You are on page 1of 198


A Journal for Teachers of English to Speakers of Other Languages

and of Standard English as a Second Dialect

SANDRA SILBERSTEIN, University of Washington
Associate Editor
SANDRA McKAY, San Francisco State University
Review Editor
HEIDI RIGGENBACH, University of Washington
Brief Reports and Summaries Editor
GAIL WEINSTEIN-SHR, University of Massachusetts at Amherst
Research Issues Editor
GRAHAM CROOKES, University of Hawaii at Manoa
Assistant Editors
DEBORAH GREEN, University of Washington
Editorial Assistant
MAUREEN P. PHILLIPS, University of Washington
Editorial Advisory Board
Roberta G. Abraham Thom Hudson
Iowa State University University of Hawaii at Manoa
Joan G. Carson Claire Kramsch
Georgia State University University of California, Berkeley
Jim Cummins Anne Lazaraton
Ontario Institute for Studies in Education The Pennsylvania State University
Graham Crookes
University of Hawaii at Manoa Michael K. Legutke
Catherine Doughty Goethe Institute, Munich
The University of Sydney David Nunan
Miriam Eisenstein Macquarie University, Sydney
New York University Teresa Pica
Yehia El-Ezabi University of Pennsylvania
United Arab Emirates University/
The American University in Cairo N. S. Prabhu
Susan Gass National University of Singapore
Michigan State University Thomas Ricento
Jean Handscombe Central Michigan University
North York Board of Education, Toronto Patricia L. Rounds
Sarah Hudelson University of Oregon
Arizona State University
Thomas Huckin Andrew F. Siegel
University of Utah University of Washington
Additional Readers
Gregory Barnes, Ruth Benander, Ulla Connor, David Eskey, Ilona Leki, Ann Raimes, Thomas Scovel,
James W. Tollefson, Carole Urzúa, Elizabeth Whalley, Richard Young
Advertising arranged by Patti Olson, TESOL Central Office, Alexandria, Virginia
Typesetting, printing, and binding by Pantagraph Printing, Bloomington, Illinois
Design by Chuck Thayer Advertising, San Francisco, California



To print, select PDF page

nos. in parentheses

The Dynamics of the Language Lesson 225 (10-26)
N. S. Prabhu
Communicative Competence and the Dilemma of
International Teaching Assistant Education 243 (28-54)
Barbara Hoekje and Jessica Williams
ESL Student Bias in Instructional Evaluation 271 (56-73)
Ann K. Wennerstrom and Patty Heiser
Beyond Comprehension Exercises in the
ESL Academic Reading Class 289 (74-103)
May Shih
See How They Read: Comprehension Monitoring
of L1 and L2 Readers 319 (104-128)
Ellen L. Block
The Effect of Speech Modification, Prior Knowledge,
and Listening Proficiency on EFL Lecture Learning 345 (130-159)
Chung Shing Chiang and Patricia Dunkel

The Cornell Lectures: Women in the Linguistics Profession 375
Alice Davison and Penelope Eckert
Reviewed by Johnnie Johnson Hafernik
Building Better English Language Programs:
Perspectives on Evaluation
Martha C. Pennington (Ed.)
Reviewed by Daniel L. Robertson

Teaching Language Minority Students in the Multicultural Classroom,
Robin Scarcella (Marjorie Terdal) 381
Multicultural Education: A Teacher’s Guide to Content and Process
Hilda Hernandez (Ilyse Rathet Post)
Communicative Strategies: A Psychological Analysis of Second-Language Use,
Ellen Bialystok (Greg Jewell)
Volume 26, Number 2 ❑ Summer 1992


Speech Rate and Listening Comprehension Further
Evidence of the Relationship 385
Roger Griffiths
A Second Look at T-Unit Analysis: Reconsidering the Sentence 390
Kathleen Bardovi-Harlig

Comments on Suzanne Irujo’s Review of Rosalie Pedalino Porter’s
Forked Tongue: The Politics of Bilingual Education 397
A Reader Reacts . . .
Keith Baker
The Reviewer Responds . . .
Suzanne Irujo
Teaching Issues
Formal Grammar Instruction 406
An Educator Comments . . .
Marianne Celce-Murcia
Another Educator Comments . . .
Stephen D. Krashen

Information for Contributors 413

Editorial Policy
General Information for Authors
Publications Received 417
Publications Available from the TESOL Central Office 419
TESOL Membership Application 423

Editor’s Note

With this issue, I welcome two new members to the editorial staff of the
TESOL Quarterly. I am pleased to announce that Sandra McKay has
accepted the newly created position of Associate Editor. Beginning with
this issue, Marilyn Kupetz of TESOL’S Central Office shares with Deborah
Green some of the duties of the Assistant Editorship. I am grateful to all
these colleagues for their significant contributions.
This issue premiers a new subsection of The Forum, Teaching Issues,
edited by Sandra McKay. In each instance, an aspect of teaching 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 Sandra McKay at the address provided in the Information for
Contributors section of the Quarterly.
Finally, I note with pleasure that the TESOL Quarterly’s 26th volume is
printed entirely on recycled paper. Readers with a taste for irony (or
mnemonics) will appreciate that this transition coincides accidentally with
the choice of a green cover for this volume.

In this Issue

This issue of the TESOL Quarterly focuses on instructional interactions.

The lead article discusses the language lesson as a classroom event. The
second addresses the training of international teaching assistants for
classroom teaching roles. The third article reports systematic bias in
student evaluations of ESL instruction. Other papers explore activities and,
processes of second language reading and listening.
● N.S. Prabhu discusses the dynamics of the language lesson. He argues
that the lesson represents at least four types of events: a curricular

event, an implementation of a method, a social event, and an arena of
human interaction. Lessons reconcile these conflicting dimensions, in
part through reliance on stable classroom routines. A change in
protective classroom routines becomes productive only when teachers’
own theories are engaged in classroom activities: “This involves
teachers being their own theorists and specialists interacting with
teachers as fellow theorists.”
● Barbara Hoekje and Jessica Williams critique a traditional skills-based
approach to the training of international teaching assistants (ITAs),
arguing that it ignores dimensions of context and role. The authors
advocate reconceptualizing language skills in terms of communicative
competence, “a theoretical model of language use that takes into
account social relationships, language appropriateness, and context.”
This recasting allows ITA programs to justify instruction that
necessarily goes beyond traditional notions of language teaching in
order to prepare ITAs for their new roles as university teachers.
• Ann Wennerstrom and Patty Heiser’s study examines ESL student bias
in instructional evaluation. The researchers find systematic bias on the
basis of ethnic background, level of English, course content, and
attitude toward the course. In light of these findings, the authors
question the use of ESL student evaluations as the basis of personnel
decisions. They recommend that student evaluation data “be used only
in conjunction with other measures. ”
• May Shih’s suggestions go beyond traditional skill-building compre-
hension activities in the ESL academic reading class. Citing research on
comprehension and study strategies, she advocates holistic strategy-
oriented activities. Practical and detailed guidelines are offered to aid
teachers in selecting and sequencing reading materials, designing
criterion tasks, and providing instruction in reading strategies. Shih
advocates activities that promote learner independence and transfer of
strategies from ESL to content reading.
• Ellen Block compares the comprehension-monitoring strategies of L1
and L2 readers with respect to two problems: locating a referent and
defining a vocabulary item. Differences seem to arise more from
readers’ proficiency level than language background. Think-aloud
protocols reveal three comprehension phases: evaluation, action, and
checking. Although she identifies some parallels, Block cautions against
applying results of L1 reading research to L2 readers. On the basis of
this comparative study, she recommends process- and strategy-
oriented L2 reading instruction in which students are taught to identify
reading problems and develop strategies to resolve them.
• Chung Shing Chiang and Patricia Dunkel investigate the listening
comprehension of high- and low-intermediate Chinese EFL students.
Their study explores the effects of speech modification, prior
knowledge, and listening proficiency on listening comprehension.


Their complex results show significant interactions between speech
modification and listening proficiency (only higher level students
benefited from redundant speech) and between prior knowledge and
test type (prior knowledge aided students’ memory for passage-
independent test items). The authors discuss the implications of both
findings for the teaching and testing of listening comprehension,
stressing the need for further research.
Also in this issue:
Reviews: Johnnie Johnson Hafernik reviews The Cornell Lectures:
Women in the Linguistics Profession, a volume edited by Alice Davison
and Penelope Eckert and Daniel Robertson reviews Building Better
English Language Programs: Perspectives on Evaluation in ESL,
edited by Martha C. Pennington.
• Book Notices
• Brief Reports and Summaries: Roger Griffiths provides evidence for
the relationship between speech rate and listening comprehension; and
Kathleen Bardovi-Harlig notes advantages of the sentence over the
T unit for describing the syntactic complexity of adult second language
• The Forum: Keith Baker comments on Suzanne Irujo’s TESOL
Quarterly review of Rosalie Pedalino Porter’s volume, Forked Tongue:
The Politics of Bilingual Education; the reviewer responds. In the
subsection Teaching Issues, Marianne Celce-Murcia and Stephen
Krashen comment on formal grammar instruction.
Sandra Silberstein


TESOL QUARTERLY, Vol. 26, No. 2, Summer 1992

The Dynamics of the

Language Lesson
National University of Singapore

The classroom lesson is an event of several different kinds: It is a

unit in a planned curricular sequence, an instance of a teaching
method in operation, a patterned social activity, and an encounter
between human personalities. Much of what happens in any given
classroom represents a stable routine which best reconciles the
varied demands of these different dimensions for the particular
teacher and learners in question. If this is so, specialist-level effort
at arriving at a good teaching method or curriculum for wide
classroom use is largely misguided since any new method or
curriculum is sure to be unsettling to the stable routine in each
classroom and likely, as a result, either to be discarded as unwork-
able or to be absorbed into a new, stable routine devoid of its con-
ceptual substance. A change in classroom routines can be produc-
tive of learning only to the extent it is motivated and sustained by
conceptual exploration by the teachers themselves. This involves
teachers being their own theorists, and specialists interacting with
teachers as fellow theorists.

This paper is an attempt to explore the nature of the language

lesson—indeed, of any lesson—as an event in the classroom. It is
speculative, in the sense that it represents an effort to make sense of
a complex phenomenon by reflecting on experience and by
perceiving a plausible pattern. In the first part of the paper, I try to
show that the classroom lesson is indeed a very complex event, with
several different dimensions to it. Here I take further a suggestion
made by Dick Allwright (1989), namely, that the lesson is not just a
pedagogic event, but a social event as well. I try to show that the
lesson is at least four different types of event. In the second part of
the paper, I examine the consequences of that perception of the
classroom lesson to specialist effort at pedagogic reform. I argue
that it is naive to think that specialists can formulate a good teaching
method and then get teachers to implement it in their classrooms. I
suggest that classroom teaching can improve only to the extent that
teachers themselves act as specialists.

Let me begin with a familiar perspective on a lesson, namely, that
it is a stage in the implementation of a course. The curriculum
normally is organised as an incremental sequence of teaching units,
the sequence as a whole meant to achieve a large objective, and
each unit in it meant to achieve a small subobjective. Classroom
teaching is seen as a steady movement from the first unit to the last,
and any given lesson is viewed as the completion of one small part
of the journey. What happens in a lesson is, from this perspective,
best understood by reference to earlier and later lessons and by
seeing the logic of the overall sequence.
It is worthwhile noticing here the nature of the connection
between curricular progression and the educative process.
Education is centrally concerned with the mind: Language
education aims to promote one form of mental development in
learners, whether we conceive of it as the development of cognitive
structures or cognitive skills, the realisation of an inherent capacity,
or the creation of a new capacity. In designing educational activity,
this concept of psychological progress is translated into one of
curricular progression, the metaphor of a journey being equally
applicable to both. The completion of a lesson as a curricular unit is
thus meant to represent the completion of a step in the learner’s
psycholinguistic development. If this is taken seriously, a given
lesson has to be understood and assessed in relation to its match
with the learner’s developmental stage at that point, rather than in
relation to preceding and following units. The reason why this is not
done is that it is extremely difficult to perceive the learner’s actual
developmental stage at each point and that the day-to-day activity
of classroom teaching needs to have a more accessible and stable
source of guidance. The provision of a curricular sequence enables
the teacher to assume, without having to ascertain, that each unit in
the sequence matches a corresponding point in the learner’s
progress. A fundamental reason for constructing a curriculum,
therefore, is to give the classroom teacher a workable alternative to
the difficult task of following the learner’s actual development.
However, this workable alternative involves the belief that the
curriculum represents a predetermination of the learner’s
development, both at particular points and over a stretch of time
(i.e., the course); and this in turn implies that the curriculum
developer has far better access than the teacher to the learner’s
mental development: The curriculum developer attempts to predict
or decide in advance the process of development before it takes
place, while the teacher has difficulty perceiving it as it happens. It


may be that curricula are necessary, even indispensable, in educa-
tional activity, but it is nevertheless important to realise that they
rest on a very large assumption and that the perception of a lesson
as a curricular unit involves an appeal to that assumption.


A second perspective on a lesson, which is also familiar, is that it
represents the implementation of a method of teaching. A teaching
method, in the sense I am using the term, has two aspects to it: a
conceptual aspect and an operational one. While the conceptual
aspect consists essentially of a theory of how learning takes place, or
can best take place, the operational aspect consists of a specification
of what should be done in the classroom, within each teaching unit.
A method in operation is commonly seen as a prespecified pattern
of activity within a given lesson (or a sequence of related lessons),
the pattern being recurrent from one stretch of teaching to another.
Given a logical link (or at least a conventional association) between
the conceptual and operational aspects of a method, it is possible to
observe the pattern of activity in a single lesson and infer the theory
of learning involved—to label the teaching, for instance, as
structural, communicative, experiential, and so forth, or as a mix of
two or more. A method is what lies behind a lesson plan—what
guides the teacher in deciding what activities are to be undertaken,
and in what order, in the course of a lesson,
Let us juxtapose briefly the two views of a lesson mentioned so
far. The lesson as a curricular unit makes reference to the teaching
content and to one stage in an incremental progression. The lesson
as implementation of a method makes reference to the teaching
procedure and to one cycle in a cyclic sequence. As a curricular
unit, a lesson needs to be seen in relation to other curricular units; as
a method in operation, it needs to be seen in relation to the theory
of learning that corresponds to it. It was noted above that in
following the curricular progression the teacher is assuming that
that progression is a dependable indicator of the learner’s psycho-
logical progress. In following a method as a pattern of activity, the
teacher is assuming that the theory of learning that informs that pat-
tern of activity is in fact a valid theory. I will have occasion to return
to this point a little later in this discussion.


Seeing a lesson in relation to a curriculum or to a teaching method
is seeing it as a pedagogic event; and this is how lessons are seen,


most of the time, in specialist-level discussions of classroom
teaching and learning. We can in fact regard this as the specialist’s
perspective on the language lesson, perhaps shared or trusted
partially by the teacher in the classroom, but rarely shared by the
learner. A different perspective on a lesson, shared by teachers and
learners alike though seldom brought into professional discussion, is
that it is a routinised social event, with roles and role relationships
established by tradition and with a ritualistic aspect to the actions
performed. There are roles assigned by custom to teachers and
learners in the classroom, though these may vary from one culture
to another, just as there are roles assigned to the priest and the
worshipers at a temple, or to the judge, the lawyers, and the
litigants in a court of law. There is also a certain agenda of events
associated with a classroom lesson, though this may be very general
in form, just as there are agendas associated with religious worship,
the courtroom, communal decision making, the annual share-
holders’ meeting, the company boardroom, or an academic
seminar. The classroom lesson, that is to say, is a kind of social
genre, stereotypic and to some extent ritualised, generally not
noticed as such when it is conformed to but disturbing when it is
violated. The ritualisation may or may not take the form of dress
regulations, standing up to show respect, the use of honorifics, first
names or last names, not speaking unless asked to, procedures for
assignment and submission of work, procedures for punishment
and reward, opening and closing moves for the lesson as a whole or
for any phase of it, and so on; but there is at least a set of shared
notions about the different phases of a lesson, legitimate and
deviant behaviour, the extent of teacher’s authority and learner’s
right, and duties and obligations on both sides. Further, these
notions are prevalent not just among teachers and learners but
among parents, educational administrators, and the general public.
There is indeed a form of classroom ethic, just as there are forms of
ethic associated with doctor-patient or lawyer-client transactions.
Although this view of the classroom lesson as a routinised social
event may look largely irrelevant to professional discussion of
pedagogy—or, at best, a possible hazard to beware of in curriculum
construction and method determination—it would be a mistake not
to see an essential function which it serves. The classroom lesson is
a recurrent encounter between people and, like all recurrent
encounters, needs the sense of security arising from shared
expectations, Human encounters which are entirely unpredictable,
with no foreknowledge or even an expectation of what is to happen,
are inherently threatening and would be intolerable as regularly
recurrent events in any social group or institution, including the


family. One needs to be able to anticipate events, in some general
form, take some things for granted, even tentatively, and have a
frame of reference and roles with which to interpret and respond to
what happens. The more recurrent the encounter, and the more
numerous its participants, the greater the need for a shared routine
and a shared set of expectations. It is only with some notion of
where one belongs and where the others belong that one can engage
in a repeated encounter with no great sense of threat. With
classroom lessons, moreover, it is not just the participants in the
encounter—teachers and learners—who need the security arising
from shared expectations; parents need it too on behalf of their
children, as do educational authorities and indeed the society as a
whole. Social routinisation is therefore a necessary support in
making the classroom lesson possible as a recurrent event.


Behind the conventionalised roles and routines of a lesson are a
group of individuals—a teacher and many learners—with varied
personalities, motives, self-images, fears and aspirations, levels of
tolerance, and degrees of maturity. A teacher perceives different
learners not just as being strong or weak in learning, but as being
interested or indifferent, modest or presumptuous, accepting or
rebellious, supportive or subversive, trusting, skeptical or mocking,
accessible to empathy or inscrutable, and so on, in varying degrees.
Conducting a lesson is, first and foremost, handling a collection of
friendly and unfriendly people in a way that maximally protects or
projects, and minimally hurts or diminishes, one’s own self-image as
a teacher. The learners in their turn perceive the teacher as being
friendly or unfriendly, helpful or hostile, tolerant or vindictive, and
so on, both to themselves and to different fellow learners, and try to
act in a way that protects or enhances their own varied self-images.
Such play of personalities takes place not just between the teacher
on the one hand and learners on the other but, in a more fierce form,
between learners themselves, in a highly complex and multilateral
form. There are likes and dislikes, loyalties and rivalries, ambitions
and desires to dominate, injured pride and harboured grudge,
fellow feeling and jealousy, all creating a continual threat to security
and self-image, and calling for protective or corrective action.
Perhaps the most immediate and pervasive concern of teachers and
learners alike in classrooms is to guard against a loss of face and, to
the extent possible, to win approval, sympathy or loyalty from
others as a safeguard against future hazards to “face.”


A classroom lesson is thus an arena of human interactions—not
the pedagogic interactions desired or elicited as a part of teaching
strategy (which may or may not occur as expected) but the more
elemental, inevitable interactions which occur simply because human
beings, with all their complexity, are involved. They are interactions
which occur equally on the playground, at a youth club, or in any
other social group, though the modes of their expression may vary
from one arena to another. One may therefore wish to ignore them
as being nothing special to the classroom, much less to the language
classroom, or try to control and manipulate them directly as in some
humanistic approaches, by reducing and simplifying them to man-
ageable proportions. My point here is simply that they constitute an
important dimension of the classroom lesson and, although it may
be naive and unproductive to try to prescribe ways of handling
them, it would be equally naive not to be aware that they operate
powerfully in the classroom and will have a large influence on what-
ever one does prescribe by way of a curriculum or method.


I have now identified four dimensions of the classroom lesson: the
lesson as a curricular unit and as a method in operation, which may
both be regarded as pedagogic dimensions; and the lesson as a
social genre and as a play of personalities, which may be considered
social and personal dimensions. A part of the value of such an
analysis, as I see it, is that it enables us to understand a range of
problems that arise in the classroom as conflicts between different
dimensions. Thus, the method being implemented may require the
teacher to be maximally supportive of learners’ efforts and
achievements in the classroom and to highlight positive feedback
rather than negative feedback. But learners as a social group often
have the notion that it is opportunism or a form of surrender to be
praised or favoured by the teacher and that there is bravery in
defying the teacher’s wishes and values. This notion of heroism,
based on rejection and rebellion, can set up a source of approval
and disapproval directly in conflict with the teacher’s approval and
disapproval, thus creating a dilemma for many learners and a threat
to the teacher’s sense of security. To take another instance, a
recommended teaching procedure may incorporate the principle
that learners’ efforts should precede the teacher’s input, such that
much of the learning takes place as a form of discovery by the
learner, and the teacher’s input is responsive to the learner’s effort,
rather than preemptive of it. But the classroom lesson as a social
genre often includes the notion that it is a part of the teacher’s role


to provide the necessary inputs and that it is therefore unfair or
incompetent of the teacher to demand effort by learners in the
absence of such inputs.
Such conflict can arise in a more acute form when the pedagogic
approach involves a principled refusal to provide a curriculum in
advance and a demand, instead, that the classroom itself engage in
evolving a curriculum through teacher-learner negotiation (Breen,
1984). Not surprisingly, proposals of this form have had to
incorporate a component of “learner-training” (Allwright, 1984)
partly in the hope of altering the expectations of a lesson as a social
genre. A related area of conflict is group work. It is common these
days to highlight the pedagogic benefits of small-group work by
learners, peer support and peer correction, in place of teacher-
fronted activity, and it is often argued that such small-group work
provides a sense of security to learners. But there is, as we have
noted, a complex play of personalities between learners, involving
dominance, rivalry, rebellion, submission, or withdrawal, and there
isn’t a socially established frame of roles and routines for small-
group interaction as there is for teacher-fronted activity. It has also
been common for some time now to give a high value to learner talk
in the classroom and to require the teacher to make a special effort
to draw out the silent learners by various means. But learners are,
like all other people, varied in their personalities—some taciturn
and others vociferous, some shy and others showy, some diffident
and others adventurous, and so on, and being drawn out conspicu-
ously can constitute for some a disconcerting exposure and a
disorientation on the lesson.
My purpose in drawing attention to such conflicts is not to
recommend any specific procedures for resolving them, but simply
to suggest a way of perceiving and understanding them as the
outcome of an interplay of different forces at work in the
classroom. I think, in fact, that it is unlikely to be fruitful for the
language teaching specialism to take on board the social and
personal dimensions of the classroom lesson, in addition to the
pedagogic dimensions, and to try to provide general procedures for
reconciling them. This is not only because the classroom lesson as a
social genre varies across societies and cultures, and the mix of
learners’ personalities in any classroom is unpredictable and
irreducible to simple patterns; it is equally because the teacher’s
own personality is a major factor in the interplay of forces, and
conflict resolution will necessarily have to vary from one teacher to


Conflict resolution, however, does take place in all classrooms, in
varied forms and to varied extents—or else it would be impossible
for the class to function. It takes place in the way it does in any other
social group—by each participant arriving at a set of relationships
which give him or her the maximum security, stability, or room to
manoeuvre attainable in the circumstances. Opposing forces ac-
commodate each other or hold each other at bay, and the different
dimensions reinforce their common ground and limit their en-
croachment on one another. Personalities come to terms, as best
they can, both with other personalities and with prevalent social and
pedagogic frames. Perhaps the social and personal dimensions are
so strong in some classrooms that they subordinate the pedagogic
ones; and perhaps it is the reverse in other classrooms. Perhaps a
concern with getting on with the curriculum (with a view to getting
ready for an examination) is sufficiently dominant to absorb or sus-
pend other concerns in some school systems; or perhaps there is a
strong concern for the socially sanctioned roles and routines such
that conflicting parts of other dimensions are subdued or sup-
pressed. Perhaps the teacher has enough status or strength of per-
sonality to ensure priority to one dimension or another; or perhaps
the teacher is sufficiently insecure to take the line of least resistance.
As a result, the balance of different dimensions and forces arrived at
is bound to differ from one classroom to another, in terms of its
composition. It is also bound to differ in terms of its relative stabil-
ity, some classrooms operating with a more precarious balance than
When language pedagogy is discussed in professional forums, it is
the pedagogic dimensions that chiefly come under focus, rather
than the social and personal dimensions. It is generally assumed that
the central task is to arrive at the best possible curriculum and the
most promising method and that, once arrived at, these pedagogic
inputs can be carried into classrooms fairly easily, given an
agreement to do so by those who have the authority to prescribe
curricula and methods. It is true that a large effort is made to
persuade authorities to adopt the putatively better curricula and
methods, and to ensure that they understand the rationale of these
new inputs as fully as possible. It is also true that various forms of
teacher training are undertaken to make sure that teachers
understand what is to be done, how, and why. But all these are
measures meant to make certain that the intended inputs do in fact
get to the classrooms, with minimal dilution or distortion on the
way. They still assume that the inputs will work in the way envis-
aged once they get to the classrooms.


Perhaps it will be objected that professional discussion of
language pedagogy does not restrict attention these days to what I
have called pedagogic dimensions but pays increasing attention to
such contextual and social factors as the local educational policy or
expectation, various characteristics of learners and teachers,
features of the school system, features of prevalent and favoured
forms of classroom interaction, and so on, and that this more
comprehensive view ensures a larger relevance or feasibility to
specialist inputs than I have suggested above. The question,
however, is not how much attention should be given to contextual
and social factors in arriving at pedagogic recommendations but, at
a more basic level, what any such attention can hope to achieve.
Making a prior study of contextual and social factors and taking
them into account in determining desirable pedagogic procedures is
at best an attempt at anticipating and preempting difficulties arising
from those factors in the classroom, not at resolving difficulties as
they arise. Anticipation depends on generalisation: One needs to
identify characteristics of sets of learners and sets of teachers in
order to accommodate those characteristics while recommending
methods. But several factors involved in the social and personal
dimension of the classroom—in particular, in what I have referred
to as the play of personalities—are characteristics of individuals
rather than of sets, or of particular groupings of individuals in
contrast to other groupings. Besides, taking account of a set of
factors is not the same as taking account of the interplay between
those factors. The latter is necessarily a process of ongoing change,
which can only be managed responsively as different situations
arise, not preemptively through an anticipation of those situations.
Indeed any attempt to deal with it preemptively can by itself
constitute a form of intervention which adds to the different forces
at play, instead of easing them. It therefore does not seem very
worthwhile for specialist inputs to try to take in more than the
pedagogic dimensions in order to be less unsettling to the
classroom. It may be more useful to reconsider the role of specialist
inputs themselves and to ask how the teacher’s own management of
the forces at play in the classroom might become pedagogically
more constructive.
An important concern for teachers and learners in the classroom,
as noted already, is their sense of security and the protection of their
self -images. In the course of conflict resolution, all participants are
likely to give first priority to those options which will most
safeguard their self-esteem and leave them to play their roles as
teachers or learners in relative security. Teachers come to terms
with learners in a way that least erodes their status, in their own


eyes, in the learners’ eyes, and perhaps in the eyes of fellow teachers
and superiors. Learners too come to terms with the teacher and with
fellow learners in a way that most protects their individual self-
images and leaves them least vulnerable to a loss of face or facade.
The different dimensions of the classroom lesson, both pedagogic
and social, are all so interpreted and accommodated as to arrive at
an overall balance which all or most participants can live with,
though not necessarily in equal comfort (and some only by building
shells around themselves, as independent protection). The
availability of conventionalised roles, routines, and agendas is a
significant help in the process, roles acting as shields to personality.
The availability of a curriculum to go through is also a help in that
it provides a nonpersonal entity to focus on, thus de-emphasizing
personal tensions. Patterns of activities dictated by the teaching
method act as a further form of routine which, being recurrent and
predictable, affords a measure of security. For these reasons, the
routines of a classroom lesson are not just routines but carriers of a
more complex though less perceptible balance of the different
forces at work. Consequently, the more well-established the
routines are, the more settled the balance is. The reverse is also true:
The more stable and comfortable the balance is to teachers and
learners, the more quickly and firmly the routines get established,
becoming less and less open to change.


The picture of the classroom I have just outlined may strike some
people as being an exaggeration of the nonpedagogic dimensions. I
hope, however, that it will be recognised as being at least partially
true, and reminiscent of some uneasy feeling we have all had about
classroom lessons on some occasions. I would like to suggest that,
because we deliberately focus on the pedagogic dimensions of the
classroom in professional discussion, we tend to exaggerate the role
of those pedagogic dimensions, at the expense of the nonpedagogic
ones. I would also like to suggest that what I have said about
classroom lessons—namely, that they go on largely as routinised
events, the routines sustaining a workable arrangement of relations
between the participants involved and making the recurrent
encounter tolerable or comfortable to them—is generally true of
other institutionalised groups in recurrent encounter, whether in
public or in private life; so it would be surprising if classroom
lessons did not share those characteristics. In any case, my purpose
in highlighting the routinised nature of classroom lessons is to point


out its implications for our professional discussion of language
pedagogy—a matter which I now turn to.
If classroom lessons take place as routinised events, they can only
provide the kinds of satisfaction that routinised events do—the
satisfaction of having performed a routine enjoined on one (as with
much religious worship) or of having gone through one more stage
in a prescribed sequence (as with many committee meetings), at a
cost not higher than necessary. Any learning that takes place, in the
sense of cognitive development, is incidental to the process and
generally independent of the specific routines engaged in. But
professional discussion of pedagogy makes the fundamental
assumption of a cause-effect relationship between the activities
performed in the classroom and the learning that results. As
specialists, we are concerned centrally with identifying those
classroom activities which promise to be most productive of
learning and with commending to teachers the outcome of our
search. We decide, from time to time, that a structure drill, or a
grammatical explanation, or a role play, or a problem-solving task,
or a discourse-processing task, or small-group interaction, and so
on, are specially productive of language learning, and urge teachers
as strongly as we can to carry out these activities in their classrooms.
If what I have said about classroom lessons is even partly true, there
is a need for us to pause to ask what effect the new procedures we
recommend are likely to have on existing classroom routines.
Since existing routines in classrooms are a source of relative
security and stability, any intrusion of new activities is a threat to
that stability. It is therefore understandable that teachers tend to be
reluctant to alter their teaching procedures despite recommenda-
tions that they do so. But let us suppose that teachers do act on the
recommendations and attempt to alter classroom procedures. There
will be a disturbance of the prevailing balance of forces,
necessitating a fresh process of conflict resolution and leading
eventually to a new balance which can accommodate the new
pattern of activity. If the new balance is a precarious one—with
which participants in the classroom lesson are not sufficiently
comfortable—the new pattern of activity continues to be a source
of insecurity and is likely to be thrown out as soon as an opportunity
arises to do so. Once again, however, let us assume that the new
balance arrived at is a stable one, with a firm assimilation of the new
pattern of activity. Inevitably, there will have been a resolution of
conflict between the old routine and the new pattern of activity,
with various adjustments and modifications in both, making use of
any common ground and toning down the conflicting parts. The
recommended pattern of activity has therefore been assimilated in


various forms in various classrooms so as to be in most harmony
with existing routines. Even where it has not been altered
significantly in the process of assimilation, it has at least become a
part of a new stable pattern, allied to a new, stable balance of the
varied forces at play.
The result of such a development is that the altered procedure
will now establish itself steadily as the new routine, serving the same
functions as did the old routine and providing little more than the
forms of satisfaction that arise from the performance of such
routines. This I think is the main reason why in the language
teaching profession there has been a succession of new pedagogic
approaches and methods on the one hand and, on the other, a
succession of disappointments at the implementation of those
methods in classrooms and, even more, with the learning outcomes
of such implementations as have taken place. While we think that
we are providing new teaching procedures capable of yielding
better learning outcomes, we are actually providing new classroom
routines to replace old routines, the general dynamics of the
classroom continuing to be what they are.
When can a new method be more than merely a new routine in
the classroom? I said earlier that a method has a conceptual aspect
as well as an operational one. The teaching procedures we
recommend to teachers constitute the operational aspect and if that
is all that teachers take from us and try to implement, the
implementation is sure to be a matter of replacing old routines with
new. But suppose we put to teachers not just the operational aspect
but the conceptual aspect as well—that is to say, our theory of how
language learning comes about and how the procedures being
recommended are likely to bring it about. We are now asking
teachers to take to the classroom not just the teaching activities we
recommend but the theory of language learning behind those
activities. We want teachers, in other words, to make our theory
their own—to be as persuaded by the theory as we are. It is not
difficult to see that this is a rather naive demand to make. It implies
that people’s concepts and beliefs are as open to a replacement as
their behavioral routines. Reference has been made above to the
difficulty of altering routines; altering beliefs can only be that much
more difficult, unsuccessful, or uncertain in its outcome. If a threat
to one’s routine is unsettling, a threat to one’s belief can only be
more unsettling. But perhaps it does not have to be an alteration of
beliefs; all that is needed, it may be argued, is that teachers
understand the theory involved, not necessarily be persuaded by it.
This means, however, that we are asking teachers to operate with a
theory that they do not themselves subscribe to. It seems to me a


formula for a profound sense of insecurity—to have to carry out,
day after day, a set of activities which upset established routines
and unsettle the prevailing balance of relationships, and to be trying
to justify those activities with a theory which is at variance with
one’s own belief. It would not be surprising if survival dictated a
discarding of the theory and a retention of just the pattern of
activity, to be assimilated into existing routines.


For classroom activities to be more than protective routines, it is
minimally necessary for teachers to be operating with their own
beliefs about the pedagogic value of those activities—with their
own notions or theories of how learning comes about and how the
teaching that is being done is bringing it about. I do not mean by
this that the theories should be original to teachers and different
from all others. They may have been received, adopted, or
inherited by teachers from various sources, with or without
conscious decision, at some time in the past. What I mean is simply
that they should be theories to which teachers see the most truth or
plausibility and with which teachers are most able to identify, at a
given time. These theories will of course vary widely in content
between teachers and vary inevitably from the theories being
advocated by specialists. Variation (and, indeed, conflict) between
theories is, after all, conspicuously prevalent among us as specialists,
who are admittedly advocating what we see the most truth or
plausibility to; so there is no reason to expect any less variation in
the theories that different teachers identify with. More importantly,
teachers’ theories are sure to vary in the degrees of their explicitness
and in the level of awareness teachers have of them. In many cases,
they may be unarticulated and ill-defined, in the form of intuitions;
in other cases, they may be strongly held but unexamined notions,
in the form of beliefs or preconceptions; in yet other cases, they
may not be easily accessible at all, having been “buried” in the
process of classroom routinisation. As a result, the foremost need, if
classroom activity is not to be just a performance of routines, is for
teachers’ own beliefs and intuitions to become more accessible to
them and more actively engaged as one of the forces in the
dynamics of the classroom.
It is common to think of specialists in language pedagogy as
people who are able to develop theories of language learning and
teaching, to support those theories with relevant evidence, and to
state them explicitly for other people’s consideration. In contrast,
classroom teachers are thought of as being ill-equipped or at least


unengaged in this mode of rational, “scientific” enquiry, though
hopefully able to make use of the products of such enquiry. I would
like to suggest, however, that there is a more basic level of mental
activity which is shared by specialists and teachers (and indeed all
human beings). All of us have commonsense views of the world,
constituting our minimal response to a human urge to make sense of
what we experience and what we do. If one can talk about this
commonsense knowledge construction by employing (for want of a
more direct means) rational theory construction as a metaphor, one
can say that the human psyche includes a large theory-building
mechanism, operating subconsciously. It takes in the data of
ongoing experience, as well as available/inherited interpretations of
such experience, and strives to arrive at what it regards as the most
satisfying view through (what may be thought of as) theory
construction, theory revision, theory replacement (or data shelving,
data doctoring, data dismissal), and so on. None of us has much idea
of the “rules” of this activity but that does not justify our ignoring its
existence. Indeed, it is possible that some (or much) of the
specialist-level, rational theory construction represents an explicit,
“rationalised” articulation of this subconscious activity, or at least
draws strength from intuitions arising from it.
Teachers have commonsense views of learning and teaching, just
as other people do of the activities they are engaged in. The data
leading to these views may have arisen from a variety of sources,
including teachers’ own experience at one time as learners, and
existing views of learning and teaching, as noted above. A
continuing source of such data is each teacher’s ongoing experience
in the classroom and, to the extent such experience is being taken in
as data by the teacher’s subconscious “theory,” there may be said to
be a further factor at play in the classroom, namely, a form of
theorising by the teacher. The process of theorising can be said to
be strong when the subconscious theory is active and able to take
account of continuing data, and weak (or “frozen” into a prejudice
or dogma) when it is not active in that way. It is the strong state of
the process that I am calling the engagement of the teacher’s theory
in the classroom.
When a teacher’s own theory is engaged in classroom activity, the
teacher has more at stake than an immediate sense of security and
the protection of self-image. The theory is something the teacher
identifies with and therefore constitutes an intellectual stake in the
activity. There is a reason to seek confirmation of the theory from
what goes on in the classroom—or to come to terms with any
disconfirmation by altering or reinterpreting the theory itself. This
makes the teacher’s theory a significant factor at play in the


classroom—a factor more powerful than the need to conform to a
prescribed curriculum or prescribed method because it is a
personal, intellectual stake for the teacher, comparable to a sense of
security, which is a personal, affective stake. The theory thus acts as
an additional, strong influence on the resolution of different
conflicts in the classroom and is in turn influenced by the other
forces at play and by the confirmation or disconfirmation it
receives. There is then an active, ongoing interaction between the
theory and its operation in the classroom—between concept and
conduct. Teaching becomes something of an intellectual explora-
tion—a process of subjecting one’s theory to an operational test, and
sustaining or modifying it in the light of outcomes (to employ again
the metaphor of scientific enquiry). To experience such a process of
exploration is to experience professional growth as a teacher—to
learn a little about teaching and learning from every lesson (or every
so many lessons) taught in the classroom. The process is also an
exciting one, in the same way that specialists find it exciting to
explore theories of teaching and learning. For specialists to “give”
methods to teachers is for them to pass on a set of outcomes devoid
of the excitement of arriving at them. For teachers to engage in
theorizing is for them to experience the excitement of the process
and to be able to continue the exploration as a part of their teaching.
Such an engagement of the teacher’s own theory will not
eliminate routinisation—indeed, it would be destructive for it to do
so, since a certain degree of routinisation is essential to day-to-day
teaching, as we have noted—but it can prevent overroutinisation to
the point where lessons become merely the performance of routines.
It can do so because the intellectual excitement of exploring and
developing a theory can act as a rival to a desire for security, and
replace it, to various extents, as the dominant factor in resolving
conflicts and arriving at a balance of the forces at play. Further, if
the teacher engages in classroom activity with a sense of intellectual
excitement, there is at least a fair probability that learners will begin
to participate in the excitement and to perceive classroom lessons
mainly as learning events—as experiences of growth for themselves.
My suggestion, therefore, is that for classroom lessons to be more
than mere routines, teachers need to be theorists—not in the sense of
being able to wield the apparatus of scholarship or the skills of
academic argument, but in the sense of operating with an active
concept of the cause-effect relationship between teaching and
learning, and of developing and modifying that concept in the light
of their ongoing experience in the classroom.
This is admittedly an idealistic view of classroom lessons, with
limits of various kinds on its realisation. I think, nevertheless, that it


is an ideal worth striving for, in order to achieve whatever success
is achievable. I think, in particular, that specialist effort in language
pedagogy needs to explore possible ways of enabling teachers to
operate as theorists, instead of providing them with new methods to
be used as new routines. If, as I have tried to show, classroom
lessons are highly susceptible to routinisation, it is more worthwhile
enabling teachers to counter the forces of routinisation than merely
helping them to change from one routine to another. It is of course
unclear at this stage what procedures might be useful in enabling
teachers to function as theorists, though it does seem clear that
teacher training will need to seek ways of activating teachers’
subconscious theories and getting them to interact with available
theories. Perhaps there are ways of helping teachers to try to
articulate their pedagogic notions and intuitions, such that the
process of articulation acts as a form of exploration, and any success
in articulation helps to increase confidence as well as ability for
further exploration. Perhaps specialists should themselves take
teachers’ theories seriously and interact with them in the way they
interact with each other’s theories. I do not think that specialists in
language pedagogy have any better source for their theories than
their own notions and intuitions, though of course they are more
highly skilled in articulating them and supporting them in
academically recognised ways. Perhaps it will be a gain for all of us
to acknowledge that specialists’ theories are on the same footing as
teachers’ theories, and that both specialists and teachers can benefit
through an interaction between their theories. Perhaps teachers will
be helped to function as theorists if those who regard themselves as
theorists begin to function as teachers.

A version of this paper was presented as a featured talk at the 25th Annual TESOL
Convention in New York, March 1991.

N. S. Prabhu teaches applied linguistics at the National University of Singapore.
His earlier work in India included the development of task-based language
teaching and the concept of a procedural syllabus, in the context of what is
generally known as the Bangalore Project.


Allwright, D. (1989, April). Interaction in the language classroom: Social
problems and pedagogic possibilities. Paper presented at Les Etats
Generaux des Langues, Paris.
Allwright, R. L. (1984). The importance of interaction in classroom
language learning. Applied Linguistics, 5 (2), 156-171.
Breen, M. P. (1984). Process syllabuses for the language classroom. In
C. J. Brumfit (Ed.), General English syllabus design. Oxford: Pergamon


TESOL QUARTERLY, Vol. 26, No. 2, Summer 1992

Communicative Competence and the

Dilemma of International Teaching
Assistant Education
Drexel University

University of Illinois at Chicago

This paper examines the curriculum of international teaching

assistant (ITA) programs within a sociolinguistic model of
language use, the “communicative competence” model proposed
by Hymes (1972) and, in its pedagogical applications, by Canale
(1983), Canale and Swain (1980), and Savignon (1983). In the face
of a growing number of ITA programs, at times with ad hoc
curricular goals, the communicative competence framework
provides an integrated, theoretically sound principle for ITA
program development and assessment, and an appropriate basis
for understanding the emerging models of ITA competence. In
order to take advantage of such a framework, however, it is
crucial that ITA program developers and instructors come to
terms with the task of preparing their students for their new role in
the classroom. This will necessarily entail training beyond
traditional ESL concerns, although ESL professionals may have
neither the mandate from their institutions nor the background to
accomplish this.

The problems associated with the international teaching assistant

(ITA) are by now well known to TESOL professionals in higher
education, particularly in the U.S. There have been conferences,
studies (see, for example, Bailey, Pialorsi, & Zukowski/Faust, 1984;
Young, 1989), and textbooks (Byrd, Constantinides, & Pennington,
1989; Pica, Barnes, & Finger, 1990; Smith, Meyers, & Burkhalter,
1992) on the topic. In this article, we have gathered these findings
and placed them in a framework from which to pursue further
research and pedagogical innovation. In order to improve ITA
education, however, we feel it is crucial to address a dilemma many
ITA program designers face.

