DOCUMENT RESUME

ED 366 184 FL 021 742

AUTHOR Long, Michael H.

TITLE Task, Group, and Task-Group Interactions.
PUB DATE 90
NOTE 22p.; In Sarinee, Anivan, Ed. Language Teaching
Methodology for the Nineties. Anthology Series 24;
see FL 021 739.
PUB TYPE Viewpoints (Opinion/Position Papers, Essays, etc.)
(120) Speeches/Conference Papers (150)

EARS PRICE MF01/PC01 Plus Postage.

DESCRIPTORS *Assignments; *Class Activities; Classroom
Techniques; Elementary Sec,,naury Education; *Group
Dynamics; *Group Instruction; Second Language
Instruction; *Second Languages; Teaching Methods.

ABSTRACT
It is proposed that "method" is not a relevant
construct for fostering change in the second language classroom, and
that concern with specific methods can divert educators' attention
from important instructional issues. Teachers usually organize
instruction around tasks rather than methods, and therefore it would
be more useful to analyze, research, and evaluate the tasks to
discover the features that promote language teaching and learning.
However, it is argued that for this to be a successful approach,
careful attention must be given to the judicious use of group work,
to the kinds of tasks teachers and learners work on, and to optimal
combinations of task type and group. (MSE)

************************* *A*************************************
Reproductions supplied by EDRS are the best that can be made
from the original document.
**********************************************************************
TASK, GROUP, AND TASK-GROUP INTERACTIONS

MICHAEL H. LONG

i'EP.IMI;,',;ON !kJ Hi 1,115

MA U S DEPARTMENT OF EDUCATION
IlIAL HAS BEL N t,HAN1E 1:0 Office of Educefierup Research and Improvement
EDUCATIONAL RESOURCES INFORMATION
CENTER (ERIC,
Tros document nas been reprOduced as
dfrotrn the person dr 0,ganaatron
I.4,nor changes have been made to .rnerove
reproducfion Sooty
7:.. flif
I,' ',QR.. A luNat RLSCUI-ILFS Po.nts Of dr OP,,Ons stated ,n !his docu
Ni OW A I CIN Ltsrt-.H RR men! do CM necessardy represent Officral
OE RI 0,,S.t.en 0, psoltry
TASK, GROUP, AND TASK-GROUP INTERACTIONS
Michael H Long

1 AGAINST METHODS
Despite the range of important issues to consider when planning a language
teaching program, books, journals and conference programs in our field reveal a
pervasive preoccupation with methodology. Not all the claims made are about
this "brand name" method or that (Audio-Lingual Method (ALM), Grammar
Translation, Silent Way, Suggestopedia, Counselling Learning, the Natural
Approach, etc), but a large percentage deal with procedural issues in classrooms
in one way or another. Much less attention overall is given to such areas as
syllabus design, testing and evaluation, despite the fact that a lot of scrious work
has been taking place there, some of it producing quite radical innovations.
Understandably, people want to know "how to teach", and as is usually the case
when demand for a product is high, there are any number of people ready to tell
them. Tips are plentiful, and can be quite useful on the rather rare occasions
they have been evaluated first.
Underlying all the prescriptions and proscriptions about how to teach, all
the books and articles advocating particular methods or reviewing methods arc
two basic assumptions. Onc is that discussion of methods makes a difference in
the classroom. Yet it might bc, for instance, that method is a useful construct in
coursework in graduate level teacher education programs (although I know of no
evidence of that), but fail to translate into changes in what teachers and learners
actually do on Monday morning. A second, more fundamental assumption is
that methods exist, outside books and discussions about methods, that is. Four
bodies of evidence suggest that neither assumption is correct.

1.1 Prescribed Overlap

Even if implemented exactly as their inventors prescribe, methods overlap a

good deal. Such superficially quite different methods as the ALM, the (Struc-
tural-Situational) Audio-Visual Method and the Silent Way (the first two of
which, along with Grammar Translation, arc probably still the most widely
used), share much in common. All three arc predominantly teacher-centered,
iuse structural grading. provide minimal input (usually just a few model sen-

31 3
BEST COPY AVP1LAM.E._
tences), attempt to elicit immediate production by learners of native-like target
language constructions, prescribe "error correction" when the attempts fail (as
they must), devote a majority of classroom time to (at best) pseudo communica-
tion, and assume communicative abilities evolve out of grammar, rather than the
other way around. All three, that is, like most other methods and the vast major-
ity of commercially published textbooks, proceed with complete indifference to
the findings of twenty years of research on naturalistic and classroom language
learning.

1.2 Lack of Conceptual Utility for Teachers

Numerous studies show that teachers of languages and other school sub-
jects plan, conduct and recall their lessons, not in terms of methods, but rather
as ser,uences of instructional activities, or tasks (for review, see Shavelson and
Stern, 198l; Crookcs, 1986). Such was the finding, for example, of an evaluation
by Swaffer, Arens and Morgan (1982) of "comprehension" and "four skills"
approaches to the teaching of German as a FL at the University of Texas.
Despite having given teachers explicit training in the different methods, and
despite the teachers then (supposedly) having taught using one or the other for a
semester, Swaffer ci al found through classroom observations and debriefing
interviews at the end of the study that there was no clear distinction between the
methods in the minds the two groups of teachers or in their classroom practices.
They conclude:

"... defining methodologies in terms of characteristic activities has led to

distinctions which are only ostensible, not real, i.e. not confirmable in
classroom practice ... Apparently, any analysis of methodologies needs to
commence with definitions of task, order, and learning strategies. This is
the way we as foreign language teachers interpret the pragmatics of the
classroom." (Swaffer, et al 1982: 32)

1.3 llomogeneity of Observed Classroom Procedures

Classroom observational studies consistently show very little difference in

what teachers actually do, as opposed to what they have supposedly been trained
to do and/or think they arc doing. The same practices are reported across all
kinds of classrooms despite differences in such factors as the "methods" teachers
have been trained in (Dinsmore, 1985; Nunan, 1987), the theoretical orientation
of the professional training they have received at the masters degree level and
profess to hold (Long and Sato, 1983), the materials they are using (Phillips and

32
4
Shettlesworth, 1975; Long, Adams, McLean and Castanos, 1976; Ross, in press),
teAching generations (I-Ioetker and Ahlbrand, 1969) and teaching experience
(Pica and Long, 1986).

