Contrastive Analysis & Error Analysis and Interlanguage

Unit 4 Contrastive Analysis, Error Analysis and Interlanguage

This unit focuses on:

* Contrastive Analysis

* Error Analysis

* Interlanguage

A starting point in the research on second language acquisition is the study of learner language because learner language can reveal insights into the process of acquisition. The description of learner language constitutes the most important source of information about how learners learn a second or a foreign language when he/she gets things wrong, because learner errors can hold vital clues abut the processes of FL learning. The errors can be analyzed to work out why they are made in order to study the developmental patterns of learner language.

Contrastive Analysis sought to predict the errors that learners make by identifying the linguistic differences between their L1 and the target language. Its underlying assumption was that errors were mainly the result of interference was believed to take place whenever the ? habits‘ of the native language differed from those of the target language. Contrastive Analysis gave way to Error Analysis because of the belief that the errors learners make can be influenced by a variety of factors. Error Analysis can be used to investigate the various processes that contribute to interlanguage development. Interlanguage is viewed as a unique linguistic system that L2 learners construct by drawing on their L1 and the target language.

Contrastive Analysis studies the comparison between mother language and target language; Error Analysis studied the comparison between interlnaguage and target language; and transfer studies the comparison between mother language and interlanguage. This unit will discuss these concepts and provide some explanations about the relationship between them.

Contrastive analysis

As a scientific comparison method, Contrastive Analysis ( CA ) has experienced a rise and fall in the last 60 years of its development. CA was a traditional method used in the field of translation and it was adopted in the 1940s as a special method to study foreign language learning and teaching. The term ? Contrastive Analysis‘ was first used by Benjiamin L. Whorf in 1941, but Charles Fries was regarded as the founder of modern contrastive linguistics and as the first man to apply contrastive analysis to foreign language learning and teaching.

Contrastive Analysis had grown to be a major concern of applied linguistics in 1950s. Robert Lado‘s book ? Linguistics Across Cultures‘, published in 1957, says that ― The teacher who has made a comparison of the foreign language with the native language of the students will know better what the real problems are and can provide for teaching them‖. He made a systematic study of contrastive analysis in his book and showed how to use the CA method to compare linguistic systematic. However, in 1960s, CA was challenged by Chomsky‘s TG grammar and it received a lot of criticism. CA became unfashionable for a time, but it was never truly killed off. CA underwent a rapid development in 1970s. ? The Pacific Conference on Contrastive Linguistics and Language Universals‘ was held in the University of Hawaii in 1971. Different schools of CA were formed in 1980s to study its principles, procedures and its applications in foreign language learning and teaching.

The Contrastive Analysis Hypothesis

Contrastive Analysis can be defined as systematic comparison of specific linguistic characteristics of two or more languages. The CA Hypothesis claimed that the principal barrier to second language acquisition is the interference of the first language system with the second language system and that a scientific, structural analysis of the two language would yield a taxonomy of linguistic contrasts between them which in turn would enable the linguist and teacher to predict the difficulties a learner will encounter. In this sense, CA is an applied contrastive study, which is concerned not only with the comparison and contrast of two or more languages to determine the differences as well as similarities between them, but also with the identification of probable areas of difficulty in learning another language.

CA was deeply rooted in the behavioristic and structuralist approaches. Behaviorists believed that human behavior is the sum of its smallest parts and components, and therefore that language learning could be described as the acquisition of all of these discrete units. Behaviorists also believed that learning is a kind of habit formation. Old habits get in the way of learning new habits and the effects of one habit on learning another is known in psychology as the study of Transfer. Language learners attempt to transfer the features of their mother tongue to the second language they are learning. Transfer will be positive when their first and second language habits are the same. When the two languages are different in structures, negative transfer, also called interference, occurs. Thus differences between the first and second languages create learning difficulty which results in errors, while the similarities between the first and sec ond languages facilitate rapid and easy learning.

Some strong claims were made for the CA hypothesis among language teaching experts and linguists. The strong version of the CA hypothesis claims that all L2 errors can be predicted by identifying the differences between the target language and the learner‘s first language. It was believed that comparing native and target languages would tell you almost everything you needed to know to devise a language-teaching program. As Lee notes, it stipulates that ― the prime cause, or even the sole cause, of difficulty and error in foreign language learning is interference coming from the learner‘s native language‖. One of the strongest claims was made by Robert Lado ( 1957 ) in the preface to ? Linguistics Across Cultures‘: ―The plan of the book rests on the assumption that we can predict and describe the patterns that will cause difficulty in leaning, and those that will not cause difficulty, by comparing systematically the language and the culture to be learned with the native language and culture of the student.‖Then, in the comparison between native and foreign language lies the key to ease or difficulty in foreign language learning. Those elements that are similar to the learner‘s native language will be simple for him and those elements that are different will be difficult. Another strong claim was made by Banathy, Trager, and Waddle ( 1967 ):― The change that has to take place in the language behavior of a foreign language student can be equated with the differences between the structure of the student‘s native language and culture and that of the target language and culture.‖

In 1970s, the strong CA hypothesis came under severe criticism. Ronald Wardhaugh called the strong version of the CA hypothesis a version that was quite unrealistic and impracticable. His viewpoint which was also shared by others led to the development of the weak version of the CA hypothesis. The weak version recognizes the significance of interference across languages and the fact that such interference can explain difficulties, but it also recognizes that not all linguistic difficulties can be predicted in advance and the later explanation of these difficulties would be more profitable. As learners are learning the language and errors appear, the teacher can utilize his

knowledge of the target and native languages to understand sources of error. Clearly, gross predictions can be made up of hundreds of thousands of items, and it is impossible to predict difficulty beyond some very noticeable phonological differences between two languages. In fact, it is only in the phonological component of language that contrastive analysis is mildly successful. Syntactic, semantic, or lexical interference is far less predictable. The hypothesis is ?weak‘because it does not claim any predictive power for CA to foresee errors with any certainty in advance.

