See

discussions, stats, and author profiles for this publication at:

https://www.researchgate.net/publication/31295480

A Framework for the

Implementation of Task-
based Instruction

Article in Applied Linguistics March 1996

DOI: 10.1093/applin/17.1.38 Source: OAI

CITATIONS READS

506 4,154

1 author:

Peter Skehan
St Mary's University, Twickenham
56 PUBLICATIONS 5,202 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these

related projects:

Rethinking language aptitude: Multiple perspectives and

emerging trends. (Book project in preparation) View project

Processing perspectives on task performance View project

All content following this page was uploaded by Peter Skehan on 24 April 2015.

The user has requested enhancement of the downloaded file.

 A Framework for the Implementation of Task-
based Instruction
PETER SKEHAN
Thames Valley University

This paper examines recent proposals for task-based approaches to instruc-

tion It reviews relevant research, before going on to examine a number of
potential problems with task-based teaching, such as a potential focus away
from form and towards lexis It renews recent developments in cognitive
psychology which support a dual-mode perspective for language processing,
and then proposes the goals of accuracy, complexity-restructuring, and fluency
as the most relevant for task-based instruction In thefinalsection, the paper
proposes aframeworkfor the implementation of task-based instruction which
draws upon relevant theory and research, and which organizes the methods by
which such instruction could be put into practice in such a way as to minimize
problems, and maximize the probability that all three above goals can be
achieved
INTRODUCTION
In recent years, a number of researchers, syllabus designers, and educational
innovators have called for a move in language teaching towards task-based
approaches to instruction (Prabhu 1987, Nunan 1989, Long and Crookes
1991, Gass and Crookes 1993a, b) Yet there are a number of critiques of task-
based instruction that could be made (Sheen 1994), and it is also unclear how
such an approach could be implemented The present article will examine the
cntiques, reviewing the theory and research that suggests limits and qualifica-
tions for task-based approaches It then proposes a framework to enable
teachers to implement task-based instruction on a more systematic and princi-
pled basis
At the outset, it is helpful to examine some preliminary questions relating to
tasks and task-based instruction Although a number of definitions of task exist
(see e g Nunan (1989)), for present purposes a task is taken to be an activity in
which meaning is primary, there is some sort of relationship to the real world,
task completion has some priority, and the assessment of task performance is in
terms of task outcome Of course, translating these criteria into reality is not
always a straightforward matter Most activities combine a number of priorities,
and it is a fine judgement to claim that the communication of meaning is a
primary goal for any particular task, or to assert that a task has a real-world
relationship Classrooms are classroms, but even so, a task which requires
personal information to be exchanged, or a problem to be solved, or a collective
judgement to be made bears a relationship to things that happen outside the
classroom in a way that separates these activities from doing, for example, a
transformation exercise One could make similar points about the other two
Applied Linguistics, Vol 17, No 1 C Oxford University Press 1996
 PETER SKEHAN 39

components of definition that are offered here (task completion and task
outcome assessment), but the point is that such a concept of task has sufficient
focus to enable it to be a viable component for what goes on in classrooms '
Long and Crookes (1991) discuss a further quality of tasks that they have a
clear pedagogic relationship to out-of-class language use, in that needs analysis
should clarify how students will need to use language in real-life, and task design
should ensure that classroom tasks bear a developmental relationship to such
non-classroom activity For this article, such a quality is regarded as desirable,
but difficult to obtain, and the more restrictive approach to characterizing tasks
will be used Similarly, there will be no discussion of the interesting possibility
that students could be involved in the negotiation of which tasks are used, and
how they are used (Breen 1987) The approach taken here is to try to develop a
framework which will help the teacher to better understand the tasks s/he is
using, and to sequence and implement them more effectively
One can, on the basis of these task characteristics, identify strong and weak
forms of the task-based approach A strong form would argue that tasks should
be the unit of language teaching, and that everything else should be subsidiary
In this view, the need to transact tasks is seen as adequate to drive forward
language development, as though second language acquisition is the result of the
same process of interaction as first language acquisition (Wells 1985) A weak
form of task-based instruction would claim that tasks are a vital part of language
instruction, but that they are embedded in a more complex pedagogic context
They are necessary, but may be preceded by focused instruction, and after use,
may be followed by focused instruction which is contingent on task per-
formance This version of task-based instruction is clearly very close to general
communicative language teaching It could also be compatible with a traditional
presentation, practice, production sequence, only with production based on
tasks (as defined above), rather than more stilted and guided production
activities (Littlewood 1981)
If we turn next to research studies into task-based learning, a range of
approaches are evident Candhn (1987) discusses criteria by which tasks may
be analysed, basing the categones used on an essentially data-free account of
task properties, an approach developed by Nunan (1989) and Skehan (1992)
Pica, Kanagy, and Falodun (1993) take a slightly different perspective, relying
on research studies more directly They analyse tasks in terms of interactional
patterns and requirements, (i e how involved each participant needs to be), as
well as the sorts of goals that underlie the tasks-to-be-transacted (e g one-way
vs two-way)
A number of studies have been more specific in their analyses of tasks,
positing particular contrasts or dimensions as the basis for characterizing tasks
Prabhu (1987) argues in support of reasoning-gap tasks, Duff (1986) contrasts
divergent and convergent tasks, arguing that the latter engage acquisitional
processes more effectively, and Berwick (1993) contrasts two dimensions
experiential-expository, and didactive-collaborative There have also been
studies of the participants within tasks such as Yule, Powers, and Macdonald
40 TASK-BASED INSTRUCTION
1
(1992) on 'hearer effects and Plough and Gass (1993) on participant (and task)
familiarity Brown, Anderson, Shilcock, and Yule (1984) have also investigated
various task design features, in an attempt to establish task difficulty on an
empirical basis They propose that static tasks (e g description) are easier than
dynamic tasks (e g narration), which in turn are easier than abstract tasks (e g
opinion giving), and that the number of elements, participants, and relationships
in a task makes it more difficult There have also been studies of processing
influences on tasks Tarone (1985) has shown that attention to form has a clear
effect on accuracy of performance Ellis (1987) reported an interaction between
the engagement of planned discourse and different forms of the past tense under
different task conditions Crookes (1989) reported greater complexity and
lexical variety for tasks done under a planning time condition, but, interestingly,
no greater accuracy Foster and Skehan (1994) report an interaction between
opportunity to plan and task type These research studies are individually
revealing, but do not currently provide the basis for more general perspectives
on task-based instruction They do, though, play a part in the wider framework
which is proposed below

PROBLEMS AND ASSUMPTIONS IN TASK-BASED LEARNING

Approaches to instruction which make meaning primary, such as task-based
instruction, obviously have considerable appeal in terms of authenticity and
linkage with acquisitional accounts of the course of language development But
there are pitfalls with such an approach, generally stemming from the con-
sequences for form of putting such an emphasis on meaning These pitfalls need
to be understood if task-based approaches are to be properly exploited
Learners (and native-speakers) will place great emphasis on communicating
meanings, but not necessarily worry about the exact form that they use (Kess
1992) In this respect, Gnce (1975) has made clear that maxims for conversa-
tion make for a considerable processing burden because of what is not said To
spell everything out in complete and well-formed sentences would soon empty
rooms, and get oneself classified as a boring pedant Much adult conversation is
elliptical and incomplete in surface form, heavy in the assumptions it makes
about background knowledge being the basis for the implicatures that are made
about intended meaning, speaker attitude, let alone propositional meaning
(Wilson 1994) It goes against the grain, in other words, to do more than use
form as one element or pressure in native-speaker communication, where the
major emphasis will be on the satisfactonness of the flow of the conversation,
not the correctness, or completeness (or the usefulness for interlanguage
development amongst learners) of what is said
Further, if we now turn directly towards learners, two other consequences of
such an emphasis on meaning become apparent There is natural and
unavoidable use of strategies of comprehension (Clark and Clark 1977), in that
non-deterministic and non-exhaustive methods are used to recover intended
meaning, with the success of this operation often being dependent on only
partial use of form as a clue to meaning (Anderson and Lynch 1987) In other
 PETERSKEHAN 41

