.1-v I

"", is paper was begun during rhe 1968-69 aColdCl1l1l: yeilr while I was a visitor at the]),pt. of Applied Linguistil:li. University of EJinuurKh. MallY studelm ilnJ te.l.Lher~ at.;!dinburgh and at Washington, through their persistent L:alls for darity, have helped me.. «:rystallizc the ideas prt:senteJ in (his paper to whatcve:r Ie:vcl of dari(y is oltlaillcli here:­)1l.• Wilih to thank them und I especially wish to thank Ruth Clark, hed LukuH, frede­rick Newmeyer, and Paul Van Buren. An earlier version of this paper was l"e:olJ al the

~~·Stoond Inrc-mational Congress of Applied Linguistics, CambridKe University J Sept., I 'jbli.

•~:;Jntroductioll

'f:'-:', '. paper t discusses some theoretical preliminaries for res~an.:hers concerned··.·th the linguistic aspects Qf the psychology of second-language learning,

theoretical preliminaries arc important because without them it is vir­· ally impossible to decide what data are relevant to a psycholinguistic theory""second-language learning,' .. It is also important to distinguish between a teaching p~rsp~ctive and a

ing one. As regards the 'teaching' perspective, one miglH very wdl write. thodology paper which would relate desired outpot to known inputs in~rincipled way, prescribing what has to be clune by the teacher in order to

",p the learner achieve learning. As regard,.. the 'learning' perspective. one. ht very well write a paper describing the process of attempted learning of

nd language, successful or not: teaching, textbooks. and other 'external• would constitute one, but only onc) important set of relevant variables.

.. ~lCr Artikcl steJlt Idilr, dall wir Jie fur dell Pru ...d! Jl:S i':wciuprol.:h~rwcllJli n:lc\Oolutcnten Jureh dic verglei.:helldc UlHenu,hun~ vuu drl·i pruJuktivcn Spr""hliptclIll"n edl""l.

n. Diesc sinJ (I) dil" Muttl"rliprill:hc Jes LemClhll"ll, (1) Jie tremdsprol..:hlidu: l\ulHpctCll1.

I Lernendcn, seine Zwisehensprache (ITlterla"gu"'8f J, (3) dols System der t-'n:nuls~lr...,he.· Ab.chliel~end werden die Pru:tessc beschrieben. die: fur die Oiftc:rellz zwis..:he:n dc:r Zwi•

nsprache des Lerne:nde:n uuJ Jel J.IlKestre:Ute:n '-:n:mdsprache:nkompctenz verillltwonlich

Publis~cd "uart~~.ly .Erschclnt vle:ndJahrltch

.'Banking Accounn: ~.No. 02/HH8 Deuuche Bank Hciddbeq:'-fiNo. 19H936 Commenbank Heiddbcr,

'~1).61)00 HeiddberglGennany, HenutnSe 6" \ . 'P. O. Box 629. .-.}.Cable Address: Groos Heidelberg .,' >

OM ",­DM 15,-

fromJulius Groos VerlagP. O. Box 6296900 Heidelberg/Germany

fromOxford Univcnity PreIS.~rcu Kuad Neasden London N,W. 1011......

!'.or from its Ovcrsea Office. In:Toronto, Melboume, Wellington, 8ombIy, .Cillcutta, Madras, Karachi. Lahore, .._Dacca, Capc Town. Job.nnc.bur,. SaliJoi­bury. Ibadan. Apapa, Nairobi. Lusaka, , .J

Addis Ababa, Kuala Lumpur. Tokyo.Hong Kong. New York

( Briaio. Canada. Auuralia.Zealand. Greece, Spain. Gibraltar,

a, CypNSleA:It. Ethiopia. Gambia, Ghana, Kenya,.I, Malawi, Nigeria, Rhodesia,hia, Siern. Leone, Somalia, South·a, Sudan, Tanzania, Uganda\R and MIDDLE EAS'I':I, Iran. Jordan, Kuwail,non, Saudi·Arabia. Syria, Turkey,.cn\:lani.tan, Burma. Ceylon,g Kool> lodia. Indonesia,n, Malayda, Paki.lan.ippiDei. Sin:rporc. Southea, Tbailan

on;

f. Dr. BertH Malmberg,venity of Lund/Sweden

f. Dr. Gerha.rd Nickel.I/euirit StuugartlGcCln,.ny

Itbcr countries:

'LCiption pcr vol.;:Ie copia:

\'\) ,'\ \ .' \ ---I:.-..,. \n ~ r 1. .~\r IJ--(/( (O.d~(CJtW (/.{ I,j Ir", "-/ () 1_J , v (" 't.. j ... "r .vl-p ..J [../V1'1UI'/-!-'CS 10 c:r 0

o\L is obtainable m\ ldj- \ "i -) J ;;~.. A-p (J (tr( ~ INTERLANGUAGE J\ -;]2.0

----from---1 TEj!1111~llrl~llliiiil~11~i~if~nr Larry Selinker Hew :tt 00

\:e: Lib..irie Marcel Didi« 3 5141 004565143

.c et 6 rue de la Sotbonnc. Pari. V r .'article demontre qUI: nous tl"OUvons les fa~teun pcrtinents Ju prOl.:CliSUli J.~ 1'.lppn:ntisSolt;e

la deuxiemc langue grace a un exalllen (:omp.ue J~ troIS SYSlCIUCS Iinguilitiquesucc-ifs, lesquds liont (I) la langUl: materndle de 1'~leve. (l) I... cQmpetcno: de I'dev.:

'nl la deuxieme langue, sa langue inlcrmeJiaire (lrllt:djJnguoIgeJ, et (.)) Ie s)'stelllC: de

langue cible. Finalemem les processus som dinits qui sonl responsablcs de I. diHereur.;.(

· tre la langue intermediaire de 1'~lhe et sa compett:n~e vuulue dans la 1.U1l:\ue ..:Iule_

19I>l

[ius Groos Verlag

210 IRAJ., VOl.. XI}, AUGUST 1972

In distinguishing between the two perspectives,' claims about the internalstructures .and processes of the le~rning organism take on a very secondarycharacter In the teaching perspective; such claims may not even be desirable.here. But s~ch claims do provide the raison dJetre for viewing second~lan~gua~ learnmg from the learning perspective. This paper is written from thelearn~ng perspective, regardless of one's failure or success in the attempted ,Jeamll1g of a second language. 0.:'

In the learning perspective, what would constitute the psychologically_'::relevant data of second-language learning?-' My own position is that such:~data would be. those behavioral events which would lead to an understanding' t?f the psychohngu.~tlc structures and processes underlying 'attempted mean-';~gful,pe~fo~mance 111 a second languagf. The term 'meaningful performance "Situation Will b~ used he~e to refer to the situation where an 'adult'"' attempts~o express meal1lngs, whIch he may already have, in a language which he is111 the pro~ess of lear~i.ng. Since performance of drills in a second-languageclass",?m IS, by defm.t.on, not meaningful performance, it follows that froma learning perspective, such performance is, in the long run, of minor interest.Also, behavior which lH..'.curs in experiments using nonsense syllables fits intothe same ca~egory. and. for the same reason. Thus, data resulting from these j'

latter ~ha~.oral SItuations are of doubtful relevancy to meaningful perform­an~ SltuattOns, and thus to a theory of second-language learning.

