Introduction On Learning a Language: Some Theoretical Perspectives In Chapter 1, various ways to define and describe language competence wore ex plored, and components of language that were thought to be Important in de- signing models of "communicative competence” and “language proficiency ‘were identified and considered. We saw that many of the same components (grammatical, lexical, phonological, pragmatic, sociolinguistic, and discourse fea- ‘tutes) were included in the various models that have been proposed. Although we have not reached complete consensus on the question of what It means to know a language, the profession is in basic agreement about the features of language ‘that are relevant to that quest “This chapter addresses another fundamental question that concerns language sesearchers and practitioners: How do,adults become proficient in a second lan- guage? Consensus about this question may be far more dificult to achieve. Ellis (1985) comments that there has been a great deal of theorizing about second- language acquisition (SLA), especially since the early 1970s, and that "the re- search literature abounds in approaches, theories, models, laws, and principles” (p. 248). He speculates that perhaps the profession has generated ‘ar too many theories, agreeing with Schouten (1979) that “tao many models have been built and taken for granted too soon, and this has stifled relevant research” (p. 4, cited in Ellis 1985, p, 248). Spolsky (1989) argues for the development of a unified macso-theory—a new general theory of second-language learning—and outlines seventy-four separate “conditions” that would need to be integrated mto such a comprehensive model, McLaughlin (1987) takes the view that although micro- theories, which try to deal with a stnaller range of phenomena and are limited In scope, may be “intsinsically more satisfactory" (p. 9}, a theory must be compre. hensive enough to explain more than a very limited range of phenomena: "A sat- 51TEACHING LANGUAGE IN CONTEXT+214 esition isfactory theory of adult seconé-language learning must go beyond accounting for how people form relative clauses” {pp.910). He adds that, given the relatively carly stage of the development of knowledge in the field of second-language ac- quisition, “it seems premature to argue for the ‘truth’ of one theory over another” (p. 6) Latsen-Freeman and Long (1991) agree, suggesting that it would be coun- texproductive for SLA researchers to espouse one single dominant theory of lan= {guage acquisition, particulary as this might discourage competing points of view. ‘We must suard against overzeaousress on the part of theorists or ther devotees who fel tha they have a monopoly on the truth. While SLA research and Language teaching will benefit rom the advantages of theoretically motivated researc... it would be dangerous at this stage for one theory to become omnipotent (p. 290). Practitioners who have been buffeted across the years by pressures to adopt differ. cent approaches to teaching, due to the changing winds of theory, may tend to agree with this resistance to theoretical “bandwagons" (Grittner 1990), ‘Why do language teachers need to know about theory, especially fit seems t likely that we can reach agreement about how language learning and acquisition, take place? One reason might be that most language teaching methodologies have ‘grown out of a particular theoretical framework of second-language acquisition, and it would be helpful for teachers to understand some of the premises underly ing those approaches in order to evaluate them, A second reason for understand ing a range of theoretical viewpoints is that it may help teachers develop and/or clan their own set of prineiples for language teaching, Fllis (1985) maintains that every teacher already has a theory of language learning, but that many teachers may have never articulated what that theory is. The fact, however, that we choose to do certain activities in the classroom and decide not to do others shows that we are working on some underlying assumptions about what is useful in promoting the development of language proficiency. Therefore, before examining some of the theories that have been influential in the field of language teaching over the years, tight be constructive to make a preliminary assessment of some of the as- sumptions that may underlie our own beliefs about language learning. iustration 2.1 presents a set of questions that can serveasa gulde fordiscussion or as an instrument for self-assessment to help teachers clarify and articulate their current beliefs about the way adults develop competenceina second language. The reader may want to consider these questions before going on to the next section. Exploring Theories of Language Learning Recent reviews of language acquisition theory (McLaughl. 