words, processing language to extract meaning does not guarantee automatic

sensitivity to form, and the consequent pressures for interlanguage develop-
ment which is assumed by supporters of the Input Hypothesis (see Skehan
(1992) for further details) Further, there is widespread use of communication
strategies These, too, help the learner succeed with meaning while having the
consequence of sometimes bypassing form Cognitive and linguistic com-
munication strategies, that is, (Kellerman 1991) can be used to handle
communicative pressure, but in so doing remove the automatic engagement that
would be required to constantly stretch interlanguage and lead to change Worst
of all, in this regard, is the possibility that reliance on comprehension and
communication strategies when meanings need to be communicated under
pressure will be loo effective What this could mean is that "solutions' to
communication problems become procedurahzed and re-used on other
occasions (Skehan 1992) This would be desirable if the solutions in question
led to interlanguage development, but it is equally likely that they will not, since
there is little tojustify the 'on the fly' improvizations that are involved being part
of a desired route for change Their procedurahzation could then become a
stumbling block for change in the future
There is, though, an even more deep-seated problemthe possibility that
much communication is lexical in nature Regarding first language acquisition,
Skehan (1992), following Peters (1983, 1985), Nelson (1981), Bates,
Bretherton, and Snyder (1989) suggests that initial progress is lexical in nature,
but then the initial stock of lexical items becomes syntactized Most important,
though, the language system which is developed in this way then becomes
relexicahzed (Skehan 1992), le language which is analysable and has been
analysed is actually stored as a repertoire of lexical items Most centrally of all,
the unit of storage is now no longer the word, but can consist of multi-word units
(Peters 1985) which, when used, are processed as a single item The cost of such
a system is that multiple storage is necessary, since each multi-word item
requires separate storage as a unit The gain, however, is considerable since
during ongoing language processing such a larger unit can be processed as a
whole, releasing resources for other aspects of the speech planning and
execution required (Bygate 1988)
Such a lexical interpretation of language performance has been argued in
recent years by both linguists and psychologists Bohnger (1975), for example,
proposes that the idiomaticity of language has been vastly underestimated, and
that much language processing is hardly creative at all, but relies upon familiar
memorized material Pawley and Syder (1983) argue that speech is planned a
clause at a time, and that language users rely on lexicahzed sentence stems, (of
which they have thousands at their command), as they improvize their way,
joining together such elements to produce connected discourse Nattinger and
DeCarnco (1992) similarly argue for the importance of lexical units in speech
production
These analyses imply that language users have available dual modes of
processing (Widdowson 1989) When accessibility and time pressure are
42 TASK-BASED INSTRUCTION

paramount, a lexical mode of communication will be relied upon, which draws

upon a capacious, well-organized, and very rapid memory system In contrast,
when exactness or creativity matter, analysability, and a concern for form, for
syntax, and for planning, will predominate (Sinclair 1991) Armed with these
two possibilities, the language user can switch between the two modes to take
account of whatever processing demands are most pressing Skehan (in press, a)
argues that 'ability for use' (Widdowson 1983) can best be understood as the
capacity which marshalls processing resources in this manner in order to
achieve communication
If we relate the discussion in this section to task-based approaches for
instruction, what this means is that tasks themselves, given their defining
properties of meaning primacy, outcome evaluation, and realism, may well
predispose those engaged in task completion to engage in a mode of
communication which does not prioritize a focus on form, either in terms of
using linguistic elements to achieve precision or to achieve accuracy As a result,
it may not be possible to rely on a task-based approach to automatically dnve
interlanguage forward Instead, it is likely that it will teach learners simply how
to do tasks better, to procedurahze strategic solutions to problems, and to
engage in lexicahzed communication These conclusions suggest that it is
necessary, if task-based approaches to instruction are to be viable, to devise
methods of focusing on form without losing the values of tasks as realistic
communicative motivators, and as opportunities to trigger acquisitional
processes Pnor to discussing a framework which may achieve this goal, how-
ever, it is necessary to try to situate task-based instruction within a theoretical
viewpoint more grounded in contemporary psychologyan information pro-
cessing perspective

COGNIIIVE APPROACHES TO LANGUAGE LEARNING

In recent years, there have been a number of studies which have clarified the
theoretical basis for a cognitive approach to language learning These have
concerned the nature of what is learned, the role of consciousness, the role of
performance factors, and the way in which attention impacts upon learning We
will examine each of these in turn
One method of addressing the question of what is learned is through
laboratory studies of artificial languages (though see Van Patten (1994) for a
critique of the relevance of such research) In the learning of such structured
material, the issue is whether progress consists of the induction of underlying
abstract rules following a process of (possibly implicit) restructuring (Reber
1989, McLaughhn 1990), or the learning of exemplars, l e specific, con-
textually coded items which may contain structure, but which are learned as
chunks (Carr and Curren 1994) The former interpretation regards develop-
ment in terms of the growth and complexity of the underlying system involved,
while the latter is more concerned with the accumulation of exemplars, and their
utility in performance
The connection with natural language learning here is clear The rule-based
 PETERSKEHAN 43
interpretation would imply that interlanguage development would be the result
of the restructuring that occurs with linguistic material (McLaughlin 1990),
motivated by the continued operation of a Universal Grammar or by other
cognitive processes The exemplar-based interpretation, in contrast, would
argue for development as being the accumulation of useful chunks of language,
i e in earlier terms, language as formulaic items Most interestingly of all, in this
regard, is that Carr and Curren (1994), following Mathews, Buss, Stanley,
Blachard-Fields, Cho, and Druhan (1989), interpret findings in this area as
being most consistent with a dual-mode of processing, in which there is evidence
for both structured learning and exemplar-based learning, but with the
operation of both modes combining in a synergistic manner to yield results, and
degrees of learning, that are more than simply the sum of the parts The parallel
with the use of a relexicahzed repertoire of language (Skehan 1992) is clear
Schmidt (1990, 1994) discusses the role of consciousness in language
learmng He distinguishes between several senses of this term, such as
awareness, control, and attention In this section, we will discuss consciousness
as awareness, moving on to consider the other two senses in the sections on
fluency and attention For Schmidt (1990, 1994), consciousness has con-
siderable importance in language learning There is accumulating evidence (see
review in Carr and Curren 1994) that explicit learning of structured material is
generally superior to implicit learning, suggesting that awareness of the learning
itself and of what is to be learned confers advantages Schmidt (1994) suggests
that, for example, awareness enables more efficient solutions to the 'matching1
problem (Klein 1986), le noticing the gap between one's current language
system and the language one encounters Similarly, Schmidt (1994) proposes
that awareness may enable learners to appreciate better the instruction that they
are receiving, especially the correction that is being given Awareness may also
(Karmilloff-Smith 1986) make it easier to transform and recombine material, to
restructure, in other words, as the structure of matenal is more available, and
other organizational possibilities become clear Finally, awareness may help
learners operate the sort of dual-mode systems outlined above, where the
learner/language user may need to combine rule-based systems and exemplar-
based systems during ongoing performance In this respect, one interesting
possibility is that on occasions where rule-based systems are used for the
generation of language, the products of such activity can themselves become
exemplars and then retrieved and used as exemplars on subsequent occasions
(cf the earlier discussion on relexicahzation, Peters (1985), and discussion
below) Clearly, consciousness-as-awareness would be of considerable help in
this process
This possibility takes us into the need to discuss fluency, and the ways in
which consciousness-as-control may operate Schmidt (1992) provides an
extensive review of psychological mechanisms underlying fluency in foreign
language performance In general, there seem to be three ways of accounting for
the development of fluency accelerating models, restructuring models, and
instance models The first approach simply suggests that there is a natural
44 TASK-BASED INSTRUCTION

