Early approaches to SLA
We begin our survey of early approaches with Contrastive Analysis (CA), which predates the establishment in the 1960s of SLA as a field of systematic study. This is an important starting point because aspects of CA procedures are still incorporated in more recent approaches, and because CA introduced a continuing major theme of SLA research: the influence of L1 on L2. The revolution in linguistic theory introduced by Noam Chomsky (1957) redirected much of SLA study to an internal focus, which is manifested in the other early (i.e. predating 1980) approaches included in this section.
Contrastive Analysis
Contrastive Analysis (CA) is an approach to the study of SLA which involves predicting and explaining learner problems based on a comparison of L1 and L2 to determine similarities and differences. It was heavily influenced by theories which were dominant in linguistics and psychology within the USA through the 1940s and 1950s, Structuralism and Behaviorism. The goal of CA (as that of still earlier theories of L2 learning) was primarily pedagogical in nature: to increase efficiency in L2 teaching and testing. Robert Lado states this clearly in his introduction to Linguistics Across Cultures ( 1957 ), a book which became a classic guide to this approach:
The plan of the book rests on the assumption that we can predict and describe the patterns that will cause difficulty in learning, and those that will not cause difficulty, by comparing systematically the language and culture to be learned with the native language and culture of the student. In our view, the preparation of up-to-date pedagogical and experimental materials must be based on this kind of comparison. (p. vii)
Following notions in structuralist linguistics, the focus of CA is on the surface forms of both L1 and L2 systems, and on describing and comparing the languages one level at a time – generally contrasting the phonology of L1 and L2 first, then morphology, then syntax, with the lexicon receiving relatively little attention, and nonverbal structures and dis course still less. A “bottom-up” priority for analysis (generally from small er to larger units) is also expressed as a priority for language learning, of structures before meaning. Charles Fries, who was a leading figure in applying structural linguistics to L2 teaching, makes this priority very clear: “In learning a new language, . . . the chief problem is not at first that of learning vocabulary items. It is, first, the mastery of the sound sys them . . . It is, second, the mastery of the features of arrangement that constitute the structure of the language” (Fries 1945 :3).
Following notions in behaviorist psychology, early proponents of CA assumed that language acquisition essentially involves habit formation in a process of Stimulus–Response–Reinforcement (S-R-R). Learners respond to the stimulus (linguistic input), and reinforcement strengthens (i.e. habituates) the response; they imitate and repeat the language that they hear, and when they are reinforced for that response, learning occurs. The implication is that “practice makes perfect.”
Another assumption of this theory is that there will be transfer in learning: in the case of SLA, this means the transfer of elements acquired (or habituated) in L1 to the target L2. The transfer is called positive (or facilitating) when the same structure is appropriate in both languages, as in the transfer of a Spanish plural morpheme -s on nouns to English (e.g. lenguajes to languages). The transfer is called negative (or interference ) when the L1 structure is used inappropriately in the L2, as in the additional transfer of Spanish plural -s to a modifier in number agreement with the noun: e.g. lenguajes moderna s to Modern s Language s (a translation which was printed at the top of a letter that I received from South America), or green s bean s (for ‘green beans,’ which I saw posted as a vegetable option in a US cafeteria near the Mexican border).
The process of CA involves describing L1 and L2 at each level, analyzing roughly comparable segments of the languages for elements which are likely to cause problems for learners. This information provides a rationale for constructing language lessons that focus on structures which are predicted to most need attention and practice, and for sequencing the L2 structures in order of difficulty. To summarize Lado’s ( 1957 ) position: the easiest L2 structures (and presumably first acquired) are those which exist in L1 with the same form, meaning, and distribution and are thus available for positive transfer ; any structure in L2 which has a form not occurring in L1 needs to be learned, but this is not likely to be very difficult if it has the same meaning and distribution as an “equivalent” in L1; among the most difficult are structures where there is partial overlap but not equivalence in form, meaning, and/or distribution, and these are most likely to cause interference. Lado gives examples in Spanish and English for some of the types of contrasts he describes, which I include in the accompanying box. I have ordered them from least to most probable difficulty for speakers of one of these languages learning the other.
The basic process of CA can be applied to any L1 and L2, but analysts cannot assume that such specifics as definitions of types of interference in Spanish and English will be relevant for languages with different typo logical features. The “same form and meaning” of -s “plural,” for example, is likely possible only in such closely related languages. A more flexible but similar category might include the use of any noun suffix for “plural,” such as Spanish -s and Hebrew -im. When an even greater level of generalization is appropriate, all languages that signal “plural” with an inflection on nouns might be contrasted with languages that inflect verbs for number, or with those that use no grammatical marker for number at all.
While CA highlighted potential learning problems, behaviorist learning theory attributed variable success by L2 learners in part to the nature of the relationship between L1 and L2 (and thus to the potential for negative versus positive transfer), but most importantly to circumstances of learning which promote poor versus good habit formation. Fries related L2 accuracy in English to the priorities he set for learning: “one can achieve mere fluency in a foreign language too soon . . . Such students, with fluency in vocabulary but with no basic control of either the sound system or the structure, are almost without exception hopeless so far as ever achieving a satisfactory control of English is concerned” (1945:3).
The CA approach of the 1940s to 1960s was not adequate for the study of SLA in part because the behaviorist learning theory to which it is tied cannot explain the logical problem of language learning that was addressed in Chapter 2 (how learners know more than they have heard or have been taught). Another problem was that CA analyses were not always validated by evidence from actual learner errors. Many of the L2 problems which CA predicts do not emerge; CA does not account for many learner errors; and much predicted positive transfer does not materialize. A major limitation in application to teaching has been that instructional materials produced according to this approach are language-specific and unsuitable for use with speakers of different native languages. Still, CA stimulated the preparation of hundreds of comparative grammars (including many unpublished master’s theses and doctoral dissertations at universi ties around the world), and its analytic procedures have been usefully applied to descriptive studies and to translation, including computer translation. Further, there has been a more recent revival and revision of CA procedures, including contrasts of languages at more abstract levels, and extension of the scope of analysis to domains of cross-cultural communication and rhetoric. There is also renewed interest in the contributions of positive transfer to SLA (see e.g. Ringbom 2007).