It has long seemed to me that one of our greatest difficulties in establishinga psychology of second-language learning which is relevant to the way'peopleactually learn ,"cond languages, has been our inability to identify unambi.guously the phenomena we wish to study. Out of the great conglomeration

2 I.t is ~o~ unfair to say thai almost all of the vast literature attempting to rdate psycho­!mguntlcs t,o srcond-Iang~agr learning, whether produced by linguists or psychologisb';IS characterized by confUSion between 'Iurning' a second language and 'teaching' a secondlanguage. (SeC' ,also ~ackC'y in Jakohovits, 1970, p, IX.). This confusion applies as welt:0 ,Imost all dl~cusslons on the topic onC' hC'ars. For example, one might hear the termpsychology l?f sC'conJ.langua~e teachill~' and not know whether the speaker is referring to'what ~hC' teacher should do, what the learner should do. or both. This terminologicalconfUSIOn makrs one rrogularly uncrrtain as to what is being claimed. I

3 TIlt answer to this qurstion is nul nhvivus since it is well known that theoretical consid.:rations help point the ~ay to relC'van~ data. See, for example, Fodor (1968, p, 48):... how we count ~havlOrs and what IS :available as a description depend, in part on

,what conceptual equipment our theorirs provide ..." . :~ ..

4~Adult' is dtfined as bC'ing ovtr t1~e age of 12. This notion is derived from Lennebrr"::i;.'(1967, e.g. pp. 156, 176) who claims that after the onset of puberty. it is difficult to' "':,.~ma~ter thr pronunciation of a second language since a "critica'" period in brain matu- I:;;

..... ~·ratlOn has been passed, and ". .. language devel~pment tends to 'freeze'" (ibid, 156). f:-

INTERLANGUAGE 211

of second-language behavioral events, what criteria and constructs should beused to establish the class of those events which are to count as relevant intheory construction? One set of these behavioral events which has elicitedconsiderable interest is the regular reappearance in second-language perfonn­ance of linguistic phenomena which were thought to be eradicated in theperformance of the learner. A correct understanding of this phenomenonleads to the postulation of certain theoretical constructs, many of which havebeen set up to deal with other problems in the field. But they also help clarifythe phenomenon under discussion. These constructs, in turn, give us a frame­work within which we can begin to isolate the psychologically-relevant dataof second-language learning. The new perspective which an examination ofthis phenomenon gives us is thus very helpful both in an identification of

. relevant data and in the formulation of a psycholinguistic theory of second­'language learning. The main motivation for this paper is the belief that itis particularly in this area that progress can be made at this time.

2. '/nterlanguage' and latent structures

Relevant behavioral events in a psychology of second-language learningshould be made identifiable with the aid of theoretical constructs whichassume the major features of the psychological structure of an adult when­

'ever he attempts to understand second-language sentences or to produce them.,. If, in a psychology of second-language learning, our goal is explanation of

some important aspects of this psychological structure, then it seems to methat we are concerned in large part with how bilinguals make what Wein­reich (1953, p. 7) has called 'interli!,gual identifications'. In his book Lan­guages in Contact, Weinreich discusses - though briefly - the practical needfor assuming in studies of bilingualism that such identifications as that of aphoneme in two languages, or that of a grammatical relationship in two lan­guages, or that of a semantic feature in two languages, have been made bythe individual in question in a language contact situation. Although Wein­

. reich takes up many linguistic and some psychological questions, he leavescompletely open questions regarding the psychological structure within which

" we assume 'interlingual identifications' existi we assume that there is such apsychological structure and that it is latent in the brain. activated when oneattempts to learn a second language.

The closest thing in the literature to the concept latent psychological struc­ture is the concept of latent language structure (Lenneberg, 1967, especially

. pp. 374-379) which, according to Lenneberg, (a) is an already formulatedarrangement in the brain, (b) is the biological counterpart to universal gram-

212 IRAL, VOL. XI). AUGUST 1972

':~- mlar, and (c) is transfonned by the infant into the realiud strHctHrt of a par.ticoiar grammar in accordance with certain maturational stages. For the pu....Poses Of this paper, I will assume the existence of the latent language structuredescribed by Lenneberg; I shall further assume that there exi.ts in the brainan already formulated arrangement which for most people is different fromand exists in addition to Lenneberg's latent language structure. It is important

.... to Ilal< that with the latent structure described in this paper as compared to. Lenneberg's, the~ is no genetic time table;' there is no direct counterpart t6

~ &111 grammatical concept such as 'universal grammar'; there is no guaranteethat thil latent structure will be activated at all; there is no guarantee that

... the lal<nt structure will be 'realized' into the actual structure of any natural.' . language (i.e. there is no guarant.. that attempted learning will prove suc­-., oessful), and there is every possibility that an overlapping exists between t~is

'.latent language acquisition structure and other intellectual structure.. ". The crucial a~!lillmption we arc making here is that those adults who 'sue.reed' in learnin~ a ..cond language so that they achieve native-speaker"compc~nce' have somehow rea(tivateJ the latent language structure whichLenneberg describes. This absolute success in a second language affects, as

.. we know from observation, a small percentage of learners - perhaps a .'. . mere 5 %. It follows from this a'mmption that this 5 % go through verydiffe... >". :~~

, ent psycholinguistic processes than do most second-language learners and that,bf''. these successful learners may be safely ignored - in a counterfactual sense'~ ,.

for the purposes of establishing the constructs which point to the psycltolog­ically-rdevant data pertinent to most second-language learners. Regardingthe study of the latter group of learners (i. e. the vast majority of second-

.,language learners who fail to achieve native-speaker competence), the notion .".of 'att<mpted learning' is independent of and logically prior to the notion of t':~.:successful learning'. In this paper, we will focus on attempted learning by"this group of learnets, successful or not, and will assume that they activate'II,~

different, though still genetically determined structure (referred to here :u.,r.the latent psychological structure) wheoever they attempt to produce a sen- FJ'tence in the second-Ian~uage, that i, whenever they attempt to express mean" ',~x

. ings, which they may already have, in a language which they are in the ,"process of learnin~. .~ .~ ·~t

"This series of 35!illmptiol1!i must be made, I think, ~cause the second-lantJ;l,t .. guage learner who actually achieves native-speaker competence cannot pos'o .."

5 Fin' pointtd out by HarnlcJ Edward~.

6 5ft Lawler and Stlinktr (forthcoming) .... htre tht rtlevance of counterfactuals to'a theoryof terond.language Irarnin~ i~ taktn up. I!.

1NTERLANGUAGE 213

sibly have been taught this competence, since linguists are daily - in almostevery generative study - discovering new and fundamental facts about par­ticular languages. Successful learners, in order to achieve this native-speakercompetence, must have acquired these facts (and most probably importantprinciples of language organization) without having explicitly been taughtthem.'

Regarding the ideal second-language learner who will not 'succeed' (inthe absolute sense described above) and who is thus representative of the "astmajority of second-language learners, we can idealize that from the beginningof his study of a second language, he has his attention focused upon one normof the language whose sentences he is attempting to produce. With this state­ment, we have idealized the picture we wish to sketch in the following ways:•the generally accepted notion 'target language' (TL), i.e. the second-languagethe learner is attempting to learn, is here restricted to mean that there isonly one nOrm of one dialect within the interlingual focus of attention of thelearner. Furthermore, we focus our analytical attention upon the only ob­servable' data to which We' can rflalr theoretical prt'dictions:9 the utterances

7 Chomsky (1969, p. 68) txpressts a Vtry similar vitw:", .. it must be recognized that one d~5 not Itarn the grammatical structure of a secondlanguage through 'explanation and instruction', beyond the molt elementary rudiments,for the simple reason that no one has enough uplicit knowledge about this uructurc toprovide uplanation and instruction."Chomsky gives as a detailed example a property ....hich is durly central to grammar:that of nominalizuion (Chomsky. 1969, pp. 68 and 52-60). I ~e no point in repeatingChomsky'. detailed arguments which dearly show that a successful learner of English asa stcond language could not have learned to make the judgments Chomsky describesthrough 'explanation and instruction'.

