Vous êtes sur la page 1sur 26

Teachers of English to Speakers of Other Languages, Inc.

(TESOL)
Current Issues in the Teaching of Grammar: An SLA Perspective
Author(s): Rod Ellis
Source: TESOL Quarterly, Vol. 40, No. 1 (Mar., 2006), pp. 83-107
Published by: Teachers of English to Speakers of Other Languages, Inc. (TESOL)
Stable URL: http://www.jstor.org/stable/40264512
Accessed: 11-08-2014 07:33 UTC

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content
in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship.
For more information about JSTOR, please contact support@jstor.org.

Teachers of English to Speakers of Other Languages, Inc. (TESOL) is collaborating with JSTOR to digitize, preserve and
extend access to TESOL Quarterly.

http://www.jstor.org

This content downloaded from 147.91.1.45 on Mon, 11 Aug 2014 07:33:07 UTC
All use subject to JSTOR Terms and Conditions

CurrentIssues in theTeachingof
Grammar:An SLA Perspective
ROD ELLIS
University
ofAuckland
Auckland,New Zealand

The studyofhowlearnersacquirea secondlanguage(SLA) has helped


to shape thinkingabout how to teach the grammarof a second
issues.
language.There remain,however,a numberof controversial
Thispaperconsiderseightkeyquestionsrelatingto grammarpedagogy
in the lightof findingsfromSLA. As such, thisarticlecomplements
Celce-Murcia's
(1991) articleon grammarteachingin the25thanniverTESOL
whichconsideredtheroleofgrammarin
issue
of
Quarterly,
sary
on a linguistic
curriculumand drewpredominantly
a communicative
theoryof grammar.These eightquestionsaddresswhethergrammar
should be taughtand ifso whatgrammar,
when,and how.Although
solutionsto thesequestions,itservesthe
SLA does notafforddefinitive
thisaspectof languagepedagogy.
valuablepurposeof problematising
This article concludes with a statementof my own beliefsabout
of SLA.
grammarteaching,groundedin myownunderstanding
articleidentifiesand discusses a number of keyissues relatingto
the teaching of grammarin a second language (L2) and, bydrawing
on theoryand research in SLA, suggestswaysto address these problems.
It points to a number of alternativesolutions to each problem, indicating
thatmore oftenthan not there are no clear solutions currentlyavailable.
The aim, therefore,is not to identifynew solutions to existingcontroversies, nor even to present new controversies.Rather it addresses within
the compass of a single articlea whole range of issues related to grammar
teaching,problematisesthese issues,and byso doing, providesa counterweight to the advocacy of specific,but also quite limited,proposals for
teaching grammarthat have originated in some SLA quarters. However,
I conclude witha statementof myown position on these issues.
The questions thatwill be addressed are
1. Should we teach grammar,or should we simplycreate the conditions
by which learners learn naturally?
2. What grammarshould we teach?
TESOL QUARTERLY Vol. 40, No. 1, March2006

This content downloaded from 147.91.1.45 on Mon, 11 Aug 2014 07:33:07 UTC
All use subject to JSTOR Terms and Conditions

83

3. Whenshouldwe teachgrammar?Is it bestto teachgrammarwhen


learnersfirststartto learnan L2 or to waituntillaterwhenlearners
havealreadyacquiredsome linguistic
competence?
4. Should grammarinstruction
be massed (i.e., the availableteaching
timebe concentratedinto a shortperiod) or distributed(i.e., the
availableteachingtimespreadovera longerperiod)?
5. Should grammarinstruction
be intensive(e.g., covera singlegramin a singlelesson) or extensive(e.g., covermany
maticalstructure
in a singlelesson)?
structures
grammatical
6. Is thereanyvalue in teachingexplicitgrammatical
knowledge?
7. Is therea bestwayto teachgrammarforimplicitknowledge?
8. Should grammarbe taughtin separatelessonsor integratedinto
communicative
activities?

DEFINING GRAMMARTEACHING
Traditionally,
grammarteachingis viewedas the presentationand
of
structures.
Thisis theviewpromulgated
discrete
practice
grammatical
in teacherhandbooks.Ur (1996), for example,in her chaptertitled
and explaininggram"TeachingGrammar"has sectionson "presenting
in her chapter
mar"and "grammarpracticeactivities."
(2000)
Hedge
titled"Grammar"similarlyonly considers"presentinggrammar"and
"practisinggrammar."This constitutesan overlynarrowdefinitionof
truethatgrammarteachingcanconsist
grammarteaching.It is certainly
of the presentationand practiceof grammaticalitems.But, as will
becomeapparent,itneed not.First,somegrammarlessonsmightconsist
of presentation
byitself(i.e., withoutanypractice),whileothersmight
entailonlypractice(i.e., no presentation).Second, grammarteaching
can involvelearnersin discoveringgrammaticalrules for themselves
and no practice).Third,grammarteachingcan be
(i.e., no presentation
conductedsimplyby exposinglearnersto inputcontrivedto provide
multipleexemplarsof the targetstructure.Here, too, there is no
and no practice,at leastin thesenseofelicitingproduction
presentation
of thestructure.
Finally,
grammarteachingcan be conductedbymeans
ofcorrective
feedbackon learnererrorswhenthesearisein thecontext
of performing
some communicative
task.The definitionof grammar
is
that
informs
this
article
a
broad
one:
teaching
Grammarteachinginvolves any instructionaltechnique that draws learners'
attentionto some specificgrammaticalformin such a waythatit helps them
eitherto understand it metalinguistically
and/or process it in comprehension
and/or production so that theycan internalize it.
84

TESOL QUARTERLY

This content downloaded from 147.91.1.45 on Mon, 11 Aug 2014 07:33:07 UTC
All use subject to JSTOR Terms and Conditions

SHOULD WE TEACH GRAMMAR?


This questionwas motivatedby earlyresearchinto naturalistic
L2
acquisition,whichshowed thatlearnersappeared to followa natural
orderand sequence of acquisition(i.e., theymastereddifferent
gramin a relatively
fixedand universalorder and they
maticalstructures
passedthrougha sequenceofstagesofacquisitionon routeto mastering
This led researchers
likeCorder(1967) to
each grammatical
structure).
suggestthatlearnershad theirown built-insyllabusforlearninggrammar.In line withthis,Krashen(1981) arguedthatgrammarinstruction
played no role in acquisition,a view based on the convictionthat
learners(includingclassroomlearners)would automatically
proceed
along theirbuilt-insyllabusas long as theyhad access to comprehensible
motivated.Grammarinstruction
could coninputand weresufficiently
tributeto learningbut thiswas of limitedvalue because communicative
abilitywasdependenton acquisition.
There followeda number of empiricalstudies designed to (a)
and naturalistic
learners
comparethe orderof acquisitionof instructed
and naturalistic
(e.g.,Pica, 1983), (b) comparethesuccessofinstructed
learners (Long, 1983) and (c) examine whetherattemptsto teach
structures
resultedin theiracquisition(e.g.,White,
specificgrammatical
& Ranta,1991). These studiesshowedthat,by and
Spada, Lightbown,
and naturalislarge,theorderofacquisitionwasthesameforinstructed
tic learners(althoughtherewere some interesting
that
differences1),
instructedlearnersgenerallyachieved higher levels of grammatical
competence than naturalisticlearnersand that instructionwas no
guaranteethatlearnerswouldacquirewhattheyhad been taught.These
resultswere interpreted
as showingthatthe acquisitionalprocessesof
instructed
and naturalistic
learningwere the same but thatinstructed
learnersprogressedmore rapidlyand achievedhigherlevelsof proficiency.Thus, some researchersconcluded (e.g., Long, 1988) that
wasbeneficialbutthatto be effective
teachinggrammar
grammarhad to
be taughtin a waythatwas compatiblewiththe naturalprocessesof
acquisition.
Subsequent research,such as Noms and Ortega's (2000) metaofgrammar
analysisof49 studies,has borneout theoveralleffectiveness
thereis evidencethat,contrary
to Krashen's(1993)
teaching.Further,
contributesto both acquired knowledge
continuedclaims,instruction
(see Ellis,2002a) as wellas learnedknowledge.There is also increasing
1For
example, Pica (1983) notes that some structures(e.g., plural-s) were used more
accuratelyby instructedlearnersand some (e.g., Verb-ing) by naturalisticlearners.In other
structures(e.g., articles)therewas no difference.
CURRENT ISSUES IN THE TEACHING OF GRAMMAR

This content downloaded from 147.91.1.45 on Mon, 11 Aug 2014 07:33:07 UTC
All use subject to JSTOR Terms and Conditions

85

evidencethatnaturalistic
learningin the classroom(as, e.g., in immersionprogrammes)does nottypically
resultin highlevelsofgrammatical
In
competence(Genesee,1987). short,thereis nowconvincingindirect
and directevidenceto supportthe teachingof grammar.Nevertheless,
doubtsremainabout the natureof the researchevidence.Manystudies
(includingmost of those reviewedby Norrisand Ortega) measure
constructed
(e.g.,fillin theblanks,
learningin termsofconstrained
responses
sentencejoining,or sentencetransformation),
whichcan be expectedto
favourgrammarteaching.Thereis onlymixedevidencethatinstruction
resultsin learningwhen it is measured by means of freeconstructed
(e.g., communicativetasks). Also, it remainsthe case that
responses
learnersdo not alwaysacquirewhattheyhave been taughtand thatfor
it needs to take account of how
to be effective
grammarinstruction
learnersdeveloptheirinterlanguages.
Aswe willsee, thereis controversy
both
how
interlanguage
developmentoccursand howinstrucregarding
tioncan facilitate
this.

