Introducing Second Language Acquisition Perspectives (PDF)

4 Theoretical Perspectives: Past and Present It is the theory that decides what can be observed. Albert Einstein Chapter outline 4.0 Chapter overview 60 4...

4 Theoretical Perspectives: Past and Present It is the theory that decides what can be observed. Albert Einstein Chapter outline 4.0 Chapter overview 60 4.1 Contrastive analysis 61 4.2 Error analysis 64 4.3 Universal Grammar 67 4.4 The Monitor Model/Input Hypothesis 70 4.5 Information processing/cognitive approach 75 4.6 Processing-related hypotheses 77 4.6.1 The Interaction Hypothesis 77 4.6.2 Comprehensible Output Hypothesis 80 4.6.3 The Noticing Hypothesis 81 4.6.4 Input processing 85 4.7 Emergentism and usage-based approaches 86 4.7.1 Connectionism 87 4.7.2 Cognitive linguistics 89 4.7.3 Language as a complex adaptive system 90 4.8 The Sociocultural approach 91 Introducing Second Language Acquisition: Perspectives and Practices, First Edition. Kirsten M. Hummel. © 2014 John Wiley & Sons, Inc. Published 2014 by John Wiley & Sons, Inc. 60   Theoretical Perspectives: Past and Present 4.9 Summing up 93 Key concepts 94 Self-assessment questions 95 Discussion questions 96 Exercises/Project ideas 97 Further reading and viewing 97 References 99 4.0 chapter overview The term “theory” has been defined in various ways, but essentially it refers to “... a set of statements about natural phenomena that explains why these phe- nomena occur the way they do” (VanPatten & Williams, 2007, p. 2). For example, the psychologist Jean Piaget, introduced in Chapter 2, proposed a theory in which children go through four distinct stages in cognitive growth, ranging from the earliest sensorimotor stage when infants react to their environment based on what they perceive through their senses, to the fourth stage of formal operations, in which the 12-year-old becomes capable of abstract thought. Piaget’s theory was therefore an attempt to explain what he observed about children’s cognitive development. The philosopher of science Thomas Kuhn (1977) specified five criteria he considered should help determine theory choice, but at the same time Kuhn pointed out that people will apply these guidelines differently, given their indi- vidual perspectives: 1. Accurate—its predictions should correspond with results found through experiments and observations. 2. Consistent—it should be internally consistent, as well as with other cur- rently accepted theories about related phenomena. 3. Broad scope—a theory’s consequences should extend beyond what it was initially designed to explain. 4. Simple—it should be the simplest explanation. 5. Fruitful—a theory should disclose new phenomena or new relationships among phenomena. Another important concept related to theory is “hypothesis.” A hypothesis is different from a theory in that it makes a prediction about a distinct phenom- enon that derives from a more general theory. To continue with the example of Piaget’s work, a hypothesis derived from his overall theory is that children who have not yet reached the formal operations stage will be unable to perform certain tasks requiring abstract reasoning.  Theoretical Perspectives: Past and Present   61 As is the case in any other discipline, there is a long tradition of theoretical views and hypotheses underlying the study of second language acquisition, and these views have changed considerably over the years. An early view tended to consider language learning largely in comparison to first language acquisition. More recent views reflect the multifaceted set of influences on L2 learning. Some views highlight the contribution of innate principles that guide acquisition. Other perspectives stress the cognitive nature of L2 acquisition, with an accom- panying focus on mental processes such as attention and memory. Another approach points out the importance of input and practice in establishing links between language forms. In yet another view, the essentially social nature of the language learning process is highlighted. Learning about various theories and the approaches that have emerged from them can help us understand the L2 learning process better. What at first might appear as disconnected observations about language learning can become clearer when viewed through the prism of a given theoretical framework, and may allow new insights. These insights may ultimately lead to more effective learning and teaching practices. 4.1 contrastive analysis In the 1950s, a structuralist view of language was predominant among linguists. As the term “structuralist” indicates, the emphasis was on language forms or structures, as opposed to their functions. The structuralist linguist sought to identify and describe the observable units of a language. What can be termed a contrastive analysis (CA) approach characterizing second language learning contrastive analysis and teaching in the 1950s and 1960s represents such a structuralist view. (CA) Comparison of the During this period, in the neighboring discipline of psychology, behaviorist linguistic structures of two or more languages, to theories were predominant. As we saw in Chapter 2, a major figure in the behav- determine their similarities iorist school of thought was B.F. Skinner, a psychologist. In the behaviorist view, and differences. The CA L2 learning took place in the same way as any other kind of learning—it involved approach used CA as tool procedures such as imitation, repetition, and reinforcement, which enable learn- for L2 teaching. ers to develop “habits” of the L2. The process of learning either the L1 or an L2 was considered to be the establishment of habits through stimulus-response connections. When one learns an L2, the learner already has in place a well- functioning linguistic system, i.e., the L1. From the behaviorist perspective, Contrastive therefore, those L1 stimulus-response patterns have to be replaced by new L2 Analysis habits. Hypothesis (CAH) Behaviorist views led to the claim that first language acquisition and second This predicts that where language acquisition were fundamentally similar processes. For instance, it was there are similarities pointed out that in both situations imitation, repetition, and practice were between the two languages, the learner will involved to some degree and comprehension normally precedes production. acquire L2 structures with Stemming from the contrastive analysis view is what became known as the ease; where there are Contrastive Analysis Hypothesis (CAH). This hypothesis consisted of the asser- differences, the learner will tion that L1 interference is what constitutes the main obstacle to L2 learning. have difficulty. 62   Theoretical Perspectives: Past and Present As a result, a main objective of the CA approach was to carry out a structural analysis of the L1 and the target language, in order to determine areas of lin- guistic contrast. Those areas of contrast were precisely where it would be pos- sible to predict that there would be learner difficulties. Areas of similarity were those where the learner was predicted to learn easily. Thus, in this strong version of the CAH, a CA was expected to predict the areas of difficulty for L2 learners. The behaviorist underpinnings of this approach are evident in the references to “language habits” that appear in the linguist Charles Fries’ preface to the influ- ential book applying the CAH to language teaching, Linguistics across Cultures (Lado, 1957): Before... the questions of how to teach a foreign language must come the much more important preliminary work of finding the special problems... to develop a new set of language habits against a background of different native language habits... Learning a second language, therefore, constitutes a very different task from learning the first language. The basic problems arise... primarily out of the special “set” created by the first language habits. Robert Lado, who wrote Linguistics across Cultures, was raised in Spain and settled in the United States at 21 years of age. His experiences in learning English as an adult contributed to his views about language learning. Along with linguist Charles Fries, he developed the CAH and wrote Linguistics across Cultures to guide language teachers in the CA approach. He also developed the Lado English Series of books for English as a Second Language (ESL) learners, still widely used today, originally based on this approach. Some strong claims for the CA approach are found in Linguistics across Cultures (Lado, 1957, p. 59): Since the learner tends to transfer the habits of his native language, we have here the major source of difficulty or ease in learning the structure of a foreign language. Those structures that are similar will be easy to learn because they will be transferred and may function satisfactorily in the foreign language. Those structures that are different will be difficult because when transferred they will not function satisfactorily in the foreign language and will therefore have to be changed. It was therefore expected that learners would be likely to make errors in areas in which L2 structures differed from the L1. Meanwhile, areas which were similar between the two languages were expected not to pose a problem for learners. For instance, for the English-speaking learner of French, one would expect no problems in learning sentences formed by the sequence Subject Verb Object, since that same sequence characterizes most English sentences: “I see the book” (French: “Je vois le livre.”) On the other hand, when a pronoun is substituted for the object (book) in the French sentence, it is normally placed before the verb as in: “Je le vois” (I see it.)  