8 We have also idealiud out of our comideration differences between individual ltarners,which makes this framework quite incomplete. A theory of Itcond-Ianguage ltarning thatdoes not provide a central place for individual differences among learners C4nnot beconsidered acceptable. See Lawler and Selinker (forthcoming) for a discussion of thistricky question in'terms of profiles of idealized learners who differ one from the otherwith respect to types of linguistic rules and types of meaningful performance in a secondlanguage.

9 There has been a great dul of misul\drrstalldin~(J'f'rsonaJ communication) of this point.I am not taking an antimentalist position herr. Neither am I ruling out on an a.prioribasis perceptual studies in a second language. However, the reader should be aware thatin addition to the usual problems with dttermining whether a subject perceives or under~

'stands an utterance, the analyst in the interlingual domain cannot rtly on intuitive gram­matical judgments since he will gain information about another system. the one thelearner is struggling with, i.e. the TL. (For a similar methodological problem in anotherdomain. see Labov. J969, p. 71 S). Another, and perhaps the most important, argumentagainst perceptual interlingual studies is that predictions based upon them are not test-

214 IRAL, VOL. XI). AUGUST 1971

.which are produced when the learner attempts to say sentences of a TL. Thisret,of utterances for most learners of a second language is not identical tothe .hypothesized corresponding set of utterances which would have beenproduced by a native speaker of the TL had he attempted to express the samemeaning a. the learner. Since we can observe that these two sets of utterancesare not identical, then in the making of constructs relevant to a theory ofsecond-language learning, one would be completely justified in hypothesizing,perhaps even compelled to hypothesize, the existence of a separate linguisticI)'Stem'· based on the observable output which results from a learner's at­tempted production of a TL norm. This linguistic system we will call 'inter­language' (IL)." One of the main points of this paper is the assumption thatpredictions of behavioral events in a theory of second-language learning.hould be primarily concerned with the linguistic shapes of the utterancesproduced in IL•. Successful predictions of such behavioral events in meaning­ful performance situations will add credence to the theoretical constructsrelated to the latent psychological structure discussed in this paper.

It follows from the above that the only observable data from meaning­ful performance situations we can establish as relevant to interlingual identi­fications are: (1) utterances in the learner's native language (NL) producedby the learner; (2) IL utterances produced by the learner; and (3) TL utter­ances produced by native speakers of that TL. The.e three sets of utterance.

. or behavioral events are, then, in this framework, the psychologically-rele­vant data of second-language learning, and theoretical predictions in a rele­vant psychology of second-language learning will be the surface structuresof IL sentences.

By setting up these three sets of utterances within one theoretical frame­work, and by gathering as data utterances related to specific linguistic struc­tures in each of these three systems, (under the same experimental conditions,if possible) the investigator in the psychology of second-language learning canbegin to study the psychuling\listic prucesses which establish the knowledgewhich underlies IL behavior. I would like to suggest that there are five central

able in 'meaningful performant:(' ~jlua(j()ns' (see definition above); a reconstruction of. the nenl upon the part of the Ie-arner would have to bt- made in a perceptual inter­'lingualnudy. Such difficulties do not exist when predictions are related to thc shape ofutterances produced as the result uf the learner attempting to exprcss in the TL meaningswhich he may alrtady havC'.

10 -Notions of such separate' lin~uistil' ~y~te'm~ have' been dC'vC'lo~d independently by Jako~

bo.iu (1969) and N,m«, (1971).I~I_ ~ notion 'interlanguagc' is intrudurec1 in SelinkC'r (1969).

INTERLANGUAGE 21S

processes (and perhaps some additional minor ones), and that they exist in. the latent psychological structure referred to above. I consider the followingto be processes central to second-language learning: first, language transfer;~econd, transfer-of-training; third, strategies of second-language learning;fourth, strategies ofsecond-language communication; and fifth, overgenerali­zation of TL linguistic material. Each of the analyst's predictions as to theshape of IL utterances should be associated with one or more of these, orother, processes.

3. Fossilization

Before briefly describing these psycholinguistic processes, another notion Iwish to introduce for the reader's consideration is the concept of fossilization,a mechanism which is assumed also to exist in the latent psychological struc­ture described above. Fossilizable linguistic phenomena are linguistic items,rules, and subsystems which speakers of a particular NL will tend to keepin their IL relative to a particular TL, I1U matter what the age of the learneror amount of explanation and instruction he receives in the TL.l1 I have inmind such fo~silizable structures as the well-known 'errors': French uvularIrl in their English IL, American English retroflex Irl in their FrenchIL, English rhythm in the IL relative to Spanish, German Time-Place orderafter the verb in the English IL of German speakers, and so on. I also havein mind less well known 'non-errors' such as Spanish monophthong vowelsin the IL of Spanish speakers relative to Hebrew, and Hebrew Object-Timesurface order .fter the verb in the IL of Hebrew speakers relative to English.Finally, there are fossilizable structures that are much harder to classify suchas some features of the Thai tone ..ystem in the IL of Thai speakers relativeto English. It is important to note that fossilizablc structures tend to remainas potential performance, reemergi ng1:l in the productive performance ofan IL even when seemingly eradicated. Many of these phenomena reappearin IL performance when the learner's attention is focused upon new anddifficult intellectual subject malter or when he is in a state of anxiety orother excitement, and strangely enough, sometimes when he is in a state ofextreme relaxation. Note that the claim is made here that, whatever thecalise, the well-observed phenomenon of 'backsliding' by second-language

12 Gillian Brown has pointed out (penonal communication) that we should worke heretowards a dynamic model where fossilization would be defined relative to various, pcr~

haps arbitrary, chronological a~C'Kroups.

13 John Lavcr has helped me to clarify this point.

216 IRAL, VOL. XI}, AUGUST 1972

leatMn from a TI. norm is not, as has been generally believed, either randomor.toward the speaker's NL, but toward an IL norm,"

.' A crucial fact, perhaps the most crucial fact, which any adequate theoryof second-language learning will have to explain is this regular reappearanceor=mergl'nce in IL productive performance of linguistic structures which~ thought to be eradicated. This behavioral reappearance is what has ledme to postulate the reality of fossilization and ILs, It should be made clearthat the reappearance of such behavior is not limited to the phonetic level.Porexample, some of the subtlest input information that a learner of a secondlanguage has to ma'itn rcg-arc1s subcategori1.ation notions of verbal comple­""",tation. Indian English as an IL with regard to English" seems to fossilizethe 'that complement' or V thaI construction for all verbs that take sententialcomplements, Even when the correct form'has been learned by the Indian.peaker of English, this type of knowledge is the first he seem. to lose whenbi. attention is diverled to new intellectual subject matter or when he hasnot spoken the TL for even a short time. Under conditions such as these, thereis a regular ~appearam..e of the 'that complement' in IL performance for allsentential complements.

4. Frot Ctntral ProcesJe5

It· i. my contention that the most interesting phenomena in IL performanceare:those item., rules, and subsystems which are fossilizable in terms of thefi~ processes listed above. If it can be experimentally demonstrated thatfosliliuble items, rules, and subsvstems which occur in IL performance area result of the NL, then we are d;aling with the process of language transfer;if these fossilizable item" rules, and subsystems are a result of identifiableit~ jn training procedure!!, then we are dealing with the process known asthe transfer-of-trainirlg; if they are a result of an identifiable approach bythe learner to the material to be learned, then we are dealing with strategies

14 St~raJ ptopl~ have ~ljntC'd (lui (~rs(lnal cummunication) that. in this paragraph, there".pptars to ~ a cunnC'l'rion soldy ~twC'C'n fossilization and errors. This connection i~

. not intended sinct if lurm oul thai 'corrr-ct' Ihin~~ can also ree-merge- whe-n thought to~.eradic.ttd.C'lptcial/)'if Iheo)' arC' caused hy proc('ues other than language tran.fcr.