1.4 Null Findings of "Comparative Methods" Studies

Large-scale "comparative methods" studies have typically found either

short-lived differences or no difference in the relative effectiveness of (supposed-
ly) quite different methods, e.g. Grammar-Translation, ALM and Cognitive
Code Learning (Scherer and Wertheimer, 1964; Smith, 1970) and inductive and
deductive approaches (von Elek and Oskarsson, 1975; Se tiger, 1975). The
reasons for the null findings are impossible to ascertain due to the absence of a
systematic observational component in most of the studies, but at least three
interpretations arc possible. One is that, while at least some methods can be
clearly differentiated in practice, the teachers in the different treatment groups
in the comparative methods studies were either simply doing the same things or
differing only in the relative frequencies with which they exhibited the same
behaviours, as has since been found to be the case in so many of thc studies
which have documented what goes in inside classrooms (e.g. Spada, 1987). A
second possibility is that methods exist hut do not matter. A third, more radical
view is that methods do not matter because they do not exist, at least, not where
they would matter if they did exist, in the classroom.
The overlap in prescribed and proscribed practices noted earlier would be
consistent with this last reading. In addition, there must be a blurring of distinc-
tions due to thc need for pedagogic variety in lessons. It is no doubt possible to
maintain potentially important differences, such as the provision or witholding of
feedback on form, for some time, e.g. the duration of a public demonstration
lesson, and fot short periods to accentuate salient (but as far as we know, psy-
cholinguistically trivial) idiosyncracies, e.g. whether feedback is provided verbally
or via hand signals. Ilowever, there must also be a natural tendency over the
course of a semester or a year for teachers to exploit most of the rather limited
range of procedural possibilities, rathcr than stick to a narrower prescribed set
of options.

15 From Methods to Methodology

In sum, there really seems to be very little justification for the continuing
debate about methods, let alone for the hunt for the single correct one. As far as
we know, 'method' is an irrelevant construct when attempting to influence e:ass-
room language teaching. Worse, it may actually be counterproductive if it di-

33 5
ens us from issues which really do makc a difference, among which, of course,
arc the many options available in methodology. Methodology is here defined
broadly as the instructional strategies and learning processes employed by both
teachers and learners in performing tasks which they engage in separately, in
groups or as a whole class.
As numerous studies have shown, classroom processes do make a differ-
ence. First, they affect other classroom processes. The kinds of questions
teachers ask affect the syntactic complexity and communicative potential of
students' speech (Brock, 1986; Tollcfson, 1988), for example. The kinds of
"simplifications" employed in listening and reading materials affect student
comprehension (Parker and Chaudron, 1987), and so on. More important in the
long run, they affect at least some (presumably many) aspects of learning, al-
though relatively little is known about learning consequences as yet (for review,
see ('haudron, 1988). The question that arises, however, is what a relevant unit
of analysis may be for examining and, where needed, altering these processes if
"method" is not that unit, and what intervention points (Long and Crookes, 1986)
we can identify to engineer such changes. 1 would like to claim that task is a
viable candidate as the unit of analysis, and that task-group interactions consti-
tute one of several potential intervention points suggested by classroom research.

2 METHODOLOGY IN AN INTEGRATED APPROACH TO

PROGRAM DESIGN

2.1 The Need for Compatibility with Other Domains

Most applied linguists would agree that there are six major areas to consid-
er in the design of a successful language teaching program: needs (and means)
identification, syllabus, materials, methodology, testing and evaluation. Of these
it can bc argued that the most important is syllabus design, and that within sylla-
bus design, as elsewhere, the central issue is choice of the unit of analysis: word,
structure, notion, function, topic, situation or task (for review, sec Long and
Crookes, 1989). The unit.selected is crucial for two reasons: first, because it
closely reflects the program designer's and teacher's theories, implicit or explicit
(Ramani, 1987), about second language learning, the process programs arc
designed to facilitate, and second, because the choice made affects decisions the
designer takes in all thc other five domains. 'Logically should affect' would
perhaps be more accurate, since many poorly designed programs exist where
theoretically incoherent options were seleck2.d. Task-based syllabuses and mate-
riak, for example, may be taught using classroom procedures, such as pattern
drills and transformation exercises, which involve structurally graded language
practice. Similarly, a needs identification may be carried out to identify the tasks

34
required in a particular occupation learners are preparing for, yet the syllabus
be based not on what the needs identification says abcut the learners' needs but
on what a linguistic analysis says about the target language's structures, notions
or functions.

2.2 Task as the Unifying Unit of Analysis

An extensive rationale for selection of task as the unit of analysis in course

design has been presented elsewhere (see e.g. Long, 1985; Long and Crookes,
1989), where a distinction is drawn between 'target tasks' and 'pedagogic tasks'.
Target taskr are the things the learners will eventually do in English, at school or
university, at work, in a vocational training program, on vacation, and so on - a
non-technical, non-linguistic definition. In task-based language teaching as
described by Long and Crookes, which target tasks are relevant for particular
groups of learners is established by a task-based needs identification. After
classification of the relevant target tasks into (target) task types, pedagogic tasks
are derived and sequenced to form a task syllabus. Pedagogic tasks are the
problem-solving activities teachers and learners work on in the classroom.
Especially in the early stages, they arc usually simpler approximations to thc
target tasks that have motivated their selection, not just linguistically, but also in
terms of the substantive content of the task, the number of steps the learners
have to take, the options they have to choose from, etc.
The rationale for choosing task as the unifying unit in program design, will
not be repeated here. Suffice to say that most other potential units, including
word, structure, notion and function - and synthetic syllabuses (Wilkins, 1972)
and Type A syllabuses (White, 1988) in general - do not sit well with what is
known about second language learning. There is no evidence that the commonly
employed target language units in such syllabuses make meaningful acquisition
units. Nor is there any evidence, contrary to what is assumed by synthetic, type
A syllabuses and materials, that structures, notions, functions, etc can be ac-
quired separately, singly, in linear additive fashion, or that they can be acquired
prior to and separate from language use. There is overwhelming evidence
against all thosc ideas, in fact. (For reviews, sec e.g. Hatch, 1983; Ellis, 1985;
Larsen-Freeman and Long, in press.) As Long and Crookes (1989) put it:

"(L)anguage learning is a psycholinguistic process, not a linguistic one, yet

synthetic syllahuscs consistently leave the learner out of the equation."