1978, 1984, 1987; Ells, 1985, 1990; Brown 1987, 1994; Larsen-Freeman 1991} have attempted to group various theoretical perspectives along a kind of continuum, ranging ftom ermpiri= cist views on one end to atfonalistor mentalist positions on the other, with theories that blend these two perspectives placed somewhere in between. This oppositionON LEARNING A LANGUAGE = 58 IMustration 2.1 ‘This et of questions is designed to hep teachers explore their assumptions and beliefs about biscussion Guide: elias second Janguage learningand teaching. The questions relate to some ofthe issues that under- about Seconé-Langu2ge {ie various theories of language acquisition in this chapter. Learning 1, Do adults learn foreign tanguages in a manner similar tothe way children acquire thelr, native language, o are the processes involved in child and adult language learning, different? 2. Ane humans bam with a special capacity for language tat i specific to our species? Oris ‘anguage learning like other kinds of learning, governed by general cognitive processes not specific to language? if we are born with a specialized capacity for acquiring a native language a children, does it work the same way with adults who are learning a foreign or second language? 3. How does our knowledge of our native language affect our learning of anew language? oes some ofthe knowledge we have transfer to the new language? I so, is this helaful, ‘or can ithe a hindrance? 4. What s the optimum type of "input? for adults who are beginning their study of a Foreign language? Do they profit best fram listening o native speakers for some initial period of time before being aked to sneak? Should the input they receive be ordeted or sequenced carefully to correspond to what they already know? Oris sufficient that the input be relatively comprehensible, even if some structures have not yet been studied? 5. What role does interaction with native speakers, teachers, or other learners have in [anguage acquisition? What kinds of information about the target language can we obtain through such interaction? What kinds of information can we obtain about our own ‘developing language proficiency when we interact with others? 6. What isthe role of explicit grammar instruction in adult foreign language learning? Can adults become proficient ina second language without having conscious knowledge of the rules ofthat anguage? Or do adults profit in some way from grammar explanations and examples of how specitc features are used? 7. Do language learners acquire grammatical features in 2 predictable order when language Fearing occurs in natural-use situations? Does instruction in formal dassrooms need to follawa “natural order” be effective? £8. What isthe role of practice in adult language learning? Is language learning ike the earning of tier “kills” such as earning to playa musical instrument, where a great deal of focused practice fs mecessary to becomne proficient? Oris language learning fundamentally differen fom other forms of human learning? 9, Do students need to have an opportunity to practice new forms and structures in “controlled activities before being asked to communicate their own meaning using those features? Or should students be encouraged te engage in conversation activities where ‘communication isthe main focus from the beginning of language instruction? When learners ate engaged in meaningful and creative communication, do they tend to make more errors than when they are doing controlled or forn-focused activities? 10. What isthe role offeelbackin language learning? How important iit to give Fearners information about whether they are making erorsas they use the new language? fit better to correct most or allo! the errors students make, or should error cortection be ‘minimal inthe language classroom? What are optimal ways to provide feedback to adult foreign language learners?54 - £3 From Empiricism To Rationalis ‘An Empiricist Perspective: Behaviorism TEACHING LANGUAGE IN CONTEXT + 3¢d edition of viewpoints is not new; Chomsky had made the rationalist/empiricist distinc tion in discussing linguistic theory in 1965, and Diller (1978) spoke of the exis- tence of a longstanding “language teaching controversy” between the rationalists ‘and the empiricists “whose roots can be traced to the beginnings of modem thought” (p. vi). The basic difference between the two positions seems tole in the prestimed locus of control of the process of language acquisition, The rationalist position includes theories that assume that humans have an innate capacity for the development of language, and that we are genetically programmed to develop ‘our linguistic systems in certain ways (Chomsky 1965). Larsen-Freeman (1991) refers to this point of view as a “nativist” or “innateness” position, which is in strong opposition to the “behaviorist” or “environmentalist” perspective. Ths lat- ter position is characteristic ofthe empiricists, who maintain that itis the learner's ‘experience that is largely responsible for language leamning and is more important ‘than any specific innate capacity (Larsen-Freeman 1991, p. 323). McLaughlin (1978, 1984) characterizes the empiricist viewpoint as one that is skeptical of any ‘explanation of language learning that cannot be observed, Learning is seer. as the result of external forces acting on the organism rather than the programmed un+ folding of language through internat biological mechanisms. Empiricsts, there- fore, assume that there is no special species-specific language ability, but that language learnings just one aspect of generat learning ability or capacity. ‘The next section provides a sampling of theories representing these