, I

.. j

The. new Applied Linguistics and Language Study series is intended. to reflect an increasing awareness by language teachers of the contribution to classroom teaching of linguistics, sociolinguistics and psycholinquistics, The volumes in the series cover areas of pedagogical application of such theoretical and descriptive studies and thus define more clearly the scope of applied linguistics and language study for the, informed language teacher and student. -#. .


One of the more important shifts in Applied Ling recent years has been from the view of the teacher of language learning towards a more leamer-cente stresses the leamer's powers of hypothesis format towards that bilingual competence sufficient for hi needs, One major result of this shift of attention has ing concern in the monitoring and analysis of lea This is an important concern both for language teac From the standpoint of practical teaching we ha aware of the long-term value of Error Analysis as both assessing the pupil's learning in general and match between his learning 'syllabus' and the teache Linguistically, the concepts of 'interlanguage' and system' present challenging areas of descriptive enq issues ofthe validity of the competence/performance suggest useful analogies more widely to the links th between the language leamer's language and 'contin in general, particularly Creoles.: Error Analysis ha Applied Linguistic justification in that data from th both serve as input to theoretical discussion (clnd, feed back to the design of remedial curricula. There is considerable value, therefore, in Dr R together in one volume a set of readings spanni fields. His important introductions and commentari the interconnections apparent. iHe has begun fr concerned with practicalities of error evaluation, gravity and deviation, and the links that can be draw and performance assessmentj From these and furt explanation posed by the determination of error looked for descriptive evidence both to clarify the suggest to the language teacher the wider connect





(iv) The need to distinguish between the descrip incorrect and the processes that were involved i of the error highlights the absence in error optimal means for linking error identification i with diagnosis in psycholinguistic terms.

Christopher N


9 Idiosyncratic Dialects and Error Analysis

10 A Non-Contrastive Approach to Error Analysis

11 Error Analysis: Source, Cause and Significance


~~~ ~~!~terference is suggested elsewhere in thi

Corder's paper rai . language teachin~.r~~t ais q~:s~~~t~~:~~~pl:;n and what is Ie d? S h . w stud . h ahrne . uc a question can on . ymg w at appens after we have

mst,;,ctionalsyll~bus. Corder suggests P::t"'::

emp oys or establishes his own syllabus for learni a more effective syllabus than one prepared from Thhae nattu~e0df this sylla~us and some of the ways C rae enze are descnbed in th bv Ri which isolates seven factors as ~t::a~:eJz~:ha~:s leamer's language u~e:.language transfer, intrali the effects of the sociolinguistic situation the m to the target language and the modali ' the leamer, the instability of the learne;'s ~~ pr~d

effects of the inherent difficulty of the parti~:f::~~





languages in terms of their native language or o whi~h they had earlier been exposed. With the notion of language as a system however, the qu language acquisition could be viewed as the juxt systems. Juxtaposition of language systems could lead system which combined features of both systems 1949), or to intersystemic interference (Weinreich, of interference between two systems struck linguist especially interesting, since it appeared to account of second language learning, particularly of adults Lado (1957), for example, tended to emphasize p between two language systems. Contrastive analy arose as a field of research. To be sure, contrast was ~nderstood not to be the only factor involved in learning, Subs~quent attempts to rectify what was see theoretical approach to language learning evolved linguists refer to as error analysis. The major defe analysis was deemed to be the attention paid to the grammars. Some linguists proposed closer study of of actual learners. Corder (1967), for instance, sugges study the process of language acquisition and the v learners may use. Strevens (1969) hypothesized th not be viewed as problems to be overcome, but r and inevitable features indicating the strategies th He conjectured that if a regular pattern of errors co in the performance of all learners in a given situation, were seen to progress through this pattern, his errors as eviden~ not ~f failure. but of success and achievem But the pioneenng studies of learner errors were s a framework which stressed interference between ~uch as they str~ssed errors per se. For example, N ~IS early work aimed at 'the collection and evaluat mterference. data.' Briere (1968) attempted to 'test am?uot of mt~rference that would ensue from com logical categones.' Errors which did not fit systema ~atIve language or target language system were, for Ignored. Both Briere and Nemser, however, no replacements occurred which were not in either the language or the (artificial) target language.


4 Intralingual Interference The second factor, termed intralingual interferen (1970), refers to items produced by the learner whi structure of the mother tongue, but generalizations exposure to the target language. In an analysis o pro~uced by speakers of a multitude of unrelated senting several l.anguage families, Richards noted error types which seem to be common to spe la~guages as they develop hypotheses about the stru LIk~ first language learners, the second language denve the rules behind the data to which he has be may develop hypotheses that correspond neither tongue nor target language. In an experiment on wor~ order, Torrey (1966) found that subjects som c~nsIstent word order different from either Rus Richards (l.971.a) fo~nd systematic intralingual e over~en~rahzatIOn, Ignorance of rule restrictio apphca~IOnof rules, and s.emantic errors. These hav for ThaI learners ~f En~hsh by Brudhiprabha (197 suggest that many intralingual errors represent the l o~ what are often low level ~ules in the target la dI~erences between the verb inflection in I walk, s be .mferred ~hat ~nce basi~ rul.es such as those con object relationships, predication, negation, etc. considerable am~unt of difficulty in second lang related to selectional restrictions and to surfac ~ontextual rules of the language. Both language tr lmg~al erro~s confirm the traditional notion of tran that IS,previous learning may influence later learni 5 Sociolinguistic Situation

third factor is the sociolinguistic situation. Diff ang~age use result in ~i~eren~ degrees and ty lea~mg. These m~y be distinguished in terms of socI~-cult~ral setting on the learner's language an relatIons~Ip holding betwe~n the learner and the com~uDl~. and the respective linguistic markers o and identities, Included here are thus the effects art"c lar ti . I.' mo vatIO~s lor learning the second lan P 1u the effects of the socio-cultural setting.



comer, When you's coming?, may depend on the the le.arner's values and attitudes, or some other s learning context. The phenomenon of simpli language contact situations, represented by th copula, r~ducti.on ~f m~rphological and inflectio grammatical simplification, may likewise be s ~Ferguso~, 19?1). When the need is for commun information WIth the help of non-linguistic clues, and word o~der may be the most crucial element as the expenen~ of t.ourists .in foreign countries folklore exemplified m comic-book caricatures Jane) testify. The influence ~f the mother tongue on the leame also vary according to the sociolinguistic situation notes that in describing interference one must acc according to the medium, style, or register in wh 0tx:ratIng. For example, the medium may be spok register may be formal or informal, and the spea one of a .num~r ~f varied social.roles. Sampso that v~~mg sltuat~ons evoke different kinds of quantines when children are trying to use the targ 6 Modality

The leamer's language may vary according to a modali~ of exposw:e to the target language and production. Production and perception may invol ?f two partially overlapping systems. Vildom mterfer~nce between the bilingual's languages is pr~uctl.ve rather than receptive side. People often of mtruslo~ of ele~ents of their mother tongue in s but rarely m t~err understanding of another lan .1963).~ems~r s research suggests that two differen mternahzed. m the target language depending on f~und that m t~e productive modality, phonologi differed depending on whether the learner was im he ~eard or producing speech spontaneously (N fact in first lan~age acquisition Lieberman (1970) ~~me p~onolo~cal features exist because their 'a match .a .partIcular neural acoustic detector. Ot because It 18 easy to produce a particular articul



ing than children. Adults have better memorie abstract concepts that can be used in learning, a to form new concepts. Children however are speech sounds. Ervin-Tripp (1970) suggests tongue development is primarily in terms of voca strategies of language learning may be more v than syntactic. Acquisition of syntax poses a task is no longer easy (Ervin-Tripp, 1970). The la device (or processes) by means of which he w induce the rules of his mother tongue in childho readily activated. Selinker suggests that this de only 5% of adults (those who achieve native spea the new language). Most language learners ( achieve native speaker competence) activate a di genetically determined structure (Selinker, this v No categorical statement about the relation learning to age can be made. As Braine (1971) rem is not so much whether chiJdren are biological acquire language, but rather a question of ho learning resembles or differs from adult language leads to an examination of the different language l which typify child and adult language learning. M and Kennedy (1973) hence see differences in results as reflecting mainly motivational and situational need to understand and speak are much more cru young children than for many adults learning a ne Research into second language learning and bili gone very far towards explaining how the rules for sentences in two or more languages by the same sp a question which becomes more crucial when the ag Swain (1971), working with children from 2.1 to 4. exposed to two languages from birth suggests tha of this age, the differentiation of two or more lingu a significantly different problem from that faced by child who acquires control of several different v informal, intimate-serious) of one language. A mode separate sets of rules for each code would be ineffi memory storage. 'More efficient would be a comm with those s~cific to a particular code tagged as process of dI~~r~ntiation' (Swain, 1971, 6). Lookin taneous acquisinon of yes/no questions, she found



1971). ~ince most studies of second language lear dealt WIth the learner~s production rather than h of ~anguage, the question also arises as to whethe which the learner understands speech is the same a produces speech, since as we saw above, modality type of system developed. Assuming the learner stands standard English but produces a significant sentences, the distinction between his receptive rules he understands) and his productive compete uses) may be useful (Troike, 1969). The learner m for example, I has a book but understand when h book: ~ic~ rule is present in his grammar? As acquisrtion, ill the development of a second lang manyelemenrs are observed to go through a stag sometimes used and sometimes omitted. A gr featu~es might ~ntain the rule but specify that Thus if rules for Items or structures unique to learne systems are to be written they will need to be embo refle~ting their probability of occurrence, if this ca ~ossI~I~ a fo~at like that proposed by Labov to hng~lStICvanable will be useful. Labov et al. (196 specific quantity be associated with each variable r proportion of cases in which the rule applies. Th of the rule structure itself. It. is to be expected too that the rules characteriz mative system may cover data (replacements) whic in e~ther native or target language. For example ~ar~ler studies (Briere, 1968; Nemser, 1971; indicates that many phonological replacements fou of seco~d language learners are unique to the approx The eXlste~ce~f such novel data is strong support f of ~ppr~X1matIvesystems as distinct from native an ThIS claim of autonomy for approximative systems ever, preclude their dependence upon either native o 9 Universal Hierarchy of Di16culty

Unlike the factors characteristic of approximative the seventh factor has received little attention in ~econd la~guage acquisition. This factor is conc inherent difficulty for man of certain phonologica



to recognize such features. Their errors indicate regularities in the French gender system. Difficulty in language learning has been de linguists in terms of such factors as sentence lengt required, derivational complexity, types of embe transformations, and semantic complexity. Expe has not however confirmed a direct relationship comprehension of an utterance by an adult listene of rules used by the linguist in describing the u comprehension seems to be defined by the type (Fodor and Garrett, 1966). Russian work on th the order of development of grammatical catego the difficulty of the concepts expressed by the cat formal grammatical complexity influences orde Categories of more concrete reference develop bef ing more abstract or relational ideas. The condi develops relatively late, though its structure is qui 1971). The learner's comprehension and efforts at co now be compared with his production. His outpu organized in terms of what he finds easiest to necessarily identifiable with what he knows. Lear word or structure they find difficult to say. This m a particular tense instead of the required one (I'm you tonight instead of I' IItelephoneyou tonight). Fac of effort may explain why first learned words and be overused and may resist replacement by later is often the case with such contrasts as simple continuous. Once the present continuous is introdu present) it is often used more frequently than nece early learned question form may replace all other the learner's speech, despite constant exposure 'questions. Words with broad semantic extension m used in preference to more specific vocabulary lea refers to this as the factor of chronology. 'We all k learned first have priority over patterns learned cause of the convenient simplicity of these first bas kind of intrastructural interference will take plac interstructural contrastive background. Thus Norw German will very often use word order of the m subordinate clauses even though conditions in the



the sort of data on which realistic predictions a teaching can be based. Instead of expecting learne to native speaker competence in the new langua goals may be set for particular learning situations, izations derived from observation of how other under similar circumstances. Teacher training familiarize future teachers with the varieties o learners may be expected, and indeed encourage given stages of learning. These will not correspond well-formed sentences in the course book, drill or Study of learners' learning systems also has int theory. The results of error analysis studies cited a theories of contrastive analysis cannot explain replacements for the target sounds occurred. As points out, discussing the replacements for Eng Javanese and Indonesian speakers, it is not clear the same linguistic background should make var of lsi, /t/, and Iff. He states that 'this is the more not f?und in Javan~se, and. in Indonesian it no functional load but IS found m some words in free and not with /t/ or /s/.' He comments that 'the /f/ to be more commonly made by younger learners. If these language problems are viewed in the co mative systems, it appears that several factors are a the age factor is relevant. As we saw above, it strategies of language acquisition differ in adults a information about language difficulties of children from the many educational studies done on oral l as a prerequisite to reading ability. Second, p shown that in sound discrimination, the secondtransitions from a fricative into the vowel are i identifying place of articulation (Delattre, et al. (1970), working with first-grade native speakers that his subjects made significantly more errors o versus /9/ and lvl versus /6/ in the context of'a front to a back vowel. The longer transition length to the to provide a better cue for the correct discriminatio Varying vowel environments for English lsi, on ~eem to be irrelevant because its noise portion pro m terms of frequency, intensity, and duration. T information is not needed. These data make it c


in the gradual acquisition of the target system may understanding of language in general and a more to language teaching.




tion of them, and here has reasonably felt that the ling little to say to him. In the field of methodology there have been two schoo in respect oflearners' errors. Firstly the school which m if we were to achieve a perfect teaching method the e never be committed in the first place, and therefore th of errors is merely a sign of the present inadequacy of techniques. The philosophy of the second school is that imperfect world and consequently errors will always o of our best efforts. Our ingenuity should be conc techniques for dealing with errors after they have occu Both these points of view are compatible with the sam standpoint about language and language learning, psy behaviourist and linguistically taxonomic. Their ap language teaching is known as the audio lingual or funda method. Both linguistics and psychology are in a state at the of what Chomsky has called 'flux and agitation' (Chom What seemed to be well established doctrine a few year the subject of extensive debate. The consequence of this teaching is likely to be far reaching and we are perhap beginning to feel its effects. One effect has been perhaps emphasis away from a preoccupation with teaching tow of learning. In the first instance this has shown itself a attack upon the problem of the acquisition of the mother has inevitably led to a consideration of the question w are any parallels between the processes of acquiring tongue and the learning of a second language. The usef distinction between acquisition and learning has been by Lambert (1966) and the possibility that the latter from a study ofthe former has been suggested by Carr The differences between the two are obvious but reason easy to explain: that the learning of the moth inevitable, whereas, alas, we all know that there is evitability about the learning of a second language; that of the mother tongue is part of the whole maturational p child, whilst learning a second language normally begi the maturational process is largely complete; that the with no overt language behaviour, while in the case o language learner such behaviour, of course, exists; that tion (if we can properly use the term in the context) fo




loses' the capacity to acquire language behaviour at does not of course carry with it the implication th learning capacity of those who have successfully le language also atrophies in the same way. It still rem that the process of learning a second language is of different nature from the process of primary acquis If we postulate the same mechanism, then we ma that the procedures or strategies adopted by the learn language are fundamentally the same. The princi then differentiates the two operations is the presen motivation. If the acquisition of the first language the predisposition to develop language behaviour, t of the second language involves the replacement of th of the infant by some other force. What this con context of this paper irrelevant. Let us say therefore that, given motivation, it is human being will learn a second language if he is language data. Study of language aptitude does i support such a view since motivation and intelligen the two principal factors which correlate significan ment in a second language. I propose therefore as a working hypothesis that the strategies adopted by the learner of a secon substantially the same as those by which a first langu Such a proposal does not imply that the course learning is the same in both cases. We can now return to the consideration of errors m When a two-year-old child produces an utterance mummy chair' we do not normally call this dev faulty, incorrect or whatever. We do not regard it a sense at all, but rather as a normal childlike comm provides evidence of the state of his linguistic deve moment. Our response to that behaviour has certain istics of what would be called 'correction' in a class Adults have a very strong tendency to repeat and ex utterance in an adult version; something like 'Y Mummy's chair'. No one expects a child learning his mother-tong from the earliest stages only forms which in adult t or non-deviant, We interpret his 'incorrect' utter evidence that he is in the process of acquiring langu



Whatever sequencing criterion is used it is one calls a 'logical' sequence. But although there are s by which sequencing can be accomplished and generally agreed that an effective sequence is meaningful to the leamer, the information s assimilated by the learner is traditionally dicta the instructor. We generally fail to consult the matter except to ask him to maximize the e whatever sequence we have already decided upo

He points out as the conclusions he draws from experiment that the next step would be to determi leamer-generated sequence, or, as we might call syllabus, is in some way more efficient than the instru sequence. It seems entirely plausible that it would be s is to determine whether there exists such a built-in describe it. It is in such an investigation that the st errors would assume the role it already plays in the language acquisition, since, as has been pointed out, in both cases is that the learner is using a definite sys at every point in his development, although it is not in the one case, nor that of the second language in leamer's errors are evidence of this system and systematic. The use of the term systematic in this context im that there may be errors which are random, or, mo systematic nature of which cannot be readily discern tion between systematic and non-systematic errors i are all aware that in normal adult speech in our nat are continually committing errors of one sort or an we have been so often reminded recently, are due to physical states, such as tiredness and psychological as strong emotion. These are adventitious artefac performance and do not reflect a defect in our knowl language. We are normally immediately aware of occur and can correct them with more or less com It would be quite unreasonable to expect the lear language not to exhibit such slips of the tongue (or subject to similar external and internal conditions w in his first or second language. We must therefore m between those errors which are the product of s


s. P.


Child: I seed him. Mother: Ah, you saw him. Child: Yes, I saw him.

Here the child within a short exchange, appears three hypotheses: one relating to the concord of s in a past tense, another about the meaning of sho third about the form of the irregular past tense of see to be pointed out that if the child had answered mediately, we would have no means of knowing merely repeated a model sentence or had already l rules just mentioned. Only a longitudinal study development could answer such a question. It is a observe the techniques used by the mother to 'co Only in the case of one error did she provide the c self: You saw him. In both the other cases, it was su query the child's utterance in such a form as: yo showed him? Simple provision of the correct form be the only, or indeed the most effective, form of c bars the way to the learner testing alternative hypo learner try to discover the right form could often be to both learner and teacher. This is the import of C already referred to. We may note here that the utterance of a correc taken as proof that the learner has learned the syst generate that form in a native speaker, since he repeating a heard utterance, in which case we s behaviour, not as language, but in Spolsky's term 'language-like behaviour'. Nor must we ?verl?ok utterance which is superficially non-deviant IS n mastery of the language systems which would ge~e speaker since such an utterance must be semantica situational context. The learner who produced 'I English' might have been uttering an unexceptionab it is more likely that he was expressing the wish to language. Only the situational context could s utterance was an error or not. Although it has been suggested that the strateg first and second language may be the same, it is n sary at this point to posit a distinction between th may suppose that the first language learner has an



approximative system for interlanguage. This term ha of implying the developmental nature of language the learners' system is continually being modified a are incorporated throughout the learning process. S systems are evident in learner's errors. Nemser note tion to the content rather than the form of utterance (cf. communication strategy) seen in learners' simp may also be characteristic of the teacher's speech 1971). He gives phonological examples of the interm independent systems developed by the learner. Su internal coherence and are not simply corrupt versio language or mother tongue and their characteristics ing to the degree of learning. Such systems are ille linguistically-their users do not normally form spee (cf. Richards, this section)-though they. are not linguistic significance, as Nemser illustrates. In the third paper I have tried to suggest how learner's developing system is determined by the soc holding between the learner and the target langua An attempt is made to relate the linguistic concepts where in this volume (interlanguage, approximative generalization, communication strategy, etc.) to an the different social settings in which the learning o place. The non-linguistic dimensions oflanguage lea include the effects of the perception of the target langu by the learner, the structure and rigidity of soc community, the economic and social strength of the form a community, and the roles for which English learner. An analysis of the learner's developing lingu be complemented by consideration of the social fac these features significance. The paper hence mak between second language and foreign language c learning of English which have important conseq view of the learner's interlanguage or developing sy



second language. The term 'meaningful performanc here to refer to the situation where an 'adult'? atte meanings, which he may already have, in a languag the process of learning. Since performance of dr language classroom is, by definition, not meaningfu it follows that from a learning perspective, such in the long run, of minor interest. Also, behavior experiments, using nonsense syllables fits into the sa for the same reason. Thus, data resulting from these l situations are of doubtful relevancy to meaningf situations, and thus to a theory of second language It has long seemed to me that one of our greate establishing a psychology of second language le relevant to the way people actually learn second lang our inability to identify unambiguously the phenom study. Out of the great conglomeration of secon havioral events, what criteria and constructs sho establish the class of those events which are to coun theory construction? One set of these behavioral e elicited considerable interest is the regular reappear language performance of linguistic phenomena whi to be eradicated in the performance of the learner. A standing of this phenomenon leads to the postula theoretical constructs, many of which have been set other problems in the field. But they also help clarify t under discussion. These constructs, in turn, give u within which we can begin to isolate the psychological of second language learning. The new perspective wh tion of this phenomenon gives us is thus very hel identification of relevant data and in the formulatio linguistic theory of second language learning. The m for this paper is the belief that it is particularly in th gress can be made at this time.