By way of contrast, analytic syllabuses and Type B syllabush (Wilkins, 1972;

White, 1988) in general, and those utilizing 'task' as the unit of analysis in partic-
ular, are at least potentially compatible with universal acquisition processes. To

35
give just one example, tasks can be combined with methodological options which
.allow for, but speed up, learners' progress through the obligatory stages in inter-
language 'development sequences', e.g. a short-term orientation to task accom-
plishment, not language accuracy, but with a focus on form when certain condi-
tions are met (Long, 1988a, 1988b). Developmental sequences have. bcen well
documented by second language acquisition researchers for such phenomena as
word order, negation, interrogatives, articles, auxiliaries and relative clauses (see,
e.g. Johnston, 1985), as has the inability of formal instruction to alter them in any
fundamental way (sec, e.g. Pienemann and Johnston, 1987; Ellis, in press).
Task-based syllabuses are also an advantage for those seeking an integrated
approach to course design. They are compatible with task-based needs identifi-
cations, which are relatively easily conducted and more likely to be valid than
identifications using linguistic units (for details and examples, see Long, 1985).
They also combine well with communicatively oriented, task-based methodology.
Indeed, Nunan (1989) has argued that the use of tasks tends to make the tradi-
tional syllabus/methodology split redundant.

...the distinction between syllabus design and methodology becomes diffi-

cult to sustain: one needs not only to specify both the content (or ends of
learning) and the tasks (or means to those ends) but also to integrate them.
This suggests a broad perspective on curriculum in which concurrent con-
sideration is given to content, methodology and evaluation.
(Nunan, 1989- 15)

Tasks lend themselves to stimulating, intellectually challenging materials, espe-

cially those of a problem-solving nature, and as noted earlier, of a kind which
seem meaningful to teachers planning and implementing lessons. They are well
evaluated with criterion-referenced tests, and the kind of tangible products typi-
cally associated with task achievement should be attractive to program evaluators
and consumers alike.
Needless to say, amidst all the advantages there are also some problems.
These include establishing valid criteria for the selection and sequencing of
pedagogic tasks (a problem with other units of analysis, too, of course), and
various aspects of evaluation (Das, 1984). Further, as is well known by now,
tasks of one sort or another have provided the basis for three distinct syllabus
types: procedural (e.g. Prabhu, 1987), process (e.g. Breen, 1984, 1987) and task
(e.g. Long, 1985; Long and Crookes, 1989), for some of which thc "advantages"
of tasks listed above would not be considered relevant at all. Some task-based
syllabuses (e.g. Prabhu's procedural syllabus) arc not derived from analyses of
learners' needs, for example, much less analyses in terms of real world 'target
tasks'; nor do they make a distinction between 'target tasks' and 'pedagogic
tasks' (sec Long, 1985, and for discussion, Nunan, 1989; Long and Crookes,

36
1989). Despite the brief history of task-based syllabuses, in fact, 'task',
'task-based' and 'task syllabus' already have a wide Variety of uses and mis-uses.
Most obvious in the latter category, several reccnt syllabuses and commercially
published textbooks which claim to be 'task-based' are nothing of the sort, at
least not in any of the senses outlined above, in which 'task' is the unit of analysis
in at least some areas of a language teaching program. In some which even
advertise themselves as structurally graded, 'task' is just a new word for 'exer-
cise'.

3 GROUP WORK AND PEDAGOGIC TASKS

Differences in the various conceptions of task aside, some important ques-
tions for all those utilizing pedagogic tasks are the grouping of participants
(teachers and/or learners) who work on tasks, the types of tasks they work on,
and task-group interactions. We will take them in order.

3.1 Group Work

At least three basic groupings of interlocutors arc possible in classrooms:

individuals, groups (including dyads), and whole class, i.e. the teacher-fronted
"lockstep" format, in which everyone (supposedly) attends to the same thing at
the same time. All three arrangements undoubtedly have unique qualities and
advantages, and our ultimate goal should be to ascertain empirically which ones
serve which purposes best. They should be viewed as complementary, in other
words, not in competition. That said, it is well known that individual and whole
class wurk predominate the world over, often to the complete exclusion of group
work, which is why it is easy to find oneself appearing to advocate group work "in
preference to" the others. While badly organized group work is no better than
badly organized lockstep work, group work is a very valuable but widely neglect-
ed asset, and also, it turns out, important for exploiting certain types of task. It is
therefore worth briefly summarizing some of its general strengths before moving
to the more subtle issues of task type and task-group interactions.

3.1.1 A pedagogic rationale

All other things being equal, group work (including pair work) has at least
five major pedagogic benefits. (1) Group work increases the quantity of lan-
guage practice opportunities. (2) Group work improves the quality of student
talk in several ways. T'-- can engage in what Barnes (1976) calls "exploratory"
talk, and practice a functionally wider speech repertoire. (3) Group work

37
 helps individualize instruction, potentially allowing students to work at their own
pace, perhaps using different materials. (4) Group work can help improve the
affective climate in the-classroom, the intimacy of the small group setting often
being especially valuable' to shy or linguistically insecure students. Finally, (5)
group work can help motivate learners because of the advantages referred to in
(1) through (4) and because of the pedagogic variety it brings to a lesson.

