CONTRASTIVE ANALYSIS AND ITS RELEVANCE TO LANGUAGE TEACHING

In tlic course of lea rn ing a second language , learners regularly p roduce u tte rances in sj)ccch a n d w ri t ing w hich ju d g e d by the rules o f the second language a rc erroneous, o r ill-formed. T rad i t ion a l ly the a t t i tu d e to errors was tha t they w ere an ind ica t ion o f the difllculties th a t the learners had with cc r ta in aspects o f the language , which could be expla ined by the persistence o f the hab i ts of the m o th e r tongue a n d the ir transfer to the new language (Lado , 1957). E rrors were the result o f interference an d in an ideal teach ing s itua tion could be avoided. F rom this notion has developed the whole indus try o f “ contras tive linguistics” , w ith research projects a n d regular publica tions of results in a n u m b e r o f countries, inc luding S ou th Africa.

Analysis" and before one can pursue its "relcvance to language teaching" it seems appropriate to define and discuss what is meant by the term.The modern view of contrastive analysis as the analysis and interpretation of interference errors is not to be confused with the Contrastive Analysis approach to interference ])henomena as an instrument of prediction, proposed by applied linguists such as Fries (1945) and Lado (1957).In his pre|)aration of teaching materials at the English Language Institute of the University of Michigan, Fries stated the guiding pi inciple to his view of the acquisition of the target language: "'I'he most efl'ective materials are those that arc based upon a scientific description ofthe language to be learned, carefully compared with a parallel description ofthe native language ofthe learner" (1945, p. 9).This view presupposes that diíTiculties in the acquisition of the target language may be predicted and if eliminated will cause the learning of the target language to be facilitated.The aspect of prediction in a CA, especially in so far as it concerns its practical application in the teaching procedure of the target language, was stressed throughout.Lado emphasizes this point when he slates: "The plan ofthe book rests on the a.ssumption that we can predict and describe the patterns that will cause difllculty by comparing systematically the language and culture to be learned with the native language and culture ofthe student" (Lado, 1957, p. vii).The principle of target language learning was based on the a.ssumption that the major problem was caused by inter-lingual identification and that the diil'erences l)ctwecn ihf sourer language and the target language once they had been predicted could be diniinishcd by exposing the learner to drills specirically designed to change his linguistic betiaviour at the relevant jjoints.'I'hcoretically, this seemed a perCectly valid argument and consequently during the 19f)0's a spate of contrastive studies appeared all over the world.
Since 1960 however, developmenl along these lines has been severely checked and for a variety of reasons of which the most important were; 1. the close links the approach had with structural linguistics which at that lime was being ousted by the transformational generative theory; and, 2. its close lies with the theory of behaviourism which also at that time seemed to be completely negated by psychologists and linguists.
3.More important, scholars working in the field of Applied Linguistics also criticized the approach, firstly on the grounds that it had very little practical application in the classroom situation (Nemser, 1971, p. 115-123), secondly that not all errors could be traced to interference by the source language (Duskova, 1969, p. 11136) and thirdly that whal GA [)redicted as likely diniculties did not always turn out to be so (Nickel, 1971, p. 219-227).
As a result of the intense debate that has raged around the value of the CA hypothesis, two distinct versions have emerged: CA apriori or predicative or strong version and aposteriori or explanatory or weak version.Initially the distinction between the two versions rests on a diilerence of point of view.The strong version, as has already i)een said, departs I'rorn a point of view in which it wishes to predicl errors which will occui in the target language.'I'he weaker version makes no such claims.It wishes to exfdain why errors in the target language occur.Assuming that learners of language A arc found by the process known as error analysis to make recurring errors in a particular construction in their attempts to learti language B, the analyst makes an analysis of the construction in language H and the comparable construction in language A in order to discover why the errors occur.In this way the analyst discovers what learners actually do in their attempts to acquire another language instead of predicting what they will do (Schachter,197.5,p. 206).'I'his approach seems to be more worthwhile to the practising teacher.
W hal then, is an error analysis?
In liis use of a language tlie language user employs a set of cognitive structures acquired by some process of clata-processing (rules of grammar) anel hypothesis formation (language exposure) in which the making of errors is evidence of the actual learning process.It may even be argued that the making of errors is a prerequisite to this learning activity (Nickel, 1973, p. 24)." You can't learn without goofmg."(Dulay and Burt, 1974, p. 95).At the same time it may also be argued that the degree of error is indicative of I he degree of competence achieved by the learner; that is, if target language acquisition is seen as the possession of a certain kind of knowledge (competence) instead of dispositions to respond to certain stimuli.
N. Chomsky's well-known distinction between what the speaker knows about his language (competence) and what actually happens when he speaks the language (performance) is of relevance here.The analyst has to distinguish between the speaker's intrinsic knowledge (or lack of knowledge) of the LT and his "mistakes" which are similar to mistakes made by native speakers and cla.ssified by other speakers as "ungrammatical" .Very often performance is an imperfect reflection of competence: e.g. the fact that people occasionally have "sli])s of the tongue" in every day conversation does nog mean that they don't know their language, or don't have fluency (i.e.competence) in it.Dusková (1969, p. 12) makes the following distinction between errors in competence and errors in performance: " . . .mistakes which are tlcfined as adventitious, random errors in performance due to memory lapses, physical states, etc. of which the speaker is inmiediately aware, and systematic errors which rcilect a defect in knowledge" .
The key words that should be noted are "random errors" and "systematic errors" ." Random errors" , such as hie for ihe may be ignored in an Error Analysis.
Once embarked on the design of an Error Analysis the analyst is usually confiontcd with the problem of what to do with the error analysis.It is agreed that a tnere listing ol errors will be of limited value only."No errors are errors in themselves" (Enkvist, 1972, p. 19).'I'he aim should be to provide an adequate linguistic explanation of the natiue of the errors found in a particular learning situation.A possible framework of a typical error analysis could include the following: (a) types of errors (cla.ssillcation with regard to the target system) i.e. classification according e.g. to grammatical categories; (b) frequency of errors, in order to determine the seriousness of various errors; (c) explanation of errors in linguistic terms; (d) degree of dislin baiicc caused by errors; and, (e) therapy (how teaching should be arranged so as to eliminate the errors).