Theoretical Perspectives: Past and Present   63 In the CAH approach, the expectation was that this SOV (Subject Object Verb) structure would be problematic for the learner, since it conflicted with the L1 structure. The learner was predicted to make the error “Je vois le” (“I see it.”) Indeed, this is a typical error on the part of English-speaking beginning learners of French. Some adherents to the CAH set out to establish what was referred to as a “hierarchy of difficulty” in order to allow predictions as to where L2 learners could be expected to make errors, based on the degree to which structures across the L1 and the L2 differed. This involved careful examination of the structures of both languages and an attempt to determine which areas constituted a higher degree of difficulty or difference. The hierarchy proposed by Stockwell, Bowen, and Martin (1965) ranged from a “zero level” of easy transfer between languages (for example, words that are cognates across languages, as in “table” in English and French), to a level 5, or “split,” in which one item in the L1 becomes two or more in the target language, predicted to be very difficult for the learner. For example, for English speakers, in learning a language such as French, the distinction between use of “tu” (“you” singular and informal) and “vous” (“you” plural and formal) must be learned where there is only one form “you” in the L1. In principle, therefore, the pro- posed hierarchy of difficulty would allow researchers to identify and predict where learners were most likely to need special instruction in order to overcome interference from the L1. Language learning in practice: the Audiolingual Method During this period in which behaviorism was largely predominant, language educators attempted to develop teaching materials that reflected the CAH, including the belief that language learning meant establishing an entire new set of habits and breaking the former L1 habits. These views largely contributed to the development of the Audiolingual teaching method throughout North America. The Audiolingual Method (ALM) focused on establishing “good” language habits through repetition, mechanical language drills and pattern practices, often carried out through language laboratory exercises. The Audiolin- gual Method will be discussed further in Chapter 5. Attempts to apply such a CA approach met with some success, but it soon became apparent that the approach was basically flawed. One problem that became evident quite rapidly is that it is very difficult, if not impossible, to make decisions about the degree of similarity or difference between languages because considerable subjectivity is involved. There was also a failure to find support for the predictions of CAH. For instance, one study (Whitman & Jackson, 1972) 64   Theoretical Perspectives: Past and Present revealed that linguists’ CA predictions applied to an English grammar test for Japanese native speakers were not supported: the Japanese speakers did not have most difficulty with the structures that differed the most from their L1. In addi- tion, other studies (e.g., Hernandez-Chavez, 1977) revealed that learners often make errors in areas where the L1 and L2 are fundamentally similar. In fact, it may be the more subtle areas of difference between languages that cause the most problems. One example is with regard to pronunciation: it may be that subtle distinctions, such as the articulation of certain sounds that vary in small ways from their counterparts in the L1, are more difficult to master than a totally new sound. For instance, for some learners, it might be more difficult to pro- nounce accurately the English counterpart of the “t” sound that occurs in their L1 since it differs in subtle ways (in French, for example, “t” is articulated with the tongue more forward in the mouth than in English), than to produce a sound absent from their L1, such as the one represented by the spelling “th.” There are other examples where cross-language similarity does not necessar- cross-linguistic ily lead to language transfer, i.e., incorporating aspects of one language into influence (CLI) This another. For instance, researchers (Hulstijn & Marchena, 1989) found that term refers to instances of Dutch L1 learners of English, whose L1 contains phrasal verbs (i.e., a verb plus phonological, lexical, grammatical, or other a particle, as in “get up“ or “give in”) just as English does, tended to avoid using aspects of transfer from them and instead opted for a single word. In fact, they particularly avoided one language to another. English items that most closely resembled their L1 phrasal verbs, i.e., “give up,” “break out,” “go off,” and “bring up.” Given the lack of supporting evidence, researchers proposed a less categorical version of the CAH. Rather than stating that CA allows one to predict areas of difficulty for language learners, it was suggested that CA can help in explaining the occurrence of certain errors resulting from transfer from the L1 into the L2. This more nuanced view of transfer across languages has come to be referred to as cross-linguistic influence (CLI). 4.2 error analysis The CA approach tried to predict and elimi- nate errors by emphasizing L1 interference. An examination of learner errors, however, ultimately led to the realization that many errors are not due to L1 interference, but rather to the learner’s own contribution in attempting to master the new linguistic system. Learners make many errors that are Cartoon 4.1 www.CartoonStock.com. due to overgeneralizing from their present  Theoretical Perspectives: Past and Present   65 understanding of the target language. For example, learners of English may tend to overgeneralize the past tense “–ed” ending on verbs: “I called” is correct, but “I goed” or “She hurted me” are not. Another point that the CA approach did not take into account is the finding that as learners advance in their mastery of the target language, the type and frequency of errors changes. Errors made in later learning are not necessarily the same ones that tend to occur in the early stages. In the 1960s, one researcher who came to be identified with a shift in thought Error Analysis (EA) about the role of errors in language learning was applied linguist Pit Corder. An approach to L2 Corder argued that errors are important in providing a glimpse into the learning acquisition research process and proposed a perspective that came to be known as error analysis involving the description (EA). Corder (1967, p. 167) pointed out: “A learner’s errors... are significant and classification of errors in that they provide to the researcher evidence of how language is learned or to gain insight into the learner’s current underlying acquired, what strategies or procedures the learner is employing in the discovery knowledge of the L2 of the language.” system. An important development during this period was recognizing learner lan- guage as a true system with its own regularities, an interlanguage (IL). Selinker interlanguage (IL) (1972, p. 214) defined interlanguage as a “separate linguistic system based on A term for the language observable output which results from a learner’s attempted production of a TL produced by a learner that differs in systematic ways [target language] norm.” A more recent, but similar, definition is Bialystok and from that of a native Sharwood Smith’s (1985, p. 101): “Interlanguage: a systematic language per- speaker. formance (in production and recognition of utterances) by second language learners who have not achieved sufficient levels of analysis of linguistic knowl- edge or control of processing to be identified completely with native speakers.” Other terms have also been used to characterize the learner’s developing system, including learner language, approximative system, and idiosyncratic dialect, but the term interlanguage has come to be widely accepted to refer to the learner’s evolving second language knowledge. As mentioned earlier, the CAH predicted that English learners of French would make errors such as “Je vois le” (“I see it”), following the sentence patterns of their L1. By the same token, it was predicted that French L1 learners of English would make the error “I it see” following the pattern of French “Je le vois.” However, it was found that French speakers do not tend to make that error. Because of wrong predictions such as this (as well as growing interest in a very different linguistic perspective, Universal Grammar, which will be covered in Section 4.3), CA grew out of favor. This led to greater attention being paid to analyzing learner errors and the recognition that one could not simply make predictions based on similarity or difference between language structures. The results of error analyses helped to discredit behaviorist views of L2 learning and increasingly the focus shifted to the learner’s own creative processes in develop- ing language. The study of learner errors by Corder and others tended to direct attention away from L1 interference and toward what was referred to as the learner’s “creative construction.” Ellis 2008 describes Corder’s 1974 proposed procedures for carrying out EA research, as outlined in Table 4.1. 