Error Analysis

Perspectives on Second Language Acquisition

Jack C. Richards

• Publisher: Routledge
• Year: 2015
• eBook ISBN: 9781317869566
• Topic: Languages & Linguistics
• Subtopic: Linguistics
• Index: Languages & Linguistics

The eleven essays in this book cover a wide range of topics from the role of 'interlanguage' and the influence of external factors on the process of language learning, to the development of syntax and the methodology of error analysis. Collectively they provide a valuable perspective on the learning process, which both enriches our theoretical understanding of the processes underlying second language acquisition and suggests ways in which teaching practice may best exploit a learner's skills.

PART ONE

Introductory

These two papers serve as an introduction to this book. The first paper offers an overview of the field of error analysis and is intended to help the reader integrate the other papers in the collection into an overall perspective; it also links current interest in learner English to earlier research on second language learning. Why study the learner’s language? Interest in the language learning process, whether it be for first language acquisition or for second language learning, is comparatively recent, and has interest both for general linguistic theory and for language teaching practice. In the field of general linguistic theory, language learning is surely a testing ground for theories of language. If language is systematically structured in particular ways, we should expect to see evidence of this in observing how the child or adult copes with the cognitive and rule governed aspects of language as he becomes a member of a linguistic community. Study of the child-learner’s errors does indeed throw light on the types of cognitive and linguistic processes that appear to be part of the language learning process (Menyuk, 1971). In second language learning, as Corder observes, the learner’s errors are indicative both of the state of the learner’s knowledge, and of the ways in which a second language is learned. Corder makes an important distinction between mistakes or performance errors, and true errors, or markers of the learner’s transitional competence. Sentences containing errors would be characterized by systematic deviancy. While the learner’s correct sentences do not necessarily give evidence of the rules the learner is using or of the hypotheses he is testing, his errors suggest the strategies he employs to work out the rules of the new language and the rules he has developed at given stages of his language development. Corder suggests that the hypotheses the learner tests will be ‘Are the systems of the new language the same or different from those of the language I know? And if different, what is their nature?’ An alternative view of the role of interference is suggested elsewhere in this volume by Dulay and Burt.

Corder’s paper raises a question crucial to an understanding of language teaching. What is the relationship between what we teach and what is learned? Such a question can only be answered by studying what happens after we have prepared and taught our instructional syllabus. Corder suggests that the learner himself employs or establishes his own syllabus for learning and this may be a more effective syllabus than one prepared from ad hoc speculation. The nature of this syllabus and some of the ways in which it may be characterized are described in the paper by Richards and Sampson, which isolates seven factors as characterizing and influencing the learner’s language use: language transfer, intralingual interference, the effects of the sociolinguistic situation, the modality of exposure to the target language and the modality of production, the age of the learner, the instability of the learner’s linguistic system, and the effects of the inherent difficulty of the particular item being learned.

1

The Study of Learner English

JACK C RICHARDS AND GLORIA P SAMPSON

This paper was written for this volume.

1 Introduction

Theories of second language acquisition traditionally have been the offspring of general linguistic theory, sometimes supplemented by insights from psychology. While current linguistic theories are more insightful than previous ones, there has not been a corresponding increase in the descriptive or explanatory powers of theories of second language acquisition. The utility of generative linguistic theories in second language research is not especially obvious at the present time. This apparent lack of a linguistic paradigm for second language research, however, may be a propitious occurrence. Lacking an obviously relevant linguistic theory, researchers in second language learning may be compelled to develop new theories relevant to the particular domain under investigation. Such theories might be specific to second language acquisition. At the same time the data gathered could perhaps provide corrective feedback to general linguistic theory and to language teaching practice. While lack of a linguistic research paradigm is frustrating for the researcher, it may yet lead to interesting and original findings.

2 The Study of Learners’ Approximative Systems

New directions of research in second language acquisition are a product of the history of research to date as a brief review reveals. Because synchronic linguistic theory was preparadigmatic in the late nineteenth century, so too were early observations about second language learning. Boaz pointed out the apparent fluctuations in learners’ (linguists’) perceptions of sounds in new languages (Boaz, 1889). He suggested that learners perceived sounds in new languages in terms of their native language or other languages to which they had earlier been exposed. With the emergence of the notion of language as a system however, the question of second language acquisition could be viewed as the juxtaposition of two systems.

Juxtaposition of language systems could lead to a new supersystem which combined features of both systems (Fries and Pike, 1949), or to intersystemic interference (Weinreich, 1953). The notion of interference between two systems struck linguists and teachers as especially interesting, since it appeared to account for the problems of second language learning, particularly of adults. The writings of Lado (1957), for example, tended to emphasize points of contrast between two language systems. Contrastive analysis subsequently arose as a field of research. To be sure, contrast between systems was understood not to be the only factor involved in second language learning.