Error Analysis
Error Analysis (EA) is the first approach to the study of SLA which includes an internal focus on learners’ creative ability to construct language. It is based on the description and analysis of actual learner errors in L2, rather than on idealized linguistic structures attributed to native speakers of L1 and L2 (as in CA). EA largely augmented or replaced CA by the early 1970s because of the following developments:
• Predictions made by CA did not always materialize in actual learner errors, as noted above. More importantly, perhaps, many real learner errors could not be attributed to transfer from L1 to L2.
• As linguistic theory changed, the exclusive focus on surface-level forms and patterns by structural linguists shifted to concern for underlying rules.
• The behaviorist assumption that habit formation accounts for language acquisition was seriously questioned by many linguists and psychologists. There was a shift to Mentalism in explanations of language acquisition, with emphasis on the innate capacity of the language learner rather than on external influences.
• The study of SLA was no longer motivated as strongly by teaching concerns as it had been for CA. L2 learning came to be thought of as independent of L2 teaching to some extent, and researchers began to separate issues in SLA from pedagogical concerns. Learning processes became an important focus for study in their own right.
The shift in primary focus from surface forms and patterns to underlying rules, and the parallel shift in efforts to explain acquisition from Behaviorism to Mentalism, are attributable in large part to the revolution in linguistics which resulted from Noam Chomsky’s introduction of Transformational-Generative (TG) Grammar (1957, 1965). Chomsky claimed that languages have only a relatively small number of essential rules which account for their basic sentence structures, plus a limited set of transformational rules which allow these basic sentences to be modified (by deletions, additions, substitutions, and changes in word order). The finite number of basic rules and transformations in any language accounts for an infinite number of possible grammatical utterances. (Note that these “rules” merely describe what native speakers say, not what someone thinks they should say.) “Knowing” a language was seen as a matter of knowing these rules rather than memorizing surface structures. Since speakers of a language can understand and produce millions of sentences they have never heard before, they cannot merely be imitating what they have heard others say, but must be applying these underlying rules to create novel constructions. Language thus came to be understood as rule-governed behavior.
Under this influence from linguistics and related developments in psychology, the study of first language acquisition adopted notions that inner forces (interacting with the environment) drive learning, and that the child is an active and creative participant in the process rather than a passive recipient of language “stimuli.” Structures of child language pro duction began to be described and analyzed as grammatical systems in their own right rather than in terms of how they are “deficient” in com parison to adult norms (Miller 1964; McNeil 1966). Similar notions began to be applied to the study of second language learning at about the same time, in part to address the issue of how L1 and L2 acquisition processes might be the same or different.
The most influential publication launching Error Analysis as an approach in SLA was S. Pit Corder’s (1967) article on “The significance of learners’ errors,” which calls on applied linguists to focus on L2 learners’ errors not as “bad habits” to be eradicated, but as sources of insight into the learning processes. Corder claimed that errors provide evidence of the system of language which a learner is using at any particular point in the course of L2 development, and of the strategies or procedures the learner is using in his “discovery of the language.” In a sense, errors are windows into the language learner’s mind. In this approach, learner language is viewed as a target of analysis which is potentially independent of L1 or L2, and the state of learner knowledge is seen as transitional competence on the path of SLA. Further, Corder claimed that the making of errors is significant because it is part of the learning process itself: “a way the learner has of testing his hypothesis about the nature of the language he is learning.” This includes testing whether aspects of existing L1 knowledge can be used in the L2. Errors are thus a sign that the learner is (perhaps unconsciously) exploring the new system rather than just experiencing “interference” from old habits.
The procedure for analyzing learner errors includes the following steps (Ellis 2008):
• Collection of a sample of learner language. Most samples of learner language which have been used in EA include data collected from many speakers who are responding to the same kind of task or test (as in Morpheme Order Studies, which are discussed below). Some studies use samples from a few learners that are collected over a period of weeks, months, or even years in order to determine patterns of change in error occurrence with increasing L2 exposure and proficiency.
• Identification of errors. This first step in the analysis requires determination of elements in the sample of learner language which deviate from the target L2 in some way. Corder (1967) distinguishes between systematic errors (which result from learners’ lack of L2 knowledge) and mistakes (the results from some kind of processing failure such as a lapse in memory), which he excludes from the analysis.
• Description of errors. For purposes of analysis, errors are usually classified according to language level (whether an error is phonological, morphological, syntactic, etc.), general linguistic category (e.g. auxiliary system, passive sentences, negative constructions), or more specific linguistic elements (e.g. articles, prepositions, verb forms).
• Explanation of errors. Accounting for why an error was made is the most important step in trying to understand the processes of SLA. Two of the most likely causes of L2 errors are interlingual (“between languages”) factors, resulting from negative transfer or interference from L1 and intralingual (“within language”) factors, not attributable to cross-linguistic influence. Intralingual errors are also considered developmental errors and often represent incomplete learning of L2 rules or overgeneralization of them. Distinguishing between interlingual and intralingual errors implicitly builds upon CA procedures, since the distinction requires comparative knowledge of L1 and L2. For example, the following passage was in a letter written to me by a native Korean speaker. I have underlined and numbered the errors.
The weather is been 1 very hot in the 2 Washington D.C. There climate 3 last week warm.
(1) Use of is instead of has with bean (intralingual/developmental error). This is evidence that the speaker/writer is learning the English auxiliary verb system, but hasn’t yet mastered the distinction between forms of be and have, which doesn’t exist in Korean.
(2) Use of the with a place name (intralingual/developmental error). This is evidence that the speaker/writer is learning to use articles in front of nouns (no articles are used in Korean) but hasn’t yet learned that they don’t occur before most place names.
(3) There climate is a direct translation of the Korean phrase which would be used in this context (interlingual/interference error).
