• Author / Uploaded
• Yumi MeeksGiz Sawabe

• Categories
• Adquisicion de una segunda lengua
• Segundo lenguaje
• Aquisición de idioma
• Inglés como segunda lengua o lengua extranjera
• Conocimiento

Citation preview

Teachers of English to Speakers of Other Languages, Inc. (TESOL)

The Structural Syllabus and Second Language Acquisition Author(s): Rod Ellis Source: TESOL Quarterly, Vol. 27, No. 1 (Spring, 1993), pp. 91-113 Published by: Teachers of English to Speakers of Other Languages, Inc. (TESOL) Stable URL: http://www.jstor.org/stable/3586953 . Accessed: 28/09/2013 12:16 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 [email protected].

.

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 81.102.155.33 on Sat, 28 Sep 2013 12:16:33 PM All use subject to JSTOR Terms and Conditions

TESOL QUARTERLYVol.27, No. 1,Spring1993

The StructuralSyllabusand SecondLanguageAcquisition ROD ELLIS University Japan Temple

This paper examinesthe case fora structuralsyllabusin the light of second language acquisitionresearch.It argues thata structural syllabuscannoteasilyserveas a basisfordevelopingimplicit knowledge of a secondlanguagebecause of thelearnability are problem-learners oftenunabletolearnthestructural theyare taughtbecause properties themannerin whichtheyare taughtdoes notcorrespondto theway learnersacquire them.It is possible,however,to envisagea role for modelof L2 acquisition. a structural syllabusbased on a weakinterface This role consistsof intake (i.e., causing learnersto pay facilitation attentionto specificformalfeaturesin theinputand tonoticethegap betweenthesefeaturesand the ones theytypically use in theirown A can also structural serve as a basis fortheconstrucsyllabus output). tion of problem-solving tasksdesigned to develop learners'explicit of grammatical knowledge properties.It is arguedthatthisknowledge may facilitatesubsequentintake.The role proposedfora structural syllabus,therefore,is a substantialone. It is recognized,however, that such a syllabuswill need to be used alongsidesome kind of meaning-basedsyllabus,whichis designedto providelearnerswith in the second language. opportunitiesforcommunicating

structuralsyllabusconsistsof a listof grammatical items,usually arranged in the order in which theyare to be taught. This kind of syllabus is probably still the most common in language teaching today. Yalden (1983) describes it as "traditional"on the grounds that it is the basis of the grammar translationand audiolingual methods. However, it also serves as a basis for more "modern" methods-Total PhysicalResponse (Asher, 1977) and The Silent Way (Gattegno, 1972), for example. The move towards a communicative approach to language pedagogy in the 1970s and 1980s resulted in alternative syllabuses (in particular, the notional-functionalsyllabus (Wilkins, 1976), the task-based or procedural syllabus (Prabhu, 1987), and the process

A

91

This content downloaded from 81.102.155.33 on Sat, 28 Sep 2013 12:16:33 PM All use subject to JSTOR Terms and Conditions

syllabus(Breen, 1984). These syllabusescontinueto attracta lot of attention,but theyhave nevertotallyreplacedthestructuralsyllabus. The problemsofa structural discussedindetailinnumerous syllabus, the 1970s 1980s and (see Krahnke,1987),havenot publicationsduring the disappeared,however.The principalproblemis thatof learnability, extenttowhichitis possibleforlearnerstolearnthestructures theyare taught.This problemhas alwaysbeen recognizedbylanguageteaching methodologists(see Palmer, 1917), but it has been given additional weightbyresearchwhichhasshownthattheacquisitionofspecificgrammaticalfeaturesis constraineddevelopmentally. Corder (1967) suggestedthatlearnerspossessa "built-insyllabus,"whichregulateswhen itis possibleforthemto acquireeach grammatical feature.Subsequent studiesofnaturalistic languagelearning(see Hatch,1978a;Meisel,Clahsen,& Pienemann,1981; Wode, 1980) havegivenempiricalsupportto thisclaim.Also,studiesdesignedtoinvestigate whetherlearnerssucceed in learningthestructures theyare taught(e.g.,Ellis,1984, 1989; Felix, 1981; Pienemann,1984, 1989) suggestthatoftentheyare unableto internalizenewstructural knowledgeina mannerthatenablesthemtouse it productively in communicationunlesstheyare readyto do so. For example, Pienemann(1984) has providedevidence thatlearnersof Germanas a secondlanguageonlyacquirea featuresuchas inversionif thatare easierto theyhave previouslyacquiredwordorderstructures process.In otherwords,in orderto acquire FeatureD, learnersmust therefore, alreadyhave acquired FeaturesA, B, and C. Learnability, remainsa centralproblemin syllabusdesign.How can thecontentof a syllabusbe selectedand gradedina waythatiscompatiblewiththelearner's abilityto learn?This is a problemforanysyllabus,butit becomes acute whenthecontentis specifiedin grammatical terms. The main purpose of thispaper is to address thisproblemand to presenta proposalforhowitmightbe overcome.The paper willbegin witha briefdiscussionof thedifference betweentwotypesoflinguistic and explicit This distinctionunderlies knowledge. knowledge-implicit muchof thediscussionin therestof thepaper. It willalso considerthe relationshipbetweenthese two typesof knowledge.There followsa detaileddiscussionof structuralsyllabusesin relationto each typeof knowledge.The main argumentof thispaper is thatthe structural syllabusis a validdeviceforraisinglearners'consciousnessaboutgrammar; thisrole is discussedin the concludingsection.

IMPLICIT AND EXPLICIT KNOWLEDGE OF AN L2 It has been hypothesizedthatthe learnerinternalizestwotypesof knowledge-implicitand explicitknowledge.As Bialystok(1981) has 92

TESOL QUARTERLY

This content downloaded from 81.102.155.33 on Sat, 28 Sep 2013 12:16:33 PM All use subject to JSTOR Terms and Conditions

is commonincognitivepsychology. pointedout,thisdistinction Explicit to thatis analyzed(in thesense thatitcan refers knowledge knowledge be describedand classified), abstract(in thesensethatittakestheform of some underlyinggeneralizationof actual linguisticbehavior),and explanatory(in the sense thatit can providea reasonablyobjective accountof how grammaris used in actual communication). Explicit knowledge is available to the learner as a conscious but it is not the same as "articulatedknowledge"(i.e., representation, or written accountsof the knowledge).A learner may have spoken constructeda consciousabstractrepresentation of a grammaticalrule on have formulated an idea that -s the end of a noun signals (e.g., morethanone) and yetnotbe able to put thisidea intowords.Often, however,explicitknowledgeis developedtogetherwithmetalinguistic knowledge(e.g., termssuch as plural),and thishelps the learnerto articulateit. Two kindsof implicit can be identified;formulaicknowlknowledge and rule-based Formulaic consistsof readyedge knowledge. knowledge made chunksof language-whole utterances,such as I don'tknowor utteranceframeswithone or more emptyslots,such as Can I have a