2 'Ioteriaoguage' and latent structures Relevant behavioral events in a psychology of s learning should be made identifiable with the aid constructs which assume the major features of th structure of an adult whenever he attempts to und



success in a second language affects, as we know fro a small percentage of learners-perhaps a mere 5 %. this assumption that this 5 % go through very dif linguistic processes than do most second language le these successful learners may be safely ignored-in a sense" -for the purposes of establishing the construc to the psychologically relevant data pertinent to mo guage learners. Regarding the study of the latter gro (i.e. the vast majority of second language learners who native speaker competence), the notion of 'attempte independent and logically prior to the notion of 'succe In this paper we will focus on attempted learning b learners, successful or not, and will assume that t different, though still genetically determined structu here as the latent psychological structure) whenever t produce a sentence in the second-language, that is, attempt to express meanings which they may alrea language which they are in the process of learning. This series of assumptions must be made, I thin second language learner who actually achieves competence, cannot possibly have been taught th since linguists are daily-in almost every generativ covering new and fundamental facts about particu Successful learners, in order to achieve this native petence, must have acquired these facts (and m important principles of language organization) w explicitly been taught them. 7 Regarding the ideal second language learner who w (in the absolute sense described above) and who is thus of the vast majority of second language learners, w that from the beginning of his study of a second lan his attention focused upon one norm of the langua tences he is attempting to produce. With this statem idealized the picture we wish to sketch in the follow generally accepted notion 'target language' (TL), language the learner is attempting to learn, is here res that there is only one norm of one dialect within the int of attention of the learner. Furthermore, we focus attention upon the only observable data to which we retical predictions.' the utterances which are produ learner attempts to say sentences of a TL. This set of



of IL utterances should be associated with one or mo other, processes.

3 Fossilization Before briefly describing these psycholinguistic proc notion I wish to introduce for· the reader's consid concept of fossilization, a mechanism which is assume in the latent psychological structure described abov linguistic phenomena are linguistic items, rules, an which speakers of a particular NL will tend to ke relative to a particular TL, no matter what the age of amount of explanation and instruction he receives have in mind such fossilizable structures as the well-k French uvular Irl in their English IL, American Engli in their French IL, English rhythm in the IL relativ German Time-Place order after the verb in the English speakers, and so on. I also have in mind less well know such as Spanish monophthong vowels in the IL of Sp relative to Hebrew, and Hebrew Object-Time surface verb in the IL' of Hebrew speakers relative to English are fossilizable structures that are much harder to c some features of the Thai tone system in the IL of relative to English. It is important to note that fossiliz tend to remain as potential performance, re-emer productive performance of an IL even when seeming Many of these phenomena reappear in IL performa learner's attention is focused upon new and diffic subject matter or when he is in a state of anxiety or oth and strangely enough, sometimes when he is in a st relaxation. Note that the claim is made here that, what the well-observed phenomenon of 'backsliding' by se learners from a TL norm is not, as has been generally b random or towards the speaker's NL, but toward an A crucial fact, perhaps the most crucial fact, which theory of second language learning will have to regular reappearance or re-emergence in IL produc ance of linguistic structures which were thought to This behavioral reappearance is what has led me to reality of fossilization and ILs. It should be made reappearance of such behavior is not limited to the



in a minor way, e.g. adding vocabulary as experienc Jain) is, it seems to me, a moot point. If these individu learn the syntactic information that goes with lexi adding a few new lexical items, say on space travel, is of little consequence. The important thing to note wi evideJlce_pres~nt~ in Coulter (1968) and Jain (1969) can entire IL competences be fossihzed in mdlVldu forming iri their own interlingual situation.l? but groups of individuals, resulting in the emergence o (here Indian English), where fossilized IIJ competenc normal situation. _. --~- ---... W~will nowprovide examples of these processes. presented in section 3 are almost certainly the result o language transfer. A few examples relating to the o should suffice for this paper.

4.1 Overgeneralization of TL rules is a phenomenon language teachers. Speakers of many languages co sentence of the following kind in their English IL. (1) What did he intended to say?l8

where the past tense morpheme -ed is extended to a in which, to the learner, it could logically apply, bu The Indian speaker of English who produces the coll bicycle in his IL performance, as in (2)

(2) After thinking little I decided to start on slowly as I could as it was not possible to d

is most probably overgeneralizing the use of drive (Jain, 1969, cf. note 26 here). Most learners of E learn the English rule of contraction which forms concert's from the concert is, but then these learn generalize this rule to produce sentences like, (3) Max is happier than Sam's these days

in their English IL. Though this sentence is hypothetic an earlier point. The learner of English who produce correctly in all environments must have learned constraint without 'explanation and instruction', since was discovered only recently; "contraction of auxilia occur when a constituent immediately following the



is at present pure conjecture. Thus, one can only ro the source of the examples presented herein to o strategy. One example of a strategy of second language l wiaespl'elI<rmmanY situations tende "of leamers to red_ucceJl!e_I1to a simpler system.Acc {I969ftne res-liltsof this strategy are manifested at all in the IL of Indian speakers of English. For example has adopted the strategy that all verbs are either trans sitive, he may produce IL forms such as


is a


(4) I am feeling thirsty (5) Don't worry, I'm hearing him

and in producing them they seem to have adopt strategy that the realization of the category 'aspect' in form on the surface is always with ing marking (for furt see Jain, 1969). Coulter (1968) reports systematic errors occurring IL performance of two elderly Russian speakers of another strategy which seems also to be widesp interlingual situations; a tendency on the part of se learners to avoid grammatical formatives such a plural forms (7), and past tense forms (8);

(6) It was ~ nice, nice trailer, ~ big one. (Coulte (7) I have many hundred carpenter my own (ib (8) I was in Frankfurt when Ifill application (i

This tendency could be the result of a learning stra fication, but Coulter (1968) attributes it to a communi due to the past experience of the speaker which has s if he thinks about grammatical processes while attemp in English meanings which he already has, then his hesitant and disconnected, leading native speakers t with him. Also, Coulter claims that this strategy of s communication seemed to dictate to these speakers th as the English plural 'was not necessary for the kind o ing they used' (ibid, p. 30). ~f t~t:se strategies, it must be pointed out, . A subconscious strategy of second language -learnin COpying' 1ias~en-experimented with by Crother ~1) on Americans learning Russian morpholog



that these psycholinguistic structures, even whe eradicated, are stiII somehow present in the brain, fossilization mechanism (primarily through one o processes) in an IL. We further hypothesize that identifications uniting the three linguistic systems (NL psychologically, are activated in a latent psychologi whenever an individual attempts to produce TL senten

5.1 The first problem we wish to deal with is: can w ambiguously identify which of these processes our ob is to be attributable to? Most probably not. It has be pointed out that this situation is quite common in ps studies on memory, for example, one often does not k one is in fact studying 'storage' or 'retrieval'. In our not know whether a particular constituent IL conca result of language transfer of transfer of training or both.26 But this limitation need not deter us, even always sort things out absolutely, By applying th suggested in this paper, I believe that relevant data can the very many second language learning situations aro

5.2 The second problem is; how can we systematiz fossilization so that from the basis of theoretical const predict which items in which interlingual situations w ized?! 0 illustrate the difficulty of attempting to answer note m the following example the non-reversibility of effect:>for no apparent reason. According to a contras Spanish speakers should have no difficulty with the he/s in English, nor should English speakers have any diffi corresponding distinction in Spanish. The facts are qu however: Spanish speakers do, indeed, regularly have this distinction, while the reverse does not seem to occur learners of Spanish."? Unlike the Serbo-Croatian e tioned above, in this case there is no clear-cut expl Spanish speakers have trouble and English speakers cases such as these, it may tum out that one process, transfer or transfer of training, overrides other consid the stating of the governing conditions may prove indeed. In principle, one feels forced to agree with Steph (personal communication) who claims that until a theo



(see note 2) can be explained by the failure to see a of second language learning in terms other than tho 'success'. For example, typical learning-theory experi done in the domain of second language learning wo knowledge of where the learner will tend to end up, n would like him to end up. Experiments of this type demand knowledge of where the second language learne would claim that prerequisite to both these types of kn detailed' descriptions of ILs-descriptions not presently us. Thus, such experiments at present are premature, wi bound to prove confusing. Specifically concerning the problem raised in the firs 5.3, it seems to me that this question, though relevant to ogy of second language learning, is one that should also for the present since its asking depends upon our un clearly the psychological extent of interlingual identifi example, before we discover how surface constituents reorganized to identify with the TL, we must have a what is in that IL, even if we cannot explain why it Selinker (1969) I believe I have shown that within a interlingual situation, the basis from which linguistic m be reorganized in order to be 'correct' has been opera unambiguously established. But I have there said nothin way in which successful learners do in fact reorganiz material from this particular IL. Here we can speculate of a definition of 'learning a second language', 'success of a second language for most learners, involves, to a the reorganization of linguistic material from an IL to id particular TL.

5.4 The fourth problem is: (a) what are the relevant hypothesized latent psychological structure within w lingual identifications exist and (b) is there any evid existence of these units? If the relevant data of the ps second language learning are in fact parallel utteranc linguistic systems (NL, IL, and TL), then it seems to m to hypothesize that the only relevant, one might say, 'psy real', interlingual unit is one which can be described sim for parallel data in the three systems, and, if possible mentally induced data in those systems. Concerning underlying linguistic structures, we shoul



in an interlingual study, if further replicated in other situations, would provide very powerful evidence for th the whole realizational unit as well as for its candidacy of realizational structure in interlingual identification. Concerning the notion of relevant units on the p level, it seems to me that Briere (1968) has demonstrated data there are several relevant units. The relevant units d correspond to known linguistic units, but rather would the sounds involved; sometimes the taxonomic phonem but the unit in other cases seems not to be describab linguistic terms. Briere evolved an experimental techn imit~ted to a large extent actual methods of teaching a apphed structural linguists; listening to TL sounds, i~itation, use of phonemic transcription, physiologic nons, and so on. If I may be allowed to reinterpret Bri seems to me that he has been working, in another situation, with exactly the three systems we are disc NL, TL, and IL: first, NL utterances which were h utterances in American English; second, TL utterances actual utterances in the 'composite language' Briere s utterance having been produced by a native speaker Arabic, or Vietnamese; third, IL utterances which utterances produced by native speakers of this NL when to produce this particular TL norm. Regarding the sou /rJ/ in his TL corpus, the unit identified interlingually three systems is the taxonomic phoneme defined dist within the syllable as opposed to within the word (Brie 73). For other sounds the relevant phonological unit is n nomic phoneme, but may be based on phonetic paramet which, he says, are probably not known (ibid, pp. 64 an If these units in the domain of interlingual identificati necessarily the same units as those in the native speak then where do they come from? An interesting bit of about native speaker performance units is provided b (1967, p. 335) who states that searching for' the unit' in na speech-perception is a waste of time. Alternative un available to native speakers, for example under noise co While other explanations are surely possible for the w fact that n?ise conditions affect performance in a second and sometimes drastically, we cannot ignore the possibl of Haggard's intriguing suggestion: that alternative lang


events which are to be counted as relevant data are not obvious. 2) These data have to be organized with the he theoretical constructs. 3) Some theoretical constructs relevant to the w 'adults' actually learn second languages are: interlingu tions, native language (NL), target language (TL), i (IL), fossilization, syntactic string, taxonomic phonem feature. 4) The psychologically relevant data of second langu are utterances in TL by native speakers, and in NL second language learners. 5) Interlingual identifications by second language lea unites the three linguistic systems (NL, TL, and IL) psy These learners focus upon one form of the TL. 6) Theoretical predictions in a relevant psycholog language learning must be the surface structures ofIL se 7) Successful second language learning, for most le reorganization of linguistic material from an IL to id particular TL. 8) There exist five distinct processes which are centr language learning: language transfer, transfer of trainin of second language learning, strategies of second lan munication, and overgeneralization of TL linguistic m 9) Each prediction in (6) should be made, if possibl one of the five processes in (8). 10) There is no necessary connection between relev linguistic theory and linguistically relevant units of a p second language learning. 11) The only linguistically relevant unit of a ps second language learning is one which is identified i across the three linguistic systems; NL, TL, and IL. 12) The syntactic string is the unit of surface struct and part of the unit of realizational transfer. 13) The taxonomic phoneme is, in the case of some unit of interlingual phonology, while in other case linguistic unit seems relevant. 14) There exists a latent psychological structure, i.e formulated arrangement in the brain, which is activate an adult attempts to produce meanings, which he ma second language which he is learning.



was read at the Second International Congress of App Cambridge University, 1969. 2 It is not unfair to say that almost all of the vast literatur relate psycholinguistics to second language learning, whet linguists or psychologists, is characterized by confusion be a second language and 'teaching' a second language. ( Jakobovits, 1970, p. IX.) This confusion applies as well t cussions on the topic one hears. For example one migh 'psychology of second language teaching' and not kno speaker is referring to what the teacher should do, what th do, or both. This terminological confusion makes one reg as to what is being claimed. 3 The answer to this question is not obvious since it is w theoretical considerations help point the way to relevan example, Fodor (1968, p. 48), ' ... how we count behav available as a description depends in part on what concep our theories provide .. .' 4 'Adult' is defined as being over the age of 12. This notion Lenneberg (1967, e.g. pp. 156, 176) who claims that af puberty, it is difficult to master the pronunciation of a s since a 'critical' period in brain maturation has been pass guage development tends to "freeze" (ibid, p. 156). 5 First pointed out by Harold Edwards. 6 See Lawler and Selinker, where the relevance of counterfac of second language learning is taken up. 7 Chomsky (1969, p. 68) expresses a very similar view; ' ... i nized that one does not learn the grammatical structure guage through "explanation and instruction" beyond the m rudiments, for the simple reason that no one has enough ledge about this structure to provide explanation and inst Chomsky gives as a detailed example a property which to grammar, that of nomina lization (Chomsky, 1969, pp. see no point in repeating Chomsky'S detailed argument show that a successful learner of English as a second lan have learned to make the judgements Chomsky describ planation and instruction'. 8 We have also idealized out of consideration differences bet learners which makes this framework quite incomplet second language learning that does not provide a central dual differences among learners cannot be considered Lawler and Selinker for a discussion of this tricky quest profiles of idealized learners who differ one from the other types of linguistic rules and types of meaningful performa language.

9 There has been a great deal of misunderstanding of this taking an antimentalist position here. Neither am I ru a-priori basis perceptual studies in a second language. How

52 L. SELINKER 21 Ian Pearson (personal communication). 22 Elaine Tarone (personal communication).

23 That is, what Corder refers to as the leamer's 'built-in 1967). 24 Example taken from Tom Huckin (personal communic

25 Example from Briana Stateman (personal communicat

26 The drive a bicycle example given in section 4 may, in fa (see Jain, 1969, p. 24). 27 Example from Sol Saporta (personal communication).

28 As was pointed out in note 7, Chomsky (1969, p. 68) al to provide native-speaker-like grammaticality judgemen 29 Note that this reactivation may be the only explanatio individual who learns any part of a second language Cheryl Goodenough (personal communication) has qualitative split between the 5 % who succeed and the language learners. Since in this paper we are not concen in a second language, as one would in the teaching app attempt to isolate the latent psychological structure wh any leamer, the system underlying attempted producti where the total effect of this output is clearly non-identi sized norm, then resolution of this issue should not aff The importance of isolating this 5 % is the speculation tha may not go through an IL. Reibel (1969) stresses the role of the latent language s language learning by suggesting that it is only when learners do the wrong things that they do not 'succeed explain differences between adult learners, not in terms o innate learning abilities, but rather in terms of the way applied'. Kline (1970) attempts to provide a point of Reibel's views and mine by suggesting that any reorgani identify with a TL must use the kinds of capacities an describes. A different opposing view to the perspective of this p sented by Sandra Hamlett and Michael Seitz (persona who have argued that, even for the vast majority of learners, there is no already formulated arrangement ex but that the latent psychological structure alluded to partly at least, by strategies which change up to the age with an individual for the rest of his life. There seems t critical empirical test for deciding between these two al 30 It is important to bear in mind that we are here working 'interlingual identifications' and thus are in a differe domain (Lawler and Selinker) than linguists who work the 'ideal speaker-listener' (Chomsky, 1965).It seems to m in the psychology of second language learning are in th tion of the language teacher who, Chomsky (1966) ad



33 The fact that Haggard is concerned with alternative elusive in larger units has no bearing on the issue und section.


LT: Target Language Ls: Source Language La: An approximative system Lai ... n: Indices referring to systems at succ proficiency.

In identifying a specific type of La' the name of the of the LT: thus 'German-English' refers to an La speakers of German communicating imperfectly in Our assumption is threefold:

1. Learner speech at a given time is the pattern linguistic system, La' distinct from Ls and Lr structured.

2. La's at successive stages of learning form an Lai ... n' the earliest occurring when a learner first Lr, the most advanced at the closest approach of L the achievement of perfect proficiency, is rare for

3. In a given contact situation, the La's of learn stage of proficiency roughly coincide, with major v able to differences in learning experience.

The speech of a learner, according to the assump ally organized, manifesting the order and cohesiven although one frequently changing with atypical rap to radical reorganization through the massive in elements as learning proceeds. As such, learner s studied not only by reference to Ls and Lr but in well. From the point of view of the history of LT undoubtedly correct in assigning interference in bilinguals, which he likens to 'sand carried by a parole of Lr along with other accidental and trans unincorporated by the community of LT speak communal language system.f However, from the po contact situation proper, to regard these same feat interference implies, exclusively as intrusive Ls elem the normal flow of LT-a kind of hiccough view of -is less rewarding, following the hypothesis, than v in terms of the learner system to which they pertain


tenders and other groups with frequent but circum ments to communicate with foreigners. 5 The term learnerpidgin can be applied to systems often employed by language students who have att the target language without mastery of its fundame arrived at a stage in instruction where attention ha from form to content. Not only do teachers often c of this system, but even participate as users (the follo were observed in a language classroom: Arabic-E Same? [i.e. Are the two wordspronounced in the same Same.; Teacher: Short answer [i.e. Use the short a conversation very good.) This variety of La is also fre other L, speakers when communicating with n apparently, even sometimes with other native speak

'Another brandy,' he said, pointing to his glass was in a hurry came over. 'Finished,' he said, sp omission of syntax stupid people employ w drunken people or foreigners. 'No more tonigh

Moreover these learner pidgins may be preserved types customarily designated as pidgins and creole systems usually incorporating Ls grammatical e lexical elements. An argument for the structural independence of source and target systems is the frequent and system in non-native speech of elements not directly attrib t., or LT. In the phonology, intermediate phon (Hungarian subjects in an experimental study often r /e/, for example, as [fe] or [se]).7 Similarly, 'intern resulting from the extension of the productive proce formations as go-ed are common in learner speech child language), and pattern confusion (observed classroom: Serbo-Croatian-English What does P occurs frequently in the grammar. More theoretically it can be argued that the de munication force the establishment of phonologic and lexical categories, and that the demands of ec imposition of the balance and order of a language Finally, there has been at least one attempt to stu directly." Customary descriptive procedures we characterize, in sui generis terms, the phonology of



other (even resulting in the creation, in some in dialects or languages as the varieties of English us speakers in India);'! and it is observable that communicate with each other more easily than w Moreover, La features are sometimes disseminated under special conditions (as was a trilled r substitute uvular phoneme, reportedly, among one group of E learners at Middlebury College), are sometimes c in La (English 161 is regularly merged with lsi in th certain schools in Germany), are sometimes trans generations (the children of native speakers of Yiddi while native speakers of English appear to frequent interference patterns from their parents' speech), a conventionalized in LT (during one era in Hunga speakers of Hungarian snobbishly replaced the Hu with the uvular variety of French to suggest prior kn language and hence higher social status; cf. also instances of substratum intrusion).

