project”: A Sasaki, Yoshinori, A Logical Difficulty
Reader Reacts, Vol. 24, No. 1, pp. of the Parameter Setting Model, Vol.
112-115. 24, No. 4, pp. 767-770.
Prabhu, N. S., There is No Best Meth- Sato, Charlene J., A Nonstandard
od—why?, Vol. 24, No. 2, pp. 161- Approach to Standard English, Vol.
176. ~, No. 2, pp. 259-282.
Prater, Doris L., and Bermudez, An- Schmid, Thomas, and Nelson, Gayle,
drea B., Using Brainstorming and ESL Reading: Schema Theory and
Clustering with LEP Writers to Standardized Tests, Vol. 2.3, No. 3,
Develop Elaboration Skills, Vol. 24, pp. 539-543.
No. 3, pp. 523-528. Schoener, Wendy, Review of Reading
Putti, Paola Sinigaglia, Review of Skills for the Social Sciences (Lou-
Longman English Grammar (Louis ann Haarman, Patrick Leech, and
G. Alexander), Vol. 24, No. 2, pp. &6~~87Murray), Vol. 23, No. 4, pp.
309-311.
Schumann, John H., Extending the
Raimes, Ann, The TOEFL Test of Scope of the Acculturation/Pidgini-
Written English: Causes for Con- zation Model to Include Cognition,
cern, Vol. 24, No. 3, pp. 427-442.
Vol. 24, No. 4, pp. 667-683.
Ramsay, Janet K., The Bilingual/ESOL
Segalowitz, Norman, and Gatbonton,
Curriculum Framework of the
Denver Public Schools, Vol. 2.3, No. Elizabeth, Comments on “Creative
1, pp. 147-152. Automization: Principles for Pro-
Ranard, Donald A., and Gilzow, Doug- moting Fluency Within a Commu-
las F., Comments on Alien Winds: nicative Framework”: The Authors
The Reeducation of America’s Indo- Respond, Vol. 23, No. 2, pp. 363-366.
chinese Refugees and the Book Selipsky, Elaine, Review of The Origin
Review: Two Readers React, Vol. of Writing (Roy Harris), Vol. 24, No.
24, No. 3, pp. 529-541. 2, pp. 311-312.
Reid, Joy, The Dirtv Laundrv of ESL Sheehan, Annette C., Review of Long-
Survey Research, ‘Vol. 24, No. 2, pp. man American Structural Readers,
323-338. vol. %, No. + pp. 733-734.
Repath, Lyn M., Review of Outsiders: Shirai, Yasuhiro, Fantuzzi, Cheryl, and
American Short Stories for Students Hatch, Evelyn, The Need for an
of ESL (Jean S. Mullen), Vol. 24, No. Integrated Theory: Connecting
4> pp. 735-736. Modules, Vol. 24, No. 4, pp. 697-716.

Siegel, Andrew F., Research Issues: Strother, Judith B., and Alford, Randall
Multiple t Tests: Some Practical L., Attitudes of Native and Nonna-
Consfierations, Vol. 24, No. 4, pp. tive Speakers Towards Selected
773-775. Accents of U.S. English, Vol. 2.3, No.
Silberstein, Sandra, Carson, Joan Eis- 3, pp. 479-495.
terhold, Carrell, Patricia, Kroll, Sun, Yilin, and Burnaby, Barbara,
Barbara, and Kuehn, Phyllis A., Chinese Teachers’ Views of Western
Reading-Writing Relationships in Language Teaching Context In-
First and Second Language, Vol. 24, forms Paradigms, Vol. 23, No. 2, pp.
No. 2, pp. 245-266. 219-238.
Snow, Marguerite Ann, Met, Myriam, Swales, John M., Comments on “A
and Genesee, Fred, A Conceptual Context-Adaptive Model of Pro-
Framework for the Integration of gram Evaluation”: A Reader Reacts,
Language and Content in Second/
Vol. 24, No. 4, pp. 761-763.
Foreign Language Instruction, Vol.
~, No. 2, pp. 201-217. Swisher, M. Virginia, The Language-
Snow, Marguerite Ann, Review of Learning Situation of Deaf Students,
Learning Through Two Languages: Vol. ~, No. 2, pp. 239-257.
Studies of Immersion and Bilingual
Education (Fred Genesee), Vol. 23, Tawake, Sandra, Review of The Pro-
No. 1, pp. 137-138. cess of Writing (Mary Stewart), Vol.
Sokolik, Margaret E., Comments on 23, No. 4, pp. 687-688.
“Bridging the Gap: A General The- Terdal, Marjorie, Review of Language
ory of Second Language Learning”: Diversity: Problem or Resource?
A Reader Reacts, Vol. 23, No. 2, pp. (Sandra Lee McKay and Sau-Ling
355-359. Cynthia Wong), Vol. 23, No. 4, pp.
Sokolik, M. E., Learning Without Rules: 685-686.
PDP and a Resolution of the Adult Thompson, Roger M., Writing-Profi-
Language Learning Paradox, Vol. ciency Tests and Remediation: Some
24, No. 4, pp. 685-696. Cultural Differences, Vol. 24, No. 1,
Spolsky, Bernard, Comments on pp. 99-102.
“Bridging the Gap: A General The- Thonus, Teresa, Review of The Long-
ory of Second Language Learning”: man Preparation Course for the
The Author Responds, Vol. 23, No. TOEFL (Deborah Phillips), Vol. 24,
2, pp. 355-359. No. 3, pp. 516-517.
Spolsky, Bernard, Introduction to a
Colloquium: The Scope and Form of Tinkham, Thomas, Rote Learning,
a Theory of Second Language Attitudes, and Abilities: A Compari-
Learning, Vol. 24, No. 4, pp. 609-616. son of Japanese and American Stu-
dents, Vol. 23, No. 4, pp. 695-698.
Spycher, Ruth, Review of Breaking
Rules: Generating and Exploring Tollefson, James W., Comments on
Alternatives (John F. Fanselow), Alien Winds: The Reeducation of
vol. 24, No. 2, pp. 301-303. America’s Indochinese Refugees,
Stanwyck, Douglas J., Kuehn, Phyllis, The Author Responds to Ranard and
and Holland, C. L., Attitudes To- Gilzow: The Economics and Ideol-
ward “Cheating” Behaviors in the ogy of Overseas Refugee Education,
ESL Classroom, Vol. 24, No. 2, pp. Voi. 24, No. 3, pp. 543-555.
313-317. Tollefson, James W., Educating for
Stempleski, Susan, Review of The Employment in Programs for South-
Media: Catalysts for Communicative east Asian Refugees: A Review of
Language Learning (Joyce Pen- Research, Vol. 23, No. 2, pp. 337-
field), Vol. 23, No. 1, pp. 142-143. 343.
798 TESOL QUARTERLY