(a) Types o f errors:
G ram m ar can be divided into siib-categorics of morphology, noun phrase, verb phrase and syntax.Lexis may have three subcategories; content words, function words and common expressions.Performance mistakes may be ignored.The system has many ailvantages as it is easy to handle, it may be extended to include new subcategories and comparisons can be made between main categories and sub-categories.It is suggest that the analysis be based on free production -i.e.short paragraphs.In other types of test material, such as grammar tests and multiple choice tests there is a danger in that they have an inherent classification built into them because a test constructor will include items which he thinks arc important for the students to know whereas the analyst wants to determine the actual conipetencc of his students.

(b) Frequency
The errors are evaluated to determine frequency of occurrence.The frequent occurrence of a specific form or the evasion of a specific form leading to circumlocution and consequently, clumsy constructions is indicative of the learner's lack of competence at this point in his learning process.The seriousness or degree of deviation from the norm is open to various interpretations.James (1974) proposes an assessment based upon the rules trangressed.Johanssen (1973) suggests that the degree of disturbance an error may have on the elFiciency of communications should be investigated.There is obviously a diiferencc in the degree of com prehensibility between "the two men is walking down the street" , and "a poem should be short and sinful" .

(c) Explanation
It is necessary to find some linguistic and psycholinguistic explanation for the occurrcnce of the errors in the analysis.

A litiguistic explanation
Back-lranslalion may be used to discover what the pupils intended to say and then the equivalent Afrikaans structure e.g. is compared to the correct English structure, to determine the type of error committed by the pupils.
E.g. * a poem should be short and sinful 'n gcdig behoort kort en sinvol te wees.
Use was made of a one-to-one relation in translating from the source language.
From a lexical entry such as: aktucel: actual, real, timely, topical, of current interest, learners select actual because it is close to the source language in sound and structure.
* He writes about .something that is actual.

A psycholinguistic explanation
'I'hrec principal cau.ses for error can be investigated: 2.1 Interference from the source language: * He walk home Hy loop huistoe.
2.2 Intralanguage interference resulting not from the source language but from the target language itself.* 'I'he girls walks home.
2.3 Faulty teaching techniques of materials or develoijmental errors referred to as a process of " hypothesizing false concepts" or as "induced errors" .e.g.'i'he u.se of ihc present progressive tense where the present indefinite tense sufllces.
(d) Degree of disturbance caused by errors A " tolerancc study" or degree of irritation that native speakers have for an error is another criterion for evaluation purposes.Edlciency ofconimuiiication is liowevei' of major importance.
(e) Therapy 'I'o my mind, language teacliing is no more than the provision ol suitalile conclit ions for language learning.A contrastive analysis should provide each learner with the right conditions at the right time.
W hat then is the rclevance of all this to language teaching?What is the " normal" sit\iation in a cla.ssroom?The average tcacher has a syllabus, from which he draws u]) a scheme of work which he proposes to leach to his pupils.
It may be argued that the teacher knows intuitively which errors his pupils arc going to make, but this is a debatable point.We so often find that teachers teach and i e-teaeh elements of grammar at Secondary Scliool which the pupils had already mastered at Primary School level.At the same time they give little or no attention to other components of the language (e.g.extension of the pupils' lexicon).If the teacher really wishes to achieve target teaching he has to determine what the needs are of his pupils.Once he h<-is determined the problem areas of a specific group of pupils, he can devise a meaningful scheme of work (and this may dilTer from group to group).Therapy or remedial work can only be attempted if and when the tcacher has determined which errors pupils really make.
Indications of the pedagogical relevance of the studies discussed fall into three categories: the problem of correction; the design of syllabusses and remedial programmes; and the writing of pedagogical grammers.All these arc related to those studies which I have called error analysis.
'I'he problem of correction is two-fold: what to correct and how to corrcct.The first c|ucstion is conccrned with the assessment of the gravity of the error in terms of its interference with comprehensibility or the degree of linguistic deviance.The need must be stressed that we have to encourage learners to communicate and that wc have to devise correction techniques with this always in mind, for example by requesting rephrasing or amplifying the message, in the way that adults react to infants' utterances in their mother longue.
The relevance of the actual performance of the students as revealed by the EA to the designing of syllabuses is based on the notion that there is some "natural" sequence ofelal)oration of the approximative system ol the .secondlanguagelearner, Nemser (1971) and that when this can be well established it would provide a psychological logic to the ordering of material in a f)0 syllabus.
As far as ihc design of pedagogical grammars is concerned, the effectiveness of the presentation and practising of linguistic materials must ultimately depend upon whal is discovered about the actual process and strategies of language learning, Allen (1973).
What is' required is a more intelligent realistic appraisal of language materials based on a Contact Analysis approach especially in a language contact situation as we have in South Africa.