66   Theoretical Perspectives: Past and Present Table 4.1 Error Analysis procedures. 1. Collection of learner language A sample corpus of learner language needed to be collected. Often, the sample came from written production tasks. 2. Identification of errors This step entailed examining the learner’s language and determining which errors were reflections of lack of knowledge and not simply momentary lapses, or mistakes. There are a number of problems that can arise in attempting to determine what actually constitutes an error (see R. Ellis, 2008, for an in-depth discussion). 3. Description of errors For some researchers, a description entailed examining sentence structures and pointing out whether an element was added, omitted, substituted or erroneously ordered (e.g., “She patted the dog big.”). Other researchers opted for describing errors in terms of which linguistic aspect was affected, i.e., phonology, vocabulary, morphology, syntax, etc. 4. Explanation of errors For error explanation the principal objective was to determine the source of the error, two main sources being transfer from the L1, and developmental developmental errors, such as L2 error An error in overgeneralization. (Note that it can be very learner language which problematic to determine the source of any given does not result from error. For instance, if an Italian L1 learner of transfer from the first English puts a final “s” on a noncountable noun language, but which such as “hair” to pluralize it, is it due to influence reflects the learner’s from the L1, as plural for “hair” in Italian requires gradual discovery of the a different final morpheme “i”: “capelli,” or from second language system. overgeneralizing the plural “s” rule in English?) 5. Evaluation of errors This stage involved establishing the effect a given error had on the learner’s conversational partner, such as whether it interfered with comprehension. Ultimately, the EA perspective was instrumental in drawing attention to the creative processes that guide the learner’s L2 development. It also contributed the view that errors can provide a glimpse into the learner’s developing inter- language: they are part of the learner’s ongoing attempts to acquire a new language Still there were a number of shortcomings in the EA approach, a very impor- tant one being that focusing on learners’ errors necessarily takes attention away from what learners are able to do correctly in acquiring a new language. Learn- ers also find ways to avoid certain structures or patterns that cause them diffi- culty, and avoidance is ignored in an EA approach. By the late 1970s and early 1980s, researchers were spending less time singling out errors for in-depth  Theoretical Perspectives: Past and Present   67 examination, and other theoretical perspectives were exerting their influence on the study of L2 acquisition. Language learning in practice: Communicative language teaching Influenced by perspectives that viewed the learner as creatively constructing an interlan- guage system, L2 classroom practices gradually changed from a rigid reliance on drills and repetitive pattern practices. Other influences included the notion of “communicative competence” (e.g., Hymes, 1972), which refers to the use of language in situationally and socially appropriate ways. Methodologies increasingly turned toward more com- municative, meaning-based activities. By the 1980s and 1990s, Communicative Lan- guage Teaching (CLT) had become widespread. Learning of grammar rules was relegated to a secondary role while teachers were encouraged to highlight interactive, conversa- tional meaning-based language in the classroom and to replicate a natural language learning environment as far as possible. CLT will be discussed further in Chapter 5. 4.3 Universal Grammar As was pointed out in Chapter 2, a new view of language emerged during the 1950s that developed out of the structuralist school of thought, but which dif- fered in important ways. Noam Chomsky, a linguist teaching at MIT, published a book (1957), Syntactic Structures, which proposed that languages are funda- mentally similar and that a primary goal of the linguist is to describe the under- lying rule system beneath the structure of all languages, known as Universal Grammar. It was proposed that a set of universal principles characterizes all human languages. This view sharply contrasted with that held by behaviorists and structuralists in that it regarded language as an innate ability, part of the human genetic endowment, and not as something that is largely crafted in response to environmental influences and pressures. Universal Grammar (UG) therefore adopts a nativist view, the view that emphasizes innate contributions to the task at hand. Another key concept in the UG approach introduced in Chapter 2 is compe- tence, the underlying knowledge that accounts for one’s language ability. The object of interest for UG scholars is competence, as opposed to performance or the actual use of that competence in real situations. Another term associated with the UG approach is “formalist” in that there is an emphasis on forms or structures, rather than behaviors, as characterized the behaviorist approach. From a UG approach, a basic question is whether L2 learners continue to have access to UG constraints. As similarly argued in L1 acquisition, the “poverty 68   Theoretical Perspectives: Past and Present of the stimulus” argument is raised, i.e., that exposure to linguistic evidence in the environment, the input, cannot on its own explain the abstract system of rules that ultimately characterizes the learner’s language. In other words, learn- ers appear to know what to do with a language even though they have not been explicitly taught certain rules; also they avoid making certain kinds of errors that are logically possible. How they do so is difficult to explain from looking only at the language they hear in their environment. A considerable body of research has gone into investigating the relation between UG and second language acquisition. White (2003, p. 38) summarized the UG view in second language acquisition (SLA) as follows: “... claims for UG operation in L2 acquisition are simply claims that interlanguage grammars will fall within a limited range, that the ‘hypothesis space,’ is specified by UG.” In this view, the interlanguage grammar, or interim grammar, does not need to be shown to be identical to the target language, only that it is constrained by the principles of UG. In terms of repercussions for L2 learning, there are a range of logical possi- bilities concerning whether or not Universal Grammar remains available for the L2 learner. We can describe three main hypotheses concerning access to UG, each of which has been supported by evidence. Note that, over the years, there have been modifications and adjustments to hypotheses, as well as about what counts as evidence, but our examples can serve to illustrate the basic issue. A first hypothesis is that the nature and availability of UG are the same in L1 and L2 acquisition: L2 learners continue to have full, direct access to UG, similar to the case of children learning their L1. There is some evidence for continuing parameters A small access to UG in the re-setting of parameters. UG parameters are a small set of set of alternatives for a alternatives for a given grammatical feature, for example, whether a comple- given grammatical feature; ment, such as a preposition (Prep), precedes or follows the main element (or for example, whether a complement, such as a the “head”), such as the Noun (N) of a noun phrase (Prep N or N Prep). For preposition (Prep), example, Japanese is considered as a “head-final” language, and the preposition precedes or follows the normally follows the head, for example: Nihon-ni (“Japan in”). Japanese speakers main element (or the easily acquire the English L2 head parameter in which the preposition occurs “head”), such as the Noun before the head (“in Japan”) and since that parameter is easily acquired, that (N) of a noun phrase (Prep evidence has been used to support the idea that UG remains available in learn- N or N Prep). ing an L2. In other words, UG guides both L1 and L2 acquisition (Flynn, 1996). A second hypothesis suggests that UG is present and available to L2 learners, but that the L1 acquisition process has to some extent affected its availability in L2 acquisition, so that learners access UG indirectly, through their L1. In this view, L2 learners need to receive some explicit information about what is ungrammatical in the L2, or they may expect L2 structures to resemble those of the L1. In a study with native Spanish speakers (White, 1985; 1986), it was found that Spanish speakers whose native language allows constructions with the initial subject pronoun omitted as in “Hablo inglès” (“I speak English”) who are learning English are more likely to accept phrases such as *“Should go” and *“Saw Suzanne,” than French L1 speakers whose native language does not permit the omission of initial subject pronouns (note that the asterisk before a sentence  Theoretical Perspectives: Past and Present   69 is a conventional notation form in linguistics to indicate something is ungram- matical). These results therefore suggest that the L1 acquisition process has indeed affected the availability of UG in learning the L2, since the Spanish speakers