Tollefson, James W., Review of Wilson, Karen Marie, Review of Focus:
Broken Promties: Reading h.struc- An ESL Grammar (Barbara Robin-
tion in Twentieth Century America son), Vol. 23, No. 4, pp. 690-691.
(Patrick Shannon), Vol. 23, No. 4,
pp. 679-681. Yap, Foong Ha, Review of The Inward
Tumposky, Nancy Rennau, Review of Ear: Poetry in the Language Class-
Contemporary World Issues: An room (Alan Maley and Alan Duff),
Interactive Approach to Reading Vol. 24, No. 4, pp. 731-732,
and Wtiting (Richard L. Light and Yasuhiro, Shirai, Fantuzzi, Cheryl, and
Fan Lan-Ying), Vol. 23, No. 4, pp. Hatch, Evelyn, The Need for an
682-684. Integrated Theory: Connecting
Modules, Vol. 24, No. 4, pp. 697-716.
van Lier, Leo, Reeling, Writhing, Young, Aileen L., and Pennington,
Drawling, Stretching, and Fainting Martha C., Approaches to Faculty
in Coils: oral Proficiency Interviews Evaluation for ESL, Vol. 23, No. 4,
as Conversation, Vol. 23, No. 3, pp. pp. 619-646.
489-508. Yule, George, and Hoffman, Paul,
Vann, Roberta J., and Abraham, Ro- Predicting Success for International
berta G., Strategies of Unsuccessful Teaching Assistants in a U.S. Univer-
Language Learners, Vol. 24, No. 2, sity, Vol. 24, No. 2, pp. 227-243.
pp. 177-198. Yule, George, Wetzel, Susan, and
Kennedy, Laura, Listening Percep-
Wetzel, Susan, Yule, George, and tion Accuracy of ESL Learners as a
Kennedy, Laura, Listening Percep- Variable’ Function of Speaker Ll,
tion Accuracy of ESL Learners as a vol. %, No. 3, pp. 519-523.
Variable Function of Speaker Ll,
vol. %, No. 3, pp. 519-523. Zarzour, Odile, Review of Frontiers:
Williams, Jessica, Review of Recent An Active Introduction to English
Publications in Sociolinguistics, Vol. Grammur (John Schmidt and Terry
~, No. 3, pp. 497-500. Simon), Vol. 23, No. 4, pp. 689-690.

Index of Books Reviewed
Alexander, Louis G., Longman English Brown, James Dean, Understanding

Grammar (Paola Sinigaglia Putti), Research in S e c o n d L a n g u a g e
Vol. 24, No. 2, pp. 309-311. Learning: A Teacher’s Guide to
Allwright, Dick, Observation in the Statistics and Research Design (Liz
Language Classroom (Courtney B. Hamp-Lyons), Vol. 23, No. 1, pp.
Cazden), Vol. 24, No. 4, pp. 717-724. 127-132.
Auerbach, Elsa Roberts, and Waller- Butler, Christopher, Statistics in Lin-
stein, Nina, ESL for Action: Problem guistics (Liz Hamp-Lyons), Vol. 23,
Posing at Work (Su-Kuei Lai), Vol. No. 1, pp. 127-132.
%, No. 2, pp. 321-322. Byrd, Patricia, Constantinides, Janet
Avery, Peter, and Ehrlich, Susan, C., and Pennington, Martha C., The
(Eds.), The Teaching of Pronuncia- Foreign Teaching Assistant’s Manual
tion: An Introduction for Teachers of (Wanda S. Fox), Vol. 24, No. 3, pp.
English as a Second Language 508-509.
(Donna M. Brinton), Vol. 23, No. 1,
pp. 135-136. Celce-Murcia, Marianne, and Hines,
Sharon, Techniques and Resources
Bachman, Lyle, Fundamental Consid- in Teaching Grammar (Peter Mas-
erations in Language Testing (Liz ter), Vol. 24, No. 3, pp. 514-515.
Hamp-Lyons), Vol. 24, No. 2, pp.
293-300. Chamot, Anna Uhl, America: The
Early Years; America: After inde-
Barndt, Deborah (Coordinator), Eng- pendence; Teacher’s Guide (Lan-
lish at Work: A Tool Kit for Teachers guage Development Through Con-
(Elsa Auerbach), Vol. 23, No. 2, pp. tent) (Joan Dungey), Vol. 23, No. 1,
322-323. pp. 133-134.
Barndt, Deborah, Cristall, Ferne, and Chaudron, Craig, Second Language
marine, dian, Getting There: Pro- Classrooms: Research on Teaching
ducing Photostories With Immigrant and Learning (Courtney B. Cazden),
W o m e n (Elsa Auerbach), Vol. 23, vol. 24, No. 4, pp. 717-724.
No. 2, pp. 323-324.
Cisneros, Sandra, The House on
Benjamin, Medea (Trans. and Ed.),
Mango Street (Andrea Nash), Vol.
Don’t Be Afraid, Gringo: A Hondu-
2.3, No. 2, pp. 326-327.
ran Woman Speaks From the Heart:
The Story of Elvis Alvarado (Andrea Clark, B. L., Talking about Writing: A
Nash), Vol. 2.3, No. 2, pp. 324-325. Guide for Tutor and Teacher Con-
Berman, Aaron, Forestville Tales: ferences (Mildred S. Rugger), Vol.
International Folk Stories (Fang- 23, No. 3, pp. 535-536.
Lian Liao), Vol. 24, No. 4, pp. 734- Cocking, Rodney R., and Mestre, Jose
735. P. (Eds.), Linguistic and Cultural
Bodman, Jean W., and McKoy, Judith Influences on Learning Mathematics
B., Spaghetti Forever: A Low-inter- (Carolyn Kessler), Vol. 23, No. 2, pp.
mediate Reader in English (Janet 317-320.
Kaplan), Vol. 23, No. 2, pp. 325-326. Coehlo, Elizabeth, Winer, Lise, and
Brindley, Geoff, Assessing Achieve- Olsen, Judy Winn-Bell, All Sides of
ment in the Learner-C entred Cur- the Issue: Activities for Cooperative
riculum (Liz Hamp-Lyons), Vol. 24, Jigsaw Groups (Candace Matthews),
No. 2, pp. 293-300. Vol. 23, No. 3, pp. 533-534.
800 TESOL QUARTERLY