? Rule-basedknowledgeconsists of generalized and abstract

structures whichhave been internalized.In bothcases,theknowledge is intuitive.Nativespeakersknowa large numberof formulaswhich theyhave learned as unanalyzedunits(see Pawley& Syder, 1983). They also know rules thatenable themto understandand produce novel sentenceswithoutconsciouseffort.Implicitknowledgeof rules is largelyhidden and we know relativelylittleabout how theyare representedin themind.It is doubtful,however,whetherthemanner of theirrepresentation correspondscloselyto thewaytheyare representedas explicitknowledge,one of thereasonswhypublishedgrammarsgenerallydo notclaimthattherulestheydescribehavepsychological validity. Because implicitknowledgebecomesmanifestonlyin actualperformance (bothcomprehensionand production),it is, perhaps,not surprisingto findthatthereis disagreementconcerningthenatureof the mechanismsresponsibleforitsacquisition,particularly whererulesare involved.Whereassome researchers(e.g., White,1987) viewrulesin both nativespeaker and learnergrammarsas primarilylinguisticin nature,others(e.g., Clahsen, 1984; McLaughlin,1978) see themas cognitive(i.e., involvingthe same generalmechanismsthatunderlie otherkindsof learning).Althoughmuchof theresearchintodevelopmentalsequencesdoes notspecifywhichtypeofknowledgeis involved, itis clearthatitis implicitknowledgethattheresearchershavein mind. For example,Wode's (1980) accountofhowGermanchildrenprogress througha seriesofstagesin acquiringEnglishnegativesand interrogaTHE STRUCTURAL

SYLLABUS AND SLA

This content downloaded from 81.102.155.33 on Sat, 28 Sep 2013 12:16:33 PM All use subject to JSTOR Terms and Conditions

93

is implicit tivesassumesthattheknowledgetheyare slowlyconstructing ratherthan explicit. Anotherdistinction fromcognitivepsychology thatis oftenreferred to in L2 acquisitionresearchis declarative and procedural knowledge. These termswereused initially byRyle(1949) and subsequentlytaken up by cognitivepsychologistslike Anderson (1983) to distinguish knowledgeas a set of facts(declarativeknowledge)and knowledge about how to do things(procedural knowledge).An example may make thisclearer.Knowledgeof the rulesof the highwaycode (e.g., would constitutedeclarativeknowledge overtaking) Alwayssignalbefore whileknowledgeof how to drivea car in accordancewiththeserules would be procedural.AndersoncharacterizesclassroomL2 learning rules(usually as beginningwithdeclarativeknowledgeof grammatical suppliedby the teacher),whichis graduallyproceduralized,resulting in the abilityto use the foreignlanguage withoutthinking. distinctions and declarative/procedural The explicit/implicit mayapshows. in but are as be to similar, fact, not, 1 they Figure very pear label thetypeof knowledgelearners Whereasthetermsexplicit/implicit possess in termsof whetherit is conscious or intuitive,the terms concernthedegreeofcontroloverL2 knowledge declarative/procedural the learnerhas, distinguishing knowledgethatcan be used onlywith effortthroughcontrolledprocessingfromknowledgethatcan be used effortlessly throughautomaticprocessing.(Bialystok,1982, also decontinua,whichshelabels knowledgeas twointersecting pictslinguistic +/- analyzedand +/- automatic.The formerrelatesto the implicit/ FIGURE 1 BetweenExplicit/Implicit The Difference and Declarative/Procedural Knowledge

Explicit

Implicit

94

Declarative

Procedural

TypeA Conscious knowledge of L2 items

TypeB

TypeC Intuitiveknowledge of L2 items

Conscious knowledge of learning, production,and communication strategies. The learner can use explicit knowledgeeasily and rapidly. TypeD Abilityto employ learning,production,and communicationstrategies automatically. The learner can use intuitive knowledgefluently. TESOL QUARTERLY

This content downloaded from 81.102.155.33 on Sat, 28 Sep 2013 12:16:33 PM All use subject to JSTOR Terms and Conditions

dimension and thelattertothedeclarative/procedural dimenexplicit refertoa knowledge whereas dimension, sion.)Thus,explicit/implicit refertoa processdimension. The keypointto declarative/procedural wecantalkaboutbothexplicit noteisthatthetwodistinctions intersect; as existing indeclarative andprocedural andimplicit form. knowledge the between and distinction is explicit implicit knowledge Although notitselfcontroversial, therelationship betweenthetwois.The main can convertinto pointof debateis whetherexplicitL2 knowledge One of the of traditional L2 knowledge. assumptions language implicit is thatexplicit methods basedon a structural knowlsyllabus teaching can become to implicit knowledge through practice. According edge

or proceduralize thisview,learnersautomatize knowledgethatis initially In activities. terms of Figure 1, this is explicitby doing grammar tantamountto claimingthatpracticeenables learnersto move from Type A knowledgeto Type D, the goal of mostlanguage programs. The notionof automatizingor proceduralizing explicitknowledgeso thatit becomes implicitis a somewhatconfusedone. It derivesfrom the failureto clearlydistinguishexplicit/implicit knowledgefromdewhereas it is legitimateto talk Thus, clarative/procedural knowledge. about the proceduralizationof declarativeknowledge,it is not legitimate to equate this with the conversionof explicitinto implicit knowledge. The keyissue-and itis here thatwe runup againstthelearnability question-is whetherwecan manipulatetheprocessbywhicha learner moves fromType A to Type D knowledge.Two positionscan be and an interface distinguished-a noninterface position.According to the

former,it is impossibleto lead learnersfromType A to Type D knowledge through practicingdeclarativeexplicit knowledge (as shownin Figure2a). This positionsees Type D knowledgeas deriving fromproceduralizingType C knowledge.Practicingexplicitknowlin usingthisknowledge edge (Type A) may resultin greaterfacility but will still involve conscious L2 knowledge.The B) (Type accessing interfacepositioncomes in a strongand a weak form.Accordingto the strongversion,Type A knowledgecan be convertedintoType D on thistaking knowledgethroughpracticeand thereare no constraints to the weak version, 2b). (see place Figure According Type A knowlinto C learners are ready edge maydevelop Type knowledgeproviding to accommodatethe new knowledgeintotheirinterlanguagesystems. Opportunitiesforformally practicingthenewknowledgeor forcomin contexts thatcall for its use willbe needed municatingnaturally beforeType D knowledgedevelops (see Figure2c). Krashen(1981) has argued strongly in favorof a noninterface position.He argues thatexplicitknowledgemayassistlearnersin certain kinds of language performance in the form of monitoringbut that it

THE STRUCTURAL

SYLLABUS AND SLA

This content downloaded from 81.102.155.33 on Sat, 28 Sep 2013 12:16:33 PM All use subject to JSTOR Terms and Conditions

95

FIGURE 2 The Noninterfaceand the Strong and Weak InterfacePositions

(a) THE NONINTERFACE Type A Explicit/ declarative

POSITION - Type B Explicit/ procedural

Practice

Type C Implicit/ declarative

Type D , Implicit/ procedural

. Communicativelanguage use

(b) THE STRONG INTERFACE POSITION Type B Explicit/ procedural

Type A Explicit/ declarative Practice P

Type C Implicit/ declarative

Type D Implicit/ procedural

(c) THE WEAK INTERFACE POSITION Type B Explicit/ procedural

Type A Explicit/ declarative

Grammar instruction

Type C Implicit/ declarative

---

Communicative language use/ practice

-

Type D Implicit/ procedural

does nothelp themto acquireimplicitknowledge.Others(e.g.,Gregg, 1984; McLaughlin, 1978; Sharwood Smith,1981) have opted for a stronginterfaceposition,accordingto whichexplicitknowledgecan change intoimplicitknowledgeas a resultof practice. It is my contentionthatthe evidenceavailablefromresearchinto on L2 learning(see Ellis,1990,and theeffectsof grammarinstruction Larsen-Freeman& Long, 1991, forrecentreviewsof theliterature)is 96