sequence in which initial declarative knowledge becomes procedurahzed

(Anderson 1989) or automatized (Schiffnn and Schrueder 1977) so that
essentially similar processes are used, but more quickly and with less need to use
mental resources to control them, I e the same steps are followed, but more
quickly and efficiently and probably less consciously Restructuring approaches
(Cheng 1985,McLaughlin 1990) regard improved performance as the result of
using better algorithms so that performance is better organized One assumes,
following this approach, that restructuring, when it occurs is rapid and
immediately available to sustain improved performance (fluency in this case)
Instance-based approaches (Logan 1988, Robinson and Ha 1993) regard
fluency as performance which is based not on rules which are applied more
quickly nor on rules which are more efficiently organized, but on contextually-
coded exemplars which function as units Such units (which may be significantly
longer than a word) are the product of previous rule applications which are now
stored in exemplar form, and so require far less processing capacity because
they are retrieved and used as wholes On this view (Peters 1985, Schmidt
1992), learning is the result of instance creation, and performance (and the
ensuing fluency) the result of instance use
We will assume here that the restructuring model is not very relevant to
second language learningfluency,although it is to interlanguage development in
general (It seems more suited to cognitive learning of a more general nature
when insight and new algorithms are what lead to problem-solving efficiency)
More important for present purposes are the two other accounts The first, the
procedurahzation model, concerns an interplay between declarative knowledge
and the fluency which arises from procedurahzation, with the cost that less
control is available over such material The contrasting approach, instance
theory, portrays the relationship between the rule-governed part of the system
and the fluency oriented component differently, giving the latter a greater
degree of autonomy
It is the latter interpretation which is taken to be relevant to foreign language
learning even though the procedurahzation account has merit This is so largely
because an instance-based interpretation fits in more effectively with the dual-
mode account of structured learning presented earlier, as well as with a
syntactic-lexical contrast in natural language learning It also provides an
interesting theoretical interpretation of the phenomenon of fossihzation, in that
one can now regard such an outcome as the premature product of a rule-based
system which is then made available as an exemplar in future language use
There is no requirement, in other words, that what are created as exemplars are
correct In beneficial circumstances, rule-created exemplars may be supplanted
by other exemplars which are created when the underlying rule-based system
has evolved more But if the underlying system does not so evolve, and if
communicative effectiveness is achieved, the erroneous exemplar may survive
and stabilize, l e become a syntactic fossil
Finally, we need to consider the role and functioning of attention, the final
meaning of consciousness mentioned earlier In information processing terms,
 PETER SKEHAN 45

attention is a process, and is capacity-robbing (Van Patten 1994) One chooses

to attend to some things at the expense of others, and the choice of attentional
direction, as well as the use of attentional resources themselves, have costs as far
as the processing of potential foregone material is concerned Attention, that is,
has both a control function for further processes, and also a direct effect m terms
of the focus of consciousness attention at any one time (Schmidt 1994)
Pursuing the information processing perspective, we can consider that, in
foreign language learning and performance, three stages can be distinguished
input, central processing, and output (Skehan 1994) We will examine the role
of attention in the first two of these, the third having been covered in the
discussion on fluency
Regarding input, Van Patten (1990,1994) has shown that meaning is primary
when attentional resources are limited He argues that under such conditions
there is attention to form only if it is necessary for the recovery of meaning On
the other hand, form can be attended to, even if it is not crucial for meaning, if
there is no pressure on attentional resources, l e if there is spare attentional
capacity Schmidt (1990) argues similarly for the importance of noticing as a
means to channel attentional capacity so that input can become intake He
proposes that various factors, e g salience, traces of previous instruction, and
task-demands (I e control of attentional capacity on the part of the task-
designer) will make it more likely that attention will be directed to form, but that
the central issue is that this noticing occurs and that the spare attentional
capacity which is involved is directed so that it attends to important aspects of
form The challenge for the task designer is, then, to engineer situations in which
this is more likely to occur with learners
Essentially, the effective use of attentional capacities during input is intended
to create a situation in which input can become intake, and effective processing
of material can be triggered so that the implications of existing input for inter-
language development can be exploited and acted upon In this respect, we are
now concerned, in Schmidt's (1994) terms, with consciousness as focal
attention, with the need to commit attentional resources to the material which is
being processed Carr and Curren (1994) propose that focal attention of this
sort does help in the learning of structured matenal They suggest that if material
is simple and unambiguous, there is no great advantage for exphcitness But if
material is more complex and ambiguous (as they characterize natural
languages), exphcitness confers an advantage In particular, Carr and Curren
(ibid) suggest that focal attention of this sort enables more effective parsing of
material and ncher coding, with such processes being helpful for language
development Focused attention of this sort does not need to lead to the ability
to consciously articulate underlying rules for the language being learned, but it is
important that there is a self-awareness about the task of learning which is being
faced In this respect, Van Patten (1994) offers the analogy of learning to
improve a backhand in tennis One may not benefit from an exposition of the
underlying physics, or of the metalanguage of tennis strokeplay, but it may be
helpful to give one's full attention to the backhands one is playing, and even to
46 TASK-BASED INSTRUCTION

benefit from 'input to intake' hints from a coach, along the lines of 'Think about
your footwork' In language, too, attention can be directed without this direction
necessarily involving detailed explanation of rules The issue, centrally, is that
limited capacity is being focused in the area which is most helpful, with the
possibility that what then happens triggers implicit processes Generalizations
in data which might otherwise be missed become more accessible with the
structuring of the learning experience that is involved

GOALS IN TASK-BASED INSTRUCTION

This discussion demonstrates that task-based approaches to instruction are
currently in a transitional position There are clear reasons for the adoption of
task-based approaches, principally associated with their potential engagement
of acquisitional processes, there is underlying psycholinguists research which,
to some extent is supportive, and there is a range of specifically task-based
second language research which is helpful in evaluating this type of instruction
On the other hand, there are arguments, both linguistic and psychological, for
why a focus on meaning may not engage such acquisitional processes, there is
psycholinguistic evidence which argues for a clear role for exphcitness and
consciousness, for the manipulation of attentional focus, and for the existence of
dual modes of processing, structural and exemplar-based This discrepancy
places proponents of task-based instruction in a difficult position, since, while it
is clear that there are advantages to using such an approach, it is difficult to
know how strongly to argue this position, and how exactly to implement such
instruction A necessary step, therefore, is to draw on the preceding discussion
to set appropnate goals for task-based approaches
It is relatively easy to identify, as a general goal in foreign language learning,
that of becoming more native-like in one's performance,2 on the grounds that
most people have such views about the levels of competence and performance
that they would like to achieve Within this general goal, however, it is proposed
that it is useful to separate learner goals into three main areas accuracy,
complexity, and fluency The first of these, accuracy, is concerned with a
learner's capacity to handle whatever level of interlanguage complexity s/he has
currently attained Complexity, and its attendant process, restructuring, relates
to the stage and elaboration of the underlying interlanguage system Fluency,
finally, concerns the learner's capacity to mobilize an interlanguage system to
communicate meanings in real time 3
To take the first of these goals in more detail, accuracy relates to a learner's
belief in norms, and to performance which is native-like through its rule-
governed nature Such a goal is desirable for a number of reasons inaccuracy
could impair communicative effectiveness, it could stigmatize, it could fossilize,
and finally self-perceived inaccuracy could be demoralizing to the learner
Turning to the causes of inaccuracy, one is that the underlying interlanguage
system is inadequate, or transitional, such that the language which has been
produced is grammatical, but to an incorrect system which needs to change
further (Ellis 1994) (see below) One could speculate that the tendency to be
 PETER SKEHAN 47