A compromise between the strong and weak versions of the hypothesis was proposed by Oller and Ziahosseiny. They claimed that ―The categorization of abstract and concrete patterns according to their perceived similarities and differences is the basis for learning; therefore, wherever patterns are minimally distinct in form or meaning in one or more systems, confusion may result.‖In other words, the learning of sounds, sequences and meanings will be the most difficult where the most subtle distinctions are required either between the target language and native language or within the target language itself. According to them, the greatest difficulties in L2 learning are neither apparent similarities nor apparent differences, but subtle distinctions between two languages. Interference can be greater where such subtle distinctions exist. Oller and Ziahosseiny‘s moderate version has more explanatory power, and it has put the CA hypothesis into some perspective, because it rightly emphasizes the generalizing nature of human learning, since minimal differences are generally overlooked. Greater differences, because of their saliency, do not always result in greater learning difficulty.

In conclusion, the strong version emphasizes the prior prediction of difficulties in learning a second language. The weak version only recognizes the significance of interference across languages and tries to explain those difficulties coming from interference. However, both versions equate differences between two languages with difficulties in L2 learning. The moderate version differs from the other two in that it emphasizes the significance of minimal distinctions, which actually cause the greatest interference and difficulties.

Procedures of Contrastive Analysis

The model initially used to compare two languages was that of structuralist linguistics. Structuralists emphasized the importance of scientific description of language—a description of the different categories that make up the patterns of a language. Structural linguistics provided the tools to accurately describe the two languages and to match those two descriptions against each other to determine the differences and similarities between them. Most of the contrastive studies have been based on surface structure characteristics. Randal Whitman (1970) noted that CA involves four different procedures.

The first of these is a formal description of the relevant features of each language compared. The linguist or language teacher, using the tools of formal grammar, explicitly describes the two languages in question.

The second is a selection of certain linguistic items, which may be entire subsystems such as the auxiliary system, since it is virtually impossible to contrast every possible facet of two languages. Whitman admits that the selection process ―reflects the conscious and unconscious assumptions of the investigator‖, which in turn affect exactly what forms are selected.

The third is the comparison and contrast itself, that is, the mapping of one linguistic system onto the other, a specification of the relationship of one system to the other, and the identification

of areas of difference and similarity.

Finally, one formulates a prediction of error or difficulty on the basis of the first three procedures. That prediction can be arrived at through the formulation of a hierarchy of difficulty or through more subjective applications of psychological and linguistic theory.

The Application of Contrastive Analysis

CA has traditionally been used to predict and explain the problems in L2 learning. Languages differ and linguistic differences in two languages cause learning problems. The greater the difference between the two languages systems, the greater the learning problem and the potential area of interference. Prediction of learning problems and diagnosis of learner errors can help teachers better their teaching and remedy their work. Knowing why a learner has committed the errors help the learner monitor and avoid the same errors in the future.

CA can also be used in developing course materials for language teaching. CA can specify those features of L2 which are different from corresponding features of the L1. This knowledge is important in the selection of teaching materials, because the identical features in two languages should take less time than the different features in teaching materials. Fries (1945) stated that ―The most effective materials are those that are based upon a scientific description of the language to be learned, carefully compared with a parallel description of the native language of the learner‖.

Criticisms of Contrastive Analysis Hypothesis

Although CA has been used by many teachers to organize their lessons and select teaching materials, it has suffered severe criticisms since the early 1970s. The criticisms of the CA hypothesis were of three major types.

First of all, there were doubts concerning the ability of CA to predict errors in language learning. It was shown in research that not all errors were caused by the interference of the mother tongue. For instance, English beginners often produce sentences like ?He speaked English‘ and ?I cuted myself‘. The interference of the mother tongue can‘t satisfactorily explain these errors. It is more reasonable to say that the learner knows some rules of English but applies them in the wrong place. Thus the leaner overgeneralizes the rules for the change of verb forms. So the error is a result of overgeneralization of the target language system. It is obvious that the psychological and linguistic basis of CA is clearly defective, because it is based on behaviorism and structuralism.

Secondly, there were a number of theoretical criticisms regarding the feasibility of comparing languages and the methodology of contrastive attacked by Chomsky and others, and the terms ?stimulus‘, ?response‘ and ?reinforcement‘ were rejected as an inadequate explanation of language learning. In addition to this, there were objections to the validity of equating ?difference‘ with ?difficulty‘on the one hand and ?difficulty‘with ?error‘on the other. Briefly, ?difference‘is a linguistic concept, while ?difficulty‘is a psychological concept. Therefore, the level of learning difficulty can‘t be inferred directly from the degree of linguistic difference between two language systems. The research also showed that items predicted to be difficult on the basis of a contrastive analysis did not in fact produce errors. So there was no necessary relationship between difficulty and error.

Thirdly, there were doubts about whether CA had any practical worth to language teaching. CA studies concentrated largely on grammatical aspects—the phonological level, the morphemic level and the syntactical level. There is practically little contrastive analysis above the sentence

level, let alone the textual or discourse level. In addition, if a majority of learner errors are not caused by interference, then CA is of limited value. However, the main doubt about CA from a pedagogic point of view has arisen from the changing attitudes to the role of errors in language learning. CA was based on the need to avoid error, but if error is seen as a positive aspect—a necessary developmental process of language learning, then why should we devise a teaching program to prevent it?

Recent research

Within the last few decades, the importance of language transfer in foreign language learning has been reassessed several times. Recently, a more balanced perspective has emerged in which the role of transfer is acknowledged and in which transfer is seen to interact with a host of other factors in ways not fully understood argues that ―native language contributes to L2 acquisition‖, and that ―the popular belied that the role of L1 is negative needs to be reconsidered in the cognitive framework‖. When the L2 learners experience difficulty in communicating an idea because they lack the necessary target language knowledge, they will resort to their L1 to make up the insufficiency. Therefore, the concept of ―interference‖should be reframed as a ―cognitive process‖, a ―strategy‖or an ―intercession‖( Lu Xiaoyong 2002). L1 should be viewed as a resource which learners can use for ad hoc translation to overcome their limitations. Lu‘s empirical studies indicate that L1 plays a positive role in L2 learning and literacy of Chinese students‘native language correlates significantly with the acquisition of literacy in the English language.