3.1.2 A psycholinguistic rationfile

A psycholinguistic rationale for group work has also been proposed. It is

noted that the precision with which input can be adjusted to an interlocutor's
comprehension abilities is likely to be greatly improved when thc listener (or
reader) is an individual (the other member of a dyad, for example) than a large
group of people of inevitably differing proficiencies, i.e. the whole class. The
more individualized negotiation for meaning which is possible in the small group
format, in other words, should increase both the quantity and quality of compre-
hensible input available to students. There is, in turn, a variety of evidence for
the necessity (although not, I believe, the sufficiency) of comprehensible input
for language learning (for review, sec Krashen, 1985; Long, 1981).
Classroom studies have shown that the negotiation work learners accom-
plish while talking together in unsupervised small groups (interlanguage talk)
does not involve a decline in grammatical accuracy compared with the same
learner's performance in lockstep work, which is exactly what would be expected
from our knowledge of how interlanguage develop. The amount of negotiation
work achieved in interlanguage talk is also greater than that in either lockstep
work or nativc/non-native speaker conversation in pairs. Finally, classroom
studies have found that the frequency of other-correction and completions by
students is higher in group work than in lockstep teaching and not significantly
different from that observed in native/non-native speaker conversations. Group
members almost never miscorrect, and there is minimal incorporation of other
students' errors. (For details and a review of studies, see Long and Porter,
1985.)
An important additional dihitnsion Co the psycholinguistic rationale has
been provided by Bygate (197fi, who begins with the observation that in conver-
sations inside and outside claAssliioms, a good deal of spontaneous native speaker
oral language production occtik not in finite sentences, hut in what he calls
"satellite unitls" (SUs). SUs are. defined as moodless utterances which lack a
finite verb gr Viands up", "Pencils down") and all other syntactically de-
pendent units,' lte or non-fitp , that have been uttered in a turn which either
(a) djes not inc Ude a main clause to which the unit in question may be
attached ("The man is riding a.- 4, , left for a student to complete), or (b)

38
14. 4. 10
t
Sminmegai
 .....

includes a related main finite clause, but one for which the dependent unit is
syntactically superfluous, as when someone makes parenthetic additions or alter-
ations to parts of an independent finite clause ("It was a boring paper, a long
boring paper ... a dreadfully boring paper, one of his worst yet"). He cites
numerous examples of SUs from a classroom study of group work to illustrate
his claim, noting that they can consist of any dependent syntactic element, such
as

Prepositional phrases:
SI: at the door
S2: yes in thc same door I think
SI: besides the man who is leaving
S2: behind him

Verb groups:
S3: and the point is that we can start
S4: compare
S3: yes

Subordinate clauses:
S2: well that man 1 think he is a robber, a thicf
SI: he might be
S2: because he is running with a handbag
SI: yeah (examples from Bygate, 1987: 68)

Bygate points out that teacher-fronted, textbook driven oral practice is

traditionally "clause-down", and advocates a reversal if thc work is to promote
language learning. Students need time and opportunities to explore thc ways
dependent SUs can be formed and then combined to make clauses. Intra-turn
repairs and cross-turn cooperative dialog, especially as induced by the need to
negotiate meaning while working on a problem-solving task in a small group,
provide both time and a place a view which fits very well, of course, with claims
by Hatch ( i978) and others to thc effect that syntax develops out of conversation,
rather than the other way round. It is not that talk containing SUs is impossible
in principle in lockstep work. It is just that descriptive studies show it rarely
occurs there, and that classroom experiments comparing the same tasks in
teacher-fronted and small group formats consistently find the small group setting
produces significantly more of it (Long, Adams, McLcan and Castaiios, 1976;
Dorghty and Pica, 1986; Rulon and McCreary, 1986; Deen, 1988).
The existence of both pedagogic and psycholinguistic rationales does not
mean, of course, that all group work is valuable. The small group setting may
simply he used to increase the quantity of work done on a useless task. Aston

39
1.
 activi-
(1986), for example, has pointed out how poorly designed problem-solving
but
ties can lead to a lot of negotiation work (what he calls "trouble-shooting),
frustration with too difficult a task, unshared
work which may reflect learner
participant backgrounds and a need to enhance rapport rather than a successful
In other
attempt to secure more appropriate input for acquisition purposes.
words, valuable though group work is, especially but not only in large classes, the
specified, i.e. until format
term itself has no real meaning until the 'work' done is
is linked to task.

3.2 Pedagogic Task Typcs

in the
Embryonic taxonomies of pedagogic tasks have begun to appear
task diffi-
second language literature, along with various proposals for assessing
culty and a even a list of (twenty) qualities of "good tasks" (for review, see
With few excep-
Crookes, 1986; Candlin, 1987; Nunan, 1988a, 1988b, 1989).
task types
tions, the proposals make no reference to the classroom research on
has been well
(or to any research findings at all, for that matter), although there
these things tend to
over a decade of such work on the topic. Given the way eventually
work in our field, it is safe to assume that many others will follow,
leading to a data-free argument about whose taxonomy or list is "best".
Most classifications so far have been pedagogic. While pedagogic criteria
surely be of limited
will eventually be important for materials design, they must
properties of tasks. I
utility until we know how they relate to the psycholinguistic
productive approach would be to continue to
would iike to suggest that a more
structural features of
search for objectively and (preferably easily) recognisable
generalizability - which can
task types - task types, not tasks, or there will be no
psycholinguistic proper-
be demonstrated to relate systematically to the relevant
learning that we
ties, since it is the kinds of task which promote second language
but at least
need to identify. There are few clear findings of this sort as yet, relating
three lines of work have begun to produce interesting results: studies
work, and to inter-
task type to quantity and quality of interlanguage negotiation
language complexity and destabilization.
'relevant', of course,
Just which of the many psycholinguistic properties arc
according to one's
and so which studies are considered 'interesting', will vary several which
views about how people learn languages. Two properties among
task type have for
interest me are the following. (' ) What potential does a
stimulating (both) teachers
encouraging negotiation work, and in particular for
and (b) to
and learners (a) to reformulate their own and others' utterances
their performance in general, not just on crrors)?
attend closely to feedback (on
(2) What is a task type'spotential for "stretching" learners' interlanguages, for
especially to
pushing them to operate at thc outer limits of their current abilities,

4019
use (a) as linguistically complex speech as possible, and (b) as much optional
syntax as possible (where each is sociolinguistically appropriate), and in these
and other ways, to expose their interlanguages to constant pressures for destabi-
lization? Not all task types are equally useful in either of these areas. I will illus-
trate with just three examples of such relationships, although many other pat-
terns arc emerging from a rapidly expanding body of research. It should be
stressed that both types of pedagogic tasks in each of the following pairs may still
be useful in the classroom, even if one type is more useful than the other in the
ways of interest here.
Where both negotiation work and interlanguage "stretching" and destabili-
zation are concerned, evidence from classroom studies is generally consistent
with the following three generalizations, assuming variables other than those
mentioned are held constant in each comparison.