inaccurate on this basis relates to how well established the particular part of the
lnterlanguage system is But it is also possible that inaccuracy is the result of the
competence-performance relationship, and of the way in which communicative
pressure has led to an error being made which, under other circumstances,
would not be a lapse, in Corder's (1981) terms And, of course, we cannot
ignore the possibility that 'inaccuracy' may itself be targeted as desirable, for
whatever reasons the learner is motivated by (Trevise and Noyau 1984)'
The reverse side of this coin is to consider what promotes accurate language
use Clearly, the use of well-integrated aspects of the lnterlanguage system will
be helpful here, I e a sort of conservative communication strategy, in which
what is well-known is used, and what is not is avoided (Schachter 1974)
Learners who dislike risk-taking will, presumably, be drawn to accuracy
because of a reluctance to use language they are not sure of A similar effect will
result from a greater concern, on the part of the learner, to be correct, to
conform to target language norms, and to value them as important But other
factors are also likely to have an impact, such as the effectiveness with which
attentional resources are mobilized, and the processing capacity which is
available given other aspects of communicative pressure, I e the more attention
is diverted elsewhere, the less attention is available for form and accuracy (Van
Patten 1990, Van Patten and Cadierno 1993)
In this regard, we need to turn to the next goal that has been proposed for
language learningcomplexity/restructuring Restructunng is concerned with
the process by which the lnterlanguage system becomes more complex,
elaborate, and structured (McLaughhn 1990), maybe more efficient and less
circumlocuitous in communication (Cheng 1985), more consistent with input
data, and more native-like (Cook 1994) It requires a learner who explicitly
accepts such developments as goals and who is driven, by whatever means, to
achieve them More complex lnterlanguage systems are desirable since they
reflect acquisition having taken place, and will enable a greater degree of
acceptance as a speaker of the language concerned Equally important, such IL
systems should enable greater precision in communication (Swain 1985), and
greater communicative efficiency in the face of difficult performance circum-
stances They should also make it more likely that more complex ideas will be
expressed effectively (Swain, in press)
If we examine why restructuring may to some extent fail to occur a number of
reasons present themselves There may be a lack of interest on the part of the
learner either in the goal of becoming more native-like or simply in making the
effort to change and re-organize an lnterlanguage system There may also be
impoverished input (or instruction), such that necessary conditions for restruc-
tunng are impaired But there might also be problems even if the above two
influences are not present For example, learners may prefer not to take risks,
relying on less elaborate lnterlanguage systems which are adapted to
communicate meanings in such a way that lnterlanguage is not pressured for
change (Schachter 1974) Such a conservative strategy would promote accuracy
at the expense of complexity Equally, there may be pressure to communicate
48 TASK-BASED INSTRUCTION

which does not provide adequate time for restructuring to occur, since
processing resources have to be excessively committed to achieve certain
communicative outcomes In such cases, it is likely that 'solutions" to com-
municative problems will be procedurahzed, and exemplar-based learning will
occur (Schmidt 1983,1992)
Reversing the above analysis makes it clear how restructuring can be more
likely to be achieved There needs to be an interest in achievmg native-like
performance, and possibly an interest in change more generally There also
needs to be helpful input, both explicit and implicit It is also important that
interactive opportunities, e g tasks which need to be transacted, will have a
stretching influence on mterlanguage, in that precision of expression should be
integral to their completion (Swain, in press) It would also be helpful if there
were time to engage restructuring processes, and to attend to things other than
immediate communication pressures Finally, it would help if there were
support for restructuring, through sequencing of teaching activities, through
relevant preparation (Crookes 1989), and through appropriate post-task
activities (Willis and Willis 1988, Skehan 1992)
We can turn finally to fluency, the last of the three goals outlined earlier
Fundamentally, this consists of the capacity to mobilize one's linguistic
resources in the service of real-time communication, l e to produce (and
comprehend) speech at relatively normal rates, approaching (but not neces-
sarily identical to) one's own native-language speech rates In particular, one
would look at features such as rate, pausing, reformulation, hesitation,
redundancy, and the use of lexical units (Bygate 1987) to establish the level of
fluency which has been achieved It is assumed that to achieve this goal requires
a capacity to use implicit knowledge systems in actual performance (Schmidt
1992)
Adequate levels of fluency are desirable if one wants to be acceptable as a
worthwhile interlocutor (Schmidt 1983) Poor fluency will lead to difficult (and
less frequent) patterns of interaction and further opportunities for learning
(Larsen-Freeman and Long 1991) It will also lead to dissatisfaction, as it
becomes difficult to express interesting ideas in real-time and the normal
orchestration of conceptualization, planning, and execution (Levelt 1989)
From a learning perspective, fluency is also desirable to the extent that it
integrates, as implicit knowledge, the results of emerging and developing
restructuring, and makes accessible the lexicahzed products of the operations of
the restructuring in actual communication (Carr and Curren 1994) More
questionably, the sort of fluency which represents the procedurahzation or
lexicahzation of transitional forms which are incorrect, and whose consolida-
tion may compromise future development, may be undesirable It may
represent communicative progress, but of a type which makes later restructur-
ing more difficult
Given these perspectives, we next need to examine what promotes (a) lack of
fluency, (b) undesirable fluency, and (c) effective fluency Lack of fluency is
clearly more likely when the speaker does not valuefluency,but instead is more
 PETER SKEHAN 49

drawn towards other goals, such as accuracy, or precision and complexity of

speech It is also more likely when there has been insufficient opportunity for the
proceduralization of language, and for the development of an adequate
repertoire of exemplars (formulaic units) to sustain the pressures involved in
real-time communication This, in turn, besides being accounted for by
individual difference factors (Skehan, in press, b), is likely to be the result of
inadequate opportunities to make language production automatic in the
necessary manner Undesirable fluency is likely to result from excessive
procedurahzation, perhaps resulting from the use of Strategic Competence to
solve communicative problems Such 'solutions' are what are automatized and
compromise future interlanguage growth Excessive pressure to communicate,
that is, may result in transitional forms fossilizing as accessible exemplars which
are easy to use, appear to have communicative effectiveness, but are incorrect
Finally, it can be proposed that effective fluency is achieved when previous
restructuring becomes automatized or becomes a (correct) exemplar In such a
case, the pressure to achieve fluency comes at the nght moment after
restructuring has occurred This implies a capacity to engage in cycles of
analysis and synthesis on the learner's part (Klein 1986), with the former focus-
ing on restructuring, and the latter on fluency (Skehan 1992) The former is
necessary to keep a system open, and capable of change The latter is vital if the
system so developed is not to be simply a rule-system which has to be applied
anew in each communication, but instead is made accessible (Widdowson
1989), lexicahzed, and automatized In other words, desirable fluency implies a
capacity to operate a dual-mode system, in which well-organized exemplars are
available to respond to real-time pressures, but a rule-based system can still be
accessed when the need for precision or creativity arises (Carr and Curren
1994) Coming from a corpus-based linguistic perspective, essentially the same
claim is made by Sinclair (1991) in his contrast between the idiom principle and
the open-choice principle the former concerns the way choices are reduced in
the service of real-time communication, while the latter principle is what
enables the language user to access a greater range of lexical choice when this is
appropriate.
If we ask how such a level of beneficial fluency can be achieved, it is clear that
there have to be opportunities to create exemplars in context which can then be
retrieved in later communicative encounters This implies giving learners
communicative problems to solve at the right level of processing difficulty, I e
avoiding excessive processing demands which would disrupt performance,
while also avoiding non-challengmg tasks which do not extend ability for use In
this way, learners are more able to bring to bear the effects of recent restructur-
ing, but at the same time achieve a level of fluency For this to happen with any
consistency, it is important that cycles of activity are organized so that there is a
balance between a focus on form and a focus on communication (Van Patten
1990, Skehan 1992)
Now, though, that we have discussed accuracy, complexity, and fluency
somewhat separately, it is important to examine the interrelationships A focus
50 TASK-BASED INSTRUCTION