Collie, Joanne, and Slater, Stephen, Hamp-Lyons, Liz, Newbury House
Literature in the Language Class- TOEFL Preparation Kit: Preparing
room: A Resource Book of Ideas and for the Test of Written English
Activities (Ilona Leki), Vol. 24, No. (Marsha Bensoussan), Vol. 24, No. 3,
2, pp. 305-306. pp. 501-506.
Connor, Ulla, and Kaplan, Robert B . Harris, Roy, The Origin of Writing
(Eds.), Writing Across Languages: (Elaine Selipsky), Vol. 24, No. 2, pp.
Anal@s of L2 Text ( M e l v i n R . 311-312.
Andr~de, Jr.), Vol. 24, ‘No. 4, pp. Heald-Taylor, Gail, Whole Language
724-727. Strategies for ESL Students (Fay
Crookall, David, Greenblat, Cathy, Pallen), Vol. 24, No. 2, pp. 306-307.
Coote, Alan, Klabbers, Jan, and Helgesen, Marc, Brown, Steven, and
Watson, D.R. (Eds. ), Simulation- Mandeville, Thomas, English First-
Gaming in the Late 1980s: Proceed- hand: Expanding Communicative
ings of the International Simuihtion Language Skills (Elliot L. Judd),
and Gaming Association’s 17th inter- vol. 23, No. 1, pp. 134-135.
national Conference (Patricia A. Henning, Grant, A Guide to Language
Dunkel), Vol. 23, No. 1, pp. 136-137. Testing: Development, Evaluation,
Curtain, Helena Anderson, and Pesola, Research (Liz Hamp-Lyons), Vol.
Carol Ann, Languages and Child- ~, No. 1, pp. 127-132.
ren—Making the Match (Noriko Hertz, Robert, Computers in the Lan-
Isogawa), Vol. 24, No. 3, pp. 509- guage Classroom (Carla Meskill),
510. Vol. 24, No. 2, pp. 308-309.
Higgins, James, and Ross, Joan, South-
Davis, Polly, English Structure in
east Asians: A New Beginning in
Focus, Book Two (Kim Haeyeon),
Lowell (Cynthia Cook), Vol. 23, No.
Vol. ~, No. 4, pp. 691-692.
2, pp. 328-329.
Edelsky, Carole, Writing in a Bilingual Higgins, John, Language, Learners and
Program: Habia una Vez (Maria Computers (Carla Meskill), Vol. 24,
Brisk), Vol. 23, No. 3, pp. 536-537. No. 2, pp. 307-308.
Huddleston, Rodney, English Gram-
Ellowitch, Azi, Tell Me A b o u t It:
Reading and Language Activities mar: An Outline (Hyun Sook Kim),
Around Multi-Cultural Issues Based Vol. ,%> No. 4, pp. 693.
on an Oral History Approach (Elsa Hughes, Arthur, Testing for Language
Auerbach), Vol. 23, No. 2, pp. 327- Teachers (Liz Hamp-Lyons), Vol.
328. 24, No. 2, pp. 293-300.
Fanselow, John F., Breaking Rules: Johnson, Terry D., and Louis, Daphne
Generating and Exploring Alterna- R., Literacy Through Literature
tives in Language Teaching (Ruth (Christine Holton), Vol. 24, No. 4,
Spycher), Vol. 24, No. 2, pp. 301- pp. 730-731.
303.
Genesee, Fred, Learning Through Two Keller, Eric, and Warner, Sylvia T.,
Languages: Studies of Immersion Conversation Gambits: Real English
and Bilingual Education (Marguerite Conversation Practices (Teresa Gra-
Ann Snow), Vol. 23, No. 1, pp. 137- nelli), Vol. 24, No. 3, pp. 510-511.
138. Kennedy, Daniel B., Kenyon, Dorry
Mann, and Matthiesen, Steven J.,
Haarman, Louann, Leech, Patrick, and Newbury House TOEFL Prepara-
Murray, Janey, Reading Skills for the tion Kit: Preparing for the TOEFL
Social Sciences (Wendy Schoener), (Marsha Bensoussan), Vol. 24, No. 3,
Vol. 23, No. 4, pp. 686-687. pp. 501-506.