• Evaluation of errors. This step involves analysis of what effect the error has on whomever is being addressed: e.g. how “serious” it is, or to what extent it affects intelligibility, or social acceptability (such as qualifying for a job). In the example I gave of the Korean L1 speaker making errors in a letter to me, the errors are not serious at all. We are friends, and the ungrammaticality of many of her sentences has no bearing on the social relationship; furthermore, there is no resulting misinterpretation of meaning.
EA continues as a useful procedure for the study of SLA, but a number of shortcomings have been noted and should be kept in mind. These include:
• Ambiguity in classification. It is difficult to say, for instance, if a Chinese L1 speaker who omits number and tense inflections in English L2 is doing so because of L1 influence (Chinese is not an inflectional language) or because of a universal developmental process (also present in L1 acquisition) which results in simplified or “telegraphic” utterances.
• Lack of positive data. Focus on errors alone does not necessarily provide information on what the L2 learner has acquired (although I have inferred from the examples I gave above what the Korean L1 speaker/writer has learned about English auxiliary verbs and articles); further, correct uses may be overlooked.
• Potential for avoidance. Absence of errors may result from learners’ avoidance of difficult structures, and this will not be revealed by EA (e.g. Shachter [1974] makes the point that Chinese and Japanese L1 speakers make few errors in English L2 relative clauses because they avoid using them).
• Influence of L2 curricula. Some L2 teachers attribute the variation in student errors to the natures of students’ prior L2 learning experiences: e.g. whether it was informal or formal in nature, and if formal, whether grammar versus communicative activities or written versus oral skills dominated. Even when very similar approaches to teaching are represented in these experiences, teachers and textbooks may have included different content, different emphases, and different sequencing. These and other possible curricular variables may have a significant influence on subsequent student errors, although most have received little attention in research that tries to explain why some learners are more successful than others.
Interlanguage
Under the same influences from linguistics and psychology as Corder, and building on his concepts and procedures for EA, Larry Selinker (1972) introduced the term Interlanguage (IL) to refer to the intermediate states (or interim grammars) of a learner’s language as it moves toward the target L2. As in EA and first language studies of the 1960s and 1970s, Selinker and others taking this approach considered the development of the IL to be a creative process, driven by inner forces in interaction with environmental fac tors, and influenced both by L1 and by input from the target language. While influence from L1 and L2 language systems in a learner’s IL is clearly recognized, emphasis is on the IL itself as a third language system in its own right which differs from both L1 and L2 during the course of its development.
An interlanguage has the following characteristics:
• Systematic. At any particular point or stage of development, the IL is governed by rules which constitute the learner’s internal grammar. These rules are discoverable by analyzing the language that is used by the learner at that time – what he or she can produce and interpret correctly as well as errors that are made.
• Dynamic. The system of rules which learners have in their minds changes frequently, or is in a state of flux, resulting in a succession of interim grammars. Selinker views this change not as a steady progression along a continuum, but discontinuous progression “from stable plateau to stable plateau” (1992 :226).
• Variable. Although the IL is systematic, differences in context result in different patterns of language use (discussed in Chapter 5).
• Reduced system, both in form and function. The characteristic of reduced form refers to the less complex grammatical structures that typically occur in an IL compared to the target language (e.g. omission of inflections, such as the past tense suffix in English). The char acteristic of reduced function refers to the smaller range of communicative needs typically served by an IL (especially if the learner is still in contact with members of the L1 speech community).
Selinker ( 1972 ) stresses that there are differences between IL development in SLA and L1 acquisition by children, including different cognitive processes involved (from McLaughlin 1987 :61):
• Language transfer from L1 to L2
• Transfer of training, or how the L2 is taught
• Strategies of second language learning, or how learners approach the L2 materials and the task of L2 learning
• Strategies of second language communication, or ways that learners try to communicate with others in the L2
• Overgeneralization of the target language linguistic material, in which L2 rules that are learned are applied too broadly. (Overgeneralizations include some of the intralingual or develop mental errors which were illustrated in the previous section.)
Also unlike L1 acquisition is the strong likelihood of fossilization for L2 learners – the probability that they will cease their IL development in some respects before they reach target language norms, in spite of continuing L2 input and passage of time. This phenomenon relates to age of learning, with older L2 learners more likely to fossilize than younger ones, but also to factors of social identity and communicative need (see e.g. Selinker 1992). Such factors are at the core of discussions concerning the basic question of why some learners are more successful than others. “Relative success” can be defined in this approach as the level of IL development reached before learning stops. A schematization of the construct is presented in Figure 3.1.
The beginning and end of IL are defined respectively as whenever a learner first attempts to convey meaning in the L2 and whenever development “permanently” stops, but the boundaries are not entirely clear. A schematization of the construct is presented in 3.1. The initial state and very early stages of L2 development in naturalistic (i.e. unschooled or untutored) settings often involve only isolated L2 words or memorized routines inserted in an L1 structural frame for some period of time. For example, we recorded the following utterances from children who were just beginning to acquire English (Saville-Troike, Pan, and Dutkova 1995):
IL probably cannot properly be said to begin until there is some evidence of systematic change in grammar. The endpoint of IL is difficult to identify with complete certainty since additional time and different circumstances might always trigger some resumption in learning.
Identification of fossilization, or cessation of IL development before reaching target language norms, is even more controversial (though primarily for social and political rather than linguistic reasons). Should individuals be considered “fossilized” in L2 development because they retain a foreign accent, for instance, in spite of productive fluency in other aspects of the target language? (One thinks of Arnold Schwarzenegger, US motion picture actor and politician, who retains a strong Austrian-German accent, or of many faculty members and students who are identifiably nonnative speakers of English although they speak and write fluently in this language – often even more fluently than many native speakers. There may even be an advantage in retaining a nonnative accent, since “sounding native” may be misinterpreted by native speakers as implying corresponding native social and cultural knowledge.)