Subsequent attempts to rectify what was seen as an overly theoretical approach to language learning evolved into what some linguists refer to as error analysis. The major defect of contrastive analysis was deemed to be the attention paid to the analysis of two grammars. Some linguists proposed closer study of the performance of actual learners. Corder ( 1967), for instance, suggested that linguists study the process of language acquisition and the various strategies learners may use. Strevens (1969) hypothesized that errors should not be viewed as problems to be overcome, but rather as normal and inevitable features indicating the strategies that learners use. He conjectured that if a regular pattern of errors could be observed in the performance of all learners in a given situation, and if a learner were seen to progress through this pattern, his errors could be taken as evidence not of failure but of success and achievement in learning. But the pioneering studies of learner errors were still done within a framework which stressed interference between languages inasmuch as they stressed errors per se. For example, Nemser (1971) in his early work aimed at ‘the collection and evaluation of relevant interference data.’ Brière (1968) attempted to ‘test empirically the amount of interference that would ensue from competing phonological categories.’ Errors which did not fit systematically into the native language or target language system were, for the most part, ignored. Both Brière and Nemser, however, noted that some replacements occurred which were not in either the learner’s native language or the (artificial) target language.

Recently it has been suggested that errors alone are of little interest ; rather the entire linguistic system of the second language learner should be investigated. Hence current research tends to focus on the learner himself as generator of the grammar of his sentences in the new language. This emphasis is reflected in a growing terminology for a field of research which deals with the learner’s attempts to internalize the grammar of the language he is learning. This terminology includes error analysis, idiosyncratic dialects, inter language, approximative systems, transitional competence, l’état de dialecte. (Corder, 1967, 1971a, 1971b; Selinker, 1972; Nemser, this volume; Richards, 1971). The learner’s partial success, reflected in the construction of rules which do not necessarily reflect those of the mother tongue or target language, is seen as representing the construction of evolving systems of grammatical and phonological rules. The small amount of research and speculation about learners’ approximative systems (term borrowed from Nemser, this volume) suggests that seven factors may influence and characterize these second language learner systems. These factors are discussed below.

3 Language Transfer

The first factor is language transfer. Sentences in the target language may exhibit interference from the mother tongue. This of course was considered to be the major, but not the only, source of difficulty by linguists doing contrastive analysis. Interference analysis tends to be from the deviant sentence back to the mother tongue. Contrastive analysis works the other way, predicting errors by comparing the linguistic systems of the mother tongue and the target language. The history and theoretical origins of contrastive analysis in behaviourist psychology are reviewed by Dulay and Burt (1972). George (1971) found that one-third of the deviant sentences from second language learners could be attributed to language transfer, a figure similar to that given by Lance (1969) and Brudhiprabha (1972). It would however be almost impossible to assess the precise contribution of systemic language interference at this time. As subsequent discussion will indicate, a number of factors interact in determining the learner’s approximative language system. Until the role of some of these other factors is more clearly understood, it is not possible to evaluate the amount of systemic interference due to language transfer alone.

4 Intralingual Interference

The second factor, termed intralingual interference by Richards (1970), refers to items produced by the learner which reflect not the structure of the mother tongue, but generalizations based on partial exposure to the target language. In an analysis of English errors produced by speakers of a multitude of unrelated languages representing several language families, Richards noted subcategories of error types which seem to be common to speakers of diverse languages as they develop hypotheses about the structure of English. Like first language learners, the second language learner tries to derive the rules behind the data to which he has been exposed, and may develop hypotheses that correspond neither to the mother tongue nor target language. In an experiment on learning Russian word order, Torrey (1966) found that subjects sometimes adopted a consistent word order different from either Russian or English. Richards (1971a) found systematic intralingual errors to involve overgeneralization, ignorance of rule restrictions, incomplete application of rules, and semantic errors. These have been quantified for Thai learners of English by Brudhiprabha (1972). These studies suggest that many intralingual errors represent the learning difficulty of what are often low level rules in the target language, such as differences between the verb inflection in I walk, she walks. It may be inferred that once basic rules such as those concerning subject-object relationships, predication, negation, etc. are acquired, a considerable amount of difficulty in second language learning is related to selectional restrictions and to surface structure and contextual rules of the language. Both language transfer and intralingual errors confirm the traditional notion of transfer of training; that is, previous learning may influence later learning.

5 Sociolinguistic Situation

A third factor is the sociolinguistic situation. Different settings for language use result in different degrees and types of language learning. These may be distinguished in terms of the effects of the socio-cultural setting on the learner’s language and in terms of the relationship holding between the learner and the target language community and the respective linguistic markers of these relations and identities. Included here are thus the effects of the learner’s particular motivations for learning the second language as well as the effects of the socio-cultural setting.