on accuracy makes it less likely that interlanguage change will occur, more likely
that speech will be slow, and probably consumes a considerable portion of
attentional resources A focus on complexity and the process of restructuring
increases the chances that new forms will be incorporated into interlanguage
systems, promotes nsk-taking, and requires attention being devoted to the new
forms of language which are being assembled (Foster and Skehan 1994)
Finally, a focus on fluency will lead to language being produced more quickly, to
an emphasis on accessibility (Bygate 1988), and with lower priority being
attached to getting language right, or to the use of new forms
If we consider the processing implications of having these three goals, it is
clear that there is not sufficient capacity for learners to devote resources to each
of them so that they can be met simultaneously As a result, decisions about the
pnontization of attentional resources have to be made during communication
and learning, leading us to need to explore the consequences of allocating
attention in one direction, and not another (Van Patten 1994) Performance is
likely to prioritize fluency, and relegate restructuring and accuracy to lesser
importance A focus on development, on the other hand, is likely to prioritize
restructuring, with accuracy and fluency being more secondary Further, to the
extent that such learning is cumulative, cycles of restructuring, followed by an
emphasis on accuracy and fluency, may be followed by contingent, develop-
mental cycles of further restructuring, as a particular interlanguage sub-system
is progressively complexified
We now need to consider how the three goals of accuracy, complexity-
restructuring, and fluency have been addressed in language teaching method-
ology, i e what decisions are explicit or implicit in existing approaches to
pedagogy A conventional presentation, practice, production sequency tacitly
assumes that change will come about through the presentation phase, and this
will be translated into accuracy and fluency through the succeeding practice and
production stages (Rivers 1981) Following the earlier discussion, we have seen
that there are problems with this approach It assumes that 'restructuring' can be
equated with whatever the teacher (or the syllabus writer) deems to be worth
presenting, and ignores thefindingsof systematicity that have emerged through
second language acquisition research (Ellis 1994) It similarly assumes that the
teacher's plan can be cumulative, with units chosen and sequenced on non-
acquisitional grounds (Breen 1984) Further, it generally assumes that there is a
linear sequence to learning the units of language when they are covered, they
are learned, and do not require cyclical revisiting and extension Finally, there is
the problem that the practice model which underlies the development of habits
in such a *3Ps' approach, has been itself widely discredited (Stern 1983) So,
although there is a sense in which the three terms of restructuring, accuracy, and
fluency could be associated with presentation, practice, and production, the
match is, on examination, clearly inappropriate
A task-based approach, in contrast, may achieve the goal of restructuring, if it
is assumed that interaction-opportunities have an 'extending' influence on
interlanguage development, and engage acquisitional processes But we have
 PETER SKEHAN 51
seen that there is also the possibility of over-pnontization of attentional
resources towards fluency, so that the proceduralization of a lexicahzed
competence emphasizing accessibility is at the expense of complexity-1
restructuring and of accuracy To overcome the difficulties of achieving such
conflicting goals, task-based instruction needs to find ways of balancing
attention allocation The next section addresses this issue

TASK-BASED INSTRUCTION AVOIDING THE DANGERS

The linguistic analysis presented earlier highlights the meaning-dnven nature of
much communication, and the problems this poses for interlanguage develop-
ment The studies from cognitive psychology and psychohnguistics reinforce
this meaning-dnven nature of communication, and link it to (a) the relevance of
a limited-capacity attentional system in which meaning is prioritized for the
consumption of scarce resources, and (b) how a dual-mode system may be used
to ease the processing burdens that are the consequence of the limited
capacities, such that a rule-based system is often implicated, but when
processing demands are high, a lexically-organized system comes into opera-
tion Finally, the task-based learmng research is relevant, but not in a systematic
way Proposals are accountable to such research but, given its incomplete
nature, it cannot be the guiding basis for proposals on task-based instruction
What is proposed in the remainder of the article is a framework to address
these problems Although it is consistent with underlying disciplines and
researchfindingsit contains an essentially speculative component, in that it tnes
to provide an orgamzational framework which can guide pedagogic decisions,
and which can structure the ways in which task-based instruction is imple-
mented It also provides a framework to which existing research can be related,
and which could be useful for decision-making, and organizing future research
The underlying themes for this discussion are that task-based learning should
work towards a constant cycle of analysis and synthesis, that this should be
achieved by manipulating the focus of attention of the learners, and that there
should be balanced development towards the three goals of restructuring,
accuracy, and fluency In practical terms, how these aims can be achieved in
task-based learning can best be discussed in terms of the traditional categones
of syllabus, 1 e sequencing of tasks, and then methodology, 1 e implementing the
tasks which have been chosen

Sequencing tasks syllabus considerations

A fundamental tension in communicative language teaching is that it tnes to
bnng together form and meaning (Van Patten 1990) in that the learner has to
have something worthwhile to say But if there is something worthwhile to say
(a) content may become of pnmary importance, and (b) concern with content
will consume attentional resources (Van Patten 1994) It is imperative,
therefore, that tasks are sequenceable on some pnncipled cntenon, since the
basis on which tasks are ordered will be a reflection of what attentional
resources they require
52 TASK-BASED INSTRUCTION

Table 1 Task sequencing features

Code complexity

Cognitive complexity
Cognitive processing
Cognitive familiarity

Communicative stress
Time pressure
Modality
Scale
Stakes
Control

Developing this general approach, the scheme shown in Table 1 is proposed

for such task sequencing, drawing on previous work by Candhn (1987) and
Nunan (1989) In brief, the scheme contrasts formal factors (code complexity)
with content (cognitive complexity) and pressure to achieve communication
(communicative stress) Code complexity is concerned with traditional areas of
syntactic and lexical difficulty and range and will not be pursued further here
Cognitive complexity is concerned with the content of what is said, and relates
to the conceptualization stage of Levelt's (1989) model It distinguishes
between the two areas of processing and familiarity Processing is concerned
with the amount of on-line computation that is required while doing a task, and
highlights the extent to which the learner has to actively think through task
content Familiarity, in contrast, involves the extent to which the task draws on
ready-made or pre-packaged solutions It is implicated when all that is required
is the accessing of relevant aspects of schematic knowledge if such knowledge
contains relevant, already-organized material, and even solutions to com-
parable tasks, e g sensitivity to macrostructures in narratives
Communicative stress concerns a group of factors unrelated explicitly to
code or meaning, but which do have an impact upon the pressure of
communication Time pressure is perhaps the most straightforward it concerns
how quickly the task has to be done, and whether there is any urgency in the
manner in which it is done (Bygate 1987) Some tasks have a time limit, while
others can be done at the speed the learners choose Modality simply concerns
the speaking/wnting, and listening/reading contrast It is assumed that speaking
leads to more pressure than wnting, and listening more pressure than reading
(Ellis 1987) Scale refers to a range of factors associated with task-based
approaches to teaching It includes the number of participants in the task, the
number of relationships involved etc (Brown, Anderson, Shilcock, and Yule,
1984) Stakes depend on how important it is to do the task, and, possibly, to do
it correctly If the process itself is the mam thing, and there are no consequences
that follow from task completion then stakes are low If, on the other hand, it is
 PETER SKEHAN 53
important not to make mistakes while doing the task, then the stakes are high
(Willis 1993) Finally, control refers to the extent to which the participants
within a task can exert an influence on the task and on how it is done Task goals
can be negotiated, or if participants can ask clarification questions to reduce the
speed of the input they receive, then one can conclude that control is higher, and
task difficulty correspondingly lower (Pica et al 1993)
The purpose of having a system such as this is that it allows tasks to be
analysed, compared, and, best of all, sequenced according to some principled
basis The rewards, if tasks are well-chosen, are
An effective balance between fluency and accuracy Prioritizing fluency as a
goal will emphasize iexicalized language production, strategic language use,
and the primacy of meaning Accuracy as a goal will require analysis, rule-
focus, and attention directed to computation It is difficult to achieve each of
these goals simultaneously, but at least tasks of appropnate difficulty will
give learners some chance of directing balanced attention to each of these
areas (Schmidt 1990) and operate a dual-mode system (Carr and Curren
1994)
The opportunity for previous restructuring to be applied By enabling
attentional spare capacity, there will be some chance that previous
restructuring can be incorporated into ongoing language use (Swain, in
press) and a wider repertoire of language be supported (Crookes 1989,
FosterandSkehanl994)
In contrast, bad task choice will probably lead to the opposite outcome Tasks
which are too difficult are likely to over-emphasize fluency, as learners only
have the attentional capacity to convey meanings, using production strategies
(Faerch and Kaspar 1983), Iexicalized language, and making meaning pnmary
(Bygate 1988) The result is that accuracy is seen as less important, or at least,
less feasible (Ellis 1987) Similarly, embryonic restructuring, which needs to be
integrated carefully into more fluent performance, will not have sufficient
attentional capacity to allow it to be exploited (Schachter 1974) Finally, there is
the contrasting danger that if tasks are too easy, they will present no challenge,
and are not likely to extend any other goals of restructunng, accuracy, or fluency
in any effective way