3 Reasons for studying L.

1. Direct and systematic examination of learner largely neglected. Classroom teachers, while aw patterns in learner behavior and often taking them their teaching, have rarely attempted comprehensive regularities within a linguistic framework. Cont specialists, on the other hand, often primarily techniques for establishing inter-systemic corresp been content for the most part to derive empirical s formulations from impressionistic observation Investigation of La data would, therefore, yield as it concrete information on learner behavior of hig classroom teacher in the planning of pedagogic stra

2. Such investigation is also a prerequisite for t both the strong and weak claims of the contrastive a a. The strong claim states that learner behavi on the basis of a comparison of Ls and LT' Ho immediately arising include: (i) different analys dictions, (ii) predictions are often ambiguou various levels of linguistic structure are interd


outset of learning, with Ls categories fusing with th parts throughout the systems. Actuall~, of cour~ exposure to LTis necessarily gradual. This fact entail contrastive analysis which can only be resolved by At post-initial stages of language learning, the prior conditions subsequent learning includes not onl knowledge of L, but his own recent experience in la tion-his knowledge of La-as well. He is no lon speaker of L s assumed by dialinguistic analysis, but a more recently acquired system. Thus the precept analysis itself force the inclusion of reference to La diction and elucidation of his subsequent learning b

4. Finally La's merit examination in their ow interest for general linguistic theory comparable on child language and on the other to the language of v types of speech disorder, as dependent systems for gradations toward specific languages but falling out dialectical and stylistic scope of these languages. 4 Summary and conclusion

Evidence suggests that the speech behavior of langua be structurally organized, and that the contact s therefore be described not only by reference to the n languages of the learner (L, and LT), but by referen system (LJ as well. Investigation of such learner systems is crucial to t of contrastive analysis theory and to its applicatio teaching. However, these systems also merit invest own right through their implications for general li In its present form, the contrastive approach see and account for learner behavior by reference to differences between Ls and LTand in terms of these this means (b) to indicate a strategy for language p ever, experimental and informal observation revea tions in the approach, in part because learner beha exhaustively described without reference to La. Theoretical and practical considerations therefo suggest the direct and systematic examination of such viewed within the general framework of the curren


Social Factors, Interlanguag Language Learning
Reprinted from Language Learning, Vol. 22, No.2,

An earlier version of -this paper was read as a gue Modem Language Center, Ontario Institute for St tion, Toronto. For comments on the earlier version ment to prepare a publishable version I am gr Selinker, Elaine Tarone and Kim Oller (University Seattle), David Lawton (Central Michigan Univ Professor R. LePage (York University, England). M here is of course tentative, and I would be grateful suggestions, or additional data.

1 Introduction A number of diverse contexts for second languag considered in this paper and the following question what conditions is standard English learned? What fa development of non-standard varieties of English, su English? What accounts for the divergence of lo English such as Nigerian or Indian English from Brit norms? Under what circumstances is more ma divergence likely to occur, such as is found in Creole generally the paper focuses on the choice of appropr the analysis of second language data. An area of resea which encompasses both psycholinguistic and dimensions. The concept of Interlanguage is pro analysis of second language learning and illustration the processes affecting language learning in the foll immigrant language learning, indigenous minori English, pidgin and creole settings, local varieti English as a foreign language. 64



personality, socialization etc., but a result of the socia possibilities made available for the group. Besides num and distribution, a number of other factors can af assimilation. These have to do with educational lev linguistic similarity to the mainstream culture, color, general factors which may determine the attitude majority group and vice versa. The evolution of lasting non-standard varieties language is a consequence of the perception by the the larger society, and a reflection of the degree to migrant groups have been admitted into the main dominant culture. Psychologists have been able between instrumental motivation, where the dom language is acquired primarily for such utilitarian getting a job, and integrative motivation, which d desire for or the perceived possibility of integration nant group. The former may lead to a functionally non-standard dialect of English. We can predict fo sort of English likely to be acquired by an immigra exclusively with his own language group and who ope catering largely, but not exclusively, to that language probably learn first to reply to a limited set of questio to manipulate a closed class of polite formulae, the some food items, and perhaps the language of si transaction. Whether he goes on to learn standar develops a functionally adequate but non-stand dialect of English will depend on the degree of in integration he achieves with the English-mainta structures. He may have very little control over the d action possible. If 100,000 such immigrants in sim reach only a minimum penetration of mainstream pow begin to perpetuate their semi-servile status, and begin among themselves, the setting for the generation of variety of English might be present. Where the learning of English is not associated penetration and upward mobility, but rather with occ economic subservience, we can expect language diverg outcome of contact with standard English. As an illu two extremes it may be useful to refer to the fate o Puerto Rican immigrants to America. A recent accoun German immigrants to Texas emphasizes that the



Whether instruction is in English or the native langua difference; rather what is important are the opportu thought available to the ethnic group themselves .. have provided the most significant evidence to de Increasingly, they have studied the relationship bet motivation and performance in school to his per society around him and the opportunities he belie there .... The crucial factor is not the relationship bet and school, but between the minority group and th Future reward in the form of acceptable occupatio status keeps children in school. Thus factors such community is socially open or closed, caste-like or n tory or not, has restricted roles or non-restricted rol for its minority group segment, become as important and other factors in the school itself, perhaps m (Leibowitz, 1970). The difficulties of some immigran result from more than simple questions of languag depend on the type and degree of interaction and ac able in the community for the immigrant group.

2.2 The Linguistic Dimension Having looked at the social background to immigra English we may tum to the linguistic problems asso description of their particular form and characteristic approach is to begin with the saurce language (Ls) language (LT) and to describe instances where the l differs from the target language as interference. Th inadequate, however, and obscures the nature of involved. Nemser proposes a three part approac leamer's approximative system as the intermediate the source and target language. 'An approximative deviant linguistic system actually employed by the le ing to utilize the target language. Such approximative in character in accordance with proficiency level; v introduced by learning experience . . . communica personal learning characteristics etc .... Our assum fold: (1) Leamer speech at a given time is the patterne linguistic system, La, distinct from Ls and 4 structured. (2) La's at successive stages of learning fro series, Lal",", the earliest occurring when a learner fi use 4, the most advanced at the closest approach o


No make any difference, but I like when I go be have too many time for buy and the little time we b to someplace and I find everything there. On being asked about trips to Puerto Rico, he gives:

I go there maybe about one and half and I find for me. But I can't work over there if I go alone a family here. I work I think 7 or 8 months in Pue

The concept of interlanguage as applied to this data a focusing on it as the learner's 'approximative system isolation of examples of language transfer, strateg munication, strategies of learning, transfer of trainin generalization. Language transfer is illustrated in sentence, which closely follows the structure of Puerto ish. In the first example however the syntax used cannot ly attributed to the effect of language transfer, since t English back into Spanish does not render the sentence Puerto Rican Spanish. To further characterize the features of the first sentence, we need to refer to the communication and learning strategies, and to overge Under communication strategies we may characterize features derived from the fact that heavy communicati may be made on the second language, forcing the lear what he has assimilated of the language into a means o he wants to say, or of getting done what he wants to g learner, isolated from close interaction with speakers language, may 'simplify' the syntax of the language i make the language an instrument of his own inte strategies affect both first and second language pe English. A child, not possessing the rule for nomi English, gave as a definition for fence: to keep the cow ... don't go out of the field.'

This process is seen in many of the constructions produc language learners (Richards, 1971b). Referring to a stud (Coulter, 1968), Selinker notes: 'Coulter reports syst occurring in the English interlanguage performance o Russian speakers of English, due to ... a tendency o second language learners to avoid grammatical forma articles, plural forms, and past tense forms .... Coulter



always a cline from minimum to full proficiency in E of Cocoliche, an immigrant interlanguage once spo by Italian immigrants in Argentina, Whinnom notes language was completely unstable in given individu was almost invariably continuing improvement i target language, and that acquisition of lexical, ph syntactic items must have been subject to chance, so of any two individuals was never identical. 'Neverthe as a whole, however ephemeral in given individuals broad a series of spectra it encompassed, was fai dictable, and was continually renewed in recogniza year to year and from generation of immigrant to immigrant' (Whinnom, in Hymes, 1971, p. 98). In interlanguages, language transfer, overgeneralization learning and communication, and transfer of train would appear to account for the basic processes invo for the analysis oflanguage learning in terms of the so under which learning and communication takes plac

3 Indigenous-minority Varieties of English In examining the social and linguistic dimensions English, we have seen that the size of the immigrant characteristics on dimensions of status, power, mobil wealth, can influence the variety of English acquired learning, whether the child learning his mother tong acquiring a second language, proceeds in terms of systems, but under certain conditions in second lan this interlingual stage may become the end point process, taking on a new role in in-group communica in ethnic identity and solidarity. The conditions un non-standard interlanguages are the outcome of cu contact are present to a greater or lesser extent i related situations. The language, educational and econ of many Mexican-Americans in the western and so States are well known, as well as the particular prob American and Canadian Indian groups. Ornstein an and Indian language questions, and notes the var clashing social systems which have contributed to maintenance in the southwest. Ornstein suggests that Hispano-Spanish and probably a Hispano-Anglo


culture and its values (Ashton-Warner, 1963). The rationalized as the child's failure, generated su cultu.r~l deprivation, restricted language developm cogmtrve deficiency, all of which are symptomatic fails to recognize the real ingredients of the child's e Recently, emphasis has been placed on the inter social and linguistic variables. Plumer points out th between knowing English and the ability to perfor cle~rly m~ch ~ore vital and complex for these groups pomt of VIewIS the same. If they see themselves locke anyway, then their motivation to learn English will ably low, especially if in so doing they risk cutting from associations they already have, namely their peer (Plumer, 1970, p. 270.) Wax et al. describe the pro drawal of Sioux Indian children from the white en pres~nted by the. school. They refer to the existence Enghsh, and point out that few Indian children ar English of the classroom (Wax et al., 1964). Darnel Indian community in Alberta, Canada, and the intera ~ree a~d English. 'Interference of Cree with the learn ISto~ sImp~ea model to account for the actual behavio Enghsh mistakes cannot be accounted for directly them to. d.ifferences between the structures of the t Rather It IS necessary to define the linguistic reperto Lake in terms of at least four, not merely two languag 1971.) She refers to Standard English, Cree Engli Cree and traditional Cree. Re~ent work by Philips highlights the role played Iearning styles and behavioral expectancies betwee child's home environment and the school, which expl ance to participate in many normal school activities (P Benton notes the role of the non-standard dialect as an self and group identification and of social perception type o~language spoken by children as reflected in their on rehable verbal tests, is often a guide to their likel performance, it may be only one of several factors whic the growth of language ability itself, and general schol m~nt. Ethnic differences also play an important par children from ~ minority or low status ethnic group may to control theIr. own des~ny than children from a dom They may find It more difficult to work with a teacher



The following samples of aboriginal English from A gest that this dialect is closer to the 'incipient pidgin' en of bilingualism, reflecting sharper social and economic of Australian Aboriginals than for comparable groups e well as omitting certain structures (verbs, auxiliaries, the copula), constructions such as the following are ob 'He bin go bump in you.' 'We bin give you a lot of shell, eh?' 'He big one, eh?' 'Ufla (we) got tee vee.' 'You know ufla (our) dog name?' 'Youfla (you) can have one.' 'Oh look at crocodile-la.' 'Look here-lao Him find this-la.' 'You mook mine-la.' (Alford,1970).

In studying the history of Cree English, Pine Ri Dormitory English and so on, it may be possible to us work proposed by Fishman for unstable bilingual soc language domain separation gradually disappears (Fish In the initial stages of contact between the native commu colonizing group, domain separation of languages English is required in certain limited roles and capacities conducive to the acquisition of a standard form of it (Le Hall, 1955). These are the conditions for the generation or a non-standard form of English characterized by st morphological simplification and by communication strategies and interference. As domain separation in l gradually disappears, English becomes an alternative t tongue, especially in family and friendship domains standard form of English now has functions related solidarity, spontaneity and informality. The standar encountered in the school and through contact with o formal functions, thus the characteristics of a diglossic obtain where complementary values L(low) and H(high realized in different varieties of English. This would app to some members of the Cree community described by is found with some monolingual Maoris, where the Maori English features increases according to the app of the domain. Features attributable to interlanguage p thus achieve stabilization through identification with



network and being associated with quite different s about their appropriateness (Stewart, 1962). In Jamaican situation, Craig proposes a model with ponent, an interlanguage, and the standard local var He uses the interlanguage concept to describe the ar creole and the standard which is the end point for t young people in Jamaica (Craig, n.d.). When the popu educational, economic, and social opportunities the c its distinctive features and becomes more like the future history of the creole hence 'depends on the soci creole vis-a-vis the standard, and the variability of the the culture' (De Camp, 1968). Cave gives details of th continuum in a creole setting (Guyana), the spectr varieties he illustrates ranging from that used by t Indian grandmother on a sugar estate, to that used b middle class urban dweller, to that of the speaker university (Cave, 1970).


Used by hlm/

/01 toold

Britons and a small number of per administrative posts imitating w social reasons

/01 to :ld hlm/ /01 to:l Im/ /01 tsl Im/ /0 tel Im/ /01 tel 1/ /0 tel 1/ /mI tel 1/

Important middle class in admin tions in government and comm professional men Ordinary middle class such as mercial employees, and teachers secondary education



Careful speech of non-clerical em assistants, hairdressers, who have but no or negligible secondary ed Alternative for 4 Relaxed form of 4 Relaxed form of 5

5 6 7 8

Rural laboring class-tradesmen, carters, etc.-who have had probab



maintained servile or semi-servile economic status fo which afforded them little chance to rise into the mid The degree and nature of contact with the upper la in the creole setting from that of the immigrant a minority examples we have looked at. The immigra interlanguages are characterized by settings whe language dominates, leading to continued opportu learning of English. Complete social assimilation i possible whereas in the pidgin setting, English was the resident or transient minority who were socially inacc the target language was not considered as a model Occupational, racial and social stratification, the pow restricted mobility of the slaves or laborers, meant t no solidarity between speakers and addressees, and that they were to become a single community. The rel of these factors in the other learning contexts we ha leads to interlingual generation which remains howev to speakers of English (Mintz, 1971, Grimshaw, 1971 4.1 Problems of description

Our basis for the analysis of immigrant and indi languages has been the concept of interlanguage, approximative system, characterized by transfer, training, strategies of communication and learnin generalization. Related concepts have been made use in their work, though we lack a complete illustration of creolization due to a relative lack of first hand data. linguistic adaptation as a product of language contac basis of pidginization and creolization, has however b from second language learning examples by a numbe interested in the explanation of creolization. Sama what we have referred to as communication and learnin discussing the grammatical adaptation and reduc creoles, drawing comparative examples from seco learning (Samarin, 1971). Whinnom illustrates how ce features of pidgin-simplification, and impoverishmen the source languages-could come about if a German schoolboy were forced to use French as a medium of co in a context where no other models for French w (Whinnom, 1971). Cassidy looks at contextual needs i situation and the use of linguistic adaptation to mee



of typical language contact situations, our understan language processes will surely be clarified by the expa creole studies, illuminating the factors involved both of a standard language and the dimensions that need to for in analysing the development of interlingual variet Typological description of the different settings detailed study of interlanguage processes in contact situ enable us to predict in instances where English comes i another language, whether the outcome will be a stan English, a non-standard form, or a new English-based 5 Local Varieties of English

The interlanguages we have looked at so far reflect dif of social, economic and political penetration of socie these structures being controlled by native speakers o language. Another related phenomenon must be reference to the generation of different dialects of English-based languages-the situation where these tures are maintained by non-native speakers of Engli phenomenon associated with countries where English natively but is widely used as a medium of instruction and informal communication. It is the case of Englis language in multilingual areas such as Commonwealth Pakistan, Malaysia, the Philippines, Fiji and so on. I English is widely used outside of native-speaking e local varieties of English have developed; Filipino Eng Nigerian English, Indian English etc. We do n phenomenon inside a political area where English is w natively. Thus we do not have 'French-Canadian E alternative to Canadian English. Deviations betw Canadians' use of English and Canadian English are sidered idiosyncratic, just as the immigrant varieties o said to be characterized by errors or by poor learnin But the educated Nigerian's or Indian's use of Engl differs from British or American English, is regarded acceptable way of speaking. In these countries English serves a variety .of formal uses (Allen Jones, 1968, Brosnahan, 1963, Fishman Hunt, 1966, Laver, 1970, Prator, 1968, Spencer, 197 used often or rarely, but no one has recourse to Engl



postmen, travel guides and bearers. The central po the register of the law courts, administration, and of of civil servants and teachers who learn English subject of study and who are able to make use of En in those restricted fields where English is used in In to which English is required in prestige settings as op functional settings is important as a standardizing countries, as it is of course in the other contexts we ha English trade pidgins represent the effects of purely for English, and trade or market English likewise rep functions for English, where it is not needed as an manipulate social behavior or the speaker's presti level in Indian usage Kachru defines as the languag are able to use English effectively for social contr social activities in which English is used in India. In the upper level may be defined by the educated uses and government officials, the middle level as that w heard in and around the secondary school, and the out-of-school uses. In West Africa the informal language may be a local pidgin. There is of course mu one country to another but the overall pattern of a cli ism with the local standard at the upper level (Stan Educated Nigerian etc.) is general. There thus appe distinct feelings of appropriateness in particular co different levels of English usage, as there are in creole appropriate use of creole or English. Ure notes that kind of English that is used as a linguafranca in the m not likely to be used in the classroom, and likewise no use classroom English in the market (Ure, 1968). Studies of the local varieties of English suggest the the concept of interference, and confirm the use approach which includes interference alongside su communication strategy, transfer of training, over and strategy of learning. Kirk-Greene, writing of Af suggests the need for study of the social role oflanguag language setting: 'Only by understanding both the s first language and the method by which English is acq the purposes for which it is used can we account f forms in bilingual usage.' (Kirk-Greene, 1971, p. suggests of the local varieties 'their grammar remains t English, with few important variations; their lexis t



in the contrast between Nigerian English and Yoruba English. Examples of mother tongue transfer in sta would be:

'I will pass by for you at 4' (for I will call for y 'How are you today? Fine. How do you do to? 'Close the light/open the light' (for tum off/on 'Go down the bus' (for Get off the bus). 'to lie on bed' (for to lie in bed) (Llamzon,

Examples of this sort of mother tongue/local-vari relationship can be found in many settings where Eng language. Many of the grammatical characteristics however must be seen as the results of overgeneral simplification and redundancy reduction etc. Exampl extension of isn't it as a question tag in Your brother isn't it, and such Indianisms as:

'I am doing it since six months'. 'It is done' (for it has been done). 'When I will come' (for when I come). 'If I will come' (for If I come). 'for doing' (for to + verb) (e.g. imprisonment fo character).