different categories or classifications, chosen to reflect some of those perspectives that have had the most influence or potential influence on classroom practice. Because there is such a profusion of competing theoretical viewpoints in the professional literature, this discussion will not be comprehensive. The interested reader would do well to consult additional sources such as Bllis (1985, 1990), Brown (1987, 1994), McLaughlin (1987), Spolsky (1989), Larsen-Freeman and Long (1991), Gass and Selinker (1994), Towell and Flawkins (1994), Cook (1996), and Mitchell and Myles (1998) for more detailed treatments of a wide spectrum of theoretical viewpoints. : A Theoretical Sampler ‘The various theories of language learning to be discussed in this section have been placed along the continuum in [lustration 2.2, which depicts in graphic form the range of viewpoints referred to in the preceding pages. The placement on the con- num Is not meant to be exact or precise, but rather locates theories ina general way in terms of their compatibility with empiricist 0: rationalist points of view, ‘The characteristics and underlying assumptions of each of these theories will be briefly summarized below. For a more thorough treatment of a particular theory, consult the primary sources in the references. ‘Since ancient times philosophers have believed that human learning and animal leaming might be slmtlar (Chastain 1976). Chastain points out that it was the publication of Darwin's Origin of the Species in 1859 that made this belief mareON LEARNING A LANGUAGE oa EMPIRICISTS RATIONALISTS (Environmentalist) (MentalistiNativist) BEHAVIORIST PSYCHOLOGY UNIVERSAL GRAMMAR (Skinner) (Chomsky) MONITOR THEORY (Krashen) CONNECTIONISM PARALLEL DISTRIBUTED COGNITIVE THEORY PROCESSING (McLaughlin; Anderson; (PDP) Shiffrin & Schneider; Ausubel) (McClelland, Rumelhart & PDP Group, Gasser) Mustration 2.2 The Rationalis-Empiricis Continasim credible, since Darwin's theory implied that there was indeed continuity be- tween the htiman species and the lower animals, and by implication between the human mind and the animal mind, In the late nineteenth and early twentieth centuries, a growing interest in animel behavior led to the growth of experime tal psychology and the school of behaviorism, ‘According to S-R (stimulus-response) psychology, all behavior is viewed as a re- sponse to stimull, whether the behavior is overt (explicit) or covert (Implicit). Ac- cording to the theory, behavior happens in associative chalns; all leazning is thus characterized as assoctative learning, or habit formation, brought about by the re- peated association of a stimulus with a response (Hilgard 1962). This process of habit formation, or conditioning, was thought to be of three basic types: (1) classi- cal conditioning, (2) operant conditioning, and (3) multiple response learning. (pp. 253-274. In classical conditioning (best knovrn trough experiments done by Pavlov), an association between a conditioned stimulus and a response was repeatedly stvengthened through the presentation of that stimulus with another, uncondi- tioned one. In Pavlov’s experiments with dogs, the unconditioned stimulus was meat powder and the response was sallvation, When Yavlay repeatedly presented ‘meat povrder with the simultaneous ringing of a bell, the dog learned to sali-: - TEACHING LANGUAGE IN. CONTEXT «30d edition vate to the sound of the bell (the conditioned stimulus), even in the absence of the meat [In operant conditioning (also known as instrumental conditioning), the response toastimulusis eamed although itis not normally a natural response to that stim- ulus. A rat pressing a bar in its cage may at first do so randomly. But if the rat dis- covers that pressing the bar releases a food pelle, it learns to push the bar again for the same reward. The operant (the random bar-pushing behavior) becomes conditioned (purposeful behavior) because it produces an effect that is rewarding. In muitipte-response learning, the animal leaens a whole chain of behaviors and performs them in succession, always in the same order, A rat that runs a maze learns a fixed sertes of turns through conditioning, rewarded by a food pellet ot ‘two for his trouble ‘What has all of this to do with language learning? As Chastain (1976) points out, behaviorism took a strong foothold in the thinking of psychologists by the middle of the twentieth century, influencing, in tum, the views of the education ‘community: ‘Soon behaviorists concluded that all learning consisted of some form of ‘conditioning, The orgazsism was conditioned to respond in a specific way to @ selected stiraslus, Complex activities were nothing more than a complex collection of conditioned responses. Since al learning i conditioned and since dumara learaing is simitar to learning in animals, the next step was to conclude ‘that human learning could be, ands, conditioned in the sae way. The belief was that haraans are reinforced by their environment in mach the same way as the rat in a maze (p. 195}. B. F Skinner (1957), perhaps the best known proponent of $.R psychology, used the term operant consitioning to describe verbal leaening. In his view, lan. guage is characterized as a “sophisticated response system! that