In an effort to provide more narrowly focused instruction than is
found in general ESL programs, ITA program developers have
frequently tried to specify the discourse functions and domains that
are most important for ITAs. Typically, course development begins
with what Hutchinson and Waters (1987) have called a language-
centered or skill-centered needs analysis. First, native-speaker
teaching assistants (NSTAS) are observed to determine the language
items and functions integral to an ITA’s success. Next, a needs
analysis is undertaken to determine the communication skills ITAs
presently hold or lack.
Early language- and skill-centered needs analyses of ITAs (see
papers in Bailey et al., 1984) revealed, among other things, that ITAs
needed to interact in colloquial spoken English inside and outside
the classroom. Interestingly, the early studies also revealed the
particular importance of the nonlinguistic factors: ITAs who were
highly authoritarian in approach were often more likely to cause
resentment among undergraduates than those who were less
linguistically skilled but more in tune with typical patterns of
classroom interaction (Landa & Perry, 1984; Shaw & Garate, 1984).
Similarly, ITAs with excellent teaching skills seemed to surpass the
language barrier, in terms of undergraduate satisfaction, whereas
those with poorer teaching skills could not (Bailey, 1984). An
understanding of the organization and rationale of the U.S.
classroom setting and undergraduates’ expectations of teaching was
seen as crucial. Thus, a focus on culture, pedagogy, and language
became an early focus of ITA instruction. In a recent survey of 26
ITA programs across the U.S. (Barnes, Finger, Hoekje, & Ruffin,
1989), program administrators continued to state ITA needs in
terms of these three categories. New ITA course texts stress the
same areas.
There appears to be a consensus in the field that adequate ITA
training should include culture, pedagogy, and language, but the
question of the relative importance of each and the relationship of
one area to the other remains. Many ITA curricula are organized
along some variation of this three-part approach; however, such
curricula are inadequate in several respects. We believe that an
underlying problem of many skills-based ITA curricula is an
insufficient recognition of context and role. Indeed, sociolinguistic
research over the past few decades has demonstrated the over-
whelming importance of analyzing communication in context. For
example, ITAs who practice 10- to 15-minute presentations in the
teaching-skills section of the ITA course, may be stonily received
by their undergraduate students who do not expect rehearsed
deliveries. ITAs who transfer the concept of U.S. classroom


informality to an overuse of slang and joking will also be seen as
behaving inappropriately.
In addition, a curricular division into culture, pedagogy, and
language immediately raises questions of turf and legality: Who
should be the appropriate training body? Some contend the ESL
program should take care of the language and culture instruction,
while the instructional development unit do the actual teacher
training. In this case, the issue of what constitutes culture becomes
problematic. ESL programs often deal with cultural topics such as
U.S. attitudes toward time, personal space, and family relation-
ships. These are important, but a familiarity with the culture of the
U.S. classroom is what is crucial for a new TA and, therefore, per-
haps better handled within a teaching methods course. Clearly,
however, it is difficult to address this problem without dealing with
the appropriate use of language, an area more typically associated
with ESL programs. This situation poses a major dilemma for ITA
education. On the one hand, the goal is to prepare new ITAs for the
roles and duties they must fulfill. On the other hand, it is precisely
this kind of specialized education, because it goes beyond standard
ESL concerns, that leaves such programs open to accusations of
The dilemma is furthered because to the general public, the ITA’s
problem is primarily seen as a linguistic one, often defined in terms
of pronunciation or fluency (Brown, Fishman, & Jones, 1989). Leg-
islation designed to respond to this problem, such as Pennsylvania
Senate Bill 539, typically requires the certification of all instruction-
al faculty in English fluency; such laws usually treat fluency as a sin-
gle construct,l and show no recognition of the fact that communica-
tion standards legitimately vary according to participants and con-
text. The notion of a unitary standard is also upheld by some univer-
sity administrators who demand a SPEAK (Speaking Proficiency
English Assessment Kit) cut-off score or other clear demarcation
between certified and noncertified speakers.
We contend that in the case of ITAs, oral skills cannot be
separated from the context in which they are practiced, that is, in
the teaching assistant role. This role encompasses a wide variety of
behaviors and skills, and implies a unique set of authority
relationships with students that are permutations of the roles played
by faculty (Hoekje & Tanner, 1987; McKeachie, 1978).
We believe that ITA curriculum and assessment are best
approached within a theoretical model of language use that takes
1 The bill requires universities to “evaluate their faculties for fluency in the English language;
providing for certifications as to that fluency. . . .”


into account social relationships, language appropriateness, and
context. Sociolinguistic models of language use, such as those of
communicative competence proposed by Hymes (1972), and more
specifically, in their more pedagogical applications, by Canale
(1983), Canale and Swain (1980), and Savignon (1983), are an
appropriate basis for developing models of ITA competence and
for examining ITA language use. Such a framework should be used
as a basis for setting goals, designing curriculum, and evaluating
We are not the first to consider communicative competence an
appropriate goal for ITA programs; in fact, ITA research is
increasingly being carried out within this or compatible perspec-
tives, as demonstrated, for example, by the presentations of such re-
searchers as Davies (1991), Myers (1991), Rounds (1990), and Tyler
(1990) in preconference symposia on ITA training at recent TESOL
Conventions. Following in this tradition, our purpose is to propose
communicative competence as a general, theoretical framework for
ITA instruction, and to review relevant studies in this light. We
intend to demonstrate the advantage of this framework as an
organizing principle for ITA curriculum and evaluation, and to
provide a theoretically sound basis for responding to language
legislation. We believe that only when ITA programs are situated
within this framework can we emerge from the dilemma regarding
turf and legality.
In the following section, we discuss the overall construct of
communicative competence and provide a detailed discussion of
the individual components of linguistic, sociolinguistic, discourse,
and strategic competence in terms of how they apply to ITA
programs. In the process, we review many studies of ITAs in light
of this model and give examples from our own observations of ITAs
in the classroom. Finally, after having come to a fuller understand-
ing of the nature of ITA language use and the goals of ITA training,
we turn to the issue of appropriate assessment.


Our framework is Hymes’ notion of communicative competence
in which language behavior is viewed in terms of its appropriate-
ness and correctness (Hymes, 1972). Of particular importance to
ITA programs is that communicative competence in this framework
includes both knowledge of linguistic rules and the ability to apply
these rules to use language appropriately. Hymes is not alone in
pointing out this crucial difference; Widdowson (1978) stresses that


in language teaching, linguistic skills and communicative abilities
must be seen as distinct. Savignon (1983) rejects a distinction
between competence and performance, stating that the only way of
determining the level of speakers’ communicative competence is
through performance. For ITAs, this ability to apply grammatical
knowledge in communicative situations must be stressed; although
they frequently display a thorough knowledge of content material
and grammatical rules, they often are unable to apply this
knowledge to communicate effectively. Any framework for
determining ITA communicative competence must minimally
include the presentation of information in such a way that students
understand it, familiarity with the speech situation and the roles of
participants, as well as development of speech and interaction
ITAs participate in complex speech events that they may not be
able to handle. It is therefore crucial that curriculum development
begin with an integrated framework for determining what abilities
ITAs need to achieve communicative competence. Canale (1983)
and Canale and Swain (1980) have proposed a theoretical basis for
communicative language teaching that is useful in the training of
ITAs. They state that a theory of communicative competence is
one in which there is a synthesis of knowledge of basic grammatical
principles, knowledge of how language is used in social contexts to
perform communicative functions, and knowledge of how utterances
and communicative functions can be combined according to the
principles of discourse. (Canale & Swain, 1980, p. 20)
They stress that the success of communication must be judged by
behavioral outcomes. Thus, although the communication problems
of ITAs are often perceived as a problem of English oral
proficiency effectiveness, we maintain that, for ITAs, the aim is
effective language usage while performing the role of TA. This
most general aim includes the objectives of language ability,
cultural awareness, and teaching skill.
In terms of language ability, what then, are the elements of
communicative success for the ITA? To begin, we again turn to
Canale and Swain’s model, as modified by Canale (1983), which
states that communicative competence consists minimally of
grammatical, sociolinguistic, discourse, and strategic competence.
Neither we nor Canale and Swain suggest that these are the only
sorts of competence necessary to successfully communicate in any
given context. Clearly, in the case of ITAs, for instance, knowledge
of the discipline is crucial. In addition, ITAs’ abilities to gauge the
know1edge and learning styles of their students may be as important


as communication skills. Finally, the desire to teach well and basic
personality variables such as confidence before a group also affect
the ability to communicate effectively. Without these basic
prerequisites, even an ITA with considerable competence in the
areas of language ability outlined below may be unsuccessful in the

Grammatical Competence
In the Canale and Swain (1980) framework, grammatical
competence refers to knowledge of the rules of morphology,
syntax, sentence-grammar semantics, lexical items, and phonology
“to determine and express accurately the literal meaning of
utterances” (p. 30). It is clear that the promotion of grammatical
competence is one of the goals of ITA training; however, it has
proved a particularly problematic area to address.
One reason is that ITA programs are usually multilevel, with
students showing widely varying proficiencies in the same class.
According to second language acquisition theory, learners acquire
the target language systematically, the process of acquisition
represents the movement toward proficiency in the target language
(Klein, 1986). Thus, a reasonable goal for the development of
second language grammatical competence would be facilitating the
acquisition of the target language through the ordered presentation
of material appropriate to the learner’s current level of proficiency
(Rutherford, 1988). But with learners at widely varying levels, such
facilitation within the classroom is a difficult and therefore,
perhaps, unrealistic goal. Moreover, there is often a lack of
reinforcing input or opportunities for English use outside the
classroom (Ard, 1987).
Perhaps most importantly, however, is the fact that the time for
training is short, whereas the improvement of grammatical
accuracy can be a time-consuming, long-term process. ITA courses
typically range from an average of 1 to 8 weeks preterm, or possibly
1 to 2 terms (several hours per week) concurrent with the academic
program (Barnes et al., 1989). During this relatively short time, ITA
programs are required to prepare students, who are often new to
the U.S. and may have limited language skills, to successfully
handle a demanding teaching role. Given the time constraints, and
keeping in mind the needs of the native English-speaking
undergraduates, the most effective means for an ITA to improve
communication may be through pragmatic rather than linguistic
means. For example, ITAs who mispronounce terms can become
more intelligible by learning to recognize which terms are


problematic for them so they can write them on the board; such
compensation techniques are quicker to master than changing
A third area of concern in teaching grammatical competence has
been the efficacy of traditional teaching methods. A case in point is
pronunciation. Although pronunciation is often the most commonly
identified problem associated with ITAs, in many cases it is
addressed only peripherally by the ITA curriculum. Stevens (1988,
1989) discusses this issue, proposing that the lack of emphasis on
pronunciation in the ITA curriculum may have stemmed from
trainers’ frustration in dealing with pronunciation problems. This is
not to deny that pronunciation still plays a major role in many
programs. One approach has been to focus on the pronunciation of
vocabulary in field-specific contexts, with change in these contexts
gradually applied to more general contexts (Anderson-Hsieh, 1990;
Byrd, Constantinides, & Pennington, 1989; Lane, 1989). Once again,
the emphasis is on success in the required role, rather than the
development of language skills in general.
Finally, there has been a confounding of linguistic and
nonlinguistic factors in how ITAs’ linguistic abilities are perceived
that has further confused the role of grammatical accuracy per se in
ITA speech. In Brown’s (1988) study, in which college students
viewed videotapes of nonnative speaker (NNS) teachers, the
perceived country of origin of the speaker affected the students’
judgments of language competence. In an earlier study, Orth (1982)
also demonstrated that students’ attitude toward the course and ex-
pected grade affected their perceptions of ITAs’ pronunciation.
Over the years, ITA training has increasingly moved to a position
where basic competence in English is seen as necessary before ITA
training can be effective. Bailey (1982) originally identified as a
threshold level an FSI (Foreign Service Institute) oral proficiency
rating of Level 2, below which ITAs were perceived as significantly
poorer performers. Some programs have instituted minimum
SPEAK scores for participation in training program activities
(Barnes et al., 1989). The relationship of these overall scores of
speaking proficiency to particular grammatical competencies has
varied, but has generally correlated more highly with pronunciation
and fluency than with grammar subsections (Clark & Swinton,

Sociolinguistic Competence
In the original Canale and Swain (1980) model, sociolinguistic
competence included both knowledge of the sociocultural rules of


speaking and rules for discourse. In Canale’s (1983) revised model,
only sociocultural rules are included. This area refers to the
appropriateness with which speakers produce and understand
language within a particular social context. In the case of ITAs, the
issue is how effectively they can use the language according to the
norms of interaction and interpretation (Hymes, 1972) of the
In this area, cultural differences in the classroom can be a primary
basis for ITAs’ failure (Landa & Perry, 1984). ITAs may have
difficulty adjusting to the teaching setting. Although some may have
held teaching positions in their own countries (Berns, 1989), the
classroom settings to which they are accustomed tend to be quite
different from those they confront as TAs in the U.S. In place of
large, impersonal lectures, they find themselves in front of small,
more intimate classes in which personal interaction is not only
possible, but expected. ITAs may nevertheless seek to transform
this new setting into one more familiar to them, leading their
undergraduates to perceive them as impersonal “mechanical
problem-solvers” (Bailey, 1984, p. 113). As these problems were
acknowledged, a culture component soon became standard fare in
ITA training programs (Turitz, 1984); this component often
included readings about U.S. universities and the goals of U.S.
education, and involved the use of undergraduates as informants
through ethnographic observations of classroom behavior (Pialorsi,
1984). Such curricular emphasis has been undertaken because of the
belief that understanding the cultural assumptions behind the U.S.
educational system and the differences between this system and the
ITA’s own will help the ITA-student communication process (Byrd
et al., 1989).
Overt attitudinal differences stemming from culturally different
educational values have proved difficult to measure, however.
Stevens (1989) developed an instrument for this purpose, but found
that ITAs showed few differences from U.S. undergraduates in
their overt attitudes towards the underlying values of U.S.
education, even in the case of ITAs from very different cultures.
Although these differences may be hard to measure by question-
naires, differences between educational systems and cultures
clearly affect most aspects of language behavior in the classroom,
including the appropriate role relationship between students and
TAs. If ITAs have never taught, a new role of authority relative to
students must be learned; if they have had teaching experience
before coming to the U. S., the role relationships must be relearned
in the U.S. context, a task that for some ITAs seems even more
difficult than for beginning teachers (Landa & Perry, 1984). The


understanding and appropriate handling of roles in teaching is more
difficult because of the many roles that instructors play in the
education process (McKeachie, 1978) and the permutations on those
roles expected of TAs. The easy phrasing of a native-speaking
chemistry TA’s “Shall we roll, folks?” as a way to signal the
beginning of class reveals the delicate balance of informality and
authority expected of TAs in the U.S. classroom (Hoekje & Tanner,
Unfortunately, providing the ITA with experience in a role of
authority is particularly difficult in the ITA course, given that the
ITA is a peer relative to the other students in the content class and
a student relative to the instructor in the training class. Example 1
illustrates the problems that may occur when the TA in a training
program assumes a role that differs from social reality.2 In this
example, a physics ITA was assigned to present an elementary
concept to his class—actually his fellow ITAs.
la. I want to ask a question about can you tell what’s the difference
speed and velocity? (General laughter followed by silence)
b. Yes, all of you are graduate students. It’s too easy.
This student was attempting to take on the role of the teacher but
was prevented from doing so because the class had not taken on its
role as students. It seems clear that practicing language items and
functions is ineffective when isolated from an appropriate context.
From this perspective, the particular value of Stevens’ (1988)
dramatic approach, in which ITAs learn to teach through acting is
that it encourages the serious adoption of the teaching role within
the training program curriculum.
Transfer of training is one of the most important issues facing ITA
programs today. It may even be necessary to go beyond role
playing to give ITAs authentic practice in their new capacities.
Increasingly, ITA programs are inviting the participation of
undergraduates. Although their perspective is not necessarily
unbiased, as noted above, the feedback that ITAs receive in a mock
lesson in front of undergraduates is far more valuable than the same
exercise in front of their fellow ITAs. One of the problems with the
inclusion of undergraduates in ITA training is, of course, cost. In
programs with restricted budgets, it may be more cost effective to
make use of actual classes. If ITAs are permitted to teach and take
an ITA preparation course concurrently, then real class material
may be used, either for real-time or delayed analysis. If ITAs are
2 All examples come from observations of ITAs made by the authors at the University of
Illinois at Chicago, the University of Pennsylvania, and Drexel University.


required to complete a preparation course prior to taking on
teaching responsibilities, it may be necessary to begin with more
receptive activities, such as observation and analysis of other
classes, taught by both NSTAS and experienced ITAs.
Bailey (1985) discusses the social differences that surface between
the training/testing program and a classroom of students. A number
of researchers who have examined ITA effectiveness conclude that
few ITAs are able to employ the range of speech styles that
contribute to classroom effectiveness (Bailey, 1984; Davis, 1984;
Orth, 1982); especially lacking is the use of humor. Bailey’s typology
of ITAs suggests that the most successful ITAs, in terms of
undergraduate evaluations, were ones who commanded a range of
speech styles, engaged in considerable interaction with students,
and used humor in their presentations. In contrast to the NSTA’S
genial class opener discussed above, we observed an ITA who had
charmed his fellow training program participants with his lucid and
entertaining presentations in the training program yet was unable in
the actual physics lab to find the appropriate tone for gathering
students to begin, other than announcing, “We start!”
In another example of formally correct language that failed
because of sociolinguistic inappropriateness, M. Tanner (personal
communication, Sept. 1990) describes an ITA circulating around the
students in the chemistry lab. She approached each student
workgroup, using the phrase, “What are you doing?” This phrase
had an accusatory tone that startled and alarmed the students but
was unintended in the situation; she was merely trying to elicit
information about what part of the procedure the students were
currently working on.3
Receptive skills are equally important for ITA effectiveness, and
it is often difficult for ITAs to know how to interpret the language
and behavior that is addressed to them or that they observe. Shaw
and Bailey (1990) describe the subtle negotiating of norms that
occurs between instructors and students in the first few weeks of a
course that establishes patterns of classroom behavior such as the
initiation and recognition of student questions. Without access to the
cultural norms of the U.S. classroom, ITAs are unable to participate
in this process effectively, neither communicating their own
expectations nor reading their students’ behavior accurately. Most
ITA programs go through the obvious points that eating in class and
3 Whether ITA preparation programs should try to modify ITAs’ personal styles is an ethical
question we will not attempt to address here. Stevens (1988) suggests that ITAs be
encouraged to take on different temporary roles rather than attempt to change long-term
behavior. It is clear, however, that a disparity is expected, and actual speech styles may
affect ITA success significantly.


tardiness are not necessarily signs of disrespect. Often, however,
misinterpretation can be more subtle and disturbing. As part of an
ITA training program, one participant taped one of his office hours.
When a student responded to his explanation with silence, followed
by some words of comprehension, the ITA interpreted the
encounter as successful. Later discussion with the undergraduate
revealed that she had simply said, “Yeah, OK, fine” in order to leave
the ITA’s office gracefully. In fact, having understood almost
nothing of the explanation, she abandoned any attempt to pursue
her questions, left frustrated, and sought an explanation from one of
her classmates. ITAs may also interpret a dwindling number of
students attending their sections as evidence of students’ lack of
seriousness in their studies rather than of a teaching problem.
More recently, there has been a growing realization that norms of
interaction and interpretation in the classroom also differ according
to academic discipline, Byrd and Constantinides (1988) have
pointed out that different disciplines have different preferred
teaching styles and warn ESL professionals not to assume that the
ESL style of teaching is appropriate in other contexts. Rounds
(1987) has described the language of the mathematics classroom in
terms of its particular routines and lesson organization. The nature
of the classroom, the assignments, and lessons affect the
organization of talk. Tanner (1991) investigated student and TA
questions in a chemistry laboratory as related to the particular
functions of that setting. In light of the differences in academic
disciplines, it is questionable whether the ESL staff of ITA
programs is in a position to give instruction. Cooperation with
faculty and experienced TAs in the disciplines is an obvious first
step toward remedying this lack. There are many ways in which this
can be accomplished. Faculty members from relevant departments
may be included at any stage: collection of materials, evaluation
and analysis of mock teaching, or as models for the ITAs to observe
and analyze, although not necessarily imitate. Some institutions
have begun to involve other faculty and NSTAs in a mentoring
system (B. Hawkins, personal communication, April 1989). During
the first term, the ITA team teaches with an experienced TA and
only in the second term do ITAs teach a class of their own and, even
then, continue to be observed by the mentor.
ESL faculty must also become more knowledgeable about field-
specific requirements. At one institution, a belated visit to the
chemistry department revealed that the faculty absolutely did not
want TAs giving presentations in the laboratory. They complained
that their new TAs were always trying to give lectures when,
instead, they should have been circulating through the lab,


answering individual questions. Because much of the ITA prepara-
tion course had focused on presenting information, departmental
needs and ITA training were clearly at cross purposes in this case. In
short, we are just beginning to understand what is involved in
adequate sociolinguistic competence for TAs. Although the infor-
mation provided in the typical culture component of the training
program is useful, it is only a very small part of the picture.

Discourse Competence
Grammatical and sociolinguistic competence, though important,
do not give a complete picture of ITA communicative competence.
It is equally important to know how to integrate these components
to produce and interpret cohesive and coherent discourse. Canale
(1983) has referred to this component as discourse competence,
encompassing both productive and interpretive abilities.
Cohesion refers to the ways in which utterances are connected so
as to produce unified oral or written text. There is evidence that
ITAs often fail to produce texts that attain the level of cohesion
necessary for easy interpretation. In the following excerpt, the
speaker, an L1 Mandarin ITA, at the end of an introduction to the
law of the conservation of energy, consistently overuses and over-
generalize the connectors and and so, often making the relation-
ships among propositions difficult to interpret. Of a group of 22
ESL professionals who listened to this presentation, 16 identified
the use of these connectors as a source of confusion:
2. A few days ago I met an American friend and he’s a very pious person
and he tell us—he told me—he recommend me study the Bible and
the first thing he told was that God create everything . . . so I think it
disobey the energy conservation law because according this law
energy cannot be create and cannot be destroy so that is a
contradiction and a very interesting problem.
Coherence is the other important element of discourse compe-
tence, according to Canale. Two areas of coherence, continuity and
progression, are essential aspects of ITA discourse. Continuity may
be indicated by repetition and rephrasing; progression by a variety
of discourse markers that identify the parts of an explanation as well
as the relationship among these parts. Adequate marking of this
kind has been shown to be an important element of comprehensi-
bility (Chaudron & Richards, 1986).
Discourse competence has been investigated in the context of
ITA production in a number of studies. The results of these studies
have yet to be fully incorporated into ITA curricula, but are


indicators of crucial elements of ITA effectiveness. Rounds (1987)
notes that in comparison to NSTAs, ITAs frequently fail to
elaborate adequately the key points of their presentations. They
often do not name important steps, mark junctures explicitly, or
make cohesive links between ideas. Tyler (1988) maintains that
unsuccessful ITAs consistently fail to orient their listeners
adequately to the relative importance of ideas and to how they are
linked to one another. According to Tyler, they misuse various cues
on which NS listeners depend in order to interpret discourse. These
include lexical, syntactic, and prosodic miscues. Taken together,
these can seriously reduce comprehensibility. Williams (in press)
found that two different sets of raters unequivocally rated as more
comprehensible ITA presentations during which discourse moves
were more overtly and explicitly marked than presentations with
fewer and less explicit discourse markers. The difference in the
level of syntactic and morphological errors between the two
productions, however, was insignificant, suggesting a relatively
greater role for discourse marking than for grammatical accuracy in
comprehensibility. It is not at all clear, however, that ITAs are
atuned to these sorts of discourse markers or to other language
features in the lectures they attend as students. Ard (1987), for
example, notes that ITAs attend to very little of these lectures,
relying instead on written information for the learning of content
material. Both the role of these discourse markers in increasing
comprehensibility and the ITAs’ possible lack of awareness of them
indicate this may be a productive area for consciousness-raising
(Rutherford, 1987). It may be that, beyond a certain threshold level
of grammatical competence, which we have yet to establish, it is
discourse competence that should be the focus of instruction.
Our knowledge of this area is far from comprehensive, however.
For example, although some rules can be taught for making an
effective oral presentation (e.g., include an introduction and
conclusion and use repetitions and clear framing statements), there
are many more features that have yet to be described, such as
appropriate encoding of topic and information focus, consistent
tracking of referents, and other elements of text cohesion, features
nonnative speakers often omit.
Even less understood is the area of receptive discourse compe-
tence. It is evident that ITAs often experience difficulties interpret-
ing questions directed at them, even if they understand the indi-
vidual words contained in them. In the following explanation of a
chemistry experiment, the ITA may have understood the lexical
items contained in the undergraduate’s (UG) question—at least he is


able to repeat them—yet he fails to interpret the meaning of the
3. ITA: Any question?
UG: Do we have to calculate the gas that’s left in the rubber tubing?
ITA: In the rubber tubing?
UG: Yeah, do we have to include it?
ITA: Pardon me?
UG: Do we have to calculate what’s left or is that given?
ITA: I don’t understand.
Although coherence and cohesion are the focus of Canale’s
discourse competence component, there are other important areas
of discourse that need to be explored in ITA program development.
For example, the difference between written and oral discourse has
been well documented (Biber, 1988; Danielewicz, 1984; Tannen,
1982). ITAs who rely on prepared notes in teaching often present
information orally in a style more appropriate for written texts. The
role of planning in the presentation of information may also be
important. The results of Williams (in press) indicate that planned
presentations, though syntactically more complex, are more
comprehensible than relatively unplanned presentations.
Other important areas of discourse competence involve
differences in competence that may be tied to specific discourse
domains. Most ITAs are well prepared in their disciplines but still
have difficulties transmitting this knowledge. Many studies of
second language production have indicated that topic may affect
accuracy, amount of production, and comprehensibility signifi-
cantly (Eisenstein & Starbuck, 1989; Gass & Varonis, 1984; Selinker
& Douglas, 1985; Woken & Swales, 1989; Zuengler, 1989).
Differences in discourse competence may be positive or negative
and have been variously attributed to expertise, perceived
expertise, and emotional investment. However, Smith (1989), in an
effort to apply some of these findings to ITA issues, devised a field-
specific version of the SPEAK.4 Her results yielded no significant
differences in language proficiency across topics for the ITAs as a
group, although individual differences did emerge. Smith focused
on domain-specific competence in terms of language ability, as
measured by the SPEAK. This is still an area for further
investigation. It may be that ITAs are relatively successful when
speaking in their area of expertise but less so in the interpersonal
interaction that is so important to successful teaching.
4 The
SPEAK test consists of older versions of the Test of Spoken English (TSE) that the
Educational Testing Service sells to institutions for in-house testing.


The studies named above have addressed differences in compre-
hensibility and accuracy across discourse domains in terms of lin-
guistic form. It is clear that discourse competence extends beyond
mere linguistic form involving the ways in which content is selected,
organized, and presented in the classroom. As noted in earlier
sections, there is growing acceptance of the fact that this may be
discipline specific (Byrd & Constantinides, 1988; Rounds, 1987;
Tanner, 1991). For example, the inclusion of subject-specialist
informants (SSIs) (Selinker & Douglas, 1989) can reveal content-
based problems with message form. During an ITA preparation
course, a chemistry SSI who attended a simulated problem-solving
session suggested that the most serious potential communication
problem of the new ITAs lay in their lack of knowledge of how the
undergraduate students expected chemistry problems to be written.
U.S. undergraduates are taught to carry their units (e.g., gm/mole)
all the way through the problem-solving sequence, then rationalize
them, and cancel where possible. Chinese students approach
problems in the opposite way: All rationalization and cancellation
of units occurs before the problem is solved. Thus, the ITAs were
presenting the problems in a form that was extremely confusing to
the undergraduates. This cannot be construed as a language
problem, but it affects message form, and thus communication, in
crucial ways. Furthermore, it was a problem that never would have
occurred to the ESL specialist in charge of the course. Most ESL
instructors do not have the background to provide this kind of
information and training. Domain-specific competence is clearly an
area that needs to be investigated further in ITA training and is of
potential importance to the fields of second language acquisition in
general and language for specific purposes (LSP) as well (see
Selinker & Douglas, 1989).

Strategic Competence
Strategic competence is described by Canale as the mastery of
verbal and nonverbal strategies that can either be used to
compensate for deficiencies in other areas of competence or to
increase communicative effectiveness in general. Because ITAs,
almost without exception, demonstrate gaps in the first three areas
of communicative competence, this component may prove to be a
crucial one. It may be possible to teach ITAs communication
strategies in order to make up for knowledge or abilities that are
weak in other areas; in this way, they may learn to use
compensatory strategies to increase their effectiveness as teachers.
On the nonverbal level, strategic competence could include the
knowledge of the advantages and disadvantages of particular


speech settings or participant structures in which communication
occurs. Tyler (1990) reports an escalating argument held in a lab
between an ITA and a student dissatisfied with his grade. The
discussion between the two takes place in the presence of other
students and in the face of time pressure on the ITA to supervise the
lab and on the student to get started with the lab work. As time goes
on, difficulty in communication and the mounting time pressure
make each speaker increasingly frantic and more rigid in his
position and increasingly unable to understand the other. As Tyler
points out, the decision to move this discussion to an office hour
setting where the two could talk more privately and leisurely is
probably the single best strategy for a more positive outcome.
On the verbal level, the notion of compensatory strategies is
somewhat alien to language-teaching programs, where targetlike
competence is usually the goal. Since many ITA courses are housed
within ESL programs that tend to use native-speaker proficiency as
the target for their students, it is not surprising that programs have
often adopted similar goals somewhat automatically. However, it is
important to stress that the goal of most ITA programs is to improve
communication skills in the teaching context, so as to better serve
the undergraduate population, not necessarily to foster the second
language acquisition of the ITAs. Although these two goals certainly
overlap, they are not identical and should not be confused.
Typically, the approach taken to second language education is
developmental; learners are guided through a series of stages of
acquisition of language rules and functions, with the eventual aim of
approaching an NS standard. Although it is clear to most ITA
program developers and instructors that this process needs to go
beyond linguistic competence, the assumptions regarding the goal
of an NS target has not been explicitly challenged. Various needs
analyses and other studies (Bailey et al., 1984; Gillespie, 1988;
Rounds, 1987; Williams, Barnes, & Finger, 1987) include the
collection of baseline data from NSs, suggesting that these studies
share the underlying assumption that ITAs should strive to emulate
effective NSTAs in terms of their linguistic, interactional, and
presentation skills. Since using an NS model as target is standard
procedure in second language acquisition research and, to an even
greater extent, in LSP, it is reasonable that such an assumption be
made in designing an ITA curriculum. Indeed, to do otherwise
smacks of assigning second class status to ITAs. However, a real
sense of urgency underlies the establishment of ITA preparation
programs since they play such an important role in higher
education. It is perhaps appropriate, therefore, to reexamine this
assumption of NS competence as the target.


First, it may be that NS-like behavior is not attainable, especially
in the short time usually allotted to ITA training. Second, and
perhaps more important, it is clear that the language production, in
terms of, for example, pronunciation, grammaticality, and lexis, of
effective ITAs does not always approach NS status; rather, ITAs are
successful as a result of compensatory strategies that they may be
taught to use. Bailey (1984) reports that one of the TAs who was
rated highest in her study was, in fact, an ITA. His English
proficiency was quite good but far from perfect. He was able to
overcome his language problems because his interpersonal and
pedagogical skills were exceptional. Bailey goes on to suggest,
however, that there may be a threshold level of language proficien-
cy below which compensatory strategies may be ineffective.
It may be that training in the use of compensatory strategies is
more effective than language instruction that uses NS behavior as a
target. Williams (in press) found that the increased use of discourse
markers and elaboration of key concepts led to significantly higher
comprehensibility ratings by undergraduate students and ESL
specialists. The use of these markers was found to be far more
significant for ITAs than for NSTAs, for whom discourse marking
and elaboration appeared to be relatively unimportant. This
disparity was presumably a result of the NSTAs’ high level of
competence in other areas as well as the ITAs’ lack thereof.
Nonlinguistic strategies may also be used; for example, the
extended use of written material, on handouts or an overhead
projector, may help to ameliorate communication difficulties. Many
of these are essentially avoidance, rather than achievement,
strategies (Faerch & Kasper, 1983); they reduce communicative
goals rather than expand linguistic resources. Again, such strategies
are at odds with the goals of general ESL programs. However,
although linguistic competence may not improve, there may be
gains in teaching effectiveness. These compensatory strategies
become particularly important in preparation courses aimed at
ITAs of widely differing linguistic competence, a situation typical
particularly of smaller institutions with limited resources (Boyd,
1989). Although it may be possible to maintain NS target models for
some of the more proficient students, the less proficient students,
too, must learn how to carry out the tasks that will be demanded of
them. The development of strategic competence may be an effec-
tive focus of instruction, especially for this population.

The assessment of ITA performance is clearly tied to the goals
and methods of training. Because the ITA problem has been viewed


historically as only a language problem—at least by some university
administrators and legislators—the assessment of ITAs’ oral English
language ability has often been relegated to oral interviews or tape-
recorded tests separate from the teaching task; however, more
performance-based tests involving teaching tasks are also
increasingly common (Plakans & Abraham, 1990). The most widely
used oral proficiency instrument in current use is the Test of Spoken
English (TSE) and its institutional version, SPEAK. This 20-minute
tape-recorded test requires the speaker to perform certain language
functions (e.g., narrate a story, give opinions, present information)
that are then rated for accuracy and fluency of grammar and
pronunciation. In addition to the subsection scores of grammar,
pronunciation, and fluency, an overall score on the TSE and SPEAK
is given for the speaker’s comprehensibility, expressed in points
from 0 to 300.
Another widely used instrument for assessing spoken language
proficiency is the Oral Proficiency Interview (OPI), originally
developed by the FSI, and later revised by the American Council on
the Teaching of Foreign Languages (ACTFL) in conjunction with
the Educational Testing Service (cf. Barnwell, 1987; Clark &
Clifford, 1988; Lowe & Stansfield, 1988, for an overview of OPI
testing in the U.S.). The OPI is a structured conversation between
two speakers in which examinees are asked to perform a variety of
speech functions during the interview—narrate and describe events,
express and support opinions—and are rated holistically on their
language, which is expressed as a language level such as
Intermediate High or Advanced Plus. Although the TSE and OPI
differ in approach, they derive from common principles and are
statistically correlated with each other. Clark and Swinton (1980)
and Stansfield and Ballard (1984) found a reasonable degree of
correlation between the proficiency rating of the OPI and the
comprehensibility rating of the TSE.
The use of these tests as measurements of spoken-language
proficiency has been challenged on a number of grounds (Valdman,
1988; van Lier, 1989). Critics of the OPI have challenged both the
tasks and the means of the interview, and the notion of proficiency
that it purports to measure (Bachman, 1988; Lantolf & Frawley,
1988). Clark and Swinton (1980) early identified the problem of
lower face and content validity with the tape-recorded TSE
compared to the FSI, a problem affecting user satisfaction with the
university-administered SPEAK as well (p. 1). Smith’s (1989)
investigation of more topic-specific versions of the SPEAK for
students in different disciplines was motivated in part by student
dissatisfaction with the SPEAK as a means of evaluation. Godfrey


and Hoekje (1990) questioned the validity of using NS norms in the
assessment of NNSs on the TSE or SPEAK on tasks where actual NS
behavior has not been measured. Another major area of concern for
both the OPI and the TSE/SPEAK is the meaning assigned to the
test result. A rating of Intermediate High on the OPI, for example,
may be interpreted as meaning that the speaker cannot control the
Advanced level function of narration and description in any setting.
Indeed, there is conflicting evidence on the extent to which
language proficiency scores on tests are representative of
proficiency in other contexts, such as the classroom. Clark and
Swinton (1980) implemented a use-validation study designed to
measure the relationship of TSE and FSI scores to the actual
communication effectiveness of ITAs in the classroom. They tested
ITAs at eight universities with TSE and FSI instruments and, for
those with classroom teaching responsibilities, surveyed the
students on the ITAs’ communication abilities with a battery of
language-related questions, for example, “When the instructor was
lecturing to the class, his or her English interfered with my
understanding of what was said” (p. 7). Questions on nonlanguage
factors such as perceived quality of textbooks or degree of
organization were also included as statistical controls. The results
showed strong correlation between scores on both the FSI and the
TSE and communication effectiveness in teaching as measured by
the language-related questions.
However, in an extensive study of ITA performance at the
University of California, Los Angeles (UCLA), Bailey (1982) found
only modest correlations between the five OPI categories and
actual undergraduate ratings of the ITAs’ oral proficiency. She
recorded even lower correlations between undergraduate ratings of
the ITAs’ teaching and certain OPI subscales. On the other hand, a
threshold proficiency level was identified by her results: As
indicated above, ITAs rated below Level 2 on the FSI were rated by
students as significantly worse in the classroom than ITAs rated
Level 2 or better. Bachman (1988) sees this issue as related to the
question of test reliability:
Rather than viewing a rating of, say, grammatical competence based on
an oral interview as a relatively unreliable indicator of this broad ability,
we must interpret it as a highly reliable indicator of a much more narrow
ability—an individual’s skill in using grammatical structures accurately
in contexts and under the conditions that are included in the testing
procedure. (p. 154)
In fact, naive judgments of language proficiency in real-life con-
texts may reflect nonlanguage factors, making their interpretation


more difficult. In the case of students rating instructors, a wide
variety of seemingly unrelated factors has been demonstrated to
affect perceptions of language proficiency. In the Clark and
Swinton (1980) study, overall ITA performance was not well
predicted by either the TSE or FSI scores (p. 45). Thus, for
example, a question on the overall value of the course to students
was not predicted by these language-test scores. However, as the
authors point out, there is no reason for which the language
proficiency scores should correlate as strongly with overall teaching
performance (to the extent that this includes nonlanguage factors)
as with communication effectiveness. As mentioned earlier, other
studies have also revealed the complexity of factors involved in
measuring student response to ITAs’ overall performance in the
classroom (Brown, 1988; Orth, 1982).
Oral proficiency tests have thus been challenged as valid
evaluation measures, both in terms of the construct they have
proposed to measure (language proficiency) and in terms of
applying the test score to other contexts, such as the classroom. ITA
programs increasingly have begun to turn to measurement of
performance in contexts that more closely resemble the actual
teaching task. The Taped Evaluation of Assistants’ Classroom
Handling (TEACH) test, in which the examinee performs a
teaching task that is videotaped and assessed by trained raters, was
developed at Iowa State University and soon widely distributed
(Abraham, Anderson-Hsieh, Douglas, Myers, & Plakans, 1988).
Many other programs have developed a variation on this activity
and begun to rely on it much more heavily than on the SPEAK as a
final assessment measure (Barnes et al., 1989). Little real
examination of the issues involved in this type of performance
testing has been raised apart from Bailey (1985). Bailey had been
involved in the development of a performance-based test for ITAs
at UCLA, in which the ITA was taped explaining a problem to a
student, and then assessed by a panel of raters. In her later
observation of ITAs in the classroom, she was struck by the
situational differences between the test and the actual task, such as
the authority difference between an examinee and the instructor,
and the one-to-one interaction in the test versus the group
interaction in the classroom. In Bailey’s assessment, situational
factors made ITAs’ performance in the classroom significantly
different from performance on even a performance-based test and
thus raised questions about the extent to which the test accurately
reflected the task. Bailey also raised the question of legal fairness in
moving toward testing that is more performance-based: “As a
screening test for foreign TAs focuses less on purely linguistic


competence and more on the specific functional skills involved in
teaching, it will become, paradoxically, less fair, and presumably,
less acceptable. At what point do we stop testing contextualized
oral English proficiency and start testing teaching ability instead?”
(p. 167). Interestingly, Bailey concluded that one alternative was a
less teaching-performance-based measure, such as the FSI or TSE/
New models of assessment such as that proposed in Bachman
(1990), however, depend crucially on the conception of communi-
cative language ability as involving the various competencies dis-
cussed here and have the potential to resolve the current perfor-
mance versus nonperformance testing dichotomy in ITA assess-

Until the ITA preparation curriculum is truly contextualized,
many ITA trainers will continue to find themselves in the double
bind to which Bailey alludes. If they pursue what may be seen as the
ethical course of limiting preparation courses to language teaching
with a bit of culture thrown in, they will surely fail. If, on the other
hand, the goal of ITA training is viewed as preparing the student to
effectively take on the role of TA with all that entails (teaching,
managing the classroom, advising), as we have proposed here, they
lay themselves open to charges of discrimination, either against
ITAs, who are forced to do extra work, or against NSTAs, who did
not have access to special training (see Brown, Fishman, & Jones,
1990). The central question that has been put to our field is whether
we as ITA educators have the right or duty to provide nonlinguistic
training to ITAs when it is not provided to NSTAs. Knowing that
such training may, in fact, be more effective in improving
classroom performance than attempts to improve linguistic
competence, is it possible or advisable to reject it in the name of a
presumed legal parity? We have tried to argue here that the notion
of language skills in a narrow sense keeps us firmly locked into this
dilemma. We have claimed that ITA education that does not
include the rich context we have described here can result only in
limited success. Therefore, the only way out of this dilemma is to
aim for parity, to provide education and support for all TAs who
lack classroom experience. It is unfortunate that, in this time of
contracting resources, this is an unlikely scenario. We hope that a
reconceptualization of language skills in terms of communicative
competence provides the theoretical means with which to begin to
recast this issue. The ITA curriculum has developed from the


perceived needs of ITAs in language, pedagogy, and culture, and
researchers have continued to add information about ITA
communication, discourse genres, and the U.S. classroom. We see a
continuing need for contextualized interlanguage research into what
Selinker and Douglas (1989) have called “real-life and important
contexts” (p. 93). As Hymes (1972) proposed 20 years ago, language
is used within a social context; the ITA classroom is such a social
context, and there is little point in teaching language items without
situating them within this context. The use of the communicative
competence framework as an organizing principle for ITA training
provides a principled basis for evaluating and using research, inte-
grating the components of curriculum, and assessing language use.
From this perspective, the improvement of language proficiency, as
mandated by state laws and educational institutions, has no mean-
ing without a direct acknowledgment of the social situation of its

Barbara Hoekje is Associate Director of the English Language Center at Drexel
University. She has written and presented on ITA education since 1984.

Jessica Williams is Assistant Professor of Linguistics and Director of the

International Teaching Assistant Program at the University of Illinois at Chicago.

Abraham, R., Anderson-Hsieh, J., Douglas, D., Myers, C., & Plakans, B.
(1988). Evaluation and instructional program for international teaching
assistants. Paper presented at the 22nd Annual TESOL Convention,
Chicago, IL.
Anderson-Hsieh, J. (1990). Teaching suprasegmentals to international
teaching assistants using field-specific materials. ESP Journal, 9, 195-214.
Ard, J. (1987). The foreign TA problem from an acquisition-theoretic
point of view. English for Specific Purposes, 6, 133-144.
Bachman, L. (1988). Problems in examining the validity of the ACTFL
oral proficiency interview. Studies in Second Language Acquisition, 10,
Bachman, L. (1990). Fundamental considerations in language testing.
Oxford: Oxford University Press.


Bailey, K. M. (1982). Teaching in a second language: The communicative
competence of non-native speaking teaching assistants. Unpublished
doctoral dissertation, University of California, Los Angeles.
Bailey, K. M. (1984). A typology of teaching assistants. In K. M. Bailey,
F. Pialorsi, & J. Zukowski/Faust (Eds.), Foreign teaching assistants in
U.S. universities (pp. 110-125). Washington, DC: NAFSA.
Bailey, K. M. (1985). If I had known then what I know now: Performance
testing of foreign teaching assistants. In P. Hauptman, R. Leblane, &
M. Wesche (Eds.), Second language performance testing (pp. 153-180).
Ottawa: University of Ottawa Press.
Bailey, K. M., Pialorsi, F., & Zukowski/Faust, J. (Eds.). (1984). Foreign
teaching assistants in U.S. universities. Washington, DC: NAFSA.
Barnes, G., Finger, A., Hoekje, B., & Ruffin, P. (1989). T o w a r d s
standards: A survey of ITA programs. Paper presented at the 23rd
Annual TESOL Convention, San Antonio, TX.
Barnwell, D. (1987). Oral proficiency testing in the United States. British
Journal of Language Teaching, 25, 35-42.
Berns, M. (1989). An ITA profile and its implications for course design. In
G. Barnes, M. Berns, & C. Madden (Chairs), The training of intern-
ational teaching assistants. Preconference symposium presented at the
23rd Annual TESOL Convention, San Antonio, TX.
Biber, D. (1988). Variation across speech and writing. C a m b r i d g e :
Cambridge University Press.
Boyd, F. (1989). Effective programs for small institutions. In G. Barnes,
M. Berns, & C. Madden (Chairs), The training of international teaching
assistants. Preconference symposium presented at the 23rd Annual
TESOL Convention, San Antonio, TX.
Brown, K. (1988). Effects of perceived country of origin, educational
status, and native speakerness on American college student attitudes
toward non-native instructors. Unpublished doctoral dissertation,
University of Minnesota, Minneapolis.
Brown, K., Fishman, P., & Jones, N. (1989). Language proficiency
legislation and the ITA. Paper presented at the 41st Annual NAFSA
Conference, Minneapolis, MN.
Brown, K., Fishman, P., & Jones, N. (1990). Legal and policy issues in the
language proficiency assessment of international teaching assistants
(Monograph No. 90-1). Houston: The University of Houston Law
Center, Institute for Higher Education Law and Governance.
Byrd, P., & Constantinides, J. (1988). FTA training programs: Searching
for appropriate teaching styles. English for Specific Purposes, 7,
Byrd, P., Constantinides, J., & Pennington, M. (1989). The Foreign
Teaching Assistant’s Manual. New York: Collier MacMillan.
Canale, M. (1983). From communicative competence to language
pedagogy. In J. Richards & R. Schmidt (Eds.), Language and
communication (pp. 2-28). London: Longman.


Canale, M., & Swain, M. (1980). Theoretical bases for communicative
approaches to second language teaching and testing. Applied
Linguistics, 1, 1-47.
Chaudron, C., & Richards, J. (1986). The effect of discourse markers on
the comprehension of lectures. Applied Linguistics, 7, 113-127.
Clark, J., & Clifford, R. T. (1988). The FSI/ILR/ACTFL proficiency
scales and testing techniques: Development, current status, and needed
research. Studies in Second Language Acquisition, 10, 129-147.
Clark, J., & Swinton, C. (1980). The Test of Spoken English as a measure
of communicative ability in English-medium instructional settings
(TOEFL Research Rep. No. 7). Princeton, NJ: Educational Testing
Danielewicz, J. (1984). The interaction between text and context: A study
of how adults and children use spoken and written language in four
contexts. In A. Pellegrini & T.. Yawkey (Eds.), The development of oral
and written language in social contexts (pp. 243-260). Norwood, NJ:
Davies, C. (1991). Pragmatic feedback in ITA training. In G. Barnes,
M. Berns, & C. Madden (Chairs), The training of international teaching
assistants. Preconference symposium presented at the 25th Annual
TESOL Convention, New York, NY.
Davis, B. (1984). A study of the effectiveness of training for foreign
teaching assistants. Unpublished doctoral dissertation. The Ohio State
University, Columbus.
Eisenstein, M., & Starbuck, R. (1989). The effect of emotional investment
in L2 production. In S. Gass, C. Madden, D. Preston, & L. Selinker
(Eds.), Variation in second language acquisition: Psycholinguistic issues
(pp. 125-140). Clevedon, England: Multilingual Matters.
Faerch, C., & Kasper, G. (1983). Plans and strategies in foreign language
communication. In C. Faerch & G. Kasper (Eds.), Strategies in
interlanguage communication. London: Longman.
Gass, S., & Varonis, E. (1984). The effect of familiarity on comprehensi-
bility of non-native speech. Language Learning, 34, 65-90.
Gillespie, J. (1988). Foreign and U.S. teaching assistants: An analysis of
verbal and nonverbal classroom interaction. Unpublished doctoral
dissertation, University of Illinois, Urbana.
Godfrey, D., & Hoekje, B. (1990). Native speaker norms on SPEAK. Paper
presented at the 24th Annual TESOL Convention, San Francisco, CA.
Hoekje, B., & Tanner, M. (1987). Establishing authority in the classroom.
Paper presented at the 21st Annual TESOL Convention, Miami, FL.
Hutchinson, T., & Waters, A. (1987). English for specific purposes.
Cambridge: Cambridge University Press.
Hymes, D. (1972). Foundation in sociolinguistics. Philadelphia: University
of Pennsylvania Press.
Klein, W. (1986). Second language acquisition. Cambridge: Cambridge
University Press.