were found more accepting of English structures resembling those in their L1 than French speakers without subject pronoun omission in their L1. In a third hypothesis, second language learners no longer have access to UG. Evidence for this position stems largely from studies indicating important dif- ferences in performance between child L1 language learners and L2 learners, particularly with regard to ultimate linguistic outcomes; adult L2 learners rarely achieve native-like ability and there is substantial variation among learners in the proficiency level that they achieve. Given this view of UG as inaccessible, learners need to call upon other learning mechanisms and processes to learn the target language, and since learners have different degrees of control over these processes, this can help explain differences in levels of proficiency. This hypothesis allows for some knowledge of language universals via the native language. (See Figure 4.1 for illustrations of the three hypotheses.) Over the past years, the UG approach has seen major revisions, most notably the proposal of what is known as the Minimalist Program, which gives a greater role to the lexicon in determining grammatical features. This revision has led to important consequences for the data that would be considered relevant as evidence for the issue of access to UG. For instance, as pointed out by Lardiere (2012, p. 118): “... the notion of parameter-setting as a useful explanatory con- struct for (second) language acquisition must be reconsidered in light of these developments within linguistic theory.” Therefore, revisions to the theory may have important implications for the continued relevance of some conclusions from previous studies. Universal Other Grammar mental processes No UG access Direct access Indirect access L1 L2 Figure 4.1 Three hypotheses for the role of UG in SLA: Hypothesis 1: Direct access Hypothesis 2: Indirect access Hypothesis 3: No UG access The UG approach to L2 acquisition does not include an exploration of the effects of social factors or the overall sociocultural context of language. The objects of interest for the UG linguist are competence and the “ideal speaker- hearer” as expressed by Chomsky (1965, p. 3): 70   Theoretical Perspectives: Past and Present Linguistic theory is concerned primarily with an ideal speaker-hearer, in a completely homogenous speech community, who knows its language perfectly and is unaffected by such grammatically irrelevant conditions as memory limitations, distractions, shifts of attention and interest, and errors (random or characteristic) in applying his [sic] knowledge of the language in actual performance. This abstraction away from “real world” contingencies is an important weakness of this approach for some, including some scholars who emphasize the impor- tance of psychological factors affecting language learning in real time, or those working within the sociocultural approach with its emphasis on the social context of learning (both of which will be explored later in the chapter). Other researchers, such as those adhering to usage-based approaches also discussed in a later section of this chapter, argue that there is no evidence for a separate language module, drawing attention to general learning mechanisms to explain SLA. Language learning in practice: UG and language teaching Chomsky himself suggested skeptism about any practical applications of UG theory to language teaching, as illustrated by these widely cited remarks: I am frankly, rather sceptical about the significance, for the teaching of languages, of such insights and understanding as have been attained in linguistics and psychology... It is possible—even likely—that principles of psychology and linguistics and research in these disciplines, may supply insights useful to the language teacher. But these must be demonstrated, and cannot be presumed...” (Chomsky, 1966/71, pp. 152–155, cited in Widdowson, 2003.) While UG theory may have led thus far to few direct applications to the classroom, its impact on theoretical aspects of second language acquisition has been major. 4.4 the Monitor Model/Input Hypothesis In the mid-1970s, in the wake of the UG approach to language acquisition, applied linguist Stephen Krashen developed an influential model of L2 learning. Initially named the Monitor Model and, later on, the Input Hypothesis, the  Theoretical Perspectives: Past and Present   71 Learnt knowledge Monitor Model Krashen’s model of second language acquisition based Affective on the concept that filter (Monitoring) learners have two systems (acquisition and learning) Comprehensible Language and that the learned Acquired input Acquisition Output system acts as a monitor knowledge Device (LAD) (editor) of the acquired system. Figure 4.2 Krashen’s Monitor Model (from Cook, 1993, p. 54). Comprehensible input, thought to occur when the learner is exposed to real messages slightly beyond his or her current level (i + 1), is hypothesized as the source of acquired knowledge allowing produc- tive output; it is only accessible when the learner’s affective filter is not blocking access to it. Learned knowledge, such as the grammar lessons the learner receives in a classroom situation, only serves to allow monitoring, under certain circumstances, such as when the learner has enough time to refer to that learned knowledge. model consisted of five separate, though interconnected, hypotheses pertaining to second language learning (see Figure 4.2). Much debate and interest has been generated from this model and it has left lasting traces in L2 classroom practices, in particular. The first of the hypotheses was referred to as the Acquisition-Learning Acquisition- Hypothesis. This hypothesis states that there is a fundamental distinction learning between ways in which individual learners can develop language competence. Hypothesis Krashen’s Monitor Model proposed One way is through subconscious exposure to language, in which the learner that only “acquisition” or engages in communicative exchanges and is not consciously focused on subconsciously acquired language features, as tends to occur in naturalistic situations. Krashen labeled knowledge leads to this process “acquisition.” Another way is through conscious exposure to productive output; language data, as in a typical language classroom in which learners are taught “learning,” the learner’s conscious knowledge of the language through grammar and vocabulary exercises. This process the rules of a language, was labeled “learning.” In a controversial view in what became known as the only serves as a monitor. “non-interface position,” Krashen stipulated that what is “learned” cannot become “acquired” knowledge. This position was sharply criticized, as it was pointed out that Krashen offers no proof, and indeed offers no clear way of distinguishing “learned” from “acquired” knowledge. We should note, however, that Krashen’s attempt to differentiate “learning” from “acquisition” has led to interest in a similar, although more clearly defined, distinction referred to as “explicit” vs “implicit” processes (we will discuss this further later in this chapter). In this second hypothesis Krashen went on to state that “learned” knowledge served as a “monitor,” leading to the earlier label for this model “Monitor Model.” According to Krashen (1982, p. 15): “Learning has only one function... as a Monitor or editor... Learning... makes changes in the form of our 72   Theoretical Perspectives: Past and Present utterance, after it has been produced by the acquired system.” The Monitor was stated to function like an editor, in that it could allow the learners to correct their utterances, under certain specific conditions: that the learner be focused on the form of the utterance, that the learner have sufficient time to use the Monitor, and that the learner know the relevant grammatical rule necessary to correct the utterance. However, it was pointed out that Krashen offered no real evidence that such a device or learning entity actually exists. And, at any rate, when Krashen refers to the necessity of “knowing the rule” it is impossible to determine whether rules are conscious or not, and therefore this aspect is untestable. In addition, at least one study (Hulstijn & Hulstijn, 1984) found no support for the necessity of the “time” condition, once a focus on form was taken into account. Natural Order The third hypothesis of the model was the Natural Order Hypothesis. This Hypothesis Monitor hypothesis stipulated that language elements are acquired in an essentially fixed, Model hypothesis pre-determined order. This position was largely based on a number of mor- stipulating that language elements are acquired in pheme order studies that had been carried out during this period, in which it an essentially fixed, was concluded that L2 learners from different L1 backgrounds tended to acquire pre-determined order. L2 (primarily English) morphemes in the same sequence; for instance, the plural “s” before the third person simple present “s.” This part of Krashen’s model was similarly criticized, since a number of flaws were associated with the morpheme studies. One important weakness was conflating accuracy order with acquisition order. The morpheme studies most commonly cited in Krash- en’s model (e.g., Dulay & Burt, 1974) were cross-sectional in design, meaning that they compared results from different learners at