In recent years, CA has been widely applied to more extensive aspects including textual, pragmatic, discourse and cross-cultural levels. CA begins to consider not only linguistic contrasts but also pragmatic contrasts such as the similarities and differences in the stylistic uses of items in the first and second language and in the form-function relationships. Contrastive pragmatics is a fairly recent development of this kind. Recent researches also prove that error is a multi-factor phenomenon and that interference, as one of the factors, interacts in complex ways with other factors.

Error Analysis

The analysis of learner errors has a long tradition. Before the early 1970s, the goal of error studies was mainly pedagogic and the information provided by errors was used to sequence linguistic items for teaching. As we saw in the previous section, CA had been the prevailing technique to study errors of L2 learners for nearly two decades. CA overstressed the interfering effects o mother language on target language learning and ignored the other factors, which interacted with interference, to result in errors. Another approach, Error Analysis, developed as a branch of applied linguistics in the late 1960s and claimed that many learner errors were not due to the learner‘s mother tongue but reflected universal learning strategies. Error Analysis was therefore offered as an alternative to CA. Attempts were made by many researchers to develop classifications for different types of errors on the basis of the different processes that were assumed to account for them. But such attempts have been problematic due to the difficulty of determining the cause of errors. By the late 1970s, Error Analysis had largely been replaced by studies of Interlanguage and Second Language Acquisition. Later Error Analysis was looked on only as an applied science to help to explain the problems in language teaching and learning.

It was believed that CA only studied the learner‘s native language and the target language,

while Error Analysis provided a methodology for investigating learner language. For this reason EA constitutes an appropriate starting point for the study of learner language an L2 acquisition.

Basic assumptions of Error Analysis

Error Analysis ( EA) is defined as ―the study and analysis of the errors made by second language learners‖( Longman Dictionary of Language Teaching ＆Applied Linguistics, 1998). As mentioned above, EA aims to identify strategies which learners use in L2 learning and to find out the causes of learner errors. And many researchers who carried out error analysis were concerned with obtaining information on common difficulties in L2 learning, which could be used as an aid to teaching or to the preparation of teaching materials. According to Brown, EA is based on the following assumptions:

Human learning is fundamentally a process involving the making o mistakes, which forms an important aspect of learning any skill or acquiring any knowledge. Learning to swim, to play tennis or to drive all involves a process in which success comes by profiting from mistakes.

Language learning is like any other human learning and L2 learning, like L1 acquisition, is a trial-and-error process in nature. L2 learners will produce ungrammatical or even ill-formed utterances if judged by the generally accepted rules of the language they are learning.

The errors made by L2 learners provide some of the keys to the understanding of L2 learning process. Therefore, L2 learners are not looked upon as producers of imperfect language full of mistakes, but as creative beings proceeding through logical, systematic stages of acquisition.

Error play an important role in L2 learning and teaching. The learners have to construct a new system of language during the process of the L2 learning. As P.Corder (1967) noted: ―A learner‘s errors…are significant in that they provide to the research evidence of how language is learned or acquired, what strategies or procedures the learner is employing in the discovery of the language.‖

The kinds of mistakes or errors

The following is a table of common mistakes made by learners. This category is popular with teachers in terms of its practicality ( 1992).

（2）Surface strategy taxonomy of errors

Applied linguists (Dulay, Burt, and Krashen 1982) tend to classify what learners get wrong in another way, describing errors using surface strategy taxonomy. There are four kinds of errors in

The above is only a simple categorization of errors. When we present grammar points, we tend to use the categories presented in Azar‘s ( 1992) table. In checking students‘ written work, teachers tend to use surface strategy taxonomy.

Procedures of Error Analysis

P. Corder (1974) proposes five stages in EA research: 1) collection of a sample of learner language; 2) identification of errors; 3) description of errors; 4) explanation of errors; and 5) evaluation of errors. Another way of describing EA procedures consists of the following five steps:

Identification of errors----Description of errors-----Explanation of errors-----Evaluation of errors-----Correction/prevention of errors

The procedures involved in each of these steps will be illustrated in detail in the following section.

1)Identification of errors

Before identifying and analyzing learner errors, it is necessary to make a distinction between errors and mistakes.

A mistake refers to a performance error which is either a random guess or a slip of tongue, that is, mistakes reflect occasional lapses in performance. Mistakes occur because, in a particular instance, the learner is unable to perform what he or she knows. All people make mistakes in both L1 or L2 situations. However, native speakers are normally capable of recognizing and correcting such mistakes, which are not the result of a deficiency in competence but the result of some sort of breakdown or imperfection in the process of producing speech. Such kind of mistakes in native-speaker production also occurs in second language speech. That is to say, mistakes are made at random and they are not systematic. An error is a noticeable deviation from the adult grammar of a native speaker, reflecting the interlanguage competence of the learner or gaps in a learner‘s knowledge. Errors occur because the learner does not know what is correct in a know n system. Therefore, errors can reveal the learner‘s competence in using the target language and they are systematic. But it is clear that identifying the exact errors that learners make is not easy, because they can only be inferred from error of performance.

P. Corder (1971) proposes an elaborate procedure for identifying errors in a second language. This procedure consists of the following steps.

The first step is to identify overt and covert errors. Overt errors refer to ungrammatical sentences, and convert errors refer to superficially well-formed sentences, but which make no sense in the context. Covertly erroneous utterances are grammatically well formed at the sentence level but are not interpretable within the context of communication. Overt error is concerned with grammaticality, while covert error deals with appropriateness. When a learner‘s sentence is analyzed, two questions should be asked: (1) Is the sentence superficially well formed in terms of the grammar of the target language? and (2) Does the normal interpretation according to the rules of the target language make sense in the context? If the answer is ?No‘ to the first question, it is an overt error and if the answer to the second question is ?No‘, it is a convert error. For instance, the sentence ?I‘m fine, thank you.‘ Is grammatically correct at the sentence level, but if it is the answer to the question ?Who are you?‘, it is uninterpretable at discourse level.

The second step is to provide interpretations and construct well-formed sentences in the target language. If a sentence contains overt or covert errors, a plausible interpretation can be put on the sentence in context. If the sentence is interpretable, reconstruct a sell-formed sentence in the target language and then compare a reconstructed sentence with the original idiosyncratic sentence to find out in what respect they differ.