TESOL QUARTERLY

This content downloaded from 81.102.155.33 on Sat, 28 Sep 2013 12:16:33 PM All use subject to JSTOR Terms and Conditions

compatibleonlywitha weakinterfaceposition.This researchsuggests the followingconclusions: 1. Grammarinstruction resultsin fasterlearningand in higherlevels of L2 grammaticalaccuracy(see Long, 1983; Pica, 1983). directedat a grammaticalfeaturethatlearn2. Grammarinstruction ers are notreadyto acquireas implicitknowledgedoes notsucceed (see Felix, 1981; Pienemann,1984, 1989). directedat a grammaticalfeaturethatlearn3. Grammarinstruction ers are ready to acquire as implicitknowledgeis successful(see Harley,1989; Pienemann,1984, 1989). The firstconclusioncannotbe easilyexplainedbya noninterface theory.The second conclusioncontradictsa stronginterfacetheory.All threeconclusionsare compatiblewitha weak interfacetheory. Figure3 providesa modelofL2 acquisitionthatincorporatesa weak interfaceposition.The modeldistinguishes input,intake,and implicit L2 knowledge.Inputrefersto thesamplesof theL2 thatthelearneris exposed to as a resultof contactwiththelanguagein communication (oral and written).Formal instructioncan also provide input (i.e., general exposure to the L2), althoughits raison d'etre is to teach FIGURE 3 A Model of L2 Acquisition Incorporatinga Weak InterfacePosition

Formal instruction

I

--

I

Noticing

Input -

aIL

=

-

I

Noticing

Explicit knowledge

of L2

I

thegap

I -

Intake

I

I

Monitoring

Monitoring

t .-

Implicit knowledge of L2 (IL' system)

I

T Output

interlanguage.

THE STRUCTURAL

SYLLABUS AND SLA

This content downloaded from 81.102.155.33 on Sat, 28 Sep 2013 12:16:33 PM All use subject to JSTOR Terms and Conditions

97

specificgrammaticalitems.Intakerefersto the linguisticpropertiesin theinputthatthelearnerattendsto. Not all of thesepropertieswillbe immediately incorporatedintothelearner'sinterlanguage system;only those featuresthatare finallyincorporatedbecome implicit knowledge of the L2. The model showsthatimplicitknowledgecan be internalizedin two ways.The mainwayis byderivingintakefromtheinput.A secondary way is directlyfromthe explicitknowledgethatis learned through This wayis consideredsecondaryfortworeasons; formalinstruction. of newgrammaticalknowledgederivedin thiswayis the amount first, because to be limited onlya smallportionofthetotalgrammatical likely of a language can be consciouslylearned (see Krashen, properties and second,explicitknowledgecan onlyfeed directlyintoim1982), readyto incorporate plicitknowledgeiflearnersare developmentally it (hence the dottedline). The model posits a numberof other uses of explicitknowledge, however: 1. Explicitknowledgeis also availableforuse in monitoring(as procan occurbeforean utterance posed byKrashen,1982). Monitoring is produced or after.Monitoredoutputconstitutesone source of can apparentlyinteract input.As Terrell(1991) putsit"monitoring withacquisition,resultingin learnersacquiringtheirown output" (p. 61). 2. Explicitknowledgecan help learnersto noticefeaturesin theinput and also to noticethe meaningsthattheyrealize. For example,if learnersknowthatpluralnouns have an -s,theyare morelikelyto noticethe -s on the ends of nouns theyhear or read in inputand also more likelyto associate the -s morphemewiththe meaning more than one. In a sense,then,as Terrell(1991, p. 58) suggests, explicitknowledgecan functionas a kindof "advance organizer" thathelps the learnerto comprehendand segmentthe inputand focuser"thatenablesthelearnerto estabalso as a "meaning-form lish meaning-form relationships. 3. Explicitknowledgemayhelp learnersto incorporatefeaturesthat have becomeintakeintotheirdevelopinginterlanguagegrammars the process by whichtheycompare theirexisting by facilitating featurewiththatactuallyobserved of a grammatical representation knowthatpluralnouns have if learners For in the input. example, betweenthis thedifference to notice are better an -s,they equipped in their own omission and its in the feature output. input are all mentalprocesses Monitoring,noticing,and noticing-the-gap box" in Figure3. Because the and hence are showninsidethe "black 98

TESOL QUARTERLY

This content downloaded from 81.102.155.33 on Sat, 28 Sep 2013 12:16:33 PM All use subject to JSTOR Terms and Conditions

availabilityof relevantexplicitknowledgedoes not guarantee their operation,all threeprocessesare representedby dottedlines. A key aspect of this model is the role thatexplicitknowledgeis Accordingto hypothesizedto play in noticingand noticing-the-gap. is frequently(and perhaps Schmidt(1990), the process of noticing necessarily)a consciousone. He definesitoperationallyas availability for verbal report.A varietyof factorscan induce learnersto notice featuresin the input-the demands of a task,the high frequencyof an item in the input,the unusual natureof a feature,the inherent thathighlights a feature.Noticingsalienceofa feature,and interaction occurs learners makethe effort & when (Schmidt Frota,1986) the-gap to establishin whatwaysa newfeature,whichtheyhave heeded in the input,is differentfromtheirexistinginterlanguagerepresentation. This entailssome formof comparisonbetweenwhatlearnerstypically do in theiroutput and what is presentin the input. Learners may noticea featurebut notbotherto noticethegap. Neithernoticingnor guaranteesthatthenewfeaturewillbe incorporated noticing-the-gap into the learner'sinterlanguagesystem,as in manycases thiswillbe constrainedby the learner'sstageof development. This model, then,envisagesthat explicitknowledgecan convert condidirectlyintoimplicitknowledgeunder certain,fairlystringent tionsrelatedto the learner'sstageof development.It also allows for explicitknowledgeto have an indirecteffecton acquisitionbyhelping to facilitate theprocessesofnoticingand noticing-the-gap. It is hypothesized thatlearnerswho knowabout a grammaticalfeaturebecause are in a better theyhave learnedaboutitthroughgrammarinstruction to heed this feature when it occurs in the input position subsequently and also are betterable to noticethedifference betweentheinputand theirown production.(Empiricalevidencein supportoftheclaimthat explicitknowledgefacilitates subsequentnoticingis providedbyFotos, functions as a kindof "acquisitionfacilitator" 1992).Explicitknowledge on whichto hang subsequent "hooks" (Seliger, 1979) by providing acquisition(Lightbown,1985).

THE STRUCTURAL SYLLABUS AND L2 ACQUISITION A structuralsyllabusemploysa synthetic definedby teaching strategy, Wilkins(1976) as "one in whichthedifferent partsof thelanguageare so thatacquisitionis a processof taughtseparatelyand step-by-step accumulation of the until thewholeof thelanguagehas gradual parts been builtup" (p. 2). The executionof thisteachingstrategy involves the course designerin makingprincipleddecisionsregardingwhich THE STRUCTURAL SYLLABUS AND SLA

This content downloaded from 81.102.155.33 on Sat, 28 Sep 2013 12:16:33 PM All use subject to JSTOR Terms and Conditions

99

partsof thelanguageto teach(i.e., selection)and whichorderto teach them in (i.e., grading). However,as Wilkinspointsout, the job of theitemswhichhavebeen presentedin smallpiecesis left synthesizing to the learner. A structuralsyllabuscan serve as a basis for the developmentof eitherimplicitor explicitknowledge.In thecase oftheformer,theaim of the syllabusis the developmentof the kindof intuitiveknowledge thatis required to communicatein the L2. In the case of the latter, the aim is knowledgeaboutthe language-some kind of conscious ofthe"rules"thatmakeup thelanguage.The structural representation in the audiolingualand oral-situational used methodsare syllabuses directed at implicitknowledge,whereas the grammar-translation methodis directedat explicitknowledge.