different proficiency levels at one point in time, rather than following one or several individuals over time (longitudinal). Such cross-sectional studies were therefore based on comparing accuracy of production of the morphemes in different learners, which may or may not reflect what a given learner does at different stages of his or her own L2 development. At least one follow-up study (Rosansky, 1976) found that longitudinal results did not correspond to cross-sectional results on morpheme order. The Natural Order Hypothesis was also criticized for failing to consider factors such as ease of articulation or mental processing requirements associated with using morphemes. For example, learners were found to correctly produce the progressive participle ending “–ing” more often than the “–ed” simple past marker. It is conceivable that a learner produces the “–ing” ending as in “talking” more frequently than the simple past “–ed” form as in “talked” due to articula- tion factors (“talking” is arguably easier to pronounce than “talked”) rather than because he or she has acquired “–ing” before “–ed.” In addition, it was pointed out that the oral production test (the Bilingual Syntax Measure—see textbox) used to collect data in many of the morpheme studies with children may have unwittingly biased certain results, due to its exclusive reliance on oral production (e.g., Larsen-Freeman, 1975). For instance, results from written tasks with adults (e.g., Larsen-Freeman, 1976) have revealed different accuracy orders.  Theoretical Perspectives: Past and Present   73 Language learning in practice: The Bilingual Syntax Measure The Bilingual Syntax Measure is a structured communication technique that elicits natural speech from children in conversation with an adult. Children are asked questions guided by colored cartoon-like diagrams; for instance, the researcher might point to a picture and ask, “Why is the boy fat?” This could lead to the child replying “because he eats too much,” “because he eats all day,” etc. The researcher will then have samples of the child’s speech. If the child omits the “–s” ending on the verb by saying something like “because he eat too much” this can be noted. The fourth hypothesis of the model, the Comprehensible Input Hypothesis, Comprehensible stated that the one most effective way to increase L2 competence was by Input Hypothesis exposure to “comprehensible input.” Krashen stated that if enough input was A Monitor Model hypothesis stating that the understood, then the necessary grammar would be automatically provided. one most effective way to Comprehensible input was defined as language that was at an “i + 1” level, i.e., increase L2 competence is one level beyond the learner’s current level. Although it is hard to disagree with by exposure to the commonsensical view that learners benefit from exposure to input they can “comprehensible input” at least partially understand, much criticism was leveled at Krashen’s attempt to (one level beyond the learner’s current level). provide an equation for comprehensible input. Researchers pointed out that Krashen provided no means to determine what constitutes the learner’s current level and therefore no way to calculate a vague level “slightly beyond” that output The language produced by the learner. current level. In addition, the emphasis on input ignores any role for output which a number of studies (e.g., Swain, 1985; 1995) argue is essential in the L2 Affective Filter learning process. Output refers to the target language that the learner produces, Hypothesis A such as, for instance, when the learner replies to a question formulated by the hypothesis of the Monitor instructor in the target language. Model suggesting that an Finally, the fifth hypothesis, the Affective Filter Hypothesis, stated that affective filter can block learners possessed a mechanism, a filter, which, under certain conditions, could access to language acquisition under certain cause access to the language acquisition device to become blocked. The lan- conditions, such as when guage acquisition device is what UG linguists originally proposed as an innate the learner is stressed or component, or mental organ, to account for language acquisition. Conditions anxious. proposed to cause the filter to block language acquisition were primarily related to the learner’s emotional state, such as high anxiety in certain learning contexts. language Krashen based this part of the model on his own observations that learners at acquisition device a certain age, such as teenagers, tend to be self-conscious in front of their peers An element that UG linguists originally and therefore not likely to be receptive to language data. Criticisms of this proposed as an innate hypothesis included the observation that similar behavior was predicted for component, or mental both unstressed, unmotivated learners and unstressed, highly motivated learn- organ, to account for ers, since the affective filter is not predicted to affect either type of learner. And language acquisition. 74   Theoretical Perspectives: Past and Present again, the Filter is too vaguely described to permit a clear way to test the exist- ence of such an entity. On the whole, Krashen’s model stimulated heated debate among theoreticians and practitioners alike. Gregg (1984, p. 94) offered a particularly critical perspective: We have seen that each of Krashen’s five hypotheses is marked by serious flaws: undefined or ill-defined terms, unmotivated constructs, lack of empirical content and thus of falsifiability, lack of explanatory power. His second lan- guage acquisition theory is not a coherent theory, it is indeed incoherent to the point that it seems inappropriate to apply the word ‘theory” to it. It is not that I think Krashen’s ideas are wrong, although of course I do strongly disa- gree with some of his ideas. Indeed, although there is much that he seems not to understand about language and language acquisition, I think he is often right on the important questions: I agree with him that most language learning is unconscious, that comprehensible input is vital for learning and that a teacher’s most important job is to provide that input, that affective barriers can prevent successful acquisition of a second language and that a teacher has the duty to try to lower those barriers wherever possible. But then, does anybody disagree? Still, as a testimony to its resilience, aspects of Krashen’s Monitor Model con- tinue to generate debate and discussion. It also led Krashen and colleagues to develop a pedagogical approach, the Natural Approach (see textbox). The dis- tinction between “learning” and “acquisition” has led to more profound queries with regard to the role of attention in learning, as well as the distinction between explicit and implicit learning. Language learning in practice: The Natural Approach Krashen and Terrell (1983) applied the principles of the Monitor Model to the L2 class- room in the form of the Natural Approach, characterized by an emphasis on communica- tive activities, including exposure to large amounts of comprehensible input, or input slightly beyond students’ current level. To reduce anxiety, students were encouraged not to speak until they were ready. Specific classroom recommendations based on the Monitor Model (Dulay, Burt, & Krashen, 1982, pp. 263–269) were: 1. Maximize students’ exposure to natural communication. 2. Incorporate a silent phase in your program.  Theoretical Perspectives: Past and Present   75 3. Use concrete referents to make input comprehensible. 4. Devise techniques to relax students and protect their egos. 5. Include time for formal grammar for adults. 6. Learn motivations of your students and incorporate this into the lessons. 7. Create an atmosphere in which students are not embarrassed by their errors. 8. Use socially useful phrases in dialogues. 9. Certain structures should be learned first. The plural, for example, will probably be acquired before the possessive in English. 10. Do not refer to the L1 when teaching the L2. 