These reflect general tendencies observable in the a second language, and I have described them elsew interlingual interference (Richards, 1971a), that is, a subsystems, derived not from the mother tongue bu English is learned and taught. Many other characteristics of local varieties of assimilation of English to the cultural mores of the c describes these in terms of transfer, colloca collocation-innovation, and register-range extensi many examples from the Indian setting. All these fac make English part of the socio-cultural structure hence it is not surprising that those who would u standard instead of the local standard are regarded artificial and subjected to ridicule and criticism. The evolution of local varieties of English is thus a the adaptation of an overseas variety of English to m ment that a second language in use as a medium of b informal communication and not native to a coun


signs of incomplete learning. There is no room here f of deviancy. since the socio-cultural basis for deviancy in the foreign language setting. The learner is generall until he has 'eradicated' traces of his foreign accen practical purposes, this may not be possible due to th available in the school course. Limitations to the standard English in the foreign language setting are hen imposed limitations, which we encountered with th domestic dialects; in the foreign language setting l rather individual. reflecting personal differences in perseverance, aptitude and so on. There are no societa leamer's progress in English. In reality those wh accentless English in a foreign language context pr because of unique personal opportunities, rather th the school programme. These motivational factors have been emphasized The desire to acquire an overseas model of the for rather than a variety which is influenced by the acquisition was 'followed by the Japanese at the two of their history, when in the 6th and 7th centurie classes learned to read and speak Chinese in order th have access to the cultural riches of China, and agai 19th century, the educated classes learned Western that they might compete on equal terms with the Occid phenomenon was seen in the Hellenistic age, when peoples, that is, certain classes, among them, administr priests, litterateurs, and traders, learned the Greek Koi language. By "learned" we are not to understand tha who set themselves to write and speak Greek cam language without an admixture of native Greek ele point is that there was an effort on the part of the classes really to learn the foreign speech, to get at its ures, as well as to use it as a mere instrument of commu the foreigner. There were limitations in the use o language, but these were due to individual limitatio limitations' (Reinecke, 1969, p. 94). In analysing the English language performance of foreign language setting all differences between performance and an overseas model may be regarded or undesirable. While in the second language setting of an interlanguage may become institutionalized at th



a marker of transitional or terminal competence, the ference, simplification, overgeneralization, collocati collocation innovation, register-range extension, or learning and communication.

7 Conclusions I have tried to suggest here that a number of differe language learning can be studied with a common mod The notion of approximative system or interlangua the learner's systematic handling of the language da has been exposed, and the particular form of the l language will be determined by the conditions under takes place. Standard English will be the outcome learning when the learner learns in order to become the community who speak that form of English (e.g. immigrant), or in order to invite perception of the lear of equal status to standard speakers (e.g. the fo motivated to learn accentless English). Non-standard be the outcome of learning when the learner learns stances which hinder his becoming a member of the standard speakers. Self-perpetuating social stratificat with color, race and other ethnic indicators, lead standard dialect taking on a new role of ethnic identit ity. Educational planning which ignores this dime standard English is unlikely to achieve success. Pa resulting from a lack of integrative interaction speakers is reflected in modifications to the gramma English, and these are best described as aspects o generation, that is, as either language transfer, transf communication strategies, learning strategies, or o tion. The extreme case of non-integrative motiva language learning is seen in pidgins and creoles, whe process contributes to the separateness of the grou while maintaining solidarity at the lower level. Non-sta differ from pidgins in that in the former, the target lan to the learner. There is no sharp break in social comm rather a gradual merging. Progress towards standard creole setting reflects changing perception of class a consequence of social mobility. A local variety of E Indian English is influenced by the perception of En



And at what stage does the notion of interference genuine characteristic of the learner's speech, requiri explanation? Only detailed studies of the sort revie posed by Dulay and Burt, but with older learners, wi questions. Dulay and Burt's paper is followed by the two stud which they discuss in several places. Ravem's two pa some of the earliest and best attempts to trace the de syntax in a second language. Ravem presents valuabl to obtain data which suggests that second language creative process not unlike that of first language Ravem's studies of syntactic development of English children tend to confirm Dulay and Burt's hypothesis approximative systems are influenced by the order o of the same structures in first language acquisition of Dulay and Burt offer alternative interpretation for som data dealing with yes/no questions, where Ravem sug ian word order was followed in the absence of do. R word of caution however concerning the relationship order of first language and second language acquisitio He sees no absolute reason why the same order should and it is not certain from first language studies that t be the same for monolingual learners of English. T however are striking and add further weight to the L2 Ll acquisition hypothesis.


Briefly the contrastive analysis (CA) hypothesis sta the child is learning a second language, he will tend to language structures in his second language speec structures in his first language (Ll) and his second differ, he will goof. For example, in Spanish, subj dropped, so Spanish children learning English shou Wants Miss Jones for He wants Miss Jones. The L2 acquisition = L 1 acquisition hypothesis ho ren actively organize the L2 speech they hear and ma tions about its structure as children learning their firs Therefore, the goofs expected in any particular L2 pro be similar to those made by children learning that as their first language. For example, Jose want Miss J expected, since first language acquisition studies ha children generally omit functors, in this case, the -s third person singular present indicative. Each hypothesis contains two levels: the level of p level of process. The level of product describes the ac example, the CA hypothesis predicts that Spanish-spe will delete subjects, as in Wants Miss Jones, while the = Ll acquisition hypothesis predicts that the child functors, as in Jose want Miss Jones. The level of pro discussed in this paper in terms of 'theoretical assumpti for the product-the CA hypothesis offers a transf L2 acquisition = Ll acquisition hypothesis offers an organization theory. Throughout this paper, the p distinction should be borne in mind. We will discuss the assumptions (process level) of ea describe their consequences in terms of predicted go and cite empirical studies from a variety of languages the issue. No study we know of analyses children's ES the purpose of testing both the above two hypothe therefore be presented separately, and then a step to that would resolve the conflict will be proposed. 2 The Contrastive Analysis Hypothesis

2.1 Statement of the Hypothesis The last two decades of enthusiasm for contrasti foreign language teaching can be traced to Charles 1945, wrote:


variety of ways that all rely on the principles of Associa is, frequency, contiguity, intensity, etc., of stimu~us in the occurrence of the event that becomes a habit, The second assumption, which follows from the first interference theory in verbal learning and memory Interference theory until at least 195911 rested on the a the association of context and/or stimulus with respons new response to the same stimulus and/or in the same c require 'extinction' of the old association. O~h.erwise would prevail. The prevalence of an old habit In attem form a new task is called 'negative transfer'. Tulving and Madigan (1970) in reviewing the relev from '350 B.C. to 1969' comment that psychologists experiencing a 'revolution in interference theory' (p. 47 mainly by several studies which have seriousl~ qu nature of the operation of transfer in paired-associate example, Slamecka (1968) found that part of a free-r not facilitate the recall of the other part of the list; (1968) reported findings which were different from th by the notion of extinction of specific A-B associa explanation of transfer is based on t~e ac~eptance o habit formation. Since habit formation IS automat response, it is theoretically impossible to get away from of unlearning as an intermediate step to new response Thus, in spite of the findings that contradicted the pred extinction notion, the notion of unlearning has been re of the theory, but has been drastically revised.P Madigan summarize the revision:

Rather than referring to the extinction of both spec term and response-term) and general (experim and specific response terms) associations, [unlea envisaged as a kind of suppression of the w repertoire of responses in the course of second During learning of the first list, the subject limit selection to those occurring in that list. When learn the second list containing different responses of selective arousal' must be established. These c the suppression of the first list repertoire. When asked immediately after learning the second lis first list, the selector mechanism cannot shift back



Are the systems of the new language the same or the language I know? And if different, what is th

'Evidence for this is that a large number, by no mea leamer's] errors are related to the systems of his mo (Corder, 1968, p. 168.) In effect, though CA proponents seldom fail to stat base for their prediction of interference goofs, the natu seems to make little difference in what they predict a language leamer's goofs. 1 7 2.3 Evidence

(a) The Child-Adult Distinction in Language Acquisi Before we discuss the evidence for the CA hypothesis, i to point out that a major issue in the field is the differ children and adults in language acquisition. This differ extensively discussed by Lenneberg (1967) who draw areas of research to support the distinction. He repor toms of traumatic aphasia ('direct, structural and loca with neurophysiological processes of language', p. 15 under age 13 are reversible, whereas those that occu not. Non-deaf children of deaf parents who are expose language environment at school age learn to speak w deaf persons who regain their hearing after puberty a spoken language. Lateralization of brain function ar of puberty seems to be the physical correlate of these After puberty the brain becomes, as it were, less plastic less able to take on certain kinds of new tasks (See Ch Ervin-Tripp (1970) suggests a difference in appro learning based on previous processing strategies. T grouping together adults who have already learned oth and children; as opposed to monolingual adults.

An adult who has changed his linguistic system o ways-by adding new vocabulary, for exampl years may not have available ready strategies fo adult who has already learned other languages, o is constantly in the process of reorganizing h system and adding to his storage at aU levels w different approaches to new input. . . . The mo sensitive language learner we can find is a young



however, it becomes clear that the phenomenon of Weinreich has documented and that of 'linguistic bo Haugen has documented are the same; and that this p quite different from that of first language interference by CA proponents and described above. The differe seen when we compare Weinreich and Haugen's defin ference, types of empirical evidence, functions of int conditions for interference with those of the CA hyp

Definition of Interference Weinreich defines interferen ' ... those instances of deviation from the norms of e which occur in the speech of bilinguals as a result oftheir with more than one language, i.e. as a result of contact .. .' (p. 1). Haugen defines linguistic borrowing as '... an exam diffusion, the spread of an item of culture from peo Borrowing is linguistic diffusion, and can be unambigu as the attempt by a speaker to reproduce in one lang which he has learned in another (p. 363) ... it is the l learner that is influenced, not the language he learns=? The CA hypothesis, on the other hand, states that due to unfamiliarity with L2, i.e. to the learner's not h the target pattern, and is manifested in the language

We know from the observation of many cases t matical structure of the native language tends to to the foreign language ... we have here the m difficulty or ease in learning the foreign languag structures that are different will be difficult. (La 58,59) Further, Weinreich's definition of interference is which language was learned first:

Throughout the analysis of the forms of linguistic conventional terms like 'mother tongue', 'first' 'native' language were avoided; for from the st of view the genetic question ... is irrelevant. (p.

On the contrary, the native-foreign language distinc to Lado's statement. Types of Evidence

Weinreich's evidence comes fro



Haugen also mentions the 'deliberate use' by a bi translations 'for the sake of enriching his language' seems similar to the notion of ,foregrounding' suggest and Hernandez (in press) in their discussion of Chic English.) This use of interference structures is quite differen notion of interference structures as unwanted forms learner cannot help but use. The use of interference cording to Weinreich and Haugen, is motivated by so the CA framework it is uncontrollable because the not yet acquired the required L2 habits.

Conditions for Interference Weinreich's observation tions between a bilingual and a monolingual, and betw leads him to state that when both speakers are bilingua runs rampant in both directions; when one speaker i and the other bilingual, interference in the bilingu 'inhibited' .

When the other interlocutor is also bilingual, the of intelligibility and status assertion are drastic Under such circumstances, there is hardly any ference. (p. 81) Haugen writes:

Linguistic borrowing . . . is something that h wherever there have been bilinguals. It is, in fact without the existence of bilinguals, and apparen where there is any considerable group of bilingual

The CA notion of interference is predicted in q circumstances: the less of a bilingual the speaker interference there will be when he attempts to comm speakers of the target language. In summary, it seems that the work of Weinreich although fundamental to research in language shift, d to the phenomenon of first language interference tha CA proponents are concerned with. What other evidence is there then? (c) Adult Studies Jakobovits in his survey of foreign language learning



(p. 16). Wolfe (1967) has more evidence for this pheno is no explanation for this goof type within the CA fra Other research on adult L2 goofs (Selinker, 1971; S Burt and Kiparsky, 1972) also focuses on non-contras (d) Child Studies

Though there are no goof analysis studies per se on ch we have extracted what seems to be relevant info studies by: Kinzel (reported by Ervin-Tripp, 1970), and Ravem (1968). A six-year-old French child studied by Kinzel prod pronouns that reflected French agreement rules:

# She is all mixed up. (pendule: feminine in Fren English) #1 got her. (serviette: feminine in French, neute # Who likes them?(epinards: countin French.mass

Valette (1964) incidentally reports that in her nine m tional study of L2 French development in a four-year-o child, the only instance of English transfer was h substitution of attendre pour for attendre in the sense o (No explanation or discussion of transfer is made in t Interpretation of this data is premature since we ha French L1 acquisition and it is possible that children lea as their native language also make those goofs. Ravem (1968) documents L1 goofs in the English 6!--year-old Norwegian child which reflect Norwegian # Drive you car to-yesterday? # Like you ice cream? # Like you me not, Reidun?

This goof type occurred before the child had acquire hypothesizes that the cue for English yes-no questions the child had not acquired, and since inverted word o quent and important question cue in Norwegian, wh have do, the child used the Norwegian question cue. reports, however, that the child's goofs in his acquisition and wh-questions were definitely not traceable to (This is discussed in 3.3.1 below.) These three studies are the only child L2 studies we present data that can be useful in our discussion.P



come a new interest in second language learning rese comparing L2 syntactic development in children language acquisition findings. A statement from Da the scene:

. . . encouraging results in the search for unive language development have led us to explore th similar phenomena in the learning of a second la

The general hypothesis has been differentiated thus specific hypotheses:

I. L2 syntactic development is characterized b sequence in which 'base structures' are learned first, the 'transformed structures' are acquired (Dato, 1970).25

2. The strategies of L2 acquisition are similar to acquisition, e.g. use of word order as the first syntactic of functors (Dulay and Burt, 1972).

3. The L2 learning sequence of certain syntactic similar to corresponding Ll syntactic development structures studied include wh-questions (Ravem, questions (Ravem, 1968), negation (Milon, 1972 and R and plural formation (Natalicio and Natalicio, 1971)

These hypotheses, interesting as they are, span more deal with here. Staying with our plan to discuss only t of L2 syntax from the point of view of goofs children acquisition = Ll acquisition hypothesis, which is state is as follows:

1. Children below the age of puberty will make syntax that are similar to Ll developmental goofs.

An example of this is the omission of functors whic number of specific syntactically unrelated goofs: lack lack of tensing, missing determiners, missing possessiv

2. Children below the age of puberty will not ma reflect transfer of the structure of their Ll onto the learning. For example, Samoan children should not tend to use order in possessive NP constructions (the reverse of order) when they are learning English.



focus of psychological research is to discover those p a learner uses to arrive at the production of gramm Psycho linguistic language learning strategies appear t of the interplay between linguistic complexity and plexity. Some researchers, like Dato, pursue the l plexity factor; some, like Bever in first language acqu the learning complexity factor. We are all, in one w grappling with the problem that Charles S. Peirce (19 his lecture on the 'logic of abduction' -that of searchin of mental organization that limit the class of possible child uses when learning a language. 3.3 Evidence

The first part of the hypothesis-that children belo puberty will make goofs in L2 syntax that are similar t mental goofs-has been shown for three syntacti wh-questions (Ravem, 1970), negation (Milon, 1972 1968), and plurals (Natalicio and Natalicio, 1971). English Ll studies have produced clear findings, mak parison straightforward. The sec~nd part of the hypothesis- that children be puberty will not make L2 syntax goofs that reflect synt~ctic structures-has been shown for the foIlowin EnglIsh-Welsh word order in adjective and possess structions (Price, 1968), Norwegian-English wh-quest 1970), and Norwegian-English negation (Ravem, 196 tional comparison of Japanese negation to Milon's (1 L2 data, and of Spanish pluralization to Natalicio an (1971) English L2 data, also confirms this part of the h Japanese-English negation and Spanish-English plurali (a) Wh-Questions

R. Rav~m (1970) ~onducted a five month study of his Norwegian speaking son Rune, using spontaneous recording plus a translation technique. The main obj study. was to test the derivational complexity hypoth questions. He found, as did Brown's group (1968) that foIlowing rough transformational sequence
Base? John will read the book tomorrow T-1 John will read the book when?


HEIDI C. DULAY AND MARINA K. BURT Ll (Adam, Eve and Sarah)

L no S -+ { not

Stage I:


t + Nucleus, or Nuc. { + no { no} no not # No wipe finger. # Not a teddy bear. # Wear mitten no.

# Not me. # Not dog. # Not cold S

Stage 2:





A ux neg -+ {neg y Neg

neg {Pred ux + Main Y neg


Aux neg


no { not



no not { no

{can't don't #1 don't sit on Cromer coffee. # He not little, he big. # He no bite you.

{no don Don't tell te #1 no queen # 1 not give # 1 no more

Japanese word order, with the verb in final position, is Ravem (1968) has similar evidence based on a parti his data. His son Rune did not produce the Norweg N+V+(N)+Neg, as in: Han arbeider ikke (He works not) Vi tok det ikke (We took it not).

Instead, he produced the English LI acquisition structu VP, as in: # I not like that. #1 not looking for edge. (p. 180) (c) Plurals

Natalicio and Natalicio (1971) add to our store with t the acquisition of English plurals by native Spanish Grades 1, 2, 3 and 10. They studied 144 males, 36 in half native English speakers and half native Spani using a test similar to the Berko (1958) 'wug' test. The both the native Spanish and the native English speakers I-sf and l-z] plural allomorphs before the /-iz/ (Berko



Unfortunately, in this paper Price did not comp acquisition to the Welsh L2 structures produced b speaking children so this evidence does not speak first part of the hypothesis. 3.4 State of the L2 Acquisition

L1 Acquisition Hy

The L2 acquisition = Ll acquisition hypothesis as s to date, explicitly refers only to the product level. A available evidence, discussed above, confirms the children's production of wh-questions, negation, plu order in NP constructions. With this empirical confirmation of the hypothesis, assumptions receive support, at the same time maki analysis theoretical assumptions even less tenable. However, two weaknesses emerge:

1. What can we say about those structures which responding L1 data analysis for comparison? No suggest the task for 'future research'.

2. There is evidence for interference structures, findings in yes-no question formation by a Norwegian English. To account for this, one might make the w that because of the limitations of natural data collecti reflecting, say, subject-verb inversion in yes-no questio been made by Adam, Eve and Sarah when Brown and were not there to collect them.

We would like to make a stronger argument, thoug require finding an explanation that accounts for both like goofs and non-interference goofs, and that would the process level of children's L2 acquisition. 4 Proposed Second Language Production Strategies

While much of the value of scientific hypotheses in their power to yield new and interesting empiri consequences, their ultimate virtue, to the extent true, is that they explain known or knowable facts them to be consequences of underlying princip generality. A scientific description therefore cann a systematic account of observable phenomena b a theory that purports to explain the phenomena,



neither reflects Spanish structure nor is found in L in English. 4.2 Goof Analysis (8) Data

Of the child L2 data presented above (section 2.3 data on yes-no questions can be categorized in o addition, we have some pilot data of English L2 Spanish-speaking children tape recorded by an E free conversation in her first grade ESL class. W collected by G. Williams (1971) during his Engl with native Spanish-speaking first grade children. T all the goof types in the data, along with an exam
Table I




Poss Pro + N number agreement not allowed in English, obligatory in Spanish Omission of obligatory how in English, optional in Spanish Use of infinitive for gerund not allowed in English, obligatory in Spanish

# Now she's putt

# I know to do a

# I finish to watch o'clock.

Verb-subject inversion not allowed in English, obligatory in Norwegian

# Like you me no

Table 2 t.l



Irregular plural treated as regular Two verbal words tensed; only one required Accusative Pro for Nominative Inappropriate Q form: no aux-NP inversion

# He took her tee # I didn't weared

# Me need crayon # What color it is



Ll acquisition findings, except for speculations w liberty to present in the final section. Therefore, a the Unique Goofs listed in Table 4 must await mo acquisition research. The Ambiguous Goofs in explained by both the CA hypothesis and the U a acquisition hypothesis; therefore, they cannot be evidence for either. The Ll Developmental Goof clearly accounted for by only the L2 acquisition hypothesis; therefore, they can be used to confirm i discussed further. We are thus left with the Interfer which, although they appear to confirm the transfe by the CA hypothesis, are, we think, explainabl acquisition = Ll acquisition framework. It is this is presented in this section, namely, that the Interfe reflect a strategy similar to one used by children a as their native language. Ll learners '... "iron out" or correct irregu language, and incidentally reveal to us the fact th learning is general rules of construction-not jus phrases they hear.' (Cazden and Brown in press example, children use the past tense inflection (verbs, such as #goed and #runned for went and been called 'overgeneralizations'. The explanation are instances of various types of overgeneralizatio types of overgeneralizations discussed below have in Ll acquisition data. So in a sense, we have now acquisition structures aside but have retained Ll proc in our explanation of L2 syntactic goofs. Spanish-English Interference-like Goofs # Now she's putting hers clothes on. # She putting hers pyjamas on. # She's gonna brush hers teeths.