The StructuralSyllabusand ImplicitKnowledge In the case of a structuralsyllabusforimplicitknowledge,the aim is to "teach the language, not about the language" (Moulton, 1961). thelanguagecan referto both the comprehension The termteaching and the productionof grammaticalitems. It is possible to teach a butin mostmethodsthatemploya structureonlyforcomprehension, is enable the learnersto produce the the to aim structuralsyllabus, it is insistence on productionthat As will this we itemscorrectly. see, of to solution thedifficulties A of the createsmany problems. possible worthwhile for the lesser but still be settle to a structural syllabusmight goal of teachinggrammarforcomprehension. When production is the goal, another distinctionis important. strictaudioWhereassomestructural syllabuses(e.g.,thoseunderlying will be fully idea that item each lingual courses) are based on the those undermasteredbeforeanotheritemis introduced,others(e.g., that such as lyingmore modern approaches to grammarteaching in the occurs describedin Ur, 1988) recognizethatmastery only long termand thateach itemwillprobablyonlybe partiallyacquiredbefore anotheris introduced.These twoviewsof structural syllabuseswillbe The and gradualmastery. problemsof referredto as immediate mastery both willnow be examined.

STRUCTURAL SYLLABUSES FOR IMMEDIATE MASTERY Ultimatelya structuralsyllabusdirectedat immediatemasterywill only work if the order in which the grammaticalitemsare taught correspondsto theorderinwhichthelearnerscan learnthem.In other 100

TESOL QUARTERLY

This content downloaded from 81.102.155.33 on Sat, 28 Sep 2013 12:16:33 PM All use subject to JSTOR Terms and Conditions

words,the syllabusmustsatisfythecriterionof learnability. Designers of structural syllabuseshavealwaysacknowledgedthis,and learnability has alwaysfiguredas one of the criteriaof selectionand grading. Mackey(1965), forinstance,identifiedfivefactorsthatcontributedto (i.e.,betweenthetargetlanguageand thenative learnability: similarity and learningload. The notionof brevity, regularity, language),clarity, factors is a rationalratherthan a that underlies these learnability one. It an externalaccountof reflects (or psycholinguistic empirical) whatought to be learnable. To whatextentdoes theorderingofitemsderivedfromtheseexternal criteriaconformwiththelearner's"built-insyllabus"?One wayof answeringthisquestionis to compare the order of itemsin sample structuralsyllabuseswiththe naturalorderof acquisitionreportedin studiesof L2 acquisition.If it can be shownthatthe orders do not match,the solutionis simple-devise a syllabuswhere theydo. It is doubtfulwhethersuch a solutionis possible,however.L2 acquisition all thefeaturesthatthelearnerwillneed researchhas notinvestigated to be taught,so thereis onlyinformation relatingto theacquisitionof available.Also, a fairlysmallnumberof grammaticalitemscurrently research that has the"natuthereis uncertainty regarding investigated ral order"of acquisition.For example,a numberof studiesproduced evidence thatlearnersof differentages and withdifferentfirstlanguages followthe same orderof acquisitionof a setof Englishgrammaticalmorphemes(Krashen,1977), but thisresearchhas been challenged on a numberof grounds(see Hatch, 1978b). In particular,itis to maintaintheviewthatL2 acquisitioninvolvesthesystematic difficult masteryof discretegrammaticalitems,as this researchappears to assume. Anotherproblemis thatthegrammatical itemsfoundin a structural do not have syllabus psycholinguistic validity.Bley-Vroman(1983) has that the arguedconvincingly categoriesof a descriptive grammar,from whichthe itemsof a structuralsyllabusare derived,bear no relation to the mentalcategorieswhichlearnersconstructin the process of learninga language. Learners appear to constructtheirown rules, and hencedo notcorrespondto anyof manyof whichare transitional the rules found in a referencegrammarof the targetlanguage. For example,thesequenceofacquisitionforGermanwordorderrules(see Meisel,Clahsen,& Pienemann,1981) containsa stagewhereadverbpreposingoccurs,as in Example 1. 1. Heute wirgehenins Kino. (Today we go to the cinema.) Such a rule

an advanceon theprevious ifused, represents stageduringwhichadverbs, theadverb-preposing rule However, onlyoccurat theendsofutterances. inan errorbecauseinthetarget results theapplication languageitobligates

of a furtherrule,inversion: THE STRUCTURAL

SYLLABUS AND SLA

This content downloaded from 81.102.155.33 on Sat, 28 Sep 2013 12:16:33 PM All use subject to JSTOR Terms and Conditions

101

wirinsKino.(Todaygowetothecinema.)Thisisa rulewhich 2. Heutegehen thelearnerdoes notyetknow.This exampleillustrates howacquisition involvestheconstruction of rulesnotfoundin thetargetlanguage(i.e., without and showsthatprogress can actually inversion) adverb-preposing resultin errorsnotevidentat an earlierstage. A thirdproblemis thatstructuralsyllabusestreateach itemas discreteand separate.It has been shown,however,thattheacquisitionof a new formcan affectthe organizationof the learner'sentiremental grammar(see Huebner, 1983). The rulesthatmakeup thisgrammar are interrelatedin complexways,so any change mayinvolvenotjust an additionof a new formbut the restructuring of the whole system This not takeplace (McLaughlin,1990). reorganization may necessarily inaccordancewiththewaythetargetlanguagegrammarisconstructed. If the implicitknowledgesystemthat a learnerbuilds is viewed as a form-function network(Ellis, 1985; Rutherford,1987), then the of a formleads thelearnernotjusttoassignita certain new acquisition functionalvalue but also to reassessthe functionalvalues assignedto formspreviouslyacquired. It followsthat the nature of the formfunctionsystemconstructedat any one stage of developmentwillbe arrivesat thetargetlanguagegrammar. unique untilthelearnerfinally In short,itis difficult to see howa structural syllabusdirectedat the immediatemasteryof grammaticalitems(definedas the abilityto use the itemsaccuratelyin production)can cope withthese learnability issues.

STRUCTURAL SYLLABUSES FOR GRADUAL MASTERY A case mightstillbe made forthe use of structuralsyllabusesas a basis forteachingimplicitknowledgeforuse in productionifitcan be showntheyare compatiblewitha viewof acquisitionas a processof gradual mastery.The structuralsyllabusesassociatedwiththe oralsituationapproach were,in fact,based on thisviewof L2 acquisition. consciousand subconscious Palmer(1917), forexample,distinguishes former as a sees the and clearly precursorofthelatter,at least learning whereadultlearnersare concerned.Palmeropposesa purely"natural" methodthatcatersto subconsciouslearningon the groundsthatit is He arguesfor"consciousstudyofthemicrocosm"through inefficient. the graded presentationof linguisticitems.He believesit possibleto guidethelearnerthrougha seriesofgeneralstagesinvolving(a) receivand (c) developing ingknowledge,(b) fixingtheknowledgeinmemory, theabilitytouse theknowledgeas skill.In otherwords,Palmeradheres to a viewof languagelearningsimilarto thatof proponentsof a strong 102

TESOL QUARTERLY

This content downloaded from 81.102.155.33 on Sat, 28 Sep 2013 12:16:33 PM All use subject to JSTOR Terms and Conditions

interfacetheory.It is probablytrueto say thatviewsclose to thoseof Palmerunderliethecontinueduse ofstructural syllabusesforteaching implicitknowledgetoday(see Ur, 1988). How can a structuralsyllabusreflectthe processby whichlearners achievegradual masteryof linguisticfeatures?Clearlya simplelinear syllabuscannotdo so. Learnersmaybe able to receivea new feature and perhapsalso fixthe knowledgein memory,but it is unlikelythat a singletreatmentwillresultin theirdevelopingthe abilityto use the knowledgeas skill.One wayaround thisproblemmightbe to design a spiralsyllabus.Howatt (1974) suggeststhatsuch a syllabusaccords betterwiththenaturalprocessof learningbecause learnershave their attentiondirectedat thesame itemson severaloccasionsbut in differand withdifferent entcombinations meanings.It is possible,therefore, thata spiralsyllabuscan caterto implicitknowledge. The key question,however,is whetherit is possibleto guide the processby whichexplicitknowledgebecomesimplicitknowledgeby items.Accordingto of grammatical meansof a cyclicalre-presentation the model of L2 acquisitionshownin Figure3, thisis onlypossibleif the presentationof an itemcoincideswiththe learner'sreadinessto acquire it. A spiral syllabusmay increase the likelihoodof this ocaffair.The onlywayto guarantee curring,but it is stilla hit-or-miss of a structural theeffectiveness syllabusdirectedat implicitknowledge is byensuringitis compatiblewiththelearner'sinternalsyllabus,and this,as we have alreadyseen, is problematic.