humans acqul ‘through automatic conditioning processes (Wacchaugh 1976, 2, 142). Some pate terns of language are reinforced (rewarded) and others are not. Only those pat- ‘tems reinforced by the community of language users will persist. In Skinnetian Psychology, the human being is likened to a machine with multiple working parts, The mind js thought to be “a tabula rasa upon which are stamped associa. tons” between various stimuli in the environment and responses chosen from ‘outside the organism for reinforcement (Chastain: 1976, p. 133) ‘Skinner's theory of verbal learning was consistent with the prevailing beliefs of ‘many applied lingaists of the 1940s and 1950s who maintained that second lar- ‘guages should be learned through extensive drill and practice without recourse to ratlonalistic explanation, In his Outline Guide for the Practical Study of Foreign Lan. $1385 (1942), Bloomfield had argued for an essentially behaviotistle appeon ‘The command ofa language is not a matter of knowledge: the speakers are quite tunable to describe the habits which make up heir language. The command ofa Iamguage is a matter of practice. ... Language leaming is overlearning: anything else Is of ro use (Bloomfield 1942, p. 12, cited in Chastain 1976, pp. 107-08),ON. LEARMING.A.LAGUAGE, : 37. Ilustration 23 summary: Behaviorist Theory (Based on Skinner 1957; Hilgard 1952; Chastain 1976; Wardhaugh 1976) CRITIQUE: Sumnmary: Behaviorist Theory 1, Hurnan learning and animal iearning are similar. 2. The child's mind is a tabula rasa, There is no innate pre-programim specifically for language learning at birth. 43, Psychological data should be limited to that which is observable, 4, All behavior is viewed as a response to stimuli. Behavior happens in associative chains; in fact, all learning Is associative in nature. 5, Conditioning involves the strengthening of associations between a stimulus and a response through reinforcement. 6, Human language is a "sophisticated response system" acquired through operant conditioning. Mlustration 2,3 summarizes the main points of the behaviorist view of lan- ‘guage learning, Behaviorist theory, in conjunction with the structuralist views of language that prevailed in the 1940s and 1950s, laid the theoretical foundations for audiolingual language teaching methodology, discussed in more detail in Chapter 3, We have seen that behavioristic theories of language learning were based on the assumption that language learning was like any other kind of learning, and, therefore, one could extrapolate heavily from general learning theory and even from animal leaming. This viewpoint was seriously challenged by Chomsky (1959) in a very critical review of Skinner's work, Chomsky maintained that lan- guage behavior was far more complex than the establishment of $-R connections, and that Skinner's theory could not possibly explain the creativity of children in generating language. ‘According to McLaughlin (1978, 1984), Skinner's 1957 treatise, Verbal Behavior, ‘was not supported by research with human subjects. There was, in fact, no sub- stantial research base ever generated by behaviorists to look at child language use, Tet alone second-language learning. He adds that evidence gleaned from subse- quent studies of child language behavior shows that a simple behavioristic per- spective does not provide a satisfactory explanation of what has been found: It ‘seems that imitation and reinforcement have a much smaller role to play in chile Janguage than Skinner and his colleagues imagined. For example, children often produce forms that they never heard their parents or other adults say ("I goed” or wo foots"). Thus, imitation of adult speech cannot completely account for the way children produce language: “The child's language is simply too strange” (McLaughlin 1984, p. 15), Furthermore, parents rarely correct their children’s ‘grammatical ertors but respond Instead to the message content (Brown and Han- Jon 1970; Brown 1973), If ungrammatical formis are thus positively sewarded (or at least ignored), how then do children eventually eliminate them? A behaviorist view of language, which would predict the need for both imitation and negative feedback in the form of overt corrections, does not seem to explain the way in which children learn. With Chomsky’ review of Skinner's theory there came a paradigm shift toward58 TEACHING LARGUAGE IN CONTEXT: 3rd edition the other end of the theoretical continuum. Hf language development was highly creative, then language learning theories needed to account for the creative pro- cessing that was taking place in the human mind. By the mld-1960s, the pendu- lun was swinging in the direction of the rationalist point of view. | Three Rationalist Perspectives of Language Learning 1. Universal Grammar Various reviews of theories of language learning (Chastain 1976; Wardhaugh 3976; MeLaughlin 1978, 1984, 1987; Bis 1985, 1990; Brown 1987, 1994; Larsen Freeman 1991) group a variety of perspectives within the “rationalist” camp. Other terms used in asocation with this perspective are “nativist,* “mentalist,” dnd “cognitive” A highly influential nativist viewpoint grew out of Chomshy’s swork, starting with the publication in 1957 of his book Syitactic