4.5 information processing/cognitive approach Walid, the language learner from Detroit, attended his Arabic heritage language classes held in the evenings after college courses and on Saturdays. There were times when, tired from his long day at the community college and his part-time job, he struggled to focus on the lessons and ended up forgetting new vocabulary and expressions the teacher had just introduced. At such times, he was still able to carry on a conversation in Arabic, but more complex patterns seemed to overtax his mental resources. He realized there were limits to his attention. The information processing approach, also referred to more generally as a information cognitive approach, or cognitivism, is a major perspective that largely devel- processing oped out of the discipline of cognitive psychology in the 1950s. Applied to approach (or cognitive second language acquisition, this approach emphasizes that the mental pro- approach) Stemming cesses used for interpreting experience are also involved in the acquisition and from cognitive psychology, use of a second language. Therefore, while the UG approach specifies that there this approach emphasizes is a separate language faculty, the information processing approach assumes that that the mental processes the same mental faculties used for general problem solving and other cognitive used for interpreting experience are also activities underlie language use. For example, the faculties you use to under- involved in the acquisition stand a math problem are generally the same ones used to learn another lan- and use of a second guage. Note that many L2 researchers working in this framework accept the UG language. view that there is an underlying template for all languages, but may or may not agree that it remains available in L2 learning. What distinguishes the informa- tion processing approach from the UG approach is the emphasis on actual performance, as well as on investigating psychological variables related to learn- ing the language, and exploring how this new information is processed and acquired. A basic premise of this approach is that the mind is limited in terms of its capacity and this has important effects on the learning process. Mental processes 76   Theoretical Perspectives: Past and Present occur in real time so that there are limits to what the mind is able to do at any given time. The SLA process is therefore viewed as affected in important ways by attention, memory, and processing limitations. controlled One important distinction in this approach is between controlled processing processing In an and automatic processing. Controlled processing characterizes new skill learn- information-processing ing; it is relatively slow, requires attention, and is limited by short- term memory view, controlled processing characterizes new skill constraints. Over time, and through practice and training, a skill can come to learning; it is comparatively be performed rapidly and with little effort, and automatic processing takes over. slow and effortful, and is This increased automatization of processing skills results in increased fluency limited by short-term and greater comprehension on the part of the L2 learner. For instance, a begin- memory constraints. ning learner generally has considerable difficulty putting new L2 words together automatic to produce a simple sentence. Articulating new sounds and at the same time processing In an searching for appropriate words from a limited set of learned vocabulary items information-processing can be an arduous process. Over time, however, processes that were at the outset view, this occurs when a painstaking and time-consuming become more fluid and automatized so that skill becomes practiced and more attention can be directed to other aspects of the learner’s message, or to can be carried out expressing more complex thoughts. For instance, once the learner has a basic relatively rapidly and without conscious effort or mastery of target language sounds, more attention can be given to the words short-term memory and grammatical patterns used to convey his or her message. Learning is there- limitations. fore viewed as moving from controlled processing to automatic processing; automatic processing is carried out rapidly and without conscious control. The information processing view is interested in how this development proceeds and attempts to describe the effects of language learning as affected by mental pro- cesses within a limited capacity system. declarative Related to controlled and automatic processing are the terms declarative knowledge knowledge and procedural knowledge. (Note that this distinction is also made Knowledge that something in relation to memory systems: declarative memory vs procedural memory.) is the case, as in knowing a grammatical rule. Declarative knowledge consists of what the learner explicitly knows, as knowing the grammar rule that you need to add an “s” to a noun to form the plural in procedural English. Procedural knowledge is knowing how to do something without having knowledge The conscious awareness, as being able to produce L2 sentences without conscious knowledge of how to do reflection on what needs to be done first, second, etc., and is thought to underlie something; underlies automatic performance. automatic performance. Traditional information processing views suggest that declarative knowledge precedes procedural knowledge. As illustrated in an early view of information processing outlined by Atkinson and Shiffren (1968; see Figure 4.3), processing is divided into distinct stages identified with different aspects of memory: incoming information (sensory information) when attended to is transferred to short-term memory; short-term memory is where information is manipulated and, if rehearsed, is transferred to long-term memory, considered to be a relatively unlimited store for infor­ mation which can be drawn upon on a more or less permanent basis. (Note that we will look at a more recent memory framework “working memory” in Chapter 8.) This approach also has its critics. Similar to criticisms of the UG approach, it has been pointed out that influences from the surrounding social context are  Theoretical Perspectives: Past and Present   77 Rehearsal Incoming information Short-term memory Long-term storage memory Transfer storage Displacement (Decay) Figure 4.3 Atkinson and Shiffren’s information processing model (1968). © Elsevier. not fully recognized and acknowledged in most research carried out in the information processing framework. Also, the heavy reliance on laboratory studies to investigate language processing under controlled conditions has the drawback that language is largely examined outside of real communicative con- texts. Still, this approach has been a rich source of ideas and has contributed to the development of a number of promising views regarding second language acquisition, as we will see in the next section. 4.6 processing-related hypotheses Several hypotheses related to the information processing/cognitivist perspective have been proposed which offer distinct views on the role of input and the learner’s interaction with that input in second language acquisition. We will take a look at a few of these views in this section. 4.6.1 the Interaction Hypothesis Linda continued to make concrete plans to immerse herself in the Italian language and culture, and managed to get a leave of absence from her teaching job. From the moment she arrived in a small village in Tuscany, she delighted in her new cultural adventure. She continued to do her best to engage in opportunities to talk with the local population as she went about daily activities in the village where she had rented an apartment. She noticed that she often had to make repeated attempts at forming sentences in order to get her message across with her conversational partners, but she came to realize that her repetitions based on feedback from her friends helped her improve her ability to express herself in Italian. When she did not understand at first, she let this be known. She discovered that, for the most part, Italians were very pleased at her attempts to use their language and did not hesitate to let her know in various creative and enthusiastic ways when they failed to understand some of her utterances. 78   Theoretical Perspectives: Past and Present An influential view largely articulated by applied linguist Michael Long (1983; 1996), but which is related to similar views voiced by other scholars (Gass, 1997; Interaction Hatch, 1978), the Interaction Hypothesis emphasizes the important role of Hypothesis A conversational interaction in second language acquisition. This view suggests hypothesis proposed by that the modifications occurring during interactions between conversational Long (1983) predicting that interactional partners contribute to comprehension, in particular when communication modification makes input breaks down. While this perspective shares Krashen’s emphasis on comprehen- comprehensible; sible input, it goes beyond arguing for simple exposure to input, by stressing the comprehensible input important role played by the learner’s interaction with that input. Linda’s experi- promotes acquisition; ences in using her rudimentary knowledge of Italian, leading to modifications therefore, interactional in feedback provided from her conversational partners which further helped her modification promotes acquisition. refine her language skills, illustrates the usefulness of engaging in interactions. Long argued that modified interaction is the necessary mechanism for making language comprehensible. The Interaction Hypothesis therefore proposes that interactional modifications make input comprehensible, ultimately enabling acquisition. Long (1983, p. 151) pointed out: “Native speakers modify not just their speech with non-native speakers, but various features of the interactional structures of their conversations with them.” Native speakers do in fact modify their speech in sustained conversation with non-native speakers. Although it is less used in current literature, the term “foreigner talk” (Ferguson, 1971) has been used to refer to the modified lan- guage often used by native speakers in conversations with non-native speakers. Long’s emphasis on modified interaction refers to both linguistic simplifications and conversational modifications. Some types of modified interactions include: elaboration, slower speech rate, gestures, contextual cues, comprehension recasts Rephrasing an checks, requests for clarification, recasts (rephrasing an incorrect utterance cor- L2 learner’s incorrect rectly as when a learner says “I buy sandwich” and the instructor follows up with utterance correctly. “Oh, you bought a sandwich”), self-repetition, and