STRUCTURAL

SYLLABUSES FOR COMPREHENSION

It is possible,however,thatthe problemsof the structuralsyllabus directedat implicitknowledgecan be overcomeif the goal of the syllabusis to enable learnersto comprehendratherthan to produce the itemswithinit. In thiscase, the teachingmaterialsbased on the thatenablelearnersto (a) hearsentensyllabuswouldprovideactivities listedin the syllabusand (b) identify ces containingthe structures the to functions the features establish form(i.e., by performed specific Forexample,tohelplearnerscomprehendthe meaningrelationships). of -s, they mightbe asked to listento sentencessuchas meaning plural He putthebookson thetableand He gavehisfriendthepenand to choose whichpicturesfroma set of picturescorrespondto the meaningsof thesentences.The picturesforeach sentencewouldincludedistractors (e.g., one showinga man puttinga singlebook on the tableor giving hisfriendtwopens)as wellas accuraterepresentations ofthesentences actuallysaid. Such activitieswould be directedat helpingthe learner to notice new grammatical features in the input and the grammatical

THE STRUCTURAL

SYLLABUS AND SLA

This content downloaded from 81.102.155.33 on Sat, 28 Sep 2013 12:16:33 PM All use subject to JSTOR Terms and Conditions

103

meaningstheyrealize.Like traditionalgrammarmaterials,theywould be speciallycontrivedto focusthelearner'sattentionon specificitems, but theywould differfromthemin thattheywould not requirethe learnerto produce sentencescontainingthe items. Pienemann's(1985) distinction betweeninputforcomprehension and a for such rationale a Pienemann input forproduction provides syllabus. argues thatthe developmentalsequence throughwhichlearnerspass ofa seriesofprocessingoperationsresponreflects thegradualmastery sible for language production.His own proposal regardingsyllabus design is as follows: 1. Do notdemanda learning at a givenstage processwhichis impossible be in linewithstagesofacquisition). (i.e. orderofteaching objectives deviantforms. 2. Butdo notintroduce 3. The generalinputmaycontainstructures whichwerenotintroduced forproduction. (p. 63) This constitutes a seriousattempttosuggesthowa structural syllabus but it runs up againsta numberof can take account of learnability, objections-our knowledge of developmental sequences remains patchyand relatesprimarilyto formalfeaturesof the language (i.e., littleis knownabout how learnersbuild form-function networks).It is not clear how teachersare supposed to identifythe developmental stageswhichindividuallearnershave reachedor whetherthiscan be practicallyachieved, and it requires teachersto constructteaching needs of individuallearners, programstailoredto thepsycholinguistic in many has as which, Lightbown(1985) pointedout,maybe unrealistic all arise because Pienemann These situations'. objections teaching viewsthe primarygoal of a structuralsyllabusas thatof providing input for production.They do not appear to apply if the syllabusis directedat providinginputforcomprehension.Pienemannsuggests that such input can be allowed to arise naturallyin the course of thatit might but he does not considerthe possibility communication, be contrivedthroughformalinstruction. It is possible,however,to envisage an approach where input for to ensurethatthe comprehensionis carefullyplannedand structured features.This to learneris systematically exposed specificgrammatical is no that the in the sense one modest is a longer the goal proposal facilitathe of but the full L2 of only implicitknowledge development 'Currently,Pienemann and his associates are workingon various waysin whicha learner's stage of developmentcan be quicklydiagnosed. In particular,theyhave developed sophisticated computer softwareto facilitatediagnosis. It is not clear to me how practicalthiswill be in manyteachingsituations,as it necessitatesteachers'obtainingreliabledata regarding the structureslearnersare able to performat any one stage of development-a painstaking and time-consumingprocess. 104

TESOL QUARTERLY

This content downloaded from 81.102.155.33 on Sat, 28 Sep 2013 12:16:33 PM All use subject to JSTOR Terms and Conditions

a substantially reducedgoal for tionof intake.Althoughthisconstitutes stilla significant one. Chaudron structuralsyllabuses,itis nevertheless second status in that intake has "has important (1985) language argued research"(p. 1), and a similarpositioncan be adopted withregardto itsimportanceforlanguage pedagogy,There is a need, as Chaudron preciselywhichfactorsinfluenceintake. emphasizes,forinvestigating One way in whichthiscan be undertakenis throughstudiesof how affectslearners'abilityto noticeand comprehend formalinstruction specificgrammaticalitems(see Van Patten& Cadierno, 1991, foran example of such a study).

SUMMARY We have consideredtwo viewsof the structuralsyllabus-one that sees it as a basis forteachingaccurateproductionand the otherthat intakethroughthe comprehensionof sees it as a basis forfacilitating In the items. case of the former,the structural specificgrammatical syllabuscan serveas a deviceforbringingabouttheimmediatemastery of grammaticalitems.We have seen thatsucha viewis notcompatible withwhatis knownabout the waylearnersacquire an L2. It can also serveas a device forensuringthe gradual masteryof items.We have seen thatthisview also runs up againstthe problemof learnability, evenifthesyllabusrecyclestheitems.In bothcasesthedifficulties arise as a resultof treatingthe structuralsyllabusas an instrumentfor itemscorrectly. It has been teachinglearnersto produce grammatical that these difficulties be if overcome the suggested might goal becomes the comprehensionratherthanthe productionof grammaticalitems. In thiscase, the goal of the syllabusis intake ratherthanthe facilitation fulldevelopmentof implicitknowledge.

The StructuralSyllabus and Explicit Knowledge Anotherwayinwhichtheproblemof learnability canbe side-stepped isbymakingthegoal ofa structural rather thanimplicit syllabusexplicit the In serves as a for basis knowledge. otherwords, syllabus developing a consciousratherthanintuitiveunderstandingof grammaticalrules, and thereis no expectancythatlearnerswillbe able to use the knowledge theyhavelearnedinfluentproduction.This amountstoa reversal of Moulton'sslogan,citedabove-we shouldteachaboutthelanguage, not the language. This proposal restson twoprincipalassumptions: 1. The acquisitionof explicitknowledgecontributesto the developmentof L2 proficiency. THE STRUCTURAL

SYLLABUS AND SLA

This content downloaded from 81.102.155.33 on Sat, 28 Sep 2013 12:16:33 PM All use subject to JSTOR Terms and Conditions