Leki, Ilona, Academic Writing: Tech- Meyer, Charles F., A Linguistic Study
niques and Tasks (Mark Dunning), of American Punctuation (K. Scott
Vol. ~, No. 1, pp. 91-95. Ferguson), Vol. 24, No. 3, pp. 515-
Light, Richard L., and Lan-Ying, Fan, 516.
Contemporary World Issues: A n Mullen, Jean S., Outsiders: American
Interactive Approach to Reading Short Stories for Students of E S L
and Writing (Nancy Rennau Tum- (Lyn M. Repath), Vol. 24, No. 4,
posky), Vol. ~, No. 4, pp. 682-684. pp. 735-736.
Lindemann, Erika (Ed.), Longman
Bibliography of Composition and Need 1 Say More: A Literary Magazine
of Adult Student Writings (Loren
Rhetoric: 1984-1985 (Michelle McGrail), Vol. 23, No. 2, pp. 331-
Chan), Vol. 2.3, No. 1, pp. 138-140. 332.
Lindemann, Erika (Ed.), Longman
Nunan, David, The Learner-Centred
Bibliography of Composition and Curriculum: A Study in Second Lan-
Rhetoric: 1986 (Michelle Chan), Vol. guage Teaching (Pat Rigg), Vol. 24,
23, No. 1, pp. 138-140. No. 1, pp. 95-98.
Logan, Robert K., The Alphabet Ef-
fect: The Impact of the Phonetic Orion, Gertrude F., Pronouncing Amer-
Alphabet on the Development of ican English (Cindy Greenberg), Vol.
Western Civilization (Valdon L. 23, No. 3, pp. 534-535.
Johnson), Vol. 2.3, No. 1, pp. 140-141. Pen field, Joyce, The Media: Catalysts
Long, Lynellyn D., and Spiegel-Pod- for Communicative Language Learn-
necky, Janet, In Print: Beginning ing (Susan Stempleski), Vol. 23, No.
Literacy Through Cultural Aware- 1, pp. 142-143.
ness (Andrea Nash), Vol. 23, No. 2, Phillips, Deborah, Longman Prepara-
pp. 329-330. tion Course for the TOEFL (Theresa
Longman American Structural Readers Thonus), Vol. 24, No. 3, pp. 516-517.
(Annette C. Sheehan), Vol. 24, No. 4, Preston, Dennis, Sociolinguistics and
pp. 733-734. Second Language Acquisition (Jes-
Lutz, William, Doublespeak (Vincent sica Williams), Vol. 24, No. 3, pp.
G. Barnes), Vol. 24, No. 3, pp. 517- 497-500.
518. Pullum, Geoffrey K., and Laduslaw,
William A., Phonetic Symbol Guide
Malamah-Thomas, Ann, Classroom (Alan S. Kaye), Vol. 23, No. 1, pp.
Interaction (Ali A. Aghbar), Vol. 23, 143-144.
No. 1, pp. 141-142.
Maley, Alan, and Duff, Alan, T h e Richard-Amato, Patricia A., Making It
Inward Ear: Poetry in the Language Happen: Interaction in the Second
Classroom (Foong Ha Yap), Vol. 24, Language Classroom (Dorothy S.
No. 4, pp. 731-732. Messerschmitt), Vol. 2A, No. 3, pp.
507-508.
The Mango Tree: Stories Told and
Retold by Children in the Cam- Robinson, Barbara, Focus: An ESL
Grammar (Karen Marie Wilson),
bridge Public Schools (Madeline
Vol. 23, No. 4, pp. 690-691.
Rhum), Vol. 23, No. 2, pp. 330-331.
McKay, Sandra, and Petitt, Dorothy, Schmidt, John, and Simon, Terry,
At the Door: Selected Literature for Frontiers: An Active Introduction to
ESL Students (Martha Dunn), Vol. English Grammar (Odile Zarzour),
24, No. 4, pp. 736-737. Vol. ~, No. 4, pp. 689-690.
McKay, Sandra Lee, and Wong, Sau- Shannon, Patrick, Broken Promises:
Ling Cynthia, Language Diversity: Reading Instruction in Twentieth
Problem or Resource? (Marjorie Century America (James W. Tollef-
Terdal), Vol. 23, No. 4, pp. 685-686. son), Vol. 23, No. 4, pp. 679-681.
802 TESOL QUARTERLY

Snyder, J. Donnie, Literary Portraits: van Lier, Leo, The Classroom and the
An Anthology of Modem American Language Learner: Ethnography
Prose and Poetry for Students of and Second-Language Classroom
English (Melinda Kodimer), Vol. 24, Research (Courtney B. Cazden),
No. 4, pp. 738-739. Vol. 2,4, No. 4, pp. 717-724.
Vasquez, Ely Patricia Martinez, The
Sobel, Betty, and Bookman, Susan, Story of Ana, La Historia de A n a
Words at Work: Vocabulary Build- (Loren McGrail), Vol. 2.3, No. 2, pp.
ing Through Reading (Elliot L. 332-333.
Judd), Vol. 24, No. 3, pp. 512-514. Voices: New Writers for New Readers
Stewart, Mary, The Process of Writing (Elsa Auerbach), Vol. 23, No. 2, pp.
(Sandra Tawake), Vol. 23, No. 4, pp. 333-334.
687-688.
Weinstein, Nina, Crazy Idioms: A
Conversational Idiom Book (Eda
Tollefson, James W., Alien Winds: The Ashby), Vol. 24, No. 3, pp. 511-512.
Reeducation of America’s Indochi- Winthrow, Jean, Brooks, Gay, and
nese Refugees (Elsa Auerbach), Vol. Cummings, Martha Clark, Changes:
24, No. 1, pp. 85-91. Readings for ESL Writers @’rank
Beyer), Vol. 24, No. 43 pp. 737-738.
Underhill, Nit, Testing Spoken Lan- Wolfson, Nessa, Perspectives: Sociolin-
guage: A Handbook of Oral Testing guistics and TESOL (Jessica Willi-
Techniques (James Dean Brown), ams), Vol. 24, No. 3, pp. 497-500.
Vol. ~, No, 1, pp. 144-145. Women’s Kit (Charo Gomez-Sanford),
vol. 23, No. 2, pp. 334-335.
Ur, Penny, Grammar Practice Activi- Woods, Anthony, Fletcher, Paul, and
ties: A Practical Guide for Teachers Hughes, Arthur, Statistics in Lan-
(Ilona Leki), Vol. 23, No. 4, pp. 688- guage Studies (Liz Hamp-Lyons),
689. Vol. 23, No. 1, pp. 127-132.