WHAT GRAMMAR SHOULD

WE TEACH?

tointerlanguage
then,thatgrammar
Assuming,
teachingcan contribute
development,the next logical question concernswhat grammarwe
should teach. This question can be broken down into two separate
questions:
1. Whatkindof grammarshouldwe base teachingon?
featuresshouldwe teach?
2. Whichgrammatical
models to choose
Linguisticsaffordsa broad selectionof grammatical
structural
from,including
grammars,
generativegrammars(based on a
of
universal
and
functional
,
grammars.Traditionally
theory
grammar)
syllabuseshave been based on structuralor descriptivegrammars.
Structural
syllabusestraditionally
emphasisedthe teachingofformover
grammeaning(e.g., Lado, 1970). Though the influenceof structural
marsis stillapparenttoday,modernsyllabusesrightly
givemore attention to the functionsperformedby grammaticalforms.Thus, for
example,lessemphasisis placedon suchaspectsofgrammaras sentence
patternsor tense paradigmsand more on the meaningsconveyedby
Some attemptwasonce
different
formsin communication.
grammatical
made to exploitthe insightsto be gleaned fromgenerativetheoriesof
grammar(see, e.g.,Bright,1965), butin general,syllabusdesignersand
to rely
teachershave not foundsuch modelsusefuland have preferred
on modern descriptivegrammars,such as Celce-Murciaand Larsen86

TESOL QUARTERLY

This content downloaded from 147.91.1.45 on Mon, 11 Aug 2014 07:33:07 UTC
All use subject to JSTOR Terms and Conditions

Book.This resourceis especiallyvaluable


Freeman's (1999) Grammar
because it not onlyprovidesa comprehensive,
clear,and pedagogically
of
but
thekinds
exploitabledescription Englishgrammar also identifies
oferrorsthatL2 learnersare knownto makewithdifferent
grammatical
is importantbecause it helps to identify
Such information
structures.
which structuresand which aspects of a structurerequire special
Bookis also ideal in thatit presentsinformation
attention.
The Grammar
formbutalso aboutthesemanticand discoursal
notonlyaboutlinguistic
realised
Williams,and Rott
byparticularforms.As VanPatten,
meanings
form
connections
between
and meaning
(2004) emphasise,establishing
is a fundamentalaspect of language acquisition.Thus, any reference
connectionsof the
grammarthatfailsto describethe form-meaning
In
must
be
necessarily inadequate. general,then,the
targetlanguage
choice ofwhichtypeof grammarto use as a basisforteachingis not a
grammarsthatdetailtheformdescriptive
majorsourceof controversy;
of
the
are
ascendant.
language
meaningrelationships
to teach is
In contrast,the choice of whichgrammaticalstructures
and variouspositions
Twopolarpositionscan be identified
controversial.
in between.At one end of this continuumis Krashen'sminimalist
position.Krashen (1982) argues that grammarteachingshould be
limitedto a fewsimpleand portablerulessuchas 3rdperson-jand past
tense-d
thatcan be used to monitoroutputfromthe acquiredsystem.
He bases hisargumenton theclaimthatmostlearnersare onlycapable
of learningsuch simplerules- thatmore complexrulesare generally
notlearnableor,iftheyare,are beyondstudents'abilityto applythrough
Krashen'sclaim,however,is not warranted.There is now
monitoring.
thatmanylearnersare capable of masteringa wide
evidence
ample
of
range explicitgrammarrules.Greenand Hecht (1992), forexample,
studentsof Englishin Germanywereable to
foundthatuniversity-level
errorstheywere
clear
explanationsfor85% ofthegrammatical
produce
askedto explain,whileoverallthelearnersin theirstudy(whoincluded
explanationsfor46%
secondaryschool students)managedsatisfactory
of the errors.Macroryand Stone (2000) reportedthatBritishcompreof the
hensiveschool studentshad a fairly
good explicitunderstanding
its
in
French
understood
tense
function,theyknew
(e.g., they
perfect
thatsome verbsused avoirand some tre,theywere familiarwiththe
formsrequiredbydifferent
pronouns,and theywereawareof theneed
fora finalaccent on the past participle).Hu (2002) foundthatadult
correctmetalinguistic
knowlChineselearnersof Englishdemonstrated
of
six
the
rules
structures
for
definite
of
(e.g.,
English
edge prototypical
constitutedthe prototypical
articlespecific
rule) but were less
reference
.
clearabouttheperipheralrulesforthesestructures
(e.g.,generic
reference)
Teach
the
whole
of
At the otherpole is the comprehensive
position:
CURRENT ISSUES IN THE TEACHING OF GRAMMAR

This content downloaded from 147.91.1.45 on Mon, 11 Aug 2014 07:33:07 UTC
All use subject to JSTOR Terms and Conditions

87

the grammarof the targetlanguage.2This is the positionadopted by


manycourse book writers(e.g., Walter& Swan, 1990) or authorsof
1994). Such a positionwould
grammarpracticematerials(e.g.,Murphy,
also seemunwarranted
becauselearnersare clearlycapableoflearninga
substantial
amountof the L2 grammarwithoutinstruction
and because
mostteachingcontextshavelimitedtimeavailableforteachinggrammar
so some selectionis needed.
What then should selectionbe based on? The answerwould seem
obvious- theinherentlearningdifficulty
strucofdifferent
grammatical
tures.The problemarisesin how to determinethis.To beginwith,it is
to distinguish
Thiscan
twodifferent
sensesof learning
necessary
difficulty.
referto (a) thedifficulty
a grammatical
learnershavein understanding
featureand (b) to thedifficulty
a grammatical
theyhavein internalising
in communication.
These
featureso thattheyare able to use itaccurately
twosensesrelateto thedistinction
betweenlearninggrammaras explicit
knowledgeand as implicitknowledge,whichis discussedlater.Clearly,
whatis difficult
to learnas explicitknowledgeand as implicitknowledge
in grasping
is notthesame.Forexample,mostlearnershaveno difficulty
in
theruleforEnglishthirdperson-sbuttheyhaveenormousdifficulty
These
two
senses
this
structure
so
can
use
it
accurately.
internalising
they
of learningdifficulty
have not alwaysbeen clearlydistinguishedin
languagepedagogy,withtheresultthatevenwhenthestatedgoal is the
developmentof implicitknowledge,it is the anticipateddifficulty
studentswillhave in understanding
a featurethatguidesthe selection
and gradingof grammatical
structures.
Thirdperson-5,forexample,is
in
a
course.
typically
taughtveryearly
How then has learning difficulty
been established?Traditionally,
in theinputand their
factorssuchas thefrequency
ofspecificstructures
to
learners
have
invoked
been
utility
(Mackey,1976), but thesefactors
wouldseem to have more to do withuse3thanwithinherentcognitive
Here I considertwoapproachesthathavefiguredin attempts
difficulty.
to delineatecognitivedifficulty.
1. Teach thoseformsthatdiffer
fromthelearners'firstlanguage(LI).
2. Teach markedratherthanunmarkedforms.
2 Of course, it is not
possible to specifythe whole grammarof a language. Though the
grammarof a language maybe determinate,descriptionsof it are certainlynot. The Longman
A Grammar
ofContemporary
English(Quirk, Greenbaum,Leech, 8c Svartvik,1972) ran to 1081
pages (excluding index and bibliography)but doubtlesslydoes not account forall the known
factsof Englishgrammar.Nevertheless,thereis a recognizedcanon of Englishstructuresthat,
in the eyesof syllabusdesignersand textbookwriters,
constitutesthe grammarof English.
3Structureslike
Englisharticlesthatare veryfrequentin the inputcan imposeconsiderable
Structuressuch as Englishconditionalsmaybe veryusefulto learnersbutare
learningdifficulty.
to learn.
also difficult
88

TESOL QUARTERLY

This content downloaded from 147.91.1.45 on Mon, 11 Aug 2014 07:33:07 UTC
All use subject to JSTOR Terms and Conditions