If the sentence is uninterpretable, translate the sentence literally into the mother tongue and see if it is understandable in the context. If it is so, retranslate the sentence into the target language to provide a reconstructed sentence. But, if the mother tongue of a learner is not known or the restructured sentence is impossible to understand, hold it in store.

However, there are a number of major methodological problems with the procedures used in error identification. The distinction between errors and mistakes is not easy to put into practice. In addition, this distinction does not take account of the possibility that learners‘knowledge is variable. Also, it is not at all clear whether Corder‘s suggestions for identifying covert errors will

work. Such a procedure assumes that learners possess the necessary metalingual knowledge to talk about their own performance. This assumption may not be justified in the case of children and some adult learners.

2)Description of errors

Once errors are identified, the next step is to describe them adequately. The description of learner errors involves two aspects: comparison and classification.

Comparing the learner‘s erroneous sentence with a reconstruction of the sentence in the target language, which the above procedure has only begun to accomplish, requires attention to surface properties of the learner‘s sentence. For instance, in the learner‘s sentence ?He comed yesterday.‘ and the reconstruction of the sentence ?He came yesterday.‘, the error ?comed‘ is compared with the correct form ?came‘ in terms or the same grammatical category.

Then errors are classified into different types, mostly based on linguistic categories.

Errors can be described from the dimension of language levels. Some general linguistic categories, such as the skeleton of English clauses, the auxiliary system is subdivided into ?do‘, ?have‘, ?be‘, modals, etc., while errors in the use of ?do‘are classified according to where they are used, in a question or in a negative sentence. A number of different categories for the description of errors have been identified in research on learner language. The more general categories are phonology, morphology, lexicon, syntax, and discourse. Of course, it is difficult to distinguish different levels of errors. A word with a faulty pronunciation, for example, might hide a syntactic or lexical error. In addition to linguistic classification, surface strategy taxonomy can be used to categorize errors into omission, addition, misformation, misordering and blends.

Errors can also be described from the dimension of learning stages. P. Corder (1974 ) distinguishes three types of error according to their systematicity. At the presystematic stage, the learner is unaware of the existence of a particular rule in the target language and the use of the particular rule is through random guessing. At the systematic stage, the learner has discovered some rules of the target language system, but they are the wrong ones. Usually he is unable to correct his own errors although he can explain the mistaken rules he used. At the postsystematic stage, the learner knows the correct rules of target language but he is inconsistent in applying them. The learner can both explain the target language rules and correct his own errors. The occurrence of errors becomes infrequent though they still exist.

The description of errors, like their identification, is problematic. Even if the error itself can be easily identified, it is often problematic to determine what the error consists of. For example, if a learner produces the sentence, ?I am worried in my mind.‘, it is not clear what constitutes the best reconstruction. One possibility is ?I am feeling worried.‘Another is ?I have a problem on my mind.‘ Even if the learner is available for consultation, it may not be possible to choose between these two reconstructions. But the description of the error will obviously vary according to which reconstruction is covertly idiosyncratic sentences will prove even more difficult.

3)Explanation of errors

The explanation of errors is concerned with establishing the sources of errors. Why are the errors made? What underlies these errors? This step is very important because some errors can reflect learners‘ attempts to perform the task of learning and to make L2 use more simple, and through this we can conclude some regular rules responsible for L2 acquisition.

As Taylor (1986) points out, the error source may be psycholinguistic, sociolinguistic, epistemic, or may reside in the discourse structure. Psycholinguistic sources concern the nature of the L2 knowledge system and the difficulties learners have in using it in production.

Sociolinguistic sources involve such matters as the learners‘ ability to adjust their language in accordance with the social context. Epistemic sources concern the learners‘lack of world knowledge, while discourse sources involve problems in the organization of information into

a coherent ?text‘. In general, however, the research on L2 acquisition has attended only to the

first of these. As Abbott puts it: ―The aim of any EA is to provide a psychological explanation‖.

Psycholinguistic sources can be transfer or interference error and intralingual error.

According to Richards ( 1971), interference errors occur as a result of ―the use of elements from one language while speaking another.‖For instance, when a German learner of L2 English says ?I go not‘because the equivalent sentence in German is ?Ich gehe nicht‘.

Intralingual errors ―reflect the general characteristics of rule learning such as faulty generalization, incomplete application of rules and failure to learn conditions under which rules apply‖.

Transfer errors can be further subdivided. Lott ( 1983), for instance, distinguishes three categories. ?Overextension of analogy‘occurs when the learner misuses an item because it shares features with an item in the L1. ?Transfer of structure‘ arises when the learner utilizes some L1 feature rather than that of the target language. ?Interlingual/intralingual errors‘ arise when a particular distinction does not exist in the L1.

Intralingual errors can also be further subdivided. Richards (1971) distinguishes the following four categories. ?Overgenralization errors‘ arise when the learner creates a deviant structure on the basis of other structures in the target language. ?Ignorance of rule restrictions‘involves the application of rules to contexts where they do not apply.

?Incomplete application of rules‘involves a failure to fully develop a structure, because th learner finds he can achieve effective communication by using relatively simple rules. ?False concepts hypothesized‘ arise when the learner does not fully comprehend a distinction in the target language.

Distinguishing transfer and intralingual errors in not an easy job, neither is identifying the different types of intralingual errors that Richards describe. In an attempt to deal with th problem of identifying sources, Dulay and Burt (1974) classified the errors they collected into three broad categories:

(1)Developmental errors, i.e. those that are similar to first language acquisition.

(2)Interference errors, i.e. those that reflect the first language structure.

(3)Unique errors, i.e. those that are neither developmental nor interference.

Dulay and Burt‘s research has often been criticized on the grounds that reliable classification of errors in terms of these categories is still not possible. However, it might be argued that by using L1 acquisition errors as a baseline they at least provide an operational procedure for establishing which errors ate intralingual.

Selinker (1972) identified five sources of errors in the development of L2 learner‘s interlanguage. They are language transfer, overgeneralization of target language rules, strategies of L2 learning, transfer of training and strategies of L2 communication.