There is also the issue of what the concept of “target language” entails as the goal of SLA, especially as it applies to English usage in parts of the world where English has been adopted as an auxiliary or official language but differs from any native variety in Britain or the USA (see Kachru and Nelson 1996). “Native-like” production is neither intended nor desired by many speakers, and assuming that it is or should be the ultimate goal for all L2 learners may be considered somewhat imperialistic.
The concept of an IL as a system of learner language which is at least partially independent of L1 and L2 has been highly productive in the study of SLA. It is generally taken for granted now, although controversies remain concerning its specific nature and whether “progress” should be measured against native-speaker norms (e.g. Eubank, Selinker, and Sharwood Smith 1995; Johnson and Johnson 1998 :174–76).
Morpheme Order Studies
One important question in the study of SLA which the concept of IL highlighted during the 1970s is whether there is a natural order (or universal sequence) in the grammatical development of L2 learners. This is interesting because if we find that the same elements of an L2 are learned first no matter what the learner’s L1 is, we might assume that transfer from L1 is less important than if we were to find that the order of acquisition is different for speakers of different native languages. If the same order of acquisition is found in L2 as in children’s L1 learning, there is the additional implication that the acquisition processes may be very much the same for all of language development.
Roger Brown (1973) provided the first baseline information on an L1 acquisition sequence by tracking the order in which three children mastered the production of a set of grammatical morphemes in English, including inflections which mark tense on verbs and plural number on nouns. His work was soon validated by studies of larger numbers of English L1 children. The claim that this sequence constituted a natural order for English L2 as well as English L1 was first made by Heidi Dulay and Marina Burt, based on studies of children learning English who were native speakers of Spanish and Chinese. A list of morphemes that were included in the Brown (1973 ) and Dulay and Burt (1974) findings is given in Table 3.1. These results indicate, for example, that the progressive suffix -ing and plural -s are the first of this set of morphemes to be mastered by both L1 and L2 learners of English; the irregular past tense form of verbs and possessive -s are acquired next in sequence for L1, but relatively later for learners of L2 (after forms of be and a/the).
Although not identical, the order of morpheme acquisition reported was similar in L1 and L2. Further, the order was virtually the same in English L2 whether children were L1 speakers of Spanish or Chinese. The existence of such a “natural order” strengthened claims for internally driven acquisition processes, which Dulay and Burt (1973) labeled creative construction. They concluded that L2 learners are neither merely imitating what they hear nor necessarily transferring L1 structures to the new code, but (subconsciously) creating a mental grammar which allows them to interpret and produce utterances they have not heard before.
A claim was originally made that this evidence of similar morpheme order supports an Identity Hypothesis (or L1 = L2): that processes involved in L1 and L2 acquisition are the same. The strong form of this hypothesis was rejected largely because the basic question of what is being acquired in SLA was limited here to a list of isolated English morphemes, with no principled relation to other aspects of English or to other languages, and also because of weaknesses in the research methodology.
The concept of natural order remains very important for understanding SLA, however, both from linguistic and from cognitive approaches. The morpheme acquisition studies were followed by research which indicated that there are also regular sequences in acquisition of some syntactic constructions by both children and adults (e.g. negation, questions, and relative clauses). These findings form part of the basis for continuing speculation that innate mechanisms for language acquisition may not be limited to early childhood.
Monitor Model
One of the last of the early approaches to SLA which has an internal focus is the Monitor Model, proposed by Stephen Krashen (1978). It explicitly and essentially adopts the notion of a Language Acquisition Device (or LAD), which is a metaphor Chomsky used for children’s innate knowledge of language.
Krashen’s approach is a collection of five hypotheses which constitute major claims and assumptions about how the L2 code is acquired. Caution is required, however, that Krashen’s model has frequently been criticized by researchers because many of its constructs (e.g. what constitutes com prehensible input) and the claimed distinction between learning and acquisition are vague and imprecise, and because several of its claims are impossible to verify (see McLaughlin 1987). The hypotheses forming the model are the following:
• Acquisition-Learning Hypothesis. There is a distinction to be made between acquisition and learning. Acquisition is subconscious, and involves the innate Language Acquisition Device which accounts for children’s L1. Learning is conscious and is exemplified by the L2 learning which takes place in many classroom contexts.
• Monitor Hypothesis. What is “learned” is available only as a monitor, for purposes of editing or making changes in what has already been produced.
• Natural Order Hypothesis. We acquire the rules of language in a predict able order.
• Input Hypothesis. Language acquisition takes place because there is comprehensible input. If input is understood, and if there is enough of it, the necessary grammar is automatically provided.
• Affective Filter Hypothesis. Input may not be processed if the affective filter is “up” (e.g. if conscious learning is taking place and/or individuals are inhibited).
In spite of being severely criticized by researchers, Krashen’s model had a major influence on language teaching in the USA in the 1980s and 1990s, including avoidance of the explicit teaching of grammar in many hundreds of classrooms. The pendulum has since begun to swing back in the opposite direction, with formal grammar teaching increasingly being introduced, especially with adults, who are able to benefit from (and may even need) an explicit explanation of grammatical structure.
The early period for linguistic study of SLA which we have just reviewed ended with some issues in rather spirited debate among proponents of different approaches, but there was widespread consensus on some important points. These include:
• What is being acquired in SLA is a “rule-governed” language system. Development of L2 involves progression through a dynamic interlanguage system which differs from both L1 and L2 in significant respects. The final state of L2 typically differs (more or less) from the native speakers’ system.
• How SLA takes place involves creative mental processes. Development of both L1 and L2 follows generally predictable sequences, which suggests that L1 and L2 acquisition processes are similar in significant ways.
• Why some learners are more (or less) successful in SLA than others relate primarily to the age of the learner.
As we reach the 1980s in this survey, new proposals in Chomskyan theoretical linguistics were about to have a major impact on the study of SLA, and Universal Grammar was to become the dominant approach with an internal focus.