paraphrasing. (Note that we will discuss interactive feedback further in Chapter 5’s section on instructional issues.) In an article on speech interaction (Gass & Varonis, 1985, p. 51), adjustments by native speakers (NS) in conversations with non-native speakers (NNS) become apparent: NNS:  How have increasing food costs changed your eating habits? NS:  Oh, rising costs we’ve cut back on the more expensive things. Gone to cheaper foods. NNS:  Pardon me? NS:  We’ve gone to cheaper foods. In the preceding dialogue, the native speaker reformulates the truncated phrase “gone to cheaper foods” into a grammatically complete sentence “We’ve gone to cheaper foods” which might be expected to facilitate understanding for the L2 speaker. Long (1996) put particular emphasis on how modified interaction contrib- utes to acquisition and underlined the importance of corrective feedback during  Theoretical Perspectives: Past and Present   79 interaction. The phrase negotiation for meaning (or negotiation of meaning) negotiation for is used for what is considered to be the opportunity for language development. meaning Process in Long (1996, p. 418) defines negotiation for meaning as: which learners and competent speakers interact in various ways, The process in which, in an effort to communicate, learners and competent making adjustments in speakers provide and interpret signals of their own and their interlocutor’s their speech until perceived comprehension, thus provoking adjustments to linguistic form, understanding is achieved. conversational structure, message content, or all three, until an acceptable level of understanding is achieved. Long (1996, p. 414) further explains how negative feedback, i.e., drawing atten- negative feedback tion in some way to learner’s incorrect utterances, contributes to second lan- Drawing attention in some guage acquisition. He also emphasizes the information processing aspects of L2 way to the L2 learner’s incorrect utterances. acquisition:... environmental contributions to acquisition are mediated by selective atten- tion and the learner’s developing L2 processing capacity, and... these resources are brought together most usefully, although not exclusively, during negotia- tion for meaning. Negative feedback obtained during negotiation work or elsewhere may be facilitative of L2 development, at least for vocabulary, mor- phology, and language-specific syntax, and essential for learning certain speci- fiable L1–L2 contrasts. As Long points out, the negative feedback that learners receive during negotia- tion helps draw attention to linguistic aspects not yet fully acquired by the learner. Learners benefit from interaction when they “notice the gap” (Schmidt & Frota, 1986) between the way they express their thoughts in the target lan- guage and the way a native speaker does so. In support of the interaction view, some research (e.g., Pica, Young, & Doughty, 1987) has revealed that conversational adjustments do indeed help L2 comprehension. Other studies (e.g., Gass & Varonis, 1994) have shown that learners involved in interaction with native speakers benefited in terms of later production on an oral task, while learners similarly exposed to discourse input, but not allowed to interact, did not. Further studies have offered additional insights; for instance, one study (Mackey, Gass, & McDonough, 2000) found that some linguistic aspects in feedback carried out in the form of negotiations and recasts were more salient to learners than others: lexical and phonological feedback was more accurately perceived than grammatical feedback from native or near-native speakers. There is also evidence that L2 learners can benefit from conversational inter- action with other L2 learners as well. In a study of conversational interaction between ESL learners (Long & Porter, 1985), 12 adult ESL learners—both inter- mediate and advanced—and six adult native English speakers served as partici- pants. Learners were found to talk more with other learners than with native 80   Theoretical Perspectives: Past and Present speakers. Furthermore, when the number of grammatical and vocabulary errors and false starts were compared, there was no significant difference between situ- ations when learners were matched with other learners or with native speakers. The authors of this study concluded that learners themselves offer genuine com- municative practice, including negotiation of meaning. 4.6.2 Comprehensible Output Hypothesis Xia Mei, studying English in her immersion program in Hong Kong, looked for any and every opportunity to use English outside school. She was introduced to a Canadian family who had a daughter her age and they began a friendship. Xia Mei found herself regularly engaged in English conversations with her new friend and discovered that she often stumbled over words and phrases that she needed to use to express her ideas. Many of those items had been introduced in her immersion classroom, and she thought she had mastered them, but she had never previously needed to actually use them in real-life conversations. As time went on and Xia Mei continued to meet with her friend, she became more skillful in expressing herself in English and found herself needing less time to search for the right words and expressions. Comprehensible The Comprehensible Output Hypothesis, originally influenced by cognitivist Output Hypothesis views, was primarily articulated by applied linguist Merrill Swain (e.g., 1985). Swain (1985) proposed This hypothesis suggests that having to communicate in the target language that having to produce the L2 encourages the learner actually contributes to improvement: “producing the target language may be the to attend to the language trigger that forces the learner to pay attention to the means of expression needed and thereby improve in order to successfully convey his or her own intended meaning” (Swain, 1985, proficiency. p. 249). To illustrate, Xia Mei’s frequent conversations with her new friend helped her improve her English speaking skills; these regular interactions helped compensate for the lack of opportunities to use the target language in her immersion classroom. Swain looked at data from English L1 children enrolled in French immersion programs in Canada who had little or no contact with the target language outside of school. She found that despite seven years of schooling, the Grade 6 immersion students differed notably from native speakers and that this was “particularly evident in those aspects of communicative performance which demand the use of grammatical knowledge” (Swain, 1985, p. 245). Swain argues that comprehensible input alone is not sufficient to ensure full L2 proficiency and suggests that what was missing from the immersion classrooms she studied were opportunities for comprehensible output. This hypothesis suggests that the demands of producing comprehensible output “push” learners ahead in their development. In contrast to Krashen’s views described earlier in this chapter, which put all the emphasis on input,  Theoretical Perspectives: Past and Present   81 learner output is given an essential role and is thought to enhance fluency and lead to greater automaticity. Another benefit is that producing output can help the learner analyze grammatical aspects of the target language, as the learner attempts to structure his or her thoughts. Output also serves to draw attention to gaps in the learner’s L2: the challenge of having to put thoughts into concrete words and sentences helps the learner to realize where there is room for improve- ment in his or her L2 abilities. Another cited benefit of output is that it elicits relevant input. When learners produce output during the course of a conversa- tion, their conversational partners provide them with input, and that input can be useful for the learner in formulating sentences and conveying their messages in more appropriate ways. 4.6.3 the Noticing Hypothesis Steve made steady progress in learning Mandarin during his year abroad in Beijing. He was particularly encouraged when he realized that much of his improvement occurred when he paid attention to the language forms used by his conversational partners. After several months, he found he had reached a stage where he could often grasp the basic meaning of a sentence, which allowed him to direct specific attention to the way in which that meaning was communicated, that is, to focus on pronunciation of words and phrases. His close attention to pronunciation appeared to transfer to his own utterances. The Noticing Hypothesis, formulated by applied linguist Richard Schmidt (e.g., Noticing Schmidt, 1990; 2001), was largely influenced by the information processing view Hypothesis Proposed of second language acquisition. An important distinction that underlies the by Schmidt (1995), it states that what learners notice Noticing Hypothesis is related to the differences among input, intake, and in the input is what output. Input normally refers to the language to which the learner is exposed; becomes intake for as, for example, when the learner listens to the teacher in a classroom who is learning. speaking in the target language. But input is not identical to intake. Since learn- ers do not necessarily register all of the input to which they are exposed, a intake The part of the further distinction is necessary: intake is “that part of the input that the learner input that the learner