Structures, and his Tritique of Skinner in 1959, As we saw earlier, Chomsky had rejected the behav- forist perspective and adopted instead a mentalist viewpoint that wes closely ¢e- tated to the basic principles and beliefs of cognitive psychology (Chastain 1976) Other theorists, such as Eric Lennenberg (1967) and David McNeil (1966) De- Hieved that language was a species-specific, genetically determined capacity and that language learning wes therefore governed by biological mechanisms. in 41965, Chomsky had concluded that chtldzen were born with some Kind of special tanguage processing ability and hacl proposed the existence ofa “language acqui- sition device” (LAD). A yeat later, McNeil (1966) characterized this LAD as having various innate linguistic properties. Brown (1994) summarized them to include: (i) the ability to distinguish speech sounds from other sounds; (2) the ability to ‘organize language into a system of structures; 3) the koovsleige of what was pos ‘Sble and what was not possblein any inguisticsystem; and (4) the ability to con- Struct the simplest possible system based on the linguistic data to which ore was exposed ‘Chomsky argued further that it must be the case that children were innately programmed to acquire language since they do itso quickly (in just afew years) find with such limited (ari less than idea!) input, He also believed that they could hot help but construct a certaln kind of linguistic systema particular transfor- national or generative grammar—any more than they could help the way theit ‘sual system percelved solid objects or lines and angles (Chomsky 1965). Al- though a child’ experience with language input could have an effect on language rearing, the “ultimate form will be a fanction of those languaye universals that existin the human mind (McLaughlin 1984, p16). Universal Grammar theory posits the existence a a set of basic grammatical e cements of "fixed abstract principles" that are common to ail natural human lan- fquages and that predispose children to organize the input in certain ways, The principles themselves are thought to be innate, a product of the “LAD." They in Glade substantive universals, which consist of fixed features of languages Ike phonemes or syntactic categories like nouns and verbs, as well as formal univer. Sale, which are more abstract, and which place limits or constraints on the poss bie rule systems or on the options children have for constructing 2 grammar (Chomsky 1965, pp. 27-30; Bills 1985, pp. 192-93)ON, LEARNING A LANGUAGE 59 Mustration 24 ‘Summary: Universal Grammar Theory (Based ‘on Chomsky 1965; Elis 1985; Metaughlin 1967 Larsen-Freeman 1991) carniaut Summary: Universal Grammar Theory 1, Language is a species-specific, genetically determined capacity 2. Language learning is governed by biological mechanisms, 3. The ultimate formn of any human language isa function of language ‘universals, a set of fixed abstract principles that are innate, 4, Bach language has Its own *parameters” whose “settings” ate learned on the Dasis of linguistic data. 5, There isa "core grammar,” congruent with universal principles, and a "peripheral geamaar,” consisting of features that are not part of universal grammar, 6. Core grammar rules are thought to be relatively easier to acquire, in general, than peripheral rules ENis (1985) provides the following example of a formal universal: One might formulate certain principles that place limits on how languages can use word onder transformations in order to form questions. All languages must operate ‘within those limited options, yet each language has its own particular “parameter settings” for question formation. The child's task isto discover which of the vari ous options applies Ia his or her language. This Is where environmental input is crucial: The child needs to hear the Janguage spoken in order to select the appro= prlate options and thus set the parameters correctly. ‘According to Chomsky, the universal principles that children discover constt- tute their “core grammar,” which is congruent with general principles operating across all languages. The “peripheral grammar” consists of rules or features that are not determined by universal grammar, but that might be derived from an older form of the language, borrowed from another language, or that might have arisen accidentally (Cook 1985, cited in Ellis 1985). Rules of the core grammar ‘might be easier to acquire than the rules of the peripheral grammar, since the lat- ter “are thought to be outside of the child's preprogrammed instructions” (McLaughlin 1987, p. 96). Wesche (1994) suggests that language leamers pro- ably acquire peripheral rules through the use of “general cognitive abilities” (p.239), CChonsky’s Universal Grammar theory and associated derivative approaches to the study of linguistic universals are quite complex. Most discussions of the ze- search in this area require some specialized knowledge of theoretical linguistics in order to fully understand the findings. (However, see Pinker [1994] for a very readable discussion of some of the important aspects ofthis theoretical approach fo language acquisition.) As was mentioned earller, the discussion of theories in this chapter is meant to be introductory in nature; readers