Early approaches to SLA
Book: Introducing Second Language Acquisition
Author: Muriel Saville-Troike
Page: c3, 36
We begin our survey of early approaches with Contrastive Analysis (CA), which predates the establishment in the 1960s of SLA as a field of systematic study. This is an important starting point because aspects of CA procedures are still incorporated in more recent approaches, and because CA introduced a continuing major theme of SLA research: the influence of L1 on L2. The revolution in linguistic theory introduced by Noam Chomsky (1957) redirected much of SLA study to an internal focus, which is manifested in the other early (i.e. predating 1980) approaches included in this section.
Contrastive Analysis
Contrastive Analysis (CA) is an approach to the study of SLA which involves predicting and explaining learner problems based on a comparison of L1 and L2 to determine similarities and differences. It was heavily influenced by theories which were dominant in linguistics and psychology within the USA through the 1940s and 1950s, Structuralism and Behaviorism. The goal of CA (as that of still earlier theories of L2 learning) was primarily pedagogical in nature: to increase efficiency in L2 teaching and testing. Robert Lado states this clearly in his introduction to Linguistics Across Cultures ( 1957 ), a book which became a classic guide to this approach:
The plan of the book rests on the assumption that we can predict and describe the patterns that will cause difficulty in learning, and those that will not cause difficulty, by comparing systematically the language and culture to be learned with the native language and culture of the student. In our view, the preparation of up-to-date pedagogical and experimental materials must be based on this kind of comparison. (p. vii)
Following notions in structuralist linguistics, the focus of CA is on the surface forms of both L1 and L2 systems, and on describing and comparing the languages one level at a time – generally contrasting the phonology of L1 and L2 first, then morphology, then syntax, with the lexicon receiving relatively little attention, and nonverbal structures and dis course still less. A “bottom-up” priority for analysis (generally from small er to larger units) is also expressed as a priority for language learning, of structures before meaning. Charles Fries, who was a leading figure in applying structural linguistics to L2 teaching, makes this priority very clear: “In learning a new language, . . . the chief problem is not at first that of learning vocabulary items. It is, first, the mastery of the sound sys them . . . It is, second, the mastery of the features of arrangement that constitute the structure of the language” (Fries 1945 :3).
Following notions in behaviorist psychology, early proponents of CA assumed that language acquisition essentially involves habit formation in a process of Stimulus–Response–Reinforcement (S-R-R). Learners respond to the stimulus (linguistic input), and reinforcement strengthens (i.e. habituates) the response; they imitate and repeat the language that they hear, and when they are reinforced for that response, learning occurs. The implication is that “practice makes perfect.”
Another assumption of this theory is that there will be transfer in learning: in the case of SLA, this means the transfer of elements acquired (or habituated) in L1 to the target L2. The transfer is called positive (or facilitating) when the same structure is appropriate in both languages, as in the transfer of a Spanish plural morpheme -s on nouns to English (e.g. lenguajes to languages). The transfer is called negative (or interference ) when the L1 structure is used inappropriately in the L2, as in the additional transfer of Spanish plural -s to a modifier in number agreement with the noun: e.g. lenguajes moderna s to Modern s Language s (a translation which was printed at the top of a letter that I received from South America), or green s bean s (for ‘green beans,’ which I saw posted as a vegetable option in a US cafeteria near the Mexican border).
The process of CA involves describing L1 and L2 at each level, analyzing roughly comparable segments of the languages for elements which are likely to cause problems for learners. This information provides a rationale for constructing language lessons that focus on structures which are predicted to most need attention and practice, and for sequencing the L2 structures in order of difficulty. To summarize Lado’s ( 1957 ) position: the easiest L2 structures (and presumably first acquired) are those which exist in L1 with the same form, meaning, and distribution and are thus available for positive transfer ; any structure in L2 which has a form not occurring in L1 needs to be learned, but this is not likely to be very difficult if it has the same meaning and distribution as an “equivalent” in L1; among the most difficult are structures where there is partial overlap but not equivalence in form, meaning, and/or distribution, and these are most likely to cause interference. Lado gives examples in Spanish and English for some of the types of contrasts he describes, which I include in the accompanying box. I have ordered them from least to most probable difficulty for speakers of one of these languages learning the other.
The basic process of CA can be applied to any L1 and L2, but analysts cannot assume that such specifics as definitions of types of interference in Spanish and English will be relevant for languages with different typo logical features. The “same form and meaning” of -s “plural,” for example, is likely possible only in such closely related languages. A more flexible but similar category might include the use of any noun suffix for “plural,” such as Spanish -s and Hebrew -im. When an even greater level of generalization is appropriate, all languages that signal “plural” with an inflection on nouns might be contrasted with languages that inflect verbs for number, or with those that use no grammatical marker for number at all.
While CA highlighted potential learning problems, behaviorist learning theory attributed variable success by L2 learners in part to the nature of the relationship between L1 and L2 (and thus to the potential for negative versus positive transfer), but most importantly to circumstances of learning which promote poor versus good habit formation. Fries related L2 accuracy in English to the priorities he set for learning: “one can achieve mere fluency in a foreign language too soon . . . Such students, with fluency in vocabulary but with no basic control of either the sound system or the structure, are almost without exception hopeless so far as ever achieving a satisfactory control of English is concerned” (1945:3).
The CA approach of the 1940s to 1960s was not adequate for the study of SLA in part because the behaviorist learning theory to which it is tied cannot explain the logical problem of language learning that was addressed in Chapter 2 (how learners know more than they have heard or have been taught). Another problem was that CA analyses were not always validated by evidence from actual learner errors. Many of the L2 problems which CA predicts do not emerge; CA does not account for many learner errors; and much predicted positive transfer does not materialize. A major limitation in application to teaching has been that instructional materials produced according to this approach are language-specific and unsuitable for use with speakers of different native languages. Still, CA stimulated the preparation of hundreds of comparative grammars (including many unpublished master’s theses and doctoral dissertations at universi ties around the world), and its analytic procedures have been usefully applied to descriptive studies and to translation, including computer translation. Further, there has been a more recent revival and revision of CA procedures, including contrasts of languages at more abstract levels, and extension of the scope of analysis to domains of cross-cultural communication and rhetoric. There is also renewed interest in the contributions of positive transfer to SLA (see e.g. Ringbom 2007).