The distinction of compound/co-ordinate bilingualism (Weinreich, 1953; Ervin and Osgood, 1954; Lambert, 1961) rests upon an assumption that different settings for language learning may motivate different processes of language learning. For example, two languages may be learned in the same socio-cultural setting or in two different settings. If the languages are learned in the same setting, the learner may develop a given type of semantic structure. Imagine the case of a child raised bilingually in the home. English bread and French pain might be identified with a single concept (compound bilingualism). On the other hand, if bread were learned in the home setting and pain later in another setting, the two lexemes might be stored separately (co-ordinate bilingualism). Although it has been criticized as too simple a model to explain real linguistic differences (Macnamara, 1971), it is still found useful for sociolinguistic analysis (Pride, 1971). More generally however the focus on the relationship between the opportunities for learning and the learner’s developing system is a useful one, since it leads to such distinctions as to whether the learning opportunities are limited to those provided by the school course (English as a foreign language) or are mainly outside of the school program (English as a second language) and to a consideration of the effects of these differences on the learner’s language (Richards, this volume).

Consideration of the sociolinguistic situation also leads to inclusion of the general motivational variables which influence language learning. Psychologists have related the types of language learning achieved to the role of the language in relation to the learner’s needs and perceptions. The ‘instrumental’ type of motivation is described as that motivating a learner to study a language for largely utilitarian purposes, and not as a means for integration with members of another cultural linguistic group. It is said to be appropriate for short term goals but inappropriate for the laborious task of acquiring a language for which an ‘integrative’ motivation is necessary. In focusing on the type of relationship holding between the learner and the target language community it would be appropriate to consider non-standard dialects (Labov, 1971), pidgins and creoles (Hymes, 1971) and immigrant language learning (Richards, this volume) as illustration of the influence of social processes on the transmission and use of language.

In both first and second language acquisition particular forms and patterns of language learning may be attributable to social variables. Whether the learner produces When are you coming? When you come?, When yous coming?, may depend on the social situation, the learner’s values and attitudes, or some other social factor in the learning context. The phenomenon of simplification in some language contact situations, represented by the absence of the copula, reduction of morphological and inflectional systems, and grammatical simplification, may likewise be socially motivated (Ferguson, 1971). When the need is for communication of simple information with the help of non-linguistic clues, vocabulary items and word order may be the most crucial elements to be acquired, as the experience of tourists in foreign countries and the linguistic folklore exemplified in comic-book caricatures (Me Tarzan, you Jane) testify.

The influence of the mother tongue on the learner’s language may also vary according to the sociolinguistic situation. Mackey (1962) notes that in describing interference one must account for variation according to the medium, style, or register in which the speaker is operating. For example, the medium may be spoken or written, the register may be formal or informal, and the speaker may play any one of a number of varied social roles. Sampson (1971) suggests that varying situations evoke different kinds of errors in varying quantities when children are trying to use the target language.

6 Modality

The learner’s language may vary according to a fourth factor, the modality of exposure to the target language and the modality of production. Production and perception may involve the acquisition of two partially overlapping systems. Vildomec observes that interference between the bilingual’s languages is generally on the productive rather than receptive side. People often report instances of intrusion of elements of their mother tongue in speech production, but rarely in their understanding of another language (Vildomec, 1963). Nemser’s research suggests that two different systems may be internalized in the target language depending on the modality. He found that in the productive modality, phonological replacements differed depending on whether the learner was imitating utterances he heard or producing speech spontaneously (Nemser, 1971a). In fact in first language acquisition Lieberman (1970) has proposed that some phonological features exist because their ‘acoustic correlates “match” a particular neural acoustic detector. Other features exist because it is easy to produce a particular articulatory manoeuvre with the human vocal apparatus ; the features “match” an articulatory constraint. Still other features may represent an optimization along both these dimensions; the features may have articulatory correlates that are “easy” to produce and result in acoustic correlates that are readily perceptible.’ It is not therefore unrealistic to assume that second language learners acquire some distinctions on the basis of auditory cues, some on the basis of articulatory cues, and others on the basis of a combination of these cues. An example of learning on the basis of auditory cues would be the discrimination between initial /f/ and /θ/ on the one hand and /v/ and /6/ in English on the basis of formant transitions. A case where visible articulatory movements would be the basis for discrimination would be the contrast between bilabial and velar stops in English. The standard second-language prescription to first aurally discriminate and only secondly to produce a given contrast in a target language may not be an optimal procedure for all classes of sounds. George describes learning difficulties derived from audiolinguai introduction to is, has, in unstressed positions, which may be realized as /z/, leading to identification as a single lexical item and to such sentences as She is a book, Her name has Sita. Written presentation of the forms may avoid this confusion (George, 1971). Spelling pronunciations, and confusions of written and spoken styles, would be other examples of modality affecting the learner’s approximative system.

7 Age

The fifth factor which may affect the approximative system of the second language learner is his age. Some aspects of the child’s learning capacities change as he grows older and these may affect language learning. The child’s memory span increases with age. He acquires a greater number of abstract concepts, and...

Table of contents

• Cover Page
• Half Title page
• Series Page
• Title Page
• Copyright Page
• Contents
• The Contributors
• Acknowledgements
• Preface
• Part One Introductory
• Part Two Towards Theories and Models
• Part Three Developmental Studies of Second Language Acquisition in Children
• Part Four Error Analyses of Adult Language Learning
• Bibliography