Selected Topic Index
BILINGUALISM/BILINGUAL CONTRASTIVE STUDIES
EDUCATION Alford, Randall L., and Strother, Judith
McKay, Sandra Lee, and Freedman, B., Attitudes of Native and Nonna-
Sarah Warshauer, Language Minor- tive Speakers Towards Selected
ity Education in Great Britain: A Accents of U.S. English, Vol. 2.3, No.
Challenge to Current U.S. Policy, 3, pp. 479-495.
Vol. 24, No. 3, pp. 385-405. Burnaby, Barbara, and Sun, Yilin,
Milk, Robert D., Preparing ESL and Chinese Teachers’ Views of Western
Bilingual Teachers for Changing Language Teaching Context In-
Roles: Immersion for Teachers of forms Paradigms, Vol. 23, No. 2, pp.
LEP Children, Vol. 24, No. 3, pp. 219-2.38.
407-426. Kuehn, Phyllis, Stanwyck, Douglas J.,
Ramsay, Janet K., The Bilingual/ESOL and Holland, C. L., Attitudes To-
Curriculum Framework of the ward “Cheating” Behaviors in the
Denver Public Schools, Vol. 23, No. ESL Classroom, Vol. 24, No. 2, pp.
1, pp. 147-152. 313-317.
Tinkham, Thomas, Rote Learning,
CALL/COMPUTER APPLICATIONS Attitudes, and Abilities A Compari-
Chapelle, Carol, The Discourse of son of Japanese and American Stu-
Computer-Assisted Language Learn- dents, Vol. 23, No. 4, pp. 695-698.
ing: Toward a Context for Descrip- Thompson, Roger M., Writing-Profi-
tive Research, Vol. 24, No. 2, pp, ciency Tests and Remediation: Some
199-225. Cultural Differences, Vol. 24, No. 1,
pp. 99-102.
CL4SSROOM-CENTERED RESEARCH Yule, George, Wetzel, Susan, and
Beretta, Alan, Attention to Form or Kennedy, Laura, Listening Percep-
Meaning? Error Treatment in the tion Accuracy of ESL Learners as a
Bangalore Project, Vol. 23, No. 2, Variable J7unction of Speaker Ll,
pp. 283-303. Vol. ~, No. 3, pp. 519-523.
Cazden, Courtney B., Review of Re-
cent Publications on Classroom CURRICULUIWSYLLABUS DESIGN
Research, Vol. 24, No. 4, pp. 717- Benson, Malcolm J., The Academic
724. Listening Task: A Case Study, Vol.
Chapelle, Carol, The Discourse of 23, No. 3, pp. 421-445.
Computer-Assisted Language Learn- Cathcart, Ruth Larimer, Authentic
ing: Toward a Context for Descrip- Discourse and the Survival English
tive Research, Vol. 24, No. 2, pp. Curriculum, Vol. 23, No. 1, pp. 105-
199-222. 136.
Kuehn, Phyllis, Stanwyck, Douglas J., Nunan, David, Toward a Collabora-
and Holland, C. L., Attitudes To- tive Approach to Curriculum Devel-
ward “Cheating” Behaviors in the opment: A Case Study, Vol. 2.3, No.
ESL Classroom, Vol. 24, No. 2, pp. 1, pp. 9-2.5.
313-317. Peirce, Bronwyn Norton, Toward a
Vann, Roberta J., and Abraham, Ro- Pedagogy of Possibility in the
berta G., Strategies of Unsuccessful Teaching of English Internationally:
Language Learners, Vol. 24, No. 2, People’s English in South Africa,
pp. 177-198. Vol. ~, No. 3, pp. 401-420.
804 TESOL QUARTERLY

Ramsay, Janet K., The Bilingual/ESOL McKay, Sandra Lee, and Freedman,
Curriculum Framework of the Sarah Warshauer, Language Minor-
Denver Public Schools, Vol. 23, No. ity Education in Great Britain: A
1, pp. 147-152. Challenge to Current U.S. Policy,
Snow, Marguerite Ann, Met, Myriam, Vol. 24, No. 3, pp. 385-405.
and Genesee, Fred, A Conceptual Peirce, Bronwyn Norton, Toward a
Framework for the Integration of Pedagogy of Possibility in the Teach-
Language and Content in Second/ ing of English Internationally: Peo-
Foreign Language Instruction, Vol. ple’s English in South Africa, Vol. 23,
~, No. 2, pp. 201-217. No. 3, pp. 401-420.
LEXICON/LEXICOGRAPHY/
DISCOURSE ANALYSIS/ VOCABULARY TEACHING
PRAGMATICS
Benson, Morton, Differences Between
Cathcart, Ruth Larimer, Authentic American English and British Eng-
Discourse and the Survival English lish: A Challenge for TESOL, Vol.
Curriculum, Vol. 23, No. 1, pp. 105- $?3, No. 2, pp. 351-355.
136.
Dolly, Martha R., Adult ESL Students’ LISTENING
Management of Dialogue Journal
Conversation, Vol. 24, No. 2, pp. Benson, Malcolm J., The Academic
Listening Task: A Case Study, Vol.
317-320.
23, No. 3, pp. 421-445.
Goldstein, Lynn, and Conrad, Susan,
Student Input and Negotiation of Blau, Eileen, The Effect of Syntax,
Speed, and Pauses on Listening
Meaning in ESL Writing Conferen-
Comprehension, Vol. 24, No. 4, pp.
ces, Vol. 24, No. 3, pp. 443-460.
746-7%.
Dunkel, Patricia, Mishra, Shitala, and
ENGLISH AS A SECOND DIALECT Berliner. David. Effects of Note
Sato, Charlene J., A Nonstandard Ap- Taking,’ Memory, and Language
proach to Standard English, Vol. 23, Proficiency on Lecture Learning for
No. 2, pp. 259-262. Native and Nonnative Speakers of
English, Vol. 23, No. 3, pp. 543-549.
Yule, George, Wetzel, Susan, and
ENGLISH FOR SPECIAL PURPOSES/
Kennedy, Laura, Listening Percep-
EST/TECHNICAL WRITING
tion Accuracy of ESL Learners as a
Canseco, Grace, and Byrd, Patricia, Variable Function of Speaker Ll,
Writing Required in Graduate Vol. W> No. 3, pp. 519-523.
Vol. ~, No. 2, pp. 305-316. LITERATURE
Yule, George, and Hoffman, Paul, Pre- Brinton, Donna (Ed.). Book Notices on
dicting Success for International the Literature-Enriched Curriculum,
Teaching Assistants in a U.S. Univer- Vol. 24, No. 4, pp. 729-739.
sity, Vol. 24, No. 2, pp. 227-243. Oster, Judith, Seeing With Different
Eyes: Another View of Literature in
LANGUAGE PLANNING/POUCY the ESL Class, Vol. 23, No. 1, pp. 86-
Davies, Alan, Is International English 103.
an Interlanguage?, Vol. 23, No. 3,
pp. 447-467. METHODS/MATERIALS
Eastman, Carol M., What is the Role of Barrett, Mary T., The Secondary ESL
Language Planning in Post-Apart- Reading Course: Rationale and Im-
heid South Africa?, Vol. 24, No. 1, plementation, Vol. 23, No. 1, pp.
pp. 9-21. 152-156.