Error Analysis
Error Analysis (EA) is the first approach to the study of SLA which includes an internal focus on learners’ creative ability to construct language. It is based on the description and analysis of actual learner errors in L2, rather than on idealized linguistic structures attributed to native speakers of L1 and L2 (as in CA). EA largely augmented or replaced CA by the early 1970s because of the following developments:
• Predictions made by CA did not always materialize in actual learner errors, as noted above. More importantly, perhaps, many real learner errors could not be attributed to transfer from L1 to L2.
• As linguistic theory changed, the exclusive focus on surface-level forms and patterns by structural linguists shifted to concern for underlying rules.
• The behaviorist assumption that habit formation accounts for language acquisition was seriously questioned by many linguists and psychologists. There was a shift to Mentalism in explanations of language acquisition, with emphasis on the innate capacity of the language learner rather than on external influences.
• The study of SLA was no longer motivated as strongly by teaching concerns as it had been for CA. L2 learning came to be thought of as independent of L2 teaching to some extent, and researchers began to separate issues in SLA from pedagogical concerns. Learning processes became an important focus for study in their own right.
The shift in primary focus from surface forms and patterns to underlying rules, and the parallel shift in efforts to explain acquisition from Behaviorism to Mentalism, are attributable in large part to the revolution in linguistics which resulted from Noam Chomsky’s introduction of Transformational-Generative (TG) Grammar (1957, 1965). Chomsky claimed that languages have only a relatively small number of essential rules which account for their basic sentence structures, plus a limited set of transformational rules which allow these basic sentences to be modified (by deletions, additions, substitutions, and changes in word order). The finite number of basic rules and transformations in any language accounts for an infinite number of possible grammatical utterances. (Note that these “rules” merely describe what native speakers say, not what someone thinks they should say.) “Knowing” a language was seen as a matter of knowing these rules rather than memorizing surface structures. Since speakers of a language can understand and produce millions of sentences they have never heard before, they cannot merely be imitating what they have heard others say, but must be applying these underlying rules to create novel constructions. Language thus came to be understood as rule-governed behavior.
Under this influence from linguistics and related developments in psychology, the study of first language acquisition adopted notions that inner forces (interacting with the environment) drive learning, and that the child is an active and creative participant in the process rather than a passive recipient of language “stimuli.” Structures of child language pro duction began to be described and analyzed as grammatical systems in their own right rather than in terms of how they are “deficient” in com parison to adult norms (Miller 1964; McNeil 1966). Similar notions began to be applied to the study of second language learning at about the same time, in part to address the issue of how L1 and L2 acquisition processes might be the same or different.
The most influential publication launching Error Analysis as an approach in SLA was S. Pit Corder’s (1967) article on “The significance of learners’ errors,” which calls on applied linguists to focus on L2 learners’ errors not as “bad habits” to be eradicated, but as sources of insight into the learning processes. Corder claimed that errors provide evidence of the system of language which a learner is using at any particular point in the course of L2 development, and of the strategies or procedures the learner is using in his “discovery of the language.” In a sense, errors are windows into the language learner’s mind. In this approach, learner language is viewed as a target of analysis which is potentially independent of L1 or L2, and the state of learner knowledge is seen as transitional competence on the path of SLA. Further, Corder claimed that the making of errors is significant because it is part of the learning process itself: “a way the learner has of testing his hypothesis about the nature of the language he is learning.” This includes testing whether aspects of existing L1 knowledge can be used in the L2. Errors are thus a sign that the learner is (perhaps unconsciously) exploring the new system rather than just experiencing “interference” from old habits.
The procedure for analyzing learner errors includes the following steps (Ellis 2008):
• Collection of a sample of learner language. Most samples of learner language which have been used in EA include data collected from many speakers who are responding to the same kind of task or test (as in Morpheme Order Studies, which are discussed below). Some studies use samples from a few learners that are collected over a period of weeks, months, or even years in order to determine patterns of change in error occurrence with increasing L2 exposure and proficiency.
• Identification of errors. This first step in the analysis requires determination of elements in the sample of learner language which deviate from the target L2 in some way. Corder (1967) distinguishes between systematic errors (which result from learners’ lack of L2 knowledge) and mistakes (the results from some kind of processing failure such as a lapse in memory), which he excludes from the analysis.
• Description of errors. For purposes of analysis, errors are usually classified according to language level (whether an error is phonological, morphological, syntactic, etc.), general linguistic category (e.g. auxiliary system, passive sentences, negative constructions), or more specific linguistic elements (e.g. articles, prepositions, verb forms).
• Explanation of errors. Accounting for why an error was made is the most important step in trying to understand the processes of SLA. Two of the most likely causes of L2 errors are interlingual (“between languages”) factors, resulting from negative transfer or interference from L1 and intralingual (“within language”) factors, not attributable to cross-linguistic influence. Intralingual errors are also considered developmental errors and often represent incomplete learning of L2 rules or overgeneralization of them. Distinguishing between interlingual and intralingual errors implicitly builds upon CA procedures, since the distinction requires comparative knowledge of L1 and L2. For example, the following passage was in a letter written to me by a native Korean speaker. I have underlined and numbered the errors.
The weather is been 1 very hot in the 2 Washington D.C. There climate 3 last week warm.
(1) Use of is instead of has with bean (intralingual/developmental error). This is evidence that the speaker/writer is learning the English auxiliary verb system, but hasn’t yet mastered the distinction between forms of be and have, which doesn’t exist in Korean.
(2) Use of the with a place name (intralingual/developmental error). This is evidence that the speaker/writer is learning to use articles in front of nouns (no articles are used in Korean) but hasn’t yet learned that they don’t occur before most place names.
(3) There climate is a direct translation of the Korean phrase which would be used in this context (interlingual/interference error).