These goofs look like they reflect modifier-noun num obligatory in Spanish but not existent in English. H unreasonable to hypothesize that these are ins generalizing the possessive -s from NP's which Tim's, Mary's. (This specific overgeneralization doe English Ll acquisition data, but is consistent w strategy.) It is also quite possible that the child was o



1. a rule's utility for rich expression with a minimum redundancy e.g. word order

Word order alone is sufficient to express the basic sem agent-action-object, genitive, possessive, locative, n in # Daddy dog for 'Daddy's dog'. This has been d analysis of children's telegraphese in their native lan 1971, Bloom, 1970), and by adult telegrams. In 12 production, word order is used to produce sentences like those in Ll, but also complex sentences long before the functors within clauses are acquired bring it to school.

e.g. a rule using a minimal number of cue speaker's semantic intention

Intonation is used alone in the beginning to signal q the wh- words are learned, they are used without obligatory aux-subject inversion, which is in fact if-then clauses, the obligatory congruence of tenses the if clause, future in the then clause, is omitted. We predict that this type of rule will be used earlies in missing functors and missing transformations. 2. the pervasiveness of a syntactic generalization e.g. the Principle of Transformational Cycle

The Principle of the Transformational Cycle roug certain rules apply to several clauses within a comp the same way that they apply to a simple sentence, e.g ment, agent deletion, reflexive, there-insertion.33 sentence types this principle is obligatorily viola indirect questions where the simple sentence question does not apply to an embedded question. We don't # I asked him when would he come. Instead we say: I asked him when he would come.

We predict that if a pervasive principle has been a in structures where its violation is obligatory will re the type mentioned above.



6 These studies include the learning of English by a J Norwegian child, and by Spanish children; the learn French by American children; along with relevant ad 7 See also Nickel and Wagner, 1968; Banathy, Trager Upshur, 1962; Lane, 1962; and the introduction to Language Series English Guide- Kindergarten 1970. 8 See Fries's Foreword to Lado, 1957,Lado, 1957,pp. 5 p. 124, Lane, 1962,Jakobovits, 1970, pp. 19+-229, an 9 See especially Rivers, 1964. 10 See Tulving and Madigan, 1970 for a review of the 1961; Postman, Stark and Fraser, 1968; Postman and treatment of the theory. II The year of 'Postman's optimistic assessment of the ference] theory ... at a major verbal learning confer Madigan, 1970, p. 470, referring to Postman, 1961.) 12 See also Siamecka, 1969, Melton, 1961 and 1967, and 13 See Postman and Stark, 1969; Postman, Stark and Fr 14 See for example Sapon, 1971, and Jakobovits, 1970, p 15 Italics added. 16 Some anthropological linguists do study the problem retention, however. 17 The theoretical assumptions stated at the beginning been the subject of intense scholarly debate. We ref Chomsky, 1964; Garrett and Fodor, 1968; Bever, 196 Garrett, 1968; and Tulving and Madigan, 1970. 18 Italics added. 19 Italics added. 20 By 'intentional' Weinreich apparently means that the s of two options available to him, i.e. he can use items languages. Whether the choice is 'conscious' or 'u significant. 21 To do justice to this statement in any way would requir of the phonological goofs, their level (deep or surface s relation to the syntax of the sentences of which these addition, the relationship between the phonological pro the syntactic processing system would have to be discus the scope of this paper. 22 Ervin-Tripp, 1970, p. 330. 23 We considered the study by Dato (1970) of America Spanish in Madrid. He includes some samples of Spa reflect English structure, but each of these samples Spanish Ll developmental goofs. In the absence of Spa data, there is no way to accurately categorize Dato's #grande oreja ('big ear') shows wrong adjective plac most adjectives follow the noun they modify, except fo


Language Acquisition in a Language Environmen

Reprinted/rom IRAL, Vol. VII2, May 1968, publishe Verlag, Heidelberg. 1 Introduction

The present paper is a report on a study of a Norwe child's acquisition of English syntax in a second la ment. The study was undertaken with my son as the was a student of applied linguistics at the Univers in 1966. It arose out of a general interest in languag se and more particularly out of an interest in re language acquisition to the teaching of foreign lang gartens and at early elementary school stages. Th about language learning the more likely we are to our teaching of a second language. However, the child acquiring his first language and a child le language, at a time when he already possesses 'lan to be so big that any direct application of our know the more so because our knowledge in the first place shaky. I hope that the present study will make a sm to filling the gap. Most recent studies of the acquisition of syntax cerned with the linguistic competence of the chil stages of their linguistic development and an effort h write generative grammars for these stages. The in ~een interested in the obtained data only to the exten light on the child's system of internalized rules for guage. According to the language acquisition mod N oam Chomsky, 1 a distinction is made between 'pe actual utt~rances-and the underlying 'competence formance IS based. Chomsky's basic tenet is the notio theory should provide an adequate characterizatio





The translation test, involving about fifty negativ tive sentences requiring an auxiliary in adult spee regular intervals. The object of the translation ex compare the utterances with the data obtained in f in order to get an indication of the validity of prom of this kind. The stimulus took the form of a reques like 'go and ask mother if .. .' or 'tell Ranny that . sentence provided Rune with less of a clue to the sy of his sentence than a direct sentence would have do further reduced by putting the sentence to him in Nor in such cases where do is used). By prolonging the t the stimulus and Rune's response it was hoped that stimulus as a clue would be further reduced. The translation experiment was supported by the obta in free conversation. There were some clear cases from Norwegian, but they were of the same kind conversation material. The conversation data were collected at four d starting on 31st December 1965 and finishing on 6 The translation test was given within a week of t recordings. It was seen to be an advantage to recor 3-4 week intervals rather than more frequently and

3 Some findings

Only negative and interrogative sentences have bee the study for analysis. However, some examples sentences were included for comparative purpose concern was with Rune's acquisition of do as a tense and Postal are correct, which they probably are, the do is to be a carrier of tense." Being semantically em appear as a morpheme in deep structure and the tas of English is to discover the particular function o carrier. This might help to explain the reason why doconstitute a particular difficulty for foreign learner this respect do has not the same status as the modal a behave, along with have and be, roughly in the s equivalent auxiliaries in Norwegian. On this basis w that Rune would acquire these auxiliaries more quic following examples show that this is in fact the cas (C.l),3 eating you dinner to-yesterday? (T.2), what y



have discovered how to realize it in English. Time rel times expressed by help of an adverb of time as sentences I singing now/yesterday/all the day (i.e. non-occurrence of -ing with such verbs as like and t that be occurs optionally only in the context of Vi indicate a beginning differentiation between the s gressive forms.

3.2 Negative Sentences In adult grammar do is used when the verb phrase another auxiliary verb. As with the modal auxiliar element, not, follows or is attached to do and not to below exemplify the similarities between the use of m and do in negative sentences in English as c Norwegian: I cannot come. I could not come. He does not work. We did not take it.

Jeg kan ikke komm Jeg kunne ikke ko Han arbeider ikke Vi tok det ikke (w

Since do is not yet available at Time 1, one pred that Rune, in keeping with Norwegian structure, le main verb and produces sentences of the form NP+ we find, however, are such sentences as I not like crying, I not looking for edge. The negative senten correspond to the pattern for declarative sentences insert not after the subject NP in our formula. 3.3 Interrogative Sentences

Th~ following types of interrogative sentences, all o do m. adult grammar, were studied: 1. sentences b question word (what, when, etc.), 2. sentences requir a~ answer, 3. negative versions of 2, 4. negative ques With why. Again we find a high degree of syntactic similarity and Norwegian in the use of modal auxiliaries and there is no equivalent to do as shown by the followi 1. What did he say? 2. Did you do it?

Hva sa han? (What said he Gjorde du det (Did you it?)



On the basis of this pattern we could make vario about the line of development for 1. and 2. That Rune, like Ll learners, makes use of the declarative sentence in Q-wh-sentences is, as we have a out, rather surprising in view of the inversion of subje Norwegian. It would be reasonable to expect both 1. out as What readingyou to-yesterday? and What doing happens with yes/no-questions, where typical examp you? and Like you food? Again we could speculate Rune in these types of sentence makes use of inversion signal from lack of a question word. Negative yes/nonegative why-questions are structured differently, but the respective affirmative sentences. These sentences in the Translation test only at Time 3. The Q-whyproved much more resistant to the do-transformation types of sentence included in the study. Typical senten introduction of the do-transformations are: I singing out yesterday. I not sitting on my chair. What you reading to-yesterday? Why we not live in Scotland? Drive you car to-yesterday? Like you me not, Reidun? 3.4 The Development of do

a Tense Marker

Do occurs from the beginning frequently in the co isolated verbs, where it is probably a lexical variant of know, I don't think, it doesn't matter. It is probably wit that it has spread by analogy to I don't will more, I don and I don't say something more. It appears also at elliptical sentence Do you?, a case incidentally where N a similar construction. I think we can safely say that t is absent from Rune's speech at this stage. The next occurrence of do is found at Time 2 in the most likely as a variant of you, pronounced [dju:]. Unf translation test for Time 2 was not recorded and it can be checked for pronunciation. However, the convers Time 2 have been carefully checked and all eight occ pronounced [dju :]. When What d'you like? was asked



4 Conclusion The study reported here was not undertaken to test hypotheses relating to certain theories oflanguage lea purpose was to conduct a study within the framewor syntax studies to find out something about developm as compared with first language learners. Recent first language syntax studies have shown exposed to a language at an early age internalize rules which they are able to generate sentences. Attempts h with some degree of success to give a characterization rules that are operative at various stages of developme have shown that a large measure of creativity enters i of language acquisition. The situation of the learner of a second language is c from that of the Ll child. The most obvious difference of the foreign learner is not to learn 'language', wh possesses and the knowledge of which must affect his a second language. The process of learning the se might therefore conceivably be qualitatively different. often exposed to 'primary linguistic data' in the se learner is, but rather to carefully graded language item small doses for a few hours a week. The present study has, I believe, shown what we w find, namely that a normal six-year-old child at all lev is greatly facilitated by the linguistic competence he alr through his first language. The six-year-old's greater m for a faster rate of learning. The first language, espec as closely related to English as Norwegian is, is a sou can draw on. The detrimental effect of first language i only be assessed properly after the learner has ac command of the second language. What is perhaps more striking is the extent to language acquisition in an environment where no form is given seems to be a creative process not unlike that o acquisition. The similarities between Rune and Ll developmental sequence of negative and interrogative in many ways more revealing than the differences. It does not follow from this that the appropriate m teaching a foreign language at an early stage is to expo to a 'language bath' and let them develop for themselv generative rules which ultimately develop into t


The Development of Wh-Qu in First and Second Langu

Learners *


Originally published as an occasional paper by the L University of Essex.

1 Introduction !l0ger Brown (1968)1 reports the result of an analysis in the speech of the three children whose language d been studied by him and his associates at Harvard U analysis was made to determine whether or not the in the spontaneous speech of pre-school children formational rules of current generative-transformati also .figure in the child's competence, in other wor mediate hypothetical strings in a transformat correspond to stages in the child's development of Such hypothetical intermediates are not, usually, act forms an~ hence not available to the child for im ?ccurred in the speech of children at a certain stage It would suggest that transformational grammar h capture psychologically real operations, and it woul doub~ on an empiricist explanation of language ac these intermediate structures are not exemplified in th the child is exposed to.

2 The Grammar of Wh-Questions Table I presents examples of types of sentences tha cerned with in this report, here given in their adult


I a~ in~ebted to my superior, Dr Terence Moore at the University of Essex, f~r critical co~ents and advice on however, ~ot responsible for the VIewsexpressed and mterpretation of Professor Brown's views.




WHAT will by (1) be changed into WHAT John will read? and further by (2) into WHAT will John read?

If morpho-phonemic rules were applied to the transformed strings, sentences like the following w John will read what? John will read the book when? Who will read the book?

If we assume that the stages in language acquisi transformational derivation in transformational would expect to find sentences in the child's gra basically of this form. I shall therefore refer to it as thetical Intermediate (H.I.!). It should be noted, however, that the H.I.l is no Brown's 'Occasional Question' (1968, p. 279). In Br the Wh-word is spoken with heavy stress and rising gives the following example. 'If someone said: "Joh telephone book" one might respond "John will rea would not in my analysis constitute an example question it is semantically different from a normal the constituent that is questioned is already known a expresses a disbelief or astonishment," The sentence to find, if our assumption were correct, would be o interrogative stress and intonation. The next stage of development would be one in phonemic rules were applied to an underlying string but in the absence of transposing. This would result i What John will read? When John will read the book?

I shall refer to this type as the second Hypothetica (H.I.2). It corresponds to the Hypothetical Intermed a general feature of the grammar of the three chil Brown and his associates. Diagram 1 is meant to illustrate some of the feat our discussion of the underlying structures of Wh-q might be represented in a child's grammar. The Q



fact that a sentence is to be interpreted as a question; to be questioned has received the feature + WHo If the the NP node has the additional feature + human, the be chosen will be Who; if it is - human, the lexical item Similarly, the features associated with the +WH of will ultimately generate the lexical items Where, W Why respectively.

The rules of the grammar will up to this point gen of the form illustrated by Table 2.
Table 2 (a) (b) (c) (d) (e) (f) Daddy come when? Mummy saying what? Adam goed where? Rune did go why? Mummy can find Reidun how? Eve are doing what there?

Brown's prediction that sentences of this type would emerge after the initial pre-transformational stage wa by the result of his analysis of the material. 'In fact, th next step-at least not the next step we could see, t performance. Occasional questions never became fr children, and the first ones appeared somewhat late III. This may be entirely a matter of grammatical p what the children found "occasion" to say rather than or what they were able to say. As we shall see, the occa these forms are used are special and may simply not the child' (1968, p. 284). This explanation seems reaso the occasional form is a true occasional form w supra-segmental features and semantic connotation implies. Our main hypothesis, however, would predi the H.I.l type as the normal interrogative form a development, but without special stress and inton superimposed on them, e.g. such sentences as: Yo Mummy?, Eve doing what, Adam?, Adam goed w Presumably no examples of this kind have been fou which means that neither Brown's hypothesis abo questions nor mine about H.1.l has been confirme study. 5 The story is different for our H.1.2, which is the s



made them more reliable." The recordings were m intervals up to July 1969, each interview averaging o The interlocutors have been either our eldest, bilin my wife and myself, native English-speaking adults, The interviews have been arranged without bein structured, as I wanted the speech during the ses spontaneous as possible. The most rewarding situatio to amount of data have been with peers or members o we were conscious about what we were looking for, the conversations in different directions and thereby of the kind we were interested in. The translation test useful instrument for eliciting the types of sentences their validity has been supported by obtained utte conversation. The collection of material was resumed after a months, when the family were in Norway. Intensive r made immediately before and after the break, as I w the degree of forgetting that might have occurred months. This, however, is only an incidental aspec Reidun seemed essentially to have caught up with h July 1969, but the collection of material has continued systematic fashion. I have only recently started analy and the present report is based on only a portion of

4 Results and Discussion There is no prima facie evidence why a child acqu language should go through a similar development a learners. Nor incidentally, is it obvious that essentially is followed (the same strategy chosen) by all learners tongue. Although a comprehensive study, the investig and associates is to my knowledge the only comple study of interrogative and negative sentences to da prises only three children. It is therefore quite conce picture will be more varied as more studies have been Nevertheless, as the actual data in Tables 4 and 5 striking similarities between my own material and study by Brown and associates (Table 6), and it is not is altogether accidental. (If one set out to search for di would probably be equally striking, but less surprisi yet found any examples corresponding to our H.L



Time 4 (c. 3 wks later) (T) (T) (T) (T) (T)

What you did after Ranny go to bed? (In response to: did.) What did you more-that night? What did you talk to them? (say/talk about) What do you doing to-yesterday? What did you do to-yesterday? When you go to bed? (past reference) What do you going to do tomorrow? Who you talking to to-yesterday? Why not that go up? (that window) Why not Mummy make meat today? (from Nor. mat Table 5 Wh-questions from Reidun

Months: weeks of exposure to English 3: I 4:0 4:4 5: 3

6:0 6: 2

6: 3

8: 0

What this? (What colour is this) Where find it? (in response to: Ask her ... where you can fi Where jeg kan jeg find it -apples? (jeg = I; kan = can) Whats that? Hvor er-my Mummy? (where is) Whats her doing? Whats er her doing? (er = is/are) Where er hers Mummy? Why that the bed er broken? (prob.: why (is it) that) What that is? Why it-Humpty Dumpty sat on a horse? What call that man? What-name that man? Why that man have that on? Whats that-is? Why-uh-him have got like that? (a jacket like that) Why her don't stand there? Hvem er that? (who) Whos that? Whos that is? Whosis.that is? What her going to make? (or: goingto?) Whats her baking? Which one you want? What you want? What do you want? (or: doyou) Who is that? Why you can't buy like that shoes? Where is it, then? Whats are they?



equivalent, question forms; the relation of the Wh-w pro-f~rms, such as it and there; and to learn the m constituent, such as NP, is an interesting and plaus show that there is more in the language data the chil th~ meets the ear and that one might profitably look might be exemplified in the input before one jumps conclusions. However, the discussion would appear rel.evant to L2 acquisition, where the abstract categorie ships are already known to the child through his first learning task of my children may have been more learning how these relations, or whatever, are realize language. It ~ppears, ~h~n, that transformational grammar stage m the child s development of Wh-questions; bu same as saying that it has captured a psychologically The latter is what Brown, in effect, assumes. 'We be ~uestions, in general, were derived by a single preposin tion out of underlying strings with dummy elements 286). Th~ preposing transformation cannot be given a stat~s without at the same time assuming some kind o reality for the underlying string (H.I.l). If this wer there would be no preposing operation to carry ou hypothesis unfortunate that no sentences have been could be said to be either an actualization of the occa or of our H.l.l. While entertaini~g the hypothesis of a preposi Brown goes on to dISCUSS the evidence for and again questions in his corpus fall into two classes, one 'Preposing Weak' and another which he calls 'Prepos general characteristic of children's early speech is t inflections and of minor word classes (functors), w what has been referred to as 'telegraphic' speech (Bro 1.963~. The class for which the evidence for a preposin tion IS weak consists of children's sentences that co learned as.a reduction of adult speech to 'telegraphese the following examples, where the omitted words are What (do) you want? How (will) you open it? What (is) his name?

The second class of Wh-questions are those for whi



that what in What you want? is the direct object of wa related to an indefinite pro-form in the declarative sen 'some thing'. It seems to me that it is quite legitimate knowledge and at the same time propose hypotheses the order of constituents in Wh-questions of the pre the child's speech. One such hypothesis suggests itself, namely that remains in initial position and is followed by a 'n retains the word order of a declarative sentence ac child's grammar at any time." This hypothesis does account for more than word order; nor is there mo seems to me, in Brown's class of weak evidence. Th preposing is weak exactly because it allows for an terms of selective imitation of an adult model and leav such questions as how the child is able to question stituents or can see the relationship between disco stituents, such as the verb and the direct object in W The Preposing Weak class constitutes weak evi posing only if the 'strong' evidence that Brown allege one is willing to concede an alternative explanation terms of an 'underlying grammatical network' for Weak class, one should do this also for the class of st On the alternative hypothesis (Wh-Nucleus) the stron turn out to be no stronger than the weak evidence; does not, in fact, distinguish between them. Sentence wants? or Why you can't open it? cannot be deriv models alone, but the 'nucleus' of the sentence (or w it) preserves the word order of the declarative sente child acquires inflections and auxiliaries in declara these will also-although usually somewhat later complicate the analysis-be incorporated in Wh-que This alternative explanation does not affect the hypo child reduces adult speech in a systematic way and i rules on the basis of this reduction; but by adopting re only criterion, one is forced into setting up a separat posing Strong evidence, which is not required by hypothesis. I like to believe that Tables 4 and 5 show, fairly con the intermediate sentence type without transposing ( feature also of my informants' acquisition of Englis language. That they already knew the transposing t



tree. It contains some morphemes that are lexical an realized as inflections, for example, present or participle (en), and present participle (ing). There combinatorial possibilities, some as complicated Past Modal have en be ing

Even without considering the cognitive proble acquiring tense and aspect, the linguistic mechan are complicated enough, and it is therefore to be e full range of auxiliary morphemes and their distribu in developing. There is probably room for some ind in the order in which children develop the AUX node picture is from no auxiliary at all through stages of to adult grammar. I have not yet done any detailed development of the auxiliary in my children, but it s in many important respects the development i learners. The main function of do is to be a carrier of tens the learner of English is to discover this particular Since the use of do is specific to English, the second is faced with very much the same learning proble included in Diagram I as a verbal element of A shares some of the distributional characteristics of could therefore on this basis predict that Wh-quest stage would have the form of sentence (d), Why Table 3, namely: WH NP Tense-DO V

When do was introduced in affirmative Wh-ques carrier, Rune used inversion, for example: What did you do before you get to bed?