In reality, L2 learner errors are so complicated and diversified that they can be traced to

various sources. Major sources of errors are interlingual transfer, intralingual transfer of learning, inference, communication strategies, cognitive and affective factors.

4)Evaluation of errors

Human learning is fundamentally a process that involves the appropriate errors at an appropriate time. So making errors is unavoidable during L2 learning, but the problem is what attitude teachers should take towards learners‘ errors. As Carl James ( 1998 ) noted: ― We do not seek to hone the analytical scalpel so as to lay the tiniest error, but the opposite: to prevent obsession with trivial errors and give priority to the ones that rally matter.‖Evaluating an error means assigning relative values and standards to it, that is, deciding which error merits attention.

Carl James (1998 ) suggests several criteria in assessing error gravity. The paramount is the linguistic criteria which judge error gravity by formal features of language. The next is a quantitative measure of error gravity: error density and error frequency. Error density is calculated by counting how many different errors occur per unit of text, while frequency is a measure of how many times the same error is repeated over say 100 words of text. Compared with high error frequency, high error density creates more difficulty in comprehension. Another criterion is the comprehensibility, which includes intelligibility and communicativity. Intelligibility, a basic standard, refers to the accessibility of the basic, literal meaning, the propositional content encoded in a sentence. Communicativity, a standard of higher level, involves access to pragmatic forces, implicatures and connotations. The noticeability of errors involves two aspects: What errors have been found? Who found the errors? Where English teaching in China is concerned, the Chinese teachers who teach English pay more attention to the errors of language forms, while the teachers who are native speakers focus on information-getting and the effect of errors on comprehension.

Up till now there is not a unified standard to evaluate errors. The key point is how the teachers treat learner errors in language learning. Some teachers are tolerant to errors because they believe that making errors is a necessary and unavoidable process of learning a second or foreign language. Thus, foreign language learners should be encouraged to take risks in correct errors. Other teachers take the intolerant attitude to errors and regard them as a sign of the learner‘s failure to master the rules being taught. They believe that errors should be corrected during the process of teaching.

5)Correction of errors

One of the EA objectives is to find the principles which can guide effective error correction.

Carl James ( 1998 ) uses the term ?correction‘ with three senses: feedback, correction and remediation. Feedback refers to ―informing the learners that there is an error, and leaving them to discover it and repair it themselves.‖ Providing treatment that leads to the same error from recurring later is called correction proper. Providing learners with information that induces them to revise the wrong rule in their mind so that this type of error does not recur.

Such processes of improvement for all future productions are called remediation.

Whether teachers should correct learners‘ errors and how they correct them are argued for

a long time in the language teaching field. Some researchers suggest distinguishing global

errors from local errors. Global errors, which violate the overall structure of a sentence, may hinder communication and badly interfere with L2 learning. For example, ?The little girl was in the corner cry…‘. The wrong sentence structure creates a processing problem and prevents the hearer from comprehending some aspect of the message, and for this reason, such an error should be corrected immediately. Local errors, which affect only processing problems

and they do not interfere the L2 learning relatively. For example, ?He go to see a film last night.‘ Such local errors need not be corrected since the message is clear and correction may interrupt learner‘s communication. But unfortunately many utterances are not clearly global or local, and it is difficult to discern the necessity for correction.

The matter of how to correct errors becomes exceedingly complex. The research shows that there is no single and effective method for error correction. Error treatment options can be classified in number of ways. One useful taxonomy is recommended by Bailey (1985), who listed seven basic options are 1) to treat or to ignore, 2) to treat immediately or to delay,

3) to transfer treatment or not, 4) to transfer to another individual, a subgroup or a whole

class, 5) to return, or not, to original error maker after treatment, 6) to permit other learners to initiate treatment and 7) to test for the efficacy of the treatment. Possible features include 1) fact of error indicated, 2) location indicated, 3) opportunity for new attempt given, 4) model provided, 5) error type indicated, 6) remedy indicated, 7) improvement indicated and 8) praise indicated. All of the basic options and features within each option are conceivably viable modes of error correction in the classroom. The teacher needs to develop the intuition for determining which option or combination of options is appropriate at given moments.

In conclusion, error analysis consists of five steps. The first step is to collect a massive sample of learner language and identify the errors in the sample. Corder proposes an elaborate procedure for identifying errors in a second language. Description of learner errors involves comparison and classification of errors. Errors can be classified by levels of language or by learning stages. The explanation of errors is concerned with establishing the sources of errors. The aim of EA is to provide a psychological explanation. Evaluation of errors provides some criteria in assessing error gravity and teachers‘attitude to learners‘errors. Correction is the final objective of EA, however, there is no single and effective method for error correction.

The reasons of errors

There is a lot of research on why learners make mistakes. The following are explorations from different perspectives.

Richards ( 1971) provides three reasons why learners make mistakes. They are interference errors, intralingual errors and developmental errors. Interference errors are also called transfer errors which refer to the use of elements from one language while speaking the other. Intralingual errors happen within one language, meaning a faulty generalization, incomplete application and failure to learn conditions under which rules apply. The last reason is about developmental errors. When learners are trying to build up hypothesis about the target language on the basis of limited language knowledge and skills, they make errors, but such errors are of progressive nature as they are part of the learning process towards proficiency.

Richards (1971) further provides a description of intralingual errors as four kinds that are

When we look at students‘ errors and mistakes longitudinally, we can notice that students make mistakes in a developmental way. The acquisition proceeds in a regular, systematic and progressive fashion.

The following table shows the general order of language acquisition in speaking and writing.