Bermudez, Andrea B., and Prater, Santos, Terry, Replication in Applied
Doris L., Using Brainstorming and Linguistics Research, Vol. 23, No. 4,
Clustering with LEP Writers to pp. 699-702.
Develop Elaboration Skills, Vol. 24,
No. 3, pp. 523-528. PROGRAM ADMINISTRATION
Burnaby, Barbara, and Sun, Yilin, AND EVALUATION
Chinese Teachers’ Views of Western Beretta, Alan, Attention to Form or
Language Teaching: Context In- Meaning? Error Treatment in the
forms Paradigms, Vol. 23, No. 2, pp. Bangalore Project, Vol. 23, No. 2,
219-238. pp. 283-303.
Carrell, Patricia L., Pharis, Becky G., Hayes, Elisabeth, Hispanic Adults and
and Liberto, Joseph C., Metacogni- ESL Programs: Barriers to Participa-
tive Strategy Training for ESL Read- tion, Vol. 23, No. 1, pp. 47-63.
ing, Vol. 23, No. 4, pp. 647-678. Lynch, Brian K., A Context-Adaptive
Dunkel, Patricia, Mishra, Shitala, and Model for Program Evaluation, Vol.
Berliner, David, Effects of Note 24, No. 1, pp. 23-42.
Taking, Memory, and Language Olsen, Roger E. W-B, A Survey of
Proficiency on Lecture Learning for Limited English Proficient Student
Native and Nonnative Speakers of Enrollments and Identification
English, Vol. 23, No. 3, pp. 543-549. Criteria, Vol. 23, No. 3, pp. 469-488.
Tollefson, James W., Educating for
Pennycook, Alastair, The Concept of
Employment in Programs for South-
Method, Interested Knowledge, and
east Asian Refugees: A Review of
the Politics of Language Teaching, Research, Vol. 23, No. 2, pp. 337-
Vol. ~, No. 4, pp. 589-618. 343.
Prabhu, N. S., There is No Best Meth- Yule, George, and Hoffman, Paul,
od—why?, Vol. 24, No. 2, pp. 161- Predicting Success for International
176. Teaching Assistants in a U.S. Univer-
Snow, Marguerite Ann, Met, Myriam, sity, Vol. 24, No. 2, pp. 227-243.
and Genesee, Fred, A Conceptual
Framework for the Integration of READING
Language and Content in Second/ Barrett, Mary T., The Secondary ESL
Foreign Language Instruction, Vol. Reading Course: Rationale and Im-
23, No. 2, pp. 201-217. plementation, Vol. 23, No. 1, pp.
Tinkham, Thomas, Rote Learning, 152-156.
Attitudes, and Abilities: A Compari- Brown, James Dean, Improving ESL
son of Japanese and American Stu- Placement Tests Using Two Per-
dents, Vol. 23, No. 4, pp. 695-698. spectives, Vol. 23, No. 1, pp. 65-83.
Carrell, Patricia L., Pharis, Becky G.,
PHONOLOGY/PRONUNCIATION and Liberto, Joseph C., Metacogni-
TEACHING tive Strategy Training for ESL Read-
Yule, George, Wetzel, Susan, and Ken- ing, Vol. 23, No. 4, pp. 647-678.
nedy, Laura, Listening Perception Carson, Joan Eisterhold, Carrell, Patri-
Accuracy of ESL Learners as a cia L., Silberstein, Sandra, Kroll,
Variable Function of Speaker Ll, Barbara, and Kuehn, Phyllis A.,
Vol. 24, No. 3, pp. 519-523. Reading-Writing Relationships in
First and Second Language, Vol. 24,
No. 2, pp. 245-266.
PROFES!3’ONAL STANDARDS Demel, Marjorie C., The Relationship
AND CONCERNS Between Overall Reading Compre-
Pennington, Martha C., and Young, hension and Comprehension of
Aileen L., Approached to Faculty Coreferential Ties for Second Lan-
Evaluation for ESL, Vol. $23, No. 4, guage Readers of English, Vol. 24,
pp. 619-646. No. 2, pp. 267-292.
808 TESOL QUARTERLY