The firstapproach was, of course,the one adopted in manyearly


structural
coursesbased on a contrastive
analysisofthelearner'sLI and
the targetlanguage. Althoughthe contrastiveanalysishypothesisas
is clearlynottenable(see Ellis,1985,chapter2), SLA
formulated
initially
at leastsomeofthe
stillgenerally
researchers
agreethatlearnerstransfer
featuresof theirLI into the L2. For example,thereis ample evidence
(Trahey& White,1993) to showthatFrenchlearnersofEnglishproduce
errorsofthekindMarykissed
JohnbecauseFrenchpermitsan
passionately
adverb to be positioned between the verb and the direct object.
contrastive
Nevertheless,
analysisdoes not constitutea sound basis for
In manyteachingcontexts,the learnstructures.
selectinggrammatical
ers come frommixedlanguagebackgroundswhereit wouldbe impossibleto use contrastive
analysisto tailorgrammarteachingto theentire
Lis. Also,we simplydo notyet
havedifferent
learners
the
because
group
does and does not translateinto
knowenough about whendifference
arises even
and in some cases, learningdifficulty
learningdifficulty,
wherethereis no difference.
has
The second approach,however,is also problematic.Markedness
been definedin termsof whethera grammaticalstructureis in some
sensefrequent,
unnatural,and deviant
natural,and basicor infrequent,
froma regularpattern(Richards,Platt,& Weber,1985). Thus,theuse of
as in He mademefollowhimcan be
withouttofollowing
an infinitive
make,
consideredmarkedbecause makeis one the fewverbsin Englishthat
takes this kind of complementand because thispatternoccurs only
The general idea is that we should teach the marked
infrequently.
featuresand leavethelearnersto learntheunmarkedformsnaturally
by
The problemis that,as thedefinition
themselves.
suggests,markedness
to apply
remainsa somewhatopaque concept,so thatitis oftendifficult
to teach.
withthe precisionneeded to determinewhichstructures
The selectionof grammatical
content,then,remainsveryproblematic.One solutionto thekindsof problemsI havementionedis to base
selectionon theknownerrorsproducedbylearners.In thisrespect,lists
ofcommonlearnererrorssuchas thoseavailableinTurtonand Heaton's
Errorsand Swan and Smith's (2001)
(1996) LongmanDictionary
ofCommon
and OtherProblemsare
LearnerEnglish: A TeachersGuide to Interference

helpful.
The problemsof selectionprobablyexplain whygrammaticalsyllabusesare so similarand havechangedso littleovertheyears;itis saferto
followwhathas been done before.Of course the selectionof whatto
teach will also depend on the learner'sstage of development.The
involveare discussedin
problemsthatthelearner'sstageofdevelopment
sections.
subsequent

CURRENT ISSUES IN THE TEACHING OF GRAMMAR

This content downloaded from 147.91.1.45 on Mon, 11 Aug 2014 07:33:07 UTC
All use subject to JSTOR Terms and Conditions

89

WHEN SHOULD WE TEACH GRAMMAR?


to the
Thereare twocompeting
answers
to thisquestion.
According
itisbesttoemphasise
intheearlystagesof
theteaching
ofgrammar
first,
tothesecond,itisbesttoemphasise
L2 acquisition.
meaningAccording
focused
instruction
tobeginwithandintroduce
later,
teaching
grammar
I will
whenlearnershavealreadybegunto formtheirinterlanguages.
forbothpositions.
considerthearguments
briefly
A keypremiseof behaviourist
theories
oflanguagelearningis that
"errorlikesin needsto be avoidedat all costs"(Brooks,1960).This
holdsthatoncelearners
haveformed
incorrect
habits,
theywill
premise
havedifficulty
themandreplacing
themwithcorrect
habits.
eradicating
habitsinthe
toensurethatlearners
Thus,itisnecessary
developcorrect
first
method
oftheaudiolingual
place.Thiswasone ofthekeypremises
ofbeginning
canbe advancedinfavour
(Lado,1964).Otherarguments
to teachgrammar
to a form-focused
early.The alternative
approach
as in task-based
language
emphasises
meaningand messagecreation,
believethatbeginning-level
(Skehan,1998),butmanyteachers
teaching
learners
activities
becausetheylack
cannotengagein meaning-centred
the necessary
of the L2 to perform
tasks.Thus,a formknowledge
focusedapproachis neededinitially
a basisofknowledge
to construct
thatlearners
can thenuse and extendin a meaning-focused
approach.
current
theoriesof L2 learning,
whichgivepriconnectionist
Finally,
exposureto the
macyto implicit
learningprocessesbasedon massive
a basisforteaching
tobeginners.
alsoprovide
target
grammar
language,
with
N. Ellis(2005) hassuggested
commences
thatlearning
necessarily
an explicit
oflinguistic
whicharethendeveloped
forms,
representation
He suggests
thatteaching
earlyis
through
grammar
implicit
learning.
thatfollows.
This
a basisforthereallearning
valuablebecauseitprovides
towhichgramseemstoechoLightbown's
(1991)metaphor,
according
learners
with"hooks"
marinstruction
facilitates
learningbyproviding
is thata
whichtheycan grabon to. The idea behindthismetaphor
of howgrammatical
features
workfacilitates
consciousunderstanding
thekindofprocessing
to linguistic
form)requiredfor
(e.g.,attention
truecompetence.
developing
The argumentagainstteachinggrammarearlyon derivesfrom
research
on immersion
(e.g.,Genesee,1987),whichshows
programmes
thatlearnersin suchprogrammes
are able to developtheproficiency
in the
neededforfluent
instruction
communication
without
anyformal
ofL2 Spanishdo notneedto be taughtthat
L2. Forexample,learners
follownounsin thislanguage;theyseemto be able to learn
adjectives
thisnaturalistically
fromexposureto communicative
input(Hughes,
clauses
learners
ofL2 English
canmaster
1979).Similarly,
simplerelative
90

TESOL QUARTERLY

This content downloaded from 147.91.1.45 on Mon, 11 Aug 2014 07:33:07 UTC
All use subject to JSTOR Terms and Conditions

(e.g., clauseswherethe relativepronounfunctionsas subjectand the


theverb). There is ample
clause is attachedto a noun phrasefollowing
evidenceto showthatlearnerscan and do learna good deal ofgrammar
withoutbeingtaughtit.This beingso, whybotherto teachwhatcan be
A second reasonfordelayinggrammarteachingto
learned naturally?
later stages of developmentis that early interlanguageis typically
agrammatical(Ellis,1984; Perdue & Klein,1993). That is, learnersrely
utterances
on a memory-based
systemoflexicalsequences,constructing
either by accessing ready-madechunks or by simplyconcatenating
lexical itemsinto simplestrings.Ellis (1984) givesexamplesof such
utterancesin the earlyspeech of threeclassroomlearners:
Me no (= I don'thaveanycrayons)
Me milkman
(= I wantto be themilkman)
Dinnertime
youout(= It is dinnertimeso youhaveto go out)

Such pidginisedutterancesrelyheavilyon contextand the use of


in simple,contextcommunicationstrategies.They are veryeffective
itis thislexicalisedknowledgethat
embeddedcommunication.
Arguably,
providesthe basisforthe subsequentdevelopmentof the grammatical
communication.This, then,is a
competenceneeded for context-free
strongargumentfordelayingthe teachingof grammaruntillearners
havedevelopeda basic communicative
ability.
In general,I have favouredthe second of thesepositions(see Ellis,
2002b). Giventhatmanyclassroomlearnerswillnotprogressbeyondthe
initialstagesof language learning,it seems to me that a task-based
approach thatcatersto the developmentof a proceduralisedlexical
structures
willensure
and simple,naturally
acquiredgrammatical
system
is tobe preferred
to an
communicative
a threshold
and,therefore,
ability
as
from
the
start
and
on
that
insists
that,
accuracy
grammatical
approach
a consequence,mayimpede the developmentof thiscommunicative
Task-based
languageteachingis possiblewithcompletebeginners
ability.
ifthefirst
tasksemphasiselistening(and perhapsreading)and allowfor
nonverbalresponses.However,it is possiblethatsuch an approachcan
be usefullycomplementedwithone thatdrawsbeginners'attentionto
features(e.g., past tense-^din English) that
some usefulgrammatical
instruction
miss.
This is the aim of input-processing
otherwise
theymight
(VanPatten,1996,2003), whichis discussedlater.

CURRENT ISSUES IN THE TEACHING OF GRAMMAR

This content downloaded from 147.91.1.45 on Mon, 11 Aug 2014 07:33:07 UTC
All use subject to JSTOR Terms and Conditions

91

SHOULD GRAMMARTEACHING
BE MASSED OR DISTRIBUTED?
Thisquestion
That
islogically
ofthepreceding
independent
question.
of whengrammar
we
need
to
is, irrespective
commences,
teaching
consider
whether
itshouldbe concentrated
intoa shortperiodoftime
hasaddressed
orspreadovera longerperiod.Remarkably
littleresearch
thisquestion.
on therelative
effects
The research
thathasbeenundertaken
reports
of massedand distributed
on
instruction
language
generallanguage
rather
thantheeffects
on grammar
Collins,
Halter,
proficiency
learning.
as
follows:
&
the
available
research
summarise
Lightbown,Spada(1999)
None of thelanguageprogramevaluationresearchhas foundan advantage
fordistributed
Althoughthe findingsthusfarlead to
languageinstruction.
thehypothesis
thatmoreconcentrated
exposuretoEnglishmaylead tobetter
studentoutcomes,theevidenceis not conclusive,(p. 659)

ESL
Collinsandcolleagues
thenreport
theirownstudy
ofthreeintensive
over
in
one
distributed
(the
taught
programmes Canada,
programme)
thefull10 monthsof one schoolyear,one (themassedprogramme)
concentrated
into5 months
buttaught
onlyto aboveaveragestudents,
andthethird(themassedplusprogramme)
into5 months,
concentrated
to
use
with
out
of
class
supplemented
opportunities Englishand taught
wasthatthemassed
tostudents
ofmixedability
levels.Themainfinding
thedistributed
and especially
themassed-plus
students
outperformed
of
students
on
most
of
the
measures
including
learning,
programme
somemeasuresof grammatical
mightin
ability,
althoughthisfinding
provided
partbe explainedbythefactthatthemassedprogrammes
moreoverallinstructional
time.
Collinsetal.'s study
research,
especially
pointstotheneedforfurther
instruction
distudiesthatcomparemassedand distributed
through
structures.
rectedat specificgrammatical
Ideallysucha studywould
in a particular
structure
spread
compareshortperiodsof instruction
into
overseveraldayswiththesameamountofinstruction
compressed
one or twolessons.4
Receivedwisdomis thata cyclicalapproachto
becauseitallowsfor
1974)istobe preferred
(Howatt,
grammar
teaching
withwhat
thatis compatible
thekindofgradualacquisition
ofgrammar
the
is knownaboutinterlanguage
However, resultsof
development.
4Giventhe
ofentire
in evaluations
extraneous
variables
thatarisein controlling
problems
studiesofmassedanddistributed
itmight
programmes,
provemucheasiertoconductrigorous
structures.
whenthesearefocusedon specific
learning
grammatical
92