Landa, M., & Perry, W. (1984). An evaluation of a training course for
foreign teaching assistants. In K. M. Bailey, F. Pialorsi, & J. Zukowski/
Faust (Eds.), Foreign teaching assistants in U.S. universities
(pp. 89-100). Washington, DC: NAFSA.
Lane, L. (1989). ITA pronunciation strategies. In G. Barnes, M. Berns, &
C. Madden (Chairs), The training of international teaching assistants.
Preconference symposium presented at the 23rd Annual TESOL
Convention, San Antonio, TX.
Lantolf, J. P., & Frawley, W. (1988). Proficiency: Understanding the
construct. Studies in Second Language Acquisition, 10, 181-195.
Lowe, P., & Stansfield, C. (Eds.). (1988). Language in education: Theory
and practice: Vol. 70. Second language proficiency assessment: Current
issues. West Nyack, NY: Prentice Hall.
McKeachie, W. J. (1978). Teaching tips. Lexington, MA: D. C. Heath.
Myers, C. (1991). Discourse and training for lab assistants. In G. Barnes,
M. Berns, &C. Madden (Chairs), The training of international teaching
assistants. Preconference symposium presented at the 25th Annual
TESOL Convention, New York, NY.
Orth, J. (1982). University undergraduate evaluational reactions to the
speech of foreign teaching assistants. Unpublished doctoral dissertation,
University of Texas, Austin.
Pialorsi, F. (1984). Toward an anthropology of the classroom: An essay on
foreign teaching assistants and U.S. students. In K. M. Bailey,
F. Pialorsi, & J. Zukowski/Faust (Eds.), Foreign Teaching Assistants in
U.S. Universities (pp. 16-21). Washington, DC: NAFSA.
Pica, T., Barnes, G., & Finger, A. (1990). Teaching matters: Skills and
strategies for international teaching assistants. New York: Newbury
Plakans, B., & Abraham, R. (1990). The testing and evaluation of
international teaching assistants. In D. Douglas (Ed.), English language
testing in U.S. colleges and universities (pp. 68-81). Washington, DC:
Rounds, P. (1987). Characterizing successful classroom discourse for NNS
teaching assistants. TESOL Quarterly, 21 (4), 643-671.
Rounds, P. (1990). Student questions: What do we know about them? In
G. Barnes, M. Berns, & C. Madden (Chairs), The training of interna-
tional teaching assistants. Preconference symposium presented at the
24th Annual TESOL Convention, San Francisco, CA.
Rutherford, W. (1987). Second language grammar: Learning and teaching.
New York: Longman.
Rutherford, W. (1988). Aspects of pedagogical grammar. In
W. Rutherford & M. Sharwood Smith (Eds.), Grammar and second
language teaching (pp. 171-185). New York: Newbury House.
Savignon, S. (1983). Communicative competence: Theory and classroom
practice. Reading, MA: Addison-Wesley.
Selinker, L., & Douglas, D. (1985). Wrestling with ‘context’ in interlan-
guage theory. Applied Linguistics, 6, 191-204.


Selinker, L., & Douglas, D. (1989). Research methodology in contextu-
ally-based second language research. Second Language Research, 5,
Shaw, P., & Bailey, K. M. (1990). Cultural differences in academic set-
tings. In R. C. Scarcella, E. S. Andersen, & S. D. Krashen (Eds.), De-
veloping communicative competence in a second language
(pp. 317-328). New York: Newbury House.
Shaw, P., & Garate, E. (1984). Linguistic competence, communicative
needs, and university pedagogy: Toward a framework for TA training.
In K. M. Bailey, F. Pialorsi and J. Zukowski/Faust (Eds.), Foreign
teaching assistants in U.S. universities (pp. 22-40). Washington, DC:
Smith, J. (1989). Topic and variation in ITA oral proficiency: SPEAK and
field-specific tests. English for Specific Purposes, 8, 155-167.
Smith, J., Meyers, C., & Burkhalter, A. (1992). Communicate: Strategies
for international teaching assistants. Englewood Cliffs, NJ: Prentice Hall
Stansfield, C., & Ballard, R. (1984). Two instruments for assessing the oral
proficiency of foreign teaching assistants. In K. M. Bailey, F. Pialorsi, &
J. Zukowski/Faust (Eds.), Foreign Teaching Assistants in U.S. Universi-
ties (pp. 101-109). Washington, DC: NAFSA.
Stevens, S. (1988). Improving the international teaching assistant
experience: An evaluative study of a training program. Unpublished
doctoral dissertation, University of Delaware, Newark.
Stevens, S. (1989). A “dramatic” approach to improving the intelligibility
of ITAs. English for Specific Purposes, 8, 181-194.
Tannen, D. (1982). Oral and literate strategies in speech and written
narrative. Language, 58, 1-21.
Tanner, M. (1991). NSTA-student interaction: An analysis of TA’s
questions and students’ responses in a laboratory setting. Unpublished
doctoral dissertation, University of Pennsylvania, Philadelphia.
Turitz, N. (1984). A survey of teaching programs for foreign teaching
assistants in American universities. In K. M. Bailey, F. Pialorsi, &
J. Zukowski/Faust (Eds.), Foreign teaching assistants in U.S. universi-
ties (pp. 43-50). Washington, DC: NAFSA.
Tyler, A. (1988). Discourse structure and coherence in international TAs’
spoken discourse. Paper presented at the 22nd Annual TESOL
Convention, Chicago, IL.
Tyler, A. (1990). Examining cross-linguistic miscommunication. In
G. Barnes, M. Berns, & C. Madden (Chairs), The training of interna-
tional teaching assistants. Preconference symposium presented at the
24th Annual TESOL Convention, San Francisco, CA.
Valdman, A. (Ed.). (1988). The assessment of foreign language oral
proficiency [Special issue]. Studies in Second Language Acquisition,
10, (2).
van Lier, L. (1989). Reeling, writhing, drawling, stretching and fainting in
coils: Oral proficiency interviews. TESOL Quarterly, 23 (3), 489-508.


Widdowson, H. (1978). Teaching language as communication. Oxford:
Oxford University Press.
Williams, J. (in press). Planning, discourse marking, and the comprehensi-
bility of second language speakers. TESOL Quarterly.
Williams, J., Barnes, G., & Finger, A. (1987). FTAs: Report on a needs
analysis. Paper presented at the 21st Annual TESOL Convention, Miami
Beach, FL.
Woken, M., & Swales, J. (1989). Expertise and authority in native—
non-native conversations: The need for a variable account. In S. Gass,
C. Madden, D. Preston, & L. Selinker (Eds.), Variation in second
language acquisition: Discourse and pragmatics (pp. 211-227).
Clevedon, England: Multilingual Matters.
Young, R. (Ed.). (1989). The training of international teaching assistants
[Special issue]. ESP Journal, 8 (2).
Zuengler, J. (1989). Performance variation in NS-NNS interactions:
ethnolinguistic differences or discourse domain? In S. Gass, C. Madden,
D. Preston, & L. Selinker (Eds.), Variation in second language
acquisition: Discourse and pragmatics (pp. 228-244). Clevedon,
England: Multilingual Matters.


TESOL QUARTERLY, Vol. 26, No..2, Summer 1992

ESL Student Bias in

Instructional Evaluation
University of Washington

This paper reports on a statistical analysis of ESL student

evaluations of teachers in two large ESL programs. Evaluation
responses were collected along with information on students’
cultural and personal backgrounds. Results indicated that
systematic bias occurs in ESL student evaluations due to ethnic
background, level of English, course content, and attitude toward
the course. Another finding was a student perception of subgroups
among the evaluation questions, implying that not all questions
should be counted equally. These results raise issues of fairness in
the use of student evaluations of ESL teachers for purposes of
personnel decisions.

In discussions of instructional assessment, educators and

researchers have expressed reservations over the use of student
evaluations of their teachers. It is not clear to what extent they
should be used to make decisions about hiring, firing, promotion,
and other personnel issues. One area of concern is that student
evaluations do not provide enough information to cover the
complexity of the instructional process. McKeachie (1979) points
out that student evaluations, though valid in measuring student
attitude and motivation, cannot provide a complete picture of a
teacher. In his words, “when using student ratings to evaluate
teaching, we should remember that students cannot judge all
aspects of teaching effectiveness equally well” (p. 390).
The Committee on the Evaluation and Improvement of Teaching
at the University of Washington (1982) agrees that, while student
evaluations do provide valuable information, they should not be
used as the sole criterion for personnel decisions. The committee
recommends self-evaluation, peer evaluation, and administrator
appraisal as other reliable measures to use in conjunction with
student evaluations. It further advises that student evaluations be
collected over an extended period of time and in several different

Pennington and Young (1989) also recommend that a variety of
measures be used in an evaluation process. They propose that
different aspects of teaching be assessed at various stages of a
teacher’s professional development and note that student
evaluations may be more appropriate at certain developmental
stages than at others.
Another area of concern regarding student evaluations is that of
instrument development. A large body of research at U.S.
universities has centered on testing the reliability and validity of
instruments (Centra, 1972; Gillmore, Kane, & Naccarato, 1978;
Marsh, 1981; McKeachie, 1979). Out of this research have come
several standardized evaluation instruments which are still widely
used, such as Students’ Evaluations of Educational Quality (SEEQ)
(Marsh, 1982), Endeavor (Frey, Leonard, & Beatty, 1975), and the
Michigan State University Student Instructional Rating System
(SIRS) (Warrington, 1973).
Also related to the issue of instrument development is whether an
evaluation questionnaire measures more than one dimension of
teaching, since teaching is obviously a complex activity with many
components working together to make a successful class. Marsh
(1984), who has done extensive research on evaluations in U.S.
universities, warns of the danger of lumping all evaluation questions
into one category:
If a survey instrument contains an ill-defined hodgepodge of items, and
student ratings are summarized by an average of these items, then there
is no basis for knowing what is being measured, no basis for
differentially weighting different components in the way most
appropriate to the particular purpose they are to serve, nor any basis for
comparing these results with other findings. (p. 709)
To address this problem, a number of studies have been
conducted in U.S. university classes to distinguish the effects of
multiple dimensions of teaching (Frey, Leonard, & Beatty, 1975;
Marsh, 1982; Warrington, 1973). Such studies use factor analysis,
which extracts a small number of useful factors from a data set. For
example, as revealed by factor analysis, the Michigan State
University SIRS instrument measures five different aspects of
teaching: instructor involvement, student interest and performance,
student-instructor interaction, course demands, and course
organization. Another factor analysis by Frances and Gruber (1981)
found two dimensions, which they call instructor evaluation and
student motivation. Factor analysis on the Endeavor and SEEQ
instruments revealed seven and nine dimensions, respectively
(Marsh, 1984).


A final area of concern is the question of student bias, or lack of
objectivity in rating. In other words, there is the fear that students
might be swayed in their ratings by circumstances outside of a
teacher’s control. Again we find that this question has been widely
investigated with respect to U.S. university students. There are
several studies using multiple regression analysis which have found
bias due to such variables as gender and academic field of the
student (Hearn, 1985; Neumann & Neumann, 1981; Pedhazur,
1982), educational level (Tollefson & Wigington, 1986), and class
size (Cashin, 1985), to name just a few.
On the other side of the bias question, many educational experts
insist that bias is not a significant problem in U.S. university student
evaluations (e.g., Aleamoni, 1981; Centra, 1979; Marsh, 1980, 1984,
1987; McKeachie, 1979). This judgment is based on extensive studies
of validity and reliability measures of evaluation instruments.
Among ESL populations, these questions about student
evaluations become even more acute. As Pennington and Young
(1989) point out, evaluation instruments which have been tested for
validity and reliability among native speakers are not necessarily
appropriate for ESL students. They recommend that “the instru-
ments and procedures [be] constructed by evaluation specialists
sensitive to the nature of the ESL context” (p. 630).
The bias question becomes particularly complex in ESL
populations, due to students’ vastly different cultural backgrounds,
which affect attitudes toward education and toward the teacher-
student relationship. Pennington and Young argue that
because of limitations of language or cultural inhibitions, these students
may be unable or unwilling to communicate as freely as native speakers
would in written or oral forms of faculty evaluation. It can also be
argued that rating scales, frequently used for faculty evaluation, are of
questionable validity when employed by those whose exposure to
English and to the American educational cultural context is limited.
(p. 629)
Concern for confidentiality may also have an influence on student
ratings, especially among immigrants and refugees, reluctant as
they may be to jeopardize their status in their new country. The
very idea of passing judgment on a teacher may be inconceivable to
someone from a culture in which the teacher is revered as an
authority figure.
To summarize these problems as they relate to TESL, evaluation
instruments may not be comprehensible, valid, or reliable; they may
not reflect the multidimensionality of teaching; and it may be the
case that ESL students approach the evaluation process with certain
cultural biases.


It was the recognition of such problems which led to the present
study, conducted at the University of Washington Extension
English as a Second Language Programs in 1987-1989. In
attempting to develop a fair performance appraisal system for the
purpose of making a variety of personnel decisions, the role of
student evaluations came under severe scrutiny. The study, outlined
in the following pages, investigated the questions of bias and
instrument development using factor analysis to look for multiple
dimensions among the evaluation questions, and multiple regression
analysis to look for bias.

Interviews with Teachers
In a series of informal, individual interviews with 16 teachers, a
skepticism about the fairness of student evaluations was revealed, as
summarized below.
The first concern was that low-proficiency-level ESL students,
because of limited English ability, would not be able to understand
questions presented to them on evaluation forms. A second concern
was that student dissatisfaction with program-wide constraints such
as testing procedures, required courses, and the like, would be
reflected negatively in student evaluations, thereby counting against
the teacher. The third, and perhaps the greatest, fear among
teachers was of student bias. Teachers wondered whether certain
cultures would rate more strictly than others, whether certain
courses were more popular than others regardless of who was
teaching them, and whether higher level students were stricter due
to a perception of a slower rate of progress in a course.

Pilot Study
A pilot study was conducted among 19 teachers in 35 classes of
ESL students in an intensive English program. Students received
surveys about their cultural and personal backgrounds, with
questions on country, age, gender, level of English, and course type;
the survey asked whether or not students were repeating the course.
They also received a course evaluation questionnaire, the language
of which was simplified for ESL students. A multiple regression
analysis similar in form to that subsequently used in the principal
study tested for bias. Afterwards teachers and students were
interviewed about the process.
Two important results were obtained from this pilot study. First,
the multiple regression analysis showed some preliminary


indications of student bias, thus validating the need for the study.
Second, through follow-up interviews with students and a survey
soliciting teachers’ comments, we found that parts of our
instrument, even in its simplified state, were being misinterpreted,
especially by low-level students. A series of individual conferences
were then held with students from various language backgrounds to
review the content of each question and determine what vocabulary
was confusing. Thus we were able to further adjust the language of
our evaluation instrument, thereby increasing the validity of its

Principal Study
The principal study was conducted over four academic quarters
(one year) in each of two ESL programs, a required academic
English program for matriculated university students scoring 500-
580 on the TOEFL (Test of English as a Foreign Language), and an
intensive English program of five levels from beginning to
advanced. Teachers were required to distribute a background
survey and 16-question evaluation form in each class. The
evaluation form, shown in Figure 1, also included two questions to
measure student attitude. Each question was rated on a 5-point scale
from 0 to 4. Students were told that the purposes of the evaluation
were both to provide feedback to their teachers and data for
research. It was made clear that none of the data would affect the
teachers’ employment status. In the academic program there were
522 evaluations collected, a 62% response rate; in the intensive
program there were 2,658, a 69% response rate. (Nonresponse was
due either to student absence or, in some cases, to teachers’ not
distributing the forms.)
Factor analysis was used to test for multidimensionality, that is, to
determine whether students distinguish certain subgroups among
the 16 evaluation questions. On the model of Marsh (1982), we
conducted the factor analysis using a Kaiser normalization and an
Oblimin rotation. Cluster analysis was also used to look for sub-
groups among the evaluation questions. While cluster analysis
preserves the identity of individual questions, grouping them into
larger and larger groups based on their similarity (see Norussis,
1988, pp. 71-101), factor analysis actually constructs single factors
from combinations of multiple questions. These two complemen-
tary approaches both shed light on the internal structure of these
multivariate data sets.
Neither of these analyses, however, incorporates the background
variables, so multiple regression analysis was used for this purpose.


Evaluation Questionnaire

We performed a multiple regression analysis to test for bias due to

the variables on the background survey. This method was chosen
because it allows for the measurement of the effect of one back-
ground variable on evaluation rating, while holding all other vari-
ables constant. The t tests for these regression coefficients provide
formal statistical hypothesis tests for bias.
The background variables measured in the intensive program
were ethnic background, age, course content, level of English,
attitude, length of time in the program, and gender. For the
academic program, three additional variables were measured: visa
status, academic level (graduate or undergraduate), and time in the
U.S. The statistical computer program used for all analyses was

Tests for Subgroups
In testing whether there were subgroups among the 16 evaluation
questions, we looked at the two programs independently. For the
intensive program, factor analysis suggests that the evaluation
questions fall into two subgroups, reflecting what we call a teacher


factor and a learning experience factor. We labeled them with these
terms because we noted that all the questions in the first subgroup
were directly related to the teacher, whereas the questions in the
second subgroup had to do with the student’s learning experience
due to other aspects of the course and the program as a whole.
The teacher factor explains 57.3% of the variance, with an
eigenvalue of 9.1718. The learning experience factor had an
eigenvalue of .9834. This value is low, indicating that this factor is
less pronounced than the teacher factor, but we included it because
it explains 6.1% more of the variance, for a total of 63.5%. The factor
loadings are presented in Table 1.

Factor Loadings for the Intensive Program

Factor 1 Factor 2

The dendrogram in Figure 2 shows the results of a cluster analysis

which lends further support to the conclusion that there are distinct
subgroups among the 16 evaluation questions. The dissimilarity
coefficient expresses how different two questions are from each
other. A well-defined group of questions falls in the middle of the
graph, which again reflects the presence of a teacher subgroup.


Cluster Analysis of Evaluation Data in the Intensive Program

Also, as in the factor analysis, the second group is less pronounced.

Note, however, that Question 14 (I have a chance to talk in this
class) falls outside of the main cluster in this analysis, unlike in the
corresponding factor analysis.
Moving to the academic program, we see similar results. Factor
analysis reveals two factors (see Table 2). Both factors together
explain 57.8% of the variance, with the teacher factor explaining
49.3%, eigenvalue 7.8775, and the learning experience factor
explaining an additional 8.5% of the variance, eigenvalue 1.3633.
Figure 3 shows the dendrogram for the cluster analysis of the
academic program. In this plot there are two distinct clusters, again
reflecting the presence of teacher and learning experience
subgroups among the questions. However, here note that Question
16 (I hope my friends will study with this teacher) is in the teacher
subgroup, whereas in the factor analysis, it fell into the learning
experience subgroup.
From these factor and cluster analyses, it appears that most
questions fall squarely into two subgroups, teacher and learning
experience, while a few questions (6, 12, 14, 16) shift between the
two depending on the nature of the ESL program and the analysis
used. Taking all these results and our knowledge of the ESL
programs into account, we established two categories of questions,


Factor Loadings for the Academic Program

teacher (T) and learning experience (L), which can be seen in

Figure 1.
This configuration was then used for further analyses in both
programs by taking averages within each subgroup. In the intensive
program, the mean evaluation rating for the teacher subgroup of
questions was 3.239 with a standard deviation of .718, on a scale of
0-4, 4 being the highest. For the learning experience subgroup,
M = 2.923, SD = .845. In the (required) academic program, for the
teacher subgroup, M = 3.392, SD = .578, and for the learning
experience subgroup, M = 2.789, SD = .835. This suggests that, on
average, negative feelings toward the required program were
vented, not at the teacher, but at the learning experience.
A score for student attitude was measured by taking an average
of two questions (see Figure 1), measured on the same 0–4 point
scale with 4 being the most positive attitude. For attitude in the
intensive program, M = 2.986, SD = .915; in the academic program,
M = 2.532, with SD = 1.02.


Cluster Analysis of Evaluation Data in the Academic Program

Multiple Regression Analysis

Tables 3-6 show the results of four multiple regression analyses,
one for each subgroup of evaluation questions within each of the
two programs.1 Forced entry regression analysis, a method which
forces all specified variables into the equation, was chosen. In the
analysis, we only used data from countries with a large number of
students represented. In the intensive program we chose those
nationalities from which we received at least 35 evaluations. In the
academic program, due to a smaller sample size, the cutoff number
was 20.
Table 3 shows the analysis for the intensive program looking only
at the teacher subgroup. Note that the regression coefficients can be
interpreted in actual point value. With numeric variables such as
level, or time in program, the regression coefficient reflects the
change in evaluation score per increment. For example, the
1 The variance explained (R2), as shown in Tables 3-6, although significant, is relatively
small. However, the goal of the study is not to explain all of the variability in student ratings
but rather to identify systematic biases, assignable to particular causes. These are directly
meaningful on the scale of rating points despite the relatively small percent of variance


regression coefficient for level is – .123. For each level that the
student moves up (low, intermediate, advanced), the rating will
decrease by .123 on average. For nonnumeric variables such as
country, or course type, the regression coefficient must be
compared to a baseline, as indicated in the tables for each category.
For example, the regression coefficient for Indonesia is .256, which
means that Indonesian students rated .256, or about 1/4 of a point
higher on average than Japanese students, the baseline.
Thus from Table 3, we see that students who self-identified as Chi-
nese (predominantly from the People’s Republic of China), Latin
American, Indonesian, and Arabic speaking, in that order,2 on aver-
age gave higher ratings than Japanese students. Ratings in reading
and writing courses were lower on average than those in grammar
courses. Students in more advanced levels gave lower ratings on aver-
age than those in lower levels. Similarly, the longer the student had
been in the program, the lower s/he rated, but only by a very small
amount. The other variables did not have significant effects.

Multiple Regression Analysis of Teacher Subgroup in the Intensive Program

Variable coefficient SE P
Ethnic background (Japan = baseline)
Latin American
Arabic speaking
Course type (Grammar = baseline)
Time in program

2 Students self-selected ethnic backgrounds from a list of categories which represented the
typical student population of the programs.


In Table 4, the effects of student background on the learning
experience subgroup of questions in the intensive program can be
seen, with Indonesians, Chinese, Latin Americans, and Arabic
speakers, again rating higher on average than the Japanese. Reading
and writing were rated lower than grammar. Higher level students
rated lower than lower level, and older students rated slightly
higher than younger ones. Between Tables 3 and 4 we see parallel
results in the area of ethnic background, course type and level, with
the exception of Arabic-speaking students and reading courses.
However, in most categories, the regression coefficients are larger
in the learning experience subgroup than in the teacher subgroup.
This indicates that students’ negative views were most strongly
expressed within the learning experience subgroup.

Multiple Regression Analysis of Learning Experience Subgroup in the Intensive Program

Table 5 shows the results of the regression analysis on the teacher

subgroup in the academic program. The only significant variable in
this category was level, the indication being that as students
progressed in level in this program, they rated higher on average.
The sample size here was rather low (298) which may account for
the small number of variables having significance in this regression


Multiple Regression Analysis of Teacher Subgroup in the Academic Program

Table 6 gives results for the learning experience subgroup in the

academic program. The strongest result in this regression was the
negative effect of the listening and speaking courses in this pro-
gram. Students rated these courses an average of 1/3 of a point
lower than the reading and writing courses.
Four additional multiple regression analyses were done with the
attitude variable included. Table 7 gives the regression coefficients
for the attitude variable in both programs. From these large
coefficients we see that attitude was a very strong predictor of
evaluation scores, with as much as 1/2 a point difference in
evaluation score per one point of attitude score. (Attitude is also
measured on a scale of 0–4, 4 being the highest). However, it is not
clear that attitude is truly an independent variable since a student’s
attitude about a course is influenced by how well he or she likes the
teacher. As a variable, it cannot be separated entirely from the
evaluation ratings themselves.

It is evident that systematic bias did occur in the evaluation
process, due most significantly to level, course type, ethnic


background, and attitude. Based on this finding, we make several
comments and recommendations.
Considering the significant variables individually, it is probable
that the regression coefficients for level and course type are specific
to our particular ESL programs. For example, we had a conference-
centered writing track in the intensive program at the time of the
study which may have influenced our students’ ratings of the
writing courses. Also, the result that level had a positive effect in
one program and a negative effect in the other may reflect the fact
that course content varies with level in both of our programs.
Despite our program particulars, it is fair to generalize this result;
that is, course content and level are likely to influence ratings in any
ESL program depending on the particular course configuration of
that program. Thus it would be necessary to evaluate a teacher in a
variety of course types and levels in order to counteract the effects
of these variables.
The effect of bias due to ethnic background depends on how
ethnic groups are distributed among classes in an ESL program.


Multiple Regression Analysis Including Attitude

Teachers working with more homogeneous classes may be differ-

entially affected by ethnic bias in comparison to teachers working
with more heterogeneous groups. However, to the extent that the
same program-wide ethnic mix is maintained in each class, the
effect of this variable will be neutralized.
Attitude, the strongest predictor of evaluation ratings, should
certainly be adjusted for, particularly in required ESL programs.
However, as mentioned before, since the attitude data were
collected near the end of the quarter along with the rest of the
evaluation data, this variable may not be truly independent.
The discovery of subgroups within the evaluation data suggests
that students do distinguish the questions which directly assess their
teacher from those which reflect the learning experience as a whole,
including effects of classmates, program-wide constraints, and
other circumstances beyond the teacher’s control. Moreover, a
comparison of the regression analyses between the two subgroups
of questions shows that ratings in the learning experience subgroup
are parallel to, yet lower than, those in the teacher subgroup. For
purposes of personnel decisions, it would be important to average
these subgroups independently, and weigh them differently.
Obviously, circumstances outside of the teachers’ control should not
be averaged into an assessment of teaching ability. However,
tabulations from the learning experience subgroup pertaining to
program-level parameters, such as tests and materials, may be
useful in providing feedback on an ESL program as a whole.


In the area of instrument development, we urge further
investigation. Although we validated the content of our question-
naire through interviews with ESL students and teachers, it has yet
to be statistically tested for validity and reliability. Such testing
would require that the form be administered in several different
ESL programs over an extended period of time and that the data be
compared across programs.
Our finding of bias should be investigated in a more extended
study, especially given a larger sample size and a greater variety of
countries represented. An ideal study could include several
different institutions to offset the effects of program particulars.
Such a study could also include a truly independent attitude
variable by measuring attitude at the time the students began a
In conclusion, if evaluation data is used for personnel decisions in
ESL programs, we urge that evaluations be collected in a variety of
courses over a period of time. We also recommend that subgroups
of evaluation questions be averaged separately, and that
adjustments be made for bias. Finally, we caution that student
evaluation data be used only in conjunction with other measures.

Support for this project came from the University of Washington Extension
English as a Second Language Programs. We would like to thank Director Bill
Harshbarger for his help and encouragement during our research. We are grateful
to Andrew F. Siegel, Gerry Gillmore, and David Carrell for their assistance with
statistics and computing. We would also like to thank all the students and teachers
who participated.

Ann K. Wennerstrom is coordinator of the international teaching assistant training
course at the University of Washington Extension English as a Second Language
Programs, where she also teaches academic English. She is currently a doctoral
student in Linguistics.
Patty Heiser, Acting Academic Coordinator at the University of Washington
Extension English as a Second Language Programs, teaches in both intensive and
academic English programs. Her work in developing the programs’ Performance
Appraisal System led to an interest in bias in student evaluations.


Aleamoni, L. M. (1981). Student ratings of instruction. In J. Millman (Ed.),
Handbook of teacher evaluation (pp. 110-145). Beverly Hills, CA: Sage.
Cashin, W. E. (1985, March-April). Student ratings: The need for
comparative data. Paper presented at the 69th Annual Meeting of the
American Educational Research Association, Chicago, IL.
Centra, J. A. (1972). Two studies on the utility of student ratings for
instructional improvement: 1. The effectiveness of student feedback in
modifying college instruction. 11. Self-ratings of college teachers: A
comparison with student ratings. (SIR Rep. No. 2). Princeton, NJ:
Educational Testing Service. (ETS Archives Microfiche No. 1147)
Centra, J. A. (1979). Determining faculty effectiveness. San Francisco:
Committee on the Evaluation and Improvement of Teaching and the
Instructional and Faculty Development Board of the University of
Washington. (1982). Evaluating teaching: Purposes, methods, and
policies. Seattle, University of Washington, Center for Instructional
Development and Research, and the Educational Assessment Center.
Frances, S., & Gruber, M. B. (1981, August). Student evaluations of
psychology instructors. Paper presented at the 89th Annual Convention
of the American Psychological Association, Los Angeles, CA.
Frey, P. W., Leonard, D. W., & Beatty, W. W. (1975). Student ratings of
instruction: Validation research. American Educational Research
Journal, 12, 327-336.
Gillmore, G. M., Kane, M. T., & Naccarato, R. W. (1978). The generaliz-
ability of student ratings of instruction: Estimates of teacher and course
components. Journal of Educational Measurement, 15, 1-13.
Hearn, J. C. (1985, March-April). Determinants of college students’ overall
evaluations of their academic programs. Paper presented at the 69th
Annual Meeting of the American Educational Research Association,
Chicago, IL.
Marsh, H. W. (1980). The influence of student, course, and instructor
characteristics in evaluations of university teaching. American
Educational Research Journal, 17, 219-237.
Marsh, H. W. (1981). The use of path analysis to estimate teacher and
course effects in student ratings of instructional effectiveness. Applied
Psychological Measurement, 6, 47-60.
Marsh, H. W. (1982). SEEQ: A reliable, valid, and useful instrument for
collecting students’ evaluations of university teaching. British Journal of
Educational Psychology, 52, 77-95.
Marsh, H. W. (1984). Students’ evaluations of university teaching:
Dimensionality, reliability, validity, potential biases, and utility. Journal
of Educational Psychology, 76, 707-754.
Marsh, H. W. (1987). Students’ evaluations of university teaching: Research
findings, methodological issues, and directions for future research.
International Journal of Educational Research, 11, 253-388.
McKeachie, W. J. (1979). Student ratings of faculty: A reprise. Academe,
62, 384-397.


Neumann, Y., & Neumann, L. (1981). Determinants of students’ satisfac-
tion with course work: An international comparison between two uni-
versities. Research in Higher Education, 14, 321-333.
Norussis, M. (1988). SPSS x Advanced Statistics Guide. Chicago: SPSS Inc.
Pedhazur, E. J. (1982). Multiple regression in behavioral research (2nd
ed.). New York: Holt, Rinehart and Winston.
Pennington, M. C., & Young, A. L. (1989). Approaches to faculty
evaluation for ESL. TESOL Quarterly, 23 (4), 619-646.
Tollefson, N., & Wigington, H. (1986). Teacher-generated and stu-
dent-generated variability in teacher effectiveness ratings. Instructional
Science, 15, 109-120.
Warrington, W. G. (1973). Student evaluation of instruction at Michigan
State University. In A. L. Sockloff (Ed.), Proceedings: The First
Invitational Conference on Faculty Effectiveness as Evaluated by
Students (pp. 164-182). Philadelphia: Temple University, Measurement
and Research Center.


TESOL QUARTERLY, Vol. 26, No. 2, Summer 1992

Beyond Comprehension Exercises in

the ESL Academic Reading Class
San Francisco State University

Recent research in L1 reading and education has focused on

cognitive and metacognitive strategies that can increase students’
comprehension and learning of academic subject matter from
written text. However, the insights from this work have had
limited impact on academic purposes ESL curricula. ESL
developmental reading courses continue to rely on such standard
activities as comprehension and skill-building exercises. Study
strategies may receive little attention or be taught generically. This
article urges English for academic purposes (EAP) programs to
devise more holistic, task- and text-specific, strategy-oriented
approaches for reading instruction. Recommendations are given
for selecting reading materials, defining criterion tasks, and
providing strategy instruction to prepare intermediate- and
advanced-level ESL students to meet the demands of reading
assignments in secondary and postsecondary content classes.

A central goal of academic purposes ESL programs is to help

students develop reading and thinking strategies needed to read
academic texts in their content classes in order to learn new subject
matter. “Study” reading, reading for in-depth comprehension and
learning, is a special type of reading, demanding a different type of
processing (in terms of focusing of attention, information encoding
and retrieval) than reading for enjoyment or reading for general
information. “Studying is associated with the requirement to
perform identifiable cognitive and/or procedural tasks . . . [to
meet] the criteria on tasks such as taking a test, writing a paper,
giving a speech, and conducting an experiment” (Anderson &
Armbruster, 1984b, p. 657).
Currently, lessons in many English for academic purposes (EAP)
reading classes consist of a prereading phase that activates and
builds background knowledge, guided reading of a text, and
postreading exercises: comprehension checks, discussion and
writing activities, exercises to develop reading skills (e.g.,

determining main ideas, finding supporting details or arguments,
and guessing unfamiliar vocabulary by using context clues), and
exercises to expand linguistic competence (e.g., word study and
sentence analysis exercises). In this skill-building approach, it is
assumed that practice of basic skills will enable students to handle
actual content-area reading assignments.
For students beyond the beginning level of language proficiency,
however, we must ask whether this discrete-skills approach is the
best we can offer to equip them to meet the demands of real
content classes. The emphasis in this traditional instruction is on
facilitating short-term reading comprehension. Study skills often are
not taught, or are taught generically, without reference to specific
academic assignments (criterion tasks), and different academic
disciplines. In academic content classes, students must not only
comprehend texts, but over the long term, critically react to the
content (e.g., in class discussion some time after reading an
assignment), recall main points and details when tested (perhaps
several weeks after initial reading), and synthesize information
from reading with other related information, such as from lectures,
discussion, and independent reading. The concentration, memory,
critical thinking, and study planning skills needed to learn and
demonstrate learning in this academic process are demanding for
native English speakers, and can be very discouraging for students
being educated in a second language and in an unfamiliar
educational system. In addition, in content classes, students face
lengthy, conceptually dense prose in textbook and supplementary
reading assignments. Typically, ESL students experience a large
leap in moving from the relatively short, varied readings of ESL
classes to the complex discourse they must manage in their content
Indeed, difficulty reading and studying content-area texts is high
on the list of problems cited in surveys of ESL students who have
exited from intensive, preuniversity ESL programs (Christison &
Krahnke, 1986; Ostler, 1980) as well as from college developmental
reading and writing programs (Smoke, 1988). ESL students
experience great difficulties when they make the transition to the
English-medium academic mainstream, whether it be at the
elementary or secondary level (Collier, 1987) or at the college level
(Benesch, 1988; Snow& Brinton, 1988a, 1988b).
EAP reading instruction should and can assist ESL students to
make the transition from “learning to read” to “reading to learn,”
but a fundamentally different approach is needed. Reading materi-
als and tasks should resemble materials and tasks students face in


academic content classes, thus encouraging students to build reper-
toires of task- and text-appropriate discourse-processing strategies.
One response is for EAP programs to shift from a skills-based
approach to a content-based approach (sheltered content courses,
theme-based language courses, and adjunct courses) (Brinton,
Snow, & Wesche, 1989; Crandall, 1987; Mohan, 1986). Content-
based curricula have led to the use of more conceptually
challenging texts and have given ESL students practice with
extended, integrated, listening, speaking, reading, and writing tasks
relevant to academic learning. Yet a challenge remains—the
challenge of how such curricula can incorporate systematic training
in strategies for learning academic content, given the nature of
learning content in an L2. CALLA (the cognitive academic language
learning approach: Chamot & O’Malley, 1987; O’Malley & Chamot,
1990) integrates training in language learning strategies with content
and language instruction, but this is not the same as developing
students’ content learning strategies.
What might be some of the strategies that both L1 and L2
students can learn and practice for improved comprehension and
retention when reading academic text? Recent research on reading-
comprehension and study strategies has much to suggest.
This article begins with a selective review of some important
conclusions from this research. Recommendations are then given
for planning EAP reading materials, tasks, and instruction at the
secondary and postsecondary levels of education in order to
prompt ESL students to develop repertoires of task- and text-
specific strategies for optimal learning from text in their content


Comprehension Strategies of Successful Readers
Skilled readers comprehend text by actively constructing mean-
ing, integrating information from the text with relevant information
from their background knowledge (Rumelhart, 1980). Conceptual
knowledge (content schemata), text-structure knowledge (formal
schemata), and knowledge about text-processing strategies are the
foundation for successful construction of meaning.
Studies have shown that L1 readers who have well-developed
conceptual knowledge on the topic of a reading selection will
understand and remember its information better than readers who
do not (Rumelhart, 1980; see review articles by Garner, 1987, and


Tierney & Cunningham, 1984). Studies in L2 reading also confirm
the facilitative effect of relevant content schemata (see reviews of
studies by Barnitz, 1986; Carrell, 1987; Carrell & Eisterhold, 1983;
and Swaffar, 1988). In academic content reading, the difficulty is
that so much of the content of a text may be new and unfamiliar. As
they begin academic reading assignments, such as a new textbook
chapter, students need strategies for activating relevant knowledge
they already have (e.g., from personal experience or from a class
lecture, handout, or worksheet) as well as for constructing a general
sense of the ideas in the text.
Text-structure knowledge helps a reader to see relations between
ideas, including hierarchical relationships between main ideas and
Expository texts . . . take predictable forms, such as the comparison and
contrast mode. . . . Authors flag important statements by such devices
as headings, subsections, topic sentences, summaries, redundancies.
. . . Expert learners know about such devices and use them as clues to
help them concentrate on essential information. (Brown, Armbruster, &
Baker, 1986, p. 54)
In both L1 and ESL studies, students who have been taught how to
identify text structure and use this knowledge to guide their reading
process have exhibited better comprehension and recall of
information than readers lacking such knowledge (e.g., for high
school and college students) (Meyer, Brandt, & Bluth, 1980; Slater,
Graves, & Piché, 1985; Slater, Graves, Scott, & Redd-Boyd, 1988;
review articles by R. Horowitz, 1985a, 1985b; and Singer & Donlan,
1989; ESL studies are reviewed by Barnitz, 1986; Carrell, 1984,1985,
1987; and Swaffar, 1988). Students who are reading expository text
need to work actively at finding and using organizational signals
and at constructing a mental (and possibly written) outline or map
of the network of ideas in the text to enhance understanding.
The third type of knowledge affecting readers’ abilities to com-
prehend text is their repertoires of cognitive strategies for process-
ing text as well as metacognitive strategies (the self-regulation of
cognition) to monitor this processing. Current interactive views of
reading emphasize “an active learner who directs cognitive
resources to complete the task” (Garner, 1987, p. 28). The research
(reviewed by Baker& Brown, 1984; Brown, Armbruster, & Baker,
1986; and Flood & Lapp, 1990) shows that successful readers are
more aware of purposes for reading and adjust their reading
process accordingly, focus attention on major ideas rather than
minor details, engage in self-questioning to determine if reading
goals are being met, notice comprehension failures, and take action


to remedy such failures, using “fix-up” strategies (Alessi, Anderson,
& Goetz, 1979) such as these: “store the problem in memory as a
pending question in the hope that clarification will be forthcoming;
reread the text; look ahead in the text; or consult another source”
(Brown, Armbruster, & Baker, 1986, p. 61). It is proposed that poor
readers can be trained in such active processing strategies in order
to improve their reading comprehension (Brown, Campione, &
Day, 1981).
Research on the reading processes of second language readers has
yielded insights on L2 readers’ cognitive and metacognitive
strategies (e.g., Block, 1986; Hosenfeld, 1977) as well as the
facilitative effects of training in metacognitive strategies (Carrell,
1989; Carrell, Pharis, & Liberto, 1989). Casanave (1988) argues
persuasively for the value of training ESL students to monitor their
reading comprehension.

Successful Study Strategies

Effective study involves several kinds of metacognitive knowl-
edge: knowledge of the criterion task (such as a multiple-choice
test, essay exam, speech, or research paper) and what needs to be
studied (task awareness); knowledge of how best to process the text
for learning, including what to focus attention on, how to
subsequently encode the information attended to, and how to
retrieve the information required by the criterion task (strategy
awareness): and self -knowledge about whether and to what extent
one has learned the material (performance awareness) (Anderson &
Armbruster, 1984b; Wade& Reynolds, 1989).
Numerous studies have examined the effectiveness of various
studying techniques such as underlining, note-taking, summarizing,
student questioning, outlining, elaboration, and diagrammatic
techniques. (See reviews by Anderson, 1980; Anderson &
Armbruster, 1982, 1984b; Caverly & Orlando, 1991; Flood & Lapp,
1990, and McAndrew, 1983.) Results have been mixed. Many of the
early studies did not demonstrate that common study techniques
were beneficial. This result might be due to the “blind training”
(Brown, Campione, & Day, 1981), in which criterion tasks and the
purpose of the study strategy were not explained to experimental
subjects who, therefore, could not regulate their strategy use.
Instructors need to provide “cognitive training with awareness,”
informing students what the strategy is and why, when, where, and
how they might use it (Armbruster & Brown, 1984).
There is very little published research on strategies actually uti-
lized by ESL students when studying subject matter in mainstream


academic content courses. An exception is Adamson’s (1990) case
study research, which reveals a variety of both effective and inef-
fective strategies used by 4 high school and 10 college ESL students.
More process-centered research of this sort is needed.
In accordance with Armbruster and Brown’s (1984) suggestions,
EAP classes should build ESL students’ metacognitive knowledge in
four areas: (a) criterion tasks; (b) basic structures of academic text
(including how to use devices such as headings, introductions, and
signal words, and how to deal with “inconsiderate” text that lacks
unity and cohesion [Armbruster, 1984]); (c) personal strengths and
weaknesses (in background knowledge, interest, and motivation)
that affect reading comprehension and learning, and how to
compensate for weaknesses when making study plans; and (d) a
range of task-specific strategies for learning from text, including
why, when, where and how to use them and how to vary strategies
to suit specific tasks. Reading-to-learn strategies need to be taught
and practiced repeatedly over a long period of time (Duffy &
Roehler, 1989).
For immigrant and international students new to the U.S.
educational system, an important, initial metacognitive topic is a
general orientation to the educational system, including information
about what happens in content classrooms, typical assignments,
teacher expectations, teacher-student relationships, and so forth.
Literacy and academic assumptions vary considerably across
societies; students from many societies may not have previously had
access to large amounts of printed materials and may not be used to
an academic learning process that emphasizes extensive reading,
synthesis of ideas from multiple sources, and critical reaction to
reading (Grabe, 1986).
In light of what we now know regarding comprehension and
study strategies, the remainder of this article presents recommenda-
tions for EAP programs concerning reading materials, criterion
tasks, and the teaching of reading-to-learn strategies.


In all classes for which textbooks and other reading materials pro-
vide a basis for learning new information, instructors should select
their materials with great care. During the text selection process,
they can make use of guidelines for evaluating a textbook’s compre-
hensibility (e.g., Clewell & Clifton, 1983; “considerateness” criteria
discussed in detail by Anderson & Armbruster, 1984a, 1986; and
Armbruster, 1984; “friendliness” criteria discussed by Singer &
Donlan, 1989, especially the “Friendly Text Inventory,” pp. 145-


Below, suggestions are given for selecting and sequencing
reading materials that lend themselves to practicing reading-to-
learn strategies, for use in pull-out academic purposes ESL classes at
the secondary and college levels.
Select independent whole texts {complete articles, entire chapters)
rather than text excerpts whenever possible. Text excerpts may
deprive readers of important context clues; also, students need to
build schemata about the format, register, and other features
associated with different types of reading materials. For students to
become familiar with textbook chapter format (including chapter
organization, headings and subheadings, typographical and
instructional aids), they need to work with entire textbook chapters,
not isolated pieces. One way to familiarize students with textbook
structure and study aids is to use a content textbook in the reading
class, for example, a study skills textbook.
In content classes, students read in depth on related topics in specific
subject areas. In the ESL academic reading class, extended reading on
a single topic, called naraow reading (Krashen, 1981) or reading-in-
depth (Dubin, 1986), more closely simulates the academic situation
than short and varied selections. It also allows reading comprehen-
sion to increase as students accumulate background knowledge
about a topic. Related readings can be sequenced so that students
first read high-interest readings written in an elaborative,
“expansive” style and then move into more information-dense,
textbook-style readings (Dubin, 1986). It should be noted that
reading-in-depth is not the same as reading a series of loosely linked
readings in the same broad topic area—an approach found in many
thematic ESL reading textbooks now in print.
An advantage of whole readings is that new concepts and
vocabulary are recycled and reinforced. “Acquisition of both
structure and vocabulary comes from many exposures in a
comprehensible context” (Krashen, 1981, p. 23). The need for
multiple encounters with new words is one of the properties of
“powerful vocabulary instruction” supported by research, along
with the need for learners to integrate new words into their existing
knowledge (Nagy, 1988). Given lengthy readings containing many
unfamiliar vocabulary words, ESL students can be guided to select
specific recurring words for study, especially topically related key
terms which can be interrelated in a semantic network tied to
students’ existing knowledge.
Reading selections should present substantial new information, on
topics appropriate to students’ ages, educational levels, and interests.