Implementing tasks methodology

It is also important to consider how tasks, once chosen, are actually imple-
mented One can distinguish three major stages in such a methodological
implementation These are shown in Table 2 The general purpose of the pre-
emptive, or pre-task activities is to increase the chance that some restructuring
will occur in the underlying language system, and that either new elements will
be incorporated, or that some re-arrangement of existing elements will take
place (Foster and Skehan 1994) Within this general purpose, there are two
more specific aims First of all, pre-task activities can aim to teach, or mobilize,
or make salient language which will be relevant to task performance This can be
54 TASK-BASED INSTRUCTION

Table 2 Methodological stages in implementing tasks

Slage Goal Typical techniques

Pre-emptive work Restructuring Consciousness-raising

establish target language Planning
reduce cognitive load
During Mediate accuracy and fluency Task Choice
Pressure Manipulation
POM 1 Discourage excessive fluency Public Performance
Encourage accuracy and Analysis
restructuring Testing
Post 2 Cycle of synthesis and analysis Task Sequences
Task Families

attempted in a number of different ways One, the most traditional, would be an

approach which simply tries to set up the relevant language for a task, in which
case one is essentially dealing with some form of pre-teaching, whether explicit
or implicit More radically, pre-task concern with language may not try to
predict what language will be needed, but instead give learners a pre-task to do,
and then equip them with the language that they need (Prabhu 1987, Willis and
Willis 1988) On this view, the task itself would be the primary factor, and task-
completion would be the aim that would dominate
The second major type of pre-task activity would be to ease the processing
load that learners will encounter when actually doing a task, releasing more
attention for the actual language that is used (Van Patten 1994) The result will
be that more complex language can be attempted (Crookes 1989) and greater
accuracy can be achieved as well (Skehan and Foster, forthcoming) A range of
activities can be used to reduce cognitive complexity in this way The cognitive
familiarity of the task can be altered by pre-task activation sessions, where
learners are induced to recall schematic knowledge that they have that will be
relevant to the task they will do. The cognitive processing load during the task to
come can also be influenced by a number of procedures Learners could
observe similar tasks being completed on video, or they could listen to or read
transcripts of comparable tasks (Willis and Willis 1988) Learners could
similarly be given related pre-tasks to do (Prabhu 1987) so that they have
clearly activated schemas when the real task is presented Finally, and very
importantly, learners could be asked to engage in pre-task planning (Crookes
1989), either of the language that they will need to use, or ot the meanings that
they want to express (Foster and Skehan 1994) Then they can devote more
attention to how they are going to carry out the task, and can thereby produce
more accurate, complex, and fluent language (Foster and Skehan 1994, Skehan
and Foster, forthcoming)
 PETERSKEHAN 55
The main factor affecting performance during the task is the choice of the task
itself, with the goal (see Table 1) being to target tasks which are of the
appropriate difficulty Tasks, that is, should not be so difficult that excessive
mental processing is required simply to communicate any sort of meaning If
they do, it may produce a reliance on ellipsis, context, strategies, and lexicahza-
tion (see above) which reduces the pedagogic value of a task-based approach
Nor should tasks be so easy that learners are bored, and do not engage seriously
with the task requirements, with the result that no gain is made in terms of
stretching interlanguage or developing greater automaticity (Swain 1985)
But in addition to task choice, as discussed in the section on syllabus, there are
implementation decisions that teachers can make to alter the difficulty of a given
task, and manipulate the way in which attention is directed As regards the code
itself, teachers can be explicit immediately before the task is done as to whether
they want accuracy to be stressed, or whether they want specific structures to be
used, (I e a pressure to conformity in structure choice (Willis 1993)) As regards
cognitive complexity, there are ways of making a task less or more difficult To
achieve the former, visual support could be provided, such as a diagram, which
can ease the amount of matenal that learners need to keep in mind while
responding to the task itself To make tasks more difficult, surprise elements can
be introduced which do not match learner expectations of what the task will
require, e g additional evidence in a 'judge' task
But perhaps the major area for adjustment while tasks are being completed is
in the area of stress (or communicative pressure) Pressure manipulation can be
based on the communicative stress factors mentioned earlier in the section on
sequencing, I e
time
modality
scale
stakes
control
Since the operation of these factors was described in that earlier section, no
additional coverage will be provided here The point is simply that they are
susceptible to variation, with consequent impact on communicative pressure
Finally, we need to look at post-task activities The assumption made here is
that learners' knowledge of what is to come later can influence how they
approach attention-management dunng an actual task The central problem is
that while a task is being done, the teacher needs to withdraw, be non-
mterventiomst, and allow natural language acquisitional processes to operate
(Brumfit 1984) But then, the danger is that communication goals will be so
predominant that lexicahzed communication strategies will become so
important that the capacity to change and restructure, to take syntactic nsks,
and to try to be more accurate, will not come into focus as serious goals, and
worthy of attention dunng the intensity of task completion (Skehan 1992) Post-
task activities can change the way in which learners direct their attention dunng
56 TASK-BASED INSTRUCTION

the task (Willis and Willis 1988,Tarone 1983) They achieve this by reminding
learners that fluency is not the only goal during task completion, and that
restructuring and accuracy also have importance
Drawing on Table 2, two phases of post-task activities can be used In 'Post 1",
the more immediately linked to the teaching which has just occurred, three
general post-task activities can be mentioned public performance, analysis, and
tests With the first of these, public performance, learners will be asked, after
they have completed a task 'in the privacy of their own group' to repeat their
performance, publicly, in front of some sort of audience The audience could be
the rest of a learning group, (who themselves may also have been doing the same
or a similar task, and who could equally well be asked to engage in the public
performance), the teacher, or even a video camera, so that the performance
could be played back later, with even the participants themselves required to
watch In this way, the knowledge while the task is being done that a task may
have to be re-done publicly will cause learners to allocate attention to the goals
of restructuring and accuracy where otherwise they would not In this way, a
concern with syntax and analysis can be infiltrated into the task work without
the heavy-handedness of teacher intervention and error correction
There are also other post-task aspects of task-based learning which are
important, as shown in 'Post 2' from Table 2 One must examine task sequences,
task progression, and generally how sets of tasks relate to one another, and to
the underlying and more important goals which are driving forward instruction
For example, there may be reasons to repeat tasks, with the idea that learners
will be more effective with the analysis and synthesis goals and that the task was
meant to embody Similarly, there may be parallel tasks Such tasks are likely to
be similar to one another in some important respect, but at the same time
contain new elements which are sufficient to engage the interest of the learner
(Plough and Gass 1993) Perhaps most generally of all, it is useful to think in
terms of 'task families', where a group of tasks resemble one another and may
well have similar language or cognitive demands (Candhn 1987) In this way,
learners will be clearer about the goals of such task groups, and there will be less
tendency for discrepancies to arise between teachers' and learners' views about
task requirements
These various methods of analysing tasks, in terms of syllabus and
methodology, are brought together in Table 3 This table shows how the three
major stages of task implementation, together with the associated goals in each
case, cross-reference with syllabus design factors In this way, it is possible to see
how systematic decisions can be made regarding the change which it is intended
should occur in learners' interlanguage systems One aspect of this table which
should be noted in passing is that it suggests that the syllabus-methodology
distinction is still relevant, even for task-based learning Nunan (1993) argues
that this is not so because we learn to communicate by communicating, we
cannot so easily separate the target from the means of achieving it The
discussion m the last few pages, however, suggests that while we cannot pretend
to offer a comprehensive sequence of tasks, there are methods of analysing
 PETER SKEHAN 57