TESOL QUARTERLY

This content downloaded from 147.91.1.45 on Mon, 11 Aug 2014 07:33:07 UTC
All use subject to JSTOR Terms and Conditions

Collinset al.'s studysuggest,at theveryleast,thatsucha positionneeds


to be investigatedempirically.
Here, then, is an issue about which
said
at
can
be
the moment.
definitive
nothing

SHOULD GRAMMARTEACHING
BE INTENSIVE OR EXTENSIVE?
refersto instruction
overa sustainedperiod
Intensive
teaching
grammar
of time(whichcould be a lessonor a seriesof lessonscoveringdaysor
structure
or,perhaps,a pairof
weeks)concerninga singlegrammatical
contrastedstructures(e.g., Englishpast continuousvs. past simple).
refersto instruction
Extensive
concerninga wholerange
teaching
grammar
withina shortperiod of time(e.g., a lesson) so thateach
of structures
structurereceivesonlyminimalattentionin any one lesson. It is the
difference
betweenshootinga pistolrepeatedlyat the same targetand
oftargets.Instruction
can be
a
firing shotgunto spraypelletsat a variety
ofwhetherit is massedor distributed.
or extensiveirrespective
intensive
distinctionrefersto how a whole grammar
The massed-distributed
distinctionrefersto
course is staged, while the intensive-extensive
whethereach singlelesson addressesa singleor multiplegrammatical
feature
(s).
viewedas entailingintensiveinstrucGrammarteachingis typically
tion.The present-practise-produce
(PPP) model of grammarteaching,
which underlies most discussionsof grammarteaching in teacher
focus
handbooks(see, e.g.,Hedge,2000;Ur,1996),assumesan intensive
discussions
such
acknowlstructures.
on specificgrammatical
Although
edge thatlearners'readinessto acquire a specificstructurelimitsthe
itis), theyalso assume
ofteaching(no matterhowintensive
effectiveness
learners
will eventually
for
thatwithsufficient
opportunities practice,
As
are
the structures
succeed in automatising
they taught. Ur says,"the
so
aim of grammarpracticeis to get studentsto learn the structures
on
them
their
will
be
able
to
that
correctly
produce
thoroughly they
own" (p. 83). Thus, the idea thatpractisemakesperfectis the primary
forthe intensiveapproach.Practise,
however,mustinvolve
justification
in
to practicethetargetstructure
bothdrillsand tasks(i.e.,opportunities
.
a communicative
context)
It is perhapsless easy to see how grammarteachingcan comprise
A teacherwouldprobablynotelectto presentand
extensiveinstruction.
withina singlelesson.
structures
a
whole
rangeof grammatical
practise
has alwayshad a place
ofa kind,however,
Extensivegrammarinstruction
in grammarteaching.Some 30 yearsago, whileteachingin a secondary
school in Zambia, I regularlygave lessons where I illustratedand
CURRENT ISSUES IN THE TEACHING OF GRAMMAR

This content downloaded from 147.91.1.45 on Mon, 11 Aug 2014 07:33:07 UTC
All use subject to JSTOR Terms and Conditions

93

explainedsome of the commonerrorsthatI had observedmystudents


in the contextof taskmakingin theirwrittencompositions.Similarly,
based teaching,some teachershave been observedto note the errors
thatlearnersmake and then to address themwhen the taskis over
(Basturkmen,Loewen, & Ellis, 2004). However,extensivegrammar
a learningactivity,
notjust as some kind of
teachingcan occur within
in the contextof both
Teachers
corrective
feedback
postscript.
provide
form-focused
and meaning-focused
lessons,and althoughfeedbackin
at thestructure
form-focused
lessonsmaybe directedprimarily
targeted
in
the
the
lessons
it
is
lesson,
by
likelyto be directedat
meaning-focused
feedback
whatever
errorslearnershappento make.Studiesofcorrective
&
&
Ranta,
Batsurkmen,
Loewen,
1997;Ellis,
2001) demon(e.g.,Lyster
forms
stratethatin communicative
lessonsa widevariety
ofgrammatical
are addressedincidentally
corrective
feedback.
through
There is littledoubt now that intensivegrammarlessons can be
effective.
Though earlierresearchshowedthatlearnersdo not always
learnwhattheyare taught,especiallywhenlearningis measuredin terms
of spontaneousproduction(e.g., Kadia, 1987), more recentresearch
(e.g., Spada & Lightbown,1999) indicatesthateven iflearnersare not
intensivegrammarteachingcan
readyto learn the targetedstructure,
them
the
help
progressthrough
sequence of stagesinvolvedin the
In
of
that
structure.
other
words,teachinga markedstructure
acquisition
can
learners
learn
associated,less markedstructures
intensively help
Intensive
evenifitdoes notresultin acquisitionofthemarkedstructure.
instructionalso helps learnersto use structuresthey have already
partiallyacquired more accurately(e.g., White,Spada, Lightbown,&
Ranta,1991).
and some empiricalevidencein
Thereare also theoretical
arguments
favourof an extensiveapproach. Cook (1989) has argued fromthe
ofuniversalgrammarthatlearnersrequireminimalevidence
perspective
to set a particularparameterforthe grammartheyare learning.Other
ofnegativeevidencethrough
researchers
haveemphasisedtheimportance
feedbackforgrammarlearningbyadults.Loewen (2002) has
corrective
feedbackare relatedto
shownthatevenverybriefepisodesofcorrective
correctnesson subsequenttests.In thatstudy,Loewen identifiedthe
in thecontextofcommunicaerrorsthatteachersaddressedincidentally
tivelanguageteachingand thendevelopedtailor-made
tests,whichhe
administered
to thelearnerswhomade thespecificerrorseitherone day
or two weekslater.These testsshowed that the learnerswere subseand correcttheirownerrors.
quentlyoftenable to identify
There are pros and cons forboth intensiveand extensivegrammar
instruction.
Some structures
maynot be masteredwithoutthe opportufor
nity
repeated practice.Harley (1989), for example, found that
anglophone learnersof L2 French failed to acquire the distinction
94

TESOL QUARTERLY

This content downloaded from 147.91.1.45 on Mon, 11 Aug 2014 07:33:07 UTC
All use subject to JSTOR Terms and Conditions

betweenthe preteriteand imparfait


pasttensesafterhoursof exposure
somecorrective
feedback)in an immersion
(and presumably
programme
butwereable to improvetheiraccuracyin usingthesetwotensesafter
is timeconsuming
intensiveinstruction.
However,intensiveinstruction
structures
were
the
(in Harley'sstudy
taughtover a 6-month
targeted
will
constrain
how
time
and
thus,
manystructurescan be
period),
on the otherhand, affords
addressed.Extensivegrammarinstruction,
structures.
to attendto large numbersof grammatical
the opportunity
the
structures
than
of
will
be
addressed
more
not,
Also,
many
likely
because
this
kind
a
of
time.
of
instrucover
Further,
period
repeatedly
tioninvolvesa responseto theerrorseach learnermakes,it is individutheskilledteacherreal-time
forthekind
alizedand affords
opportunities
ofcontextualanalysisthatCelce-Murcia(2002) recommendsas basisfor
grammarteaching.However,it is not possible to attend to those
thatlearnersdo notattemptto use (i.e., extensiveinstruction
structures
withavoidance). Also, of course, it does not
cannot deal effectively
the
mayrequirebefore
in-depthpractisethatsome structures
provide
can
be
fullyacquired.
they
grammarteachingneedsto be conceivedofin termsofboth
Arguably,
grammarteachingneedsto be reconceptualised
approaches.Therefore,
in teacherhandbooksto include the kind of extensivetreatmentof
feedback.
throughcorrective
grammarthatarisesnaturally

IS THERE ANY VALUE IN TEACHING


EXPLICIT GRAMMATICALKNOWLEDGE?
The distinctionbetweenexplicitand implicitknowledgewas menconsistsof thefactsthatspeakers
earlier.Explicit
tionedbriefly
knowledge
ofa languagehavelearned.These factsare oftennotclearlyunderstood
witheach other.Theyconcerndifferent
and maybe in conflict
aspectsof
is
languageincludinggrammar.Explicitknowledgeis held consciously,
and is typically
accessedthroughcontrolled
learnableand verbalisable,
in
difficulty
processingwhenlearnersexperiencesomekindoflinguistic
needs to be drawnbetweenexplicitknowlusingthe L2. A distinction
explanation.Analysed
edge as analysedknowledgeand as metalinguistic
featureworks,
entailsa consciousawarenessofhowa structural
knowledge
consistsof knowledgeof grammatical
while metalinguistic
explanation
metalanguageand the abilityto understandexplanationsof rules.In
and can
is procedural,is held unconsciously,
contrast,implicit
knowledge
onlybe verbalizedifit is made explicit.It is accessedrapidlyand easily
MostSLA
and thusis availableforuse in rapid,fluentcommunication.
a matterof
researchersagree thatcompetencein an L2 is primarily
implicitknowledge.
CURRENT ISSUES IN THE TEACHING OF GRAMMAR