interested in a more de- tailed treatment should consult the sources cited in this section. For a summary of some of the main premises of Universal Grammar theory that have been pre~ sented here, ee llustration 2.4 “Although Chomsky’s generative grammar theory has had a wide-ranging influ- feice on the field of linguistics and on theorles of how children acquite a native language, Universe] Grammar theory has not won universal acclaim, Beedham60 TEACHING LANGUAGE LN CONTEXT +3¢d edition (1998) reviews the work ofa number of scholars who have been critical of genera tive models (such as Gross 1979; Hall 1987; Moote and Carling 1982, among oth. 1s) and criticizes the basic methodology of the generative approach to lenguage acquisition. He maintains that all models of “generative grammar” have at Teast two basic flaws: (1) confusion of “mathematical notation” with linguistic form and (2) circularity of argumentation: ‘The principles and criteria of Principles and Parameters theory ave merely assumptions, with nothing to back them up except the circular argument that ‘without thers language would be unexplained ... (Beedham 1995, p. 209) Ieedham also steongly ctlticlzes generative grammar theory because of what he claims to be its “complete lack of applicability.” Although he recognizes that the- retical subjects are different from applied subjects, he maintains that at some point, a theory needs to be empirically tested in some type of application: Ths is yet to happen to generative grammar, Certainly itis now universally recognized that generative grammar cannot be appliee to language teaching (p-214). However, the question of how Universal Grammar might play a role in adult language learning is still subject of much debate inthe field of second-language acquisition. McLaughlin (1987) states that “Universal Grammar theory does not concem itself with second-language acquisition” (p. 91), but that a number of second-language researchers have applied principles of Universal Grammar to this domain in an effort to find sufficiently sophisticated explanations ofthe very complex. characteristics of interlanguages. Wesche (1994) maintains that al- ‘hough Universal Grammar theory is widely accepted in fist language acquis tion, second-language acquisition specialsts disagiee about whether Universal Grarnmar continues to operate in adult learners o play any significant role. She audds thet ever iF does play a roe in second-language acquisition, Is Imited 10 the core grammar and does not help explain how leamers acquire such important features of language as the elements of peripheral grammar, vocabulary, discourse competence, oF other performance features Ttalso does not help explain "the dre- ‘matic individual differences found in the rate and ultimate mastery ofthe second language” (p. 239) Some theorists operate on the assumption that the same universls that chil ‘dren use to consttuct their native language aze available to adults; others believe that they are no longer available, and that different cognitive processes must be involved In adult secone/foreign language learning (sce Larsen-Freeman 1991), Gass and Selinker (1994), for example, discuss the “Fundamental Difference Hy- pothesis,” proposed by scholars such as Schachter (1988) and Biey-Vroman (1989), wino argue that adults no longer have cigect access to UG principles, Inv stead, they maintain that child language acquisition, especially ofa irs language, and adult language acquisition of a second language are quite different in several important ways. First, adults rorely achieve native levels of proficiency or full competence in a second language, whereas children normally do achieve this inON LEARNING. A LANGUAGE a1 2. Krashen’s Monitor Theory: First- and Second- Language Acquisition Are Similar their native language, Children can learn any of the world’s languages equally ‘wel; acts expestence differing levels of iffcaly, depending on how closely the foreign language is related to their native language. (We sa this, for example, tn the difficulty hierarchies outlined by the Foreign Service Institue in Chapter L) Gass and Selinker further point out that adults and children have differing levels of knowledge about how languages work, given that adults already have full com petence in their native language wen they begin to learn a foreign Language. One additional difference isthe role thet motivation and attitude toward the target language can play in adult language acquisition, Differences in motivation do not seem (o have any appreciable impact on the chila's learning his or ier native [an- sage; however, motivation and attitude are important factors for adults learning. 