S -Ktith Brown (J'f'tlnn:tl communiC'l\fion) has arKurd that the sociolinguistic status of the."languages' or 'dillle'cts' callC'd Indian English, Filipino Englillh. Wesl African English,Welt AfriC2n F~n(h. and m on. places them in a different category from that of theIL .ituatktn which I havt httn tftsaibillK. From th~ sociolinguistic point of virw this

'.tiunxnt might be jUHified. hUI I am t:ont:e-rne-d in ,his pa~r with a psychological per­IpKtive and the rde-vanl idulizat;ons st'em to me II) be identical in all of these! cases.

INTERLANGUAGE 217

of second-language learning; if they are a result of an identifiable approachby the learner to communication with native speakers of the TL, then weare dealing with strategies of second-language communication; and, finally,if they are a result of a dear overgeneralization of TL rules and semanticfeatures, then we are dealing with the overgeneralization of TL linguisticmaterial. I would like to hypothesize that these five processes are processeswhich are central to second-language learning, and that each process forcesfossilizable material upon surface IL utterances, controlling to a very largeextent tht' surface structures of tl1esc.' utterances.

Combinations of these processes produce what we might term entirelyfossilized IL competences. Coulter (1968) presents convincing data to demon­strate not only language transfer but also a srratl'gy of communication com­

"''- mon to many second-language learners. This strategy of communication dic~

. . 'tates to them, internally as it were, that they know enough of the TL in orderto communicate. And they stop learning.'· Whether they stop learning en­tirely or go on to learn in a minor way, e,g. adding vocabulary as experiencedemands [Jain (1969 insists they must] is, it seems to me, a moot point. Ifthese individuals do not also learn the syntactic information that goes withlexical items, then adding a few new lexical items, say on space traveJ, is, Iwould argue, of little consequence, The important thing to note with regardto the evidence presented in Coulter (1968) and Jain (1969) is that nOt onlycan entire IL competences be fossilized in individual learners performing intheir own interlingual situation," but also in whole groups of individuals,resulting in the emergence of a new dialect (here Indian English), wherefossilized IL competences may be the_normal situation.

We will now provide examples of these processes. The examples presentedin 'ection 3 are almost certainly the result of the process of language transfer.A few examples relating to the other processes should suffice for this paper.

4.1 Overgcneralization ofTL rules is a phenomenon well-known to languageteachers. Speakers of many languages could produce a sentence of the follow­ing kind in their English IL:

\

16 To describe- this lituation. Jain (1969) speaks of functional comptUnCt, Corder {I 967)using the te-rm transitional comptUnct focuses on the- provisional aspect of developing'competen~' in a second language. Both these notions owe- tht'ir uistence- in the firstplace, to Chomsky's (1965) notion of linguistic com~tence which is to be distinguishedfrom actual linguistic ~rformance.

17 An 'interJin~ual situation' is defined as a specific combination of NL. n, and JL

218 IRAl. VOl.. XiJ, AUC;UST 19n

(I) What did he intended to say?"

where the past ten~e morpheme -ed is extended to an environment in which,to the learner, it could IOKically apply, hut just docs not. The Indian speakerof English whu proJuces the collocation drive a bicycle in his IL performance,

:u in (2):

(2) After thinking little I decided to start on the bicycle as slowly as I

could as it was not possible to drive fast.

iJ most probably overgeneralizing the use of drive to all vehicles aain, 1969,pp. 22 Ilc 24; but see footnote 26 here). Most learners of English quickly learnthe English rule of contraction which forms things like the concert's fromtht conelr/ is. but then these learners max overgeneralize this rule to produce

sentences like:

(3) Max is happier than Sam's these days.

in their English IL. Though this sentence is hypothetical, it illustrates anearlier point. The learner of English who produces contractions correctly inaU environments must have learned the following constraint without -ex"planation and instnu:tion'. since this constraint was discovered only recently:"contraction of auxiliaries . .. cannot occur when a constituent immediatelyfollowing the auxiliary to be contracted has been deleted," e.g. 'happy' in(3) (Lakoff, in press). Dozens of examples of overgeneralization of TL rules

are provided in Richards (1970).

".2 The transfer-ol-training is a process which is quite different from lan­guage transfer (see Selinker, 1969) and from overgeneralization of TL rules.It underlies the source of a difficulty which Serbo-Croatian speakers at alllovels of English proficiency regularly have with the he/she distinction,producing in their En~lish IL he on almost every occasion wherever he.orshe would be called for according to any norm of English. There is no lan­guage transfer dfect hen,.' sjll«~. with reganJ to animateness, the distinctionbetween he and sbe is the same in Serbo-Croation as it is in English. '9 Accord..ing to a standard contrastive analysis then there should be no trouble. Itseems to be the case that the resultant IL form, in the first instance, is duedirectly to the transfer-of-training; textbooks and teachers in this interlingualsituation almost always present drills with he and never with she. The extent

IS This ICnt~nl'l,' And sentt'lll"t'S likt' it wert 111 f;t(l produced consistently by a middle-aged

IsrU'li who was v~ry Ouenl in En/;:lidL. .Jt I am indept~ 10 Wayles Browne (pasona Immmunication) fUI" clarification of this point.

INTERLANGUAGE 219

. of this fossilization can be seen with respect to speakers of this IL over theage of J 8. who even though they are const:iously aware of the distinction and

, of their recurrent error. in fact. regularly produce he for both he and she.stating that they feel they do not necd to make this distincfion in order tocommunicate. 20 In this case, theil, the fossilizable error is due originally toa type of transfer-of-training and later to a particular strategy of second­language communication,

4.3 Concerning the notion 'strategy' little is known in psychology about whatconstitutes a strategy; and a viable definition of it does not seem possible atpresent. Even less is known about strategies which learners of a second lan­guage use in their attempt to master a TL and express meanings in it. It hasbeen pointed out21 that learner strategies are probably culture-bound to someextent. For example, in many traditional cultures, chanting is used as a learn­ing device, clearly relating to what is learncd in these situations. Crucially,it has been argued" that strategies for handling TL material evolve when­ever the learner realizes, either consciously or subconsciously, that he has nolinguistic competence with regard to some aspect of the TL. It cannot bedoubted that various internal strategies" on the part of the second-languagelearner affect to a large extent the surface structurc..'!i of sentences underlyingIL utterances. But exactly what these strategies might be and how they mightwork is at present pure conjecture. Thus, one can only roughly attribute thesource of the examples presented herein to one or another strategy.

One example of a strategy ofsecond-language learning that is widespreadin many interlingual situations is a tendency on the part of learners to reducethe TL to a simpler system. According to Jain (1969, pp. 3 & 4), the resultsof this strategy are manifested at all levels of syntax in the IL of Indianspeakers of English. For example, if the learner has adopted the strategy thatall verbs are either transitive or intransitive, he may produce IL forms suchas:

4) 1 am feeling thirsty.or

(5) Don't worry, I'm hearing him.

and in producing them seems to have adopted the further strategy that therealization of the category 'aspect' in its progressive form on the surface isalw'ays with -ing marking (for further discussion, see Jain, 1969, p. Hf.).