In beginning reading classes, it is appropriate for ESL students to
build decoding skills and language competence while reading
selections on familiar content. But to prepare for academic content
reading, they need to read challenging selections that teach new
information. Within the EAP curriculum, it makes sense to start with
readings on topics of greater familiarity and personal concern (e.g.,
cross-cultural adjustment) and then to move to readings on more
remote and abstract topics.
At the elementary and secondary levels, Chamot and O’Malley
(1987) suggest that an ESL class serving as a bridge to the academic
mainstream should use a content-based curriculum that incorpo-
rates the major topics of the school’s mainstream curriculum. In
order to design such curricula, ESL instructors need to communi-
cate with their colleagues in all the subject areas and become very
familiar with their materials and approaches.
Similarly, for university English for science and technology (EST)
contexts, Hudson (1991 ) describes a successful, content-based
reading curriculum that uses authentic reading material from science
and engineering courses students have already taken or are taking.
A logical theme for readings at the beginning of the term is
orientation to the assumptions and demands of high school/college.
Students who have been previously educated in a different culture
need to gain an explicit understanding of the academic learning
process in their new school system. Readings on such topics as
cross-cultural differences in education (e. g., Levine & Adelman,
1982; Pugh & Fenelon, 1988), high school versus college (for
entering college students), principles of learning and memory, and
time management (as covered in introductory portions of certain
study skills texts such as McWhorter, 1988, 1989; Nist & Diehl,
1990a; and Pauk, 1989) can begin to build students’ metacognitive
understanding of what it will take to be a successful learner in their
new educational environment. School documents (e.g., college
catalogs, class schedules, and information pamphlets) are also
suitable materials both for the information to be learned and as
material with which to practice reading strategies,
Materials should lend themselves to the particular criterion tasks that
students need to develop strategies to handle. For example, if one
objective is to build study and review strategies for satisfactory
performance on multiple-choice tests, the reading(s) must be dense
enough in informational content to construct a sufficient number of
such questions. If the criterion task is writing an essay stating and
supporting one’s position on an issue, readings should provide
sufficient background material to stimulate critical thinking.


In college developmental reading courses, a needs assessment
should inform materials selection. By communicating with their
colleagues across academic disciplines as well as surveying students,
ESL curriculum developers can find out about the reading-related
academic tasks that students need to handle in content courses at
various levels as well as about specific difficulties facing ESL
students (e.g., Carson, 1990; Fox, 1986).
Chose texts that exhibit discourse patterns and devices that students
need to recognize. Readings should provide repeated exposure to
organizational patterns commonly found in academic expository
prose (e.g., see Cheek & Cheek, 1983; Roe, Stoodt, & Burns, 1987),
so that with guidance students will build formal schemata and strat-
egies for using text structure knowledge as a tool for comprehend-
ing and remembering information. Instructors also need to under-
stand differences in text characteristics emphasized in different aca-
demic disciplines (e.g., see Hayes & Peters, 1989; Santa, Havens, &
Harrison, 1989; Singer & Donlan, 1989).
In efforts to find easily accessible text, narrative, and descriptive
texts are sometimes favored over expository texts. However, ESL
students who can easily follow narrative and descriptive prose have
greater difficulty perceiving the conceptual relationships in
expository prose, and they need training and practice analyzing the
superordinate, subordinate, and coordinate relationships between
ideas (Blanton, 1984). Because they are easier for the instructor to
find, magazine and newspaper articles are often chosen as opposed
to textbook material. However, if students need to build textbook
reading and study skills, they need to work with textbook-style
readings. Certainly an interesting, easily accessible periodical article
can be a good way to introduce a reading unit, but students then
need to move on to the more information-dense, organizationally
complex format of content-area texts. Newspaper articles have a
format that is quite different from that of academic text
(Fredrickson & Wedel, 1984).
ESL curriculum developers should become familiar with a range
of general-interest (nonspecialist) magazines and their characteris-
tics, because some utilize the expository text features central to
academic text (e. g., section headings, informative introductions,
explicit organizational signals). High quality informational trade
books are a possible source of in-depth, up-to-date reading material
(see Schallert & Roser, 1989); many are written in textbook style.

Content-area instructors guide and assess learning by the kinds of
tasks that they assign to students (Doyle, 1983). Academic tasks may


be oral (e.g., participation in class discussion, oral presentations,
panel discussions, role plays), written (e.g., quizzes and tests of
various types, problem sets, essays, research papers), and activity-
based (e.g., observations, labs) and are evaluated by certain criteria.
Most often, grades are given to written products (e.g., quizzes,
exams, lab reports, papers). Thus, research on writing assignments
assigned in content classes (e.g., Applebee, Auten, & Lehr, 1981;
Bridgeman & Carlson, 1984; D. M. Horowitz, 1986a, 1986b; Rose,
1983) reveals the kinds of understanding, thinking, and learning
(e.g., recognition, recall, synthesis, evaluation) that instructors hope
to see emerge from class instruction and students’ independent
reading. Close observation of academic class sessions and students
at work, interviews with students, and analyses of class-related texts
and handouts are a valuable source of information about academic
tasks and the reading and studying strategies utilized by students as
they work to fulfill these tasks (e.g., Adamson, 1990; Carson, 1990;
Langer, 1986).
For all academic tasks, especially tests, specific knowledge of the
task and evaluation criteria are crucial to effective study and an
optimal final result. “The more specific students’ knowledge of the
criterion task, the more able and likely they are to focus attention on
relevant material and engage in processing activities appropriate to
performing that task” (Anderson & Armbruster, 1984b, p. 664).
The concept of task-appropriate encoding suggests that certain types of
study techniques are best suited to certain types of tests. This . . . helps
to explain why some students who put forth considerable effort to read,
underline, or take notes still fail the test. . . . Students who write
summaries . . . are likely to be successful on an essay test that requires
them to generate main ideas, but they may be less successful on an
objective test which requires them to recall specified detailed facts.
(Schmidt, Barry, Maxworthy, & Huebsch, 1989, p. 431)
Using needs assessment information, ESL curriculum developers
can decide on the specific types of criterion tasks on which they will
focus instruction. They can encourage students to set themselves
learning goals in line with these task demands. Below are
recommendations for EAP reading classes concerning criterion
Reading assignments in EAP classes should be guided by criterion
tasks like the tasks students are assigned in content classes, rather
than by a focus on reading skills. The instructor should clearly
announce the criterion task before students start the reading
assignment. Tasks should guide students to focus attention and


encode information in task-appropriate ways. Students’ thinking
process should be centered on the general meaning of the text and
strategies for meeting task demands, rather than on reading-skills
concepts such as “main idea” or “contextual clues.”
A strategic-content approach . . . stresses the importance of viewing
reading as a “whole” act rather than as a series of isolated skills. That is,
just because students can select a main idea or make an inference from
a brief passage does not necessarily mean that they can extract the major
concepts from a complete chapter or critically analyze a theory. One of
the major reasons that reading labs or classes are often unsuccessful in
preparing students for content courses is that they teach students how to
break reading into small, isolated fragments but do not teach them how
to put it back together again. College students are not going to be
required anywhere in their postsecondary education to answer test
questions such as “The main idea of this passage is. . . .” (Nist & Diehl,
1990b, p. 3)
Tasks in a content-based, criterion-task-driven reading class are
quite different from tasks in a skills-based reading program.
Instructions accompanying reading assignments in a typical skills-
based class might run something like this: “Read the selection,
answer the comprehension questions (and be ready to justify your
answers), and do the vocabulary-in-context exercises.” An example
of instructions more in keeping with real academic criterion tasks
would be assigning a textbook chapter on a topic under study,
giving out study questions, and announcing that at the end of the
chapter, there will be a test consisting of multiple-choice and short-
essay questions. (Before reading, the ESL class could brainstorm
possible reading strategies and long-range study strategies, and the
instructor might model a strategy on a portion of the assigned
chapter. Subsequent classes would explore the chapter content in
depth; the test would follow.)
Students need to learn and practice strategies for task analysis (i.e.,
understanding an instructor’s directions and expectations/
evaluation criteria) and analyzing steps needed to complete the task
successfully. For effective and efficient reading and studying,
students should develop the habits of checking to see that they have
adequate and correct information about the task and of seeking
clarification where needed.
ESL students need to clearly understand the modes in which
content instructors expect them to demonstrate their knowledge
and original thinking. They must have a clear grasp of academic
terminology such as the language that is used to phrase essay
examination prompts (D. M. Horowitz, 1986a) and formal writing


assignments. Rose’s (1983) analysis of college essay and take-home
examination questions and paper topics revealed that formal
writing tasks could be classified into eight basic “superframes” for
arranging ideas: listing, definition, seriation (chronology),
classification, summary, comparison/contrast, analysis, and
academic argument. Along with repeated work on recognizing such
thought patterns in their assigned reading, then, students also need
repeated practice using such organizational patterns in their writing
(Kiniry & Strenski, 1985). Just as reading and writing are tightly
integrated in their academic learning process, there are many
arguments for integrating reading and writing instruction in EAP
Students should develop a habit of using their knowledge of task
demands to set reading and study goals and to monitor their reading
and studying process accordingly. Effective learners are goal
directed and have developed metacognitive knowledge and control
concerning strategies best matched to specific task demands.


Strategy Instruction
“Research in classrooms confirms that the most effective way to
bring about student control of a strategy is through an instructional
sequence in which independent use is preceded by direct
explanation and guided practice” (Anthony & Raphael, 1989,
p. 254). (See Brown, Campione, & Day, 1981; Nist & Kirby, 1986;
Pehrsson & Denner, 1989; Singer & Donlan, 1989.) In the first phase,
direct explanation and modeling, the teacher should explain the
strategy to the students:
(1) a definition of what the strategy is, (2) a reason why they are learning
the strategy (i.e., how it will help them in their content area reading), (3)
the steps in how to use the strategy, (4) the appropriate times when the
skill is useful, and (5) ways to decide how well they used the strategy.
(Anthony & Raphael, 1989, p. 254)
Then the teacher models the strategy by thinking aloud; this is a
way “to render the covert cognitive and metacognitive processes in
overt form” (Garner, 1987, p. 132). A teacher can verbalize the
thought processes of a skilled reader, including strategies that are
frequently lacking among poor comprehenders, such as forming
hypotheses about the text’s meaning before beginning to read it,
organizing information into mental images while reading, making


analogies to one’s own experiences in order to link prior knowledge
to new information in text, monitoring how well one is compre-
hending as one goes along, and “fixing up” lagging comprehension
(Davey, 1983). Instructors can model both “immature” and
“sophisticated” forms of strategies and ask the class to assess their
relative effectiveness (Garner, 1987, p. 132).
During the second phase, guided practice, students take on more
responsibility as they move toward greater control of the strategy.
This stage involves repeated practice, feedback, and possible
reteaching. At this stage, students can share their own thought
processes and written products such as text underlining, text
annotations, and graphic organizers, with each other and with the
teacher. Palincsar and Brown’s (1986) reciprocal teaching illustrates
how control of comprehension and comprehension-monitoring
strategies (self-questioning, summarizing, clarifying, and predict-
ing) can be shifted from teacher to students, as students play the
role of “teacher” (see Casanave, 1988; Irvin, 1990). Teachers should
provide specific feedback on students’ attempts with a new strategy
(see, e.g., Simpson & Nist’s, 1990, checklists).
The last stage of the instructional sequence is independent
application. Students are asked to apply the strategies to whole
texts, on their own. Many study-skills texts (e.g., McWhorter, 1989;
Nist & Diehl, 1990a) have students first practice a strategy together
in class and then apply it to materials in their concurrent content
EAP reading classes can introduce ESL students to a variety of
prereading, during-reading, and postreading strategies to enhance
their reading and learning process as they work toward fulfilling
specific criterion tasks. Current literature describes both strategies
with which content teachers can mediate in students’ reading
processes and strategies that students can learn to apply on their
own. Because one important goal in EAP reading classes is to
improve students’ independent text interaction strategies, the
discussion below will outline examples of student-initiated
strategies that can be taught and practiced in EAP classes.

Before Reading
When the goal of reading is in-depth comprehension and good recall
of information from expository text, previewing (surveying,
overviewing) facilitates the process because it familiarizes a reader
with the basic content and organization of the text and helps to
activate relevant prior knowledge. Students need to learn why and
how to preview texts of various types (e.g., textbook, textbook


chapter, section of a chapter, articles with and without headings).
(They should first become familiar with the parts of a textbook,
e.g., chapters. Adams, 1989, is one source on textbooks; Fredrickson
& Wedel, 1984, is a source on newspaper articles.) For example, to
survey a textbook chapter, the reader examines only certain highly
informative items (a very important strategy for those ESL readers
who treat all parts of a text as equally important): title and subtitle,
introduction or first paragraph, section headings, material
emphasized typographically, summary or last paragraph, and end-
of-chapter material (such as study questions and vocabulary lists).
“The survey step . . . is . . . an excellent way for students to derive
independently the microstructure of the text as well as to build their
background knowledge of the text prior to reading it in its entirety”
(Jakobowitz, 1988, p. 128). Previewing establishes the important
topics and subtopics to be covered; reading will be easier because
the reader knows the direction of the discussion and what concepts
are important and should be especially attended to. Previewing
activates thinking and brings to mind what the reader already
knows about the subject; new information can then be integrated
into the reader’s existing knowledge structure.
The more relevant background knowledge that readers can relate to
a reading assignment, the more prepared they will be for reading.
Some content instructors, aware of techniques to facilitate
comprehension and learning from text, take care to prepare
students for reading assignments and to set up reading and studying
goals, through study aids such as worksheets and study/discussion
questions. Students need to develop a habit of consulting such aids,
relevant lecture notes, and so forth, as they begin a reading
assignment. These aids provide background and are helpful in
setting reading goals appropriate to criterion tasks. ESL reading
classes should guide students in the use of such study aids.

During Reading
Knowledge of the organization of a particular text and of common
textual signals can help a reader identify important information as
well as relationship between ideas in the text. Van Dijk (1979)
provides a list of types of signals which can be used to identify
important ideas:
1) graphical: type size, italics, underlining; 2) syntactical: word order,
topicalization, 3) lexical: words like important, relevant, the subject is,
the conclusion is; 4) semantic: thematic words and sentences,
summarizing or introductory sentences, repetition; 5) schematic/


superstructural: story grammars, narrative schema, expository text
structures. (cited in Winograd & Bridge, 1986, p. 25)
ESL instruction can familiarize students with such useful cues as
well as with signals of important relationships between ideas (e.g.,
see Adams, 1989, on markers which signal purpose, sequence and
enumeration, examples, restatement, detours and resumption, cause
and effect, comparison and opposition).
Self-questioning during reading is a strategy to actively monitor
comprehension. When readers detect a comprehension failure, they
should use a “fix-up strategy.” Research using the think-aloud
technique has shown that skilled readers ask themselves mental
questions before, during, and after reading (e.g., Jakobowitz, 1990).
Content questions (e.g., Now how exactly does this procedure
work?) are a way to check meaning derived from text content;
process questions (e.g., What did I just read? Do I understand this?
What will the author explain next?) are a way to monitor compre-
hension (Davey, 1985; Irvin, 1990).
The Question step of the well-known SQ3R (Survey, Question,
Read, Recite, Review) study technique advises students to turn
chapter subheadings into questions (Robinson, 1946). However, this
approach to constructing questions may or may not result in good
questions; for example,, headings may not reflect main ideas.
Students may miss asking important questions (Anderson &
Armbruster, 1982; Jakobowitz, 1988). Anderson and Armbruster
(1982) argue that criterion task demands should be a more
important determinant of attention focusing and questioning than
the author’s headings, which may or may not reflect ideas an
instructor might emphasize and test.
Self-questioning can help to motivate a reader, “arousing interest
and directing the reader’s attention, especially under adverse
conditions such as reading a lengthy or uninteresting passage”
(Balajthy, 1984, p. 409).
When readers detect a comprehension problem, they need to
have a repertoire of fix-up strategies to turn to. For example, instead
of ignoring the problem, the reader can look back at previously
read text (Alessi, Anderson, & Goetz, 1979), look ahead in the text,
consult another source, or make a note of the problem as one to be
resolved in the future (possibly in later reading or by questioning a
classmate or the teacher).
ESL students need a sense of priorities concerning the numerous
unfamiliar vocabulary words they encounter in content-area
reading. Many unfamiliar words can be ignored and others
interpreted through intelligent use of context clues; words for which


meanings are not clear from context yet are crucial to understanding
central ideas in the text should be looked up in a reliable dictionary
or, for technical terms, a specialized glossary. Within the holistic
approach being advocated in this article, vocabulary strategies such
as word analysis and context clues would not be taught from a
reading skills text, but rather taught in the context of the texts being
read and studied. To construct accurate meaning from academic
texts, ESL students should be neither overly dependent nor overly
lazy about dictionary use. Class time could be spent identifying key
words essential to a particular text. Students should have strategies
for identifying key words (e.g., words in the title, section headings,
topic sentences, words which recur). The instructor can demon-
strate techniques for learning the selected vocabulary, such as
sketching semantic networks to connect the new words and
concepts with words and concepts students already know.
Textbook authors constantly introduce new, discipline-specific
terminology. Such new terms are defined and explained within the
text and often are listed in end-of-chapter vocabulary lists and final
glossaries. ESL students need to recognize the context cues that
signal definitions of new terms (e.g., punctuation signals, appositive
sentence structure). Content-area instructors sometimes test stu-
dents on terminology. EAP classes could do the same. A study
strategy would be to make vocabulary cards with terms on one side
and definitions and examples on the other.
Selective underlining and annotating of a text help a reader to
actively interact with text to monitor comprehension, and to create
a record for future review. What and how one underlines and
annotates a particular text should suit the criterion task. When a
student owns a book or has a copy of a text and is able to write
directly in it, two popular study-reading strategies can be
underlining selected segments of text and writing annotations in the
margins. Underlining and annotating serve two functions—as
encoding mechanisms during reading (helping to focus the reader’s
attention on selected points) and as external storage mechanisms for
later review.
Studies on students’ underlining of text have generally found that
relevant marking of important points increases comprehension, and
extraneous marking impairs it (McAndrew, 1983). This suggests that
students can benefit from training in underlining so that they avoid
underlining too much or underlining insignificant points. Also,
studies have shown that both underlined and nonunderlined
material is better recalled if higher level, superordinate sentences
are underlined. “It may be that the amount and depth of processing


required to mark higher level sentences increases students’ recall”
(McAndrew, 1983, p. 104). Students should proceed section by sec-
tion through a text, first reading, then rereading and underlining, as
they need a sense of the whole to guide their underlining.
To increase the value of underlining as an external storage
mechanism, students need to develop a habit of first analyzing the
task requirements and using this to help them to determine how
much and what type of material to underline (McWhorter, 1988,
p. 206). Teachers can emphasize the importance of task awareness
by, for example, demonstrating the differences between what
might be useful to underline for a multiple-choice test as opposed to
an essay test (Blanchard, 1985, p. 201).
To deal with situations in which instructors assign reading and do
not indicate how students will need to recall or apply the
information, students need to learn ways to obtain clarification
from the instructor as well as ways to setup their own reading goals.
Underlining alone “does not separate main ideas from exam-
ples . . . nor does it provide any opportunity to comment on or
react to material” (McWhorter, 1988, p. 210). Marginal notes can
create a more meaningful written record for later review. In fact,
good annotation may make underlining unnecessary (Nist & Diehl,
1990a). Annotations can be of various types. Summary words and
phrases are especially helpful. For test review purposes, certain
kinds of information should be targeted, such as definitions,
examples, names, dates, events, lists, causes/effects, and likenesses/
differences (Nist & Diehl, 1990a; Simpson & Nist, 1990). Many
study-skills textbooks provide useful lists of abbreviations and
symbols (e.g., “def .,” numbering, arrows). It is best to preview first,
then reread and annotate, one section at a time. Stating the
information in one’s own words and in short phrases makes a reader
process text at a deeper level. Annotating, like underlining, is an
excellent way to monitor comprehension. Readers can mark
confusing passages with question marks and later seek clarification.
One useful strategy a reader can use to analyze the sequence of
subtopics in a text (i.e., to derive the text’s microstructure) is to
mark off and label “notional blocs” that comprise the conceptual
units of the text, blocs that extend across physical paragraph
boundaries (Blanton, 1984). Each group of related paragraphs can
be labeled with a phrase stating key idea, subtopic, or function of
that portion of the text.
For assignments that require students’ reactions to and evalua-
tions of ideas in a text, annotations can be in the form of personal
comments as well as summary words and phrases. A good way to


divide the two types of annotations is to use a system of double-
entry notes: summary notes in the left margin, reaction notes (e.g.,
agreement or disagreement, surprise, personal associations) in the
right (Spack, 1990).

After Reading
After initial reading of a text (immediately afterwards and in the
days preceding a test or assignment due date), students can make
use of a variety of strategies to organize, reduce, and rehearse
important information from their reading to facilitate recall on a test
and to stimulate thinking in order to tackle other assignments such
as papers.
Note-taking and summarizing are useful strategies for organizing
and condensing information to be remembered for a test or ideas to
be applied to an oral presentation or written assignment.
“Theoretically, notetaking has great potential as a study aid, for it
allows the student to record a reworked (perhaps more deeply
processed) version of the text in a form appropriate for the criterion
task (Anderson & Armbruster, 1984b, p. 666). In general, studies
report significant positive effects on recall for students who both
take notes and review them in a timely fashion (see studies
reviewed by Hidi & Klaiman, 1983; McAndrew, 1983; and Smith &
Tompkins, 1988). Students can benefit from formal training in note-
taking with an emphasis on paraphrasing and on connecting ideas to
prior knowledge.
Notes may take many possible forms. Note cards or concept
cards (term, person, concept on one side of an index card, its
definition, description, or related ideas on the other) are useful for
reviewing factual material (Nist & Diehl, 1990a). Note cards are
also useful when collecting material for a multiple-source research
paper, as they are easy to manipulate. An informal outline form for
notes is a good way to show relationships between ideas. Two-
column notes (narrow lefthand column for key concepts, righthand
column for elaboration), also called the Cornell notetaking system,
facilitate organization of ideas and self-testing (see Santa, Havens,
& Harrison, 1989, who highly recommend this system for science
classes, and Pauk, 1989). Another note-taking format is “structured
note-taking,” a blend between graphic organization and an informal
outline (Smith & Tompkins, 1988). If students are aware of alterna-
tive formats for note-taking, they can choose formats suitable to
specific tasks and personal learning styles.
Writing summaries, that is, succinct statements of the main
idea(s) and key supporting points of a text or text segment, is an


especially good strategy to prepare for a test that will include essay
questions, since such questions typically require students “to take a
great deal of information and organize it concisely in new ways”
(Nist & Diehl, 1990a, p. 154). Summarizing is also useful in
preparing to write a research paper or other oral or written report
requiring synthesis of ideas or viewpoints from several sources.
Composing a summary is a complex task, and typically a difficult
one for ESL students. Students can improve their summarizing skills
if given instruction, modeling, repeated practice, and feedback. It is
best to teach summarizing as a step-by-step process in which
marginal annotations and underlining of main ideas and key
supporting points are used as the basis for taking selective and
careful notes (paraphrasing as much as possible, and placing
quotation marks around exact quotations), and then writing a
summary from notes (without referring back to the original text—
thereby helping to avoid plagiarism, in the case of notes taken for a
paper using secondary sources).
Conceptual mapping promotes comprehension, retention, and
retrieval of ideas, equally ideas from expository text containing
many new terms and complex relationships between concepts. A
conceptual map (also called a graphic organizer, semantic map,
cognitive map, semantic organizer, and network) is a visual repre-
sentation of the relationships between concepts in a text:
It is a graphic arrangement showing how the major and minor ideas are
related in a written work. The map consists of nodes which can be
drawn as circles, rectangles, or squares containing key words or phrases,
and connecting links in the form of lines or arrows drawn between the
nodes. . . . A major classroom value of such maps lies in the way that
they holistically conceptualize a content. (Sinatra, Stahl-Gemake, &
Morgan, 1986, pp. 4-5).
When diagramed by an instructor using student input before
students read a text, it is a way to introduce key vocabulary words,
to activate students’ prior knowledge and focus their attention on
relevant schemata, and to motivate them to read the selection; it is
also a way for the instructor to assess students’ prior knowledge on
the topic and to fill in crucial gaps in this knowledge (Heimlich &
Pittelman, 1986). When constructed independently by students
during and after reading a text, a map helps students to work out the
meaning and relationships of concepts in the text and can serve as a
study guide for test preparation. The degree to which students will
find mapping useful will depend on the types of classes they are
taking, since “some types of information are more easily learned by
using-visual or organizational maps than others” and also will


depend on individual learning-style preferences (McWhorter, 1988,
p. 217). Carrell, Pharis, and Liberto (1989) found that ESL students
who were trained to use semantic maps in prereading and
postreading performed significantly better on open-ended short-
answer comprehension questions on test passages than students who
hadn’t received such training; effectiveness of training varied with
students’ learning styles.
Written information is organized into identifiable patterns in
English expository prose, and these patterns can be related to types
of conceptual maps. Thus, maps can take the form of comparison
and contrast charts, time lines, flow charts, and classification
networks, for example. Armbruster and Anderson (1984),
Armbruster, Anderson, and Ostertag (1989), Clarke (1991), Johnson
(1989), Miccinati (1988), Pehrsson and Denner (1989), Peresich,
Meadows, and Sinatra (1990), Piccolo (1987), and Sinatra et al.
(1986), among others, describe forms of maps which correspond to
text structures, and discuss how students can be trained to identify
text structures while/after reading and to draw maps to represent
this underlying organization. Johnson (1989) describes five types of
visual displays: stars (to list facts about a concept—facts are written
on lines radiating out from the central concept), charts (to list or
compare attributes of concepts—in tabular form), chains (to show a
sequence of steps or events—connected with arrows), trees (to
exhibit hierarchies of concepts—e.g., a classification in the form of
a branching tree), and sketches (pictures, to visualize a concept).
She observes that “creative graphing” absorbs both the artists and
problem solvers among students; the power of this strategy is in the
learning gained in the process.
Vaughan (1982, 1984) designed his ConStruct (Concept Struc-
turing) Procedure specifically for use in difficult reading situations,
in which the reader faces a complex expository text, has limited
(perhaps nonexistent) knowledge of the concepts covered in the
text, and will need to recall the concepts on a future test. After or
during each of three readings of the text, readers depict their under-
standing of the text’s important concepts on a “graphic overview,”
which becomes more complete each time. Students trained to use
this strategy outperformed other students on recall tests, in three
studies conducted in lst-year university medical school classes,
10th-grade biology classes, and 8th-grade history classes. This is the
type of powerful strategy that ESL students need when tackling
their difficult reading tasks. ESL teachers, students, and researchers
need to explore, refine, and share versions of such a strategy.
Formulating and implementing a task-specific study plan with active
planning, rehearsed, monitoring, and evaluation of one's content


knowledge and study process lead to optimal performance on tests
and other major assignments such as research papers. An example of
such a set of strategies is PORPE (Simpson, 1986a), for use when the
criterion task is an essay test. Students predict potential essay
questions to guide their study (also recommended by Finley &
Seaton, 1987, and many study skills texts); organize key ideas using
their own words; rehearse the key ideas; practice recalling the key
ideas in self-assigned writing tasks; and evaluate their written
products on such tasks. PLAE (Nist & Diehl, 1990a; Nist & Simpson,
1989; Simpson & Nist, 1984) is a set of strategies for test preparation
in general. Students preplan study sessions by asking and answering
a number of task-specific questions (e. g., when is the test and when
can I study for it, what does the test cover, what kind of test is it,
what is my goal for a grade); list specific rehearsal strategies
appropriate for each section of the test (e.g., concept cards for a
definition section, conceptual maps and summaries for an essay
section) and draft a study plan showing when and how each
strategy would be used; activate the study plan and monitor
understanding and retention of concepts, modifying the plan if
necessary; and evaluate the test outcome after the test is returned,
analyzing mistakes so that strategies can be modified in study
sessions for future tests.
EAP classes can enable students to develop metacognitive control
over their academic assignments through practice of such strategies
in the context of tests given on the content of reading materials.
Students should also try applying these strategies to their concurrent
content classes.
ESL students need to monitor their academic performance and
know about available outside help they can turn to when they need
it. Such sources may be services available at a learning center, or
printed sources they can find on their own. For example, if students
feel lost with a particular textbook, they can try to find easier, more
basic texts on the subject from the school library or public library,
look up difficult terminology in special dictionaries or encyclope-
dias, and consult other library reference materials ( McWhorter,

Support Strategies
Successful academic learning involves not only effective use of
cognitive and metacognitive strategies, but also techniques to deal
with problems due to affective factors and practical constraints,
problems such as lack of an appropriate study environment,


insufficient time to devote to school work due to job or family
commitments, poor time management, problems in coping with
distractions, and culture stress. Weinstein’s (1987) classification of
learning strategies includes “affective strategies” as an important
category. “Examples of affective strategies include using positive
self-talk to reduce performance anxiety, finding a quiet place to
study, setting a time schedule, using rewards, and setting goals.
Using these strategies helps to focus attention and eliminate internal
and external distractions that can adversely affect concentration”
(p. 593). Dansereau (1985) proposes that students need, not only a
repertoire of “primary strategies” (comprehension and memory
strategies for dealing with text), but also “support strategies” to
achieve and maintain an appropriate frame of mind for studying.


Learning-from-text strategies must be practiced regularly over a
sustained period of time before students show signs of self-
regulation and transfer (Duffy & Roehler, 1989; Garner, 1987;
Herber & Nelson-Herber, 1987; Nist & Simpson, 1987; Simpson,
1984, 1986b; Simpson & Nist, 1990). Reading teachers need to
facilitate transfer in a variety of ways, for example, by modeling
strategies frequently (not just teaching them once), demonstrating
variations of a strategy for different types of texts and tasks,
strengthening students’ metacognitive control (such as through
journal entries analyzing their own reading processes and peer
“micro-teaching” (Nist & Simpson, 1987) as well as through
individualized feedback (Simpson & Nist, 1990)), and assigning
students to apply strategies under discussion to reading assignments
in concurrent content classes and to evaluate the effectiveness of
strategy use in these cases. Ideally, reading-to-learn strategies would
be taught and reinforced by both reading teachers and content
teachers across grade levels and content areas (e.g., Peresich et al.’s
(1990) teaching of the cognitive mapping strategy in one school
Reading-to-learn strategies can be systematically taught and
frequently practiced in a content-based approach. There is some
evidence that students taught with a content-based, strategy-
oriented approach show increased metacognitive control over their
reading and study process and later use these strategies in academic
content classes (e.g., Nist & Simpson, 1987; Simpson, 1986b). L1
adjunct study-skills and learning-strategies courses (attached to


content courses) have been successful (e.g., Langer & Neal, 1987;
Weinstein, 1982, 1988).
In summary, recent research and curriculum innovations support
a movement away from the discrete-skills approach that has been
common in academic purposes ESL reading programs. It is time to
rethink reading materials, tasks, and instructional foci. EAP reading
courses need to use whole texts reflective of real academic
discourse, assign tasks like those assigned by content teachers, and
guide students to develop repertoires of cognitive and metacogni-
tive strategies for optimal learning from text.

May Shih is Assistant Professor of English at San Francisco State University, where
she teaches in the ESL and MA TESL programs. She has coordinated and
developed materials for a two-semester, freshman ESL course in reading,
composition, and study skills.

Adams, T. W. (1989). Inside textbooks: What students need to know.
Reading, MA: Addison-Wesley.
Adamson, H. D. (1990). ESL students’ use of academic skills in content
courses. English for Specific Purposes, 9, 67-87.
Alessi, S. M., Anderson, T. H., & Goetz, E. T. (1979). An investigation of
lookbacks during studying. Discourse Processes, 2, 197-212.
Anderson, T. H. (1980). Study strategies and adjunct aids. In R. J. Spiro,
B. C. Bruce, & W. F. Brewer (Eds.), Theoretical issues in reading
comprehension (pp. 483-502). Hillsdale, NJ: Lawrence Erlbaum.
Anderson, T. H., & Armbruster, B. B. (1982). Reader and text-studying
strategies. In W. Otto & S. White (Eds.), Reading expository material
(pp. 219-242). New York: Academic Press.
Anderson, T. H., & Armbruster, B. B. (1984a). Content area textbooks. In
R. C. Anderson, J. Osborn, & R. J. Tierney (Eds.), Learning to read in
American schools: Basal readers and content texts (pp. 193-226).
Hillsdale, NJ: Lawrence Erlbaum.
Anderson, T. H., & Armbruster, B. B. (1984b). Studying. In
P. D. Pearson, R. Barr, M. L. Kamil, & P. Mosenthal (Eds.), Handbook
of reading research (pp. 657-680). New York: Longman.
Anderson, T. H., & Armbruster, B. B. (1986). Readable textbooks, or,
selecting a textbook is not like buying a pair of shoes. In J. Orasanu
(Ed.), Reading comprehension: From research to practice (pp. 151-162).
Hillsdale, NJ: Lawrence Erlbaum.


Anthony, H. M., & Raphael, T. E. (1989). Using questioning strategies to
promote students’ active comprehension of content area material. In
D. Lapp, J. Flood, & N. Farnan (Eds.), Content area reading and
learning: Instructional strategies (pp. 244-257). Englewood Cliffs, NJ:
Prentice Hall.
Applebee, A. N., Auten, A., & Lehr, F. (1981). Writing in the secondary
school: English and the content areas. Urbana, IL: National Council of
Teachers of English.
Armbruster, B. B. (1984). The problem of “inconsiderate text.” In
G. G. Duff y, L. Roehler, & J. Mason (Eds.), Comprehension instruc-
tion, perspectives and suggestions (pp. 202-217). New York: Longman.
Armbruster, B. B., & Anderson, T. H. (1984). Mapping: Representing
informative text diagrammatically. In C. D. Holley & D. F. Dansereau
(Eds.), Spatial learning strategies: Techniques, applications, and related
issues (pp. 189-209). Orlando, FL: Academic Press.
Armbruster, B. B., Anderson, T. H., & Ostertag, J. (1989). Teaching text
structure to improve reading and writing. The Reading Teacher, 43,
Armbruster, B. B., & Brown, A. L. (1984). Learning from reading: The
role of metacognition. In R. C. Anderson, J. Osborn, & R. J. Tierney
(Eds.), Learning to read in American schools: Basal readers and content
texts (pp. 273-281). Hillsdale, NJ: Lawrence Erlbaurn.
Baker, L., & Brown, A. L. (1984). Metacognitive skills and reading. In
P. D. Pearson, R. Barr, M. L. Kamil, & P. Mosenthal (Eds.), Handbook
of reading research (pp. 353-394). New York: Longman.
Balajthy, E. (1984). Using student-constructed questions to encourage
active reading. Journal of Reading, 27, 408-411.
Barnitz, J. G. (1986). Toward understanding the effects of cross-cultural
schemata and discourse structure on second language reading
comprehension. Journal of Reading Behavior, 18, 95-116.
Benesch, S. (1988). Introduction. In S. Benesch (Ed.), Ending remediation:
Linking ESL and content in higher education (pp. 1-5). Washington,
Blanchard, J. S. (1985). What to tell students about underlining . . . and
why. Journal of Reading, 29, 199-203.
Blanton, L. L. (1984). Using a hierarchical model to teach academic
reading to advanced ESL students: How to make a long story short. The
ESP ]ournal, 3, 37-46.
Block, E. (1986). The comprehension strategies of second language
readers. TESOL Quarterly, 20 (3), 463-494.
Bridgeman, B., & Carlson, S. (1984). Survey of academic writing tasks.
Written Communication, 1, 247-280.
Brinton, D. M., Snow, M. A., & Wesche, M. B. (1989). Content-based
second language instruction. New York: Newbury House.
Brown, A. L., Armbruster, B. B., & Baker, L. (1986). The role of
metacognition in reading and studying. In J. Orasanu (Ed.), Reading
comprehension: From research to practice (pp. 49-75). Hillsdale, NJ:
Lawrence Erlbaum.


Brown, A. L., Campione, J. C., & Day, J. D. (1981). Learning to learn: On
training students to learn from text. Educational Researcher, 10, 14-21.
Carrell, P. L. (1984). The effects of rhetorical organization on ESL
readers. TESOL Quarterly, 18 (3), 441-469.
Carrell, P. L. (1985). Facilitating ESL reading by teaching text structure.
TESOL Quarterly, 19 (4), 727-752.
Carrell, P. L. (1987). Content and formal schemata in ESL reading.
TESOL Quarterly, 21 (3), 461-481.
Carrell, P. L. (1989). Metacognitive awareness and second language
reading. Modern Language Journal, 73, 121-134.
Carrell, P. L., & Eisterhold, J. C. (1983). Schema theory and ESL reading
pedagogy. TESOL Quarterly, 17 (4), 553-573.
Carrell, P. L., Pharis, B. G., & Liberto, J. C. (1989). Metacognitive
strategy training for ESL reading. TESOL Quarterly, 23 (4), 647-678.
Carson, J. G. (1990, March). Curriculum development: Discovering the
academic interface for intensive courses. Paper presented at the 24th
Annual TESOL Convention, San Francisco, CA.
Casanave, C. P. (1988). Comprehension monitoring in ESL reading: A
neglected essential. TESOL Quarterly, 22 (2), 283-302.
Caverly, D. C., & Orlando, V. P. (1991). Textbook study strategies. In
R. F. Flippo & D. C. Caverly (Eds.), Teaching reading and study
strategies at the college level (pp. 86-165). Newark, DE: International
Reading Association.
Chamot, A. U., & O’Malley, J. M. (1987). The cognitive academic
language learning approach: A bridge to the mainstream. TESOL
Quarterly, 21 (2), 227-249.
Cheek, Jr., E. H., & Cheek, M. C. (1983). Reading instruction through
content teaching. Columbus, OH: Charles E. Merrill.
Christison, M. A., & Krahnke, K. J. (1986). Student perceptions of
academic study. TESOL Quarterly, 20 (l), 61-81.
Clarke, J. H. (1991). Using visual organizers to focus on thinking. Journal
of Reading, 34, 526-534.
Clewell, S. F., & Clifton, A. M. (1983). Examining your textbook for
comprehensibility. Journal of Reading, 27, 219-224.
Collier, V. P. (1987). Age and rate of acquisition of second language for
academic purposes. TESOL Quarterly, 21 (4), 617-641.
Crandall, J. (Ed.). (1987). ESL through content-area instruction:
Mathematics, science, social studies. Englewood Cliffs, NJ: Prentice
Hall Regents.
Dansereau, D. F. (1985). Learning strategy research. In J. W. Segal,
S. F. Chipman, & R. Glaser (Eds.), Thinking and learning skills: Vol. 1.
Relating instruction to research (pp. 209-239). Hillsdale, NJ: Lawrence
Davey, B. (1983). Think aloud—Modeling the cognitive processes of read-
ing comprehension. Journal of Reading, 27, 44-47.
Davey, B. (1985). Helping readers think beyond print through self-
questioning. Middle School Journal, 17, 26-27.


Doyle, W. (1983). Academic work. Review of Educational Research, 53,
Dubin, F. (1986). Dealing with texts. In F. Dubin, D. E. Eskey, &
W. Grabe (Eds.), Teaching second language reading for academic
purposes (pp. 127-160). Reading, MA: Addison-Wesley.
Duffy, G. G., & Roehler, L. R. (1989). Why strategy instruction is so
difficult and what we need to do about it. In C. B. McCormick,
G. E. Miller, & M. Presdey (Eds.), Cognitive strategy research: From
basic research to educational applications (pp. 133-154). New York:
Finley, C. D., & Seaton, M. N. (1987). Using text patterns and question
prediction to study for tests. Journal of Reading, 31, 124-132.
Flood, J., & Lapp, D. (1990). Reading comprehension instruction for
at-risk students: Research-based practices that can make a difference.
Journal of Reading, 33, 490-496.
Fox, L. (1986). Gateway: A cross-curricular reading and writing book. San
Diego: Harcourt Brace Jovanovich.
Fredrickson, T. L., & Wedel, P. F. (1984). English by newspaper: How to
read and understand an English language newspaper. Rowley, MA:
Newbury House.
Garner, R. (1987). Metacognition and reading comprehension. Norwood,
NJ: Ablex.
Grabe, W. (1986). The transition from theory to practice in teaching
reading. In F. Dubin, D. E. Eskey, & W. Grabe (Eds.), Teaching
second language reading for academic purposes (pp. 25-48). Reading,
MA: Addison-Wesley.
Hayes, B., & Peters, C. (1989). The role of reading instruction in the social
studies classroom. In D. Lapp, J. Flood, & N. Farnan (Eds.), Content
area reading and learning: Instructional strategies (pp. 152-178).
Englewood Cliffs, NJ: Prentice Hall.
Heimlich, J. E., & Pittelman, S. D. (1986). Semantic mapping: Classroom
applications. Newark, DE: International Reading Association.
Herber, H. L., & Nelson-Herber, J. (1987). Developing independent
learners. Journal of Reading, 30, 584-588.
Hidi, S., & Klaiman, R. (1983). Notetaking by experts and novices: An
attempt to identify teachable strategies. Curriculum Inquiry, 13,
Horowitz, D. M. (1986a). Essay examination prompts and the teaching of
academic writing. English for Specific Purposes, 5, 107-120.
Horowitz, D. M. (1986b). What professors actually require: Academic
tasks for the ESL classroom. TESOL Quarterly, 20 (3), 445-462.
Horowitz, R. (1985a). Text patterns: Part I. Journal of Reading, 28,
Horowitz, R. (1985b). Text patterns: Part II. Journal of Reading, 28,
Hosenfeld, C. (1977). A preliminary investigation of the reading strategies
of successful and nonsuccessful second language learners. System, 5,


Hudson, T. (1991). A content comprehension approach to reading English
for Science and Technology. TESOL Quarterly, 21 (1), 77-104.
Irvin, J. L. (1990). Reading and the middle school student: Strategies to
enhance literacy. Boston: Allyn & Bacon.
Jakobowitz, T. (1988). Using theory to modify practice: An illustration
with SQ3R, Journal of Reading, 32, 126-131.
Jakobowitz, T. (1990). AIM: A metacognitive strategy for constructing the
main idea of text. Journal of Reading, 33, 620-624.
Johnson, L. L. (1989). Learning across the curriculum with creative
graphing. Journal of Reading, 32, 509-519.
Kiniry, M., & Strenski, E. (1985). Sequencing expository writing: A
recursive approach. College Composition and Communication, 36,
Krashen, S. D. (1981, December). The case for narrow reading. TESOL
Newsletter, p. 23.
Langer, J. A. (1986). Learning through writing: Study skills in the content
areas. Journal of Reading, 29, 400-406.
Langer, J. A., & Neal, J. C. (1987). Strategies for learning: An adjunct
study skills model. Journal of Reading, 31, 134-139.
Levine, D. R., & Adelman, M. B. (1982). Beyond language: Intercultural
communication for English as a second language. Englewood Cliffs, NJ:
Prentice Hall.
McAndrew, D. A. (1983). Underlining and notetaking: Some suggestions
from research. Journal of Reading, 27, 103-108.
McWhorter. K. T. (1988). Study and thinking skills in college. Glenview,
IL: Scott, Foresman.
McWhorter, K. T. (1989). College reading and study skills (4th ed.).
Glenview, IL: Scott, Foresman.
Meyer, B. J., Brandt, D. M., & Bluth, G. J. (1980). Use of top-level
structure in text: Key for reading comprehension of ninth-grade
students. Reading Research Quarterly, 16, 72-103.
Miccinati, J. L. (1988). Mapping the terrain: Connecting reading with
academic writing. Journal of Reading, 31, 542-552.
Mohan, B. A. (1986). Language and content. Reading, MA: Addi-
Nagy, W. E. (1988). Teaching vocabulary to improve reading comprehen-
sion. Urbana, IL: National Council of Teachers of English.
Nist, S. L., & Diehl, W. (1990a). Developing textbook thinking: Strategies
for success in college (2nd ed.). Lexington, MA: D. C. Heath.
Nist, S. L., & Diehl, W. (1990b). Instructor’s guide: Developing textbook
thinking (2nd ed.). Lexington, MA: D. C. Heath.
Nist, S. L., & Kirby, K. (1986). Teaching comprehension and study
strategies through modeling and thinking aloud. Reading Research and
Instruction, 24, 254-264.
Nist, S. L., & Simpson, M. L. (1987). Facilitating transfer in college
reading programs. Journal of Reading, 30, 620-625.