2 foreign Imguage (pp. 124-25), Nevertheless, many scholars in the feld of second-language acquisition stil feet that Universal Grammar can play a role, and Gass and Selinker (1994) maine tain that much of second-language acquisition research “is criven by the notion ‘hat frst and second language acquisition involve the same processes” (p. 124) Tie rationalist theories that ace discussed in the next two sections represent two Afferent perspectives on how frst and second-language acquisition are related ‘One of the most influential and widely discussed models of language learning/ac quisition in recent years is Stephen Krashen’'s “Monitor Model.” The most con plete description of the theory (1982) describes five centzal hypotheses 1. The acquisition-learning distinction, which "states that adults have two distinct and independent ways of developing competence in a second language": ‘acquisition, which is a subconscious process “similar, ifnot identical, to the way children develop ability in thelr fest language"; and learning, which refers to conscious knowledge of the rules of grammar of a second language and their application in production (p. 10), 2. The natural order hypothesis, which maintalns that acquisition of grammatical structures (primarily morphemes) follows a predictable order when that acquisition {s natural (Le., not via formal learning) 3. The monitor hypotkesis, which states that acquisition is the sole initiator of all second-language utterances and is responsible for fluency, while learning (conscious knowledge of rules) can function only as en “editor” or “monitor” for the output. This monitor operates only when there is sufficient time, the focus is on form, and the language user knows the rule being applied. 4. The input hypothesis, which matntains that we acquire more language only when we are exposed to "comprehensible input"—language that contains structures that are "alittle beyond” our current level of competence (i+ 1), bbut which is comprehensible through our use of context, out knowledge of the world, and other extralinguistic cues directed to us. According to this, hypothesis, acquitets “go for meaning” fist, and, asa result, acquire structure as well. A third part of this hypothesis states that input need not be deliberately planned to contain appropriate structures (1+ 1): IfCRITIQUE: TEACHING LANGUAGE IN CONTEXT #34d edition ‘communication is successful and there is enough of it,/+ 1 Is provided automatically. A final part of the input hypothesis maintains that speaking fluency cannot be taught directly, but rather “emerges” naturally over time. Krashen maintains that although early speech is not grammatically accurate, accuracy will develop over time as the acquirer hears and understands more Input. S. The affective filter hypothests states that comprehensible input can have its effect on acquisition only when affective conditions are optimal: (1) the acquirer is motivated; (2) he has self-confidence and a good self-image; and (3) his level of anxiety Is low. When learners are “put on the defensive” (see Stevick 1976), the affective filter is high, and comprehensible input can not "get in.” (For a fuller account of these five hypotheses, see Krashen 1982, pp. 9-32.) ‘Krashen suggests that there are certain Implications for classroom practice if i language instruction Is to be consistent with his theory. Among these are: 1. The main function of the classroom may be to provide comprehensible input in an environment conducive to a low affective filter (Le,, high motivation, low anxiety). 2, The classtoomn Is most useful for beginners, who cannot easily utilize the informal environment for input, That Is, ILis useful for foreign language students who do not have input sources outside of class ot for those whose competence is so low that they are unable to understand the language of the outside world (pp. 32-37). 3. The requirements for optimal Input are that it be (a) comprehensible, (b) interesting and relevant, (c) not grammatically sequenced, (d) provided in sufficient quantity to supply #1, and (¢) delivered in an environment where students are “oif the defensive” (p. 127) 4, Error correction should be minimal in the classroom; itis of some limited use ‘when the goal is learning, but of no use when the goal is acquisition. Ercor comrection talses the affective filter and should, therefore, not be used in free conversation or when acquisition is likely to take place (pp, 116-117). 5, Students should never be required to produce speech in the second language ‘unless they are ready to do so, Speaking fluency cannot be taught, but “emerges” naturally in time with enough comprehensible input. Ifustration 2.5 summarizes the main premises of Monitor Theory. A more com: pletely developed model of language teaching using Krashen’s theory asa basis is given by Teneell (1977, 1982), His "Natural Approach” is discussed in detail in Chapter 3 ‘A number of the hypotheses and assertions in Krashen's theory of second- language acquisition have been challenged in recent years. In an early review of the Monitor Model, Munsell and Carr (1981) questioned the distinetion betweenOM LEARIUING LANGUAGE 6 Uustration 2.5 Summary af Monitor “Theory (Based on Krashen 1982) Summary: Monitor Theory 1, Adults have two distinet ways to develop competence in a second language: acquisition, which is a subconscious process, and Jeamning, which Is conscious. 2, Acquisition is similar to the process by which children acquire their native language. Learning Involves conscious knowledge of rules. ‘When acquisition is natural, the order in which certain grammatical features of the language are acquired is predictable. 