Nelson, Gayle, and Schmid, Thomas, Santos, Terry, Replication in Applied
ESL Reading: Schema Theory and Linguistics Research, Vol. 23, No. 4,
Standardized Tests, Vol. 23, No. 3, pp. 699-702.
pp. 539-543. Siegel, Andrew F., Research Issues:
Multiple t Tests: Some Practical
Considerations, Vol. 24, No. 4, pp.
REFUGEE ISSUES/ 773-775.
ADULT EDUCATION
Auerbach, Elsa (Ed.). Book Notices on SECOND LANGUAGE ACQUISITION
Nontraditional Materials for Adult Bialystok, Ellen, The Competence of
ESL, Vol. ~, No. 2, pp. 321-335. Processing: Classifying Theories of
Auerbach, Elsa, Comments on Alien Second Language Acquisition, Vol.
Winds: The Reeducation of Ameri- 24, No. 4, pp. 635-648.
ca’s Indochinese Refugees and the Blau, Eileen K., The Effect of Syntax,
Book Review: The Reviewer Re- Speed, and Pauses on Listening
sponds, Vol. 24, No. 3, pp. 541-542. Comprehension, Vol. 24, No. 4, pp.
Hayes, Elisabeth, Hispanic Adults and 746-753.
ESL Programs: Barriers to Participa- Collier, Virginia P., How Long? A
tion, Vol. 2.3, No. 1, pp. 47-63. Synthesis of Research on Academic
Ranard, Donald A., and Gilzow, Doug- Achievement in a Second Language,
las F., Comments on Alien Winds: vol. ~, No. 3, pp. 509-531.
The Reeducation of America’s In- Hatch, Evelyn, Yasuhiro, Shirai, and
dochinese Refugees and the Book Fantuzzi, Cheryl, The Need for an
Review: Two Readers React, Vol. Integrated Theory: Connecting
24, No. 3, pp. 529-541. Modules, Vol. 24, No. 4, pp. 697-716.
Tollefson, James W., Comments on Long, Michael H., The Least a Second
Alien Winds: The Reeducation of Language Acquisition Theory Needs
America’s Indochinese Refugees, to Explain, Vol. 24, No. 4, pp. 649-
The Author Responds to Ranard and 666.
Gilzow: The Economics and Ideol- McLaughlin, Barry, “Conscious” versus
ogy of Overseas Refugee Education, “Unconscious” Learning, Vol. 24,
vol. %, No. 3, pp. 543-555. No. 4, pp. 617-634.
Tollefson, James W., Educating for Sasaki, Yoshinori, A Logical Difficulty
Employment in Programs for South- of the Parameter Setting Model, Vol.
east Asian Refugees: A Review of 24, No. 4, pp. 767-684.
Research, Vol. 23, No. 2, pp. 337-343. Schumann, John H., Extending the
Scope of the Acculturation/Pidgini-
zation Model to Include Cognition,
RESEARCH METHODOLOGY Vol. 24, No. 4, pp. 667-684.
Brown, James Dean, Research Issues: Sokolik, M. E., Learning Without Rules:
The Use of Multiple t Tests in Lan- PDP and a Resolution of the Adult
guage Research, Vol. 24, No. 4, pp. Language Learning Paradox, Vol.
770-773. 24, No. 4, pp. 685-696.
Hamp-Lyons, Liz, Review of Recent Spolsky, Bernard, Introduction to a
Publications on Statistics, Language Colloquium: The Scope and Form of
Testing, and Quantitative Research a Theory of Second Language
Methods I, Vol. 23, No. 1, pp. 127-132. Learning, Vol. 24, No. 4, pp. 609-616.
Hamp-Lyons, Liz, Review of Recent Swisher, M. Virginia, The Language-
Publications on Statistics, Language Learning Situation of Deaf Students,
Testing, and Quantitative Research Vol. 26, No, Z, Pp. 239-25’7.
Methods II, Vol. 24, No. 2, pp. 293-300. Vann, Roberta J., and Abraham, Ro-
Reid, Joy, The Dirty Laundry of ESL berta G., Strategies of Unsuccessful
Survey Research, Vol. 24, No. 2, pp. Language Learners, Vol. 24, No. 2,
323-338. pp. 177-198.

SOCIOLINGUISTICS/CULTURE Milk, Robert D., Preparing ESL and
Alford, Randall L., and Strother, Judith Bilingual Teachers for Changing
B., Attitudes of Native and Nonna- Roles: Immersion for Teachers of
tive Speakers Towards Selected LEP Children, Vol. 24, No. 3, pp.
Accents of U.S. English, Vol. 2.3, No. 407-426.
3, pp. 479-495. Pennington, Martha C., and Young,
Eastman, Carol M., What is the Role of Aileen L., Approaches to Faculty
Language Planning in Post-Apart- Evaluation for ESL, Vol. 23, No. 4,
heid South Africa?, Vol. 24, No. 1, pp. 619-646.
pp. 9-21. Prabhu, N. S., There is No Best Meth-
Kunnan, Antony John, DIF in Native od—Why?, Vol. 2.4, No. 2, pp. 161-
Language and Gender Groups in an 176.
ESL Placement Examination, Vol.
24, No. 4, pp. 741-746. TESTING
Sato, Charlene J., A Nonstandard
Approach to Standard English, Vol. Brown, James Dean, Improving ESL
%, No. 2, pp. 259-282. Placement Tests Using Two Per-
Swisher, M. Virginia, The Language- spectives, Vol. 23, No. 1, pp. 65-83.
Learning Situation of Deaf Students, Hayward, Malcolm, Evaluations of
Vol. 23, No. 2, pp. 239-257. Essay Prompts by Nonnative Speak-
Williams, Jessica, Review of Recent ers of English, Vol. M, No. 4, pp.
Publications in Sociolinguistics, Vol. 753-758.
24, No. 3, pp. 497-500. Hamp-Lyons, Liz, Review of Recent
Publications on Statistics, Language
Testing, and Quantitative Research
SPEAKING
Methods I, Vol. 23, No. 1, pp. 127-
van Lier, Leo, Reeling, Writhing, 132.
Drawling, Stretching, and Fainting
Hamp-Lyons, Liz, Review of Recent
in Coils: Oral Proficiency Interviews
Publications on Statistics, Language
as Conversation, Vol. 23, No. 3, pp.
Testing, and Quantitative Research
489-508.
Methods II, Vol. 24, No. 2, pp. 293-
300.
SYNTAX/GR4MMAR TEACHING Jonz, Jon, Another Turn in the Conver-
Master, Peter, Teaching English Arti- sation: What Does Cloze Measure?
cles as a Binary System, Vol. 24, No. Vol. 24, No. 1, pp. 61-83.
3, pp. 461-478. Kunnan, Antony John, DIF in Native
Language and Gender Groups in an
ESL Placement Examination, Vol.
TEACHER PREPARATION 24, No. 4, pp. 741-746.
Benson, Morton, Differences Between
Nelson, Gayle, and Schmid, Thomas,
American English and British Eng-
ESL Reading Schema Theory and
lish: A Challenge for TESOL, Vol.
Standardized Tests, Vol. 23, No. 3,
23, No. 2, pp. 351-355.
pp. 539-543.
Brinton, Donna, and Holten, Christine,
What Novice Teachers Focus On: Raimes, Ann, The TOEFL Test of
The Practicum in TESL, Vol. 2.3, No. Written English: Causes for Con-
2, pp. 343-350. cern, Vol. 24, No. 3, pp. 427-442.
Freeman, Donald, Teacher Training, van Lier, Leo, Reeling, Writhing,
Development, and Decision Making: Drawling, Stretching, and Fainting
A Model of Teaching and Related in Coils: Oral Proficiency Interviews
Strategies for Language Teacher as Conversation, Vol. 23, No. 3, pp.
Education, Vol. 2.3, No. 1, pp. 27-45. 489-508.
808 TESOL QUARTERLY