105

2. The acquisitionofexplicitknowledgecan takeplaceas an accumulationof discreteentities. Assumption1 derivesfromand is supportedbythe weak-interface model discussedearlier.Assumption2 isjustifiedifitis acceptedthat explicitknowledgeconsistsof a body of consciousknowledgeabout isolatedgrammatical itemsand rules,a viewadoptedbymanylearners, as thisquotationfromMoore (1989) illustrates: Asa learnerI havetwomajorproblems toovercome-accumulating a large number ofpartialentities, whether rules, theybe lexicalitems, grammatical or whatever; to tryout combinations of secondly, finding opportunities inbothstructured andunstructured situations. OnceI haveacquired, them, I ambetterplacedtocommunicate than say,twothousandpartialentities, ifI haveacquiredonlya hundred.(p. 157) As Moore recognizes,accumulatingthe "factsof language"is not the whole of acquisition,but it can help to get the learnerstarted. There are educationalas wellas psycholinguistic argumentsin favor of teachinggrammaticalfacts.Breen (1985) addressesthequestionof whatis authenticforthe social situationof the classroomand argues thatbecause theraisond'etreof thissituationis languagelearning,the contentof the teachingprogramshould be drawnfromthe "culture and social of the classroom."Breen's idea is thatthe communicative serve of should content for work. as One language aspects learning source of such contentis the linguisticand pragmaticsystemsof the language. A syllabusthatisolatesvariousformaland functionalfeatureswitha viewto makingthesethetopicsoflearningactivities might a accordwiththeexpectationsof manylearners.Grammarconstitutes content. seriousand intellectually challenging Whatwilla syllabusforexplicitknowledgeconsistof?On whatbasis contentproceed? shouldtheselectionand gradingof thegrammatical We willconsidersomewhatbrieflya numberof possibilities. Perhaps the mostobviousone is to make use of the criteriawhich twogeneral used. Widdowson(1968) identifies have been traditionally drawn on: relative have that (a) difficulty syllabusdesigners principles and (b) usefulness(i.e.,thecoveragevalueofan itemand theclassroom value of the item). However, as Widdowsonpoints out, these two principlesare oftenin conflictas whatis usefulis oftennot relatively to whatextenttheycan be applied in simple.It is notclear,therefore, mannerand Halliday,McIntosh,and Strevens(1964) are a systematic probablyrightin claimingthatitis "practicalteachingexperience"that oftenservesas a basis forselectionand grading,althoughthisrather thereis begs thequestionas to whatthisactuallyconsistsof. Certainly, 106

TESOL QUARTERLY

This content downloaded from 81.102.155.33 on Sat, 28 Sep 2013 12:16:33 PM All use subject to JSTOR Terms and Conditions

considerableagreementregardingboth whatstructuresto teach (at leastin generalcourses)and in whatordertheyshouldbe taught(see Yalden, 1983). criteriamightbe sharpenedis by One wayin whichthesetraditional in usingthe insightsobtainedfromthe studyof linguisticmarkedness' the second possibility. The notion languagelearning.This constitutes of markednessis notitselfnew-adherentsof structural gradinghave long workedwitha similarnotionto thatunderlyingmuch current discussionof the concept(see Mackey's,1965,discussionof "regularity,"forexample),but recentstudiesdo givegreaterprecisionto the concept.For example,the NP + PP patternafterdativeverbssuch as give in sentenceslike Bob gave a gifttoIsabelcan be consideredless markedthan the NP + NP patternafterthe same verbsin sentences likeBobgaveIsabela gifton a numberof grounds.The lattersentence is more transparent, of the verband thedirectobjectis the integrity maintained,and it is more regular (i.e., just about all dative verbs permitthe NP + PP pattern,but only some permitthe NP + NP pattern).This kind of informationcan be used to make decisions regardingwhichlinguisticfeatureto introduceearlyand whichlate. But it is not yet clear how this informationshould be used. It can be argued thatlearnersgenerallyfindit easier to handle unmarked but it has also been sugfeatures,so theseshouldbe introducedfirst, gestedthatlearnerswillbe able to projecttheirknowledgeof marked featuresto associatedunmarkedfeatures(see Eckman,1985), which constitutesan argumentin favorof focusingattentionon marked features.Also, unmarkedfeaturesmaybe learned by mostlearners do notrequireexplicitattention.In contrast, naturallyand, therefore, markedfeaturesare oftennotacquired(see Bardovi-Harlig& Bofman, 1989; Long, 1988) unlessthe learners'consciousattentionis directed at them. On balance, the argumentsfavorthe selectionof marked ratherthan unmarkedfeaturesin a syllabusforexplicitknowledge. A thirdalternative fororganizingthecontentofa structural syllabus is derived fromanother old idea-that of remedial teaching.The contentof a remediallanguage programis establishedthroughthe identification and descriptionof learners'errors.It restson thesimple idea thatformallanguageteachingwillbe moreefficient ifitconcen2Markedness is not an altogetherclear notionas it has been definedin differentways.In one definitiona feature is considered marked if it can be shown to be less common in the world'slanguages thansome other,relatedfeature(i.e., typologicalmarkedness).In another definition,a feature is considered marked if its use is in some way more restrictedthan another related feature(e.g., an is more markedthan a because it occurs onlybeforenouns and adjectivesthatbegin witha vowel). UniversalGrammarsupplies yetanotherdefinition of markedness.The concept of markednessis perhaps bestconsidered at thispointin time as of potentialratherthan realized value to the designer of a structuralcourse.

THE STRUCTURAL SYLLABUS AND SLA

This content downloaded from 81.102.155.33 on Sat, 28 Sep 2013 12:16:33 PM All use subject to JSTOR Terms and Conditions

107

trateson whatthelearnerhas notlearnedratherthanon teachingthe whole grammar.However,as Corder (1981) has noted, thereis no reasonwhyremedialteachingshouldworkanybetterthaninitialteachcauses of errorsare takenintoaccount. ingunlessthe psycholinguistic Once again,then,we seemtocomeup againstthelearnability problem. In thecase ofa structural syllabusdesignedtoteachexplicitknowledge, thisproblemis side-stepped,however,if,as we have argued, it only ariseswhereimplicitknowledgeis involved.A remedialsyllabusmight consistof a listof structures whichhave been shownto be problematic to eitherlearnersin generalor, betterstill,to the particulargroup of learnersforwhomthe syllabusis intended.This constitutes a record of thepotentialdeviationsand serves,therefore, as a checklist.Armed withthislist,the teacherwould need to observethelearners'errorsin order to establishwhetherthe potentialdeviationsactuallyoccur in their productionand, if so, when. The teacherwould then devise activitiesto draw the learners'attentionto errorsand help themcompare the errorsto the correcttargetlanguage forms. To sum up, theaim of a structural syllabusforexplicitknowledgeis to raise learners'consciousnessabout how the targetlanguage grammarworks.As Larsen-Freeman(1991) has pointedout,thiswillinvolve formsare formed,(b) devel(a) drawingattentionto howgrammatical oping an understandingof how particulargrammaticalformssignal particulargrammaticalmeanings,and (c) helpinglearnersrealizewhat constitutes appropriateuse of the formsin context.The rationalefor thisuse of a structuralsyllabusis thatexplicitknowledgemay help learners to notice featuresin the input that they mightotherwise ignore and also to notice the gap betweenthe input and theirown interlanguageproductions.The contentof such a syllabusmightbe determinedon the basis of traditionalcriteriafor the selectionand grading of grammaticalstructures,by the principledselectionof markedlinguisticfeatures,or remedially,by identifying gaps in the learners'implicitknowledgethrougherroranalysis.