Nist, S. L., & Simpson, M. L. (1989). PLAE, a validated study strategy.
Journal of Reading, 33, 182-186.
O’Malley, J. M., & Chamot, A. U. (1990). Learning strategies in second
language acquisition. New York: Cambridge University Press.
Ostler, S. E. (1980). Survey of academic needs for advanced ESL. TESOL
Quarterly, 14 (3), 489-502.
Palincsar, A. S., & Brown, A. L. (1984). Reciprocal teaching of compre-
hension-fostering and comprehension-monitoring activities. Cognition
and Instruction, 1, 117-175.
Pauk, W. (1989). How to study in college (4th ed.). Boston: Houghton
Pehrsson, R. S., & Denner, P. R. (1989). Semantic Organizers: A study
strategy for special needs learners. Rockville, MD: Aspen.
Peresich, M. L., Meadows, J. D., & Sinatra, R. (1990). Content area
cognitive mapping for reading and writing proficiency. Journal of
Reading, 33, 424-432.
Piccolo, J. A. (1987). Expository text structure: Teaching and learning
strategies. The Reading Teacher, 40, 838-847.
Pugh, S. L., & Fenelon, J. (1988). Integrating learning, language, and
intercultural skills for international students. Journal of Reading, 31,
Robinson, F. P. (1946). Effective study. New York: Harper & Row.
Roe, B. D., Stoodt, B. D., & Burns, P. C. (1987). Secondary school
reading instruction: The content areas. Boston: Houghton Mifflin.
Rose, M. (1983). Remedial writing courses: A critique and a proposal.
College English, 45, 109-128.
Rumelhart, D. E. (1980). Schemata: The building blocks of cognition. In
R. J. Spiro, B. C. Bruce, & W. F. Brewer (Eds.), Theoretical issues in
reading comprehension (pp. 33-58). Hillsdale, NJ: Lawrence Erlbaum.
Santa, C., Havens, L., & Harrison, S. (1989). Teaching secondary science
through reading, writing, studying, and problem solving. In D. Lapp,
J. Flood, & N. Farnan (Eds.), Content area reading and learning:
Instructional strategies (pp. 137-151). Englewood Cliffs, NJ: Prentice
Schallert, D. L., & Roser, N. L. (1989). The role of reading in content area
instruction. In D. Lapp, J. Flood, & N. Farnan (Eds.), Content area
reading and learning: Instructional strategies (pp. 25-33). Englewood
Cliffs, NJ: Prentice Hall.
Schmidt, C. M., Barry, A., Maxworthy, A. G., & Huebsch, W. R. (1989).
But I read the chapter twice. Journal of Reading, 32, 428-433.
Simpson, M. L. (1984). The status of study strategy instruction:
Implications for classroom teachers. Journal of Reading, 28, 136-142.
Simpson, M. L. (1986a). PORPE: A writing strategy for studying and
learning in the content areas. ]ournal of Reading, 29, 407-414.
Simpson, M. L. (1986b). Teaching university freshmen to employ,
regulate, and transfer study strategies to content areas. Forum for
Reading, 17, 61-71.


Simpson, M. L., & Nist, S. L. (1984). PLAE: A model for planning
successful independent learning. Journal of Reading, 28, 218-223.
Simpson, M. L., & Nist, S. L. (1990). Textbook annotation: An effective
and efficient study strategy for college students. Journal of Reading, 34,
Sinatra, R., Stahl-Gemake, J., & Morgan, N. W. (1986). Using semantic
mapping after reading to organize and write original discourse. Journal
of Reading, 30, 4-13.
Singer, H., & Donlan, D. (1989). Reading and learning from text (2nd ed.).
Hillsdale, NJ: Lawrence Erlbaum.
Slater, W. H., Graves, M. F., & Piché, G. L. (1985). Effects of structural
organizers on ninth-grade students’ comprehension and recall of four
patterns of expository text. Reading Research Quarterly, 20, 189-202.
Slater, W. H., Graves, M. F., Scott, S. B., & Redd-Boyd, T. M. (1988).
Discourse structure and college freshmen’s recall and production of
expository text. Research in the Teaching of Engish, 22, 45-61.
Smith, P. L., & Tompkins, G. E. (1988). Structured notetaking: A new
strategy for content area readers. Journal of Reading, 32, 46-53.
Smoke, T. (1988). Using feedback from ESL students to enhance their
success in college. In S. Benesch (Ed.), Ending remediation: Linking
ESL and content in higher education (pp. 7-19). Washington, DC:
Snow, M. A., & Brinton, D. (1988a). The adjunct model of language
instruction: An ideal EAP framework. In S. Benesch (Ed.), Ending
remediation: Linking ESL and content in higher education (pp. 33-52).
Washington, DC: TESOL.
Snow, M. A., & Brinton, D. (1988b). Content-based language instruction:
Investigating the effectiveness of the adjunct model. TESOL Quarterly,
22 (4), 553-574.
Spack, R. (1990). Guidelines: A cross-cultural reading-writing text. N e w
York: St. Martin’s Press.
Swaffar, J. K. (1988). Readers, texts, and second languages: The
interactive processes. Modern Language Journal, 72, 123-149.
Tierney, R. J., & Cunningham, J. W. (1984). Research on teaching reading
comprehension. In P. D. Pearson, R. Barr, M. L. Kamil, & P. Mosenthal
(Eds.), Handbook of reading research (pp. 609-655). New York:
van Dijk, T. A. (1979). Relevance assignment in discourse comprehension.
Discourse Processes, 2, 113-126.
Vaughan, J. L. (1982). Use the ConStruct Procedure to foster active
reading and learning. Journal of Reading, 25, 412-422.
Vaughan, J. L. (1984). Concept structuring: The technique and empirical
evidence. In C. D. Holley & D. F. Dansereau (Eds.), Spatial learning
strategies: Techniques, applications, and related issues (pp. 127-147).
New York: Academic Press.
Wade, S. E., & Reynolds, R. E. (1989). Developing metacognitive
awareness. Journal of Reading, 33, 6-14.


Weinstein, C. E. (1982). Learning strategies: The metacurriculum in
community college teaching. ]ournal of Developmental and Remedial
Education, 5 (2), 6-7, 10.
Weinstein, C. E. (1987). Fostering learning autonomy through the use of
learning strategies. Journal of Reading, 30, 590-595.
Weinstein, C. E. (1988). Assessment and training of student learning
strategies. In R. R. Schmeck (Ed.), Learning strategies and learning
styles (pp. 291-316). New York: Plenum Press.
Winograd, P. N., & Bridge, C. A. (1986). The comprehension of
important information in written prose. In J. F. Baumann (Ed.),
Teaching main idea comprehension (pp. 18-48). Newark, DE:
International Reading Association.


TESOL QUARTERLY, Vol, 26, No. 2, Summer 1992

See How They Read:

Comprehension Monitoring
of L1 and L2 Readers
Baruch College, The City University of New York

This paper illustrates the comprehension-monitoring process used

by first and second language readers of English as they read
expository prose. The think-aloud protocols of 25 college
freshmen were collected. Sixteen (8 L1 and 8 L2 readers) were
classified as proficient, 9 as nonproficient (3 L1 and 6 L2 readers).
The monitoring process is discussed with respect to two specific
problems: one involving a search for a referent, the other, a
vocabulary problem. Three phases and six specific steps are
defined: evaluation phase (problem recognition and problem
source identification), action phase (strategic plan and action/
solution attempt), and checking phase (check and revision). The
responses indicated that monitoring was most thorough with the
referent problem when the problem was explicitly signaled. The
process was somewhat truncated with the vocabulary problem.
Proficient L2 readers performed similarly to proficient L1 readers;
less proficient L2 readers performed similarly to less proficient L1
readers. Although the general trends shown in L1 research were
supported, there were some discrepancies in developmental
trends. Caution is advised in applying the results of L1 research to
L2 readers.

Reading is such a hidden process that it often goes unnoticed in

the language classroom. Still we know much more about reading in
both a first and a second language than we did 15 years ago. We
have ceased debating whether reading is a bottom-up, language-
based process or a top-down, knowledge-based process. Most
people now accept that the two processes interact (Carrell, Devine,
& Eskey, 1988; Grabe, 1991; Rumelhart, 1977; Stanovich, 1980). We
now also appreciate the influence of background knowledge on
both L1 and L2 readers (e.g., Adams & Collins, 1985; Anderson &
Pearson, 1984; Carrell, 1983, 1988; Carrell & Eisterhold, 1983;
Steffensen, Joag-dev, & Anderson, 1979). And, furthermore, we

believe that readers actively control this hidden process, and that
this control directly affects their ability to understand and to learn
from text. (See Baker & Brown, 1984; Garner, 1987; Paris, Wasik, &
Van der Westhuigan, 1988, for reviews of research on metacogni-
This control, commonly called metacognition, involves thinking
about what one is doing while reading. It is an ability that develops
relatively late because it involves the ability to stand back and
observe oneself. It is also an ability often related to effective
learning and to competent performance in any area of problem
solving. Experts plan, predict outcomes, and monitor their
performance more consistently than do novices.
Although “most characterizations of reading include skills and
activities that involve metacognition” (Baker & Brown, 1984,
p. 354), little is known about what happens during the metacogni-
tive process. This paper investigates one aspect of that process,
comprehension monitoring, to see how it operates for first and
second language readers.
To date, most of the research involving comprehension
monitoring has been conducted with native speakers of English, but
there are reasons to believe that comprehension monitoring is of
particular importance for L2 readers. Indeed, Casanave (1988) has
called comprehension monitoring “a neglected essential” in L2
reading research. For one thing, L2 readers can be expected to
encounter more unfamiliar language and cultural references while
reading authentic or unadapted texts than L1 readers would. They,
therefore, may have to “repair” more gaps in their understanding
than L1 readers.
Information from related areas supports the idea that second
language learners may be particularly able to reflect on their
cognitive processes. Metalinguistics awareness seems to play a
special part both in learning to read and in learning a second
language (Bialystok & Ryan, 1985). Hakuta, Ferdman, and Diaz
(1987) report that bilingual profit from sensitivity to metalinguistic
information. Although metalinguistic data is not synonymous with
metacognition, both areas provide evidence for a kind of internal
control that increases the efficiency of some cognitive processes.
Yet, to date, we know little about how detection and regulation
processes operate in L2 readers.
The results of research involving monolingual first language
speakers of English indicate that comprehension monitoring
operates rather automatically, particularly in good readers, and is
not readily observable until some triggering event, some confusion
or failure to comprehend, occurs (Baker & Brown, 1984). A reader


needs to have some plan for repairing the perceived difficulty.
Research has consistently shown that older and more proficient
readers have more control over this monitoring process than
younger and less proficient readers (Baker, 1985). Good readers are
more aware of how they control their reading and more able to
verbalize this awareness (e.g., Forrest-Pressley & Waller, 1984;
Gambrell & Heathington, 1981; Hare, 1981; Smith, 1967; Winograd
& Johnston, 1982). They also appear to be more sensitive to
inconsistencies in the text and to respond to a wider range of
inconsistencies than poorer readers do (August, Flavell, & Clift,
1984; Baker, 1979; Canney & Winograd, 1979; Garner & Kraus, 1982;
Paris & Myers, 1981; Zabrucky & Ratner, 1989). However, even
good readers do not always report or may not recognize
inconsistencies (Baker, 1979; Baker & Anderson, 1982; Winograd &
Johnston, 1982). Baker (1985) speculates that good readers’ failures
to report or even consciously recognize inconsistencies may be the
result of their desire to develop a coherent reading of the printed
Although most L1 research on this topic has concentrated on
establishing the existence and usefulness of comprehension
monitoring, some has also pointed out that readers may differ in the
standards or cue systems used to evaluate whether they have
understood, and that the use of different cue systems is usually
associated with reading proficiency. For example, proficient
readers tend to use meaning-based cues to evaluate whether they
have understood what they read, whereas poorer readers tend to
use or overrely on word-level cues (Baker, 1985; Canney &
Winograd, 1979; Garner & Kraus, 1982), to focus on the decoding
part of reading (Garner & Kraus, 1982; Myers & Paris, 1978), and to
focus on intrasentential rather than intersentential consistency
(Garner, 1981).
It is not clear where L2 readers fit into this process. Although it is
generally accepted that L1 reading ability transfers to L2 reading,
there is considerable debate about how and when it does so. Some
research suggests that reading ability and strategy use is dependent
on language proficiency. The findings of Clarke (1979, 1980),
Cummins (1980), Cziko (1980), and Yorio (1971) are representative
of those researchers who suggest that reading ability is heavily
dependent on L2 language proficiency.
Some researchers believe that L2 readers tend to read more
locally than native speakers (Cohen, Glasman, Rosenbaum-Cohen,
Ferrara, & Fine, 1979) and fail to recognize cohesive ties such as
conjunctions and referents (Carrell, 1988; Chapman, 1979; Cowan,
1976; Devine, 1988b; Mackay, 1979). Other researchers agree with


Alderson & Urquhart (1985), who “find it difficult to draw a clear
distinction between first and foreign language reading” (p. xv).
They believe that L2 readers use the same strategies as L1 readers
(Block, 1986; Goodman, 1973; Hudson, 1982).
There is some tentative evidence suggesting that metacognitive
control, a globally oriented process, distinguishes more and less
skilled L2 readers. For example, Devine (1988a) studied two
readers with different models of the reading process. Her discussion
suggests that meaning-oriented readers can overcome the effects of
limited language proficiency. Carrell’s (1989) study of metacogni-
tive resources suggests that L2 readers with greater L2 proficiency
(readers she defines as ESL) favored a global process, whereas
those with less L2 familiarity (readers she defines as foreign
language readers) used a more localized process. For all readers,
reading locally in their first language was “negatively correlated
with reading performance” (p. 127). In both studies, better L2
comprehenders favored a top-down, meaning-based approach, just
as more proficient L1 readers do.
Casanave (1988) has recommended that “research . . . exam-
ine . . . the monitoring and repair strategies that L2 populations use
and might be taught to draw on as they read” (p. 296). Yet little has
been done to answer that call. We know little about the processes
that L2 readers use to monitor or evaluate their comprehension and
to repair gaps in comprehension, or about the cues to which they
attend in this evaluation and regulation process. We do not know
whether the monitoring process follows the same steps in L1 and L2
readers or whether different types of triggering events produce
different versions of the process in either L1 or L2 readers.
One reason we know so little is that most studies of comprehen-
sion monitoring have focused on whether readers perceive
inconsistencies or problems that have been introduced into the text
by a researcher (e.g., Baker, 1979; Garner& Kraus, 1982; Winograd
& Johnston, 1982). Such studies treat comprehension monitoring as
a “unitary phenomenon” (Baker, 1985) and fail to show what
readers are attending to when they indicate that a problem exists.
Even when researchers control the types of errors or inconsistencies
inserted into a text, they still don’t know the reader’s degree or
depth of recognition when a problem is recognized.
Think-aloud protocols provide a chance to examine the compre-
hension-monitoring process in some depth. This technique, devel-
oped by Newell and Simon (1972), has been used to study the read-
ing process by a growing number of L1 (e.g., Afflerbach, 1986;
Caron, 1989; Olshavsky, 1977; Olson, Duffy, & Mack, 1984) and L2
researchers (Block, 1986; Hosenfeld, 1977; Sarig, 1987). By using


think-aloud protocols to peer into the minds of readers, we can
begin to see whether L1 and L2 readers use similar processes and re-
sources for solving the comprehension difficulties they perceive.
When using think-alouds with L2 readers, certain cautions must
be added. Of special concern is that reports may be incomplete due
to lack of language proficiency or additional processing demands.
In spite of this caveat, much useful information has been collected
when using think-alouds to study the reading of second language
This paper uses think-alouds to explore and compare the compre-
hension-monitoring processes of first and second language readers
of English as they read a passage of expository text. To focus the
discussion, two problematic sentences were identified by surveying
the think-aloud transcripts. Probably not coincidentally, the
sentences identified contain language features long thought to be
problematic for L2 readers. The first involves locating an
appropriate referent for a demonstrative pronoun; the second, that
of defining an unknown word. Thus, by looking at two simple
examples, we will be able to gain some perspective on how L2
readers fit into the pattern established by L1 research.


Think-alouds were collected from 25 first-semester students
attending an urban college. Four groups of readers were defined
based on their scores on a standardized reading test—the
Descriptive Test of Linguistic Skills (Educational Testing Service,
1978) administered to all entering students. Sixteen participants
were considered proficient readers of English; 8 of these were
monolingual native speakers of English (proficient native speakers);
8 were nonnative speakers of English of which 4 were native
speakers of Chinese and 4 of Spanish (proficient ESL readers). All
these proficient nonnative speakers had attended at least 2 and no
more than 5 years of school in the U.S.
After a training session, all participants were asked to “say
everything they understood and everything they were thinking as
they read each sentence.” Red dots were placed after each sentence
to remind participants to respond.
The 9 nonproficient readers had failed to pass the standardized
reading test and were enrolled in reading classes at the college. Six
were nonnative speakers of English of which 3 were native speakers
of Spanish and 3 of Chinese (less proficient ESL readers); 3 were
monolingual native speakers of English (less proficient native


speakers). (See Block, 1986, for a more complete description of
these readers.)
The passage which these participants read was taken from an
introductory psychology textbook (Rubin & McNeil, 1981). (See the
Appendix for a copy of the text.) The article, entitled “Talking to
Babies,” contained 589 words (30 sentences) and had a readability
level of ninth grade according to the Fry readability formula. It
begins by stating that no matter what language is being spoken
(from English to Urdu), the talk people use with babies has similar
characteristics. People tend to use a sing-song intonation, short
sentences, limited vocabulary and straight-forward grammar, and
few references to the past or future. They do this naturally and
because babies prefer sounds in higher pitch ranges. The babies’
reactions help to shape baby talk. The grammatical simplicity and
concreteness of baby talk help to teach the baby how to talk. An
additional function of baby talk is that it conveys affection in ways
that words alone cannot.
Unlike passages traditionally used in comprehension-monitoring
studies (e.g., Baker, 1979; Scardamalia & Bereiter, 1984; Winograd
& Johnson, 1979), “Talking to Babies” contained no “planted”
inconsistencies or errors. A survey of the protocol responses showed
that many readers experienced two types of language-based
problems: finding appropriate pronoun referents and defining
unknown words.
In “Talking to Babies,” referent problems involved indefinite
pronouns for which there were no specific referents in the previous
sentence (e. g., What is the point of these peculiarities?; It marks the
end of the adult’s verbal offering and invites the child to make a
response; And this may be as important a function as language
learning and communication).
Lexical problems were of several types. One word was unknown
to all readers (All the world’s languages, from English to Urdu share
one special kind of speech: baby talk). Several others were
unfamiliar to many readers (For the toddler who has begun to utter
a few words, the rising voice at the end of the sentence serves as a
signal: Your turn) or were used in unfamiliar ways (Babies
themselves help to shape baby talk, through their reactions to adult
utterances). Complex forms of relatively simple words (peculiari-
ties, repetition, melodious, simplicity, and concreteness) also pro-
vided problems for nonnative speakers and less proficient readers.
Two sentences, one containing a referent problem and one a
lexical problem, were selected for in-depth study of monitoring


Responses to one sentence, And this may be as important a
function as language learning and communication, revealed how
the participants evaluated and regulated their comprehension when
faced with an obscure referent. They illustrate the differences in the
ways proficient and less proficient readers, both first and second
language speakers, control the monitoring process.
Solving this problem was particularly difficult since the referent
is not found in the previous sentence. The precise antecedent, the
function of baby talk to convey affection, is explicitly stated at the
beginning of the paragraph. Thus, the reader must recognize the
syntactic complexity of as important a function and synthesize
information to search for the appropriate function.
Overall, the monitoring process seemed to comprise three parts:
evaluation of comprehension, action, and the checking of action.
When readers evaluated comprehension, they recognized that they
were having trouble understanding the sentence either explicitly by
making a statement (problem recognition) or implicitly by taking
some action. Some readers also explicitly identified the precise
source of their confusion or problem (source identification); that is,
they verbalized the need to find the referent of this. Others
identified the problem source implicitly by searching for the
referent without naming the problem. None of the readers, whether
more or less proficient, seemed to mistake the nature of the
problem, although some failed to notice a problem. A few of the
less proficient readers realized that a problem existed, but couldn’t
identify its source. Still none thought that the problem was
something other than a referent search.
The action stage seemed to have two potential parts: strategic
plan statement (or report), in which readers reported the strategy
they had used or would be using; and action/solution, in which
readers attempted to solve that problem. The final stage involved
checking the result of the action, revising the initial hypothesis
where necessary, and providing an appropriate strategy or actual
Control of the various stages of this process seemed to depend
more on reading ability than on whether the reader was a first or
second language reader of English. That is, proficient native readers
(PNs) used the process most completely and explicitly; proficient
ESL readers (PESLs) used the process almost as completely, but not
as explicitly. The less proficient readers (both less proficient native
readers (LNs), and less proficient ESL readers (LESLs) ) were not as
adept at recognizing that a problem existed or in identifying its


source and, for the most part, seemed to lack resources even to
attempt to solve this referent problem. (See Table 1 for a summary
of these responses.)

Steps in Monitoring a Referent Problem

Proficient Native Readers

Most PNs (6 out of 8) both recognized that a problem existed and
identified its source, the unclear referent of this. While only 3 (out
of these 6) explicitly questioned the word this, the other 3 searched
for a referent; thus it can be assumed that they had correctly
identified the problem’s source. Two did not comment on this
sentence at all. Two thirds (4 out of the 6 identifying the problem)
verbalized the strategy or plan they had used or were using to
complete their search. Most often this involved looking back to the
previous sentence:
1. I had to go backup to find out what this is talking about.
2 . . . . because if you take it with this other sentence, I’m just reading
this with the one I read before.
3. I think it’s talking about the preceding sentence.
4. How does this pertain to the last sentence.
All who recognized a problem took some action; however they did
not always identify the specific referent. All used a rather lim-
ited “look back” strategy. They looked at the preceding sentence,


identifying either baby talk or baby talk . . . between adults as the
antecedent. Only one eventually revised this initial hypothesis and
identified the affectionate part of baby talk as the referent.
Two thirds (4 out of 6) checked their solution by substituting their
choice of antecedent for this in the sentence. Two of these
(Examples 5 and 7 below) seemed dissatisfied with their under-
standing but didn’t take any more action. Their checks compared
their understanding of the text with their knowledge of the real
world. Only one (the fourth) compared his interpretation with the
succeeding text as well as with real-world knowledge. He was also
the only one to revise his initial solution.
Examples of these solutions follow. (The stages of the process are
indicated in capital letters within brackets.)
5. [SOLVE] It’s kind of funny. It seems as if speaking this way helps the
person learn to communicate. [CHECK] I’m saying that if you have
two lovers here, I’m sure they know how to communicate so it
couldn’t be that they’re talking about that. [REVISE] It must be the
children, the babies that they’re talking about. [RECHECK] Baby
talk is important in language communication, we established that
6. [SOLVE] I guess baby talk. [CHECK] To figure it out, I put baby
talk in place of this to see if it works. I think so.
7. And it said this may be as important a function as language learning
and communication. I don’t see how. [SOLVE] I guess baby talk
between lovers is just as important as learning the language and
communicating, [CHECK] but I don’t see how.
8. [SOLVE] They’re talking about baby talk between adults and its
important function as language learning, communication. By the time
they’re adults, they already had learning communication down; you
can’t teach them any more. [CHECK] Can’t understand that, ’cause
the next sentence is Children need to learn to talk.
Only the fourth reader seemed to pursue an adequately thorough
checking/revision strategy. He was puzzled by his conclusion
because it did not connect with the rest of the text. He finally read
on to the end of the passage, saying [REVISE AND RECHECK],
“Oh, I see, it’s feeling I love you that’s as important as language
In spite of the limitations of their checking strategies, all these
readers seemed intent upon creating a connected text. This
particular sensitivity is expressed explicitly by the last reader’s
comment. “Seems like they’re jumping over, from one topic to
another, seems like they’re jumping from talking about adults and
baby talk, then going back to children.”


Proficient ESL Readers
The PESLs tended to follow the same monitoring process, with
three major differences: (a) they did not explicitly state or reflect on
plans or strategies; (b) they checked solutions less consistently and
less explicitly than did the PNs; (c) some identified a more inclusive
referent than most PNs had.
Like the PNs, most PESLs (6 out of 8) recognized the problem
and its source. However, only 2 identified that source explicitly.
Three supplied a referent apparently automatically and did not
comment on the process they used. Two others showed no
awareness of a problem and simply read the sentence aloud.
Only one PESL verbalized a strategic plan (“Cause when they
say, when they usually start sentences And this, they’re talking about
something they just spoke about in the last sentence”). Still half of
the solutions attempted (3 out of the 6) supplied a more specific and
appropriate referent than those of PNs:
9. [SOLVE + CHECK] It’s saying that this form of talking, it’s not just
important for babies to learn but it’s also a good way to communi-
10. [SOLVE] Oh I guess that the this means the baby talk, the
gentleness. I think it means the baby talk. [CHECK] So baby talk is
very important to the learning of communication.
11. [SOLVE + CHECK] Now it’s saying that the function of expressing
feelings without language but through baby talk is needed,
important function, it's important function just as learning the
language, communicate,
Three others identified baby talk or baby talk between lovers as
the antecedent (just as the PNs had) and failed to recognize or
resolve the inconsistencies of their solution. Two of these tried to
check their solutions, experienced some confusion, but didn’t revise
their solution:
12. [SOLVE] Between lovers, I think. I think between adults. [CHECK]
I think, maybe, I’m not sure. If you are fooling around, but I think
they should understand each other by then. So communication is
13. [SOLVE] Baby talk between adults. [CHECK] So like I said, I don’t
see the importance. So, it’s all right for them to talk like . . . what’s
the point?
14. [SOLVE + CHECK] I don’t consider baby talk an important
function for learning language and communication so we have
different views.
Thus, the proficient L1 and L2 readers seemed to experience
some problem with the sentence. They identified this as the source


of that problem, attempted to solve it, usually by looking back, and
finally checked that solution, usually by substituting the proposed
referent. It is surprising that PESLs were more successful in
identifying the referent than PNs were. Their responses suggest that
their checks were focused more on evaluating consistency within
the text than on relating their conclusions to the real world.

Less Proficient Readers

Both first and second language less proficient readers (LNs and
LESLs) seemed to recognize that a problem existed but showed less
explicit awareness of the source of the problem. Four (of the 6)
LESLs but only 1 (of the 3) LNs expressed some confusion about
the meaning of the sentence. Only one (a LESL) explicitly
questioned the referent of this; but did not take any subsequent
action. Two LESLs and the 1 LN supplied baby talk as the
15. LESL: I don’t know what it’s saying about the way the baby talk is
some kind of language and that helps to communicate between a
baby and an adult.
16. LESL: Baby talk is some kind of language and that helps to
communicate between a baby and an adult.
17. LN: I guess it’s relating to the baby talk that it’s important for
language learning to use a certain way of talking cause that’s the way
the young children learn.
The other readers in this category merely read the sentence aloud
or went on to the next sentence. The only less proficient reader to
verbalize anything akin to a strategic plan was the LN who
identified baby talk as the referent. He said, “I’m reading this over
again. Usually if I read something over again, I can usually get the
full understanding. It might take me a while, two or three times to
read it over.” He has a general, all-purpose strategy, rather than one
specific to this problem.
Thus, with respect to this particular problem, employment of
monitoring processes varied with reading proficiency. The readers
designated as more proficient tended to recognize that a problem
existed, identified the source of that problem, solved the problem
with varying degrees of success, and checked their solutions. Some
even revised and rechecked their solutions. Several also explicitly
verbalized their strategic plan. Most of the less proficient readers
failed to identify the source of the problem, solve the problem, or
check their solution.
L2 readers appeared to use a more careful, more text-centered,
language-based approach than did L1 readers. The more proficient


ESL readers tended not to report their planning process. This may
reflect an actual difference in the process or it may be the result of
reporting in a second language. Further research is necessary to
resolve this issue. In any case, their use of the monitoring process, as
evidenced by their awareness of the referent problem and by their
ability to take some remedial action, seemed as complete as that of
native speakers. The less proficient ESL readers seemed more
aware of their failures to understand than the less proficient native
speakers were; however, both groups lacked resources to take
remedial action.


It is often difficult to compare lexical processing of native speak-
ers and L2 readers since native speakers can be expected to have
many fewer problems understanding specific words than L2 read-
ers have. However, the first sentence of the passage, All the world’s
languages, from English to Urdu, share one special language: baby
talk, provided an ideal illustrative example since most of the read-
ers would not be expected to automatically know what Urdu
meant. Because of this, strategic resources and manipulations of the
reader are evident.
In general, proficient and less proficient readers used the
monitoring process very differently. The proficient readers used the
process less consciously, verbalizing fewer strategic plans and
checking less frequently than they had with the referent problem.
The less proficient readers participated more actively, identifying
the lexical problem appropriately, but still evidencing limited
resources for defining this unknown word. (See Table 2 for a
summary of these results.)

Proficient Native Speakers

All PNs identified Urdu as an unfamiliar word. Thus, they both
recognized that a problem existed and realized its source. All PNs
mentioned Urdu in their responses to the first sentence. Some (3 of
the 8) seemed to solve this lexical problem automatically. Although
they did not explicitly indicate that they had a problem, their
responses indicated that they had sought some definition of Urdu:
18. Baby talk is a universal language ranging from English to a language
called Urdu.
19. Who made up these extremes? How did they get the facts on all the
worlds languages?


Steps in Monitoring a Lexical Problem

20. It says from English to Urdu, as languages, and I don’t know where
that originated from.
These 3 readers seemed unconcerned with defining this unknown
word and focused on other issues. The first reader summarized and
generalized from the sentence; the second and third questioned the
information presented.
A second group (3 more of the 8) identified the problem more
explicitly and guessed that Urdu had something to do with Africa.
(Steps in the process are indicated within brackets.)
don’t know what Urdu is. [SOLVE] At first I thought it was Africa
‘cause it just has that sound. [REVISE] Then I saw that it must be
22. [SOLVE] It’s just probably another language I never heard of, like
African, or something like that.
23. [SOLVE] I guess it [Urdu] might be in the Middle East or in Africa.
Only the first reader explicitly checked his response and de-
scribed his strategic plan.
24. Before I chopped the sentence up because of the word Urdu. Now
I understand that they’re saying we all have something in common,
baby talk.
The other two readers mentioned that they didn’t understand Urdu,
but took no action.
Thus, for most of these proficient readers, the monitoring process
for this vocabulary problem involved the same steps as that used for


the referent problems (evaluation of comprehension, action, and
checking); however, PNs didn’t use the process as consciously or
verbalize as completely as they had when searching for a referent.
In addition, for PNs, this lexical problem seems to be resolved using
both text-based lexical and knowledge-based semantic cues.

Proficient ESL Readers

It’s surprising that while all the PNs commented on this unusual
word, most PESLs did not. Most seemed to allocate their attention
selectively, probably realizing that they did not need to understand
all the words in a sentence, or perhaps avoiding pronouncing an
unfamiliar word. Two articulated the strategy that seemed to have
influenced all of them:
25. [PLAN] I knew it was not important. [SOLVE] It’s a language ’cause
it says from English to something but I knew it was not important.
thought, what was Urdu? I never thought of that before. [PLAN]
Just went on ’cause I didn’t see it meant too much, just another
language. They’re talking about languages.
Only 3 explicitly mentioned the word Urdu at all. The other 5
PESL readers appeared to ignore the phrase from English to Urdu,
focusing instead on the main clause of the text. Their lack of
knowledge of Urdu neither interfered with their understanding nor
evoked a comment, suggesting that they were more concerned with
overall meaning than with understanding each word.
Several paraphrased the content, elaborating on it and making the
information more concrete. The first example below is from a
native speaker of Chinese:
27. I think baby talk is saying, how they saying that Chinese baby and
other babies have things in common. I thought when the Chinese
baby communicates with the other baby, do they understand each
28. I was trying to imagine someone from another country trying to talk
to a baby.
One reader, a native speaker of Spanish, was the only PESL to
revise her understanding. She initially guessed that the passage
would be about language learning but corrected her impression
once the text made it clear this assumption was not appropriate:
29. I thought this was going to be about languages, Spanish, English, but
then it says that they share one special kind of speech: baby talk. It’s
really that every language begins with a basic way of speaking,
which all of them share in common, which is baby talk.


Thus, among the proficient readers, both the PNs and PESLs, this
obvious lexical problem caused no interruption in understanding.
Readers had different degrees of apparent awareness of the
problem and different types of strategies for dealing with the
problem. Once again, the PESLs showed efficient and sophisticated
strategies for coping with this language-based problem.

Less Proficient Readers

Both LNs and LESLs identified Urdu as a problem; yet their
attempts at problem solving were extremely limited. Proportion-
ately more LESLs (5 of the 6) identified the problem than LNs (1
out of 3). However, only two LESLs attempted to define the word.
One thought Urdu referred to a language, the other thought it
referred to some kind of people:
30. I don’t know what Urdu is, but is a kind of language, that share a
special kind of speech by baby talk.
31. From English to Urdu mean talking in English to the people, Urdu
means a kind of people.
The other LESL readers questioned the word but made no attempt
to solve the problem they had recognized.
The LNs tended to ignore Urdu. Only one recognized the
problem and attempted to solve it:
32. From English to Urdu. I don’t know what is Urdu. Some kind of
baby talk. It’s similar to baby talk.
One LN expressed confusion but did not specifically identify the
problem. The other one did not comment on the problem at all.
33. My mind just wanders when I read that. I don’t know why. I don’t
get understanding.
34. I think it’s going to be interesting because it has something to do
with babies and I love little babies.
We see similar patterns with other lexical problems throughout
the passage. When proficient readers had difficulty understanding a
word, they elaborated on the text, fleshing it out and making it
concrete, providing an example, or searching for reasons why the
information was given. Although PNs did not understand every
word they read, their lexical problems did not interfere with their
on-going reading. The PESL readers experienced more word-level
problems. They tended to read sentence by sentence and to
specifically question vocabulary problems, but few of these readers
had trouble solving the problems they identified. Thus, their ability
to solve their problems enabled them to continue reading fluently.


Some of the strategies they used were
1. Grammatical approach: “To-d-d-l-e-r is the subject of the
sentence so I have to know this word. ”
2. Deciding on importance: “I don’t think I need this.”
3. Omitting/focusing on what is understood
4. Reading to look for examples: “After reading the last sentence, I
understand what your turn is.”
5. Rereading
6. Using the base form of a word: “When I got to the word
melodious, I knew it was melody.”
7. Using background knowledge: “I try to think of an example
where I talk to baby this way.”
Less proficient ESL readers were also aware of their lexical
problems, but they seemed to have few resources for solving their
problems. Most merely noted the problem or used their real-world
knowledge to guess, sometimes wildly, at the meaning. In the case
of vocabulary, LESLs had recognition but not resources.
In general, the less proficient native speakers experienced the
same lexical problems as the second language readers and seemed
to rely almost exclusively on a lexical standard for evaluating their
comprehension. When they didn’t understand words, they felt they
didn’t understand the sentence; when they did, they felt they had
understood the sentence. One LN, in particular, was frequently
stymied by decoding and vocabulary problems. When he didn’t
understand a word, he seemed to feel totally defeated:
35. I’ll get to a word that I won’t know and then I try to figure the word
out. I’m nervous ’cause I won’t know it. I’m afraid I won’t get the
complete understanding.
Another LN also seemed to feel defeated by unfamiliar words,
but was much less able to articulate her problem:
36. Just like a bunch of words, like all of a sudden switch to another
language. Just hope it will clear up.
The less proficient readers identified many vocabulary problems,
but they had limited resources for solving them. The emphasis of
the less proficient readers on identifying unfamiliar words suggests
their word-level model of reading; the more proficient readers
seemed to have a more meaning-based approach and didn’t worry
about the meaning of words if they could extract the gist of the


The purpose of this paper was to illustrate the comprehension-
monitoring process used by first and second language readers of
English in dealing with two types of language-based problems
commonly met when reading expository prose. The data suggest
that there is a regular process that operates similarly for native
speakers of English and second language readers. This study
revealed that process in progress, at least as it operated in two
specific instances. Three phases and six specific steps of the process
were identified: the evaluation phase (problem recognition and
problem source identification), the action phase (strategic plan and
action/solution attempt), and the checking phase (check and
These phases were used with different degrees of completeness
and explicitness for each of the two language-based problems
examined. The process operated more fully with respect to the
referent problem than the lexical problem. When searching for an
appropriate referent, the most proficient readers tended to
verbalize strategic plans and to check their solutions. When dealing
with lexical problems, however, the process was truncated, and
there were fewer statements of strategic plans and fewer checks of
Differences that existed in monitoring seemed due more to read-
ing proficiency than to the language backgrounds of the readers.
Most readers seemed to recognize when a problem existed;
however, the proficient readers identified the problem’s source
more frequently and more explicitly than did the less proficient.
They also verbalized their strategic plans more frequently;
proficient L1 readers seemed more likely to do so than proficient L2
readers. Whether this was the result of cognitive awareness or of
relative comfort with verbalizing in English is not clear since
readers were not explicitly questioned about this.
The less proficient readers used the process incompletely. Even
when they expressed a problem with understanding, they did not
seem to know what to do next, supporting Baker’s (1985) contention
that evaluation. and regulation of comprehension are two separate
processes. Two of these less proficient readers, one native speaker
of English and one native speaker of Chinese, seemed to have more
control of the monitoring process than the others in this group.
Block (1986) categorized those readers as “integrators,” readers who
were concerned about connecting text and who used “good reader”
strategies. It may be that their placement as less proficient was a
result of slow rather than unskilled reading.


Differences did exist between the monitoring of L1 and L2
readers. The proficient ESL readers seemed at least as able as the
proficient native speakers to recognize and solve the problems that
the text presented. Although, during the referent search, they did
not verbalize their internal process as much as native speakers did,
their solutions were as pertinent. When solving the vocabulary
problem, several ignored an unknown word they saw as unessential,
a phenomenon Baker (1989) found to be indicative of competent
adult readers. “Adults were more likely than younger students to
recognize that some unknown words are more disruptive to overall
comprehension than others” (p. 258). Language-based/text-based
problems may exist for the proficient ESL readers; however, these
problems did not interfere with their basic comprehension of the
text because they had the resources to solve them. Strategic
resources, thus, seem more important than specific linguistic
knowledge for these readers.
Both the less proficient L1 and L2 readers lacked awareness of
problems and the ability or inclination to take action when they
were aware of a problem. The reading process of L2 readers may
be less automatic and slower than that of L1 readers of similar
ability, but adequate resources for evaluation and regulation of
reading can help compensate for this.
Moreover, the less proficient readers, particularly the native
speakers, seemed to favor a local, word-based processing strategy
while the more proficient readers tended to prefer a more global
meaning-based one. This supports the research conducted with
native speakers of English (e.g., Garner, 1981; Garner & Kraus,
1982; Myers & Paris, 1978) and second language speakers of English
(Cohen et al., 1979).
Ten years ago, Cohen et al. (1979) suggested that a major ques-
tion for L2 reading researchers should be “How do learners go
about solving problems in reading?” (p. 563) rather than “What is
problematic for nonnative readers when reading in English?”
(p. 552) The present examination of a few readers indicates that the
problem-solving process to which Cohen et al. allude has regular
steps but may differ as the problem differs. We still have a lot to
discover about this process. We need to know how L1 and L2
readers deal with a variety of apparent reading problems, both
those that are language based and those that are knowledge or
schema based. Hopefully, this paper has begun that exploration.
What does all of this say to the classroom teacher? First, it
suggests the value of process-oriented instruction. First and
foremost is the need to encourage an awareness of the nature of the
reading process. Active readers do not expect to understand


everything as they read. They are prepared to monitor their
understanding and to question what they understand. ESL readers
should be made aware that questioning and monitoring is a part of
good reading, not the result of imperfect knowledge of their second
Second, students should be taught to search for the sources of
their problems. A prime difference between the more and less
proficient readers we observed was their awareness of the source of
the problems they encountered. Such awareness is a necessary
predecessor to taking action. Finding the source may be as
important as having the resource.
L1 studies of strategy training suggest that especially important to
training is the awareness that specific strategies can be helpful in
solving particular problems. The ability to attribute a comprehen-
sion difficulty to a source rather than to lack of skill is an important
part of effective reading. The same may be true for second
language readers. As Carrell (1989) says,
Too often students in second language reading programs, who receive
instruction only in the skills or strategies, fail to use them intelligently
and on their own volition because they do not appreciate the reasons
why such strategies are useful nor do they understand where and when
to use them. Adding instruction in “awareness” or knowledge about a
strategy’s evaluation, rationale, and utility should greatly increase the
positive outcomes of instruction. (p. 129)
Third, the results of this investigation of monitoring run counter
to the common sense wisdom of the reading classroom. For years,
ESL teachers have sought to predigest printed material for students
believing that comprehension depends on understanding of all the
language features of the text. Thus, reading traditionally enters the
ESL classroom late and the texts used are most often adapted so
that students will understand every word they read. When graded
texts are not used, difficult vocabulary, structures, and concepts are
pretaught so that the process will be smooth. This investigation
reveals that the process is not a smooth one and suggests that, by
chewing up the text for students, we are distorting the process and
not preparing them to eat on their own. The proficient L2 readers in
this study did not have to understand all the words or structures to
understand what they read. Part of the strength of their reading was
in being able to decide which problems they could ignore and
which they had to solve.
The responses of students in this paper suggest that certain types
of monitoring may promote a more global processing style than
others. L1 research has shown instructional approaches may


actually reinforce or encourage the use of negative reading
strategies such as focusing on word-level processing. (See
Stanovich, 1986, for a review of this literature.) ESL reading
teachers who focus on developing vocabulary instead of on
building cognitive and metacognitive resources may similarly
reinforce their students’ word-oriented processing style.
It may be that it is more important to monitor for certain types of
difficulties than for others; however, we do not as yet know enough
about how different types of L2 readers deal with specific problems
and about how these affect comprehension to draw any definite
conclusions. What is clear is that it is important to develop students’
resources for identifying and solving their own reading problems.
We cannot expect that ESL readers will continue to meet easy-to-
read texts as they go through school. Teaching them awareness of
the metacognitive process and of their strategic resources will give
them ways to keep on nourishing themselves through school.
For those of us interested in improving the way ESL students
function in academic settings, the appropriate area of instruction
seems to be process-oriented strategy use rather than presentation
of content-oriented background knowledge or instruction in
particular linguistic features in isolation. Teaching students that
problems exist when reading and that there are ways of solving
them may be more important than teaching the meaning of specific
words, phrases, and concepts. If we continue to teach background
knowledge and linguistic features, we continue to apply only a
bandaid to the problem. We are like the cold remedy that masks the
symptoms without curing the illness. Students will have to keep
taking us, so to speak, because they haven’t learned to cure
themselves, that is, to become independent learners.

This paper is a revision of one originally presented at the 23rd Annual TESOL
Convention, San Antonio, TX. The research for this paper was supported in part by
grants from the Research Award Program of the Professional Staff Congress and
The City University of New York (CUNY), and the CUNY Office of Special
Programs. I thank both for their support. I also thank Marta Martino and Jack
Gantzer, fellow CUNY ESOL teachers, for their careful reading of versions of this
paper and for their supportive friendship throughout.


Ellen L. Block is Associate Professor in the Academic Skills Department at Baruch
College, a branch of The City University of New York. She received the 1986
TESOL Research Interest Section/Newbury House Award for Distinguished
Research and has published articles and made conference presentations on reading
and thinking aloud. She is currently on the Editorial Advisory Board of a new
journal, College ESL.

Adams, M. J., & Collins, A. (1985). A schema-theoretic view of reading. In
H. Singer & R. B. Ruddell (Eds.), Theoretical models and processes of
reading, third edition (pp. 404-425). Newark, DE: International Reading
Afflerbach, P. (1986). The influence of prior knowledge on expert readers’
importance assignment processes. In J. A. Niles & R.V. Lalik (Eds.),
Solving problems in literacy: Learners, teachers, and researchers
(pp. 30-40). Chicago, IL: National Reading Conference.
Alderson, J. C., & Urquhart, A. H. (1984). Introduction: What is reading?
In J. C. Alderson & A. H. Urquhart (Eds.), Reading in a foreign
language (pp. xv-xxviii). New York: Longman.
Anderson, R. C., & Pearson, P. D. (1984). A schema-theoretic view of
basic processes in reading. In P. D. Pearson (Ed.), Handbook of reading
research (pp. 255-292). New York: Longman.
August, D. L., Flavell, J. H., & Clift, R. (1984). Comparison of compre-
hension monitoring of skilled and less skilled readers. Reading Research
Quarterly, 20, 39-53.
Baker, L. (1989). Developmental change in readers’ responses to unknown
words. Journal of Reading Behavior, 21, 241-260.
Baker, L. (1985). How do we know when we don’t understand? Standards
for evaluating text comprehension. In D. L. Forrest-Pressley, G. E.
MacKinnon, & T. G. Waller (Eds.), Metacognition, cognition, and
human performance (Vol. 1, pp. 155-205). Orlando, FL: Academic
Baker, L. (1979). Comprehension monitoring: Identifying and coping with
text confusions (Tech. Rep. No. 145). Champaign: University of Illinois,
Center for the Study of Reading. (ERIC Document Reproduction
Service No. ED 177 525)
Baker, L., & Anderson, R. I. (1982). Effects of inconsistent information on
text processing: Evidence for comprehension monitoring. Reading
Research Quarterly, 17, 281-294.
Baker, L., & Brown, A. (1984). Metacognitive skills and reading. In
P. D. Pearson (Ed.), Handbook of reading research (pp. 353-394). New
York: Longman.
Bialystok, E., & Ryan, E. (1985). A metacognitive framework for the
development of first and second language skills.In
D. L. Forrest-Pressley, G. E. MacKinnon, & T. G. Waller (Eds.),
Metacognition, cognition, a n d h u m a n p e r f o r m a n c e ( V o l . 1 ,
pp. 207-252). Orlando, FL: Academic Press.