20 Repllrh'd hy GtorJ.:t' Ml:Cready (p~nonall'omnHllljl:;lti(ltl)"

21 Ian Pear~on (personal communication).22 Elaine Tarone (ptnon,,1 communicOllinn).23 That is. ",hal Cordt'r refers to a~ the learner's "built-in syllabus" (Corder, 1967).

--,"

220 IRAL. VOL XI). AUGUST 1972

Coulter (1968) reports systematic errors occurring in the English IL per­fonnance of two elderly Russian speakers of English, due to another strategywhich seems also to be widespread in many interlingual situations: a tendencyon the part of second-language learners to avoid grammatical formatives suchat articles (6), plural forms (7), and past tense forms (8):

(6) It was'" nice, nice trailer, '" big one. (Coulter, 1968, p. 22)(7) I have many hundred carpenter my own. (ibid, p. 29)(8) I was in Frankfort when I fill application. (ibid, p. 36)

This tendency could he the result of a learning straltgy of simplification, butCoulter (1968, p. 7 fl.) attrihutes it to a communicatiOtl Slraltgy due to thepast experience of the speaker which h~s shown him that if he thinks aboutgrammatical processes while attempting to express in English meanings whichhe already has, then his speech will be hesitant and disconnected, leadingnative speakers to be impatient with him. Also, Coulter claims that this stra-

. ugy of stcond-language communication seemed to dictate to these speakenthat a form such as the English plural "was not necessary for the kind ofcommunicating they used" (ibid, p. 30).

Not all of these strategies, it musl be pointed out, are conscious. A sub­conscious strategy of second-language learning called Hcue-copying" has beenexperimented with hy Crothers and Suppes (1967, p. 211) on Americans learn­ing Russian morphological con~epts. This "copy the cue" strategy is mostprobably due tu what they call "probability matching", where the chancethat the learner will solect an alternative morphological ending related tothe cue noun is not random. Crothers and Suppes do not provide examplesof the result uf thi, 'trate~y in meaningful performance situations; an examplewould be the r at till' end of words like California and saw which foreignstudents of English who have had teachers from the Boston area regularlyreproduce in their Engli,h It.

4.4 To cunclude thi, 'eetion, it ,hould be pointed out that beyond the five50-called central processes, there exist many other processes which accountto some degree for th(' surface form of IL utterances. One might mentionspelling pronunciations. e.g. speakers of many languages pronounce final-tr on English words as I ~'l plus some form of r; cognate pronunciation, e.g~

English athlete pronounced as [atlit] by many Frenchmen whether or notthey can produCt· Ifll in other English words;" holophrase learning Gain,19(9), e. g. lor h,dr,lII-hollr the Indian learner of English may produce

24 Eumpl~ frtlm Tum Huckin (pt'f$onal communication).

INTERLANGUAGE 221

one ~nd half-an-hour; hypercorrection, e.g. the Israeli who in attempting to

get tid of his uvular fricative for English retroflex (r] produces (w ] befort.front vowels, "a vocalization too far forward';15 and most assuredly otherssuch as long exposure to signs and headlines which according to Jain (1969)affect by themselves the shape of English IL utterances of Indians, or at leastreinforce more important processes such as language transfer.

5. Problems with this perspective

There are certainly many questions one might wish to ask regarding the per­spective presented so far in this paper; I ,hall attempt to deal with five(5.1-5.5). The reader should bear in mind that we are here calling for ti,ediscovery, description and experimental testing of fossilizable items, ruk-sand subsystems in interlanguages and the relating of these to the above­mentioned processes - especially to the central ones. What seems to be most~rom.isi~g for study is the observation concerning fossilization. Many ILImgUlstlC structures are never really eradicated for most second-languagelearners; manifestations of these structures regularly reappear in IL productiveperformance, especially under conditions of anxiety, shifting attention, andsecond-language performance on subject matter which is new to the learner.It is this observation which allows us to claim that these psycholinguisti ..:structures, even when seemingly eradicated, are still somehow present in thl'brain, stored by a fossilization mechanism (primarily through one of the'"five processes) in an IL. We further hypothesize that interlingual identifica­tions uniting the three li.nguistic systems (NL, IL, and TL) psychologically,are activated in a latent psychological structure whenever an individualattempts to produce TL senteIOce,.

5.1 The first problem we wish to deal with is: can we always unambiguousl"identify which of these processes our observable data is to be attributable to',Most probably nol. It has been frequently pointed out (personal communi­cation) that this situation is quite common in psychology. In studies on mem­ory. for example, one often does not know whether one is in fact studying"storage' or "retrieval'. In our case, we may not know whether a particularconstituent IL concatenation is a result of language transfer or of transfer­of-training or, perhaps, of both. 26 But this limitation need not deter us, even

2S Example- from Briana Statt'man (personal ~mnmullj(ali()n).

26 Th~ drivt 4 bicyclr t'xample Rivell in u·c,ion 4 may. in fact, fit this situation (Ji~e Jain.1969, p. 24).

222 IRA!.. VOl.. XI). AU<;UST 1_71

if we cannot always sort things out absolutely. By applying the constructssuggeoted in this paper, I believe that relevant data can be found in the very-many 5e'cond-Janguage-learning situations around us.

5.2 The second problem i:'\: how call we systematize the notion fossilizationso that from the basi~ of theoretical constructs, we can predict which itemsin which interlingual situations will be fossilized? To illustrate the difficultyof attempting to answer this question, note in the following example thenon-rnersibility of fossilization effects for no apparent reason. Accordingto a contrastive analysis, Spanish speakers should have no difficulty withthe he/she distinction in English, nor should English speakers have anydifficulty with the corresponding distinction in Spanish. The facts are quitedifferent, however: Spanish speakers do, indeed, regularly have trouble withthis distinction, while the revene does not ~eem to occur with English learnenof Spanish." Unlike the Serbo-Croatian example mentioned above, in thilcase there is no clear-cut explanation why Spanish speakers have trouble andEnglish speakers do not. In cases such as these, it may turn out that one pro­cess, e.g. language transfer or transfer-of-training, overrides other considera~

tions, but the stating of the governing conditions may prove very difficultindeed.

In principle, one feels forced to agree with Stephanie Harries (personalCommunication) who claims that until a theory of second-language learning'can ~nswer questions like: "How do I recognize fossilizable structures inadvance?" or "Why do some things fossilize and others do not?", all experi­ments conducted within the framework provided in this paper must be re­garded as lexploratory' in nature. (To put things in more familiar jargon: withregard. to fossilization, our results are 'descriptive' and not 'explanatory' innatu"".) But this task of prediction may prove to be impossible; certainly asP""d Lukoff points out (personal communication) this task, on the face ofit, may be even tougher than trying to predict errors in second-languageptrformance - a task notably lat.:kin~ in success. .

The major justification one has for writing about the construct 'fossili­zation' at this stage of knowledge is that descriptive knowledge about ILs

, 'which turns out to suggest predictions verifiable in meaningful performancelituations, leads the way to a systematic collection of the relevant data; thiltask, one which is impossible without this construct, is expected to be relevant

,'in serious theory construction in a psychology of second-language learning.