CONCLUSION: CONSCIOUSNESS-RAISING AND THE STRUCTURAL SYLLABUS This paper has soughtto presenta new rationaleforthe structural syllabus.The need for thishas arisen fromthe recognitionthatthe traditionalrationale,whichderivesfrombehavioristlearningtheory, solutionto the is inadequate because it cannotprovidea satisfactory learnability problem.The new rationalerestson theclaimthatgramratherthan mar teachingshould be directedat consciousness-raising refersto a delibpractice(see Fotos& Ellis,1991). Consciousness-raising 108

TESOL QUARTERLY

This content downloaded from 81.102.155.33 on Sat, 28 Sep 2013 12:16:33 PM All use subject to JSTOR Terms and Conditions

erate attempton the partof the teacherto make the learners'aware of specificfeaturesof the L2; it entailsan attemptto instillan understandingof the formaland functionalpropertiesof these features of them. by helpingthe learnersdevelop a cognitiverepresentation to the learner involves an on the other hand, Practice, attempt supply in conwithplentifulopportunitiesforproducingtargetedstructures trolledand freelanguageuse in orderto developfullyproceduralized implicitknowledge.It is notintended,however,to suggestthatpractice has no role at all in languageteaching.Practicemaystillbe important as a meansof helpinglearnersgaincontroloverformulaicknowledge, and it probablyalso has some place in the teachingof pronunciation. What is being challengedhere is the traditionalrole it has played in the teachingof grammaticalitems. In accordancewiththeprecedingdiscussion,twokindsofconsciousfor ness-raisingcan be identified.In the case of consciousness-raising the aim is to focusthe learners'attentionon the meancomprehension, ing(s) performedby specificgrammaticalproperties.It has been suggested that this is tantamountto helping the learner to intake-a of thefeatureas step forinternalization necessary(but not sufficient) of willbe achieved This type consciousness-raising implicitknowledge. and that induce a learner to notice of activities understand means by the featurein the input (i.e., activitiesthatrequire receptionrather thanproductionin theL2). In thecase ofconsciousness-raising forexplicit theaim is to help thelearnerlearnabouta particulargramknowledge, matical featureby developingan explicitrepresentationof how it worksin the targetlanguage.In manycases,thiswillinvolveteaching thelearnerthemetalanguageneeded to talkabout grammaticalrules. It has been hypothesizedthatexplicitknowledgealso aids the process of intakeformationbyfacilitating This noticingand noticing-the-gap. can be of achieved means of traditional by type consciousness-raising grammarexplanationof the kind found in the grammar-translation method.Anotherway,however,is to make use of problem-solving tasksthatsupplythe learnerswiththe data theyneed to discoverthe rule forthemselves.An example of such a taskis providedin Fotos and Ellis (1991). If such tasksare carriedout in the targetlanguage, theyservethe double purposeof raisinglearnersconsciousnessabout a specificgrammatical itemwhileprovidingopportunities forcommuin the will learners be nicating targetlanguage-the communicating about grammar. Traditionally,a structuralsyllabushas been used as the basis for designinga completelanguage course.This was possiblebecause the goal was proceduralimplicitknowledge,whichunderliesthe actual abilityto use the L2 in communication.However,if the goal of a structuralsyllabus is the lesser one of consciousness-raising,it can no

THE STRUCTURAL

SYLLABUS AND SLA

This content downloaded from 81.102.155.33 on Sat, 28 Sep 2013 12:16:33 PM All use subject to JSTOR Terms and Conditions

109

longerserveas a completecourse,as theultimategoal of mostcourses willcontinueto be the abilityto use the L2 in production.It follows thatthe structuralsyllabuscan only providepartof a course. It will need to be complementedby otherkindsof syllabusesthatare based on.the provisionof inputof the kind thathas been hypothesizedto or a task-basedsyllabus,for promoteimplicitknowledge-a functional The & between Crookes, 1991). preciserelationship example(see Long the structuralcomponentand these othercomponentsof an overall syllabusremainsto be decided. In restatingthe case fora structural we have also acknowledgeditsreducedvalue. The syllabus,therefore, ratherthanas a prime new structuralsyllabuswillserveas a facilitator moverof L2 acquisition. Finally,it needs to be acknowledgedthat many of the claims of boththemodelofL2 acquisitionshownin Figure3 and thepedagogical argumentsbased on ithaveonlylimitedsupportfromexistingempirical thatthevariousclaims research.Clearly,thereis a need todemonstrate are valid. Howthathave been made regardingconsciousness-raising it to that is advance here I am ever, legitimate pedagogical assuming proposalswithoutwaitingforthenecessaryempiricalsupporttobe collected.Indeed, empiricalL2 studiesprovideonlyone wayofvalidating such proposals.The other,whichmaywellbe moreimportant,is the methodof tryingthemout in theclassroomand using well-established them. accepting,or refining practicalexperienceas a basisforrejecting, Atthemoment,thebestthatcan be saidisthattheseclaimsare compatible withthe L2 researchthathas been carriedout to date. The essential syllabusas a pointsare (a) thatitis prematureto dismissthestructural oftherole basisforL2 acquisitionand (b) thatconsiderablemodification needed. to be is a to such syllabus likely traditionally given

ACKNOWLEDGMENTS readers,whose constructive I would like to thanktwoanonymousTESOL Quarterly commentshelped me greatlyto sharpen my arguments.

THE AUTHOR Rod Ellis is currentlyProfessorof Applied Linguisticsat Temple University Japan, where he works in the graduate programs in TESOL. Previouslyhe worked in Spain, Zambia, and the United Kingdom. He has published several books on second language acquisition.Currently,his main interestis the utilizationof SLA theoryand research in grammarteaching. 110

TESOL QUARTERLY

This content downloaded from 81.102.155.33 on Sat, 28 Sep 2013 12:16:33 PM All use subject to JSTOR Terms and Conditions

REFERENCES Anderson,J. (1983). Thearchitecture ofcognition. Cambridge,MA: Harvard UniversityPress. actions:The complete teachers' Asher, J. (1977). Learninganotherlanguage through Los Gatos, CA: Sky Oaks. guidebook. Bardovi-Hardlig,K., & Bofman,T. (1989). Attainmentof syntacticand morphologicalaccuracybyadvanced language learners.StudiesinSecondLanguageAcquisition,11, 17-34. Bialystock,E. (1981). The role of linguisticknowledge in second language use. Studiesin SecondLanguageAcquisition, 4, 31-45. Bialystok,E. (1982). On the relationshipbetween knowingand using linguistic forms.AppliedLinguistics, 3, 181-206. Bley-Vroman,R. (1983). The comparativefallacyin interlanguagestudies: The case of systematicity. LanguageLearning,33, 1-17. Breen, M. (1984). Process in syllabusdesign and classroomlanguage learning. In C. M. Brumfit(Ed.), GeneralEnglishsyllabusdesign(ELT Documents No. 118, pp. 47-60). Oxford: Pergamon Press. Breen, M. (1985). Authenticityin the language classroom.AppliedLinguistics, 6, 60-70. Chaudron, C. (1985). Intake: On models and methods for discoveringlearners' 7, 1-14. processingof input. Studiesin SecondLanguageAcquisition, Clahsen, H. (1984). The acquisitionof Germanwordorder: A testcase forcognitive approaches to L2 development.In R. Andersen (Ed.), Secondlanguages:A crosslinguistic approach(pp. 219-242). Rowley,MA: NewburyHouse. Corder, S. P. (1967). Thesignificance IRAL, 5, 161-169. oflearners'errors. Corder, S. P. (1981). Error analysisand remedial teaching.In S. P. Corder (Ed.), Erroranalysisand interlanguage (pp. 45-55). Oxford: Oxford UniversityPress. Eckman, F. (1985). Some theoretical and pedagogical implications of the markedness differentialhypothesis.Studiesin SecondLanguage Acquisition, 7, 289-308. Ellis,R. (1984). Can syntaxbe taught?A studyof the effectsof formalinstruction on the acquisition of WH questions by children.AppliedLinguistics, 5, 138155. Ellis, R. (1985). Sources of variabilityin interlanguage.AppliedLinguistics, 6, 118131. Ellis, R. (1989). Are classroom and naturalisticacquisitionthe same? A studyof theclassroomacquisitionof Germanwordorder rules.StudiesinSecondLanguage 11, 305-328. Acquisition, secondlanguageacquisition. Oxford: Blackwell. Ellis, R. (1990). Instructed Felix, S. (1981). The effectof formalinstructionon second language acquisition. Language Learning,31, 87-112. tasks:Negotiating Fotos,S. (1992). Grammar consciousness-raising meaningwhilefocussingonform.Unpublished doctoral dissertation,Temple UniversityJapan. Fotos, S., & Ellis, R. (1991). Communicatingabout grammar: A task-basedap25(4), 605-628. proach. TESOL Quarterly, Gattegno,C. (1972). Teaching foreignlanguagesinschools:Thesilentway.New York: Educational Solutions. 5, 79Gregg, K. (1984). Krashen's monitorand Occam's razor.AppliedLinguistics, 100. and language sciences Halliday, M., McIntosh,A., & Strevens,P. (1964). Thelinguistic teaching.London: Longman. THE STRUCTURAL SYLLABUS AND SLA