With Reidun the situation is not clear-cut. The material that has not been analysed, and although affirmative Wh-questions have inversion of do and the translation tests show isolated examples of no

(8:4) Where we did livdfor we come here? (for

I would have tended to interpret the few examples of do and subject NP in Wh-questions found so performance mistakes had it not been for Reidun's



sentence, which he largely echoes (using his own g which he preposes Why or Why not. As for the two o the study, they did not start producing Why (not they had reached the stage when they could give appr to them. This appears, from a survey of parts of my corpu the case also with Reidun. Early Why-questions di appropriate response.

RUNE: Why do you put the telephone on (3:2) REIDUN: Yes.

Reidun's acquisition of Why-questions and th responses cannot in the same way as for first langu related to her cognitive development, that is, to explanation is'. She knew this, relative to her age, a Norwegian equivalents for some time. As Table 7 shows, all Reidun's early sentences lac first attested occurrences of Why-questions were in th exposure to English. One half-hour recording at 6: 3 27 Why-questions and 8 Why (not) -questions. In the there were three 'because' -responses and two embe clauses. The Why (not) -questions corresponded b Preposing Weak and Preposing Strong types. Since now acquired auxiliaries and inflections required b Strong class, a large percentage of her sentences w
Table 7 Why (not)- questions from Reidun Months: weeks of exposure to English 6:2 Why that man have got it? Why uh that horses have that-that on-foot? Why-that man are over there? Why her don't stand there? 7:2 Why that man take-hang clothes on the-on Why that go up? 7:3 Why you can't eat it? 8:0 Why I sitting there? Why Daddy hold me? Why we can't go to London now-today?



Father: Why you didn't go to Colchester? Reidun: Why didn't you go to Colchester? Father: Why she has got trousers on? Reidun: Why have you got trousers on?

Father: Why Mummy doesn't sit on the table? Reidun: Why doesn't Mummy sit on the table?

It appears, then, that the transposing operation too short period of time and seems to have affected both negative Why-questions with different auxiliaries s All Adam's early negative Why-questions we sentences preposed by Why not. The introduction of not has tentatively been suggested by Bellugi as a stage in the formation of negative questions, which have been responsible for the temporary use of doub the children, such as, Why not me can't dance ?12 A cited from Reidun show, the basis for her negative W Why followed by a negative nucleus, and I have fo negatives at this stage. Rune, however, produced negative questions of bo Why or Why not for example: Why you not come home? Why not that window go up? Nucleus Nucleus neg

Either type was produced throughout Times 3-5 in lation test, apparently in a random fashion. Althoug data on Why not-questions from Rune's first stay in there is supporting evidence in the-as yet unanalys the beginning of his second stay to suggest that they w patterns. Although the auxiliary do had appeared in questions at Time 4, there were no occurrences of do Why not- questions. This might be accidental, or due negative questions are more complicated, involvi transformation in addition. Rune's further development could have been based



Why you don't like and going skiing? Why you don't going to school to-yesterday? Why Mummy don't play piano now?

It is tempting to speculate that I have accidental intermediate stage in Rune's development of Why a productive rule between the last test in March termination of Rune's first stay in Great Britain in J does it suggest that the process of 'forgetting' has be learning-a regressive process?
-+ Why -+ Why

Why don't you like ice-cream? you don't like ice-cream? you not like ice-cream?

7 Conclusion The purpose of this paper has been to present some concerning the development of Wh-questions in children acquiring English as a second language and those of a similar study of first language acquisition tion has been somewhat biased in that I have chosen on the similarities between first and second lan Taking the age and maturity levels into consideratio that my children already know one language, the quite striking and not necessarily what one would e The findings have been discussed in the light of the forward by Brown (1968). Brown has been concerne ing or disconfirming a development of Wh-quest which reflects the transformational derivations in tra generative grammar, in order to find out if these m represent psychologically real operations. Brown is interpretation of the evidence and recommends that to have a second look at empiricist explanations, as throw light on the process of language acquisition. Although I think nothing conclusive can be said ab logical reality of the transformational rules discusse the transformational description itself has made it p testable hypotheses. Whether Brown is right or not conclusions is ofless importance. At the present stag child language development it is of interest to fin regularities are across children with regard to the ord


Idiosyncratic Dialects a Error Analysis

Reprintedfrom IRAL, Vol. IX/2, May 1971,published Verlag, Heidelberg.

What has come to be known as 'Error Analysis' has investigation of the language of second language lea taking the point of view in this paper that the lang learner, or perhaps certain groupings of learners, is dialect. This is based on two considerations: firstly, a speech intended by the speaker to communicate is.me sense that it is systematic, regular and, consequently describable in terms of a set of rules, i.e. it has a spontaneous speech of the second language learner has a grammar. Secondly, since a number of sen language are isomorphous with some of the sentenc language and have the same interpretation, then so the rules needed to account for the learners' langu same as those required to account for the target langu the learner's language is a dialect in the linguistic guages which share some rules of grammar are diale

Set of rules of language B

Languages A and B are in a dialect relation. (I am not whether or not all languages can be regarded as being i It is, of course, usual to apply a further non-linguisti language in order to establish its dialect status, namel be the shared behaviour of a social group, i.e. that 158

160 s.


All idiosyncratic dialects have this characteristic in some of the rules required to account for them are p individual. This has, of course, the result that some of are not readily interpretable, since the ability to inter depends in part upon the knowledge of the conventi that sentence. The sentences of an idiolect do not th the same problems of interpretation since somewh member of that social group who shares the conven speaker. It is in the nature of idiosyncratic dialects that the unstable. The reason for this is obvious. The obje normally to communicate, i.e. to be understood. If u only partial, then a speaker has a motive to bring his line with conventions of some social group, if h instability accounts for part of the difficulty expe linguist in describing idiosyncratic dialects. The da description is made is fragmentary. This means verification procedures required in the construction grammar are not readily available. The other difficulty the linguist experiences is tha interpretation on some of the sentences of the di interpretation, of course, analysis cannot begin. The language of a second language learner is not t idiosyncratic dialect. 'Error Analysis' is not applicab language of second language learners. One class o dialects is the language of poems, where this canno for wholly in the terms of the rules of some social dial (1965) says: 'given a text like Cummings' poem "An pretty how town" containing sequences which resist grammar of English, it might prove more illuminating a sample of a different language, or a different dialect, English' (My italics). That the language of this poem is idiosyncratic is because of the difficulty of interpretation. It is s Thome's approach to the analysis of the language essentially that of 'error analysis', a type of bilingua That is, he attempts to discover the rules which wou the idiosyncratic sentences' in terms of the same syn uses to account for the social dialect to which it most in this case, Standard English. The idiosyncratic sentences of a poetic text can pe




Language A

An alternative name might be transitional dialect, e unstable nature" of such dialects. I have suggested that it would be reasonable to syncratic sentences of a poet's dialect deliberately de writer is assumed to know the conventions of a social he deliberately chooses not to follow them. Similarly I that the aphasic's idiosyncratic sentences might be logically deviant since he too was presumably a speake dialect before his disease. We cannot, however, ref syncratic sentences of a child as deviant, since he, o yet a speaker of a social dialect; and indeed it is not goes to school) to call a child's sentences devian ungrammatical. For precisely the same reason I su leading to refer to the idiosyncratic sentences of the se learner as deviant. I also suggest that it is as undesirab erroneous as it is to call the sentences of a child errone implies wilful, or inadvertent breach of rules which, ought to be known. Whereas, of course, sentences ar precisely because the rules of the target dialect are n The only sentences in anyone's speech which could, justice be called erroneous are those which are the failure of performance. These may contain what are o of the tongue, false starts, changes of mind, and s (1948) refers to these as lapses. They may be the re in memory. A typical example in English would b problem which I don't know how to solve it' (Reibe estingly such erroneous sentences do not normally pr of interpretation. The reason that suggests itself for t may be, in any social dialect, 'rules for making m clearly, is a field for investigation (Boomer and Laver are not yet in a position, I think, to set up a fifth class o dialects to account for the regularities of erroneous


s. P.


a sentence as: 'And hearkened as I whistled the beside' is perhaps unique to verse in modern Eng accounted for by a convention accepted by all English is not true, however, of 'Up so many bells .. .' of Cu This is not part of the poetic dialect of English, is diff tation and I doubt whether the rules which account account for any other poetic utterances by any other idiosyncratic. The situation is, I think, different in the case of classes of idiosyncratic dialects. Aphasics do not form in any sociological sense, and yet there is strong evid that the idiosyncracies of their speech may be cla number of dimensions (Jakobson, 1956). No one w attempt to describe the speech of aphasics unless 'h some general statements of classification were possi of such investigations is to find what relations the the medical signs, symptoms, history and the set of account for the idiosyncratic aspects of the aphasic's Similarly, no one would undertake the 'study of acquisition unless he had reason to believe that al certain dialect environment followed a course of deve was more or less similar (Smith and Miller, 1968). T little point in describing the speech of a three-year-ol expected ultimately to throw light on the speech of old. Therefore, there is an underlying assumption th of all three-year-olds in a certain language environm certain features in common." May it be that the situation is similar in the case of second language? It is certainly the case that teache assumption that a group of learners having the same and having had the same experience of learning the se speak more or less the same interlanguage at any learning career, and that what differences there are c to individual variation in intelligence, motivation attitude. This belief is inherent in the notion of 'teach opposed to an 'individual', and indeed, it is difficult could proceed otherwise. Can we assume that such learners all follow a sim development in acquiring a second language? We cer can to see that they do. That is what a syllabus is f of the route the learners are to follow. But supposing i




poetic dialect, and not an indefinite pronoun, and ho and not an interrogative adverb. The first stage in 'error analysis' then is recognition We can enunciate a general law. Every sentence is t idiosyncratic until shown to be otherwise. As I ha learner's sentence may be superficially 'well-form idiosyncratic; these sentences I shall call covertly idi may also, of course, be overtly idiosyncratic, in that ficially 'ill-formed' in terms of the rules of the targ they may, of course, be neither. If the 'normal' i acceptable in context, then that sentence is not purposes idiosyncratic. If, however, the sentence ficially well-formed in terms of the rules of the targ nevertheless cannot be interpreted 'normally' in co sentence is covertly idiosyncratic and a plausible inte be placed upon it in the light ofthe context. We then a reconstructed sentence to compare with the original. sentence is, roughly speaking, what a native speak language would have said to express that meaning i.e. it is a translation equivalent. Let us take another possibility: that the sen idiosyncratic, that is, it is superficially 'ill-formed' a rules of the target language. We must then ask whe interpretation can be placed upon it in the context. I good, and we can proceed to make a 'well-formed sentence to compare with the original. If we cannot plausible interpretation of the overtly idiosyncratic our problem is much greater. Somehow or other we make a plausible interpretation. We can first see w ence to the mother tongue of the learner, we can a interpretation. If the mother tongue is not known t of that sentence may have to remain in abeyance learned more of the idiosyncratic dialect of the learn the mother tongue is known, we may be able, by a p translation, to arrive at a means of interpreting plausibly. If we can do that, then, by translating the sentence back into a well-formed sentence of the targ have available a reconstructed sentence which onc compare with the original overtly idiosyncratic learner. The end point of the process of identifying idiosy




production of a reconstructed sentence is two sen syncratic sentence and a well-formed sentence, wh have the same meaning. I need hardly say that the picture I have given is id decision point in the algorithm it is unlikely that a c answer can readily be made. The first decision formedness' is in itself a problem in view of the i grammar (Lyons, 1968). But more acute is the prob tation. How can we be sure when interpretatio Frequently there may be two equally plausible inter for example such an overtly idiosyncratic sentenc know the word so he asked a dictionary'. In the cont tation 'He askedfor a dictionary' is perhaps as likely a dictionary'. There is not always in the context any make one interpretation more plausible than anothe often be had to the mother tongue, if known. But pointing out that the problem of interpretation loom the classroom than in. The teacher has almost certa idiosyncratic dialect of his class and, of course, th possibility of asking the learner in his mother tong authoritative interpretation. The recourse to the mother tongue of the learner that is) is in fact also a highly intuitive process depends on the degree of knowledge of that dialect investigator. Furthermore, we cannot assume that nature of the leamer's dialect is solely explicable mother tongue; it may be related to how and w taught. Here again the teacher is in a privileged pos the idiosyncratic sentence, though teachers may admit that idiosyncracy can be accounted for by re they have done or not done! We have now arrived at the second stage: accountin idiosyncratic dialect. The first stage, if successf provides us with the data of a set of pairs of sen definition have the same meaning, or put another w tion equivalents of each other: one in the leamer's d in the target dialect. This is the data on which the des The methodology of description is, needless to say that of a bilingual comparison. In this, two languag in terms of a common set of categories and relat terms of the same formal model. The technical prob


s. P.


the process of learning. Sufficient that they were that the learning task was not yet complete. The teaching process had been perfect, no errors woul The alternative view would suggest that the maki inevitable and indeed necessary part of the learni 'correction' of error provides precisely the sort of n which is necessary to discovery of the correct Consequently, a better description of idiosyncratic tributes directly to an account of what the learner not know at that moment in his career, and should u the teacher to supply him, not just with the info hypothesis is wrong, but also, importantly, with information or data for him to form a more adequ rule in the target language. It is not, I think, therefore, a pure coincidence t interest in error analysis at the present time co increased interest in formulating some alternative h habit-formation theory of language learning.


1 Strictly speaking, of course, all the sentences of the poem since the dialect is idiosyncratic and as de Saussure say systeme dont tous les termes sont solidaires et OU la valeu que de la presence simultanee des autres', This means tha idiosyncratic' sentences should not receive the same int would if they were sentences of Standard English. Thi value of any term in a system is a function of all the term since by definition his system is different, then the value idiosyncratic, even when surface realization appears t however, hereafter use the term 'idiosyncratic sentenc sentence which superficially is not a sentence of any soc any sentence which, while it superficially resembles a s dialect, cannot receive the interpretation that such a sent in that dialect. '

2 We may note in passing that one of the problems of desc speech is an absence of information about the precise n before illness. 3 Theoretical objections to the concept of Interlanguag Dialect) apply with equal force to the concept of 'a lang course, unstable. The concept of etat de langue underlies languages. The objections to the concept of interlanguag inasmuch as they are regarded as practical objections t making comprehensive descriptions of an etat de diale


A Non-Contrastive Approa Error Analysis *

Reprinted from English Language Teaching, Vol. 2 published by Oxford University Press, London.


This article is based on a paper presented at the TESO at San Francisco in March 1970. I am grateful to W Bernard Spolsky and John Macnamara for comments of it.

1 Introduction

The identification and analysis of interference betw contact has traditionally been a central aspect bilingualism. The intrusion of features of one langu in the speech of bilinguals has been studied at the leve morphology and syntax. The systems of the co themselves have sometimes been contrasted, and an come of contrastive studies has been the notion tha prediction of the difficulties involved in acquiring a s 'Those elements that are similar to the (leamer's) will be simple for him, and those areas that are difficult." In the last two decades language teach considerable impetus from the application of contra recently as 1967, Politzer affirmed: 'Perhaps the leas least questionable application of linguistics is the contrastive analysis. Especialiy in the teaching o which no considerable and systematic teaching available, contrastive analysis can highlight and p culties of the pupils.'? Studies of second language acquisition, however imply that contrastive analysis may be most predicti phonology, and least predictive at the syntactic leve of Spanish-English bilingualism, for example, state



and developmental errors within the framework second language learning, and through examining the teaching of the forms from which they are de possible to see the way towards teaching proce account of the learner's strategy for acquiring a s

3 Types and causes of intralingual and development

An examination of the errors in tables 1-6 suggests errors are those which reflect the eneral charac earning, such as faulty genera lzabon, mcomp et rules, and failure to learn conditions under wh Developmental errors illustrate the learner attemp hypotheses about the English language from his li of it in the classroom or textbook. For convenience tables 1-6 will be discussed in terms of: 1. ove 2. ignorance of rule restrictions, 3. incomplete app 4. false concepts hypothesized. 4 Over-generalization

Jakobovits defines generalization or transfer as 'the available strategies in new situations .... In s learning ... some of these strategies will prove help the facts about the second language, but others, superficial similarities, will be misleading and inap generalization covers instances where the learner c structure on the basis of his experience of other target language. For example (see table 1, 1.3.4. we are hope, it is occurs, he comefrom. Over-generali involves the creation of one deviant structure in plac structures, It may be the result of the learner reduc burden. With the omission of the third person -s, ove removes the necessity for concord, thus relieving considerable effort. Duskova, discussing the. om person -s, notes 'Since (in English) all grammatical same zero verbal ending except the third person present tense ... omissions of the -s in the third pers be accounted for by the heavy pressure of all other e The endingless form is generalized for all persons, was is generalized for all persons and both numb



Analogy seems to be a major factor in the misuse (table 4). The learner, encountering a particular one type of verb, attempts by analogy to use the s with similar verbs. He showed me the book leads to the book; he said to me gives he asked to me; we therefore we discussed about it, ask him to do it pro to do it; go with him gives follow with him. Some p appear to encourage incorrect rules being applied t Here is part of a pattern exercise which practises ena cause, permit.

Expansion jOintS} {the PipeS} {xpand or Safety valves permit the steam escape fro allow to boiler. We the metal 001 slowl

melt The heat caused ~he metal to fract { It } { ten Weakness in the metal }

The heat } d {the metal melt. Weakness in the metal ma e it fracture under ten

It is followed by an exercise in which the student complete a number of statements using verbs and p the table: The rise in temperature-the mercury-ri The risk of an explosion-the workers-leave the fa of the train-it-Ieave the rails on the curve.... From a class of 23 with mixed language backgro than 13 produced sentences like The rise in tempe mercury to rise up the tube. Practising make in th as allow it to, permit it to, enable it to, precipitates c instances of analogous constructions may be less Table 3, 2 includes this is not fit to drink it, the man By analogy with the learner's previous experience of object constructions, the learner feels that ther incomplete about that's the man who I saw, and so after the verb, as he has been taught to do elsewher Failure to observe restrictions in article usage from analogy, the learner rationalizing a deviant previous experience of English. This may happen mother tongue is close to the English usage. F. G. F following example of how a common article mistake




Teacher's Question Do you read much? Do you cook very much? Ask her what the lastfilm she saw was called. What was she saying? What does she tell him? What's he doing? Ask her how long it takes. Will they soon be ready? How much does it cost? What does he have to do? What does he ask his mother?

Student's Yes, I read muc Yes, I cook ver What was called saw? She saying she She tell him to He opening the How long it tak Yes, they soon It cost one dolla He have to do w He ask his mothe

As the above sample illustrates, when a question sentences, the answer often has to be corrected b counteract the influence of his question. Some course almost entirely through the use of questions; others use of questions by utilizing signals to indicate the required. These may reduce the total number of de produced.