3.2.1 Two-way tasks produce more negotiation work and more useful negotia-
tion work than one-way tasks

The one-way/two-way distinction (Long, 1980) refers to the way informa-

tion is distributed at the outset of a task and the requirement that the structure
of the task imposes on participants to exchange that information if they are to
complete the task succesfully. It is insufficient that information exchange can
facilitate or improve task complebon; for a task to be two-way, information
exchange must be required for completion to be possible at all. One-way/two-
way also has nothing to do with the number of participants. Nor is two-way the
same as "information gap". One-way ard two-way tasks are both information
gap tasks, as that term is used in the pedagogic literature, but research has
shown that it is two-way tasks, that are more conducive of negotiation work, for
which many one-way tasks, and hence many information gap tasks, are quite
useless, it turns out. Several studies support this generalization: for NS/NNS
conversation, see, e.g. Long (1980), and for interlanguage talk, e.g. Doughty and
Pica (1986).
An example may help clarify the distinction. A task in which one person
(teacher or student) describes a picture which only he or she can sec so that
others can draw it is one-way. A task ;n which each member of (say) a four
person group has exclusive access to information about a crime, all of which
must be pooled before a villain can be identified, is two-way. For example, one
student might hold information about the motives or lack of motive of six sus-
pects, a second about their alibis, a third about thc way the crime was commit-
ted, and a fourth about certain externally verified facts which, taken together,
support some alibis but not others, and so on. None of the separate pieces of
information is interpretable without the others, meaning that the group must

41 13
work cooperatively to exchange their information if the crime is to be solved.

3.2.2 Planned tasks "stretch" inter-languages further and promote destabilization

more than unplanned tasks

Building on work by Ochs (1979) on linguistic differences between planned

and unplanned native language discourse, Ellis (1987) and Crookes (in press)
have manipulated degree of planning to assess its effects on written and oral
work in a second language. Ochs found planned discourse, such as prepared
lectures or expository texts, typically to contain more complex language and a
wider variety of linguistic constructions (more relative clauses, noun modifiers,
passives, and so on) than unplanned discourse,,such as informal face to face
conversations and personal letters. The L2 studies to date have shown that
learners produce syntactically more complex language when given planning time
than when performing the same tasks without planning or with less planning
time.
Using a counterbalanced repeated measures design, Crookes (in press)
studied the monologic speech of forty Japanese learners of English on two oral
production tasks involving instructions on how to assemble Lego pieces to make
a model house and thc siting of buildings on the map of a town. All subjects
compleied one of two equivalently complex versions of each task. Half the
subjects did one task after a ten-minute opportunity to plan the words, phrases
and ideas (but not the exact sentences) they would use, and the other task after
no opportunity to plan. The other subjects reversed the order of tasks and
conditions. Crookes was careful to obtain evidence that learners actually had
used the alloted time for planning in the form of writtcn notes they were in-
structed to make but which were removed before they recorded their instruc-
tions. That is, he made sure the treatment had been "delivered.% As predicted,
Crookcs found that the learners' speech was syntactically more complex on
various measures, e.g. words, subordit,Ate clauses and s-nodes per utterance, on
both tasks after planning time. He also noted trends towards more target-like
use of particular linguistic forms (articles) and use of a greater variety of words
(higher type-token ratio) under the planning condition.
Crookes is cautious about drawing pedagogical implications from what was,
after all, a tightly controlled laboratory experiment. He suggests, however, that
whereas many writers on communicative language teaching (for good reasons)
advocate provision of spontaneous language practice opportunities, teachers
(and materials designers) might well consider systematically building planning
opportunities for at least some tasks into their lessons, given the evidence that
the same students on the same day can operate at a higher level, both quantita-
tively (the linguistic complexity measures) and, it appears likely, qualitatively

42
.14.
(developmentally more advanced in terms of percentage target-like use) if given
planning opportunities. It seems quite reasonable to assume that, other things
being equal, learners will improve faster if they engage in language work nearer
the upper bounds of what they are currently capable than practice at levels below
their current capacity.
This finding is of a different order from those concerning one-way and two-
way tasks and other task types because it concerns a quality, degree of planning,
which can in principle be manipulated, and fairly easily, for vii tually any task
with (potentially) the same results. One wonders what other features and condi-
tions might be superimposed in this way to alter tasks in the classroom, possibly
thereby creating new task types with significance for language performance and,
presumably, language learning.

3.23 Closed tasks produce more negotiation work and more useful negotiation
work than open tasks

Unlike the previous two claims, which have each been explicitly formulated
and then tested in a number of studies, the proposal I am about to make con-
cerning the relative merits of 'closed' and 'open' tasks reflects my own post hoc
interpretation of a number of results, and should therefore be treated more
cautiously. It has not, to my knowledge, been
addressed in a second language
study thus far. The argument, briefly, is as follows.
Negotiation for meaning is usually both fun and intellectually stimulating
for teachers and learners alike if the materials writei is clever enough. It can
also be hard work, however, most obviously when a task is too difficult for a
particular group of learners in one or more ways. A least effort outlook will
mean students (and some teachers, I suppose) will tend to avoid negotiation if
the task itself does not demand it. Some tasks, even within other categories, such
as one-way and two-way, elicit more negotiation work than others, some less,
and some aspects of negotiation are probably more beneficial for language
development than others.
The last point is well illustrated in research findings by Pica and her associ-
Pica (1987) noted a tendency for NS
ates. In a study of NS/NNS conversation,
interlocutors to model corrcct versions NNSs problematic utterances as
of
confirmation checks following communicative trouble. While potentially very
valuable in some respects, the seeming disadvantage was that the NNSs then had
only to acknowledge in order to complete the discourse repair, rather
than to
attempt their own reformulations, as shown in this example:

NNS: 1 many fren

NS: You have many friends?
NNS: Ycs

43 15
Pica, Holliday; Lewis and Morgenthaler (in press) found clarification requests,
on the other hand, to be more successful at eliciting reformulations from learn-
ers, especially on tasks in which they had some control over the topic, a condition
which can be built into a (two-way) ta64 if the designer wishes, of course. Pica
provides the following example:

NNS: ... you have a three which is ...

white square of which appears sharp
NS: Huh?