This content downloaded from 147.91.1.45 on Mon, 11 Aug 2014 07:33:07 UTC
All use subject to JSTOR Terms and Conditions

95

Whetherthereis anyvaluein teachingexplicitknowledgeofgrammar


has been and remainstodayone of the most controversial
issues in
To
make
sense
of
the
different
teachinggrammar.
positionsrelatingto
the teachingof explicitknowledge,it is necessaryto considerthree
separatequestions:
1. Is explicitknowledgeof anyvalue in and of itself?
the developmentof
2. Is explicitknowledgeof value in facilitating
implicitknowledge?
3. Is explicitknowledgebesttaughtdeductively
or inductively?
I partlyaddressedthefirst
questionwhenI consideredwhatgrammar
to teach.I notedthatresearchers
disagreeoverlearners'abilityto learn
some
with
explicitknowledge,
(e.g., Krashen,1982) seeingthisas very
limitedand others(e.g., Green& Hecht, 1992) producingevidenceto
a separateissuerelated
suggestthatitis considerable.Thereis,however,
to thefirst
This
issue
concerns
the
extent
towhichlearnersare
question.
able to use theirexplicitknowledge(whatever
thatconsistsof) in actual
performance.
Again,one positionis thatthisabilityis limited.Krashen
that
can onlyuse explicitknowledgewhentheymonitor,
learners
argues
whichrequiresthattheyare focusedon form(as opposed to meaning)
and have sufficient
timeto access the knowledge.There is also some
evidence that teachingexplicitknowledgeby itself(i.e., withoutany
forpractising
thetargetfeature)is noteffective.
Studiesby
opportunities
VanPattenand Oikennon(1996) and Wong(2004) indicatethatexperimentalgroupsthatreceivedexplicitinformation
alone performedno
on
and
than
a controlgroup
tests
differently interpretation production
did. But otherpositionsare also possible.I have argued thatexplicit
knowledgeis used in the processof formulating
messagesas wellas in
in
and
that
learners
are
adroit
accessingtheirexplicit
monitoring
many
memoriesfor these purposes,especiallyif the rules are, to a degree,
automatised.However,this does require time.Yuan and Ellis (2003)
if
showed that learners'grammaticalaccuracyimprovedsignificantly
a
narrative
had
on-line
while
time
for
task,a
they
performing
planning
result most readilyexplained in terms of their accessing explicit
knowledge.
of whetherexplicitknowledgehas anyvalue in and of
Irrespective
it
thedevelopment
assist
itself, may
byfacilitating
languagedevelopment
of implicitknowledge.This issueis addressedbythe second of the two
hypothesis,
questions.It concernswhathas become knownas the interface
whichaddressesthe role explicitknowledgeplaysin L2 acquisition.
Three positionscan be identified.
position
Accordingto the noninterface
and
(Krashen,1981), explicit
implicitknowledgeare entirelydistinct
withtheresultthatexplicitknowledgecannotbe convertedintoimplicit
knowledge.This position is supportedby researchsuggestingthat
96

TESOL QUARTERLY

This content downloaded from 147.91.1.45 on Mon, 11 Aug 2014 07:33:07 UTC
All use subject to JSTOR Terms and Conditions

explicitand implicitmemoriesare neurologicallyseparate (Paradis,


1994). The interface
arguesthe exactopposite.Drawingon skillposition
learningtheory,DeKeyser(1998) argues that explicitknowledgebeforplentiful
comesimplicitknowledgeiflearnershave the opportunity
The
weak
claims
communicative
(Ellis,
1993)
position
interface
practice.
that explicitknowledgecan convertinto implicitknowledgeif the
learneris readyto acquirethe targetedfeatureand thatthisconversion
occursbypriminga numberofkeyacquisitionalprocesses,in particular
thegap (Schmidt,1990). That is,explicitknowledge
and noticing
noticing
makesit morelikelythatlearnerswillattend
structure
of a grammatical
to the structurein the inputand carryout the cognitivecomparison
betweenwhattheyobservein the inputand theirown output.These
positionscontinueto be arguedat a theoreticallevel.Althoughthereis
is effective
in promotingL2
plentifulevidencethatexplicitinstruction
&
no
Noms
Ortega,2000)
publishedstudyhas directly
learning(e.g.,
intoimplicitknowltestedwhetherexplicitknowledgeconvertsdirectly
One
reason
forthe lack of
its
facilitates
or
development.
edge
simply
of
researchis the problem of measurement,that is, the difficulty
learners
when
of
which
theypertype knowledge
employ
ascertaining
forma languagetaskor test.
The threepositionssupportverydifferent
approachesto language
to
a
zerogrammar
leads
The
noninterface
position
approach,
teaching.
and
thatis,itprioritizes
meaning-centred
approachessuchas immersion
PPP
the
idea
that
The
interface
task-based
positionsupports
teaching.
a grammaticalstructureshould be firstpresentedexplicitlyand then
The weakinterface
positionalso
practiseduntilitis fullyproceduralised.
lendssupportto techniquesthatinducelearnersto attendto grammatical features.It has been used to providea basisforconsciousness-raising
tasksthatrequirelearnersto derivetheirown explicitgrammarrules
fromdatatheyare providedwith(Ellis,1993;Fotos,1994). It is likelythat
all threeapproacheswill continueto attractsupporters,drawingon
differenttheoriesof L2 acquisitionand citing research that lends
indirectsupportto the preferredapproach. It is unlikelythat this
willbe resolvedthroughresearchin the nearfuture.
controversy
The thirdquestionassumesthereis value in explicitknowledgeand
addresseshow best to teach it. In deductiveteaching,a grammatical
and thenpractisedin one wayor another;
is presentedinitially
structure
thisis thefirstP in the present-practise-produce
sequence. In inductive
first
to
learners
are
exposed exemplarsof the grammatical
teaching,
and are asked to arriveat a metalinguistic
structure
generalisationon
oftherule.
theirown;theremayor maynotbe a finalexplicitstatement
A numberofstudies(see Erlam,2003,fora review)haveexaminedthe
ofthesetwoapproachesto teachingexplicitknowlrelativeeffectiveness
been mixed.Forexample,Herronand Tomosello
The
results
have
edge.
CURRENT ISSUES IN THE TEACHING OF GRAMMAR

This content downloaded from 147.91.1.45 on Mon, 11 Aug 2014 07:33:07 UTC
All use subject to JSTOR Terms and Conditions

97

Robinson
instruction,
(1992) founda clear advantageforinductive
while
Rosa
was
more
effective,
(1996)foundthata deductive
approach
and O'Neill (1999) foundno significant
in effectiveness.
difference
Erlam's(2003)ownstudy
forthegroup
revealed
a significant
advantage
deductive
instruction.
to
be learned
the
main
lesson
receiving
Perhaps
fromtheresearch
to dateis theneedfora differentiated
approachto
bothresearching
It
andteaching
is
thatmany
knowledge. likely
explicit
variables
affect
mostfrom,including
whichapproachlearnersbenefit
the specificstructure
and the
thatis the targetof the instruction
learners'aptitudeforgrammatical
rules
maybestbe
analysis.
Simple
be taught
while
more
rules
best
taughtdeductively,
may
complex
in
Learners
are
skilled
inductively.
grammatical
analysis likelyto fare
better
withan inductive
than
those
lessskilled.
approach
IS THERE A BEST WAYTO TEACH
GRAMMARFOR IMPLICIT KNOWLEDGE?
To answerthisquestionit is necessary
to identify
theinstructional
I haveattempted
thisin a numberof
optionsforteachinggrammar.
publications
justtwo:the
(e.g.,Ellis1997,1998,2002b).51willconsider
difference
instruction
and
betweeninput-based
and production-based
between
different
ofcorrective
feedback.
types
The case forthe input-based
optionis based on a computational
modelofL2 acquisition,
towhichacquisition
takesplaceas a
according
and processing
productof learnerscomprehending
input.Such apattention
to
whendirected
atgrammar,
seektodrawlearners'
proaches,
for
thetargeted
structure
(s) in one or moreways:simply
bycontriving
numerousexemplars
of the structure
(s) to be presentin the input
thetargetstructure
materials,
(s) in someway(e.g.,by
byhighlighting
tasks
, orbymeansofinterpretation
texts)
usingboldoritalicsinwritten
to form-meaning
learners'attention
(Ellis,1995) directedat drawing
oftheinputVanPatten
(1996,2003)hasdevelopeda version
mappings.
at
Thisis directed
basedoptionthathe callsinput
instruction.
processing
thatare a
strategies
helpinglearnersto overcomethe default
processing
featureof interlanguages
(e.g., assumingthatthe firstnoun in a
sentence
is always
theagent). A casefortheoutput-based
optioncanbe
foundin bothskill-building
theory(see previousdiscussion)or in a
arises
towhichlearning
sociocultural
ofL2 learning,
according
theory
51
distinguishbetween psycholinguisticand methodological options (cf. Ellis, 1998).
optionsare
Psycholinguistic
optionsare related to some model of L2 acquisition.Methodological
evidentin instructionalmaterialsforteachinggrammar.Here I consideronlypsycholinguistic
options.
98