↑

Fossilization

↑

proficiency in grammar and pragmatic

( acquired accuracy and fluency )

↑

use English

( self-correction)

↑

Correct grammar/inappropriate in pragmatic

↑

some mistakes disappear; some others appear

↑

more practice ( more mistakes)

↑

independent practice/watched over

(making mistakes)

↑

controlled practice/ non-acquired accuracy

↑

silent period

The above is only a rough picture of a second language acquisition order, with variations for different learners. From this we can see learners first are quiet, but it does not mean they are not learning. Actually they are taking in language input which is being processed in the black box in their mind. After some time, learners begin to produce language which can be relatively correct in controlled or well-planned situations, but as they begin to use English, they start to make mistakes as they are unable to control grammatical and vocabulary forms and means. Those mistakes emerge together with their attempting. As learners further use the language, some mistakes disappear while others appear, but those mistakes are of a higher level, which means a comparatively less proficient learner has no ability or no chance to make those mistakes. Throughout such a process, a language learner develops his/her proficiency both grammatically and pragmatically. However, there is a saddening thing that does not exist in the first language. That is fossilization, which is a process that sometimes occurs in which incorrect linguistic features become a permanent part of the way a person speaks or writes a language. Aspects of pronunciation, vocabulary usage, and grammar may become fixed or fossilized in second or

foreign language learning.

Specifically, acquisition of a particular language item also follows a systematic sequence. Let us study the following table.

Sequence of acquisition (Ellis 2000) This table tells us that a learner does not acquire a particular item overnight. It goes through a process with different stages, each of which has its own characteristics. First of all, learners fail to produce the correct past form of the verb ―eat‖, and then they can be correct in the second stage. After some time, however, they are incorrect again which seems that their English is even worse than that at stage 2, but actually their English is better. They go through different stages until they really have acquired the past form of the verb ―eat‖. This sequence is like a U shaped model.

Non-acquired accuracy Acquired accuracy

U

Regression

The developmental nature of learners‘ language acquisition tells teachers that we should be patient, tolerant and optimistic in our teaching. We should be patient because making mistakes is an unavoidable process and sooner or later students will get better. We should be tolerant because no one can reach acquired accuracy immediately. They all have to undergo trying, regressing and rising up again. And we should be optimistic that making mistakes helps learning. However, that does not mean we will ignore students‘ mistakes without principles.Balanced attitudes and behavior are needed in dealing with students‘ mistakes.

Implications of Error Analysis in Language Teaching

EA is the first serious attempt to investigate learner language in order to discover how learners acquire an L2. It has proved to be a direct and effective method of studying the learning process and examining teaching situations. The significance of EA lies in its theoretical and pedagogical application. EA is part of the methodology of investigating the language learning process and its function is to guide the remedial activities to correct an unsatisfactory state of affaires in both learning and teaching. Therefore, EA is useful not only to the researcher but to the teacher as well.

EA provides feedback to the applied linguists and psycholinguists about the feasibility of theories of language learning. It also provides researchers with evidence of how the target language is acquired or learned, and what strategies the learner is using in his discovery of the language. EA is a useful way for teachers to study learners‘ difficulties systematically and to treat their errors in the classroom. As P.Corder (1967) noted ―Errors provide feedback, they tell the

teacher something about the effectiveness of his teaching materials and his teaching techniques, and show him what parts of the syllabus he has been following have been inadequately learned or taught and need further attention.‖

EA can be applied to foreign language teaching in the following ways. First, EA provides the teacher with explanations of the nature of errors, the possible sources of the errors and the way to improve them. Thus it can help the teacher to determine what errors to correct and how to correct them in the classroom. Secondly, the findings from EA can provide useful information for the designing of curricula and the selection and compilation of teaching materials although errors cannot be precisely predicted for each learner or a group of learners. Lastly, EA is beneficial to the organization of remedial teaching when a mismatch is discovered between the learners‘knowledge of the target language and the linguistic demands of some situations they have to enter.

In the end, one general implication that can be drawn from EA is that learners are constructing, either consciously or unconsciously, a system for understanding and producing sentences in the target language.

Criticisms of Error Analysis

EA has been most successful in accounting for learners‘errors that cannot be predicted by CA. It has also succeeded in raising errors from undesirable status to legitimate status in both teaching and research. However, there have been a number of criticisms of EA. The criticisms fall into two main categories: limitations in the scope of EA an methodological problems involving procedures of analysis. The major weaknesses of EA are as follows:

EA might affect the teacher‘s work with the result that he pays too much attention to learners‘errors in the classroom. Although errors can reveal a system of learner language in L2 learning, EA fails to provide a complete picture of his language. The foreign language teacher may become so preoccupied with noticing errors that the correct utterances in the second language go unnoticed. The teacher must beware of placing too much attention on errors and not lose sight of the value of positive reinforcement of clear, free communication. While the diminishing of errors is an important criterion for increasing language proficiency, the ultimate goal of L2 learning is the attainment of communicative fluency in a language.

Another criticism of error analysis is an overstressing of production data. Language is speaking and listening, writing and reading. The comprehension of language is as important as production. It so happens that production lends itself to analysis and thus becomes the prey of researchers; but comprehension data is equally important in developing an understanding of the process of second language acquisition. Production errors are only a subset of the overall performance of the learner.

In the explanation of errors, simple classification is inappropriately used to delineate sources of errors. The taxonomic formats used in the description imply that a particular error has only one source. However, it is generally accepted that language acquisition is an interaction between internal and external factors, then the same error must have at least two sources: one in the environment and one internal to the learner, and multiple sub-sources can be described within these two sources.

There is a confusion between error description and error explanation. Error description deals with the product of L2 acquisition whereas error explanation, which determines the origins of errors, deals with the process of L2 acquisition. The process of language acquisition involves the

interaction of the learner‘s internal processing mechanism with the external environment, whereas the product of language acquisition involves the learner‘s verbal performance. While the product can be directly observed and described, the process can only be inferred. However, many reports on error analysis confuse the causes of errors and their description. The lack of distinction results in controversy in the development of criteria to describe different error types and in the formulation of theories to account for those errors.

Interlanguage

As was noted in the previous section, the learners are, consciously or unconsciously, constructing a system when understanding and producing sentences in the target language. This mental system of L2 knowledge is often referred to as interlanguage, which refers to a systematic development of learner language.

The term ?interlanguage‘ was first used by Selinker in his thesis ―Language Transfer‖( 1969). V arious alternative terms have been used by different researchers to refer to the same phenomenon. Nemser (1971) created the term ?approximative systems‘, and Corder (1971) used ?idiosyncratic dialects‘and ?transitional competence‘. In 1972, Selinker wrote a paper ―Interlanguage‖, which established the status of ?interlanguage‘ in L2 acquisition. This section concentrates on the concept of interlanguage, its essential characteristics and some issues on interlanguage.