TEXT ANALYSIS Dolly, Martha R., Adult ESL Students’
Hall, Chris, Managing the Complexity Management of Dialogue Journal
of Revising Across Languages, Vol. Conversation, Vol. 24, No. 2, pp.
24, No. 1, pp. 43-60. 317-320.
Jonz, Jon, Another Turn in the Conver- Goldstein, Lynn, and Conrad, Susan,
sation: What Does Cloze Measure? Student Input and Negotiation of
Vol. ti, No. 1, pp. 61-83. Meaning in ESL Writing Conferen-
ces, Vol. 24, No. 3, pp. 443-460.
WRITING
Bermudez, Andrea B., and Prater, Hall, Chris, Managing the Complexity
Doris L., Using Brainstorming and of Revising Across Languages, Vol.
Clustering with LEP- Writers to .24, No. 1, pp. 43-60.
Develop Elaboration Skills, Vol. 24, Hayward, Malcolm, Evaluations of
No. 3, pp. 523-528. Essay Prompts by Nonnative Speak-
Canseco, Grace, and Byrd, Patricia, ers of English, Vol. 24, No. 4, pp.
Writing Required in Graduate 753-758.
Vol. ~, No. 2, pp. 305-316. Raimes, Ann, The TOEFL Test of
Carson, Joan Eisterhold, Carrell, Patri- Written English: Causes for Con-
cia L., Silberstein, Sandra, Kroll, cern, Vol. 24, No. 3, pp. 427-442.
Barbara, and Kuehn, Phyllis A., Thompson, Roger M., Writing-Profi-
Reading-Writing Relationships in ciency Tests and Remediation: Some
First and Second Language, Vol. 24, Cultural Differences, Vol. 24, No. 1,
No. 2, pp. 245-266. pp. 99-102.

Vol 24 4 PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Vol 24 4 PDF

Uploaded by

Copyright:

Available Formats

TABLE OF CONTENTS

TESOL QUARTERLY Volume 24, Number 4 ❑ Winter 1990

A Journal for Teachers of English to Speakers of Other Languages

Longman American Structural Readers (Annette C. Sheehan)

Information for Contributors 777

TESOL QUARTERLY 603

■ On behalf of the TESOL Quarterly, I am pleased to congratulate the

Bernard Spolsky’s interpretive essay examines the scope and form of a

606 TESOL QUARTERLY

This paper introduces a colloquium on the scope and form of a

The history of language teaching is sometimes written as though

610 TESOL QUARTERLY

If one looks at the complexity of the circumstances under which second

Understanding the conditions under which learning occurs does not

INTRODUCTION TO A COLLOQUIUM 611

612 TESOL QUARTERLY

INTRODUCTION TO A COLLOQUIUM 613

614 TESOL QUARTERLY

INTRODUCTION TO A COLLOQUIUM 615

616 TESOL QUARTERLY

“Conscious” versus “Unconscious”

This article examines the concept of consciousness in second lan-

In this article, I discuss the theoretical ramifications of a persistent

618 TESOL QUARTERLY

CONSCIOUS VERSUS UNCONSCIOUS LEARNING 619

620 TESOL QUARTERLY

The Reber/Dulany Debate

CONSCIOUS VERSUS UNCONSCIOUS LEARNING 621

622 TESOL QUARTERLY

The McClelland, Rumelhart/Pinker, Prince Debate

CONSCIOUS VERSUS UNCONSCIOUS LEARNING 623

624 TESOL QUARTERLY

On the Utility of the Conscious-Unconscious Distinction

CONSCIOUS VERSUS UNCONSCIOUS LEARNING 625

626 TESOL QUARTERLY

The Case for Abandoning the Conscious-Unconscious Distinction

CONSCIOUS VERSUS UNCONSCIOUS LEARNING 627

has been considerable research on the question of learning without

628 TESOL QUARTERLY

CONSCIOUS VERSUS UNCONSCIOUS LEARNING 629

630 TESOL QUARTERLY

CONSCIOUS VERSUS UNCONSCIOUS LEARNING 631

632 TESOL QUARTERLY

CONSCIOUS VERSUS UNCONSCIOUS LEARNING 633

634 TESOL QUARTERLY

The Competence of Processing:

Theories of second language acquisition are frequently divided

THE COMPETENCE OF PROCESSING:

COMPETENCE THEORIES AND PROCESSING THEORIES

636 TESOL QUARTERLY

CLASSIFYING THEORIES OF SECOND LANGUAGE ACQUISITION 637

638 TESOL QUARTERLY

CLASSIFYING THEORIES OF SECOND LANGUAGE ACQUISITION 639

640 TESOL QUARTERLY

CLASSIFYING THEORIES OF SECOND LANGUAGE ACQUISITION 641

IDEALIZATION AND THEORY BUILDING

642 TESOL QUARTERLY

CLASSIFYING THEORIES OF SECOND LANGUAGE ACQUISITION 643

644 TESOL QUARTERLY

CLASSIFYING THEORIES OF SECOND LANGUAGE ACQUISITION 645

646 TESOL QUARTERLY

CLASSIFYING THEORIES OF SECOND LANGUAGE ACQUISITION 647

648 TESOL QUARTERLY