~2"" Eumplt' from Sol Sarnru (rt'r~llIlal communication),

INTERLANGUAGE 223

5.3 The third problem to be treated here concerns the apparent difficulty offitting the following type of question into the idealized domain I have beensketching: how does a second-language-learning novice become able to pro­duce IL utterances whose surface constituents are correct, i. e. 'correct" withrespect to the TL whose norm he is attempting to produce? This questionfinally brings us face-to-face with the notion of 'success' in absolute terms:productive performance in the TL by the second-language learner which isidentical to that produced by the native speaker of that TL.'· We noted thisin section 2 so as to exclude from our idealized domain of inquiry those learn­ers of second languages who reactivate 2Q the latent language structure thatis realized into a native language. In this paper, we are concentrating onattempted learning of a second language, unsuccessful in this absolute sense.Of course, 'success' in second-language learning need not be defined so ab­solutely. The teacher or the learner can be satisfied with the learner's achiev­ing what hal been called 'communicative competence' (see, for example,Jakobovits. 1970, or Hymes, in press). But this is not the issue here. As was

28 A! wa~ pointed out in footnote 7, Chomsky (1969, p. 68) also adds the ability to providenative-speaker-like grammaticality judgments.

29 Note that this reactivation may be the only explanation possible for an individual wholearns any part of a second language wdl. In this light, Cheryl Goodenough (personalcommunication) has objectt'd to tht' qualitativt' split bt'tween the 5 % who succt'ed and therest of all second· language lurllt'rs. Since in this paPf'r we art' not concentrating on successin a second language. as one would in the teaching approach, but on tht' attempt to iso­late the latent psychological structure which determines, for any learner, the systemunderlying attempted production of a TL norm where the total t'fft'ct of this output ,isclearly non-identity to the hypothesiud Tl norm, then resolotion of this issue shouldnot affect the discussion, The importance {lfisolating this 5 % is tht' speculation that thest'individuals may not go through an II..Reibtl (1969) stresses the role of tht' latent lan~uage structure in second-language learningby ~uggesting that it is only when sec(lnd-Iall~lla~t' It'arnt'u do the wrong things th,u lht'ydo nut "succt'ed," i.e. "we s('ek to f'xplain dif!'t'renct's betwt't'n adult learnt'rs, not interms of difft'rences in the innate learnin~ abilities, but rather in terms of tht' way inwhich they are applied." (p. 8). Kline (1970) attt'mpu to providt' a point of contact bt'­tween Reibel's views and mine by suggesting that any rt'or~anintion of an Il to idt'ntit)·with a TL must use the kinds of capacities and abilities Rt'ibel dt'scribes.A different opposinl': vit'w to the penpectivt' of this papt'r has been prt'st'nted by SandraHamlett and Michael Seitz (pt'uonal communication) who have argued that, even forthe vast majority of second·language learners, there is no already formulatt'd arrange­ment existing in the brain, but that the latt'llt psychological structure alluded to hert' isdevdoped, partly at least, by strategies which ,hangt' up to tht' age of 12 and r~'ll1ain

with an individual for the rest of his lift'. Tht're st't'ms to be at present no critical eOlpiril:altest for deciding bt'tween the~e two altern;uives.

224 IRAL. VOL. XIJ. AUGUST 1972

point<d out in section I, tho emphasis upon what the t<acher has to do inordor to help the learner ad,iev< successful learning belongs to the 'teaching'ponpoctive, which i, not the perspectiv< of this paptr. Porhaps tho rath..curious confusion in the literature of 'learning a second language' with 'teach·ing a second language' (see footnote 2) can be <xplained by tho failure to soca psychology nf 'econd-Ianguage learning in terms oth<r than thos< rdat<dto '.uccess'. For example, typical learning-theory exp..iments when done intho domain of second-language learning would d<mand knowledge of whoretho l<arner will tend to end up, not where we would like him to end up,Exporiments of this type would also demand knowledge of wher< tho s<cond­language lea rnt' r begins. We would claim that prerequisite to both th"etypo. of knowledge are detailed descriptions of ILs - description' not pre·~ntly available (0 us. Thus, such experiments at present are premature, withthe results bound (0 prove conrusin~.

Specifically concerning the problem raised in the first sentence of 5.3, it'eem, to me that this question, though relevant to the psychology of second­language learning. i, one that should also not be asked for the present sinceits asking depend, upon our under'tanding clearly the psychological extentof interlingual identifications. For example, before we can discover how sur­fa~ constituents in all IL get reorganized to identity with the TL, we musthave a c1..r idea of what i, in that IL, even if we cannot explain why iti'thore, In Selinker (1969) I believe I have shown that within a very limit<dinter/jngual situation, the basis from which linguistic mat..ial must be r__

organized in order to be 'correct' has been operationally and unambiguouslyntablished. But I have there 'aid nothing about the way in which successfulleamen do in fact reorgani7_e linguistic material from this particular IL. Here'We can speculate that as part (If a definition of 'learning a second language','sucoessfullearning' of a 'econd language for most learners, involves, to a largoextent, the reorganization of linguistic material from an IL to identity witha particular TL.

5.4 Th< fourth problem is: (a) what are the relevant units of this hypothesiz<dlatent psychological stnKture within which interlingual identifications existand (b) is there any evidence for the existence of these units? If the relevantdata of the psychology of ,econd-Ianguage learning are in fact paralld ult<r­anees in three linguistic sy,tem, (NL, IL, and TL), then it seems to me reason­abl< to hypothesize that the only relevant, one might say, 'psychologicallyreal', interlingual unit is one which can be described simultaneously for par­allel data in the three 'ystem'. and, if possible, for experimentally-induced.data in those systems.

INTERLANGUAGE 225

Concerning underlying linguistic structure, we ,hould perhaps not be toosurprised if it turns out not to matter whose model we need, if an eclecticone will do, or even if such notions as the 'cycle" 'tree pruning', or even 'deri­vation' prove not to have much relevance. If it is reasonable to assume thatthe only linguistically-relevant unit of a theory of ,econd-Ianguage learningis one which is identified interlingually across three linguistic systems (NL,TL, and IL) by means of fossilization and the processes described in section 4,then it follows that no unit of linguistic theory, as these units are currentlyconceived, could fit this criterion. More generally, we should state that thereis no necessary connection between relevant units of linguistic theory andlinguistically-relevant units of a p'ychology of second-language learning,·'"That this assumption is obviously c..'orrcct is dear to me; that many linguistswill not be convinced is also dear.

For evidence of the relevant unit of surface syntactic structure, applyingat one and the same time to these three linguistic systems, I refer the readerto experimental evidence appearing in my paper on language transfer (Selin­ker, 1969). In those experiments subject' re'ponded orally in their native lan­guage to questions presented orally in their NL and attempted to respond inEnglish to parallel questions presented in English. The questions came froman interview designed to elicit manifestations of specific types of surfacestructures in certain syntactic domains. The only experimental instructiongiven was for each subject to speak in a 'complete sentence'. Replicated resultsshowed that the interlingual unit of ,urface syntactic structure transferredfrom NL to IL (not to TL) was a unit roughly equivalent to the traditionaldirect object or to an adverb of place, an adverb of time, an adverb of degree,and so on. I would claim that this unit, a ,urface constituent labelled thesyntactic string, has a behaviora~ unity both in the experimental situation and

30 It is important to b~ar in mind that w~ ar~ here workinK in the domain of 'interlingualidentifications' and thus are in a different counterfaetual domain (Lawler and Selinker.forthcoming) than linguists who work in the domain of the "ideal sp~ak~r-listener"

(Chl)msky. 1965), It se('ms to me that researchers in the psychology of sN:ond-languagelearning a~ in the analo~IHlS pnsition of the language teacher who, Chomsky (1966)admonishes, has the burden of deciding what in linguistics and psychology is relevant

to his needs.Nevertheless, the Jinguistic status of ILs has still to be dl'termined. One would like toknow, for example, whether such things as transformatiolU (lCCUr ill IL grammars, Watkin(1970) asks whether the rules of IL are of the same general construction or shape as therules for the same phenomena ill the second lan~uage. "or are they in a 'reroded' form?".Watkin's data implies the same type of fossilization rl.'lated to som~ similarity among

rules of different ILs.