TESOL QUARTERLY

This content downloaded from 147.91.1.45 on Mon, 11 Aug 2014 07:33:07 UTC
All use subject to JSTOR Terms and Conditions

out of social interactionwhichscaffoldslearners'attemptsto produce


new grammaticalstructures(Ohta, 2001). A numberof studieshave
ofinput-based
and production-based
comparedtherelativeeffectiveness
withmixed results,resultingin ongoingdebate about the
instruction,
meritsofthesetwooptions(VanPatten,
relative
2002;DeKeyser,
Salaberry,
Robinson,& Harrington,2002). It may be that,in classrooms,this
meaninglessbecause,in practise,bothoptions
comparisonis ultimately
and production.For example,it is
are likelyto involveinput-processing
in
that
an
conceivable
input-basedapproach,individualstudents
quite
the
while in a production-based
structure,
target
silentlyproduce
apan
utterance
one
studentservesas input for
produced by
proach,
another.It is, therefore,not surprisingthatboth options have been
shownto resultin acquisition.6
There is a rich descriptiveliteratureon correctivefeedback (i.e.,
fewstudieshave
teacherresponsesto learnererrors)but remarkably
the
relative
effects
of
different
of
feedback
on acquisiinvestigated
types
tion.Keyoptionsare (a) whetherthefeedbackis implicitor explicitand
occurs
(b) whetherthefeedbackis inputor outputbased.Implicit
feedback
forceoftheresponseto learnererroris masked,for
whenthecorrective
a deviantutterancecorrectingit
whichreformulates
example,a recast,
whilekeepingthesame meaning:
NNS: Whyhe isveryunhappy?

NS:

Whyis he veryunhappy?

NNS: Yeah whyis veryunhappy?(Philp,2003)


Or, as in this contrivedexample, a request for clarification:
NNS: Whyhe is veryunhappy?
NS:

Sorry?

NNS: Whyis he veryunhappy?


takesa numberof forms,such as directcorrectionor
Explicit
feedback
metalinguistic
explanation.There is some evidencethatexplicitfeedin both elicitingthe learner'simmediatecorrect
back is moreeffective
and in elicitingsubsequentcorrectuse,forexample,
use ofthestructure
in a post-test
(Carroll& Swain1993;Lyster2004). Butsomeevidenceand
6There is also
of these two (and
controversy
regardinghow to measure the effectiveness
of
other) instructionaloptions. Norrisand Ortega (2000) have shown that the effectiveness
instructionvaries depending on whetherit is measured using metalinguisticjudgements,
selected response,constrainedconstructedresponse,or freeconstructedresponse.Most SLA
researchers(and teachers,too, perhaps) would consider the last of these the most valid
measure.Ellis (2002a) revieweda numberofstudiesthatexaminedtheeffectsofdifferent
kinds
of instructionon learners'freeconstructedresponses,reportingthatinstructioncan have an
effecton thistypeof language use.
CURRENT ISSUES IN THE TEACHING OF GRAMMAR

This content downloaded from 147.91.1.45 on Mon, 11 Aug 2014 07:33:07 UTC
All use subject to JSTOR Terms and Conditions

99

some strongtheoreticalreasonsexistto supportimplicitfeedback(see


Long 1996,in press).Indeed, thistypeof feedbackis morecompatible
withthefocus-on-form
approachdiscussedearlierbecauseitensuresthat
learnersare more likelyto stayfocused on meaning. However,as
when
Muranoi(2000) notes,implicitfeedbackis probablymoreeffective
at a preselectedformthan when it occurs
it is targetedintensively
in incidentalfocuson form.In thelatter,
explicitattentionto
extensively
formmaybe moreeffective.
modelsthe correctformforthe learner(e.g., by
Input-based
feedback
elicitsproductionof thecorrect
meansofa recast).Output-based
feedback
formfromthe learner (e.g., by means of a requestforclarification).
ofthesetwo
abouttherelativeeffectiveness
Again,thereis disagreement
feedbackoptionsand no clear evidencefor choosingbetweenthem.
feedbackis more
Some descriptive
studieshaveshownthatoutput-based
erroneousuttertheir
own
initial
to
lead
to
learners
likely
correcting
to as uptake.
ances in whatis referred
However,uptakeis notthesameas
acquisition.
In short,although considerableprogresshas been made toward
identifyingthose instructionaloptions that are likely to be of
as yet,fewconclusionscan be drawnabout
significance,
psycholinguistic
foracquisition.It is possibleto pointto
whichones are themosteffective
studiesand theoreticalargumentsthatsuggestthateach of the major
optionsdiscussedcan contributeto acquisition.

SHOULD GRAMMARBE TAUGHT


IN SEPARATE LESSONS OR INTEGRATED INTO
COMMUNICATIVE ACTIVITIES?
In Ellis (2001) I considered three broad typesof form-focused
as shownin Table 1. "Focuson forms"refersto instruction
instruction,
a
approach,wherethe students'primary
involving structure-of-the-day
focusis on form(i.e., accuracy)and wherethe activitiesare directed
at a single grammaticalstructure.This approach, then,
intensively
involvesteachinggrammarin a seriesof separatelessons.Focusonform
entailsa focuson meaningwithattentionto formarisingout of the
wherea focusedtaskis
Thisfocuscan be planned,
communicative
activity.
a
for
to
elicit
occasions
using predeterminedgrammatical
required
structure,
as,forexample,in Samuda (2001). In thisapproach,attention
to the predeterminedgrammaticalstructureswill also be intensive.
whereattentionto formin
focuson formcan be incidental
Alternatively,
butrather
is
not
thecontextofa communicative
predetermined
activity
as
the
needs
occursin accordancewiththeparticipants'
activity
linguistic
100

TESOL QUARTERLY

This content downloaded from 147.91.1.45 on Mon, 11 Aug 2014 07:33:07 UTC
All use subject to JSTOR Terms and Conditions

TABLE 1
Types of Form-FocusedInstruction

Type
1. Focuson forms
2. Plannedfocuson form
3. Incidental
focuson form

Focus
Primary
Form
Meaning
Meaning

Distribution
Intensive
Intensive
Extensive

Note.
Thistableis-adaptedfromEllis(2001,p. 17).

proceeds.In thisapproach,it is likelythatattentionwillbe givento a


structures
widevariety
ofgrammatical
duringanyone taskand thuswill
be extensive.Focus on formimpliesno separategrammarlessonsbut
rathergrammarteachingintegratedinto a curriculumconsistingof
tasks.
communicative
There is considerabletheoreticaldisagreementregardingwhichof
is mosteffective
in developingimplicitknowlthesetypesof instruction
and
havearguedstrongly
that
1991)
(2001)
edge. Long (1988,
Doughty
focuson formis bestequipped to promoteinterlanguage
development
because the acquisitionof implicitknowledgeoccurs as a resultof
format thesame timetheyare engaged
learnersattendingto linguistic
withunderstandingand producingmeaningfulmessages.Other researchers,however,have argued that a focus-on-forms
approach is
for
has
effective.
that
DeKeyser(1998),
example,
argued
grammatical
structures
are learnedgraduallythroughthe automatisation
of explicit
this
can
be
of
and
that
achieved
means
a
focus-on-forms
by
knowledge
approach.This approach acknowledgesthe value of teachingexplicit
it by means of activities
knowledgeand subsequentlyproceduralising
behaviours
involve
and
that
tasks)
(i.e.,
(drills
practise
meaning)rather
It is worthnoting,however,one pointof agreementin
than structures.
needs to ensurethatlearnersare
thesedifferent
positions:Instruction
able to connect grammaticalformsto the meaningstheyrealise in
So far,thedebatehas addressedthedifference
communication.
between
focuson formand focuson forms.Therehas been littlediscussionofthe
relativemeritsof planned and incidentalfocuson form.In effect,this
discussionwouldinvolvea considerationofwhetherinstruction
should
a questionwe havealreadyconsidered.
be intensiveor extensive,

CONCLUSION
Grammarhas held and continuesto hold a centralplace in language
teaching.The zero grammarapproachwas flirtedwithbut neverreally
CURRENTISSUES IN THE TEACHINGOF GRAMMAR

This content downloaded from 147.91.1.45 on Mon, 11 Aug 2014 07:33:07 UTC
All use subject to JSTOR Terms and Conditions

101

tookhold,as is evidentin boththecurrenttextbookmaterialsemanating


frompublishinghouses (e.g., Whitney& White,2001) and in current
that
theoriesofL2 acquisition.There is ample evidenceto demonstrate
works.
teachinggrammar
thata traditional
approachto
Althoughthereis nowa clearconviction
and
drill-like
based
on
practice
teachinggrammar
explicitexplanations
is unlikelyto resultin theacquisitionof theimplicitknowledgeneeded
forfluentand accuratecommunication,
therecontinuesto be disagreementregardingwhatshouldreplacethis.It seemsappropriate,then,to
finishwitha statementof my own beliefsabout grammarteaching,
thatmanyof themremaincontroversial:
acknowledging
1. The grammartaughtshould be one thatemphasisesnotjust form
structures.
but also the meaningsand uses of different
grammatical
structures
2. Teachersshouldendeavourto focuson thosegrammatical
thatare knownto be problematic
to learnersratherthantryto teach
thewholeof grammar.
3. Grammaris besttaughtto learnerswho havealreadyacquiredsome
level) ratherthanto
abilityto use the language (i.e., intermediate
can
be taughtthrough
complete beginners.However,grammar
correctivefeedbackas soon as learnersbegin to use the language
productively.
4. A focus-on-forms
approachis validas long as itincludesan opportutasks.
for
learners
to
nity
practisebehaviourin communicative
witha massed
5. Considerationshould be given to experimenting
to
ratherthandistributed
approach teachinggrammar.
instruc6. Use shouldbe made of bothinput-basedand output-based
tionaloptions.
7. A case existsforteachingexplicitgrammatical
knowledgeas a means
of
of assistingsubsequentacquisition implicitknowledge.Teaching
explicitknowledgecan be incorporatedintoboth a focus-on-forms
and a focus-on-form
approach. In the case of a focus-on-forms
sometimesdeductive
approach,a differentiated
approachinvolving
work
best.
and sometimesinductiveinstruction
may
8. An incidentalfocus-on-form
approachis of specialvalue because it
affordsan opportunityfor extensivetreatmentof grammatical
afforded
bya focusproblems(in contrastto theintensivetreatment
on-forms
approach).
9. Correctivefeedbackis importantfor learninggrammar.It is best
conductedusinga mixtureof implicitand explicitfeedbacktypes
thatare bothinputbased and outputbased.