This content downloaded from 81.102.155.33 on Sat, 28 Sep 2013 12:16:33 PM All use subject to JSTOR Terms and Conditions

111

Harley,B. (1989). Functionalgrammarin Frenchimmersion:A classroomexperiment.AppliedLinguistics, 10, 331-359. Hatch, E. (1978a). Secondlanguageacquisition.Rowley,MA: NewburyHouse. Hatch, E. (1978b). Acquisitionof syntaxin a second language. In J. Richards(Ed.), secondandforeignlanguagelearning(pp. 401-435). Rowley,MA: Understanding Newbury House. Howatt,T. (1974). The background to course design. In P. Allen & S. P. Corder (Eds.), The Edinburghcoursein applied linguistics:Vol 3. Techniquesin applied linguistics (pp. 1-23). Oxford: Oxford UniversityPress. Huebner, T. (1983). A longitudinal analysisoftheacquisition ofEnglish.Ann Arbor, MI: Karoma. Hyltenstam,K., & Pienemann, M. (Eds.). (1985). Modellingand assessingsecond languageacquisition.Clevedon, England: MultilingualMatters. tosyllabus Krahnke,K. (1987). Approaches designforforeignlanguageteaching.Englewood Cliffs,NJ: Prentice Hall. Krashen, S. (1977). Some issues relatingto the monitor model. In H. Brown,' C. Yorio, & R. Crymes.(Eds.). On TESOL '77: Teachingand learningEnglishas a secondlanguage: Trendsin researchand practice(pp. 144-158). WashingtonDC: TESOL. and secondlanguagelearning.Oxford: Krashen,S. (1981). Secondlanguageacquisition Pergamon Press. Krashen, S. (1982). Principlesand practicein secondlanguageacquisition.Oxford: Pergamon Press. Larsen-Freeman,D. (1991). Teaching grammar.In M. Celce-Murcia(Ed.), Teaching language(pp. 279-295). Rowley,MA: NewburyHouse. Englishasasecondorforeign tosecondlanguageacquisiLarsen-Freeman,D., & Long, M. H.( 1991). An introduction tionresearch.London: Longman. Lightbown, P. (1985). Can language acquisition be altered by instruction.In K. Hyltenstam& M. Pienemann (Eds.), Modellingand assessingsecondlanguage acquisition(pp. 101-112). Clevedon, England: MultilingualMatters. Long, M. H. (1983). Does second language instructionmake a difference?A review of the research. TESOL Quarterly, 17(3), 359-382. Long, M. H. (1988). Instructedinterlanguagedevelopment. In L. Beebe (Ed.), Issuesin secondlanguageacquisition:Multipleperspectives (pp. 113-141). Rowley, MA: Newbury House. Long, M. H., & Crookes, G. (1991). Three approaches to task-based syllabus 26(1), 27-56. design. TESOL Quarterly, Mackey,W. (1965). Language teachinganalysis.London: Longman. McLaughlin,B. (1978). The monitormodel: Some methodologicalconsiderations. LanguageLearning,23, 309-332. 11, 113-128. McLaughlin, B. (1990). Restructuring.AppliedLinguistics, Meisel,J., Clahsen, H., & Pienemann, M. (1981). On determiningdevelopmental stagesin naturalsecond language acquisition.Studiesin SecondLanguageAcquisition,3, 109-135. Moore, A. (1989). [Review of Secondlanguagegrammar:Learningand teaching]. EnglishLanguage TeachingJournal,43, 155-157. Moulton,W. (1961). Linguisticsand language teachingin the United States,19401960. In C. Mohrman,A. Sommerfelt,& J. Whatmough(Eds.), Trendsin Euro1930-1960 (pp. 82-109). Utrecht,Netherlands: pean and Americanlinguistics, Spectrum. Palmer, H. (1917). The scientific studyand teachingoflanguages.London: Harrap. Pawley,A., & Syder,F. (1983). Two puzzles forlinguistictheory:Nativelikeselec112

TESOL QUARTERLY

This content downloaded from 81.102.155.33 on Sat, 28 Sep 2013 12:16:33 PM All use subject to JSTOR Terms and Conditions

tion and nativelikefluency.In J. Richards & R. Schmidt (Eds.), Language and communication (pp. 191-225). London: Longman. Pica, T. (1983). Adult acquisitionof Englishas a second language under different conditionsof exposure. LanguageLearning,33, 465-497. Pienemann, M. (1984). Psychologicalconstraintson the teachabilityof languages. Studiesin SecondLanguageAcquisition, 6, 186-214. Pienemann, M. (1985). Learnabilityand syllabusconstruction.In K. Hylstenstam & M. Pienemann (Eds.), Modellingand assessingsecond language aquisition (pp. 23-75). Clevedon, England: MultilingualMatters. Pienemann, M. (1989). Is language teachable? Psycholinguistic experimentsand 10, 52-79. hypotheses.AppliedLinguistics, Prabhu, N. S. (1987). Secondlanguagepedagogy.Oxford: Oxford UniversityPress. Rutherford,W. (1987). Secondlanguagegrammar:Learningand teaching.London: Longman. Ryle,G. (1949). The conceptofmind.London: Hutchinson. Schmidt,R. (1990). The role of consciousnessin second language learning.Applied 11, 17-46. Linguistics, Schmidt,R., & Frota,S. (1986). Developing basic conversationalabilityin a second language: A case studyof an adult learner. In R. Day (Ed.), Talkingto learn (pp. 237-326). Rowley,MA: NewburyHouse. Seliger, H. (1979). On the nature and functionof language rules in language 13(3), 359-369. teaching. TESOL Quarterly, Sharwood Smith, M. (1981). Consciousness-raisingand the second-language learner. AppliedLinguistics, 2, 159-169. Terrell, T. D. (1991). The role of grammarinstructionin a communicativeapproach. ModernLanguageJournal,75, 52-63. Ur, P. (1988). Grammar practiceactivities. Cambridge: Cambridge UniversityPress. Van Patten,B., & Cadierno, T. (1991). SLA as inputprocessing: A roleforinstruction. Unpublished manuscript. White, L. (1987). Against comprehensibleinput: The input hypothesisand the 8, 95-110. development of second language competence.AppliedLinguistics, Widdowson, H. (1968). The teaching of English through science. In J. Dakin, B. Tiffen,& H. Widdowson (Eds.), Language in education(pp. 115-175). London: Oxford UniversityPress. Oxford: Oxford UniversityPress. Wilkins,D. (1976). Notionalsyllabuses. viewoflanguageacquisition. Wode, H. (1980). Learninga secondlanguage:An integrated Tiibingen, Germany: Gunter Narr. Yalden, J. (1983). The communicative syllabus:Evolution,designand implementation. Oxford: Pergamon Press.

THE STRUCTURAL SYLLABUS AND SLA

This content downloaded from 81.102.155.33 on Sat, 28 Sep 2013 12:16:33 PM All use subject to JSTOR Terms and Conditions

113