226 IRAL. VOL. XIJ, AUGUST 1971

~.. 'in meaningful performance situations,31 and thus, if the results were repli­;~.ated in other 'interlingual situations' (i.e. other combinations of NL, TL,

In<f IL), would account for a large class of IL events.With regard to a 'realizational unit', i.e. a syntactic string tied to a specific

semantic notion, replicated results from this same series of experiments showthat responses concerning a topic such as 'subjects studied in school', as 0P"posed to other topics such as 'buying and receiving things' and 'seeing moviesand parades', affected very drastically the surface concatenation of the abov..mentioned strings." This <emantic effect on surface syntactic order in aninterlingual study, if further replicated in other interlingual situations, wouldprovide very powerful evidence for the transfer of the whole realizational

, 31 11le' _urfa~ domain considered wa~ constitutnt concatenation after the verb. Samplere.ults showed Hatislically-si~nirlcant paralld trends for Nt (Hebrew) and It (English)Objtet and T,mt constituents on the one hand and (direct) Objtct and AdfJerb (ofdeCr'ff) on the other. That is, whenner an C?biect constituent and a Ti"!t co~stituent

occurrtd after rhto v~rh. thto statistically-dommant surface order was ObJ~ct-T,me, and

not the revtrH', both concerning NL respon5es, e. Jt. (9), and IL usponse5, e.g, (10):

(9) raiti ret ha..eret hauJ [Iifney huaimJ'I saw that movie two weeks ago'

(10) I m" [Mrs. Cosman) [today]But whenever an Object con~tituent and an Adverb constitucont occurred after thco verb,the statistically-dominant surface ordcor waS Adverb-Obj~Cl. and not the reverse. bothconcerning NI. rr'pon,r', co.~. (11) and II. ~sponAeS. e. g. (12):

(I I) ani ohev Imcoodl luatim] 'I like movies very much'(12) I like Inry mu(:h I Imovies1

Importantly, these and all other t'xptrimental results were controlled informally by ob­serving speakers of all ages over 12, from this interlingual situation, producing IL utter­

ances in meaningful prrformance situations.32 That is, wh~n th~ respons~~ wncrrned th~ tl)pic 'subjecu studied at school', there occured

an almost absolute trt'nd toward both the NL (Hebrew) oder Place-Object noun afterthe verb, e.~. (D), and toward the ~ame IL (English) order of surface constituent" e.,_(14):

(13) ani nlca lilmlHJ Ihaunivenita] [biologiaJ'I wan I to stud}" biolll~)' at the university'

(14) I will study lin the university] (biology]But when the responses concerned topics such as the other two topics mentioned in the,teo_to thert occurrt'd an almost absolute trend toward both the NL order ObiKIItOv..PI.ct after the verb, e.g. (15) and toward the same 11. order of surface constituent.,

•. g. (16):(l5) kaniti ret hahonl (haxanut]

'I bought the watch in the store'

(16) 1bouKht [my walchl [in Tel Av;v)For further details, see Stlinkt'r (1969) sections 3.,41 and 3."2.

" ',

INTERLANGUAGE 227

unit as well as for its candidacy as the unit of realizational structure in inter·lingual identifications.

Concerning the notion of relevant units on the phonological level, it seemsto me that Briere (1968) has demonstrated that for his data there are severalrelevant units. The relevant units do not always correspond to known lin·guistic units, but rather would depend on the sounds involved; sometimes thetaxonomic phoneme is the unit, but the unit in other cases seems not to bedescribable in purely linguistic terms. Briere evolved an experimental tech­nique which imitated to a large extent actual methods of teaching advocatedby applied structural linguists: listening to TL sounds, attempted imitation,use of phonemic transcription, physiological explanations, and so on. If Imay be allowed to reinterpret Briere's data, it seems to me that he has beenworking, in another interlingual situation, with exactly the three systems weare discussing here, NL, n, and IL: first, NL utterances which were hypo­thesized utterances in American English; second, TL utterances which wereactual utterances in the 'composite language' Briere set up, each utterancehaving been produced by a native speaker of French, Arabic, or Vietnamese;third, IL utterances which were actual utterances produced by native speakersof this NL when attempting to produce this particular TL norm. Regardingthe sounds Ii.! and I~I in his TL corpus, the unit identified interlinguallyacross these three systems is the taxonomic phoneme defined distributionallywithin the syllable as opposed to within the word (Briere, 1968, p. 73). Forother sounds the relevant phonological unit of interlingual identifications isnot the taxonomic phoneme, but may be based on phonetic parameters someof which, he says, are probably not known (ibid., pp. 7J & 64).

If these units in the domain of interlingual identifications are not neces­sarily the same units as those if!. the native-speaker domain, then where dothey come from? An interesting bit of speculation about native-speaker per­formance units is provided by Haggard (1967, p. 335) who states that search­ing for «the unit" in native-speaker speech·perception is a waste of time.Alternative units may be available to native speakers, for example under noiseconditions." While other explanations are surely possible for the wellknownfact that noise conditions affect performance in a second language, and some­times drastically, we can not ignore the possible relevance of Haggard'sintriguing suggestion: That alternative language units are available to indi­viduals and that these units are activated under certain conditions. It fits invery well with the perspective outlined in this paper to postulate a new type

33 The fact that Haggard is concerned with alternative units which are inclusive in lar~r

units has no bearing on the issue under discussion in this section.

~':1:l 1/\:\1,.\\1[ ..\ ..1, ;\lH,U'lj I'//!.

vi' psycholinguistic unit, available to an individual whenever he attempts toproduce senten~t."s in a second language. Thisinterlingual unit stretches, we

hypothesize, across three linguistic systems: NL, IL, and TL, and becomesavailable to the idealized second-language learner who will not achieve native­speaker competence in the TL, whenever he attempts to express meanings,which he may already have, in a TL he is learning, i.e. whenever he attemp.tsto produce a TL norm. These units become available to the learner only afterhe has switched his psychic set or state from the native-speaker domain tothe new domain of intedingual identifications. I would like to postulatefurther that these relevant units of interlingual identifications do not comefrom anywhere; they are latent in the brain in a latent psychological structure,available to an individual whenever he wishes to attempt to produce thenorm of any TL. •

5.5 The final diffi<ulty with this perspective which we will treat here is thefollowing: how can we experiment with three linguistic systems, creating the!lame experimental 'onditions for ea,h, with ont: unit which is identifiedinterlingually across tht."se systems? J can only refer the reader once again tomy own experiments on language transfer (Selinker, 1969) where manifesta·lions of desired con,atcilations of particular surface syntactic structures wereobtained in what, I believe, was an efficient and valid manner. An oral inter­view technique was used i the purpose of the interview was to achieve a similarframework in the three systems whi,h served the interviewer as a guid~ inhis attempt to elicit certain types of sentences from the subjects. Upon request,I am prepared to make available a transcript of this interview as well as somethuughts for its improvement. r~lIturc experimental work, to be undertakenwithin this perspel.:tive, will gu toward investigating the kind and extent oflinguistic structures amenable to this partkular technique.

6, Summar)'

The following are some assumptions which are necessary for research into thelinguistic aspects of the psy<hology of second-language learning and whichhave been suggested by the above discussion.

1) In a theory of second-language learning, those behavioral events whichare to be counted as relevant data are not immediatdy obvious.

2) These data have to be organized with the help of certain theoreticalt:onstructs.

3) Some theoretical constructs relevant to the way in which 'adults· ac~

tually learn second languages are: interlingual identifications, native language

IN IlKLANGUAtiJ! U\h-...