102

TESOL QUARTERLY

This content downloaded from 147.91.1.45 on Mon, 11 Aug 2014 07:33:07 UTC
All use subject to JSTOR Terms and Conditions

should take
10. In accordancewiththesebeliefs,grammarinstruction
the formof separategrammarlessons(a focus-on-forms
approach)
intocommunicative
activities
and shouldalso be integrated
(a focuson-form
approach).
are open to challenge.They
not
all) of these statements
(if
Many
of whatthe researchto date has
constitutea personalinterpretation
shown.It mayalso seem thatI am hedgingmybetsbyencompassinga
thatanything
widenumberof optionsand thatI am suggesting
goes. It
is certainlytruethatI do not believe (and do not thinkthe research
demonstrates)that thereis just one preferredapproach to teaching
grammar.The acquisitionof the grammaticalsystemof an L2 is a
can be assistedbestbya variety
of
complexprocessand almostcertainly
is
is
to
what
are
But
what
important
recognize
options
approaches.
available,whatthetheoreticalrationalesfortheseoptionsare,and what
the problemsare withthese rationales.This is the startingpoint for
developinga personaltheoryof grammarteaching.
existspointsto the need formore
The factthatso muchcontroversy
research.One ofthegreatestneedsis forresearchthataddressestowhat
resultsin implicitknowlextentand in whatwaysgrammarinstruction
of
methods
this
would
measuringacquisitionthat
require
edge. Ideally,
to
use
the
into
learners'
grammaticalstructures
theyhave
ability
tap
oral
Studies
been taughtin communication
communication).
(especially
thatemploysuchmethodsare stillfewand farbetween.Anotherneed is
the effectsof instruction
over
forlongitudinalstudiesthatinvestigate
time.Althoughmostrecentlypublishedstudiesinclude delayedposttreatments
of a relatively
tests,theytypically
incorporateinstructional
shortduration.Longitudinalstudiesthatemployqualitativeas well as
methodswillhelp to shownotjust ifthereis a delayedeffect
quantitative
but also itsaccumulativeeffect.The effects
of corrective
forinstruction
when
feedback,forexample,are mostlikelyto becomeevidentgradually
learnersare repeatedlyexposed to feedbackon the same grammatical
Furtherresearch,even if it does not succeed in providing
structures.
clear-cutanswersto the questionsraisedin thisarticle,willdeepen our
of the issuesinvolvedand affordbetterdefinedproviunderstanding
sionalspecifications
(Stenhouse,1975), whichteacherscan experiment
within theirownclassrooms.
THE AUTHOR
in the Departmentof AppliedLanguageStudiesand
Rod Ellis is a professor
ofAuckland,
in
NewZealand.He has publishedwidely
at theUniversity
Linguistics
the fieldof SLA. His latestbooks are Analyzing
LanguageLearningand Planningand
in a SecondLanguage.
TaskPerformance

CURRENT ISSUES IN THE TEACHING OF GRAMMAR

This content downloaded from 147.91.1.45 on Mon, 11 Aug 2014 07:33:07 UTC
All use subject to JSTOR Terms and Conditions

103

REFERENCES
Basturkmen,H., Loewen, S., & Ellis, R. (2004). Teachers' stated beliefs about
incidentalfocus on formand theirclassroompractices.AppliedLinguistics,
25,
243-272.
and skillsin English.Arusha,Tanzania: Longman.
Bright,J. (1965). Patterns
Brooks, N. (1960). Language and languagelearning.New York: Harcourt Brace &
World.
Carroll,S., & Swain,M. (1993). Explicitand implicitnegativefeedback:An empirical
studyof the learning of linguisticgeneralizations.Studiesin SecondLanguage
15, 357-386.
Acquisition,
Celce-Murcia,M. (1991). Grammar pedagogy in second and foreign language
25, 459-480.
teaching.TESOL Quarterly,
Celce-Murcia,M. (2002). Whyitmakessense to teach grammarthroughcontextand
on grammar
throughdiscourse. In E. Hinkel 8c S. Fotos (Eds.), Newperspectives
teachingin secondlanguage classrooms(pp. 119-134). Mahwah, NJ: Lawrence
Erlbaum.
book(2nd d.). Boston:
D. (1999). Thegrammar
Celce-Murcia,M., & Larsen-Freeman,
Heinle 8cHeinle.
Collins,L., Halter,R., Lightbown,P., & Spada, N. (1999). Time and distributionof
timein L2 instruction.TESOL Quarterly,
33, 655-680.
17, 169-182.
Cook, V. (1989). Universalgrammartheoryand the classroom.System,
Reviewof
Corder, S. P. (1967). The significanceof learners' errors.International
5, 161-169.
AppliedLinguistics,
DeKeyser,R. (1998). Beyondfocuson form:Cognitiveperspectiveson learningand
practicingsecond language grammar.In C. Doughty8c]. Williams(Eds.), Focuson
secondlanguageacquisition(pp. 42-63). Cambridge,England:
formin classroom
Press.
CambridgeUniversity
DeKeyser, R., Salaberry,R., Robinson, P., 8c Harrington,M. (2002). What gets
A commentaryon Bill VanPatten's"Processprocessedin processinginstruction?
An update." LanguageLearning,52, 805-824.
ing instruction:
Doughty,C. (2001). Cognitiveunderpinningsof focus on form.In P. Robinson
(Ed.), Cognitionand secondlanguageinstruction
(pp. 206-257). Cambridge,EnPress.
gland: CambridgeUniversity
Ellis,N. C. (2005). At theinterface:How explicitknowledgeaffectsimplicitlanguage
27, 305-352.
learning.Studiesin SecondLanguageAcquisition,
secondlanguagedevelopment.
Oxford,England: Pergamon.
Ellis,R. (1984). Classroom
secondlanguageacquisition.
Oxford,England: Oxford
Ellis, R. (1985). Understanding
Press.
University
Ellis, R. (1993). Second language acquisition and the structuralsyllabus.TESOL
27, 91-113.
Quarterly,
tasksforgrammarteaching.TESOL Quarterly,
29, 87Ellis,R. (1995). Interpretation
105.
Ellis, R. (1997). SLA researchand language teaching.Oxford, England: Oxford
Press.
University
Ellis, R. (1998). Teaching and research: Options in grammar teaching. TESOL
32, 39-60.
Quarterly,
instruction.In R. Ellis (Ed.), Form-focused
form-focused
Ellis,R. (2001). Investigating
instruction
and secondlanguagelearning(pp. 1-46). Maiden, MA: Blackwell.
instructionaffectthe acquisition of implicit
Ellis, R. (2002a). Does form-focused
24,
knowledge?A reviewof the research.Studiesin SecondLanguageAcquisition,
223-236.
Ellis,R. (2002b). The place of grammarinstructionin the second/foreignlanguage
104

TESOL QUARTERLY

This content downloaded from 147.91.1.45 on Mon, 11 Aug 2014 07:33:07 UTC
All use subject to JSTOR Terms and Conditions