7 False concepts hypothesized In addition to the wide range of intralingual errors w with faulty rule-learning at various levels, there is a c mental errors which derive from faulty comprehen tions in the target language. These are sometime gradation of teaching items. The form was, for ex interpreted as a marker of the past tense, giving happened (table 1,2) and is may be understood to be ing marker of the present tense: he is speaks Frenc table 2, 4 we find the continuous form instead of th narrative; elsewhere we encounter confusion betw very, between come and go, and so on. In particular traced errors of this sort to classroom presentation, tion which is based on contrastive analysis of Engl language or on contrasts within English itself. Here is an example of how the present continuo understood as a narrative tense. The simple present is the normal tense used for actions seen as a whole,



is = present state, is + ing = present action.l

The contrast is in fact quite false to English. W introduced, it is often introduced as a past state. lays the groundwork for the learner to comple present and past in English by analogy: is = present state, is + ing = present action, was = past state .'. was + ing = past action.

Thus was or was + ing may be used as past marke with the verb + ed this produces such sentences the tree. Interpreted as the form for 'past actions' i down town yesterday instead of I went down town y Table 3.4 shows examples of the confusion of Other substitutions are common, such as the use of do for make, of come for go, of bring for take. L that the members of such pairs are synonyms, desp to demonstrate that they have contrastive mea fusion is sometimes attributable to premature con tion. Here are the occurrences of too and very in a tells the story of a group of children who light a f front of an old house: The house is empty because i cold. England is too cold ... The fire is very big It's a very big fire. The firemen are going to put because it's too big. The course designers intended to establish a too and very, but in so doing they completely con of the two forms. From the presentation-and fr of a young learner-they have the same meanings. parallelism between: It's too big and it's dangerous. Thefire is dangerous. It's very big.

How could a child, following such a presentatio This is a too big house? Too would be more saf association with very, and in contexts where it did a substitute for very, as, for example, in a stru adjective + infinitive-this box is too heavy to lift. Other courses succeed in establishing confusion and very by offering exercises like these:




It may be that the child's strategy of learning is to independent of the methods by which he is being ta Interference from the mother tongue is clearly a difficulty in second language learning, and contrast proved valuable in locating areas of interlangua Many errors, however, derive from the strategies e learner in language acquisition, and from the mut of items within the target language. These cannot b by contrastive analysis. Teaching techniques and pr take account of the structural and developmental c come about in language learning.


Tables 1-6- Typical Intralingual and Developmenta Table I Errors in the Production of Verb Groups I. be + verb stem for verb stem We are live in this hut The sentence is occurs ... We are hope ... He is speaks French The telegraph is remain ... We are walk to school every day. 2. be + verb stem + ed for verb stem + ed Fanners are went to their houses He was died last year One day it was happened They are opened the door 3. Wrong form after do He did not found . He did not agreed . The man does not cares for his life He did not asks me He does not has ...


JACK C. RICHARDS be + verb + ing for verb stem She is coming from Canada I am having my hair cut on Thursdays


be+not+ verb+ing for do +not+ verb I am not liking it Correct rules are not existing In French we are not having a present continuous tense and when to use it


be + verb + ing for verb + ed in narrative

... in the afternoon we were going back. On Saturday w town, and we were seeing a film and after we were meeti 5. verb stem for verb + ed in narrative

There were two animals who do not like each other. One wood and there is no water. The monkey says to the ele 6. have + verb + ed for verb + ed They had arrived just now He had come today I have written this letter yesterday Some weeks ago I have seen an English film He has arrived at noon I have learned English at school 7. have+be+ verb-s-ed for be+ verb+ed He has been married long ago He has been killed in 1956 8. verb ( + ed) for have + verb + ed We correspond with them up to now This is the only country which I visited so far 9. be s-verb-s-ed iot verb stem This money is belonged to me The machine is corned from France


JACK C. RICHARDS at instead of fP by in to for reached at a place, at last year held him at the left arm at the evening; interested at it went at Stratford at the first time


for instead of rP in of from since on instead of fP in at with of to fP in by on for

serve for God one bath for seven days suspected for, the position for C a distance for one country to an been here for the 6th of June


played on the piano for an hour on many ways, on that place, go on the end angry on him countries on the world pays attention on it aged of 44, drink less of wine rich of vitamins book of Hardy depends of civilisation a reason of it


of instead of


to instead of

for of

join to them, went to home, reac an occupation to them his love to her

Table 5 Errors in the Use of Articles 1. Omission of the (a) before unique nouns (b) before nouns of nationality (c) before nouns made particular in context (d) before a noun modified by a participle (e) before superlatives (0 before a noun modified by an of-phrase

Sun is very hot Himalayas are ... Spaniards and Arab At the conclusion o She goes to bazaar She is mother of tha Solution given in th Richest person Institute of Nuclear


JACK C. RICHARDS Omission of do Where it happened? How it looks like? Why you went? How you say it in English? How much it costs? How long it takes? What he said?


Wrong form of auxiliary, or wrong form after auxiliary Do he go there? Did he went? Did he finished? Do he comes from your village? Which road did you came by?


Inversion retained in embedded sentences Please write down what is his name. I told him I do not know how old was it I don't know how many are there in the box


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

Lado, 1957,p. 2. Politzer, 1967, p. 151. Lance, 1969. Cf. Cook, 1969; Stem, 1969; Menyuk, 1969. Corder, 1967, pp. 161-9. Major sources for Tables I-IV are: French, 1949; Duskov 1968; C. Estacia, 1964; Bhaskar, 1962; Grelier; Aguas, Jacobovits, 1969, p. 55. Duskova, ibid. Ervin-Tripp, 1969,p. 33. Wolfe, 1967, p. 180. French, ibid., p. 9. Aguas, ibid., p. 49. Hirtle, 1967,pp. 40-1; Close, 1959, p. 59. Ritchie, 1967, p. 129. Hok, 1964,p. 118. George, 1962,pp. 18-31. Dakin, 1969, pp.107-11.



variables related to second language learners operate d each second language learning situation. Learning different training procedures, individual differences learners, text books-all seem to operate to make ea situation different from the other. Inevitably, L, indepen across diverse language backgrounds differ to the exten variables operate differently. In presenting error evid an attempt is made to avoid simple inventories, for inv essentially unrevealing if they do not explain why the list made. Of the many factors causing L, independent erro strategies, teaching techniques, folklore about the secon the age of bilingualism, i.e. the period over which language has been used by the speech community t learner belongs, the learner's sociolinguistic situation, the more easily identifiable variables. In this paper the on some learning and teaching strategies; the other only been touched upon in passing, because they represe for which we do not have systematic knowledge. The pa suggest that there is a system in learners' errors in s apparent arbitrariness in performance data, but this me be captured through a simple binary opposition betwee and non-systematic errors.' Finally, though the eviden the paper offers some tentative observations about the of learners' errors: what light they seem to throw on the of second language learning and what implications th language pedagody. If then errors are traceable to both L, and other-than a single conceptual framework uniting both these s seems to provide a meaningful frame for inquiry into and error significance. Not that it is an easy task: errors to submit themselves to any precise systematic analysis; between errors traceable to L, interference and those t pendent of LI interference is not invariably clearcut; menon of errors caused by the cross-association of bot also seems to exist; the identification and establishing o independent interference factors is far from easy; psychological processes of second language learning learning strategies can at best be marginally inferred f formance data. Beset though the task is with nume dological difficulties, it appears that some tentatively tion is feasible for fruitful explorations in this area.



by plural morphemes; that the latter have no plural they marked by 'a'. Hence the following occur in performance data: 1. They have bought a new chair. 2. They have bought new chairs. And also: 3. She wants drinking water. 4. She wants some drinking water.

'Chair' and 'water', like many other members of the c conform to the generalization. A rule strategy of this however block the following if the learner is writing plural sense apparently under pressure from non-lingu mental cues. (The asterisk marks a 'wrong' sentence.)

* * *

5. Entry will be by tickets. 6. He is a man of his words. 7. She is always finding faults with me.

To avoid the plural he has to sub-categorize certain 'co 'uncount' for these specified contexts. In the face whelming experience with 'word', 'ticket' and 'fault' language corpus as 'count' nouns, and also in terms of strategy, this sub-categorization is particularly difficu pressure of these nouns as 'count' occurrences nearly Occasional appearances of these words in the singular fore become 'significant' for him. The generalization count constitutes for him the 'creative'S aspect for the accounts for the plural forms of 'leg', 'pass', 'book' 'stone' (table 1, section 1). Thus, whenever the situati seems to demand a plural idea, the generalized rule p well as 9 unless of course the use of 'word' in idiomatic is deliberately learnt outside the generalization: 8. He had no words to express his gratefulness. 9. He will not go back upon his words.

The nonce singular form occurrence o( these 'count context where non-linguistic features of'the situation demand the plural, e.g. 10. He sent the message by word of mouth.



on the noun seems out of place when the singular thr so overtly indicated as in the following (see table 1, se


17. It is one of the interesting book I have read.

As a majority of noun occurrences for number in structure of English are regularly governed by this 'co generalization, the economy-seeking second langu readily adopts it for all occurrences, excluding at best only which have been given the status of 'exception particular teaching strategy or a popular school gramm generalization crystallized as a rule, and no longer a me open to further modification, the learner multiplies his thousand times on its basis in the 'creative' mode;with good communication results. It is possible that LI may categorization of some of the nouns in the category 'c than the other. What is more important is that on categorized, their occurrences in the "learner's English by this generalization. Subsumed under the category nouns 'work', 'bedding', 'fruit', 'scenery', 'hair', 'lugg ment', 'data', 'criteria', 'criterion', 'phenomenon', 'pape tion', (see table 1, sections 6 and 7) take plural morp according to what sense the learner wishes to convey:

* * * *
* *

18. 19. 20. 21. 22. 23.

She has a very important work to do. We have many important works to finish toda He offered a fruit to the deity. The doctor has advised him to eat plenty of I pulled out a white hair from my head this m Her hairs are black.

It is possible that the errors in 18and 20 are due to LI but it is difficult to explain the errors in 19 and 21 on the particularly when the pull of the mother tongue in th should have forced the singular form of 'work' and learner is not operating merely in terms of the mo equivalents. Similarly, one may attribute 'a' in 22 to tongue, and this time the interference proves a blessing but we are again put to great difficulty in explaining 'ha the same ground because the mother tongue in this co demand the singular form. The 'count-uncount' generalization seems to have oth tions too. The learner's experience of the regularity o



drift through various stages by successively modif generalizations through 'exceptions' Dr sub-classif direction of the accepted standard of an ideal speak but for some learners the generalizations premature status of rules. These rules, used creatively, are ov and give rise to' the many inflexional errors so well kn language teachers: mouse : 'mouses': machinery: equipment: 'equipments'; hit: 'hitted'; dig: 'digged their productive character they are very often a source novel creations; e.g.

* *


28. Witnessers and supporters also. thought of energy ... 29. A discussion started about pick-pocketers. 30. He is a big cheater.

If the 'count-uncount' generalization constitutes aspect fDr number and also.the use of 'a' for members the generalization that the verb must agree with the mines the number of the verb. Hence, if the learn generates 'datas', 'criteria' - 'criterias', 'scissor' subject-verb agreement generalization leads to':



* * *

31. 32. 33. 34. 35.

The datas given in the table are very revealin This criteria is questionable, The criterias are objective, This scissor is of good quality. These two. scissors are of better quality.

Similarly, as 'police' is not overtly marked fDr numbe the teaching strategy nor the SChDDI grammar has giv of 'exception', the generalization systematically d singular verb:


36. The police was unable to. control the rioting

The errors in table 2 are the product of the generaliz members of the class verb in English are either tra transitive, and for the progressive aspect they are mar on the surface. Hence the occurrences: 37. They are drinking tea. 38. They are playing chess. 39. They are running.



about learning economy through reduction of the se along one dimension or another. Limited vocab structures, abridged and simplified texts, simplified oversimplified school grammar books are attempts direction. Thus simplified generalizations would se into a second language teaching situation. For examp hardly be a teaching course that does not base itself uncount' generalization; nor is there a teaching te dispenses with it. We all know that this type of gen helpful teaching device '" but it does not truly reflect t second language. In fact nearly all count nouns uncount.P But the learner has very little choice in stricted generalization, if that is the only one available so he faces the problem of missing the generaliza sequently learning is affected. The learner does not generalize along the dimensio regularities alone. Other kinds of regularities also see Some generalizations seem to be derived from a consistency between form and function in the surfa English. For example subject-verb inversion is near property of the interrogative sentence, while sub nearly always marks declarative sentences. If the generalizes that a sentence overtly marked in the initi the question word or the anomalous finite alone has subject-verb inversion, and gives it the status of ru from capturing a very frequent feature of the Engl terms of a form related to a particular function. A along these lines rules out inversion in the following word is not prefixed to these sentences (see table 3)



43. Under no circumstances we will accept the 44. They didn't like it; nor I liked it. 45. He was unhappy at the development; so I

Yet another way in which the learner genera experience of this sentence type is to associate the with the subject-verb inversion. Either his own expe sentence type or a particular teaching strategy or bo the generalization. Over-generalized to indirect que leads to errors of the type cited in table 4:


46. Now I see why did he behave like this?



We said to him, The weather is stormy, and th The teacher told us, 'The prize will be presente We heard him say, 'I will agree to what you My son exclaimed, 'Someone has taken the reading.' 64. He made a promise, 'I will come, if 1 can.' What is to be noted is that by its very nature 'that' as a clause marker. The exercise in the from indirect to direct narration, also tends learner is asked to convert sentences of the Indirect to Direct: 65. 66. 67. 68. 69.

60. 61. 62. 63.

the exe reverse to emp followi

He made them understand that he would soo They affirmed that he was the best worker t He admitted that he had not worked so hard He heard them say that he did not deserve th He promised them that he would do it as soo

To these are added examples of the following type:

70. He reminded me, 'When the cat is away, the 71. He reminded me that when the cat is away, 72. Pilate replied to the Jews, 'What 1 have w written.' 73. Pilate replied to the Jews that what he had w written. 1 6

Exposed to this language data, coupled with text less around the principle of one-thing-at-a-time in which th is again separately highlighted as a 'teaching' point, the to induce that 'that' is an invariant clause marker for In terms of relative position of 'that' to other items in it implies that the clause after the verb is to be prece His non-error performance data as well seem to bring fixation; his utterances are cluttered with 'that'. Take sequence:

74. He conceives that with the help of science ... that with the help of the Bokanovsky's proces that all the discomforts ... He thinks that .. ... The fact that ...

When the learner adds to this generalization, anoth



The errors in table 7 highlight the learner's failure restrictions on the co-occurrence of items with Restrictions are of two types: lexical and gramm restrictions deal with the co-occurrence of wholly they rule out the co-occurrence of some words with of the incompatibility of pairs of words, e.g.


83. He drives a scooter.

Grammatical restrictions, on the other hand, determ rence of items in grammatical contexts; they do no words to enter at all into certain grammatical cons


84. We want that Hindi should be the medium

is not acceptable because 'want' cannot be follow ordinate clause. Very often the learner's failure to observe thes restriction is attributed to L, interference, particular such as 83 and 84. However, the violation of th restrictions in the following sentence (for more exa 7) cannot be defended on the same ground:


85. She delivered a male child.

The failure hardly seems to stem from L, interference suggested here is that most. of these errors are root learning strategy of speech reduction as discussed ea to many other errors. By not going into the idiosync of individual words, the learner cuts down the categorizing. It may, however, be noted that the categorizing at the level of grammatical and lexical the co-occurrence of items is highest. While a rule either transitive or intransitive takes care of many m class V, the rule, for example, that 'deliver' is to passive and followed by 'of takes care of a single m rules at the level of restrictions very often reflect th property of individual words, they cannot be used mode; so the failure to apply them correctly may nonce errors. So far we have discussed errors which seem to fal patterns; they show a consistent system, are intern and free from arbitrariness: they are therefore sys



to settle for him is where to use the article and where when he has to make the choice for members of t language is taken to be rule-governed behaviour, then data of this type in these selected areas would constit like behaviour. In other words he has no producti article usage, not even that type of control which he where he has over-generalized. The result is: asystema in the context of such errors that it is difficult to stat for the occurrence of errors in the learner's perfo However, he has some kind of receptive control of the ate areas, largely owing to the total context of the discourse. It appears that in a particular learning situation there are too many cues to frustrate his reaching a fi tion. Teaching techniques, the learner's experience w language- the facts of surface structure at times m obscure the generalization-the popular school gram about the second language, the learner's sociolingu the various practices and traditions regarding the se usage obtaining in a bilingual situation, all may pr selected areas of indeterminacy in the learner's syllabus let us take the learning situation in which both the sc and the teaching techniques offer for article use generalization; 'a' is to be used before singular coun should not be used in front of 'abstract' nouns. Exam 86. 87. 88. 89. 90. 91. 92. 93. 94. It is a big white egg. It is a toy train. I see a man in the distance. It is a good story. You can buy a pencil; she can buy a pen. He is a painter. This is a bus, that's a doll, and that's a cat. They jumped for joy. Books give pleasure.

Before the proffered generalization crystallizes into learner is confronted with the data of the following 'cognitive clutter' occurs: 86a. The child likes egg. 87a. We'll go by train.



systematic rules in these areas: 'learning' is 'hoped' through the leamer's 'experience' of and 'practice' language. Quite often the 'hope' is not fulfilled becau indeterminacy seems to mar the leamer's ability to forms, patterns and rules from new linguistic context Articles, prepositions and the tense system in Eng do not exhaust the likely areas of indeterminacy. A tea or anyone or more of the other error-causing L variables referred to above may transfer this oppos rules and arbitrariness to many other areas, even thou structure of English may not so obviously carry it f making it difficult for him to accept the proffered rule This is apparent from the leamer's clause handling in graph sequence quoted in table 8, section 2. In spite o ately using 'that' in the first sentence of the first pa in the first sentence of the second paragraph, 'what' sentence of the fourth paragraph as clause markers use of 'that' after arise, becomes and is a problem seem his grammar has not been able to spell out for him th restrictions which hold for the 'that' clause, nor h use of 'how' and 'what' explicit as clause markers permits ungrammatical occurrences of 'that what' a Through error examples of this nature it is not u suggest that these are not random errors; they s product of indeterminacy in the leamer's syllabus construction rules for noun clauses. Besides systematic and asystematic errors, the l native speaker, seems to make unsystematic errors to slips of the tongue or pen caused purely by psycholog such as intense excitement, and/or physiological f tiredness, which change from moment to moment an to situation: errors under these circumstances are systematic. From the pedagogic point of view, on dismiss them as 'mistakes'. 19 As these 'mistakes' ar the discussion in this paper, they have not been con From the above explanation of the source and certain implications for language acquisition and lan follow. 1. Errors discussed under the category of system to establish that in certain areas of language use the l construction rules. The debatable question whether



through intensive drills of correct forms. Depending up of the error-causing generalizations, it may be absolu to re-examine and modify teaching techniques, the p the corpus of the primary linguistic data, the disa learner's language acquisition device of the heresies o erroneous generalizations. If placed in this larger conte tion of actual errors would have meaning. 5. The fact that errors cited in this paper are from ance data of those who have studied English from 1 would suggest that their learning has largely stoppe petence is marked partly by indeterminacy and partly generalizations crystallized into rigid rules hardly op in the face of new evidence. In areas of language learner possesses rules, he is no longer discoverin language; he has arrived at a system. And howsoever rules may be from the standpoint of the accepted g native speaker, he uses them in the 'creative' mode an are grammatical in terms of his own grammar. Th where it seems reasonable to hypothesize that the 'inp learning the L2 system is not what is available for goi he takes in through his over-generalized rules. 23 Iflear place in these areas, teaching materials with an expl in them to motivate the learner to test his 'rules' inadequacies would seem to be required. An implic kind of approach is not to regard errors as exemplifie sin' to be subjected to 'intensive remedial drill'; consider them as an essential condition of learning, teacher with clues on how to take the learner fro schemata to more generalized ones, establishing the 'significant' than the former.

Table I I.

To convey the idea of plurality, the underlined 'count' n leading to errors. Entry will be by tickets. He is a man of his words. She is always finding faults with me. He will not go back upon his words. Don't pull my legs. --



This study describes some rare phenomenons in the field of m Let me have two papers. (Instead of two sheets of paper.) These informationsare not very reliable. Her hairs are black. 7.