Block, E. (1986). The comprehension strategies of second language
readers. TESOL Quarterly, 20 (3), 463-494.
Canney, G., & Winograd, P. (1979). Schemata for reading and reading
comprehension performance (Tech. Rep. No. 120). Champaign, IL:
University of Illinois, Center for the Study of Reading. (ERIC
Document Reproduction Service No. ED 169 520)
Carrell, P. L. (1983). Background knowledge in second language
comprehension. Language Learning and Communication, 2, 25-34.
Carrell, P. L. (1988). Interactive text processing: Implications for ESL/
second language reading classrooms. In P. L. Carrell, J. Devine, &
D. Eskey (Eds.), Interactive approaches to second language reading
(pp. 239-259). New York: Cambridge University press.
Carrell, P. L. (1989). Metacognitive awareness and second language
reading. Modern Language Journal, 73, 121-133.
Carrell, P. L., Devine, J., & Eskey, D. (1988). Interactive approaches to
second language reading. New York: Cambridge University Press.
Carrell, P. L., & Eisterhold, J. (1983). Schema theory and ESL reading
pedagogy. TESOL Quarterly, 17, 553-573.
Caron, T. (1989). Strategies for reading expository prose. In
S. McCormick & J. Zutell (Eds.), Cognitive and social perspectives for
literacy research and instruction (pp. 293-300). Chicago, IL: National
Reading Conference.
Casanave, C. P. (1988). Comprehension monitoring in ESL reading: A
neglected essential. TESOL Quarterly, 22 (2), 283-302.
Chapman, L. J. (1979). Confirming children’s use of cohesive ties in text:
Pronouns. The Reading Teacher, 33, 317-322.
Clarke, M. A. (1979). Reading in Spanish and English: Evidence from
adult ESL students. Language Learning, 29, 121-150.
Clarke, M. A. (1980). The short circuit hypothesis of ESL reading—or
when language competence interferes with reading performance.
Modern Language Journal, 27, 135-154.
Cohen, A., Glasman, H., Rosenbaum-Cohen, P., Ferrara, J., & Fine, J.
(1979). Reading for specialized purposes Discourse analysis and the use
of student informants. TESOL Quarterly, 13 (4), 551-564.
Cowan, J. R. (1976). Reading, perceptual strategies and contrastive
analysis. Language Learning, 26, 95-109.
Cummins, J. (1980). The cross-lingual dimensions of language proficiency:
Implications for bilingual education and the optimal age issue. TESOL
Quarterly, 14 (2), 175-187.
Cziko, G. A. (1980). Language competence and reading strategies: A
comparison of first- and second-language oral reading errors. Language
Learning, 30, 101-114.
Devine, J. (1988a). A case study of two readers: Models of reading and
reading performance. In P. L. Carrell, J. Devine, & D. Eskey (Eds.),
Interactive approaches to second language reading (pp. 127-139). New
York: Cambridge University Press.


Devine, J. (1988b). The relationship between general language compe-
tence and second language reading proficiency: Implications for
teaching. In P. L. Carrell, J. Devine, & D. Eskey (Eds.), interactive
approaches to second language reading (pp. 261-277). New York:
Cambridge University Press.
Educational Testing Service. (1978). Descriptive Test of Language Skills.
Princeton, NJ: Author.
Forrest-Pressley, D. L., & Waller, T. G. (1984). Cognition, metacognition,
and reading. New York: Springer-Verlag.
Gambrell, L. B., & Heathington, B. S. (1981). Adult disabled readers’
metacognitive awareness about reading tasks and strategies. Journal of
Reading Behavior, 13, 215-222.
Garner, R. (1987). Metacognition and reading comprehension. Norwood,
NJ: Ablex.
Garner, R. (1981). Monitoring of passage inconsistency among poor
comprehenders: A preliminary test of the “piecemeal processing”
explanation. Journal of Educational Research, 74, 159-162.
Garner, R., & Kraus, C. (1982). Good and poor comprehended differences
in knowing and regulating reading behaviors. Educational Research
Quarterly, 6, 5-12.
Goodman, K. S. (1973). Psycholinguistic universals of the reading process.
In F. Smith (Ed.), Psycholinguistics and reading (pp. 21-29). New York:
Holt, Rinehart and Winston.
Grabe, W. (1991). Current developments in second language reading
research. TESOL Quarterly, 25 (3), 375-406.
Hakuta, K., Ferdman, B. M., & Diaz, R. D. (1987). Bilingualism and
cognitive development: Three perspectives. In S. Rosenberg (Ed.),
Advances in applied psycholinguistics: Vol. 2. Reading, writing, and
language learning (pp. 284-319). New York: Cambridge University
Hare, V. C. (1981). Readers’ problem identification and problem solving
strategies for high and low knowledge articles. Journal of Reading
Behavior, 13, 359-365.
Hosenfeld, C. (1977). A preliminary investigation of the reading strategies
of successful and nonsuccessful second language learners. System, 5,
Hudson, T. (1982). The effects of induced schemata on the “short circuit”
in L2 reading: Non-decoding factors in L2 reading performance.
Language Learning, 32, 1-31.
Mackay, R. (1979). Teaching the information-gathering skills. In
R. Mackay, B. Barkman, & R. Jordan (Eds.), Reading in a second
language (pp. 79-90). Rowley, MA: Newbury House.
Myers, M., & Paris, S. G. (1978). Children’s metacognitive knowledge
about reading. Journal of Educational Psychology, 70, 680-690.
Newell, A., & Simon, H. A. (1972). Human problem solving. Englewood
Cliffs, NJ: Prentice Hall.


Olshavsky, J. (1977). Reading as problem-solving: An investigation of
strategies. Reading Research Quarterly, 12, 654-674.
Olson, G., Duffy, S., & Mack, R. (1984). Think-out-loud as a method for
studying real-time comprehension processes. In D. E. Kieras &
M. A. Just (Eds.), New methods in reading comprehension research
(pp. 253-286). Hillsdale, NJ: Lawrence Erlbaum.
Paris, S. G., & Myers, M. (1981). Comprehension monitoring, memory,
and study strategies of good and poor readers. Journal of Reading
Behavior, 13, 5-22.
Paris, S. G., Wasik, B. A., & Van der Westhuizen, G. (1988). Meta-
metacognition: A review of research on metacognition and reading. In
J. E. Readence, R. S. Baldwin (Eds.), Dialogues in literacy research
(pp. 143-166). Chicago, IL: National Reading Conference.
Rubin, Z., & McNeil, E. B. (1981). The psychology of being human (3rd
ed.). New York: Harper & Row.
Rumelhart, D. E. (1977). Toward an interactive model of reading. In
S. Dornic (Ed.), Attention and performance (Vol. 6, pp. 573-603). New
York: Academic Press.
Sarig, G. (1987). High-level reading in the first and in the foreign language:
Some comparative process data. In J. Devine, P. Carrell, & D. Eskey
(Eds.), Research in reading in English as a second language
(pp. 105-120). Washington, DC: TESOL.
Scardamalia, M., & Bereiter, C. (1984). Development of strategies in text
processing. In H. Mandl, N. L. Stein, & T. Trabasso (Eds.), Learning
and comprehension of text (pp. 379-406). Hillsdale, NJ: Lawrence
Smith, H. K. (1967). The responses of good and poor readers when asked
to read for different purposes. Reading Research Quarterly, 3, 53-83.
Stanovich, K. E. (1986). Matthew effects in reading: Some consequences
of individual differences in the acquisition of literacy. Reading Research
Quarterly, 21, 360-407.
Stanovich, K. E. (1980). Toward an interactive-compensatory model of
individual differences in the development of reading fluency. Reading
Research Quarterly, 16, 32-71.
Steffensen, M. S., Joag-dev, C., & Anderson, R. C. (1979). A cross-
cultural perspective on reading comprehension. Reading Research
Quarterly, 15, 10-29.
Winograd, P., & Johnston, P. (1982). Comprehension monitoring and the
error detection paradigm. Journal of Reading Behavior, 14, 61-76.
Yorio, C. (1971). Some sources of reading problems for foreign language
learners. Language Learning, 21, 107-115.
Zabrucky, K., & Ratner, H. H. (1989). Effects of reading ability on
children’s comprehension evaluation and regulation. Journal of Reading
Behavior, 21, 69-84.



All of the World'S languages, from English to Urdu, share one special kind of speech: baby
talk. Recent research has confirmed that in spite of the great differences among cultures and
languages, the general properties of speech used with babies who are learning to talk remain
the same.
Baby talk sounds different from adult speech. When talking to 1- and 2-year olds, adults
usually raise the pitch of their voices and adopt a “sing-song” intonation, in which the voice
rises and drops dramatically, often ending a sentence at a high point. (Imagine the way you
would say to a baby, “Hi, Johnny. You’re playing with your teddy, aren’t you.”)
What is the point of these peculiarities? Research has shown that babies prefer sounds in
higher pitch ranges (Kearsley, 1973). Adults may quickly learn that they are more likely to get
a smile or a satisfied gurgle from a baby when they raise their voices a bit. And the melodious
rise and fall of the speech signal may hold the baby’s attention—something that isn’t easy to
do. For the toddler who has begun to utter a few words, the rising voice at the end of the
sentence serves as a signal: “Your turn.” It marks the end of the adult’s verbal offering and
invites the child to make a response.
Adult speech to toddlers is also characterized by short sentences, limited vocabulary, and
straight-forward grammar. There are lots of questions and there is plenty of repetition (Snow,
1972). Furthermore, speech to beginning talkers tends to be tied to the here and now, with
few references to the past or future. A father is much more likely to say, “See the birdie,
Franny?” than “Do you remember the bird we saw yesterday?” The grammatical simplicity
and concreteness of baby talk help make the structure and rules of language clearer to
someone just starting to learn it, and they help ease communication with a small person who
cannot yet understand much speech.
Adults seem to catch on to baby talk quite naturally. Catherine Snow (1972) found that
non-mothers (who had almost no experience with babies) made the same speech changes
when they talk to babies that mothers did. And Marilyn Shatz and Rochel Gelman (1973)
found that even 4-year-old children will make similar speech modifications when talking to
2-year-olds. Babies themselves help to shape baby talk, through their reactions to adult
utterances. When mothers were asked to talk to an imaginary baby, they did not simplify
their speech as much as when they spoke to a real one (Snow, 1972). The child’s presence—
giving evidence of comprehension, boredom, or pleasure—was necessary to elicit “true”
baby talk from the mothers. True baby talk, with its particular grammatical simplifications,
does not appear in parents until the baby is about 18 months old and begins to demonstrate
some understanding of what is being said (Phillips, 1973).
Roger Brown (1977) suggests that there is something else baby talk can do besides helping
babies learn to talk: It can express affection in ways that normal speech can’t. He points out
that sometimes baby talk occurs between adults, but that such behavior is usually limited to
lovers. And this may be as important a function as language learning and communication.
Children need to learn to talk. They need to understand “Stay away from the stove” and
“Don’t eat the Swedish ivy.” But they also need to hear “I love you” and to feel the meaning
of these words even before the words themselves are actually understood.

1 The
sentences discussed are italicized. Other problems mentioned are underlined. The text
was presented in normal type for the readers in the study.


TESOL QUARTERLY, Vol. 26, No. 2, Summer 1992

The Effect of Speech Modification, Prior

Knowledge, and Listening Proficiency
on EFL Lecture Learning
The Pennsylvania State University/World Journalism and Communication College

The Pennsylvania State University

This study investigates the listening comprehension of 388 high-in-

termediate listening proficiency (HILP) and low-intermediate lis-
tening proficiency (LILP) Chinese students of English as a foreign
language. These students listened to a lecture, the discourse of
which was (a) familiar-unmodified, (b) familiar-modified, (c)
unfamiliar-unmodified, or (d) unfamiliar-modified. The modified
discourse contained information redundancies and elaborations.
After the lecture, the EFL subjects took a multiple-choice exam test-
ing recognition of information presented in the lecture and general
knowledge of the familiar (“Confucius and Confucianism”) and
unfamiliar (“The Amish People”) topics. A significant interaction
between speech modification (redundant vs. nonredundant speech)
and listening proficiency (HILP vs. LILP) indicated that the HILP
students benefited from speech modification, which entailed elabo-
ration/redundancy of information, but the LILP students did not. A
significant interaction between prior knowledge (familiar vs.
unfamiliar topic) and test type (passage-independent vs. passage-
dependent items) was also found. For both the HILP and LILP sub-
jects, prior knowledge had a significant impact on subjects’ memory
for information contained in the passage-independent test items on
the postlecture comprehension test. Those EFL subjects who
listened to the familiar-topic lecture on Confucius had higher
passage-independent than passage-dependent scores. There was no
difference in the performance on the passage-independent and
passage-dependent items of those who listened to the lecture on an
unfamiliar topic (the Amish). However, the passage-independent
performance of subjects who listened to the familiar topic lecture
was superior to that of those who listened to the lecture on the
unfamiliar topic. Subjects’ performance on passage-dependent
items did not differ significantly whether the familiar or unfamiliar
topic was presented. Implications of the findings for assessing and
teaching EFL listening comprehension are suggested.

Although understanding lectures in English is perceived to be a
vital component of academic survival for international students
studying in English-speaking universities (Dunkel & Davey, 1989),
many students whose native language is other than English appear
to have difficulty understanding and retaining lecture information
in English (Dunkel, Mishra, & Berliner, 1989). Their attempts to
comprehend and retain English lecture information may be
thwarted by a number of cognitive and linguistic factors as well as
academic and cultural issues, including: (a) inability or lack of
opportunity to engage in communicative interaction with the
second/foreign language teacher or lecturer (Pica, Young, &
Doughty, 1987); (b) inability to detect the main points of the
lecture, or “to grasp the usual goals of particular genres of discourse
situation of which the discourse is a part” (Olsen & Huckin, 1990,
p. 41); (c) unfamiliarity with the structure and type of the discourse
(Barnitz, 1986); (d) inability to apprehend discourse markers and
logical relationships in the English lecture (Chaudron & Richards,
1986; DeCarrico & Nattinger, 1988); (e) inability to comprehend
lecture speech delivered at faster rates of speed (i.e., discourse
delivered at “broadcast” speed, or 165 words per minute (wpm),
rather than “lecturelike” speed, or 107 wpm) (Dunkel, 1988a); (f)
limited short-term memory for English input (Dunkel, 1985;
Dunkel, Mishra, & Berliner, 1989); (g) failure to use appropriate
cognitive or learning strategies (Dunkel, 1988b; Oxford, 1990;
Padron, Knight, & Waxman, 1986; Wenden & Rubin, 1987); (h) poor
inferencing abilities in English (Carrell, 1984a); (i) limited
proficiency in English (Dunkel, Mishra, & Berliner, 1989); (j) lack of
prior knowledge about the content of the spoken or written text
(Aron, 1986; Carrell, 1984d, 1987; Carrell & Eisterhold, 1983;
Connor, 1984); and (k) inability to process L2 input devoid of
speech modification such as elaborations or redundancies (cf.
Chaudron, 1983b). In the present study, we explored the effects of
three of the many difficulties confronting L2 learners when they (a)
have limited listening proficiency in English; (b) lack prior
knowledge about the topic of the L2 lecture; and (c) are not
supplied with modified speech (i.e., speech containing elaborated
and redundant information) during the L2 lecture presentation.
We focused on these particular issues for several reasons.
Whereas it has been demonstrated that certain types of speech
modification, such as restatements and redundancy, have a
facilitative effect on listeners’ comprehension of L2 input (see
Chaudron, 1983b), it is not clear that the beneficial effect of
redundancy holds equally well for higher and lower levels of L2
proficiency alike, and for learners of English as a foreign language
and English as a second language alike. The latter is not known


since the majority of the studies on the effect of speech modifica-
tion carried out to date have involved subjects studying English in
a second, rather than foreign, language environment. We also de-
cided to investigate the influence of prior knowledge on the lecture
information processing of L2 listeners because the majority of the
studies examining the effect of prior knowledge on comprehension
of L2 information have involved reading comprehension (cf. Aron,
1986; Carrell, 1984a, 1984b, 1984c, 1984d, 1987; Johnson, 1981, 1982;
Nelson, 1987) rather than listening comprehension. It was thought
that an investigation of the role prior knowledge plays in the pro-
cessing of L2 oral discourse would help shed light on the processing
of information by L2 learners, in general, and by EFL learners,

Speech Modification and L2 Discourse Processing
In communicative interactions with nonnative speakers of
English, native speakers of English often adjust or modify their
speech in their attempt to make it more comprehensible to the
second language listener (Long, 1983). According to Long, the need
to provide comprehensible input for nonproficient listeners is a
central component of Krashen’s input hypothesis, which postulates
that mere exposure to L2 input does not ensure comprehension and
intake of the L2 information; rather learners need to have
comprehensible (oftentimes “modified”) input for second language
comprehension and acquisition to occur. Speakers modify their
speech to help the nonnative listener better understand the
information conveyed or to help the nonnative listener continue to
communicate with the native speaker (cf. Brown, 1977; Brown &
Hilferty, 1986; Chaudron, 1983a, 1986; Faerch & Kasper, 1986;
Hatch, 1983; Henzl, 1973, 1975; Kelch, 1985; Krashen, 1980a, 1980b;
Long, 1983; Pica, Doughty, & Young, 1986; Snow & Hoefnagel-
Hohle, 1982; Wesche & Ready, 1985; Wong Fillmore, 1982, 1985).
The speech modifications can entail a simplification of linguistic
form involving an alteration in the syntax (e.g., the use of two
simple sentences rather than one complex sentence) and/or a
change in the lexis of the message (e.g., using “to hold on tightly to”
rather than “to cling to”). (See Parker & Chaudron, 1987. )
Although the notion of providing L2 listeners with modified input
is intuitively appealing and widely accepted, relatively little is
known about precisely which types of modifications actually


augment, or possibly obstruct, the intake of aurally received
information. In the attempt to clarify the relationship between
speech modification and L2 listening comprehension, researchers
have conducted studies in the hope of identifying some of the
specific variables that make spoken input, such as academic
discourse, more comprehensible to L2 learners. For example, in a
study of L2 listeners’ comprehension of modified and unmodified
lectures, Dunkel (1988a) found that presentation of the lecture in
conversational (i. e., modified) style rather than so-called broadcast
(i.e., unmodified) style led to higher achievement on a postlecture
recognition quiz and to an increase in the quantity of lecture notes
taken under this conversational style lecture condition. In this case,
the modified speech augmented L2 comprehension.
Although positive effects of speech modification have emerged
from the research on L2 academic and conversational listening, the
efficacy of all forms of speech modification has not been
established. In a study of the effects of different types of input
modifications on the L2 listeners’ comprehension of academic
discourse, for example, Parker and Chaudron (1987) found that
linguistic simplifications in the form of simplified syntax and
vocabulary failed to have a significantly positive effect on
comprehension of the information. However, an elaborative
modification (a repetition of the information) and clear segmenting
of the thematic structure of the communication was found to
enhance comprehension of the L2 information presented orally.
Blau (1990) investigated the effect of syntax, speed, and pauses
on EFL and ESL listening comprehension. She found that
mechanically reducing the velocity of speech with the aid of a
variable-speed tape recorder from faster (170 wpm) to slower (145
wpm) rates of speech did not enhance the comprehension of Polish
or Puerto Rican listeners “except at the lowest levels of L2
proficiency” (p. 752). She also found that reducing the number of
complex sentences and increasing the number of simple sentences
contained in the input did not augment L2 listeners’ comprehension
of read text. However, Blau did find that the inclusion of pauses at
constituent boundaries in the input enhanced the L2 listeners’
comprehension. In another study of speech modification, Chaudron
and Richards (1986) found that inclusion of macromarkers signaling
major propositions or the important transition points within the
lecture improved listeners’ comprehension and retention of lecture
information. However, micromarkers signaling intersentential
relations, framing of segments, and pause fillers did not aid the
learners’ retention of the lecture information. Pica, Young, and


Doughty (1987) noted that input simplified a priori proved to
support L2 learners’ comprehension to a lesser extent than did
elaborations made to the input during negotiated interaction.
Although some researchers, such as Parker and Chaudron (1987),
have determined that certain modifications made to the input (e.g.,
linguistic simplifications) did not augment L2 comprehension to a
statistically significant degree, other researchers have, in fact,
concluded that the modification of speech or text may actually
hinder the reading and listening comprehension of second language
learners (cf. Blau, 1982; Chaudron, 1983a, 1983b; Honeyfield, 1977;
Schinke-Llano, 1983, 1986; Snow & Ferguson, 1977; Speidel,
Roland, & Kobayashi, 1985). Honeyfield’s (1977) study of the effect
of simplification on reading comprehension, for example, disclosed
that simplification of linguistic structure and content material could
lead students to develop reading strategies that are inappropriate
for unsimplified English. Chaudron (1983b) concluded that certain
types of speech adjustments made for L2 listeners (e. g.,
simplification of surface structure) actually impeded rather than
facilitated comprehension. Schinke-Llano (1983, 1986) has also
argued against ESL teachers’ practice of modifying their speech in
the elementary school classroom, suggesting that simplifying the
input for limited English proficient (LEP) students in the classroom
may be detrimental to the LEP students’ self-esteem as well as to
their second language acquisition. Echoing Schinke-Llano’s
concern, Klein (1986) speculates that if native-speaker teachers
modify their speech, the learner may interpret the modification as a
sign of social distance and condescension on the part of the teacher
and feel insulted. Furthermore, the modification may hinder
comprehension if the learner is fairly advanced in the language.
Finally, there always remains the possibility that modifying
expressions in the utterance in order to simplify the meaning may
actually alter the meaning of the message, thereby supplying
listeners with a distortion of the intended message.
It is clear that native speakers modify their speech in com-
municative interactions with nonnative speakers to augment L2
listeners’ comprehension (cf. Chaudron, 1983a, 1983b, 1986, 1988;
Clyne, 1981; Freed, 1980). What is not clear in an academic context,
however, is (a) whether certain types of speech modifications
support L2 lecture information processing to a greater extent than
others, (b) whether speech modification of L2 lecture input benefits
all levels of L2 proficiency in equal fashion, and (c) whether the
listener’s familiarity with the topic of the L2 lecture plays a role in
catalyzing the positive or negative effects of speech modification.


Prior Knowledge and L2 Discourse Processing
EFL or ESL students have difficulty understanding and recalling
information contained in spoken and written discourse when they
lack familiarity with the topic of, or the cultural elements contained
in, the discourse (Carrell, 1983; Connor, 1984). Carrell and
Eisterhold (1983) have suggested that for ESL learners, if there is a
mismatch between the prior knowledge assumed in the discourse
and the prior knowledge possessed by the reader or listener, a
breakdown in communication will occur.
The critical role prior knowledge plays in language comprehen-
sion has been articulated in schema theory and documented in the
work of Anderson (1984), Anderson and Barnitz (1984), Anderson,
Spiro, and Anderson (1978), Barnitz (1986), Bartlett (1932), Carrell
(1983, 1984a), Connor (1984), Jenkins (1987), Kintsch and Greene
(1978), Long (1989), Rumelhart and Ortony (1977), Rumelhart
(1980), Schallert (1982), and Swaffar (1988). The basic tenet of
schema theory posits that written text, or spoken discourse, does not
carry meaning in and of itself. Rather, meaning occurs as a result of
the interaction between the reader’s or listener’s prior knowledge
about the world and the text or speeeh. This world knowledge is
rooted in life experiences and enables individuals to make infer-
ences and form expectations about commonplace situations, ac-
cording to Long (1989). Steffensen and Colker (1982) note that
through membership in a culture, an individual has “privileged in-
formation” about the world that is “represented in a rich system of
schemata” (p. 2) (i.e., previously acquired knowledge structures).
According to Rumelhart (cited in Adams & Bruce, 1982), when the
cultural background of an author/speaker and the reader/listener
differ, the reader or listener may utilize inappropriate schemata. If
this occurs, communication or comprehension of the message con-
tained in the text fails. Several researchers have demonstrated the
strong and interactive effect of background knowledge on L1 infor-
mation processing of written text (Allen, Bernhardt, Berry, &
Demel, 1988; Aron, 1986; Hudson, 1982; Johnson, 1981, 1982; Lip-
son, 1983; Markham & Latham, 1987; Nelson, 1987).
Although the positive influence of prior knowledge on compre-
hension seems well documented in the L1 and L2 reading compre-
hension research literature, there has been a dearth of schema-theo-
retie L2 listening research. Long (1989) echoes Pearson and Field-
ing’s (1982) concern that little has been written about listening from
the schema-theoretic perspective, even though much of the (Ll) cog-
nitive research supporting schema theory has used listening as the
mode through which information has been transmitted to subjects.


Long supports Pearson and Fielding’s contention that schema-theo-
retic models of comprehension should apply to L2 listening as well
as reading because they are based on L1 listening research.
Expressing her dismay that applications of schema theory to L2
listening are so few in number, Long (1989) points out that between
1969 and 1989 only two empirical studies, other than her own, had
attempted to assess the effect of prior knowledge on L2 listening
comprehension, those of Mueller (1980) and Weissenreider (1987).
Mueller, who investigated the effects of visual context clues on the
listening comprehension of beginning students of German, found
that more language-proficient learners needed fewer visual contex-
tual cues than less language-proficient learners. The cues were
needed to help them “to activate appropriate scripts” that, in turn,
increased their level of listening comprehension. In an examination
of the interaction of cognitive strategies with background knowl-
edge, Weissenreider assessed the importance of textual schemata
(knowledge about the newscasting process) and content schemata
(knowledge about specific themes in the newscast) relative to the
listening comprehension of intermediate-level students of Spanish.
She found that their listening comprehension was enhanced when
they possessed the necessary textual and content schemata,
especially when they were able to apply effective listener strategies,
such as anticipating related themes (see Long, 1989). Smyth, Morris,
Levy, and Ellis (1987) note that in cognitive psychology, the com-
mon core of ideas underlying the concepts of schemata, frames, and
scripts is the assumption that
there are organized structures of knowledge in memory. These are
derived from past experience, and can be retrieved to guide
comprehension and recall. These schemata provide the framework for
the interpretation of the new experience. They specify what to look for
in the input, and they indicate what is likely to happen next, which
allows inferences to be drawn. (p. 188)
Providing listeners with a lecture containing information that
activates stored knowledge structures may advantage those listeners
who already possess the structures as a result of previous experi-
ence, and disadvantage those who have no experiential knowledge
to draw upon to support the processing of the incoming discourse.


The present study attempted to enhance what is known about the
influence of prior knowledge on L2 listening comprehension and to
extend this by examining the effect of prior knowledge, speech
modification, and listening comprehension proficiency on how well


EFL listeners understood lectures. Specifically, we attempted to
determine whether speech modification (supplying redundant and
elaborated information in the lecture), prior knowledge of the
lecture topic, and the level of listening comprehension in English
affected in positive, negative, or neutral ways the comprehension of
information presented orally to Chinese EFL listeners. We
anticipated that the investigation would shed additional light on the
processes involved in EFL listeners’ comprehension and retention
of academic discourse and would add to the growing data base of
information about the input variables that support (or reduce) L2
listeners’ comprehension of academic discourse, in general, and
EFL listeners’ comprehension, in particular.
The following main-effect research questions were posed
regarding the effects of speech modification, listening proficiency,
and prior knowledge on L2 lecture learning as measured by a
postlecture test of information recognition:
1. Would recognition by the information contained in the L2 lecture
be greater when the EFL students listen to a lecture containing
redundant information (modified speech) than when they listen
to the nonredundant (unmodified) lecture? (We hypothesized
that the answer to this question would be affirmative.)
2. Would the high-intermediate listening proficient (HILP) students
of EFL achieve higher scores on the postlecture information
recognition quiz than the low-intermediate listening proficient
(LILP) students of EFL? (We expected that they would.)
3. Would the scores on the postlecture recognition quiz be higher
when the intermediate-level EFL students are familiar with the
cultural content of the lectures than when they are not? (We an-
ticipated that the answer to this question would be affirmative.)
In order to assess the effect of the EFL students’ prior knowledge
on their comprehension of lecture information, it seemed worth-
while to include two types of recognition test items on the post-
lecture, information-recognition quiz: (a) passage-independent
items that would provide an indication of the listener’s prior knowl-
edge of the lecture topic and (b) passage-dependent items that
would probe the listeners’ understanding of the information pre-
sented in the lecture. As a result, the final research question posed
was the following:
4. Would the passage-independent item scores of the intermediate-
level EFL students be higher than their passage-dependent item
scores? (We did not formulate a directional answer to this ques-
tion a priori.)


The study involved 388 EFL subjects. They were male under-
graduate students at the Chinese Naval Academy in the Republic of
China. The subjects had studied English as a foreign language for at
least 6 years during high school and were categorized generally as
intermediate-level EFL students (Test of English as a Foreign
Language [TOEFL] scores between 400 and 500) in the Academy.
On the basis of their scores on the listening section of the
Comprehensive English Language Test (CELT) (Harris & Palmer,
1970), the range of which is 0-50, the subjects were divided into two
listening proficiency levels: HILP and LILP. Those who scored
between 20 and 35 on the CELT were assigned to the HILP level;
those who scored between 8 and 18 were assigned to the LILP
level. 1 Students at each proficiency level then were randomly
assigned to one of the four experimental conditions of lecture
presentation: (a) familiar-modified, (b) familiar-unmodified, (c)
unfamiliar-modified, and (d) unfamiliar-unmodified. There were
180 in each of the two proficiency levels and 45 in each of the four
experimental groups. Twenty-eight students who scored 19 on the
CELT were also randomly assigned to one of the four experimen-
tal groups, but their performance was excluded from the statistical
analysis in order to separate the HILP and LILP groups by at least
1 score point. In all, the performance of 360 subjects on the post-
lecture test of information recognition comprised the data set.

The unfamiliar-unmodified lecture. An unmodified version of the
650-word lecture on “The Amish People and the Pennsylvania
Dutch Country” was adapted from a reading passage in Duffy
(1986). The passage was used by Markham (1988) in a study on the
effect of a speaker’s gender and perceived level of expertness on
ESL students’ listening recall of a speaker’s message. It was assumed
1 Twenty eight subjects who scored 19 on the CELT test were randomly assigned to one of
the four experimental conditions because they could not be excluded from taking part in the
experiment conducted during a regularly scheduled class session. However, the data
accrued on these 28 subjects was not included in the ANOVR procedure. These were
omitted in order to separate the HILP from the LILP levels of listening proficiency by at
least 1 score point.


that the topic of the lecture would be unfamiliar to the Chinese
subjects. 2
The familiar-unmodified lecture. An unmodified adaptation of the
680-word lecture “Confucius and Confucianism,” adapted from
The New Encyclopedia Britannica (1985), constituted the familiar
lecture. The basic schematic knowledge contained in the lecture
was presumed, and later tested and found, to be familiar to the
Chinese subjects (see Footnote 2). See Table 1 for a comparison of
the structural features of the four lectures.

Composition of the Four Lectures

Confucius The Amish

(familiar) (unfamiliar)
Feature Unmodified Modified Unmodified Modified
No. of words 680 880 650 850
No. of T units 40 56 40 54
No. of S nodes 69 88 69 86

The modified lectures. The familiar and unfamiliar lectures were

modified for the study to include elaborated information in the
form of paraphrase and repetition of information. For example, in
the unmodified version of the lecture, students heard: “The food of
the Pennsylvania Dutch Country is very hearty and delicious.” In
the modified version, they heard: “The food of the Pennsylvania
Dutch Country is very hearty and delicious. Hearty and delicious
food is nourishing and tasty.” Fifteen instances of redundant
information were injected into each modified lecture. The decision
2 In a pilot study, 36 native English speakers found the lecture on “The Amish People and the
Pennsylvania Dutch Country” to be more familiar, t (35) = 5.23, p <.01, and easier to
understand, t (35) = 3.59, p < .01, than the lecture on “Confucius and Confucianism.”
Forty-four nonnative speakers of English from the Defense Language Center of the
Republic of China rated the lecture on “Confucius and Confucianism” to be significantly
more familiar, t (42) = 7.78, p <.01, and easier to comprehend, t (42) = 2.30, p <.05, than
the lecture on the Amish.


to provide redundancy of certain pieces of information contained in
the lectures was informed by the perceived need to elaborate (a)
information that proved difficult for students when tested in the
pilot study (i.e., information contained in items whose difficulty
indices were less than .80) and (b) information that was tested in the
passage-dependent items on the postlecture comprehension test.
The four lectures were recorded by a male native speaker of U.S.
English. The two unmodified lectures contained approximately the
same number of propositional ideas (i.e., one S node equaled one
propositional idea), and the same number of T units.3 (See Table 1.)

The Postlecture Comprehension Tests

To test the EFL students’ comprehension of lecture information
presented on the familiar and unfamiliar topics, two 30-item
multiple-choice tests were constructed, one for the modified and
unmodified versions of the familiar-topic lecture on “Confucius and
Confucianism,” and one for the modified and unmodified versions
of the unfamiliar-topic lecture, “The Amish People and the
Pennsylvania Dutch Country.” (There were 15 items in the passage-
independent subtest, and 15 items in the passage-dependent
subtest.) These tests were adapted from two original 30-item
multiple-choice tests pilot tested on Chinese EFL students in
January 1989 in Taiwan. The test data obtained in the pilot study
had been analyzed, and the earlier test of L2 information retention
was subsequently revised.4

The experiment was conducted in the language laboratory at the
Chinese Naval Academy in Taiwan. Before the experiment began,
3 A T unit is defined by Hunt (l965) as a single main clause plus whatever other subordinate
clauses or nonclauses are attached to, or embedded within, that main clause. For example,
the sentence “The man who persuaded John to be examined by the doctor was fired”
contains only one T unit and three S nodes: “The man was fired; “The man persuaded
John”; and “John was examined by the doctor.” An S node is defined here as a clause,
whether it is a main clause or a subordinate, or an embedded clause (e.g., “John was
examined by the doctor”). An S node may also be a simple sentence (e.g., “I like you”).
4 For the final form of the Amish lecture test, the average difficulty index of the passage-
independent items was .83; for the passage-dependent, it was .53. For the Confucius test,
the average difficulty index for the passage-independent items was .91; for the passage-
dependent items, it was .66. The Kuder-Richardson reliability indexes of the tests were the
following: “’Confucius and Confucianism,” K-R 20 = .89 “The Amish People and Pennsyl-
vania Dutch Country,” K-R 20 = .88.
It should be noted that, technically speaking, in this study, listening comprehension was
operationalized as lecture comprehension, and lecture information comprehension as
performance on the 15 passage-dependent test items since the 15 passage-independent
items test prior knowledge of the topic, for the most part.


subjects were asked to give their informed consent to participate in
a study. All agreed to participate.
Harris and Palmer’s (1970) CELT was administered 1 week prior
to presentation of the lecture. On the basis of the scores, the 390
subjects were randomly divided into HILP and LILP levels.5
Students of each level were then randomly assigned to one of the
four experimental groups.
The lecture and comprehension test were given 1 week after the
administration of the CELT. The postlecture quiz contained items
testing recognition of main ideas and details contained in the lecture
(via text-dependent items) and general information about the two
topics (via text-independent items). Each completion item
contained four response choices. Before the test, a short audiotaped,
warm-up lecture was provided. Then, subjects in each group
listened to one of the recorded lectures. After a 5-minute rest
period, the subjects took the postlecture test of information
recognition. During the rest period, the researcher distracted the
students from discussing the lectures by discussing his experiences
as a graduate student at The Pennsylvania State University.

Statistical Analysis
The data were analyzed by a 2 x 2 x 2 x (2) mixed analysis of
variance with repeated measures (ANOVR). The between-subjects
factors of speech modification (i.e., information redundancy), prior
knowledge, and listening proficiency combined with the within-
subjects factor of test type. There were two levels of speech
modification (redundant vs. nonredundant speech), two levels of
prior knowledge (familiar vs. unfamiliar topic), two levels of
listening proficiency (high-intermediate vs. low-intermediate
listening proficiency) and two levels of test type (passage-
independent vs. passage-dependent items). The analysis was
conducted using the ANOVR software package (Games et al.,

A synopsis of the findings is first presented and is followed by an
in-depth explanation of the results of the statistical analysis. A
5 In all 390 students took part in the study, including 2 non-Taiwanese nationals who were
excluded from the statistical analysis although they took part in the experiment. Of the 388
Taiwanese participants, 28 of these subjects who scored 19 on the CELT were excluded
from the ANOVA procedure (see Footnote 1); this left a data set on 360 subjects for the
ANOVA procedure.


significant interaction between speech modification (redundant
speech vs. nonredundant speech) and listening proficiency (HILP
vs. LILP) indicated that the HILP students benefited from speech
modification, but the LILP student did not benefit to a significant
degree from the speaker’s use of the heuristic. A significant
interaction between prior knowledge (familiar vs. unfamiliar topic)
and test type (passage-independent vs. passage-dependent items)
was also detected. For both the HILP and LILP subjects, prior
knowledge had a significant impact on subjects’ scores on the
passage-independent test items on the postlecture comprehension
test. Those EFL subjects who listened to the familiar-topic lecture
on Confucius had higher passage-independent than passage-
dependent scores. There was no difference in the performance on
the passage-independent and passage-dependent items for those
who listened to the lecture on the unfamiliar topic of the Amish.
However, the passage-independent performance of subjects who
listened to the familiar-topic lecture proved to be superior to that of
those who listened to the lecture on the unfamiliar topic. Subjects’
performance on passage-dependent items did not differ signifi-
cantly whether the familiar or unfamiliar topic was presented.

Results of the Statistical Analysis

The analysis uncovered significant main effects for the following
variables: prior knowledge, F (1,352) = 110.68, p < .001; speech
modification, F (1,352) = 12.38, p < .001; listening proficiency, F
(1,352) = 50.42, p < .001; and test type, F (1,352) = 231.30,
p <.001. Significant effects for the following interactions were
found: speech modification and listening proficiency, F
(1,352) = 8.78, p < .01; and prior knowledge and test type, F
(1,352) = 183.67, p <.001 (see Table 2 for the ANOVR results; see
Table 3 for the cell means).
Tukey’s WSD (wholly significant difference) procedure was used
in post hoc analyses of the two-way interactions (alpha set at .05).

Interaction of Speech Modification and Listening Proficiency

The results of the post hoc analysis on the interaction between
speech modification and listening proficiency indicated that HILP
students who listened to the speech containing redundant
information, M = 7.27, outperformed the LILP students who
listened to the modified academic discourse, M = 5.56. The HILP
students who listened to the nonredundant lecture, M = 6.17, also
outperformed LILP students who listened to the lecture without


ANOVR Summary of Postlecture Comprehension Test Scores
for Variables of Prior Knowledge, Speech Modification,
Listening Proficiency, and Test-Item Type

redundancy, M = 5.56. The results also showed that the HILP

students who listened to the modified speech, M = 7.27, performed
significantly better on the postlecture quiz than those HILP students
who listened to the unmodified speech, M = 6.17. However, the
difference in the performance of the LILP students in the modified .
and unmodified lecture conditions did not differ to a significant
degree, M = 5.56 versus M = 5.46. (See Table 4, and Figure 1.) Note
that the maximum score for the passage-independent as well as the
passage-dependent items was 15.

Interaction of Prior Knowledge and Test Type

The statistics for the interaction between prior knowledge and
test type are summarized in Table 5 and graphed in Figure 2.
Subjects who listened to the familiar-topic lecture achieved
significantly higher scores on the passage-independent items,
M = 8.84, than on the passage-dependent items, M = 5.18. Subjects
who listened to the unfamiliar-topic lectures (whether it was
modified or unmodified), did not differ significantly in respect to
the passage-independent items, M = 5.32, and passage-dependent


Cell Means for the Passage-Independent Score and Passage-Dependent Score of the Postlecture
Comprehension Test
Means of the Postlecture Comprehension Test Score
for the Interaction Between Modification
and Listening Proficiency

Means of the Postlecture Comprehension Score
for Interaction Between Modification
and Listening Proficiency


items (M = 5.11) in the postlecture comprehension test. The post
hoc analyses, however, did disclose a significant difference in the
passage-independent item scores between those listening to the
familiar-topic lectures (M = 8.84) and those hearing the unfamiliar-
topic lectures (M = 5.32), (See Table 5 and Figure 2).

Interactive Effect of Speech Modification and L2 Listening
Proficiency on L2 Lecture Comprehension
The HILP students of EFL benefited from speech modification
to a significantly greater extent than the LILP students did. In the
present study, speech modification only involved repetition of the
lecture information tested in the passage-dependent items of the
postlecture comprehension test. However, it was noted that the
HILP students who listened to the modified lecture achieved higher
aggregate scores than the HILP students who listened to the
unmodified lecture. This was not the case for the LILP students.
The comprehension of the lecture information was significantly
improved among the HILP students as a result of the discourse
redundancies. It may be that repetition of even selected pieces of
information by the lecturer allows students more time to process the
information so that, among other things, the relationship of
syntactic forms and the meaning of the lexical items in the lecture
become clearer, as has been suggested by Hatch to be a potential
benefit of speech modification (cited in Chaudron, 1983b). The
results of the present study also revealed that the HILP subjects
scored higher than the LILP subjects on the postlecture comprehen-
sion test whether they listened to the modified or unmodified lec-
ture. It appears that a “language competence ceiling” (Clarke, 1979,
p. 138) prevented the LILP subjects from performing as well as the
HILP subjects, even when redundant information was supplied.
The empirical findings suggest one possible answer to the
question Chaudron (1988) has raised, “If input to learners must be
comprehensible, what factors make teacher talk in classrooms [or
lecturer talk in the lecture hall] appropriate for L2 learners’
differing proficiency levels?” (p. 8). For higher levels of
intermediate EFL listening proficiency (i.e., CELT scores greater
than 19), the provision of redundant information in extended
discourse appears to augment listeners’ comprehension of the
academic discourse, whereas it does not appear to affect as
supportively the comprehension of EFL listeners of lower levels of
listening proficiency (i.e., CELT scores less than 19). It may be that


Means of the Postlecture Comprehension Test Scores
for the Interaction Between Prior
Knowledge and Test Type

Means of the Postlecture Comprehension Score
for the Interaction Between Prior
Knowledge and Test Type


additional modifications of the input are needed to catalyze greater
understanding of the extended lecture material by the LILP
students, modifications such as those outlined by Hatch (1983),
Chaudron (1986), and Parker and Chaudron (1987), including use of
the following: (a) syntactic simplification (e.g., use of short mean
length of utterances, simple propositional syntax, fewer words per
T unit, fewer clauses per T unit, fewer S nodes per T unit, omission/
deletion of sentence elements, omission of inflections); (b) clearer
articulation (i. e., use of fewer reduced vowels and fewer
contractions); (c) shorter utterances (fewer words per utterance,
fewer words per T unit, fewer T units per utterance); (d) slower
speech; (e) emphatic stress; (f) topic-left dislocation; and (g)
rhetorical signaling or framing, or optional syntactic markers.
Perhaps inclusion of more macromarkers in the discourse (e.g.,
What I’m going to talk about today is . . .) would have helped
enhance the comprehension of the LILP, as well as the HILP
listeners, as Chaudron and Richards (1986) have suggested.
It may be that the LILP students in the present study lacked
adequate command of the lexicon of the lecture to a greater extent
than did the HILP students, and this lack of vocabulary could have
vitiated the positive effect of the speech elaboration for these
students. Use of less complex lexis (i.e., use of higher frequency
vocabulary, and smaller type-token ratio) might have enhanced the
effects of redundancy/elaboration on the comprehension of the
lower proficiency level listeners. Further research is needed to
confirm or disavow this supposition.
The results suggest that different proficiency levels need
different types of modified or simplified speech. As Chaudron
(1983b) notes, “the definition of what is simplified speech can be
quite different for different learners therefore, no single form of
simplification would be an appropriate method of presentation for
a group of learners representing a range of proficiency levels”
(p. 451), Additional research is needed to pinpoint the types of
modifications and simplifications that aid the comprehension of
extended discourse by different levels of EFL proficiency.