Table 3 Factors influencing task implementation

Stage Code complexity Stress Cognitive complexity

Goal

Pre-emptive
Restructuring Pre-teach Processing
establish target Consciousness-raising i
) Observe
language Practice Solve similar tasks
reduce cognitive conventional Plan
load parallel tasks cognmvely
rehearsal of linguistically
elements Familiantv
Activate
During TASK C H O I C E
Mediating accuracy Accuracv focus Time Support available
and fluencv Conformity pressure Modalitv Surprise elements
Scale additional
Stakes conflicting
Control
Postl
Increases accuracy Public performance
Encourages teacher
restructuring - group
Discourages excessive camera
synthesis Degree of analysis
Testing
Post 2 The task sequence
Cycle of synthesis and repeating
analysis parallel tasks
Task families

tasks, both for difficulty and for type, and that as a result, we can try to work with
syllabus units in a well-defined and principled way Similarly, viewing task
implementation in terms of the three phases of pre, during, and post clearly
indicates where methodological choices are relevant in task-based learning, that
these choices take as input the units from a syllabus specification, and that the
choices themselves are methodologically motivated So it is argued that the
syllabus-methodology distinction can still be relevant, provided that a
framework such as that advocated here is used

CONCLUSIONS
Task-based learning is an area which has grown in importance enormously
during the last ten years, and can now be approached from a number of perspec-
tives The present paper has taken a processing-pedagogic viewpoint and its
main ideas are
58 TASK-BASED INSTRUCTION

Task-based learning, a current vogue in communicative language teaching,

contains dangers if implemented without care In particular, it is likely to
create pressure for immediate communication rather than interlanguage
change and growth In the process, it may encourage learners to use
excessively and prematurely lexical modes of communication
It is possible to draw on cognitive psychology and second language
acquisition research, at least of the sort that emphasizes processing factors,
to propose a framework which avoids or at least minimizes these dangers
Task-based learning is an attempt to address one of the dilemmas of language
teaching how, on the one hand, to confront the need to engage naturalistic
learning processes, while, on the other, to allow the pedagogic process to be
managed in a systematic manner The proposals outlined in this paper contain
partial, but not complete, solutions to this dilemma The paper accepts that
language learning is not any sort of simple, linear, cumulative process Instead,
learners must be able to develop their mterlanguage^systems in more complex
ways, through cycles of analysis and synthesis revisiting some areas as they are
seen to require complexification, learning others in a simple, straightforward
manner, developing others by simply relexicalizing that which is available
syntactically, but which need not be used on such a basis The proposals
presented here attempt to offer such systematization as is possible within such a
complex situation It attempts, that is, to structure the freedom which learners
need to have' It does so, above all, by trymg to address the issue of attention, or
learners' capacities to focus their attention In this way, it is hoped that, however
inexact our understanding of language learning, the greatest chance is being
created for naturalistic mechanisms and processes to come into play
Clearly, this analysis is more programmatic than based on a range of
completed studies The framework which was presented in Table 3 is an attempt
to synthesize what is known about the influence of task variation on learning and
performance Some parts of this table are supported by empirical work In other
places, this is much less true The table, that is, attempts to be consistent with the
evidence that is available, but goes beyond it to try to establish a more general,
and therefore useful, framework But in other ways, what the table does is to
demonstrate how much research is needed to investigate the claims that it
makes If it serves any function in addition to that of utility, it has to be that it
provides some sort of organizational framework which can stimulate research,
and within which future research can be located
(Revised version received April 1995)

ACKNOWLEDGEMENTS
The author would like to thank Antony Bruton, Graham Crookes, Pauline Foster, and
three anonymous reviewers, who read earlier versions of this article Needless to say, any
remaining errors or shortcomings are the responsibility of the author
 PETERSKEHAN 59

NOTES
1
As an Applied Linguistics reviewer pointed out, this represents a narrow
interpretation of what a task is The justification for this, which should become clearer as
the argument develops, is that the narrower approach draws upon acquisition research
more directly, assuming thai mterlanguage development is a key goal for pedagogy,
particularly in relation to the development of spoken language ability
2
Although it is recognized that to accept this goal makes a number of assumptions, it is
accepted here for expository purposes, but is challengable on a number of grounds First
of all, there is the issue of what native-like means There is also the issue that many
language learners may have other models of competence that they aspire to, rather than a
particular native-speaker version Finally, there are learners who reject a native-speaker
model completely or partially, which complicates the picture considerably
1
Again, for expository purposes, the concentration here is on what Bachman and
Palmer (in press) would term Organizational Competence, and does not concern their
Pragmatic Competence

REFERENCES
Anderson, A andT Lynch 1987 Listening Oxford Oxford University Press
Anderson, J R. 1989 'Practice, working memory, and the ACT theory of skills
acquisition A comment on Carlson, Sullivan, and Schieder" Journal of Learning,
Memory, and Cognition 15 527-30
Bachman, L. and A Palmer, (in press) Language Jesting in Practice Oxford Oxford
University Press
Bates, E , I. Bretherton, and L Snyder 1988 From First Words to Grammar
Cambridge Cambridge University Press
Berwick, R. 1993 "Towards an educational framework for teacher-led tasks' in S Gass
and G Crookes (eds ) 1993a
Bohnger, D. 1975 "Meaningand memory Forum Ltnguisticum 1 2-14
Breen, M. 1984 "Process in syllabus design' in C Brumfit (ed ) 1984 General English
Syllabus Design Oxford Pergamon
Breen, M. 1987 "Contemporary paradigms in syllabus design (Parts 1 and 2)' Language
reaching2Q 91-2 and 157-74
Brown, G , A Anderson, R. Shilcock, and G Yule. 1984 Teaching Talk Strategies for
Production and Assessment Cambndge Cambridge University Press
Brumfit, C 1984 Communicative Methodology in Language Teaching Cambndge
Cambndge University Press
Bygate, M. 1987 Speaking Oxford Oxford University Press
Bygate, M. 1988 'Units of oral expression and language learning in small group
interaction' Applied Linguistics 9/1 59-82
Candlin, C 1987 Towards task based language learning" in C Candlin and D Murphy
(eds) 1987 Language Learning Tasks Englewood Cliffs, NJ Prentice Hall
Carr, T. and T Curren. 1994 'Cognitive factors in learning about structured sequences
Applications to syntax,' Studies in Second Language Acquisition 16/2 205-30
Cheng, P W. 1985 'Restructuring versus automaticity Alternative accounts of skill
acquisition' Psychological Review 92 414-23
Clark, H. H. and E. Clark 1977 Psychology and Language Harcourt, Brace,
Jovanovich
Cook, V. J. 1994 Linguistics and Second Language Acquisition London Macmillan
60 TASK-BASED INSTRUCTION