notices. notices” (Schmidt, 1990, p. 139). There are many reasons why the input might not be noticed or registered: for example, lack of attention, incomprehension, and cognitive overload. In the case of Steve studying in Beijing, his first months of immersion in a Mandarin environment were overwhelming in terms of his exposure to the language, as well as to new cultural concepts and experiences. It was only after he was able to better comprehend basic sentences, along with the cultural context underlying certain acts of speech, that he was able to focus attention on how native speakers pronounced certain sounds that were difficult for him. 82   Theoretical Perspectives: Past and Present Schmidt formulated the Noticing Hypothesis based on the experience he had in following the L2 progress of “Wes” (a pseudonym), a Japanese native speaker and an artist, who moved to live in Hawaii, where the main language spoken is English. Schmidt observed Wes over a three-year period, leading to 18 hours of recorded oral transcripts (Schmidt, 1983; 1984). The following are some samples of Wes’s L2 utterances: I like talk to people you know / I’m always listen then start talk/ then listen/ always thinking my head/ then talk / some people you know only just talk, talk, talk ,talk I’m never learning, I’m only just listen then talk. Yesterday I’m go beach. If you back to room, can I bring cigarette? So tonight, Tim and me we are come back here early, we are apartment. She’s name is Izumi Ukimura, she’s working is beautiful. Schmidt found the following examples where Wes appeared not to make any grammatical progress: Excessive use of “–ing” Past form only in very frequent irregular forms (“went”, “saw”) Absence of “–ed” or correct past form, instead, use of adverbs to indicate past (“Yesterday I write letter.”) Overall, Schmidt noted a general lack of sensitivity or attention to grammatical form in Wes’s language. He points out that Wes’s manifestly positive attitude toward using English, as well as his exposure to a facilitating environment, were evidently insufficient to allow him to make significant progress. He concluded that learners must also actively process the input. Schmidt’s observations led him to formulate the “Noticing Hypothesis.” According to Schmidt (1995, p. 20): “The noticing hypothesis states that what learners notice in the input is what becomes intake for learning.” Noticing does noticing Paying attention to language not of itself result in acquisition, but it is the essential starting point. Therefore, input with some from the Noticing Hypothesis viewpoint, L2 learners do not begin to acquire a awareness. language feature until they become aware of it in the input. There are a number of different factors which lead to noticing (Schmidt, perceptual salience 1990): The noticeability of a feature (e.g., if a given Frequency feature is given more emphasis in an utterance, Perceptual salience (the noticeability of a feature) it is more perceptually Instructional strategies salient). Individual processing ability  Theoretical Perspectives: Past and Present   83 Influences upon noticing Input how frequent Working memory noticeability Noticing Output Long-term Instruction memory Figure 4.4 Influences upon noticing (Skehan, 1998, p. 49). © Oxford University Press. Readiness to notice Task demands (what the learner needs to accomplish using the target language) Figure 4.4 illustrates the place of noticing in SLA. Another study by Schmidt provided further evidence for the Noticing Hypothesis (Schmidt & Frota, 1986). Schmidt followed his own progress in learning Portuguese during a stay in Brazil. By keeping a diary of his L2 devel- opment and comparing his remarks with the input he received both in and outside a language class, and with samples of his oral production, he found links between what he noticed in the input and what he actually produced. Some forms were taught and frequent, but not learned if not noticed. Schmidt cites the example of the phrase “o que é que,” a long question form in Portuguese. This form was present in the input he heard for five months, but it was not until he noticed the form that he began to use it himself. Schmidt (2010, p. 725) makes a distinction between noticing as “a technical term limited to the conscious registration of attended specific instances of lan- guage,” and a higher level of awareness (“understanding”) that includes gener- alizations across instances which would characterize rule knowledge as well as metalinguistic awareness. According to Schmidt (1990), awareness at the level of “noticing” is necessary for SLA, and awareness at the level of “understanding” is facilitative but not required: [what is noticed are] “elements of the surface structure of utterances in the input–instances of language, rather than any abstract rules or principles of which such instances may be exemplars” (Schmidt, 2001, p. 5). In addition, Schmidt (1995, p. 26) specifies that “learning requires awareness at the time of learning” which allows for later forgetting: the Noticing Hypothesis therefore does not suggest that what is noticed is necessarily retained in memory. Whether learners must be aware they are “noticing” something in the input in order to acquire linguistic features is considered debatable. Some researchers 84   Theoretical Perspectives: Past and Present argue that awareness may be necessary for certain types of learning, but not others:... although I (among many others) am perfectly willing to agree that learning individual words (the lexicon), individual sounds (the phonetic inventory), and writing systems must be via attentional focus, I am not the least willing to say that learning phonological, morphological and syntactic rules requires this attentional focus. (Schachter, 1998, p. 574)... some linguistic items are too rare, abstract, complex, or semantically opaque, or have too many irregularities to be explicitly noticeable by learners. The article system in English is a typical problem for learners of many different native languages. But the fact that learners with no apparent metalinguistic knowledge are often successful with complex linguistic structures such as the English article system suggests that implicit learning can occur in such cases. Indeed, with low literate adolescents and adults, this type of implicit learning of L2 systems may be the norm. (Tarone, Bigelow, & Hansen, 2009, pp. 23–24) Therefore, while there appears to be agreement that noticing is conducive to L2 learning, there remain doubts as to the extent to which noticing in the form of explicit awareness is a fundamental necessity for all aspects of L2 learning. There is recognition that rule generalization or understanding (recall Schmidt’s dis- tinction above between “noticing” and “understanding”) may occur without explicit awareness. Schmidt has responded to the previous objections with the argument that most evidence suggests attention is in fact necessary for learning, although he admits some kinds of learning require more focused attention than others (Schmidt, 2010). The Noticing Hypothesis is linked to the distinction between implicit and explicit learning in SLA. We recall that Krashen made a similar, although less clearly explained distinction, between subconscious “acquisition” and conscious implicit learning “learning.” Implicit learning has been defined by DeKeyser (2003, p. 314) as Learning without “learning without awareness of what is being learned” as opposed to explicit awareness of what is being learning, or learning with explicit awareness. There are problems in finding learned. ways to accurately test for implicit learning. For instance, if a learner fails to explicit learning report awareness of a given linguistic feature, it may be that he/she no longer Learning with explicit recalls it, despite awareness at time of exposure. There are various views on the awareness of what is being role of implicit learning in SLA. For instance, N. Ellis (2005, p. 306) argues that learned. most language acquisition “... is implicit learning from usage: the vast majority of our cognitive processing is unconscious.” Others argue that it only occurs under very specific conditions, if at all (see DeKeyser, 2003, for a review). Also related to the implicit–explicit learning distinction in SLA is the sugges- tion (Paradis, 2004) that the automatization of grammatical rules results in implicit computational procedures that differ in nature from explicitly learned grammatical rules.  Theoretical Perspectives: Past and Present   85 Another important, but related, distinction is between implicit knowledge implicit knowledge and explicit knowledge. These terms refer to the results of the learning process. In SLA, knowledge of the Rod Ellis (2009, p. 6) describes the distinctions between the two sets of concepts L2 that underlies the learner’s performance, but (implicit/explicit learning vs implicit/explicit knowledge): of which he or she is not explicitly aware. It is possible, for example, that learners will reflect on knowledge that they have acquired implicitly (i.e., without metalinguistic awareness) and thus, explicit knowledge subsequently develop an explicit representation of it. Also, it is possible that In SLA, knowledge of the L2 (vocabulary, grammar explicit learning directed at one linguistic feature may result in the incidental rules, etc.) of which implicit learning of some other feature. learners