(NL), target language (TL), intedanguage (IL), fossilization, syntactic string,taxonomic phoneme, phonetic feature.

4) The psychologically-relevant data of second-language learning areutterances in TL by native speakers, and in NL and lL by second-languagelearners.

5) Intedingual identifications by second-language learners is what unitesthe three linguistic systems (NL, TL, and IL) psydlOlogically. These learnersfocus upon one norm of the TL.

6) Theoretical predictions in a relevant psychology of s«olld-Ianguagelearning must be the surface structures of IL sentences.

7) Successful second-language learning, for most learners, is the rcorganiazation of linguistic material from an IL to identity with a particular U.

8) There exist five distinct processes which are «ntralto second-languagelearning: language transfer, transfer-of-training, strategies of second-languagelearning, strategies of second-language communication, and overgeneraliza­tion of TL linguistic material.

9) Each prediction in (6) should be made, if possible, relative to one ofthe five processes in (8).

10) There is no necessary connection between rdevant units of linguistictheory and linguistically-relevant units of a psychology of second-languagelearning.

11) The ollly linguistically-relevant unit of a psychology of second-lan­guage learning is one which is identified interiinguaJly a~ross the three lin­guistic systems: NL, TL, and IL.

12) The syntactic string is the unit of surface strUl.:ture transfer and partof the unit of realizational transfer.

13) The taxonomic pboneme is. in the case of some sounds, the unit ofintedingual phonology, while in oth... cases no purely linguistic unit seemsrelevant.

14) There exists a latent psychological structure, i.e. an already formu­lated arrangement in the brain, which is activ~ted whenever an adult attemptsto produce meanings, which he may have, in a second language which he islearning.

15) lntedingual identifications, the units mentioned in (12) und (13),and the processes listed in (8) exist in this latent psychological structure.

16) Fossilization, a mechanism whid\ also exists in this latent psychologaical structure, underlies surface linguistic material which speakers will tendto keep in their IL productive performance, no lUaller what the age of thelearner or the amount of instruction he receives in the TL.

230 IRA!. VOL. X13. AU<;UST 1972

17) The fossilization mechanism accoonts for the phenomenon of theregular reappearance in IL productive performance of linguistic materialwhich was thou~ht to be eradicated.

18) This latent psy{;holo~ical structure, for most learners, is differentfrom and exi,t, in addition to the latent language structure described byLenneber~ (1967, pp. 374-379).

19) The,e two latent structure' differ in the following ways: (a) the latentp'ychological structure has no genetic time-table; (b) it has no direct counter­part to any grammatical concept; (c) it may not be activated at all; (d) it maynever be realized into a nalural language; and (e) it may overlap with otherintellectual structures.

20) The qualification ('for most learoers') in (7) and (18) is necessary,slnee those adults who seem to achieve native-speaker 'competence', i.e. thosewho learn a second language so that their cperformance' is indistinguishablefrom that of nalive ,peakers (perhaps a mere 5 % of all learoers), have notbeen taught this performance through 'explanation and instruction' but havesomehow reactivated this latent language structure.

21) Since it is assumed that the two structures mentioned in' (18) aredifferent and sinl" we know very little about the latent language structureand its activation, then the 5 % mentioned in (20) should be ignored in setti~g

up the idealizations which guide us to the psychologically-relevant data ofsecond-language Ifarning. '

Larry SelinkerDepartment of LinguisticsUniversity of WashingtonSeattle, Washington 98105USA

REFERENCES

Br~~. Eugene J.: A rSyfholi"gI4;stir Study of Phonological Inttrftrtnct. Mouton, 1968.Chomsky, Noam: AJptets oftltf' rhtory oJSyntax, M,I.T. Press, 1965.Chomsky, Noam: "Lin~uisti(: ThC'ory", Northtast ConJtrC'nCC' on thC' Ttaching oj Forri".

ungNagtS, 43-49,1966.Chomsky, Noam: "Linguistics and Philmophy", il1l.angu~gtimd Philosophy, ed. by Sidney

Hook. NC'w York UnivC'rsity PrC'55. 1969, 51-94.ConlC't. S. Pit: "The SignificancC' of LC'arnC'r's Errou", (RAL, 5 (1967), 161-170.Couhtr, KC'nnC'th: "I.in~ujstil.· F.rrur- Analy"j50f the Spoken English ofTwo Native Russian'."

," Unpublished M. A. thesis, UnivC'tsity of Washington, 1968. ' ',J

•

INTERLANGUAGE 2 \

Crothers, Edward and Suppes. Patrick: F.xpniments in St'cond-Langudgt' Ltarning. AcaJcllll'Press. 1967.

Fodor. Jerry A,: Psychological Explan.lIion: An ItltrodllCtion to the Philosphy oj Psycho/oKlRandom HousC'. 196ft

Ha~~~rd, Mark P.~ "Mo~lels and f);ua ill Spel.·ch Pen:eption··. in Models/or the Ptraptl'"oj Sptfch and Vm~al form, ed. oy Weiant Wathen-Dunn. M.I.T. PrC'ss, 1967. 3JI _ 3Y,l

Hymes, Dell: On Communicative Competence, PC'nguin, (in press).

Jain, M~havir: "Error Analysis of all Indian English Corpus." Unpublished paper, Uni\'ersil \of Edmburgh, 1969.

Jakobovits, Leon A.: "Second Language Learning and Transfl"r Theory: a Theoretical Asse~,Illent", Language Ltarning, t 9 (Junl". 1969), 55 -86.

jakobovits, Leon A.: Fortign Languagt Lrarning: A Psycholinguistic Analysis oj the lUll,'Newbury House, t 970,

Klint', Hden: "Resl"arch in the PsydlUlngy uf Secorld-I.anli:uat:e Learning." Ullpuhli~h,'.,paper, UnivC'rsity of Minnesota, 1970,

Labov, William: "Contractiun, DeI(·til1n, and Illhert"nt Variability llf the English COpUI.l'·Languagt, 4S.4 (t 969),715-762.

LakuU. George: "On GC'nerativt Sem.uuks", in SemanticJ - An Inttrdiuiplin",ry Rt~d,',in Philosophy: Lingliist~cs, Anthropology and Psychology, ('d. by Danny Steinberg :tIll!

Leon Jakobovlts. Cambndge University I'n:ss, (in press).

Lawler, ~ohn and Selinker. larry: "On Paradoxes, Rub. and Research in Second-Langu.1L:,Ll'arnmg", Languagt Learning (in press). '

Lenllt"berg, .Eric H.: Bi%gical FOImdatiom of 1,''''RII.lge. John Wiley and Sons Inc., 1967.NemsC'r, William: "Approximative Systt'llH (If Furei~n l.anguage Learners".IRAL 9 (1971,

115-123.

Reibel., D. A.: "Language Lt"arning Strategies for the Adult", paper read at Second hllt'l

. natIonal CongrtJS of Applied l.inguiuics. CambridJi:e UnivC'rsity. Sept., 1969..Richards, Jack c': "A NUIl-Contrastivt' Approach to Error Analysis." Paper de/iq'red .It

TESOl Convention, San Francis'o, March, 1970.

Selin~C'r, larry: "Language Transfer", Ge"rrall.inguistics, 9 (1969).67-92.

Watkll,l' "K. 1..: ",Fossilil.ation and its Implications Re~arding the Interlanguage HH'Othem. Unpublnhed papt'r, Uni1't'rsit)' of Washington, 1970. .

Weinreich, Urit"l: Languages in CoraMt. ThC' Linguistic Circle of New York, 1953.