• Evaluation of errors. This step involves analysis of what effect the error has on whomever is being addressed: e.g. how “serious” it is, or to what extent it affects intelligibility, or social acceptability (such as qualifying for a job). In the example I gave of the Korean L1 speaker making errors in a letter to me, the errors are not serious at all. We are friends, and the ungrammaticality of many of her sentences has no bearing on the social relationship; furthermore, there is no resulting misinterpretation of meaning.
EA continues as a useful procedure for the study of SLA, but a number of shortcomings have been noted and should be kept in mind. These include:
• Ambiguity in classification. It is difficult to say, for instance, if a Chinese L1 speaker who omits number and tense inflections in English L2 is doing so because of L1 influence (Chinese is not an inflectional language) or because of a universal developmental process (also present in L1 acquisition) which results in simplified or “telegraphic” utterances.
• Lack of positive data. Focus on errors alone does not necessarily provide information on what the L2 learner has acquired (although I have inferred from the examples I gave above what the Korean L1 speaker/writer has learned about English auxiliary verbs and articles); further, correct uses may be overlooked.
• Potential for avoidance. Absence of errors may result from learners’ avoidance of difficult structures, and this will not be revealed by EA (e.g. Shachter [1974] makes the point that Chinese and Japanese L1 speakers make few errors in English L2 relative clauses because they avoid using them).
• Influence of L2 curricula. Some L2 teachers attribute the variation in student errors to the natures of students’ prior L2 learning experiences: e.g. whether it was informal or formal in nature, and if formal, whether grammar versus communicative activities or written versus oral skills dominated. Even when very similar approaches to teaching are represented in these experiences, teachers and textbooks may have included different content, different emphases, and different sequencing. These and other possible curricular variables may have a significant influence on subsequent student errors, although most have received little attention in research that tries to explain why some learners are more successful than others.
Interlanguage
Under the same influences from linguistics and psychology as Corder, and building on his concepts and procedures for EA, Larry Selinker (1972) introduced the term Interlanguage (IL) to refer to the intermediate states (or interim grammars) of a learner’s language as it moves toward the target L2. As in EA and first language studies of the 1960s and 1970s, Selinker and others taking this approach considered the development of the IL to be a creative process, driven by inner forces in interaction with environmental fac tors, and influenced both by L1 and by input from the target language. While influence from L1 and L2 language systems in a learner’s IL is clearly recognized, emphasis is on the IL itself as a third language system in its own right which differs from both L1 and L2 during the course of its development.
An interlanguage has the following characteristics:
• Systematic. At any particular point or stage of development, the IL is governed by rules which constitute the learner’s internal grammar. These rules are discoverable by analyzing the language that is used by the learner at that time – what he or she can produce and interpret correctly as well as errors that are made.
• Dynamic. The system of rules which learners have in their minds changes frequently, or is in a state of flux, resulting in a succession of interim grammars. Selinker views this change not as a steady progression along a continuum, but discontinuous progression “from stable plateau to stable plateau” (1992 :226).
• Variable. Although the IL is systematic, differences in context result in different patterns of language use (discussed in Chapter 5).
• Reduced system, both in form and function. The characteristic of reduced form refers to the less complex grammatical structures that typically occur in an IL compared to the target language (e.g. omission of inflections, such as the past tense suffix in English). The char acteristic of reduced function refers to the smaller range of communicative needs typically served by an IL (especially if the learner is still in contact with members of the L1 speech community).
Selinker ( 1972 ) stresses that there are differences between IL development in SLA and L1 acquisition by children, including different cognitive processes involved (from McLaughlin 1987 :61):
• Language transfer from L1 to L2
• Transfer of training, or how the L2 is taught
• Strategies of second language learning, or how learners approach the L2 materials and the task of L2 learning
• Strategies of second language communication, or ways that learners try to communicate with others in the L2
• Overgeneralization of the target language linguistic material, in which L2 rules that are learned are applied too broadly. (Overgeneralizations include some of the intralingual or develop mental errors which were illustrated in the previous section.)
Also unlike L1 acquisition is the strong likelihood of fossilization for L2 learners – the probability that they will cease their IL development in some respects before they reach target language norms, in spite of continuing L2 input and passage of time. This phenomenon relates to age of learning, with older L2 learners more likely to fossilize than younger ones, but also to factors of social identity and communicative need (see e.g. Selinker 1992). Such factors are at the core of discussions concerning the basic question of why some learners are more successful than others. “Relative success” can be defined in this approach as the level of IL development reached before learning stops. A schematization of the construct is presented in Figure 3.1.
The beginning and end of IL are defined respectively as whenever a learner first attempts to convey meaning in the L2 and whenever development “permanently” stops, but the boundaries are not entirely clear. A schematization of the construct is presented in 3.1. The initial state and very early stages of L2 development in naturalistic (i.e. unschooled or untutored) settings often involve only isolated L2 words or memorized routines inserted in an L1 structural frame for some period of time. For example, we recorded the following utterances from children who were just beginning to acquire English (Saville-Troike, Pan, and Dutkova 1995):
IL probably cannot properly be said to begin until there is some evidence of systematic change in grammar. The endpoint of IL is difficult to identify with complete certainty since additional time and different circumstances might always trigger some resumption in learning.
Identification of fossilization, or cessation of IL development before reaching target language norms, is even more controversial (though primarily for social and political rather than linguistic reasons). Should individuals be considered “fossilized” in L2 development because they retain a foreign accent, for instance, in spite of productive fluency in other aspects of the target language? (One thinks of Arnold Schwarzenegger, US motion picture actor and politician, who retains a strong Austrian-German accent, or of many faculty members and students who are identifiably nonnative speakers of English although they speak and write fluently in this language – often even more fluently than many native speakers. There may even be an advantage in retaining a nonnative accent, since “sounding native” may be misinterpreted by native speakers as implying corresponding native social and cultural knowledge.)