Implication of the Interaction between Speech Modification

and Listening Proficiency for EFL Pedagogy
The results revealed that level of listening proficiency played an
important role in the comprehension of the L2 lecture. In this study,
both HILP and LILP EFL students benefited from their prior
knowledge, but only HILP students benefited from speech
modification. This finding seems to support Ur’s (1984) contention


that second language students comprehend and learn best if their
level of listening ability is taken into consideration when planning
listening materials. Ur (1984) suggests that L2 students learn best
“from listening to speech which, while not entirely authentic, is an
approximation to the real thing, and is planned to take into account
the learners’ level of ability and particular difficulties” (p. 23).
Accurate assessment of L2 students’ levels of listening proficiency is
sorely needed in EFL programs to determine what types of speech
modification as well as what kind of content should be included in
the materials designed for learners of different proficiency levels. If
we are trying to improve the ability of EFL students to understand
and remember extended discourse presented in English, it is crucial
that we understand what types of materials should be created and
used for EFL learners at the beginning, intermediate, and advanced
levels. In other words, two basic questions about L2 listening
pedagogical practice and materials development need to be
addressed: (a) what types of speech modifications, if any, should be
made in the stimulus lecture materials? and (b) what types of
content should be included in the lectures? This study provided
only partial answers to these questions. For higher levels of listening
proficiency (CELT scores greater than 19), the lectures should
contain redundant information and should be on topics familiar to
the EFL listeners. We do not know exactly what the lower levels of
listening proficiency need in terms of extended-lecture discourse.
We do not know, indeed, whether low-level listeners should even be
exposed to extended-lecture-type materials. Continued investiga-
tion of those elements that enhance beginning-, intermediate-, and
advanced-level EFL learners’ listening comprehension is called for.
We attempt to address the second question, What types of content
should be included in the lectures? in the following section.

The Interaction of Prior Knowledge and Test Type on

the Comprehension and Assessment of L2 Lecture Learning
The main effect for prior knowledge on subjects’ performance on
the postlecture comprehension test indicated that Chinese EFL
listeners scored higher in their postlecture comprehension test when
they listened to the familiar-topic lecture than when they listened to
the unfamiliar-topic lecture. However, when the within-subjects
variable of test type was considered in the subjects’ performance on
the postlecture comprehension test, the main effect was nullified
because a significant interaction between prior knowledge and test
type was detected. The results indicated that the significant effect
of prior knowledge only appeared on subjects’ performance on the


passage-independent items of the postlecture comprehension test.
There was no significant effect of prior knowledge on the subjects’
performance on the passage-dependent items of the postlecture
comprehension test. The post hoc analysis of the interaction
between prior knowledge and test type indicated that those EFL
subjects who listened to the familiar-topic lecture scored higher on
the passage-independent items than on the passage-dependent
items and also scored higher than those who listened to the unfamil-
iar-topic lecture on the passage-independent items. The findings
support Johnston’s (1983) contention that passage-independent
items provide a measure of a listener’s or reader’s prior knowledge
(Johnston, 1983).
Long (1989) highlights the need to investigate how schemata
affect auditory comprehension in a second language. In this study,
prior knowledge was the information subjects possessed as a result
of membership in the Chinese culture. These schemata facilitated
the listeners’ processing of information received via text or oral
discourse. The fact that the Chinese subjects performed better on
the familiar-topic lecture on Confucius than on the unfamiliar-topic
lecture on the Amish may be attributed to their being able to
activate schemata about Confucius and his teachings. Anderson
(1984, p. 248) explains the role of schema in comprehension as
1. A schema provides ideational scaffolding for assimilating text
2. A schema facilitates selective allocation of attention.
3. A schema enables inferential elaboration.
4. A schema allows orderly searches in memory.
5. A schema facilitates editing and summarizing.
6. A schema permits inferential reconstruction.
Thus, if a listener has a schema of the lecture content, as had the
Chinese listeners who listened to the lecture on Confucius, the
listener will be able to process the information and achieve better
comprehension of the lecture. In other words, the more prior
knowledge the listener has about the topic of the lecture, the easier
it is for that listener to comprehend the lecture and retain general
points of information. The present study, then, lends some
additional support to the notion that schemata play a vital role in
auditory comprehension (the EFL subjects who listened to the
familiar-topic lecture scored higher on the passage-independent
items than on the passage-dependent items and also scored higher
than those who listened to the unfamiliar-topic lecture on the


passage-independent items). However, the effect will be most
evident when students are tested on passage-independent

Implication of the Interaction between Prior Knowledge and Test

Type for Testing/Teaching EFL Listening Comprehension
The significant interaction between prior knowledge and test
type suggests that the EFL students’ listening comprehension will
be higher if they are tested on the information presented in a
familiar-topic lecture. It also suggests that it may be inappropriate
to judge the listening comprehension ability of nonnative speakers
of English on the sole basis of the results of standardized listening
tests, such as the TOEFL. One reason for this is that some EFL/ESL
students may not be familiar with the topic of the minilectures to
which they listen and on which they are tested. The findings also
suggest that more accurate assessment of EFL listeners’ listening
proficiency may result if the comprehension test that follows
presentation of a segment of extended discourse is divided into
items testing passage-independent and passage-dependent
information. Such a division would identify the amount of prior
knowledge the listener possessed about the topic and the amount of
new information acquired from the lecture. It would also yield a
better profile of the listening abilities and background knowledge
of the test taker.
Johnston (1983) suggests that subjects’ lecture comprehension
should be tested by incorporating both passage-independent and
passage-dependent items into a test because without some measure
of what background knowledge readers or listeners have, it is
difficult to estimate their integration of new information with the
old. In addition, as Johnston (1983) notes, background knowledge
enhances the listener’s ability to make inferences during information
processing. Thus, to eliminate all passage-independent items would
leave only what have been called low-level questions on the test. On
the other hand, to eliminate all the passage-dependent items in a test
would leave only the effect of cultural background knowledge
without testing listening or reading comprehension. Hence, a test
containing both passage-independent and passage-dependent items
may be a better assessment of listening or reading comprehension
than one that contains one or the other.
In sum, the findings of the present study have several pedagogical
implications, particularly for those teaching and learning in Chinese
EFL settings:


1. The fact that LILP students did not benefit from speech
modification (i. e., the redundancy of lecture information)
indicates that they may need types of speech modifications other
than elaborations and redundancies alone or several types of
speech modification combined. They may need, for example, to
listen to minilectures containing more pauses or syntactic
simplification (cf. Parker & Chaudron, 1987). Further research is
necessary to pinpoint the appropriate types of modifications
needed for learners with lower levels of listening proficiency.
2. Comprehension of a lengthy English lecture (more than 9
minutes) may be beyond LILP students’ abilities. They may
need shorter lectures as well as more familiar-topic lectures. If
they are required to listen to extended English lectures,
providing them with a written version of the lecture before they
listen may be helpful for them.
3. Redundancy of information can be used effectively to help
enhance the HILP students’ comprehension of lecture informa-
tion, and it is possible that redundancy of information is also an
effective method of speech modification for more advanced
EFL learners in college. This supposition remains to be tested in
future investigations, however.
4. In this study, subjects’ passage-independent item scores were
higher than their passage-dependent item scores in the familiar-
topic lecture but not in unfamiliar-topic lecture. The passage-
independent item score thus proved to be a very good indicator
of the degree of prior knowledge that a student possesses.
Listening tests for EFL students should contain items that assess
the listeners’ degree of familiarity with the topic if accurate
assessments of listening skills are a major goal.


Several limitations of the study should be noted:
1. The subjects in the study were male EFL students studying at the
Chinese Naval Academy in Taiwan. Thus, caution should be
exercised in relating the findings of the study to other groups of
EFL or ESL listeners. Since the subjects were a homogeneous
and select sample of Chinese EFL students (naval cadets), the
findings may not generalize to other groups of Chinese-speaking
EFL students (e.g., those from nonmilitary institutions of higher
learning). Replication of the study with different groups of EFL
students as well as with ESL students is called for.


2. The subjects in the study were not majoring in English. Their
scores on the CELT test were fairly low, with the mean score
being only 39% and the median score being only 38%. It may be
that the CELT proved to be too difficult for the intermediate-
level EFL students. Other listening tests (e.g., the listening
section of the TOEFL) might have given a more accurate
measure of the subjects’ listening abilities. Furthermore, the
students’ attitude toward English, their aptitude for learning
language and their motivation to participate in the experiment
were not assessed. Such factors could have influenced, possibly
confounded, the results.
3. Only one type of speech modification (i.e., elaboration of
information in the form of repetition of constituents, paraphrase,
and use of synonyms) was considered, so the results of the study
cannot be generalized to include the wide range of teachers’
speech modifications (e. g., syntactic simplification, slower rate
of speech). In addition, modifications of lecturer-student
interaction (e.g., clarification requests, comprehension checks, or
confirmation checks) were not studied. Research on the effects
of lecturer-student interactions on lecture comprehension is
sorely needed.
4. In the study, subjects were not allowed to take any notes during
presentation of the lecture. However, in the real-life lecture
situation, students customarily take notes during a lecture and
review the notes before taking a test (cf. Dunkel, 1988b). Perhaps
if note-taking had been permitted, the results of the study might
have been different. Speech modification might also have
benefited the LILP students if note-taking had been allowed.
Additional research is needed to test this possibility.
The relative effects of prior knowledge, speech modification, and
listening proficiency on EFL listeners’ comprehension of English
lectures are not clear. This study uncovered only a fraction of the
issues involved in determining the effect of a lecturer’s speech
modification and some of the issues involved in testing EFL
listeners’ listening proficiency in English. Additional experimental
studies of the processes involved in the cognitive processing of
foreign language material are sorely needed to help illuminate the
ways in which EFL students learn to comprehend and retain
information contained in lectures given in English, and to help
materials writers create more appropriate listening comprehension
materials for EFL learners.


The authors would like to thank Professors Dennis Gouran, Frank De Vesta, and
Gerard Hauser of The Pennsylvania State University for their scholarly
contributions to the formulation of the study and the reporting of the results. We
would also like to thank the two anonymous TESOL Quarterly reviewers, whose
suggestions led to revision and strengthening of the research report. Finally, we
would like to acknowledge the critical role that Sandra Silberstein and the editorial
staff of the TESOL Quarterly played in helping bring the paper into its final form.
Their meticulous editing of the paper is appreciated and admired.

Chung Shing Chiang is Associate Professor of Speech Communication at the World
Journalism and Communication College, Taiwan. His research interests focus on
listening comprehension and teacher talk.

Patricia Dunkel is Associate Professor of Speech Communication and Director of

the Center for English as a Second Language at The Pennsylvania State University.
She has published scholarly articles and instructional texts on listening
comprehension and is currently conducting research on computer-adaptive testing
of L2 listening comprehension proficiency.

Adams, M., & Bruce, B. (1982). Background knowledge and reading
comprehension. In J. A. Langer & M. Trika-Smith (Eds.), Reader meets
author: Bridging the gap (pp. 2-25). Newark, DE: International Reading
Allen, E. D., Bernhardt, E. B., Berry, M. T., & Demel, M. (1988).
Comprehension and text genre: An analysis of secondary school foreign
language readers, The Modern Language Journal, 72, 163-172.
Anderson, B. V., & Barnitz, J. G. (1984). Cross-cultural schemata and
reading comprehension instruction. Journal of Reading, 28, 102-108.
Anderson, R. A. (1984). Role of reader’s schema in comprehension,
learning and memory. In R. A. Anderson (Ed.), Learning to read in
American schools (pp. 243-272). Hillsdale, NJ: Lawrence Erlbaum.
Anderson, R. C., Spiro, R. J., & Anderson, M. C. (1978). Schemata as
scaffolding for the representation of information in connected discourse.
American Educational Research Journal, 15, 443-440.
Aron, H. (1986). The influence of background knowledge on memory for
reading passages by native and nonnative readers. TESOL Quarterly,
20 (1), 136-140.
Barnitz, J. (1986). Toward understanding the effects of cross-cultural
schemata and discourse structure on second language reading
comprehension. ]ournal of Reading Behavior, 18, 95-116.
Bartlett, F. C. (1932). Remembering. Cambridge: Cambridge University


B1au, E. K. (1982). The effect of syntax on readability for ESL students in
Puerto Rico. TESOL Quarterly, 16 (3), 517-528.
Blau, E. K. (1990). The effect of syntax, speed, and pauses on listening
comprehension. TESOL Quarterly, 24 (4), 746-753.
Brown, J., & Hilferty, A. (1986). Listening for reduced forms. TESOL
Quarterly, 20 (4), 759-763.
Brown, R. (1977). Introduction. In C. Snow & C. Ferguson (Eds.),
Talking to children (pp. 1-3). New York: Cambridge University Press.
Carrell, P. L. (1983). Three components of background knowledge in
reading comprehension. Language Learning, 33, 183-207.
Carrell, P. L. (1984a). Inferencing in ESL: Presuppositions and implica-
tions of factive and implicative predicates. Language Learning, 34, 1-21.
Carrell, P. L. (1984b). Evidence of a formal schema in second language
comprehension. Language Learning, 34, 87-112.
Carrell, P. L. (1984c). The effects of rhetorical organization on ESL
readers. TESOL Quarterly, 18 (3), 441-469.
Carrell, P. L. (1984d). Schema theory and ESL reading: Classroom
implications and applications. The Modern Language Journal, 68,
Carrell, P. L. (1987). Content and formal schema in ESL reading. TESOL
Quarterly, 21 (3), 461-481.
Carrell, P. L., & Eisterhold, J. C. (1983). Schema theory and ESL reading
pedagogy. TESOL Quarterly, 17 (3), 553-573.
Chaudron, C. (1983a). Foreigner talk in the classroom—An aid to
learning? In H. W. Seliger & M. H. Long (Eds.), Classroom oriented
research in second language acquisition (pp. 127-145). Rowley, MA:
Newbury House.
Chaudron, C. (1983b). Simplification of input: Topic reinstatements and
their effects on L2 learners’ recognition and recall. TESOL Quarterly,
17 (3), 437-458.
Chaudron, C. (1986). The role of simplified input in classroom language.
In G. Kasper (Ed.), Learning, teaching and communication in the
foreign language classroom (pp. 99-110). Aarhus, Denmark: Aarhus
University Press.
Chaudron, C. (1988). Second language classrooms: Research on teaching
and learning. New York: Cambridge University Press.
Chaudron, C., & Richards, J. (1986). The effect of discourse markers on
the comprehension of lectures. Applied Linguistics, 7, 113-127.
Clarke, M. A. (1979). Reading in Spanish and English: Evidence from
adult ESL students. Language Learning, 29, 121-150.
Clyne, M. (1981). Foreigner talk. International Journal of the Sociology of
Language, 28, 1-29.
Connor, U. (1984). Recall of text: Differences between first and second
language readers. TESOL Quarterly, 18 (2), 239-256.
DeCarrico, J., & Nattinger, J. (1988). Lexical phrases for the comprehen-
sion of academic lectures. English for Specific Purposes Journal, 7,


Duffy, P. (1986). Variations: Reading skills/oral communication.
Englewood Cliffs, NJ Prentice Hall.
Dunkel, P. (1985). The immediate recall of English lecture information by
native and non-native speakers of English as a function of notetaking.
Unpublished doctoral dissertation. University of Arizona, Tucson.
Dunkel, P. (1988a). The effect of delivery modification on notetaking and
comprehension of L2 lectures. Journal of Intensive English Programs, 2,
Dunkel, P. (1988b). The content of L1 and L2 students’ lecture notes and
their relation to test performance. TESOL Quarterly, 22 (2), 259-281.
Dunkel, P., & Davey, S. (1989). The heuristic of lecture notetaking:
Perceptions of American and international students regarding the value
and practice of notetaking. English for Special Purposes Journal, 8,
Dunkel, P., Mishra, S., & Berliner, D. (1989). Effects of note taking,
memory, and language proficiency on lecture learning for native and
nonnative speakers of English. TESOL Quarterly, 23 (3), 543-549.
Faerch, C., & Kasper, G. (1986). The role of comprehension in second
language learning. Applied Linguistics, 5, 257-274.
Freed, B. F. (1980). Talking to foreigners versus talking to children:
Similarities and differences. In R. C. Scarcella & S. D. Krashen (Eds.),
Research in second language acquisition (pp. 19-27). Rowley, MA:
Newbury House.
Games, P. A., Gray, S. G., Herron, W. L., Pentz, A., Sinatra, R., &
Pitz, G. F. (1985). Analysis of variance with repeated measures, main
program [computer program]. Computation Center, The Pennsylvania
State University.
Harris, D. P., & Palmer, L. A. (1970). A Comprehensive English Language
Test for Speakers of English as a Second Language (CELT). New York:
Hatch, E. (1983). Simplified input and second language acquisition. In
R. Andersen (Ed.), Pidginization, creolization as language acquisition
(pp. 64-86). Rowley, MA: Newbury House.
Henzl, V. M. (1973), Linguistic register of foreign language instruction.
Language Learning, 23, 207-222.
Henzl, V. M. (1975). Foreigner talk in the classroom. International Review
of Applied Linguistics, 17, 159-167.
Honeyfield, J. (1977). Simplification. TESOL Quarterly, 11 (4), 431-440.
Hudson, T. (1982). The effects of induced schemata on the “short circuit”
in L2 reading Non-decoding factors in L2 reading performance.
Language Learning, 32, 1-31.
Hunt, K. (1964). Grammatical structures written at three grade levels
(NCTE Research Rep. No. 3). Champaign, IL: National Council of
Teachers of English. (ERIC Document Reproduction Service No. ED
113 735)
Jenkins, S. M. (1987). The effect of language proficiency and prior
knowledge on the reading comprehension of graduate student native
and non-native speakers of English. Unpublished doctoral dissertation.
The Pennsylvania State University, University Park.


Johnson, P. (1981). Effects on reading comprehension of language
complexity and cultural background of a text. TESOL Quarterly, 15 (2),
Johnson, P. (1982). Effects on reading comprehension of building
background knowledge. TESOL Quarterly, 16 (2), 503-516.
Johnston, P. H. (1983). Reading assessment: A cognitive basis. Newark,
DE: International Reading Association.
Kelch, K. (1985). Modified input as an aid to comprehension. Studies in
Second Language Acquisition, 7, 81-89.
Kintsch, W., & Greene, E. (1978). The role of culture-specific schemata in
the comprehension and recall of stories. Discourse Processes, 1, 1-13.
Klein, W. (1986). Second language acquisition. Cambridge: Cambridge
University Press.
Krashen, S. D. (1980a). The theoretical and practical relevance of simple
codes in second language acquisition. In R. Scarcella & S. Krashen
(Eds.), Research in second language acquisition (pp. 7-18), Rowley,
MA: Newbury House.
Krashen, S. D. (1980b). The input hypothesis: Issues and implications. In
J. E. Alatis (Ed.), Current issues in bilingual education (pp. 168-180).
Washington, DC: Georgetown University Press.
Lipson, M. Y. (1983). The influence of religious affiliation on children’s
memory for text information. Reading Research Quarterly, 18, 448-457.
Long, D. (1989). Second language listening comprehension: A
schema-theoretic perspective. Modern Language Journal, 73, 32-41.
Long, M. (1983). Linguistic and conversational adjustments to non-native
speakers. Studies in Second Language Acquisition, 5, 177-193.
Markham, P. L. (1988). Gender and the perceived expertness of the
speaker as factors in ESL listening recall. TESOL Quarterly, 22 (2),
Markham, P. L., & Latham, M. (1987). The influence of religion-specific
background knowledge on the listening comprehension of adult
second-language students. Language Learning, 37, 157-170.
Mueller, G. (1980). Visual contextual cues and listening comprehension:
An experiment. Modern Language Journal, 64, 335-340.
Nelson, G. L. (1987). Culture’s role in reading comprehension: A schema
theoretical approach. Journal of Reading, 30, 4 2 4 - 4 2 9 .
The New Encyclopedia Britannica. (1985). Chicago: Encyclopedia
Olsen, L., & Huckin, T. (1990). Point-driven understanding in engineering
lecture comprehension. English for Specific Purposes, 9, 22-47.
Oxford, R. (1990). Language learning strategies: What every teacher
should know. New York: Newbury House.
Padron, Y., Knight, S., & Waxman, H. (1986). Analyzing bilingual and
monolingual students’ perceptions of their reading strategies. The
Reading Teacher, 39, 430-433.
Parker, K., & Chaudron, C. (1987). The effects of linguistic simplification
and elaborative modifications on L2 comprehension. University of
Hawai’i Working Papers in ESL, 6, 107-133.


Pearson, D. P., & Fielding, L. (1982). Research update: Listening
comprehension. Language Arts, 59, 617-629.
Pica, T., Doughty, C., & Young, R. (1986). Making input comprehensible:
Do interfactional modifications help? I. T. L. Review of Applied
Linguistics, 72, 1-25.
Pica, T., Young, R., & Doughty, C. (1987). The impact of interaction on
comprehension. TESOL Quarterly, 21 (4), 737-758.
Rumelhart, D. E. (1980). Schemata: The building blocks of cognition. In
R. J. Spiro, B. Bruce, & W. Brewer (Eds.), Theoretical issues in reading
comprehension (pp. 35-58). Hillsdale, NJ: Lawrence Erlbaum.
Rumelhart, D. E., & Ortony, A. (1977). The representation in memory. In
R. C. Anderson, R. J. Spiro, & W. E. Montague (Eds.), Schooling and
the acquisition of knowledge (pp. 99-135). Hillsdale, NJ: Lawrence
Schallert, D. (1982). The significance of knowledge: A synthesis of
research related to schema theory. In W. Otto & S. White (Eds.),
Reading expository material (pp. 13-48). New York: Academic Press.
Schinke-Llano, L. (1983). Foreigner talk in content classrooms. In
H. W. Seliger & M. H. Long (Eds.), Classroom oriented research in
second language acquisition (pp. 146-164). Rowley, MA: Newbury
Schinke-Llano, L. (1986). Foreigner talk in joint cognitive activities. In
R. Day (Ed.), Talking to learn: Conversation in second language
acquisition (pp. 99-117). Rowley, MA: Newbury House.
Smyth, M., Morris, P., Levey, P., & Ellis, A. (1987). Cognition in action.
Hillsdale, NJ: Lawrence Erlbaum Associates.
Snow, C., & Ferguson, C. (1977). Talking to children: Language input and
acquisition. Cambridge, MA: Cambridge University Press.
Snow, C. E., & Hoefnagel-Hohle, M. (1982). School-aged second
language learners’ access to simplified linguistic input. Language
Learning, 32, 411-430.
Speidel, G. E., Roland, G. T., & Kobayashi, L. (1985). Is there a
comprehension problem for children who speak nonstandard English?
A study of children with Hawaiian-English background. Applied
Psycholinguistics, 6, 83-96.
Steffensen, M. S., & Colker, L. (1982). The effect of cultural knowledge
on memory and language (Tech. Rep. No. 248), Urbana, IL: University
of Illinois, Center for the Study of Reading. (ERIC Document
Reproduction Service No. ED 217 405)
Swaffar, J. K. (1988). Reader, texts, and second languages: The interactive
processes. The Modern Language Journal, 72, 123-149.
Ur, P. (1984). Teaching listening comprehension. New York: Cambridge
University Press.
Weissenreider, M. (1987). Listening to the news in Spanish. Modern
Language Journal, 71, 18-27.
Wenden, A., & Rubin, J. (Eds.). (1987). Learner strategies in language
learning. Englewood Cliffs, NJ: Prentice Hall.


Wesche, M. B., & Ready, D. (1985). Foreigner talk in the university
classroom. In S. M. Gass & C. G. Madden (Eds.), Input in second
language acquisition (pp. 89-114). Rowley, MA: Newbury House.
Wong Fillmore, L. (1982). Instructional language as linguistic input:
Second language learning in classrooms. In L. C. Wilkinson (Ed.),
Communicating in the classroom (pp. 283-296). New York: Academic
Wong Fillmore, L. (1985). When does teacher talk work as input? In
S. Gass & C. Madden (Eds.), Input in second language acquisition
(pp. 17-50). Rowley, MA: Newbury House.


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.


University of Washington

The Cornell Lectures: Women in the Linguistics Profession.

Alice Davison and Penelope Eckert (Eds.). Washington, DC: The
Committee on the Status of Women in Linguistics of the Linguistics
Society of America, 1990. Pp. xi + 255. (Available at no cost from:
Linguistic Society of America Secretariat, 1325 18th St. NW, Suite
211, Washington, DC 20036-6501)

❑ In June 1989, a two-day conference on women in linguistics was

held at Cornell University. This volume, a collection of the papers
given at that conference, transmits the energy, openness, struggles,
yet hope of the conference and its participants. The contributors,
both female and male, have a variety of research interests and
positions in academe. The papers focus on the problems and com-
plexities faced by women and others who do not fit the model of
the successful (male) academic because of such factors as socializa-
tion, cultural expectations, and personality. Also included are con-
crete suggestions for making changes and improving the education-
al setting for all persons equally, regardless of gender.
The volume is divided into five parts with descriptive titles:
“Introduction,” “Women in Linguistics: the Numbers,” "Problems
and their Sources," “Finding and Giving Support,” and “Places in a
Woman’s Career.” There is a useful annotated bibliography of
feminist linguistics included in the introductory article by Sally
McConnell-Ginet as well as helpful references at the end of each
article. Each reader will be struck by different sections of the
collection and by different details and analyses. It is unimportant
whether or not one views oneself as a linguist, for the papers speak
to anyone in higher education—junior or senior faculty members,
adjunct or part-time ESL faculty members, administrators, or
graduate students.
Barriers to success in academia range from overt discrimination
such as sexual harassment to subtle discrimination such as lack of

informal contact with and access to male networks. This volume
presents a myriad of examples of discrimination, many pervasive
and unconscious; yet there is a commonality in experiences that
women in academe have had, and this commonality runs through all
the articles.
For example, two articles share a common theme—that interpre-
tation of events and data are often gender based. Molly Hite, in her
article “Sexual Harassment and the University Community” con-
tends that males and females tend to view sexual harassment differ-
ently. Men tend to see more gray areas and tend to be unable to
grasp the concept of sexual harassment and unable to identify with
the victim, generally a woman in a lower position than the male who
sought sexual favors. Hite sees the differing versions of incidents of
sexual harassment as tending “to stratify that entity we call the uni-
versity community—first and most obviously along the lines of gen-
der, and only second along the lines of institutional status” (p. 160).
Another interpretation and evaluation that is gender related is the
way women’s and men’s work is evaluated differently. Virginia
Valian in “Success and Failure: Expectations and Attributions”
contends that women are less successful in terms of salary, rank, and
time needed to get tenure for two reasons: Women’s performance is
evaluated more negatively than men’s, and women’s successes tend
to be attributed to luck and hard work, whereas men’s successes
tend to be attributed to ability and effort. “To oversimplify a bit:
We see men as deserving their successes, and women as deserving
their failures” (p. 132).
The theme of gender-based communicative styles and language
use is another theme found in this volume and commonly found in
popular books and media as well as in feminist literature. Linguists
Robin Lakoff (1975, 1990) and Deborah Tannen (1986, 1990) have
done much to explain and popularize this issue. In addition, sex
discrimination in academic settings has recently gained wide media
coverage with the cases of Jenny Harrison, a mathematician denied
tenure at the University of California, Berkeley, who has filed suit
charging sex bias, and Francis Conley, the top female neurosurgeon
at Stanford Medical School, who resigned abruptly, citing rampant
sexism in the School of Medicine.
However, the message from this volume is not one of despair but
one of cautious hope and movement toward change of institutional
settings. Even Hite’s article on sexual harassment shows guarded
optimism as she points out that more and more women are now on
university campuses as administrators, professors, and graduate and
undergraduate students. Concrete suggestions for change are pre-
sented in each paper. Examples given in some detail are (a) setting


up support and networking systems for women, (b) setting up men-
toring programs for junior female faculty and students, (c) men and
women talking together about such issues as gender-based interpre-
tation of events and data, and (d) women not compromising their
feminist vision as they achieve success, reaching “a balance between
caution and confidence in the rightness of one’s stand” (p. 235).
This last point seems especially important for those of us who
have achieved some stability in academe. We must be ever mindful
of our responsibility to help improve the educational setting for all
by fostering equality for all. We have this responsibility to others in
the profession, to students whether they be graduate or undergrad-
uate, and to native and nonnative English speakers alike.
In reading this engrossing and timely volume, I repeatedly asked
myself, How does the status of women in TESOL and in higher
education specifically compare to the status of women in the
linguistics profession? Is TESOL in higher education as Pennycook
(1989) suggests “a gendered division of the workforce, a hierarchi-
cally organized division between male conceptualizers and female
practitioners” (pp. 610-611)? TESOL professionals need to discuss
these issues openly, and the recent survey in TESOL Matters (April/
May 1991) should provide data for such discussions.
As Amy Sheldon stated in introductory remarks to the conference
on women held at Cornell, “We need to make discussions of
women . . . an acceptable part of our daily public discourse, if we
hope to figure out how to make our work places and work lives
more sensitive to women’s needs” (p. 40). The Cornell Lectures is
important reading for those of us who want to make such public
discourse common in the TESOL profession.

Lakoff, R. (1975). Language and women’s place. New York: Harper &
Lakoff, R. (1990). Talking power: The politics of language in our lives.
New York: Basic Books.
Pennycook, A. (1989). “The concept of method, interested knowledge,
and the politics of language teaching.” TESOL Quarterly, 23 (4), 589-618.
Tannen, D. (1986). That’s not what I meant! Conversational style makes or
breaks your relationships with others. New York: Morrow.
Tannen, D. (1990). You just don’t understand: Women and men in
conversation. New York: Morrow.


University of San Francisco

Building Better English Language Programs: Perspectives on
Evaluation in ESL.
Martha C. Pennington (Ed.). Washington, DC: NAFSA: Association
of International Educators, 1991. Pp. xi + 259. (Available from
Publications Order Desk, NAFSA, 1860 19th St. NW, Washington,
DC 20009)

This refreshing collection of articles edited by Martha C. Ben-

nington should interest and be of use to evaluators, program
administrators, and other ESL professionals.
The book is a welcome addition to my bookshelf, and is highly
recommended for several reasons. First, it is the first ESL
evaluation book. While there are many other works that touch on
evaluation areas with reference to ESL programs, this book brings
together twelve articles on different topics directly related to
evaluation. Second, the book addresses specific aspects of
administrative concern which are common to ESL professionals at
all levels. The articles in the book include extensive appendices,
which provide models for instrumentation and sources of
information which can be adapted and used immediately. Finally,
this volume sets an example for future collections of articles on ESL
program evaluation.
The book provides general theoretical perspectives from authors
concerned with evaluation models, and the more specific, practical
perspectives of authors concerned with the various areas of ESL
programs and services. It is divided into four parts, each of which
has three articles. The first part is devoted to approaches to
evaluation in ESL programs. The first article presents a traditional
approach to program evaluation in which sources of information,
the role of the administration, and decision making are of primary
importance. The second article describes the self-study approach
using procedures suggested by two professional organizations,
TESOL and NAFSA. The third suggests that comparison with
groups outside the program may provide a valuable perspective on
the program’s success.
In the second part, articles address aspects of the program
curriculum. The first suggests that curriculum design should be
conceived of as an ongoing process and that excellence in the
program is very much dependent on that process. The second
describes the development of a participatory placement procedure
which incorporates an unstructured, student-controlled oral
placement interview, a writing sample, and a series of short C-test
passages. The third article of this section discusses the cultural


components of ESL programs and suggests how various elements of
culture might be incorporated into the program.
In the third part, articles address noninstructional program
aspects, such as student services, recordkeeping, and promotional
materials. The first emphasizes the importance of recognizing the
“responsibilities and obligations” (p. 136) that ESL programs
assume when they admit international students. Student recruit-
ment, admissions, orientation, advising, employment counseling,
financial aid, housing, health services, community services, and
alumni relations are discussed, with guidelines for the evaluation of
these service components. The next article describes the use of data
bases to collect evaluative data. Through the use of example cases
common to ESL programs, it illustrates how carefully collected and
analyzed data can provide useful information for formative
evaluation. The final article in this section describes the design and
use of promotional materials for ESL programs.
In the final part, articles address the evaluation of program per-
sonnel. The first article in this section describes the many methods
used for faculty evaluation in ESL programs. The second article
describes evaluation by external review of administrator perfor-
mance. The author outlines some general areas desirable in program
directors, including vision, communication skills, monitoring skills,
team-building skills, and decision-making skills, providing sugges-
tions as to how performance in these skill areas might be assessed.
The final article illustrates how administrators might conduct eval-
uations of their own performance. This kind of self-evaluation is
seen as being more appropriate for the characteristics and needs of
individual programs as well as more responsive to specific,
immediate needs for improvement or program development.
In addition to its obvious usefulness for program administrators,
the book could also be used in professional training programs in
ESL. Supplemented by additional sources, particularly in the areas
of testing, personnel management, and computer applications, the
book could provide an introduction to several of the issues and
areas of concern for prospective ESL administrators.
This book represents a very important first step in the
dissemination of information regarding ESL program evaluation in
a form that is easily accessible to ESL professionals worldwide.
Pennington has noted in the introduction that her wish is for the
book to be “only the first of many” on evaluation in ESL. Many of
us working in ESL share that hope.

University of Guam

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.

Teaching Language Minority Students in the Multicultural Classroom.

Robin Scarcella. Englewood Cliffs, NJ: Prentice Hall Regents, 1990.
Pp. xvi + 271.

Scarcella’s intent in this book is to provide content-area teachers with

information about the cultural backgrounds of language minority students
and with culturally responsive teaching strategies. Her larger purpose is to
promote cultural pluralism, which she sees as the true goal of multicultural
education. Her perspective is congruent with that of others who recognize
linguistic and cultural diversity as a resource.
This book fills a need for integrating in one source information from
many different fields relevant to the education of language minority
students. It synthesizes “information from second language acquisition and
pedagogy, bilingual education, multicultural education, ethnography of
education, and critical pedagogy” (viii), with frequent references to the
work of Jim Cummins, Shirley Brice Heath, Michael Long, Stephen
Krashen, Henry Trueba, Lily Wong Fillmore, and many others.
Each chapter develops 1 of 11 general principals for teaching language
minority students:
1. Know your students.
2. Understand language development.
3. Provide comprehensible lessons.
4. Encourage interaction.
5. Appeal to diverse learning styles.
6. Provide effective feedback.
7. Test fairly.
8. Encourage minority parent participation.
9. Appreciate cultural diversity.
10. Incorporate your students’ languages and cultures.
11. Reduce prejudice.
Scarcella explains each principle clearly and supports it with references
to relevant research. Numerous tables, including characteristics of simpli-
fied input, factors encouraging language retention and language loss, and

incongruities between teachers’ and minority parents’ expectations,
provide concise summaries of the research. Quotations from language
learners and teachers enliven the discussion. In addition to the theoretical
framework, Scarcella points out implications for teaching with specific
strategies, which classroom teachers will appreciate. For example, the
chapter on feedback lists 15 practical techniques to help teachers provide
more effective feedback to students of diverse cultural backgrounds. And
the chapter on interaction includes not only suggested activities for
developing speaking proficiency but also specific strategies for reducing
Because this book pulls together so much information in a very readable
form, it is also well suited as a textbook for teachers in training. Discussion
questions and activities at the end of each chapter provide opportunities to
respond to the ideas presented, to use ethnographic techniques of
observing and interviewing, and to develop class activities. Annotated
recommendations for further reading and an extensive bibliography are
valuable to those wanting to explore a subject in more depth. Chapter
appendices include lists of useful journals, pleasure reading books for
students, and resources for information about various minority groups in
the United States.
I used this book quite successfully as a text in a culture learning class for
ESL teachers. It could also be used in a class on multicultural education for
K-12 teachers to broaden their understanding of what multicultural
education should accomplish. I recommend it highly for teachers who
work, or plan to work, with language minority students of all ages,
including those in adult and higher education.

Portland State University

Multicultural Education: A Teacher’s Guide to Content and Process. Hilda

Hernandez. Columbus, OH: Merrill, 1989. Pp. xi + 225. (Available from
Macmillan, New York)

Today’s teachers find themselves in a multicultural classroom culture.

Multicultural Education is intended as a guide to issues and techniques for
teachers with beginning and advanced students in multicultural classrooms
at elementary to secondary levels. I find it also a valuable resource for all
ESL or refugee teachers.
The book is broad in scope, focusing on both content and process in
multicultural education in an attempt to “help teachers see the big picture
of multicultural education and to provide them with the means to make it
a truly integral part of teaching and learning in their classrooms” (p. v).
In the first four of eight chapters, the book provides a broad overview
of important issues in multicultural education. Chapters 6, 7, and 8


(“Instructional Materials in Multicultural Education,” “Development of a
Multicultural Curriculum,” and “Beyond the Classroom: Home,
Neighborhood, Community”) will probably be of greatest interest to ESL
teachers. They provide suggestions for evaluating cultural bias in texts and
materials, setting guidelines for student needs assessment, and doing
ethnographic investigation of a teacher’s own student population. Some of
the procedures outlined in the book could readily be applied in any
multiethnic ESL classrooms, such as those in publicly funded refugee
programs. For example, the “Survey of Cultural Group Characteristics” in
Chapter 8 is a well-developed set of questions for coming to know students
and their ethnic identities. ESL instructors might also choose to apply it in
determining areas of students’ own native cultures that might be readily
incorporated into the ESL curricula. The later chapters also look at student
subpopulations—such as bilingual, gifted, and handicapped—that may
demand special consideration, and provide practical suggestions for
development and implementation of multicultural instructional materials
and teaching strategies.
Each chapter contains a well-organized discussion of the issues related
to the topic of the chapter, followed by a summary and useful reference
list. A valuable feature of each chapter is a section called, “Did you know
that . . . ?” which presents statistics on minority groups, trends in
education and society related to multiculturalism, and examples pertinent
to the chapter. This section adds needed practical perspective on the
theoretical issues discussed in each chapter. It begins to introduce the
reader to some of the surprising realities and challenges of multicultural
education. In addition to presenting an overview of concepts in
multicultural education, this resource book suggests ways to ensure a
multicultural focus in the classroom for both teachers and students.


University of Washington

Communication Strategies: A Psychological Analysis of Second-Language

Use. Ellen Bialystok. Oxford: Basil Blackwell, 1990. Pp. viii + 163.

Since communicative competence became a central tenet of ESL,

communication strategies have received considerable attention within
second language acquisition research. Bialystok provides a comprehensive
and insightful survey of this research subarea, arguing that it has failed to
explain how and why learners use these strategies. She then proposes a
broader level of inquiry encompassing general (first and second) language
processing. These theoretical issues are related to issues in instruction in the
final chapter.
The limitations of L2-focused research are found rooted in two fallacies:
“the uniqueness fallacy” (p. 81), which assumes that there are communica-
tion strategies unique to second language contexts, and “the modularity


fallacy” (p. 82), which assumes that learners’ strategic intentions can be
inferred from surface linguistic forms. The first assumption is claimed to
foster unsupportable definitional criteria of the strategies as problem-
focused, consciously used, and intentionally systematized procedures. The
modularity fallacy is shown to underlie the sometimes tenuous division of
raw language data into mutually exclusive functions of language use,
learner strategies, and communication strategies.
Bialystok notes that inferentially derived taxonomies have greatly
coincided in their descriptions of communication strategies, yet they
cannot represent mentally organized options for the learner. Most
critically, she finds their strategic distinctions to lack both reliability for
labeling strategic utterances in new data and validity, since subjects
identifying pictures based on collected reliability test descriptions made
errors of similar frequency in response to most of the strategic utterances
they heard.
She then proposes a cognitive processing model consisting of an
“analysis-based strategy” (p. 132), which selects linguistic forms from their
available relational mappings to concepts (i.e., the learner’s analyzed
linguistic knowledge), and a “control-based strategy” (p. 133), which
directs attention among sources assisting the expression of meaning. The
former would more elegantly account for the circumlocutions,
paraphrases, transliterations, and word coinages of the taxonomies,
whereas the latter would subsume the use of external resources,
extralinguistic cues, and language switch. Moreover, this model would
include the development of knowledge and control, during which one
strategy may compensate for inadequacies in the other vis-à-vis
communicative demands. Herein Bialystok claims the model’s explanatory
power, drawing support from an inverse relationship shown between
lower proficiency levels and greater use of the control-based strategy.
However, more research is needed to show how this model might further
explain the communicative behaviors of language learners.
The closing discussion casts doubt upon the efficacy of instruction in
communication strategies. Evidence that learners have failed to properly
employ taxonomical distinctions as specific techniques reflects the
conceptual ambiguities in the taxonomies themselves. Training in the
activation of executive skills is mentioned as a potential, yet so far
unproven, alternative. Bialystok therefore suggests that learning and
practice alone may be all that can realistically contribute to the strategic
competence which learners already possess.

Tokai University, Numazu Campus, Japan


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.


University of Massachusetts at Amherst

Speech Rate and Listening Comprehension:

Further Evidence of the Relationship

Nagoya University of Commerce and Business Administration

In terms of temporal variables and comprehension, the outstanding

question (in the sense of being prominent as well as unsettled) is still that
of the relationship between speech rate (SR) and listening comprehension.
The overall nature of the relationship is however clear—after a certain
level, comprehension declines as SR increases—and recent studies
(Conrad, 1989; Griffiths, 1990a) have begun to supply details of the
relationship. It now remains to indicate for specific groups, on specific
texts (a) how much, and at what rates, comprehension decreases as SR
rises; (b) if there is a lower SR limit at which reduced comprehension is
also observed,
The experiment described here is, therefore, one of a series of
investigations which aim to provide more detailed information on the
relationship beween SR and comprehension. The purpose is to indicate
specific SRS or a range of rates which facilitate comprehension and to
indicate rates at which comprehension is likely to decline rapidly.
A previous study, upon which the present investigation is modelled
(Griffiths, 1990a), has indicated that rates of speech faster than 200 words
per minute (wpm) / 3.8 syllables per second (sps) impair comprehension
for lower intermediate learners. However, recent descriptive studies
(Griffiths, 1990b; Higgins & Griffiths, 1990; Tauroza & Allison, 1990) have
shown that the language learner is likely to meet a far greater spread of
rates than the 1.93–3.8 sps rates investigated in that earlier study. A new set
of recordings of different materials was consequently prepared for the
investigation described here. These were recorded at rates of approxi-
mately 5 sps, 3.75 sps, and 2.5 sps. (Tauroza & Allison, 1990, report
“average” SR as 2.30-280 spin, i.e., 3.8-4.7 sps).

Hypotheses tested were as follows:
1. Mean listening comprehension test scores for the passages delivered at
slow SRS (approximately 127 wpm / 2.5 sps) would be significantly
higher than for passages delivered at fast SRS (approximately 250
wpm /5 sps).
2. Mean listening comprehension test scores for the passages delivered at
slow SRS (approximately 127 wpm / 2.5 sps) would be significantly
higher than for the passages delivered at average SRS (approximately
188 wpm / 2.75 sps).
3. Mean listening comprehension test scores for the passages delivered at
average SRS (approximately 188 wpm / 3.75 sps) would-be significantly
higher than for the passages delivered at fast SRS (approximately 250
wpm /5 sps).

The subjects (Ss) for the study were 24 young adult Omani Elementary
School teachers taking part in a university in-service training course. When
they entered the 5-week course, their language proficiency was estimated
as varying between upper elementary and intermediate on a university
examination. Their average level can best be described as lower

Three passages, actually stories, of 454, 451, and 442 words (544, 537,
and 520 syllables; word/syllable ratio = 1.19) were selected from the SRA
Mark 2 Reading Laboratory 2a Power Builders materials (Parker, 1978).
(The passages were designed for native speakers (NSs) of reading aged
8.5.) Using the same conventions as described in Griffiths (1990a), type-
token ratios for the three texts were calculated as 0.42 for Text 1, 0.41 for
Text 2, and 0.46 for Text 3.
As in the previous experiment, actual SRs in the nine recordings came
extremely close to the targets. These, and other information on the texts,
are shown in Table 1.

Testing and Research Design

Fifteen true/false questions for each text were used in testing. (For
details of and rationale for the testing and research design, see the original
study, Griffiths, 1990a). The questions were NS validated by 5
nongraduate NSs (average score on the three tests heard at the fast rate for


Titles and SRs of Recorded Experimental Texts

5 NSs = 13.53). A K-R 20 test indicated an internal reliability of .80. A

randomized complete-block design was again employed, and different
recordings of each text were delivered to each group of Ss from audio
cassettes in a language laboratory. The procedure was repeated three
times—once with each text—following an initially randomized sequence
which ensured that all groups, and therefore all Ss, heard one text recorded
at each speech rate.

Mean scores on the individual texts and over the three texts, are given in
Table 2. This clearly demonstrates the relationship (at these SR levels and
with these subjects) between SR and comprehension.
Mean Scores and Standard Deviation on the Three Texts
and on the Combined Texts
A two-way ANOVA was performed to assess the significance of the
observed difference in means between the treatments, the blocks, the
texts, and the interactions between them. The results of this analysis
together with the F test and p values are given in Table 3. All two-factor
interactions were nonsignificant.

Two-way Analysis of Variance for Scores Related to SRs,
Blocks, and Texts

The significant finding on rate is graphically shown in the notched box-

and-whisker plot (Figure 1).
The plot shows that scores obtained at the slow rate are significantly
different (at the 5% level) from scores at both the average and the fast rates.
This was confirmed by a least significant difference (LSD) range analysis.
Hypotheses 1 and 2 were, therefore, confirmed; but, as the difference
between the average and fast rates was not significant, Hypothesis 3 was
not confirmed.

Input studies might be expected to provide the data from which
practical teaching recommendations can be drawn on language-classroom
performance dimensions. In terms of SR, this study, and those upon which
it builds, indicate the direction these recommendations may follow, but
clearly further experimentation will be necessary before these can be
made with the required level of conviction on a wider scale.
However, should the present finding and that of the previous study
(Griffiths, 1990a) be confirmed in future studies, the oft repeated teacher-
training direction to speak at normal rates to beginners (e.g., see Hatch,
1983) looks increasingly suspect, as even lower intermediate students can
be seen to benefit (in terms of comprehension) from below-average rates
(in the region of 2.5 sps). The issue appears important enough to merit
continued investigation.