NNS: ... you have a three houses ...

one is no-no-not-one is not square
and one is square, but with a little
bit- a little small house

"Free conversation" is a particularly poor task in most respects where nego-

tiation work is concerned, as can be seen in the consistent finding of a tendency
for speakers in free conversation to treat topics briefly, to drop them altogether
when serious trouble arises, to provide feedback to their interlocutors less often,
to incorporate feedback from their interlocutors less often, and to recycle lin-
guistic material less often than when the same speakers work together on various
other "problem-solving" tasks.
Quantified demonstrations of these relationships can be found, among
other places, in a study comparing the performance of the same fifteen dyads in
free conversation and working on two so-called "problem-solving" tasks, Spot
the Diffc -ce and Odd Man Out, by Crookes and Rulon (1985, 1988), who also
provide a detailed discussion of possible casual relationships. Related findings of
higher quantity and quality negotiation work have been reported favouring
"convergent" over "divergent" tasks by Duff (1986), and (social, cooperative
problem-solving) tasks combining "non-teaching goals and experiential process-
es" (e.g. construction of a Lego toy) over tasks emphasizing "teaching goals and
expository processes" (e.g. instruction in the string-searching function of a lap-
top computer with the computer physically present) by Berwick (1988). Ber-
wick's is the largest scale, most comprehensive study of these issues to date, and-
involved careful manipulation (singly and in combinations) of several variables,
including task types, native language of teacher and learners, and the availability
of visual support for tasks. There are several other relevant studies, too many
to review here, and I have not done justice to the wealth of detail available even
in those few I have mentioned.
The claim I would like to make is that a distinction between "open" and
"closed" tasks would account for many (although by no means all) of the findings,
and would be worth manipulating in a future study. The distinction is a modifi-

44
1G
cation of one for a classification of questions first proposed, I believe, by Robin-
son and Rackstraw (1972), applied now to the classification of pedagogic tasks.
By an open task, I mean one in which participants know there is no prede-
termined correct solution, but instead a wide (in some cases, infinite) range of
acceptable solutions. Free conversation, a debate, ranking favourite leisure time
activities, explaining how something works (how you think it works, with no form
of "test" of your interlocutor's competence after your explanation - not necessari-
ly how it really works), and discussing and eventually choosing (individually or by
consensus) the ten greatest world figures, would all be examples of 'open' tasks.
By a closed task, I mean one in which the task itself (as opposed to some
construal put on it by the participants) requires that the speakers (or listeners,
readers and writers, of course) attempt to reach either a single correct solution
or one of a small, finite set of correct solutions determined beforehand by the
designer of the task and again (crucially) known to the participants to have been
so determined. There may only be one possible correct answer to who commit-
ted the crime, for example, exactly four differences between two otherwise iden-
tical pictures, only three countries out of ten whose GNP rose every year from
1975 to 1984, and so on. It is crucial that participants know whether the task is
open or closed.
The idea is that the quantity and quality of negotiation for meaning will be
higher on closed tasks, when participants know that task completion depends on
their finding the answer, not settling on any answer they choose when the going
gets rough and moving on to something else. The prediction is that, all other
things being equal, closed tasks as defined above will elicit more topic and
language recycling, more feedback, more incorporation, more rephrasing, more
precision, and so on. These adjustments involve the kinds of reformulations
noted earlier and are likely to lead to provision and incorporation of feedback,
and hence, to interlanguage destabilization.

3.3 Task-group Interactions

In a study of Mexican university students of EFL working on (supposedly)

"communicative" materials written by the research team, Long, Adams, McLean
and Castarios (1976) compared language use in teacher-led, whole class discus-
sions and in unsupervised pair work when participants were engaged in solving
the same problem. This was to decide which of a list of characteristics (tool use,
thought, etc) wcrc unique to humans (I am ashamed to admit we wrote 'unique
to man' in 1976) and which ones could be found in other animals. Contrary to
our predictions, the materials had no apparent effect on the kinds of language
use that transpircd in thc whole class discussions. Teachers continued to ask
display questions, correct errors, drill target language constructions, and general-

45

7
ly to focus on forms, not communication. In the small group condition, however,
which involved the student pairs doing the same task in an adjoining room at the
same stage in the lesson, the materials seemed to "work". Students produced
more talk, all of it with a focus on meaning, not language, a functionally wider
range of talk, and more "exploratory" talk in Barnes' sense. Wc concluded,
tentatively, that it was thc combination of materials and grouping that had
produced the result.
Similar findings have since been obtained in at least one other study. Pica
and Doughty (1985) and Doughty and Pica (1986) compared various features of
teacher and student talk, focussing primarily on negotiation work, on one-way
and two-way "decision-making" tasks conducted in small group and teacher-
fronted lockstep formats. Like other researchers (c.g. Porter, 1986), they found
student speech was equally grammatical in both formats (as measured by the
percentage of grammatical T-units), and that the students talked more and
provided more other corrections and completions in the small groups. The two-
way task involved each student planting flowers on a feltboard garden to which
only he or she had access and which differed slightly from every other student's
board, the object bcing for everyone to finish with the same final picture. Pica
and Doughty report that the two-way task produced significantly more negotia-
tion work than the one-way task in the small group setting, but found no effect
for task type in the teacher-led lessons. When task type was held constant, sig-
nificantly more negotiation work (the ratio of conversational adjustments to total
T-units and fragments) was found in the small groups (four person groups and
pairs) than in the lockstep, but differences between thc pairs and the four person
groups themselves were not significant.
On the basis of these two sets of results, it would seem that the amount and
quality of language practice can sometimes depend not simply on the tasks or
format employed, but, at least where some tasks and possibly some task types
arc concerned, upon the interaction of either task or task type and grouping.
Both studics find the combination of communicative task with small group set-
ting necessary to bring out the full potential of the task itself, and both find that a
task's true potential may not he realised at all in a lockstep format.
How generalizable are these findings? At this point, we simply do not
know. We need further studies of task-group interactions. It may be that well
designed tasks are protected against the effects of one grouping arrangement or
another. That would be the optimistic view, certainly. It might also turn out that
these two studies were providing an early warning of a phenomenon that vie
would do well to investigate further. It would be a shamc, after all, if we spent
the next few years learning to do clever things with tasks only to have the effects
of our work unintentionally preempted by the way those tasks were used in the
classroom.