The concept of interlanguage

The term ?interlanguage‘was coined by the American linguist, Selinker, to refer to the interim grammars constructed by L2 learners on their way to the target language. Generally speaking, the term ?interlanguage‘reflects two related but different concepts: (1) the structured system which the learner construct at any single point in time, and (2) the range of interlocking systems that characterizes the development of learners over time. The interlanguage is thought to be distinct from both the learner‘s first language and from the target language. It evolves over time as learners employ various internal strategies to make sense of the input and to control their output. Hence, interlanguage is a unique linguistic system that L2 learners construct by drawing on their L1 and the target language. During the process of L2 acquisition, the learners find the similarities between L1 and L2, combine them to special rules of L2, and finally from their own rules of language learning, which are different from both L1 and L2.

According to Rod Ellis ( 1997), the concept of interlanguage involves the following premises about L2 acquisition.

(1)The learner constructs a system of abstract linguistic rules which underlies

comprehension and production of the L2. This system of rules is viewed as a

?mental grammar‘ and is referred to as an ?interlanguage‘.

(2)The learner‘s grammar is permeable. That is, the grammar is open to influence

from the outside ( i.e. through the input). It is also influenced from the inside . For

example, the omissions, overgeveralization, and transfer errors constitute evidence

of internal processing.

(3)The learner‘s grammar is transitional. Learners change their grammar from one

time to another by adding rules, deleting rules, and restructuring the whole system.

This results in an interlanuage continuum. That is, learners construct a series of

mental grammars or interlanguages as they gradually increase the complexity of

their L2 knowledge.

(4)Some researchers have claimed that the systems learners construct contain variable

rules. That is, learners are likely to have competing rules at any one stage of

development. However, other researchers argue that interlanguage systems are

homogeneous and that variability reflects the mistakes learners make when they try

to use their knowledge to communicate. These researchers see variability as an

aspect of performance rather than competence. The premise that interlanguage

systems are themselves variable is, therefore, a disputed one.

(5)Learners employ various learning strategies to develop their interlanguages. The

different kinds of errors learners produce reflect different learning strategies. For

example, omission errors suggest that learners are in some way simplifying the

learning task by ignoring grammatical features that they are not yet ready to

process. Overgeneralization and transfer errors can also be seen as evidence of

learning strategies.

(6)The learner‘s grammar is likely to fossilize. Selinker suggested that only about five

per cent of learners go on to develop the same mental grammar as native speakers.

The majority stop some way short. The prevalence of backsliding ( i.e. the

production of errors representing an early stage of development) is typical of

fossilized learners. Fossilization does not occur in L1 acquisition and thus is

unique to L2 grammars.

This concept of interlanguage offers a general account of how L2 acquisition takes place. It incorporates elements from mentalist theories of linguistics ( for example, the notion of a ?language acquisition device‘ ) and elements from cognitive psychology ( for example, ?learning strategies‘).

The notion of L1 interference discussed in the previous section is reconstituted as one factor among many of the cognitive processes responsible for L2 acquisition. Selinker ( 1972) suggested five central cognitive processes operated in interlanguage: (1) language transfer ( Some items, rules and subsystems of the interlanguage may result from the transfer from the L1.), ( 2) overgeneralization of the target language rules ( Some elements of the interlanguage may be the product of overgeneralization of the rules and semantic features of the target language.), (3) transfer if training ( A rule enters the learner‘s system as a result of instruction.), (4) strategies of L2 learning ( Some elements of the interlanguage may result from the specific approach to the material to be learned.) and ( 5) strategies of L2 communication ( Some elements of the interlanguage may result from the specific ways people learn to communicate with native speakers of the target language.) Interference was seen as one of several processes responsible for interlanguage. The five processes together constitute the ways in which the learner tries to reduce the learning burden to manageable proportions. Learners have limited processing space and , therefore, cannot cope with the total complexity of a language system, so they limit the number of hypotheses they test at any one point of time.

The essential characteristics of interlanguage

An interlanguage, a version of a foreign language spoken by its learners, is a natural human language in its own right and should be described independently. As a natural language system, interlanguage is intermediate between the learner‘s mother tongue and the target language. The

subsequent discussion of interlanguage focuses on its three principal features, all of which were raised by Selinker in one way or another.

The L2 learner‘s interlanguage system is permeable, in the sense that rules that constitute the learner‘s knowledge at any one stage are not fixed, but are open to amendment. That is, interlanguages are constantly subject to a number of impinging forces. Target rules or forms may be partially acquired or improperly generalized, while the native language may permeate the emerging interlangauge grammar at various levels. In many respects, permeability is a general feature of natural languages, which evolve over time in ways not dissimilar to the developments that take place in language-learner language. All language systems are permeable. Interlanguage differs from other language systems only in he degree of permeability.

The L2 learner‘s interlanguage is dynamic or constantly changing . However, the learner does not jump from one stage to the next, but rather slowly revises the interim systems to accommodate new hypotheses about the target language system. This takes place by the introduction of a new rule, first in one context and then in another, and so on. A new rule spreads in the sense that its coverage gradually extends over a range of linguistic contexts. The process of constant revision and extension of rules is a feature of the inherent instability of interlanguage and its built-in propensity for change ( Ellis 1985). Corder (1981) claims that the interlangauge system is a dynamic one . The learning of the system is a process of interaction between the learner and the data provided. Since it is impossible to feed all the data into the learner at one time and leave him to process it, like digesting a heavy meal, there will be a theoretically infinite number of states of his grammar.

The L2 learner‘s interlanguage is systematic, which comes from the hypothesis that interlanguage is natural language. This signifies that an interlanguage cannot be a random collection of entities. Like any other natural language, an interlanguage is systematic from the start and it has an internally coherent structure. The L2 learner does not select haphazardly from his store of interlanguage rules, but in predictable ways. He bases his performance plans on his existing rule system in much the same way as the native speaker bases his plans on his internalized knowledge of the L1 system. It has often been pointed out that evaluating L2 performance in terms of the target language grammar is unsatisfactory, because the learner behaves ?grammatically‘ in the sense that he draws systematically on his interlanguage rules. The term ?error‘ itself is, therefore, doubtful. A learner‘s utterance can be classified as erroneous only with reference to the norms of the target language. For the L2 learner, however, the true norms are contained in the interlanguage system he has constructed. The study of interlanguage is, then, the study of language systems of language learners.