Subsumed under the category 'count', the underlined nouns are (also underlined) to convey the sense of 'one'.

She has a very important work to do. Each one of you is advised to carry ~ bedding with you. He offered a fruit to the deity. As you enter this mountain city, you are greeted by ~ beautifu You can't charge so much cartage for carrying one luggage o A sound criteria is to be fixed for judging it. - --For the time being I want a paper and a pencil. He was irritated to find it hair in his food. (Though the source sentences in this section of the table, 'a' here turns out to usage.) 8. Subsumed under the category 'count', the underlined nouns numerals (also underlined) to give the idea of number.

These are two very good news you have brought. The girls need three more scissors. You may buy two woollen trousers. I always carry two spectacles with me, one for reading and the use. In the theft we lost about 15 clothes. 9. To use the 'count' nouns of his grammar in the singular, the plural marker -~and puts 'a' in front of the noun.

She wants to buy a scissor. It seems to be a new trouser. Why didn't he-buy a new spectacle? In the theft they see-medto have lost only one clothe. It's a good news. (The -s seems to stick because of the meaningful contrast be 'new' in the learner's English.)

Table 2 Verbs are marked with -ing for the progressive aspect.

I am having a very heavy work-load this semester. We make mistakes because we are not knowing the rules. The main reason for my leaving the hostel was that I was no



God will ask him that why he did not make any use of his t In these lines the poet has described that how the nature will g He tells us that he has moments when he considers that how He knows well that what a blind man can do. -- -His behaviour indicates that whom he likes more. He thought that how people behaved in romantic ways. He says that when he thinks that whether God wants hard w man, the answer is 'no'. -- ---He further says that what can be done when one's death ha --God. It looks to him very strange that why man should fear deat It is fantastic that how a blind man can write so well. It is wonderful that how they have been able to explore the o Since it is not predictable in most of the cases that who is g 2.

'That' is used as a noun clause marker, and the subject-verb question word or the anomalous finite.

He points out that how will he escape punishment? He asks that is it possible to expect work from a blind man? But thenhedeeply thinks that does God expect work from Then he complains to God that does God demand full work f blind? -----

Do you think that will God punish a blind man for not doin Now it is clear that why will God not punish him?

Table 6 l. Mutual interference between items in English.

In this poem Wordsworth has described about nature. He has explained to him about the whole matter. In this way Huxley conceives about the future of the modem 2.

Mutual interference between items in English in conjunctio Wh generalizations.

He explains to Raina about the battle that how they were a In this poem Wordsworth has described about nature that Lucy. ----

Table 7

Miscellaneous items used without the grammatical restrictions th She delivered a male child this morning. You didn't avail of the opportunity.



modity, then who will consume it? ... The same reason can be used to produce different comm it is decided what to produce, the problem arises that them because the resources are limited .... The production less than the demand for it, so the problem becomes tribute among people the commodities produced.


See French, 1949; Corder, 1967, pp. 161-70; Wolfe, 19 Lee; Duskova, 1969, pp. 11-36; Richards, 1971, pp. 204 1971,pp. 220-29; George, 1971, pp. 270-77.

2 The data (see Appendix, Tables 1-8) pertain to the w English of Indian students at the university level who h years' instruction in English at school. At the university th 2 to 3 years' more instruction. Of the learners who figure in are about to complete and some have just completed the formal instruction in English, the total instruction period to 14years from learner to learner. From now onwards th with English on their own in whatever situation they wish 3 In the context of this paper Hindi is Lt, the leamer's f mother tongue; English is L2, the second or foreign langu 4 In collecting the evidence the context of each error was account. For economy of presentation, errors from I cited devoid of the context of the discourse; while in Tab the nature of the errors, they have been cited in longer stre course. 5 Cf. S. P. Corder, ibid., p. 166: The opposition between syst systematic errors is important.' Also see Duskova, ibid.

6 Non-error performance data do carry evidence for red ample, between two free variants, Help me to lift the box the box, if a learner has made a choice for the one and not reducing speech but without creating possibilities for er when he settles on a restricted generalization and stret over-application. See the discussion below. 7 See Klima and Bellugi, 1966. 8 Used here in the Chomskian sense. Chomsky, 1965. 9 It is interesting to note that if the second language lear under the pull of the mother tongue, the mother tongue w singular form of 'ticket' and 'word' in 5, 6 & 9. 10 They aid the identification of nouns for number. Cf. Brown II See Ervin, 1964. She points out how under the pressure o i.e. adding 'ed' for the past, the irregular forms of past te 'went', 'sat', often correctly produced early on in the child are dropped and forms like 'corned', 'goed' and 'sitted' a


Afolayan, A. 'Contrastive Linguistics and the Teaching of Eng Language.' English Language Teaching, XXV, 1971,220-9. Aguas, E. F. English Composition Errors of Tagalog Speake UCLA,1964. Alatis, James (ed.). Bilingualism and Language Contact: Linguistic, Psychological and Sociological Aspects. Mono Languages and Linguistics, 23, Washington, 1970. Alford, N. D. Research Report on some Aspects of the Langua of Pre-School Children. Brisbane, 1970. Allen Jones, J. 'English Language Teaching in a Social/C Situation.' English Language Teaching, xxn, 3, 1968, 19 1969, 18-23. Allsopp, Richard. 'The English Language in British Guinea.' E Teaching, xn, 2, 1958, 59-65. Anisfeld, M. and Tucker, C. R. 'English pluralization rules children.' Child Development, 38, 1967, 1201-18. Arabski, J. 'A Linguistic Analysis of English Composition Polish Students.' Studia Anglica Posnaniensia, I, 1 and 2, 1 Ashton-Warner, Sylvia. Teacher. New York, 1963. Bamgbose, A. 'The English Language in Nigeria.' In Spence Banathy, B., Trager, E. C. and Waddle, C. D. 'The use of co foreign language course development.' In A. Valdman Language Teaching. New York, 1966. Bazell, C. E. et al. In Memory of J. R. Firth. London, 1966. Bellugi, Ursula. 'The Development of Interrogative Structur Speech.' In K. Riegel (ed.), The Development of Lang Ann Arbor: Michigan Language Development Program, 1965, 103-37. Benton, Richard. Research into the English Language Diffic School Children, 1963-1964. Wellington, Maori Educatio 1964. Berko, J. 'The child's learning of English morphology.' W 159-77. Bever, T. 'Associations to stimulus-response theories of lang Dixon and D. R. Horton (eds.), Verbal Behavior and G Theory. Englewood Cliffs, N.J., 1968. Bever, T., Fodor, J. A. and Garrett, M. 'A formal limitation




Cazden, C. a~d Brown, R. .'The early development of the ~raft of article to appear In Lenneberg, Eric and Elizabet ~wnsof Language Development: A Multidisciplinary Appr m press. Ceraso, J. 'Specific interference in retroactive inhibition.' Jo ogy, 58, 1968,65-77. Ceraso, J. and Tendler, M. 'Pair vs. list interference.' Ame Psychology, 1968a, 47-52. Chomsky, Carol. The Acquisition of Syntax in Childrenfrom chusetts, 1969. Chomsky, N. 'A review of B. F. Skinner's Verbal Behavior. J. Fodor. The Structure of Language. Englewood Cliffs, N. Chomsky, N. 'Research on Language Learning and Linguisti Northeast Conference, 1966. Chomsky, N. Language and Mind. New York, 1968. Chomsky, N. 'Linguistics and Philosophy,' in Language ed. by Sidney Hook. New York, 1969. Christian, Jane and Christian, Chester. 'Spanish Language an Southwest.' In Fishman, 1966. Close, R. A. 'Concerning the Present Tense.' English Lang XU, 1959, 59. Cook, Vivian. 'The Analogy Between First and Second Lang IRAL,7, 1969,207-16. Corder~ S. P..'Th~ Significance of Learners' Errors.' IRAL, 5 (repnnted In this volume). Corder, S. P. 'Describing the Language Leamer's Language.' I Approaches.t0.Languag~, C.ILT Reports and Papers, 6, 197 Cor~er, S'.P. I.dlOsyncrancdialects and error analysis.' IRAL, printed In this volume). Coult~r, Kenn~th. Linguist~c Error-Analysis of the Spoken E Native Russians. Unpubhshed M.A. Thesis University o 1968. ' Craig, Dennis. 'Linguistic and Sociolinguistic Problems i Language Education and Policy in the English-Speaking C Crothers, Edward and Suppes, Patrick. Experiments in Sec Learning. New York, 1967. Curti~, J. '<?on~ptual Models of Speech Production and the Articulation, Paper presented at the 1968 ASHA Conven Colorado, 1968. Dak!n, J: '!he Teaching of Reading.' In Fraser and O'Donnell Linguistics and the Teaching of English. London, 1969. Darnell, Regna. 'The Bilingual Speech Community: A Cree Darnell, 1971. Darn.ell, Regna. Linguistic Diversity in Canadian Society. Edm Da SIlva, Z. S. Beginning Spanish, Second Edition. New York, Dato, ~. P. American Children's Acquisition of Spanish Syntax Env~ronment: Preliminary Edition. U.S. H.E.W., Office o Institute oflnternational Studies, Final Report, Project No. 3 no. O.E.C. 2-7-002637, May 1970.



Fishman, Joshua, Cooper, Robert L., Ma, Roxana et al. the Barrio. Yeshiva University, New York, 1968. Fodor, J. and Garrett, M. 'Some Reflections on Competenc ance.' In J. Lyons and R. J. Wales (eds.), Psycholinguistic P 1966. Fodor, Jerry A. Psychological Explanation; An Introduction t of Psychology. New York, 1968. French, F. G. Common Errors in English. London, 1949. Fries, C. C. Teachingand Learning English as a ForeignLangua 1945. Fries, Charles C. and Pike, Kenneth L. 'Coexistent Phon Language, 25.1, 1949, 29-50. Garrett, M. and Fodor, J. 'Psychological theories and lingui In T. R. Dixon and D. R. Horton (eds.), 1968. Garvey, Catherine and Rocklyn, Eugene H. Development an a Tactical Mandarin Chinese Language Course. Washingt George, H. V. 'Teaching Simple Past and Past Perfect.' Bullet Institute of English (Hyderabad, India), 2, 1962, 18-31. George, H. V. 'English for Asian Learners: Are we on t English Language Teaching, XXV, 1971,270-77. George, H. V. Common Errors in Language Learning. Rowley, 1972. Gilbert, Glenn (ed.). The German Language in America-A Austin, University of Texas, 1971. Gradman, Harry L. 'What Methodologists Ignore in Contras Paper presented at the Pacific Conference on Contrastive Language Universals, Honolulu, 1971. Grelier, S. 'Recherche des principales interferences dans les sy de l'anglais du wolof et du francais,' Senegal. Centre d Appliquee de Dakar. n.d. Grimshaw, Allen D. 'Some Social Forces and Some Socia Pidgin and Creole Languages.' In Hymes, 1971. Gumperz, John and Hymes, Dell. 'The Ethnography of Co American Anthropologist, 66, pt. 2, 1964. Gumperz, J. and Hernandez, E. 'Bilingualism, bidialectalism interaction.' In C. B. Cazden, D. Hymes and V. John (eds Language in the Classroom. New York, in press. Haggard, Mark P. 'Models and Data in Speech Perception,' in Perception of Speech and Visual Form, ed. by Weiant Massachusetts, 1967. Hall, Robert. Pidgin and Creole Languages. Ithaca, 1966. Halle, M. The Sound Pattern of Russian. The Hague, 1959. Halle, Morris and Keyser, Jay. English Stress: Its Form, Its Role in Verse. New York, 1971. Halliday, M. A. K. 'The Users and Uses of Language.' In F Hammer, John H. and Rice, Frank A. (eds.). A Bibliography Linguistics. Washington D.C., 1965. Hart, J. T. 'Memory and the memory-monitoring process.' Jo Learning and Verbal Behavior, 6, 1967,685-91.



Regional Laboratory, Los Angeles, 1970. Kennedy, G. 'Conditions for Language Learning.' In J. W. O C. Richards (eds.), Focus on the Language Learner. N Rowley, Massachusetts, 1973. Kessler, Carolyn. The Acquisition of Syntax in Bilingual Childr University Press, Washington D.C., 1971. Kinzel, P. Lexical and Grammatical Interference in the Speec Child. Seattle, 1964. Kirk-Greene, A. 'The Influence of West African Languag In Spencer, 1971. Klima, E. S. and Bellugi, U. 'Syntactic regularities in the spee In J. Lyons and R. J. Wales (eds.), Psycholinguistic Pap 1966. Kline, Helen. 'Research in the Psychology of Second Langu Unpublished paper, University of Minnesota, 1970. Kloss, Heinz. 'German-American Language Maintenanc Fishman, 1966. Labov, William, Cohen, Paul, Robins, Clarence and Lewis, Jo the Nonstandard English of Negro and Puerto Rican Speake City.' ERIC: ED. 028 423, 1968. Labov, William. 'Contraction, Deletion, and Inherent Va English Copula.' Language, 45, 4, 1969, 715-62. Labov, William. 'The Study of Language in Social Context.' St 23, 1970,30-87. Labov, William. 'Variation in Language.' In Carroll E. R Learning of Language. New York, 1971, 187-221. Labov, William. 'The Notion of System in Creole Languag 1971. Lado, Robert. Linguistics Across Cultures. Michigan, 1957. Lakoff, George. 'On Generative Semantics,' in Semantics-An ary Reader in Philosophy, Linguistics, Anthropology and Danny Steinberg and Leon Jakobovits. Cambridge Univers 232-96. Lambert, W. E. 'Behavioral Evidence for Contrasting Forms o In M. Zarechnak (ed.), Monograph Series on Languages and 1961. Lambert, W. A. 'Some Observations on First Language A Second Language Learning.' (Mimeograph), 1966. Lance, D. A Brief Study of Spanish-English Bilingualism. Research Project Orr-Liberal Arts-15504. College Station A. and M. University, 1969. Lane, H. 'Some differences between first and second lang In Language Learning, XU, 1, 1962. Laver, John. 'Assimilation in Educated Nigerian English.' En Teaching, XXII, 2, 1968, 156-60. Laver, J. D. M. 'The Production of Speech' in John Lyons (ed. in Linguistics. Baltimore, 1970, 53-75. Lawler, J. and Selinker, L. 'On the nature of paradoxes, rules second language learning.' Language Learning, XXI, 1, 197



Mimeo. Hartford, Connecticut Board of Education, 1970 Marckwardt, Albert H. 'Phonemic Structure and Aural Perce Speech, 21, 1946, 106-11. Melton, A. W. 'Comments on Professor Postman's paper.' In and Verbal Behavior. C. N. Cofer (ed.). New York, 1961. Melton, A. W: an~ I~n.' J. 'The influence of degree of inter on retroactive inhibition and the overt transfer of spe American Journal of Psychology, 53, 1961, 173-203. Melton, A. W. 'Repetition and retrieval from memory.' Sc 532. Menken, H. L. The American Language (4th edition). New Y Menyuk, Paula. Sentences Children Use. Massachusetts, 196 Menyuk, Paula. The Acquisition and Development of Langua 1971. Michigan Oral Language Series. English Guide-Kindergar Council on the Teaching of Foreign Languages, 1970. (O Foreign Language Innovative Curricula Studies, Ann A



Miller, G. A. 'Language and Psychology.' In E. H. Lenne Directions in the Study of Language. Boston, 1966. Miller, G. A. The Psycholinguists.' Encounter, 23.1, 1964. Milon, J. 'A Japanese child learns English.' Paper presen Convention, March 1972. Mintz, Sidney. 'The Socio-Historical Background to Pid Creolization.' In Hymes, 1971. Nababa~, P. W. J. 'A Note on Transfer and Interference in Fo Learning.' Working Papers in Linguistics, 3.4, 1971, 17-23. Natalici?, S. and .Natalicio, L. F. S. 'A comparative st pluralization by native and non-native English speakers.' In ment, 42, 1971, 1302-6. Neisser, U. Cognitive Psychology. New York, 1966. Nemser, William J. 'Hungarian Phonetic Experiments.' Amer Lea~ned Societies, Research and Studies in Uralic and Alt Project No. 32, 1961. N~mser, WilliamJ. and Juhasz, Francis S. 'A Contrastive Anal Ian and Enghsh Phonology.' American Council of Lea Research and Studies in Uralic and Altaic Languages, Projec Nemser, William. 'Approximative Systems of Foreign Langu IRAL,9.2, 1971, 115-23 (reprinted in this volume). Nemser, William. An Experimental Study of Phonological Int English of Hungarians. Bloomington, 1971. Nemser, William. 'Prob~ems and Prospects in Contrastive Lin presented at International Conference on Modem Lingui guage Teaching. Hungary, April 1971. Nesfield, J. C. 'Idiom, Grammar and Synthesis.' English G Book IV, 1953, 290-1. Newmark, Leonard. 'How Not to Interfere With Language Mark Lester (ed.), Readings in Applied Transformational G York, 1970,219-27.




Language Learning, xxn.2, 1972, 159-188 (reprinted in Richards, Jack C. 'Some Social Aspects of Language Le Quarterly, 6.3, 1972,243-254. Ritchie, William C. 'Some Implications of Generative Gram Learning, XVll, 1967. Rivers, W. The Psychologist and the Foreign Language Teach Rocklyn, Eugene H. Self-Instructional Tactical Language C Washington D.C., 1965. Rosenbaum, P. The Grammar of English Predicate Compleme Cambridge, 1967. Rudegeair, Robert E. 'The Effect of Contextual Influen Discrimination oflnitial Consonants.' ERIC: ed. 047 307 Samarin, William. 'Lingua Francas, with Special Referen Rice, 1962. Samarin, William J. 'Salient and Substantive Pidginization.' Sampson, Gloria P. 'The Strategies of Cantonese Speakers L In Regna Darnell (ed.), Linguistic Diversity in Canadian So 1971, 175-99. Sapon, S. M. 'On defining a response: a crucial problem i verbal behavior.' In Pimsleur and Quinn (eds.), The Psyc Language Learning. Cambridge, 1971. Saporta, S. 'Applied Linguistics and Generative Grammar. (ed.), Trends in Modern Language Teaching. New York, 1 Sebeok, Thomas (ed.). Linguistics in South Asia: Current Tre 5, The Hague, 1969. Selinker, Larry. 'Language Transfer.' General Linguistics, 9 Selinker, Larry. 'Interlanguage.' IRAL, X.3, 1972,219-231. volume.) Selinker, L. 'The psychologically relevant data of second la In Pimsleur and Quinn (eds.). The Psychology of Second La Cambridge, 1971. Slamecka, N. J. 'An examination of trace storage in free r Experimental Psychology, 76, 1968,504-13. Slamecka, N. J. 'A temporal interpretation of some rec Psychological Review, 76, 1969,492-503. Smith, F. and Miller, G. A. (eds.), The Genesis of Languag Cambridge, Massachusetts, 1966. Spencer, John. Language in Africa. Cambridge, 1963. Spencer, John. 'West Africa and the English Language.' I 1-35. Spencer, John. The English Language in West Africa. Lo 1971. Spolsky, B. 'A Psycholinguistic Critique of Programmed F Instruction.' lRAL, 4.2, 1966, 119-29. Stem, H. H. 'Foreign Language Learning and the New View o Acquisition.' Child Study, 30.4, 1969, 25-36. Stewart, William. 'Creole Languages in the Caribbean.' In R Stewart, William. 'An Outline of Linguistic Typology for D lingualism.' In Rice, 1962.




Wax, Murray, Wax, Rosalie and Dumont, Robert. 'F American Community.' Social Problems, 2,4, 1964 Weinreich, Uriel. Languages in Contact. New York, 1 Whinnom, Keith. 'Linguistic Hybridization and the S and Creoles.' In Dell Hymes (ed.), Pidginization an guages. Cambridge, 1971. Wilkins, D. A. Review of A. Valdman, Trends in Langu 1968,99-107. Williams, Frederick. Language and Poverty. Chicago, Williams, G. 'Analysis of Puerto Rican English.' Language Research Foundation, Cambridge, Mass Wolfe, D. L. 'Some theoretical aspects of language teaching.' Language Learning, xvn, 3 and 4,1967,



Sign up to vote on this title
UsefulNot useful