4 6 13
 4 CONCLUSION
There is no evidence that "method" is a relevant construct for those inter-
ested in fostering change in classrooms. Worse, a concern with "methods" can
divert us from methodological issues, which clearly arc important. Methodology,
however, will be treated more effectively as part of an integrated approach to
program dcsign, and thc task has many advantages as the unit of analysis if that
is the goal. It can serve in needs identification, syllabus design, materials writing,
methodology, testing and evaluation alike. The potential of task-based language
teaching for harnessing instructional and learning strategies in ways consistent
with second language acquisition research findings is also considerable. If that
potential is to be realised, however, careful attention needs to be given to the
judicious use of group work, to the kinds of tasks teachers and learners work on,
especially thc psycholinguistic properties of task types, and to the optimal
combinations of task types and groups, that is, to task-group interactions.

REFERENCES
ASTON, G. 1986. 'Woub le-shooting in Interaction with Learners: The More the Merrier?" Applied
Linguistics 7, 2: 128-143.
BARNES, D. 1976. From Communication to Curriculum. Harmondsworth: Penguin.
BERWICK, R F. 1988 The Effect of Task Variation in Teacher-Led Groups on Repair of English as
a Foreign Language, PhD Dissertation. Vancouver: University of British Columbia.
BREEN, M P. 1984 "Process Syllabuses for the Language Classroom." General English Syllabus
Design, edited by C I Brumfit. English Language leaching Documents 118. acfbrd: Pergamon .0^"

Press.
BROCK, C A. 1988 '77:e Effects of Referential Questions on ESL Classroom Discourse." TESOL
Quarterly 20, 1: 47-59.
BYGA7'E, Al. 1988. "Units of Oral ExpressiGn and Language Learning in Small Group Interaction."
Applied Linguistics 1: 59-82.
CANDLIN, C N. 1987. 'Towards Task-Based Language Learning." Language Learning Tasks,
edited by C N Candlin and D F Murphy. Lancaster Practical Papers in English Language Educa.
lion 7. Englewood Cliffs, NJ: Prentice-Hall.
CHAUDIWN, C. 1988. Second Language Classrooms: Research on leaching and Learning. New
York: Cambridge University Press.
CROOKES, G. l9,SO. Task Classification: A Cross-Disciplinary Review. Technical Report No. 4.
Honolulu: Center for Second l.anguage Classroom Research, Social Science Research Institute
University of Hawaii at Manoa.
. 1989. In press. "Planning and Interlanguage Variation." Studies in Second Language
Acquisition 11.

BEST COPY MAILABLE 19

 and Rulon, K A. 1985. Incorporation of Corrective Feedback in Native Speaker/Non-Ncuive
Speaker Conversation. Thchnical Report No. 3. Iknolulu: Center for Second Language Class-
room Research, Social Science Research Institute, University of Hawaii at Manoa.
. 1988. 'Topic and Feedback in Native-Speaker/Non-Native Speaker
Conversation." MSOL
Quarterly 22, 4: 675 681.
DAS, B K. 1984. "Process-Oriented Language Curricula and their Implications." Trends in Language
Syllabus Design, edited by! A S Read. Singapore: Singapore University Press/RELC.
and Thacher-
DEEN, J Y. 1987. An Analysis of Classroom Interaction in a Cooperative Learning
Centered Setting. MA TESL Thesis. Los Angeles: UCLA.
Teaching
DINSMORE, D. 1985. Waiting for Godot in the EFL Classroom." English Language
Journal 39: 225-234.
DOUGHTY, C and Pict,, Il 1986. informotion Gap Risks: An Aid to Second I anguage Acquisi-
tion?" TESOl. Quarterly 20: 303-325,
Talking to Learn:
DUFF, P A. 1986. "Another Look at Interlanguage Talk: Taking Task to Task."
Conversation in Second Language Acrishion, edited by R R Day. Rowley, MA: Newbury House.
ELLIS, R. 1985. Understanding Second Language Acquisition. Oxford: Oxford University Press.
. 1987 "Interlanguage Variability in Narrative Discourse: Style
Shifting in the Use of the Past
Tense." Studies in Second Language Acquisition 9, I: 1-20.
Study of the Classroom
. 1989. "Are Classroom and Naturalistic Acquisition the Same? A
Acquisition of German Word Order Rules." Studies in Second Language Acquisition, in Press.
HATCH, E Af. 1983. Psycholinguistics: A Second Language Perspective. Rowley, MA: Newbury
House.
HOETKER, J P and Ahlbrand, W. 1969. '77w Persistence of the Recitation." American Educational
Research Journal 6: 145-169.
Canberra:
JOHNSTON, M. 1985. Syntactic and Morphological Progressions in Learner English.
Department of Immigration and Ethnic Affairs.
KRASHEN. S D. 1985. 771e Input Hypothe.gis. New York: Longman.
LARSEN-FREEMAN, D and Long, M II. In press. An Introduction to Second Language Acquisition
Research. London: Longman.
Dissertation. Los
LONG, Al H. 1980. "Input, Interaction, and Second Language Acquisition. PhD
Angeles: MLA.
Native Language and For-
. 1981. Input, Interaction and Second Language Acquisition."
eign 1 anguage Acquisition, edited by II Winitz. Annals of the New York Academy of Sciences 379:
250.278.
. 1985. 'A Role for Instruction in Second Language Acquisition: TOsk-Based Language
Teaching." Modelling and Assessing Second Language Acquisition, edited by K Ilyhenstan and M
Pumemann, Clevedon, Avon: Multilingual Matters.
Methodology." Paper
. 1988a. "Focus on Form: A Design Featur, in Language Teaching
presented at the National Foreign Language Cemer/European Cultural Foundation Conference on
Empirical Research on Second Language Learning in Institutional Settings. Bellagio, Italy: Rocke-
feller ( 'enter. June 20-24.
Language Acquisition:
19886. "instructed Inteilanguage Development." Issues in Second
.