Notched Box-and-Whisker Plot of Mean Scores
Obtained at Slow, Average, and Fast Speech Rates

Conrad, L. (1989). The effects of time-compressed speech on native and EFL
listening comprehension. Studies in Second Language Acquisition, 11, 1-16.
Griffiths, R. T. (1990a). Speech rate and NNS comprehension: A preliminary study
in time-benefit analysis. Language Learning, 40,311-336.
Griffiths, R. T. (1990b, March). Language classroom speech rates: Life in the fast
lane. Paper presented at the 24th Annual TESOL Conference, San Francisco,
Hatch, E. M. (1983). Psycholinguistics: A second language perspective. Rowley,
MA: Newbury House.
Higgins, J. M. D., & Griffiths, R. T. (1990, September). Listening comprehension
materials: Why don’t we rate them. Paper presented at the British Association of
Applied Linguistics Conference, Swansea, England.


Parker, D. H. (1978). Mark 2: Reading laboratory, 2A. Chicago, IL: Science
Research Associates (SRA).
Tauroza, S., & Allison, D. (1990). Speech rates in British English, Applied
Linguistics, 11, 90-105.

Editor’s Note. The editor is deeply saddened to report the death of Roger

A Second Look at T-Unit Analysis:

Reconsidering the Sentence
Indiana University

■ Although originally designed to assess syntactic development in the

written work of children learning their first language, the T unit has been
widely used in second language research and is particularly useful for oral
production (see, for example, Beebe, 1983; Gaies, 1980; Larsen-Freeman,
1983). A variety of measures of complexity have been based on the T unit,
e.g., words per T unit, words per error-free T unit, and clauses per T unit
(Gaies, 1980; Hunt, 1965; Larsen-Freeman, 1983). Nevertheless, for the
description of syntactic complexity in the writing of adult second language
learners, sentence-based analyses may be superior.
Hunt (1965) introduced the T unit, or minimal terminable unit, to
measure development of sentences in the writing of grade-school children.
Each T unit contains one independent clause and its dependent clauses.
Hunt (1970) described T units as “the shortest units into which a piece of
discourse can be cut without leaving any sentence fragments as residue”
(p. 189). A sentence has two (or more) T units when independent clauses
(with subjects and finite verbs) are conjoined as in Example 1, but a single
T unit when one or more clauses are embedded in an independent clause
as in Example 2:
1. There was a woman next door, and she was a singer.
(S + S) = 2 T units
2. There was a woman next door who was a singer.
[S (S)] = 1 T unit
Because children in the fourth grade (Hunt’s youngest group) frequently
produce what are known in the classroom as run-on sentences, as in
Example 3, younger children appear to produce longer sentences than
older children or than adults. (In Example 3 the T-unit boundaries have
been marked.)
3. I like the movie we saw about Moby Dick, the white whale [1] the captain
said if you can kill the white whale, Moby Dick, I will give this gold to the one
who can do it [2] and it is worth sixteen dollars [3] they tried and tried [4] but
while they were trying they killed a whale and used the oil for the lamps [5]


they almost caught the white whale [6]. (Ll English, 4th grade) (6 T units/
1 sentence)
By dividing sentences such as Example 3 into T units, however, Hunt
showed that children write increasingly longer units as they get older.
Although the T unit has advantages for certain language samples, in
evaluating the syntactic complexity of compositions written by advanced
adult second language learners, T-unit analysis does not seem to reflect
accurately the knowledge of the learner. As the unit directly produced by
the learner, the sentence has a certain degree of psychological reality in
that it allows researchers to glimpse how the learner views the structure of
the English sentence. In contrast, T-unit analysis artificially divides
sentences that were intended to be units by the language learner, imposing
uniformity of length and complexity on output that is not present in the
original language sample. By treating all conjoined sentences as if they
were not conjoined, a T-unit analysis discounts the learner’s knowledge of
Although a T-unit analysis divides a cumbersome run-on sentence such
as Example 3 into manageable pieces, this same division breaks up
legitimately coordinated sentences, as in Example 4, which are not only
grammatical but which reflect a certain rhetorical sophistication that
would be absent if the two clauses were merely juxtaposed as they are in
Example 5. Example 4 and all subsequent examples with multiple T units
originated in essays written by advanced second language learners
(Bardovi-Harlig & Bofman, 1989).
4. Hundreds of schools were built, and tens of institutions are starting to join in
providing technical education to the public. (Ll Arabic) (2 T units/1
5. Hundreds of schools were built. Tens of institutions are starting to join in
providing technical education to the public. (2 T units/2 sentences)
Not only does the T-unit analysis divide sentences exhibiting
coordination with an additive function as in Example 4, but it also divides
sentences with asymmetrical conjunction. In asymmetrical conjunction
(Lakoff, 1971; Schmerling, 1976), the conjuncts cannot be reversed and do
not have the same meaning individually as the conjoined unit. The
asymmetrical conjunction in Example 6 functions as a conditional sentence
(Quirk, Greenbaum, Leech, & Svartvik, 1965) similar to the if-then condi-
tional in Example 7.
6. Take notice on his characteristics, and you’ll find that there is different
between a Chinese and an Asian from other nations. (Ll Chinese) (2 T units/
1 sentence)
7. If you take notice of his characteristics, then you’ll find that there is a
difference between a Chinese and an Asian from other nations. (1 T unit/
1 sentence)
Using the sentence as the unit of analysis, Examples 6 and 7 receive the
same analysis; using T units, they receive a different analysis. In other


words, a sentence analysis reflects the complexity of a conjunc-
tive-conditional such as Example 6, but a T-unit analysis does not.
Furthermore, by dividing conjuncts, the T-unit analysis treats all
conjunctions as semantically null, breaking up syntactic and semantic
units. Just as a T-unit analysis divides conjunctions with and, it also divides
conjuncts joined by but and or. Separating sentences such as Example 8
into 2 units divides a semantic whole: the contrast between conjuncts is
part of the information conveyed by the sentence.

8. They do not encourage their children to mix with the opposite sex, but they
do let them interact to certain extent. (Ll Malay) (2 T units/1 sentence)

If the T-unit analysis seems to underestimate a learner’s knowledge of

coordination on the one hand, on the other, it can also give the learner too
much credit by breaking up sentences. The sentence in Example 9 is overly
complex and cumbersome, yet a T-unit analysis misses its awkwardness by
separating the conjuncts.

9. When you see a group of Asians talking, if they are speaking their native
language, I think you do not know which native they come from, but if you
hear that among their conversation there are some English single words
between their native language, I can say that they are Chinese from Hong
Kong. (Ll Chinese) (2 T units/1 sentence)

A sentence analysis captures the full effect of this overloaded sentence,

whereas a T-unit analysis, by breaking the sentence into two T units,
reduces any measure of complexity by one-half. 1 As an illustration,
consider one measure of complexity, the subordinate clause index
(number of clauses per syntactic unit; Hunt, 1970). A sentence analysis of
Example 9 finds 10 clauses/sentence, but a T-unit analysis counts an
average of 5 clauses/T unit (6 clauses/T unit and 4 clauses/T unit,
respectively). Thus, a T-unit analysis quantitatively reduces the excessive
sentence combining on the part of the learner, obscuring the inappropri-
ateness of so many combinations.
A third difficulty lies in the differential treatment of full clause
coordination by environment. When the conjuncts are main clauses as in
Example 4 and Example 10, a T-unit analysis divides the conjuncts, but not
when the conjuncts are subordinate clauses as in Examples 11 and 12.

10. Main clause coordination 11. Subordinate clause coordination

(2 T units/1 sentence) (1 T unit/1 sentence)
[[S2 + S3]]S, [S2 [S3 + S4]]S1

1 Thatis not to say that a sentence-based analysis simply changes the intervals on the scale,
however. Nonconjoined sentences such as those in Example 2 are unaffected. In Example
2 there are 2 clauses/T unit and 2 clauses/sentence. The mean number of clauses per unit
of analysis changes in direct proportion to the number of main clause coordination in a
language sample.


12. s l [They believed that the remaining one character is very important
S 2 [because S 3 [ s 4 [the relationship between these characters influence their
life greatly ]s4 and s5 [they can decide only one character among their
name. ]s5]s3]s2]sl (LI Korean)
As a result, a T-unit account treats coordination in two different ways
although the coordination of any two full clauses presumably involves the
same linguistic knowledge.
If we wish to investigate coordination specifically, it can be described
quantitatively by measures such as a coordination index. The coordination
index proposed here indicates the degree to which a learner achieves
syntactic complexity through coordination (as opposed to subordination).
To calculate the coordination index, the number of clauses and sentences
in the language sample are tabulated. Because each sentence has at least
one independent clause (or attempted independent clause in the case of a
fragment), the number of sentences is subtracted from the total number of
clauses to give the number of clauses that are combined (by either
coordination or subordination). Next, the number of independent-clause
coordination is counted. The number of independent-clause coordina-
tion divided by the number of combined clauses yields the coordination
index in decimal form. Multiplied by 100, the coordination index gives the
percentage of syntactic complexity achieved by independent-clause
coordination. A sample calculation is given below:
Calculating the Coordination Index:
Clauses Sentences Clauses Combined = Indep.-Clause Coordination
Clauses – Sentences Coordination Index
47 22 47-22=25 6 6/25=.24x100=24%
The coordination index differs from Hunt’s (1970) “main clause coordi-
nation index” (p. 189) in that our coordination index considers only
multiclausal sentences and reflects the frequency with which coordination
is used by a learner relative to the total number of combinations produced.
In contrast, Hunt’s main clause coordination index measures the number of
T units per sentence, giving a ratio of the total number of T units to the
total number of sentences.
The coordination index quantitatively distinguishes between learners
who employ coordination and subordination to attain syntactic complex-
ity. Thus, if beginning learners (Ll and L2) favor coordination over subor-
dination (Monroe, 1975), then the coordination index should be higher for
beginners than for advanced learners. Analysis of written text supports this
Eighty-six compositions written by ESL learners at seven levels of
proficiency were examined. The 56 learners in Levels 1-6 were enrolled in
the Intensive English Program of the Center for English Language
Training at Indiana University. The compositions, written on the topic
How do you expect your life to be different from your grandparents’?
were part of the program examination and were allotted 35 minutes. Data
for the 30 advanced learners in Level 7 come from an earlier study


conducted under comparable conditions (Bardovi-Harlig & Bofman,
1989). Placement was determined by the program independent of this
The compositions were coded far clauses and T units by the author and
a trained assistant with an interrater reliability of 96%. Discrepancies were
resolved by discussion. The coordination index in Table 1 shows that the
number of combinations resulting from coordination decreases with
learner proficiency. Less proficient learners rely cm coordination to
combine sentences to a greater extent than more advanced learners do.2

Percentage of Sentence Combinations Resulting from Coordination

An additional benefit of using sentence-based analyses is the increased

opportunity for the teacher-researcher dialogue that is so important in our
field. By stating our findings in sentence units, we make them most
relevant to the classroom where the sentence is used in instruction and
In conclusion, a sentence analysis has several advantages over a T-unit
analysis for the description of syntactic complexity in written production
of adult second language acquisition. A sentence analysis allows
knowledge of coordination to be represented in quantitative descriptions,
thereby more accurately reflecting learner knowledge. The addition of the
coordination index captures the shift from the use of coordination by
beginning learners to the use of embedding by more advanced learners.

An earlier version of this report was presented by the author and Theodora Bofman
at the 22nd Annual TESOL Convention in Chicago, March 1988. I thank Cheryl
Engber for her assistance in coding the compositions.
2 Theincrease in coordination at Level 6 may be an artifact of program placement. In an
unrelated study with different students from the same program, Level 6 showed a similar
deviation (Bardovi-Harlig, 1990). Although this merits further investigation, the scores of
Level 6 do not detract from the clear overall pattern.


Bardovi-Harlig, K. (1990, April). The acquisition of form and meaning: The syntax
and semantics of tense. Paper presented to the 9th World Congress of AILA,
Chalkidiki, Greece.
Bardovi-Harlig, K., & Bofman, T. (1989). Attainment of syntactic and morphologi-
cal accuracy by advanced language learners. Studies in Second Language
Acquisition, 11, 17-34.
Beebe, L. (1983). Risk-taking and the language learner. In H. W. Seliger &
M. H. Long (Eds.), Classroom oriented research (pp. 39-65). Rowley, MA:
Newbury House.
Gaies, S. J. (1980). T-unit analysis in second language research: Applications,
problems and limitations. TESOL Quarterly, 14 (1), 53-60.
Hunt, K. W. (1965). Grammatical structures written at three grade levels (Research
Report No. 3). Urbana, IL:, National Council of Teachers of English.
Hunt, K. W. (1970). Recent measures in syntactic development. In M. Lester (Ed.),
Readings in applied transformation grammar. New York: Holt, Rinehart and
Lakoff, R. (1971). U’s, and’s, and but’s about conjunction. In C. J. Fillmore &
D. T. Langendoen (Eds.), Studies in Linguistic semantics (pp. 115-149). New
York: Holt, Rinehart and Winston.
Larsen-Freeman, D. (1983). Assessing global second language proficiency. In
H. W. Seliger & M. H. Long (Eds.), Classroom oriented research (pp. 287-304).
Rowley, MA: Newbury House.
Monroe, J. H. (1975). Measuring and enhancing syntactic fluency in French. The
French Review, 48, 1023-1031.
Quirk, R., Greenbaum, S., Leech, G., & Svartvik, J. (1985). A comprehensive
grammar of the English language. London: Longman.
Schmerling, S. {1976). Asymmetric conjunction and rules of conversation. In
P. Cole & J. L. Morgan (Eds.), Speech acts (pp. 211-256). New York: Academic
Author’s Address: Program in TESOL and Applied Linguistics, Department of
Linguistics, Memorial Hail 313, Indiana University, Blooming-
ton, IN 47405


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.

Comments on Suzanne Irujo’s Review

of Rosalie Pedalino Porter’s
Forked Tongue: The Politics of
Bilingual Education
A Reader Reacts. . .

Silver Spring, MD

While I generally agree with Irujo’s TESOL Quarterly (Vol. 25,

No. 1) review of Porter’s Forked Tongue, there are some flaws in
the review. Irujo’s most devastating criticisms of Porter are that (a)
Porter gives a biased presentation of the research by ignoring
studies that contradict her opinions, and (b) Porter misrepresents
the research of others. I fully agree but take issue with Irujo in that
she barely scratched the surface of Porter’s unscholarly misrepre-
sentations. Consider these examples which Irujo did not mention.
Porter’s discussion of the Fairfax County ESL program illustrates
the biased way she treats the research literature. Referencing one re-
port from the Fairfax school system, Porter concludes “These evalu-
ations portray a program that is working well for its students”
(p. 148). No mention is made of Collier’s (1987, 1987/1988) extensive
evaluation of the program which concluded by recommending it be
replaced with a transitional bilingual education (TBE) program.
Porter describes a report from El Paso (1987) which, according to
Porter, concluded that an all-English immersion program was su-
perior to TBE. The report makes no such finding. The experiment-
al El Paso program is a dual bilingual immersion program in which
the students are taught both Spanish and English: "BIP [Bilingual
Immersion Program] is a bilingual program, students receive in-
struction in both English and Spanish" (El Paso Independent School

District, 1990). There are differences between BIP and the program
to which it was compared, a typical Texas transitional bilingual
education program, but nothing in the El Paso study supports the
use of all-English immersion, as Porter claims.
Porter describes a structured immersion program in San Diego
(p. 151). Since I invented the term structured immersion (see Baker&
de Kanter, 1981), I am comfortable in saying there is not now, nor was
there ever, a structured English immersion program in San Diego. Like
El Paso, San Diego had a dual, two-language, immersion program.
There is some evidence that the dual immersion program was better
than TBE. The San Diego study, if anything, argues for maintenance
bilingual education programs instead of TBE. Nothing in either the San
Diego or El Paso studies support Porter’s claims.
Porter describes a program in Paramount, California, in which
she says students in an English immersion program “show a far
higher rate of English-language learning than the children in the
Spanish bilingual program” (p. 152). Not recognizing this study, I
consulted Porter’s reference for it, a report prepared for the Reagan
Administration by a group of educators opposed to bilingual
education programs. There was no data in that report either, nor
was there any reference to any study supporting Porter’s claim.
I have test scores from this program for the year in which Porter
wrote her book. The school system did have both a bilingual
education program and an all-English immersion program. Of 12
comparisons between the two programs (reading, language, math,
and total scores for three grades), the bilingual education program
students had the higher score for 8 of the 12 comparisons. The
evidence points to a conclusion just the opposite of Porter’s.
Porter misrepresents some of my research, writing “federal
statistics show that up to 60 percent of the children in bilingual
education programs are English dominant” (p. 66). I quote the
sentence from the article she cited where the figure 60% appears:
“Title VII requires by law a minimum of 60 percent LEP [limited
English proficient] enrollment” (Baker & Rossell, 1987, p. 263).
In discussing the English language amendment, Porter says
“Among my professional colleagues, I have heard unexpectedly
inflammatory opinions on the topic” (p. 217). To support this
conclusion, she quotes at some length from an article by Elliott Judd
and then says “The logic in Mr. Judd’s reasoning seems absurdly
inverted” (p. 217). Judd (personal communication, September 24,
1991) told me that by selectively omitting parts of his statement,
Porter severely altered and misrepresented his meaning. Porter also
misrepresents findings of the U.S. Department of Education’s
survey of language minority parents (I was the U.S. Department of
Education project director for this survey). Porter says “The report


on that study contains two major conclusions that have important
implications for public education policy: (1) parents agreed that
special language programs should be provided to language minority
children, but opinions varied widely both within and among ethnic
groups as to what the most desirable programs are” (p. 1.23). In the
discussion that follows, Porter distinguishes among the different
“Hispanic” ethnic groups.
The first problem with Porter’s discussion of the findings of the
report is that representative samples were achieved for only
Mexican American and Asian parents. Thus, as is obvious in the
report, comparisons among the different Hispanic groups should
not, indeed cannot, be made. As for Porter’s first conclusion, the
findings of the study are presented in Table 1.
Should LMs Get Extra Help at School Learning English by Home Language Use
Percentage of Parents Answering Yes

Home language Mexican American Asian

English only 77.4 57.2
Both 81.3 62.5
Other only 95 80.2
Total 82 63.7

The most striking finding in Table 1 is that a large minority of Asian

parents do not think the schools should provide special help in
learning English. Moreover, even the Mexican American parents are
far from unanimous on the issue. Porter fails to point out the big
differences within ethnicities by home language use (see Table 1).
This is an important variable, since the study sample was of
language minority parents, not of parents of LEP students, another
point Porter neglected.
Moreover, Porter does not give her readers the important finding
of the study: parents from homes with heavy use of L1 at home
want their children to learn English and retain Ll; these parents
think bilingual education programs are the best way to teach
English. Moreover, they prefer maintenance bilingual education
programs to TBE.
Irujo errs in accepting Porter’s claims for the Newton program.
Porter repeatedly and vociferously proclaims that the all-English
program in one school district, Newton, Massachusetts, shows all-
English programs for LEP students are superior to bilingual
education programs. However, there is no evidence offered in


the book to support these sweeping claims. There are no test scores,
no data on dropout rates, or any other empirical measures. Instead,
all we have is the author’s personal opinion that she is right and
everybody else is wrong. In fact, the book is so bereft of any
empirical evidence about the effects of the Newton program that
there is no reason to conclude it is even as effective as is submersion.
This is not research. This is a classic propaganda technique: If you
say it often enough and loudly enough, many people will believe it,
regardless of the truth.
Irujo says that the Newton program is “a model” for schools where
native language programs are either “not possible or not necessary”
(p. 152). There seems to be no basis for Irujo’s conclusion since there
is no evidence of the effectiveness of the Newton program.
In short, Porter’s misrepresentations of the bilingual education
research in the United States is much more extensive than Irujo
noted. Likewise, Porter’s discussion of the Canadian research is
more seriously flawed than Irujo noted. In reviewing Porter’s
discussion of the Canadian research, Cummins (1991, pp. 787-792)
Many of the arguments advanced by Porter are confused to the point of
almost total incoherence. The text is full of inaccuracies and
contradictions that reveal an extremely shallow grasp of the
research. . . . [The book is characterized by] limited understanding of
the research and the extent to which data have been misrepre-
sented. . . .
[Porter praises dual immersion programs] but she never addresses the
fact that the documented academic success of Latino students in such
programs, despite reduced exposure to English, refutes her cherished
“time-on-task principle. . . .
The reason transitional bilingual programs have mixed results is that
many fail to provide an adequate L1 conceptual foundation in early
grades as a result of insufficient time devoted to L1 promotion. . . .
[Porter says the Canadian heritage language programs data shows
delay in the use of L2 is harmful, but] there is no citation to back up this
claim for the simple reason that there is no “available evidence.” In the
first place, no Canadian program . . . comes even close to the
description offered by Porter. . . .
Contrary to Porter’s assertion, the rationale for heritage language
programs in Canada is very different from that of transitional bilingual
programs in the United States.
[Her] claim that English immersion is not offered to minority students
is quite bizarre. . . .
[She accuses researchers of dishonesty to advance their own careers,
but] it is significant but not surprising that there is no citation to back up
this “fact.”


She fails to realize that the French immersion programs are bilingual
programs. . . .
Porter also overestimates the outcomes of French immersion
programs. . . . The data are clear that immersion students remain
considerably below native-like levels. . . .
Porter devotes only one sentence to late immersion, possibly because
the results of late immersion refute the “time-on-task principle. . . . and
also refute Porter’s unqualified claim that [it’s best to learn L2 early]. . . .
It is certainly not a book that reveals great depths of scholarship.
Porter sees the hand of the bilingual conspiracy reaching into the
colleges. She claims that the state bilingual education program put
pressure on the university where she was a graduate student to
block her dissertation, which was critical of bilingual education
programs. After Porter complained to the dean, she says the
university administrator who applied the pressure was promoted
while Porter’s advisor, who supported Porter, was demoted.
These are obviously serious allegations, accusing public
education officials of conspiring to stifle free thought and a major
university of acquiescing to political pressure. If Porter is right, was
not the university’s response reprehensible? Should we believe
Porter’s tale? She presents no collaborating evidence of these
events. She did in fact get her degree, and her dissertation, highly
critical of bilingual education programs, was accepted. She has
established a record of biased and misleading reporting of the
empirical literature in the book. If she cannot accurately summarize
a written document, can we trust her interpretation of these events?
Given Porter’s pervasive misrepresentations and utter lack of
evidence supporting her claims, Irujo lets Porter off too easily.
Irujo’s strongest condemnation is that “Unfortunately, the flaws of
the book far out weigh its strengths” (p. 152). They do at that, but
the book is also an affront to the dignity of science. Reviewers in
scholarly journals, as well as readers, must perforce trust the author
of a supposedly scholarly work to be accurate, especially in
summarizing other research. Irujo found that Porter misrepresented
the research with which Irujo was familiar. Cummins found that
Porter misrepresented the research with which he was familiar. I
found that Porter misrepresented the research with which I was
familiar. No reviewer or reader should be expected to have the
degree of familiarity with the literature needed to spot these
misrepresentations. Porter violated the most basic trust between
reader and author of scholarly works. Porter’s level of scholarship
falls far short of the most minimal standards of acceptable research.
Porter’s unmerited scholarly pretensions led to large numbers of
people on the far right and in the English-only movement who share


Porter’s biases giving the book wide publicity as evidence of both
the ineffectiveness and political corruption of the bilingual
education movement. Neither conclusion is justified, and only time
will tell the full extent to which the book has damaged the quest for
effective instruction for LEP children.
In some places, Irujo is inaccurate in her review. Irujo correctly
notes Porter ignores research contrary to her beliefs, citing as an
example Porter’s use of Baker and de Kanter while ignoring Willig.
But Irujo fails to mention Baker (1987, 1990), in which I show that
most of Willig’s analysis is invalid and that Willig’s major
conclusion, repeated by Irujo, that bilingual education students out-
performed comparison students taught only in English results from
Willig averaging together a large positive effect of bilingual
education programs on L1 with a small (trivial) negative effect on
English. In other words, what Willig shows is that bilingual
education programs advance L1 at no cost to the students’ English
development. Willig does not contradict Baker and de Kanter, as
Irujo implies.
Moreover, Irujo misstates Baker and de Kanter’s conclusions. We
did not conclude “there was no evidence for the effectiveness of
bilingual education programs” (p. 152). Our conclusion was that the
evidence supporting the effectiveness of bilingual education
programs was not sufficiently strong to justify a federal mandate of
TBE as the only allowed way to teach LEPs. About one third of the
studies we reviewed found positive effects for bilingual education
Although this may appear to be hairsplitting, it is important. Think
of the analogy of investing for retirement. Some people will take
chances and invest in stocks, others will be conservative and put their
money in the bank. Our view was that the federal government should
be very conservative in taking the unprecedented step of mandating
a curriculum (something the U.S. Department of Education is
forbidden by law to do in general, but apparently is possible in the
case of LEP students because of a loophole).
Whether a local school board should adopt bilingual education
programs was not considered by Baker and de Kanter. My view is
that the research clearly shows bilingual education programs are
superior to all-English programs, such as those advocated by Porter,
in the early stages of learning English, but I still would not say that
a federal mandate is justified.
Finally, Irujo did not mention two recent studies (Burkheimer,
Conger, Dunteman, Elliott, & Mowbray, 1989; Ramirez, Pasta,
Yuen, Billings, & Ramey, 1991) that show beyond any question that
all-English programs such as that advocated by Porter are not better


than bilingual education programs in teaching English to LEP
students. In fact, in the early stages of learning English, bilingual
education programs are better than all-English instruction.

Baker, K. (1987). Comment on Willig. Review of Educational Research,
57(3), 351-362.
Baker, K. (1990). Bilingual education: Partial or total failure? Proceedings
of the 10th Second Language Research Forum. Eugene: Department of
Linguistics, University of Oregon.
Baker, K., & de Kanter, A. (1981). The effectiveness of bilingual education
programs: A review of the literature. Final draft report. Washington,
DC: U.S. Department of Education.
Baker, K., & Rossell, C. (1987). An implementation problem: Specifying
the target group for bilingual education. Education Policy Analysis, 1 (2),
Burkheimer, G. J., Jr., Conger, A., Dunteman, G., Elliott, B., & Mow-
bray, K. (1989). Effectiveness of services for language minority limited
English proficient students. Raleigh-Durham, NC: Research Triangle
Collier, V. (1987). Age and rate of acquisition of second language for
academic purposes. TESOL Quarterly, 21 (4), 617-641.
Collier, V. (1987/1988). The effect of age on acquisition of a second
language for school (FOCUS: Occasional Papers in Bilingual Education
No. 2). Wheaton, MD: National Clearinghouse for Bilingual Education.
Cummins, J. (1991). Forked tongue: A critique. Canadian Modern
Language Journal, 47 (4), 786-793.
El Paso Independent School District. (1987). Interim report of the five year
bilingual education pilot project. El Paso, TX: Author.
El Paso Independent School District. (1990). Bilingual education
evaluation: The sixth year. El Paso, TX: Author.
Ramirez, J. D., Pasta, D. J., Yuen, S. D., Billings, D. K., & Ramey, D. R.
(1991). Longitudinal study of structured English immersion strategy,
early-exit, and late-exit bilingual education programs for language
minority children (Vols. 1-2, U.S. Department of Education Report,
Contract No. 300-87-0156). San Mateo, CA: Aguirre International.

The Reviewer Responds. . .

Boston University

Keith Baker takes issue with my review of Rosalie Porter’s book

Forked Tongue: The Politics of Bilingual Education because it
“barely scratched the surface of Porter’s unscholarly misrepresenta-
tions.” I couldn’t agree more and am grateful to Baker for calling
attention to many of the defects of the book that I did not include
in my review because of space constraints. Because Baker and I are
essentially in agreement on the merits of the book in question, I will
not respond to his comments point by point. However, I would like
to clarify my position with regard to certain statements he makes.
Baker says that I err “in accepting Porter’s claims for the Newton
program” because “there is no evidence offered in the book to
support these . . . claims.” I did not accept Porter’s claims; in fact,
I specifically rejected her assertion that students in Newton or in
any other all-English program could achieve grade-level equiva-
lence in 1 to 3 years. My basis for rejecting these claims was the
conflict with other research, particularly Collier (1987), and the fact
that “she supplies no data to substantiate her claim” (p. 154). My
statement that Newton could serve as a model for programs where
native language instruction is not possible or not necessary was
based on my own experience supervising student teachers in
Newton’s ESL classrooms. I have seen excellent teaching, a well-
developed content-based curriculum, and students obviously
making progress in English and in content areas. I did not imply that
this program was more effective than a bilingual program would
have been.
Baker mentions Porter’s allegations concerning efforts to block
her dissertation. Because I had no personal knowledge of this
incident, and could only have responded to it indirectly, as Baker
does, I decided not to include it in my review. Instead, I directly
refuted an incident that I was personally involved in (the MATSOL
Executive Board meeting) where I knew Porter had twisted the
facts to suit her own purposes. I hoped in this way to cast doubt on
the veracity of everything Porter claimed.
I apologize to Baker if I misstated the conclusions of Baker and
de Kanter (1981). I should have reread the report, rather than
relying on my memory of it, which may have been influenced by
other people’s interpretations (or misinterpretations) of this widely
cited source. In returning to the original, however, I note a revealing
change of perspective. In 1981, Baker and de Kanter claimed that


“carefully conducted second language instruction in all subjects
may well be preferable to bilingual methods” (p. 16). Baker now
states that “the research clearly shows bilingual education programs
are superior to all-English programs . . . in the early stages of
learning English.” It may be that improved research, called for by
Baker and de Kanter (1981), has helped to clarify the issues.
I also appreciate Baker’s calling attention to two recent studies
showing the superiority of bilingual programs. I was not familiar
with Burkheimer, Conger, Dunteman, Elliott, & Mowbray (1989),
and I chose not to mention Ramirez, Pasta, Yuen, Billings, & Ramey
(1991) because it had not yet been published at the time I wrote my
review, although the major findings of the study had been widely
I might argue with some of the other issues raised by Baker, such
as whether Title VII funding of bilingual education programs
constitutes a “federal mandate,” as he implies. I will not do so
because we are in agreement on the most important issues. Did I let
Porter off too easily, as Baker claims? I may have. In attempting to
avoid the impression that I was presenting a biased, one-sided view
(something I had criticized Porter for doing), I was perhaps less
forceful in my review than I might have been. Thanks to Keith
Baker for adding some of that force.

Baker, K. A., & de Kanter, A. A. (1981). The effectiveness of bilingual
education programs: A review of the literature. Final draft report.
Washington, DC: Department of Education, Office of Planning, Budget,
and Evaluation.
Burkheimer, G. J., Jr., Conger, A., Dunteman, G., Elliott, B., & Mow-
bray, K. (1989). Effectiveness of services for language minority limited
English proficient students. Raleigh-Durham, NC: Research Triangle
Collier, V. (1987). Age and rate of acquisition of second language for
academic purposes. TESOL Quarterly, 21 (3), 617-641.
Ramirez, J. D., Pasta, D. J., Yuen, S. D., Billings, D. K., & Ramey, D. R.
(1991). Longitudinal study of structured English immersion strategy,
early-exit, and late-exit bilingual education programs for language
minority children (Vols. 1–2, U.S. Department of Education Report,
Contract No. 300-87-0156). San Mateo, CA: Aguirre International.


Teaching Issues
The TESOL Quarterly publishes brief commentaries on aspects of English lan-
guage teaching. For this issue, we asked two educators to address the following
question: Under what circumstances, if any, should formal grammar instruction
take place?

Edited by SANDRA McKAY

San Francisco State University

Formal Grammar Instruction

An Educator Comments. . .

University of California, Los Angeles

To frame this discussion, two clarifications are in order. First, I will be

addressing grammar instruction only with the ESL context in mind. For
the EFL context, a somewhat different answer is needed. Second, I would
highlight my underlying assumption that working towards grammatical
accuracy does not mean sacrificing fluency; grammatical competence is
one component of communicative competence—the others being
sociolinguistic competence, discourse competence, and strategic
competence (Canale, 1981).
What do we mean by “formal grammar instruction”? For some this
means the teacher lecturing about grammar or preparing the students for
a test like the TOEFL (Test of English as a Second Language); for others
perhaps it evokes an image of the teacher answering learners’ questions
about grammar or carrying out a correction activity. Such perspectives
entail too narrow a view of “formal grammar instruction.” For me, any
learning activity that focuses the learner’s attention on the form of a
message (ideally, in the context of the meaning and function of the
message) constitutes formal grammar instruction. Such attention focusing
can, of course, be done both deductively and inductively. It can be done
by a teacher or a tutor; it can be part of an initial presentation, a practice
activity, or a follow-up error-correction session. Any such instruction is
more effective if it is discourse-based and context-based than if it is
sentence-based and context-free.
Given this broad view of formal grammar instruction, one must still
consider the learner’s age, proficiency level, and ultimate objectives in
studying a second language in order to answer the question of when to
teach grammar formally in a satisfactory manner.
Research in second language acquisition (for a summary, see Larsen-
Freeman & Long, 1990) indicates that generally only young, prepubescent
learners, and then only those with good access to native-speaking peers


and sufficiently rich and varied native speaker input, can—in the absence
of formal grammar instruction-learn a-foreign or second language with
nativelike proficiency and accuracy. Postpubescent adolescents and adults
need to pay some attention to the form of the target language. If they do
not, they ultimately develop an incomplete and imperfect interlanguage
that reflects learning problems such as negative transfer from the native
language, simplification, overgeneralization, erroneous rule formation,
and so forth (see Richards, 1974). Indeed, Higgs and Clifford (1982) argue
from their years of experience at the Foreign Service Institute that adults
who learn a foreign language without any formal grammar instruction
during the basic- language learning stage can never achieve high
proficiency in the target language. Such learners may become fluent, but
in terms of their grammatical development, they plateau at an
intermediate or low-intermediate level and are unable to progress even if
they are provided with formal grammar instruction at a later time.
The research of Cummins (1979), among others, tells us that immigrant
school-age children can learn basic interpersonal oral communication skills
in a second language -within a few years. However, it takes them up to 7
years (and sometimes even up to 10 when additional factors are
considered; cf. Collier, 1989) to acquire the second language literacy skills
needed to achieve academic parity with native speakers. This suggests that
in optimal second language learning, the spoken language and the written
language are different: it is easier to learn to understand and speak a
second language for basic oral communication than it is to learn to read
and write in this language for academic purposes. Grammar, likewise,
seems to be more crucial if advanced proficiency is desired and a high
level of literacy is required.
McGirt’s (1984) study supports this distinction between spoken and
written language by showing that nonnative ESL students at the University
of California, Los Angeles, many with 7 or more years of residence in the
U.S. and virtually no accent in their speech, still tend to make significantly
more morphosyntactic errors in their academic writing than do native
speakers (i.e., 7 errors per 100 words vs. 1 error per 100 words). Sixty
percent of the ESL students McGirt studied had acceptable writing from
the point of view of organization and logic. However, only 20% were rated
as overall acceptable writers; faulty grammar made the writing of the
other 40% unacceptable to the composition faculty.
When a similar population of ESL students at the University of
California, Davis (i.e., students who had a high rate of faulty grammar and
other problems in their writing) was surveyed and interviewed by
Schwabe (1989), she found that most had never received any grammar
instruction (i.e., they were not provided with the rules of English
grammar) or grammar correction on their written work while in middle
school or senior high school. Thus even under optimal environmental
conditions (i. e., when the learners are using the second language to learn
content and to interact with native speakers), the grammar needed for
acceptable academic writing is not well acquired in the total absence of


any feedback or formal grammar instruction. Fortunately, Frodesen
(1991), and others offer ESL teachers suggestions on how to effectively
provide contextualized instruction in grammar as one component of a
writing course for nonnative speakers.
Before concluding, I must point out that grammar instruction carried out
for its own sake, totally divorced from activities that involve using it as a
resource to convey meaning is as irresponsible and counterproductive as
not teaching grammar at all. The challenge for language teachers is to
develop effective ways of focusing learner attention on form at critical
moments while learners are using the second language for purposeful
communication, especially written communication. This is not easy, but it
is necessary if one is teaching postpubescent learners who need to achieve
a high level of proficiency for professional or academic purposes,
especially if the learners need to become reasonably effective and accurate
writers in their second language. Grammatical accuracy is important
because it marks a second language learner as competent; it helps open
academic, social, and economic doors for them.
Yes, there are definitely circumstances where formal grammar
instruction is necessary. And there are also circumstances where it is not,
for example, with learners who want to acquire only basic, rudimentary
oral communication such as older immigrants, or preliterate immigrants
who want a hands-on job, or young children who are learning the second
language in an optimal environment with lots of native-speaking peers
around them. Obviously, the question of when to teach grammar formally
has no simple answer.

Marianne Celce-Murcia is Professor of TESL and Applied Linguistics at the
University of California, Los Angeles. She has coauthored two books used to
prepare ESL teachers to teach English grammar.

Canale, M. (1981). From communicative competence to language pedagogy. In
J. Richards & R. Schmidt (Eds.), Language and communication (pp. 2-27).
London: Longman.
Collier, V. P. (1989). How long? A synthesis of research on academic achievement
in a second language. TESOL Quarterly, 23 (3), 509-531.
Cummins, J. (1979). Linguistic interdependence and the educational development
of bilingual children. Review of Educational Research, 49(2), 222-251.
Frodesen, J. (1991). Grammar in writing. In M. Celce-Murcia (Ed.), Teaching
English as a second or foreign language (pp. 264-276). Boston: Newbury House/
Heinle & Heinle.
Higgs, T. V., & Clifford, R. (1982). The push toward communication. In
T. V. Higgs (Ed.), Curriculum, competence, and the foreign language teacher
(pp. 57-79). Lincolnwood, IL: National Textbook Company.


Larsen-Freeman, D., & Long, M. H. (1990). An introduction to second language
acquisition research. London: Longman.
McGirt, J. D. (1984). The effect of morphological and syntactic errors on the
holistic scores of native and non-native compositions. Unpublished master’s
thesis in TESL, University of California, Los Angeles.
Richards, J. C. (Ed.). (1974). Error analysis: Perspectives on second language
acquisition. London: Longman.
Schwabe, G. T. (1989). Appendix E: UC ESL students’ evaluation of their high
school English instruction: Examples of grammar problems in the writing of UC
ESL students. In English as a second language at the University of California (A
report to the University-Wide Committee on Undergraduate Preparatory and
Remedial Education from the ESL Subcommittee). Berkeley, CA: Office of the
President, University of California.

Another Educator Comments. . .

University of Southern California

It is significant that this question should be asked. Not long ago, it was
thought that formal grammar instruction was the only means of
developing second language competence. Current research, however,
shows that second language competence is developed in another way.
My interpretation of the research is that we acquire language by
understanding messages, by obtaining comprehensible input. Direct
evidence supporting the input hypothesis includes studies showing that
when acquirers obtain more comprehensible input, they acquire more of
the target language. This is the case both outside of school (exposure and
length of residence studies) and inside of school (method comparison
studies) and holds for both second language acquisition and the
development of literacy (Krashen, 1991).
There are also several serious problems with the hypothesis that direct
instruction plays a major role in developing language competence. It has
been argued that language is too complex to be deliberately taught and
learned, and there is evidence that people develop high levels of second
language competence without formal instruction (Krashen, 1991).
Does grammar study have any effect? My interpretation of the research
is that grammar learning does have an effect, but this effect is peripheral
and fragile. I have argued (Krashen, 1982) that conscious knowledge of
grammar is available only as a monitor, or editor, and that there are three
necessary conditions for monitor use: Performers need to know the rule,
have enough time to apply the rule, and need to be focused on form. When
these conditions are met, application of grammar rules can indeed result in
increased accuracy, but the performer pays a price in decreased
information conveyed, and a slower, more hesitant speech style. There are
other risks, such as editing one’s next sentence while the other person is
talking, which results in grammatically improved but sometimes


inappropriate speech, and, when rules are complex, diminished instead of
increased accuracy.
Current research confirms that the effect of grammar is peripheral and
fragile; this research shows that direct instruction on specific rules has a
measurable impact on tests that focus the performer on form, but the
effect is short-lived (e.g., Harley, 1989; White, 1991).1
Optimal use of the monitor occurs when application of conscious rules
does not interfere with communication. For most people, this means using
the monitor in writing, but delaying its use until all of their ideas are on the
page. Because of the fragility of conscious grammatical knowledge,
optimal users may also refer to handbooks occasionally. There may be a
few mental gymnasts who can remember many rules and monitor while
they speak, but I suspect that even these virtuosi rely mostly on acquired
knowledge and consciously monitor only a few aspects of gramrnar.2 Of
course, a significant number of students will not use the monitor at all, such
as young children, unschooled adults, and those who simply have no
interest in grammar.
If this view is correct, it implies that formal grammar teaching can be
done when students know the limits of conscious grammatical knowledge:
When they know it is not the major source of second language compe-
tence, when they understand that they will learn only a subset of the rules
of a language, when they understand the restricted function of grammar,
and when they understand when to use conscious knowledge of language.
1 lthas been claimed that direct instruction, when timed so that it is exactly at the acquirer’s
level of development, has positive effects (Pienemann, 1984, 1989). From the research
available to me, this so-called success is based on very little data. Pienemam rests much of
his case on the performance of one student (Giovanni) acquiring German as a second
language. Moreover, only one rule was taught (inversion), and Pienemann only provides
Giovanni’s percentage correct in using inversion just before and just after instruction; we are
not told how many times inversion was attempted. Another student “learned inversion in a
similar manner as Giovanni” (Pienemann, 1984, p. 197), but no details are provided other
than this assertion. Also, Pienemann claims that Giovanni generalized his learning of the
principles underlying inversion to another structure, but his mastery of this structure was “in
its very beginnings” (p. 205) and the “number of such instances is rather small” (p. 205).
Once again, no data is provided. Pienemann (1989) claims that instruction improved
accuracy in the use of the copula in three acquirers, but once again only percentages are
provided, and the effects appeared to be short-term.
White, Spada, Lightbown, & Ranta (1991) maintain that knowledge gained through
direct instruction in their study was not peripheral but was integrated into their subjects’
acquired systems because, after instruction, gains were found on an oral test as well as on
written tests. It is quite plausible, however, that conditions for monitor use were met on the
oral test. The oral test focused the subjects on rules they had just studied and subjects
probably suspected they were being tested on grammatical form (all other tests used
obviously focused the student on the target rules). In addition, White et al. do not indicate
that there was any time pressure on the test.
2 French-speaking children in an intensive ESL class in Lightbown (1991) showed
remarkably high accuracy in the English there is construction, compared to other classes,
and maintained their high level of performance when tested one year later. Lightbown’s
interview with their teacher revealed that she had placed special emphasis on this structure
(“‘I drummed it into their little heads’” [p. 207]). I suspect that these children, after so
much drill, simply specialized in this rule, monitoring it all the time. Interestingly, this class
was not significantly better- than comparisons in overall reading and listening (Lightbown &
Spada, 1990).


Although there is a role for grammar, research and theory show that the
best way of increasing grammatical accuracy is comprehensible input.
Studies also suggest that the most effective kind of comprehensible input
for advanced grammatical development is reading (Elley, 1991; Krashen,
1991). Getting students interested in books will insure continued
grammatical development (as well as improvement in vocabulary,
spelling, and writing style) long after the language course ends.

I thank Joe Allen and Eula Krashen for their helpful comments on earlier versions
of this paper.

Stephen D. Krashen is Professor of Education and Linguistics at the University of
Southern California.

Elley, W. (1991). Acquiring literacy in a second language: The effect of
book-based programs. Language Learning, 41, 375-411.
Harley, B. (1989). Functional grammar in French immersion: A classroom
experiment. Applied Linguistics, 10, 331-359.
Krashen, S. D. (1982). Principles and practice in second language acquisition. New
York: Prentice Hall.
Krashen, S. D. (1991, April). The input hypothesis: An update. Paper presented at
the Georgetown University Round Table on Languages and Linguistics,
Washington, D.C.
Lightbown, P. (1991). What have we here? Some observations on the influence of
instruction on L2 learning. In R. Phillipson, E. Kellerman, M. Sharwood Smith,
&M. Swain (Eds.), Foreign/second language pedagogy research (pp. 197-212).
Clevedon, England: Multilingual Matters.
Lightbown, P., & Spada, N. (1990). Focus-on-form and corrective feedback in
communicative language teaching: Effects on second language learning. Studies
in Second Language Acquisition, 4, 429-448.
Pienemann, M. (1984). Psychological constraints on the teachability of languages.
Studies in Second Language Acquisition, 6, 186-214.
Pienemann, M. (1989). Is language teachable? Psycholinguistic experiments and
hypotheses. Applied Linguistics, 10, 52-79.
White, L. (1991). Adverb placement in second language acquisition: Some effects
of positive and negative evidence in the classroom. Second Language Research,
7, 133-161.
White, L., Spada, N., Lightbown, P., & Ranta, L. (1991). Input enhancement and
L2 question formation. Applied Linguistics, 12, 416-432.