There is also the issue of what the concept of “target language” entails as the goal of SLA, especially as it applies to English usage in parts of the world where English has been adopted as an auxiliary or official language but differs from any native variety in Britain or the USA (see Kachru and Nelson 1996). “Native-like” production is neither intended nor desired by many speakers, and assuming that it is or should be the ultimate goal for all L2 learners may be considered somewhat imperialistic.
The concept of an IL as a system of learner language which is at least partially independent of L1 and L2 has been highly productive in the study of SLA. It is generally taken for granted now, although controversies remain concerning its specific nature and whether “progress” should be measured against native-speaker norms (e.g. Eubank, Selinker, and Sharwood Smith 1995; Johnson and Johnson 1998 :174–76).
Morpheme Order Studies
One important question in the study of SLA which the concept of IL highlighted during the 1970s is whether there is a natural order (or universal sequence) in the grammatical development of L2 learners. This is interesting because if we find that the same elements of an L2 are learned first no matter what the learner’s L1 is, we might assume that transfer from L1 is less important than if we were to find that the order of acquisition is different for speakers of different native languages. If the same order of acquisition is found in L2 as in children’s L1 learning, there is the additional implication that the acquisition processes may be very much the same for all of language development.
Roger Brown (1973) provided the first baseline information on an L1 acquisition sequence by tracking the order in which three children mastered the production of a set of grammatical morphemes in English, including inflections which mark tense on verbs and plural number on nouns. His work was soon validated by studies of larger numbers of English L1 children. The claim that this sequence constituted a natural order for English L2 as well as English L1 was first made by Heidi Dulay and Marina Burt, based on studies of children learning English who were native speakers of Spanish and Chinese. A list of morphemes that were included in the Brown (1973 ) and Dulay and Burt (1974) findings is given in Table 3.1. These results indicate, for example, that the progressive suffix -ing and plural -s are the first of this set of morphemes to be mastered by both L1 and L2 learners of English; the irregular past tense form of verbs and possessive -s are acquired next in sequence for L1, but relatively later for learners of L2 (after forms of be and a/the).
Although not identical, the order of morpheme acquisition reported was similar in L1 and L2. Further, the order was virtually the same in English L2 whether children were L1 speakers of Spanish or Chinese. The existence of such a “natural order” strengthened claims for internally driven acquisition processes, which Dulay and Burt (1973) labeled creative construction. They concluded that L2 learners are neither merely imitating what they hear nor necessarily transferring L1 structures to the new code, but (subconsciously) creating a mental grammar which allows them to interpret and produce utterances they have not heard before.
A claim was originally made that this evidence of similar morpheme order supports an Identity Hypothesis (or L1 = L2): that processes involved in L1 and L2 acquisition are the same. The strong form of this hypothesis was rejected largely because the basic question of what is being acquired in SLA was limited here to a list of isolated English morphemes, with no principled relation to other aspects of English or to other languages, and also because of weaknesses in the research methodology.
The concept of natural order remains very important for understanding SLA, however, both from linguistic and from cognitive approaches. The morpheme acquisition studies were followed by research which indicated that there are also regular sequences in acquisition of some syntactic constructions by both children and adults (e.g. negation, questions, and relative clauses). These findings form part of the basis for continuing speculation that innate mechanisms for language acquisition may not be limited to early childhood.
Monitor Model
One of the last of the early approaches to SLA which has an internal focus is the Monitor Model, proposed by Stephen Krashen (1978). It explicitly and essentially adopts the notion of a Language Acquisition Device (or LAD), which is a metaphor Chomsky used for children’s innate knowledge of language.
Krashen’s approach is a collection of five hypotheses which constitute major claims and assumptions about how the L2 code is acquired. Caution is required, however, that Krashen’s model has frequently been criticized by researchers because many of its constructs (e.g. what constitutes com prehensible input) and the claimed distinction between learning and acquisition are vague and imprecise, and because several of its claims are impossible to verify (see McLaughlin 1987). The hypotheses forming the model are the following:
• Acquisition-Learning Hypothesis. There is a distinction to be made between acquisition and learning. Acquisition is subconscious, and involves the innate Language Acquisition Device which accounts for children’s L1. Learning is conscious and is exemplified by the L2 learning which takes place in many classroom contexts.
• Monitor Hypothesis. What is “learned” is available only as a monitor, for purposes of editing or making changes in what has already been produced.
• Natural Order Hypothesis. We acquire the rules of language in a predict able order.
• Input Hypothesis. Language acquisition takes place because there is comprehensible input. If input is understood, and if there is enough of it, the necessary grammar is automatically provided.
• Affective Filter Hypothesis. Input may not be processed if the affective filter is “up” (e.g. if conscious learning is taking place and/or individuals are inhibited).
In spite of being severely criticized by researchers, Krashen’s model had a major influence on language teaching in the USA in the 1980s and 1990s, including avoidance of the explicit teaching of grammar in many hundreds of classrooms. The pendulum has since begun to swing back in the opposite direction, with formal grammar teaching increasingly being introduced, especially with adults, who are able to benefit from (and may even need) an explicit explanation of grammatical structure.
The early period for linguistic study of SLA which we have just reviewed ended with some issues in rather spirited debate among proponents of different approaches, but there was widespread consensus on some important points. These include:
• What is being acquired in SLA is a “rule-governed” language system. Development of L2 involves progression through a dynamic interlanguage system which differs from both L1 and L2 in significant respects. The final state of L2 typically differs (more or less) from the native speakers’ system.
• How SLA takes place involves creative mental processes. Development of both L1 and L2 follows generally predictable sequences, which suggests that L1 and L2 acquisition processes are similar in significant ways.
• Why some learners are more (or less) successful in SLA than others relate primarily to the age of the learner.
As we reach the 1980s in this survey, new proposals in Chomskyan theoretical linguistics were about to have a major impact on the study of SLA, and Universal Grammar was to become the dominant approach with an internal focus.
الاكثر قراءة في Linguistics fields
اخر الاخبار
اخبار العتبة العباسية المقدسة