48 20
 Multiple Perspective edited by L M Ikebe. New York: Newbury House/Harper and Row.
; Adams, L; McLean, M and Castanos, F. 1976. "Doing 771ings with Words: Verbal Interac-
tion in Lockstep and Small Group Classroom Situations." On TESOL '74 edited by J Fanselow
and R C'rymes. Washington, DC:: TESOL.
and Crookes, G. 1987. "Intervention Points in Second Language Classroom Processes."
Patterns in Clas.sroom Interaction in Southeast Asia, edited by B K Das. Singapore: Singapore
University Press/RELC.
and Crookes, G. 1989. "Units of Analysis in Syllabus Design." Unpublished ms. Honolulu:
University of Hawaii at Manoa.
and Porter, P A. 1985. "Group Work, Interlanguage Thlk, and Second Language Acquisi-
tion." TESOL Quarterly 19, 2: 207-228.
and Sato, C J. 1983. "Classroom Foreigner Thlk Discourse: Forms and Functions of Teach.
crs' Questions." Classroom Oriented Research in Second Language Acquisition, ediwd by II IV
Sehger and M 11 Long. Rowley, MA: Newbury House.
NUNAN, D. 1987. "Communicative I,anguage Teaching: Making it Work." English Language
Thaching Journal 41, 2: 136-185.
. 1988a. Syllabus Design. Ovford: Oxford University Press.
. 1988h. The Learner-Centered Curriculum. Cambridge: Cambridge University Press.
. 1989. Designing Tasks for the Communicative Classroom. Cambridge: Cambridg: Univers,
ty Press.
OCHS, E. 1979. "Planned and Unplanned Discourse." Discourse and Syntac edited by T Givon.
Ncw York: Academic Press.
PARKER, K and Chaudron, C. 1987. "The Effects of Linguistic Simplification and Elaborative
Modifications in L2 Comprehension." University of Ilawau Working Papers in ESL 4 2: 107-133.
PHU J.1PS, M and Shelties-worth, C 1975 "Questions in the Design and Implementation of Courses in
English for Specialized Purposes." Proceedings of the Fourth International Congress of Applied
Linguistics Volume 1, edited by G Nickel. Stuttgart: liochselude Verlag.
PICA, 7'. 1987. "Interlanguage Adjustments as an Outcome of NS NNS Negotiated interaction."
Language Learning 3Z 4: 563-593.
and Doughty, C 1985. "Input and Interaction in the Communicative Language Classroom:
A Companson of 7i'acher-Fromed and Group Actissties." Input and Second Language Acquisi
lion, edited by S Gass and Madden. Rowky, MA: Newbury House.
. 1985. '77ie Role of Group Work in Classroom Second Language Acquisition." Studies in
Second Language Acquisition 7: 233-248.
; lolliday, I.; Lewis, N and Morgenthaler, L. 1989. "Comprehensible Output as an Outcome
of Linguistic Demands on the Learner." Studies in Second Language Acquisition 11, I, in pres%
and and Long, M H. 1986. '77w Linguistic and Conversational Performance of Experienced
Inexperienced Teachers." Talking to Learn: Conversation in Second Language Acquisition, edited
by R R Day. Rowley, MA: Newlin'?" House.
; Young, 12 and Doughty, C. 1987. "Ow impact of Inter-tic-turn on Comprehension !LSO!
Quarterly 21, 4: 737-758.
PIENEMANN, M and Johnston, M. 1Q87. "Factors Influencing the Development of language Profi

49 ?
 4

ciency." ApAring Second Language Acquisition Research, edited by D Nunan. Adelaide: Nation-
al Curriculum Resource Centre.
and hueiaction in Task-Centered
PORMR, P A. 1986. "How Learners Thlk to Each Other: Input
Second Larguage Acquisition, edited by R R Day.
Discussions." Talking to Learn: Conversation in
Rowley, MA: Newbury House.
PR.A&HU, N S. 198Z Second Language Pedagogy. Oxford: Oxford University Press.
Teaching Journal 41, 1: 3-
RAMAN!, E. 1987. '77reorifing from the Classroom." English Language
H.
and Kegan
ROBINSON, W P and Rackstraw, S J. 1972. A Question of Answers London: Roudedge
Paul.
Education
ROSS, S. lb appear. 'Praxis and Product in the EFL Classroom." Second Language
Program.% edited by C Alderson and A Berretta. Cambridge: Cambridge University Press.
Teacher-Fronted and Small Group
RULON, K and McCreary, J. 1986. "Negotiation of Content:
Second Language Acquisition, edited by R R
Interaction." Talking to Learn: Conversafion and
Day. Rowley, MA: Newbury House.
Foreign Language
SCHERER, G A C and Wertheimer, M. 1964. A Psycholinguistic Experiment in
Teaching. New York: McGraw,Hill
Thaching: A Re-Examina-
SEL1GER, II W. 1975. "Inductive and Deductive Method in Language
tion." International Review of Applied Linguistics a I: 1-18.
and
SHAVELSON, Ri and Stern, P. 1981. "Research on Teachers' Pedagogical Thoughts, Judgments
Behaviour." Review of Educational Research 51, 4: 455-498.
SMITH, P 1970. A Comparison of the Cognitive and Audiolingual Approaches to Foreign Lan-
for Curricu-
guage Instruction: The Pennsylvania Foreign Language Project Philadelphia: Center
lum Development.
Instructional Differences and Learning Outcomes: A
Sl'ADA, N. 1987. "Relationships Between
137-161.
Process-Product Study of Communicative Language Teaching." Applied Linguistics
Redefining Method
SWAEFER. J K; Arens, K and Morgan, M. 1982. "Teacher Classroom Practices:
as Task Hierarchy." Modern Language Journal 66: 24-33.
ESL/EFL Classes." Cross Currents 15, I:
OLLEFSON, J 1988. "Measuring Communication in
37.46.
Language
1.-ON E1,1X, T and Oskinrson, M. 1975. Comparative Methods Experiments in Foreign
'leaching. Gothenburg: Molndal School of Education.
WHTIE, R W 1988. The ELT Curriculum: Design, Innovation and Management. Orford: Basil
Blackwell.
WILKINS, D A. 1972. Linguistics in Language Teaching. London: Edward Arnold.
1976. Notional Syllabuses. Otford: Oxford University Press.
.

. 1081 'Notional Aylabuses Revisited." Applied Linguistics Z 1: 83-90.

:?2
50