Stages of interlanguage development

L2 learners experience similar stages on the way to the target language, though the spans of time spent by the learners are different. Ellis (1994) points out that L2 learners with different language backgrounds follow a remarkably similar path of interlanguage development. According to the systematicity, Corder ( 1973) divides the development of interlanguage into presystematic error stage, systematic error stage and postsystematic error stage. Borrowing some insights from Corder‘s model, Brown (1994), based on observations of what the learner does, proposed a four-stage model:

The Random Error Stage is the first stage in the development of interlanguage. At this stage,

the learner has some forms of the target language, but he is only vaguely aware that there is some systematic order to a particular class of items. The learner does know there is something for him to learn, but he always gets things wrong. He has a particular intention to communicate with others, but he does not know the way to realize his intention. At this stage, the errors made by the learner, to a great extent, are random or arbitrary and they indicate a stage of experimentation and inaccurate guessing.

The Emergent Stage of interlanguage finds the learner growing in consistency in linguistic production. The learner has begun to notice a system and to internalize certain rules. These rules may not be correct according to target language standards, but they are legitimate in the learner‘s mind. At this stage, the learner seems to have grasped a rule or principle and the n regresses to some previous stage. Generally, the learner, at this stage, cannot give the reasons why he has made errors and he is unable to correct the errors when they are pointed out by someone else.

The Systematic Stage is the stage in which the learner is able to show more consistency in producing the second language, but his understanding is incomplete. For example, the learner knows that the general rule of forming the past tense is to add ― ed‖ to the rule cannot be applied. Therefore, the learner will produce such forms as ― comed‖ and ― goed‖. Such kinds of errors are not arbitrary and they reflect the system of interlanguage. At this stage, the learner himself cannot correct the errors until they pointed out, because he does not know some special verbs exist. However, the learner can explain why he has produced ― goed‖ and ― comed‖ etc.

The final stage in the development of the interlanguage system is the Stabilization Stage, which Corder (1973) calls the postsystematic stage. The learner, at this stage, knows the correct target rule and has relatively few errors. He has mastered the system to the point that fluency and intended meanings are not problematic. The system is complete enough and attention can be paid to those few errors that can be corrected by the learner without being pointed out. At this stage, the interlanguage system becomes stable.

The most salient difference between the second and the third stage is the ability of learners to correct their errors when they are pointed out. The fourth stage is characterized by the learner‘s ability to self-correct.

Some issues on interlanguage

1)Fossilization

The process of interlanguage development is different from that of L1 development because of the possibility of fossilization in L2 acquisition. Selinker noted that many L2 learners(perhaps as many 95℅) fail to reach target language competence. That is, they don not reach the end of the interlanguage continuum. They stop learning when their interlanguage contains at least some rules different from those of the target language system. He referred to this as fossilization. Fossilization is the state of affairs that exists when the learner ceases to elaborate the interlanguage in some respect, no matter how long there is exposure to new data, or new teaching.

Fossilization occurs in most language learners and cannot be remedied by further instruction. Fossilized structures can be realize as errors or as correct target language forms. If, when fossilization occurs, the learner has reached a stage of development in which feature x in his interlanguage has assumed the same form as in the target language, then fossilization of the correct form will occur. If, however, the learner has reached a stage in which feature y still does

not have the same form as the target language, the fossilization will manifest itself as an error. Selinker maintained that such fossilization results especially from language transfer, but fossilization may also be the result of other processes. For example, strategies of communication may indicate to some individuals that they stop learning the language once they have learned enough to communicate. Because fossilization does not occur in L1 development, the acquisition of the interlanguage is thought to be different from L1 acquisition.

Fossilized structures may not be persistent, however. On occasions the learner may succeed in producing the correct target language form, but when the learner is focused on meaning—especially if the subject matter is difficult—he will ?backslide‘towards his true interlanguage form. Selinker and Lamendella (1978) argue that the causes of fossilization are both internal and external. It can occur both because the learner believes that he does not need to develop his interlanguage any further in order to communicate effectively whatever he wants to, and because changes in the neural structure of his brain as a result of age restrict the operation of the hypothesis-testing mechanism.

2)Hypothesis-testing

In the research on L2 acquisition, many errors made by L2 learners were found to be similar to those made by L1 acquirers, which led the researchers to study how closely the processes of developing L2 acquisition have proposed that L2 learners may develop their understanding of the L2 structure rules in the same way as L1 learners, by testing out hypotheses about how the language works. The research shows that L2 learners just like L1 children, can build up their knowledge of the language structure from scratch and eventually reach native speaker proficiency. In the process of L2 learning, learners have a partially correct mental picture of the L2 structure, which is made up of a collection of more or less correct hypotheses about the L2 grammatical system, named ?interlanguage‘ by Selinker (1972).

As we all know, different learners may have different mental pictures of the L2 structure, i.e. their own set of varyingly accurate hypotheses or interlanguage. L2 learners vary not forms but also in the way they use them in real –time communication. Tarone (1983) introduced the idea of the Continuum of Competence, which indicates that each correct hypothesis about the grammatical system, may be used in real—time production correctly or may help comprehension. He labels the two ends of the continuum Careful Style and V ernacular Style.

The competence continuum ( Tarone 1983)

V ernacular Style Careful Style

Using language naturally in Paying close attention to Interactions without worrying about correctness and appropriateness Accuracy or correctness too much. of the language being used.

Any grammatical form in an interlanguage can potentially be used at any point along the continuum. Most L2 learners are likely to be able to use the rules they know with varying accuracy at different points along the continuum. Tarone suggests that for any given learner it is easier consistently to apply a known rule correctly ate the careful end of the continuum, in, for example, a piece of formal writing. At the vernacular end, on the other hand, the same learner may often incorrectly apply the same rule. This is not because he /she does not know the rule. It is