Corder, S Pit 1981 Error Analysis and Interlanguage Oxford Oxford University
Press
Crookes, G. 1989 'Planning and interlanguage variation' Studies in Second Language
Acquisition 11 367-83
Duff, P. 1986 'Another look at interlanguage talk Taking task to task' in R Day (ed )
1986 Talking to Learn Rowley, MA Newbury House
Ellis, R 1987 'Interlanguage variability in narrative discourse Style shifting in the use of
the past tense' Studies in Second Language Acquisition 9 12-20
Ellis, R. 1994 The Study of Second Language Acquisition Oxford Oxford University
Press
Faerch, C and G. Kasper. 1983 Strategies in Interlanguage Communication London
Longman
Foster,P.andP Skehan 1994 'Themfluenceofplanmngonperformanceintask-based
learning' Paper presented at the British Association of Applied Linguistics, Leeds
Gass, S and G. Crookes (eds) 1993a Tasks in a Pedagogical Context Integrating
Theory and Practice Clevedon, Avon Multilingual Matters
Cass, S and G. Crookes (eds) 1993b Tasks and Languge Learning Integrating Theory
and Practice Clevedon, Avon Multilingual Matters
Gnce, H P 1975 'Logic and conversation'in P Cole and J Morgan (eds) 1975 Syntax
and Semantics 3 Speech Acts, New York Academic Press
Karmilloff-Smith, A. 1986 'From meta-processes to conscious access Evidence from
children's metalinguistic and repair data" Cognition 23 95-147
Kellerman, E. 1991 'Compensatory strategies in second language research A critique, a
revision, and some (non-)implications for the classroom' in R Phillipson, E
Kellerman, L Sehnker, M Sharwood-Smith, and M Swain (eds) 1991 Foreign/
Second Language Pedagogy Research Clevedon, Avon Multilingual Matters
Kess, J E. 1992 Psycholinguistics, Amsterdam John Benjamins
Klein, W. 1986 Second Language Acquisition Cambridge Cambridge University Press
Larsen-Freeman, D andM. Long. 1991 An Introduction to Second Language Acquisi-
tion Research London Longman
Levelt, W. J. M. 1989 Speaking From Intention to Articulation Cambridge, MA MIT
Press
Littlewood, W. 1981 Communicative Language Teaching An Introduction Cambridge
Cambridge University Press
Logan, G. D. 1988 'Towards an instance theory of automatisation' Psychological
Review 95 492-527
Long, M. and G. Crookes. 1991 "Three approaches to task-based syllabus design'
TESOL Quarterly 26/1 27-55
Loschky, L and R. Bley-Vroman. 1993 'Grammar and task-based methodology' in S
Gass and G Crookes (eds) 1993b
Matthews, R C , R. R Buss, W. B. Stanley, F Blachard-Fields, J R. Cho, and B
Druhan. 1989 'Role of implicit and explicit processes in learning from examples A
synergistic effect' Journal of Experimental Psychology Learning, Memory, and
Cognition 15 1083-100
McLaughlin, B. 1990 'Restructuring' Applied Linguistics 1 1 1 1 3 - 2 8
Nattinger, J. R. and J. S. DeCarrico. 1992 Lexical Phrases and Language Teaching
Oxford Oxford University Press
Nelson, K. 1981 'Individual differences in language development Implications for
development and language' Developmental Psychology 17 170-87
 PETERSKEHAN 61

Nunan, D 1989 Design Tasks for the Communicative Classroom Cambridge

Cambridge University Press
Nunan, D 1993 Task-based syllabus design Selecting, grading, and sequencing tasks'in
S Gass and G Crookes (eds) 1993a
Pawley, A. and F. Syder 1983 'Two puzzles for linguistic theory Nativehke selection
and nativehke fluency' in J C Richards and R Schmidt (eds) 1983 Language and
Communication London Longman
Peters,A 1983 The Units of Language Acquisition Cambndge Cambridge University
Press
Peters, A 1985 "Language segmentation Operadon principles for the perception and
analysis of language' in D Slobin (ed ) 1985 The Crosshnguistic Study of Language
Acquisition Vol 2 Theoretical Issues London Erlbaum Associates
Pica,T., R Kanagy, andJ Falodun. 1993 'Choosing and using communicative tasks for
second language instruction' in S Gass and G Crookes (eds) 1993b
Plough, I and S. Gass. 1993 'Interlocutor and task familiarity Effect on interactional
structure1 in S Gass and G Crookes (eds) 1993b
Prabhu, N S 1987 Second Language Pedagogy Oxford Oxford University Press
Reber, A. 1989 implicit learning and tacit knowledge* Journal of Experimental
Psychology General US 219-35
Rivers, W. 1981 Teaching Foreign Language Skills Chicago University of Chicago
Press
Robinson, P J.andM A. Ha. 1993 'Instance theory and second language rule learning
under explicit conditions' Studies in Second Language Acquisition 15/4 413-38
Schachter, J. 1974 'An error in error analysis' Language Learning 24 205-14
Schiffrin, R. M. and W Schnieder. 1977 'Controlled and automatic information
processing II Perceptual learning, automatic attending, and general theory Psycho-
logical Review 84 127-90
Schmidt, R. 1983 'Interaction, acculturation, and the acquisition of communicative
competence" in N Wolfson and E Judd (eds) 1983 Sociohngutstics and Second
Language Acquisition Rowley, MA Newbury House
Schmidt, R. 1990 'The role of consciousness in second language learning' Applied
Linguistics 11 17-46
Schmidt, R. 1992 'Psychological mechanisms underlying second language fluency'
Studies in Second Language Acquisition 14 357-85
Schmidt, R 1994 'Deconstructing consciousness In search of useful definitions for
Applied Linguistics ' AILA Review 11 11-26
Sheen, R. 1994 'A critical analysis of the advocacy of a task-based syllabus' TESOL
Quarterly 28/1 127-51
Sinclair, J. 1991 Corpus, Concordance, Collocation Oxford Oxford University Press
Skehan, P 1992 'Strategies in second language acquisition' Thames Valley University
Working Papers in English Language Teaching, No 1
Skehan, P 1994 'Foreign language learning ability Cognitive or linguistic9' Thames
Valley University Working Papers in English Language Teaching, Vol 2
Skehan, P. 1995 'Analysabihty, accessibility, and ability for use1 in G Cook and B Seidl-
hofer (eds) 1995 Principle and Practice in Applied Linguistics Studies in Honour of
H G Widdowson Oxford Oxford University Press
Skehan, P (in press) implications of SLA research for language teaching methodology'
in D Willis and J Willis (eds) in press
Skehan, P. and P. Foster (forthcoming) The influence of planning and post-task
 62 TASK-BASED INSTRUCTION

activities on accuracy and complexity in task based learning' Thames Valley University
Working Papers in English Language Teaching, Vol 3
Stern, H. H. 1983 Fundamental Concepts of Language Teaching Oxford Oxford
University Press
Swain, M. 1985 'Communicative competence Some roles of comprehensible input and
comprehensible output in its development' m S Gass and C Madden (eds) 1985
input in Second Language Acquisition, Rowley, MA Newbury House
Swain, M. 1995 Three functions of the output m second language learning' in G Cook
and B Seidlhofer (eds) Principle and Practice in Applied Linguistics Studies in
Honour of H G Wtddowson
Tarone, E 1983 'On the variability of mterlanguage systems' Applied Linguistics 4
143-63
Tarone, E. 1985 "Variability in mterlanguage use A study of style-shifting in
morphology and syntax' Language Learning 35 373-403
Trevise, A. and C. Noyau. 1984 'Adult Spanish speakers and the acquisition of French
negation forms Individual variation and linguistic awareness' in R W Andersen (ed)
1984 Second Languages Rowley, MA Newbury House
Van Patten, B. 1990 'Attending to content and form m the input An experiment in
consciousness' Studies in Second Language Acquisition 12 287-301
Van Patten, B. 1994 'Evaluating the role of consciousness in SLA Terms, linguistic
features, and research methodology' AILA Review 11 27-36
Van Patten, B. and T Cadierno. 1993 'Explicit instruction and input processing'
Studies in Second Language Acquisition 15 225-43
Wells, G. 1985 Language Development in the Preschool years Cambridge Cambridge
University Press
Widdowson, H. G 1983 Learning Purpose and Language Use Oxford Oxford
University Press
Widdowson, H. G 1989 'Knowledge of language and ability for u s e ' Applied
Linguistics 10/2 128-37
Willis, D. 1993 'Syllabus, corpus, and data-driven learning' in C Kennedy (ed) 1993
Plenanesfrom the 1993IA TEFL Conference Swansea, Wales, IATEFL
Willis, D. and J.Willis. 1988 COBUILD Book 1 London Collins
Willis, D. and J. Willis, in press Language Teaching Alternatives Methodological
Innovation in English Language Teaching Oxford Heinemann
Wilson, D. 1994 'Context and relevance' in G Brown, K Malmkjaer, A Pollm, and J
Williams (eds) 1994 Language and Understanding Oxford Oxford University Press
Yule, G., M. Powers, and D. Macdonald. 1992 The variable effects of some task-based
learning procedures on L2 communicative effectiveness' Language Learning 42/2
249-77

View publication stats