ongrammar
in
curriculum.In E. Hinkel 8cS. Fotos (Eds.), Newperspectives
teaching
secondlanguageclassrooms
(pp. 17-34). Mahwah,NJ:LawrenceErlbaum.
Ellis,R., Basturkmen,H., & Loewen, H. (2001). Learner uptake in communicative
ESL lessons.LanguageLearning,51, 281-318.
Erlam, R. (2003). The effectsof deductive and inductive instructionon the
acquisitionof directobject pronounsin Frenchas a second language. TheModern
LanguageJournal,87, 242-260.
Fotos,S. (1994). Integratinggrammarinstructionand communicativelanguage use
tasks.TESOL Quarterly,
28, 323-351.
throughgrammarconsciousness-raising
twolanguages:Studiesofimmersion
and bilingual
Genesee, F. (1987). Learningthrough
education.
Rowley,MA: NewburyHouse.
Green, P., 8c Hecht, K. (1992). Implicitand explicitgrammar:An empiricalstudy.
13, 168-184.
AppliedLinguistics,
Harley,B. (1989). Functionalgrammarin French immersion:A classroomexperiment.AppliedLinguistics,
19, 331-359.
Oxford,England:
Hedge, T. (2000). Teachingand learningin thelanguageclassroom.
Press.
OxfordUniversity
Herron, C, & Tomosello, M. (1992). Acquiringgrammaticalstructuresby guided
induction.FrenchReview,65, 708-718.
Howatt,T. (1974). The backgroundto course design. In J. Allen 8c S. P. Corder
(Eds.), The Edinburghcoursein appliedlinguistics:Vol. 3. Techniquesin applied
Press.
linguistics
(pp. 1-23). Oxford,England: OxfordUniversity
on theutility
ot metalinguistic
Hu, G. (2002). Psychologicalconstraints
knowledgein
second language production.Studiesin SecondLanguageAcquisition,
24, 347-386.
Hughes,A. (1979). Aspectsof a Spanish adult's acquisitionof English.Interlanguage
StudiesBulletin,4, 49-65.
Kadia, K. (1987). The effectof formalinstructionon monitoredand spontaneous
naturalisticinterlanguageperformance.TESOL Quarterly,
22, 509-515.
Krashen,S. (1981). Secondlanguageacquisitionand secondlanguagelearning.Oxford,
England: Pergamon.
Krashen, S. (1982). Principlesand practicein secondlanguageacquisition.Oxford,
England: Pergamon.
Krashen, S. (1993). The effectof grammar teaching: Still peripheral. TESOL
27, 717-725.
Quarterly,
A scientific
Lado, R. (1964). Languageteaching:
approach.New York:McGrawHill.
Lado, R. (1970). Lado Englishseries:Book1. Montreal,Canada: Centre Educatifet
Culturel.
Lightbown,P. (1991). What have we here? Some observationson the effectof
instructionon L2 learning.In R. Phillipson,E. Kellerman,L. Selinker,M. Sharwood Smith,8cM. Swain (Eds.), Foreign/
secondlanguagepedagogy
research
(pp. 197212). Clevedon,England: MultilingualMatters.
of incidentalfocuson formin
Loewen, S. (2002) . The occurrenceand effectiveness
of
meaning-focusedESL lessons. Unpublished doctoral dissertation,University
Auckland,New Zealand.
A review
Long, M. H. (1983). Does second language instructionmake a difference?
of the research.TESOL Quarterly,
17, 359-382.
Long, M. H. (1988). Instructedinterlanguagedevelopment.In L. Beebe (Ed.), Issues
in secondlanguageacquisition:Multipleperspectives
(pp. 115-141). Rowley,MA:
NewburyHouse.
Long, M. H. (1991). Focus on form: A design feature in language teaching
methodology.In K. de Bot, R. Ginsberg,8c C. Kramsch (Eds.), Foreignlanguage
research
in cross-cultural
perspective
(pp. 39-52). Amsterdam:
John Benjamins.
CURRENTISSUES IN THE TEACHINGOF GRAMMAR

This content downloaded from 147.91.1.45 on Mon, 11 Aug 2014 07:33:07 UTC
All use subject to JSTOR Terms and Conditions

105

Long, M. H. (1996). The role of the linguisticenvironmentin second language


acquisition. In W. Ritchie & T. Bhatia (Eds.), Handbookof secondlanguage
acquisition(pp. 413-468). San Diego, CA: Academic Press.
Long, M. H. (in press). Recastsin SLA: The storyso far.In M. H. Long (Ed.) , Problems
in SLA. Mahwah,NT:LawrenceErlbaum.
Lyster,R. (2004). Differentialeffectsof prompts and recasts in form-focused
instruction.Studiesin SecondLanguageAcquisition,
26, 399-432.
Lyster,R., 8cRanta, L. (1997). Correctivefeedbackand learneruptake:Negotiation
of formin communicativeclassrooms.Studiesin SecondLanguageAcquisition,
19,
37-66.
Mackey,W. (1976). Languageteaching
analysis.London: Longman.
Macrory,G., 8cStone,V. (2000). Pupil progressin theacquisitionof theperfecttense
in French: The relationshipbetween knowledge and use. Language Teaching
Research,
4, 55-82.
Muranoi,H. (2000). Focus on formthroughinteractionenhancement:Integrating
formal instructioninto a communicativetask in EFL classrooms. Language
Learning,50, 617-673.
bookforintermediate
and practice
reference
Murphy,R. (1994). Englishin use:A self-study
Press.
students
(2nd d.). Cambridge,England: CambridgeUniversity
A researchsynthesis
of L2 instruction:
Norris,J.,8c Ortega,L. (2000). Effectiveness
and quantitativemeta-analysis.
LanguageLearning,50, 417-528.
Ohta, A. S. (2001). Secondlanguageacquisitionprocessesin theclassroom:
Learning
Japanese.Mahwah,NJ:LawrenceErlbaum.
Paradis,M. (1994). Neurolinguisticaspectsof implicitand explicitmemory:Implications for bilingualismand SLA. In N. Ellis (Ed)., Implicitand explicitlearningof
languages(pp. 393-419). San Diego, CA: Academic Press.
Perdue,C, 8cKlein,W. (1993). Concludingremarks.In C. Perdue (Ed.) , Adultsecond
Vol. 2. The results(pp. 253-272).
languageacquisition:Cross-linguistic
perspectives:
Press.
Cambridge,England: CambridgeUniversity
Philp,J. (2003). Constraintson "noticingthe gap": Non-nativespeakers'noticingof
recastsin NS-NNS interaction.Studiesin SecondLanguageAcquisition,
25, 99-126.
Pica, T. (1983). Adult acquisitionof Englishas a second language under different
conditionsof exposure. LanguageLearning,33, 465-497.
Quirk,R., Greenbaum,S., Leech, G., & Svartvik,
ofcontemporary
J. (1972). A grammar
English.London: Longman.
Richards,J., Platt,J., 8c Weber, H. (1985). Longmandictionary
ofappliedlinguistics.
London: Longman.
Robinson,P. (1996). Learningsimpleand complex rules under implicit,incidental
rule-searchconditions, and instructedconditions. Studiesin SecondLanguage
18, 27-67.
Acquisition,
Rosa, R., 8cO'Neill, M. (1999). Explicitness,intakeand theissue ofawareness.Studies
in SecondLanguageAcquisition,
21, 511-556.
Samuda, V. (2001). Gettingrelationshipbetween formand meaning during task
performance:The role of the teacher.In M. Bygate,P. Skehan,& M. Swain (Eds.),
Task-basedlearning:Language teaching,learningand assessment
(pp. 119-140).
Harlow,England: Pearson.
Schmidt,R. (1990). The role of consciousnessin second language learning.Applied
11, 129-58.
Linguistics,
Oxford,England: Oxford
Skehan,P. (1998). A cognitive
approachtolanguagelearning.
Press.
University
Spada, N., & Lightbown,P. (1999). Instruction,firstlanguage influence, and

106

TESOL QUARTERLY

This content downloaded from 147.91.1.45 on Mon, 11 Aug 2014 07:33:07 UTC
All use subject to JSTOR Terms and Conditions

developmentalreadiness in second language acquisition. The ModernLanguage


Journal,83, 1-22.
tocurriculum
research
and development.
London:
Stenhouse,L. (1975). An introduction

Heinemann.

and other
Swan, M., & Smith, B. (2001). LearnerEnglish: A Teacher'sguide to interference

Press.
(2nd d.). Cambridge,England: CambridgeUniversity
problems
Trahey,M., 8c White,L. (1993). Positiveevidence and preemptionin the second
15, 181-204.
language classroom.Studiesin SecondLanguageAcquisition,
errors.
London:
Turton,J.,8cHeaton, N. (Eds.). (1996). Longmandictionary
ofcommon
Longman.
Ur, P. (1996). A coursein language teaching.Cambridge, England: Cambridge
Press.
University
and grammarinstruction
in secondlanguage
VanPatten,B. (1996). Input processing
Norwood,NJ:Ablex.
acquisition.
An update. LanguageLearning,
VanPatten,B. (2002). Processinginstruction:
52, 755804.
VanPatten, B. (2003). Frominput to output:A teacher'sguide to secondlanguage
New York:McGraw-Hill.
acquisition.
VanPatten,B., & Oikennon,S. (1996). Explanationvs.structuredinputin processing
instruction.Studiesin SecondLanguageAcquisition,
18, 495-510.
connections
in second
VanPatten,B., Williams,J., 8c Rott, S. (2004). Form-meaning
Mahwah,NJ.:LawrenceErlbaum.
languageacquisition.
course.
Walter,C, & Swan,M. (1990). ThenewEnglishCambridge
Cambridge,England:
Press.
CambridgeUniversity
White,L., Spada, N., Lightbown,P., & Ranta, L. (1991). Input enhancementand
12, 416-432.
questionformation.AppliedLinguistics,
book1. Oxford,England: Oxford
Whitney,N., & White,L. (2001). Teamup: Students
Press.
University
in French:The rolesof explicitinformation
Wong,W. (2004). Processinginstruction
and structuredinput.In B. VanPatten(Ed.), Frominputtooutput:A teacher's
guide
tosecondlanguageacquisition(pp. 187-205). New York:McGraw-Hill.
Yuan, F., & Ellis,R. (2003). The effectsof pre-taskplanningand on-lineplanningon
and accuracyin L2 oral production.AppliedLinguistics,
24, 1fluency,complexity,
27.

CURRENT ISSUES IN THE TEACHING OF GRAMMAR

This content downloaded from 147.91.1.45 on Mon, 11 Aug 2014 07:33:07 UTC
All use subject to JSTOR Terms and Conditions

107

Vous aimerez peut-être aussi