4, Learning can function only as an “editor” of what is produced, since acquisition is the sole initiator ofall second-language utterances. Learning can serve as a “monitor” of performance only under certain conditions, 5, We acquire new structures only when we are exposed to “comprehensible input” ((+ 1), Input does not need to be deliberately structured or planned for the acquirer, If communication is successful, i+ 1 will happen automatically. For acquisition to take place, the learner must be motivated, have a good self- mage, and be free from anxiety. 7, bttor correction should be minimized in the class:oom, where the main purpose of instruction should be to provide comprehensible input. “Yearning” and “acquisition” and the notion of “conscious” and “unconscious” rules, The reviewers also seem to abject to the underlying nativist assumptions of the model and the implications that language learning is distinct from other kinds of learning, In their view, language skill Is much like other kinds of skilled performance: Krasien may not wish to extend Monitor Theory to chess, yet the measured characteristics ofthe knewledge of skilled chess players bear some striking similarities to the characteristics of linguist knowledge... Sinifarly, suiclt disparate areas of skill as sports annd mathematics seem to benefit fra early cenaphasis on conscious and systematic learning despite the fact that expert performances ta these areas also display a rumnber of characteristics that formally resemble expert performance in language. We cannot imagine trying to earn basketball, monopoly, bridge, or quantum mechanics simply by watching people do them, trying them, and creatively constructing the rules. It is much easier to start with some conscious exposition ofthe rules end build one's skill ‘upor: that founclation (pp. 498-99), Munsell and Care imply that Krashén should incorporate language learning the- ry into a wider context where the nature of human skilled performance in gen: eval is explored, This point of view is congruent with the commentary on Monitor ‘Theory made by McLaughlin (1987) who leans toward a more cognitive perspec- tive, ‘McLaughtln’s objections to Monitor Theory are summarized in the following five points:64 3. Cognitive Theory: First- and Second: Language Learning Differ TEACHING LANGUAGE IN CONTEXT+3ed edition “The acquisition-learning dlstinction is not clearly defined.” Therefore, the centval claim that Krashen makes that “Ieacniing” cannot become “acquisition” cannot be tested (p. $6) Various studies have shown that the Monitor does not work the way Krashen, originally thought it would, and he has had to place more and more restrictions on the conditions under which it would be used effectively. ‘McLaughlin believes that these restrictions make Krashien’s, ‘conceptwalization of “learning” of limited usefulness in explaining a learner's conscious knowledge of grammar. ‘The case for the Natural Order Hypothesis is quite weak due to methodological problems. “If the Natural Order Hypothesis is to be accepted, it must be ina weak form, which postulates that some things are learned. before others, but not always” (p. 56) 4, Since no clear definition of “comprehensible input” is given, McLaughlin believes the Input Hypothesis s also untestable. “The Affective Filter Hypothesis is also questionable, not only because Krashen has not explained how this filter develops, but also because it does not take individual differences among learners into account. McLaughlin states that this hypothesis s incapable of predicting the course of Unguistic development with any precision, Although Krashen’s theory has been criticized on a variety of points by a num- ber of scholars, it has also had a strong influence on thinking in the field over the past twenty years. Virtually everyone who talks about language learning in recent ‘years seems compelled to consider whether itis “learning” or “acquisition” that is the focus of attention in one’s remarks. Many people feel that the distinction has at least an intuitive appeal and that it represents some psychological reality. In the same way, many practitioners recognize the need to provide learners with "com- prehensible input” and find Krashen’s recommendation that affective considera- tions be primary in the classroom very appealing. In many ways, Krashen has articulated in his Monitor Theory hypotheses about language learning that have touched a responsive chord for many practitioners, This is not to say, however, that the criticisms reviewed above should not be considered seriously as one eval ates the merits of Monitor Theory. ‘As mentioned earlier, some theorists prefer a view of language learning that recognizes essential differences between the way children and adults process in- formatlon. Although there may be some similarities between child and adult lan- guage learning, Cognitive theory predicts that adult second-language learning will differ in some imporiaat ways from the way In which children acquire their native tongue, Larsen-Freeman and Long (1991) categorize various cognitive approaches to lan ‘guage acquisition as “interactionist” views, where both external and intemal fac- tors are considered in accounting for language acquisition (p. 266). Although this characterization may be valid, the emphasis on environmental factors seemsOW LEARNING A.LANGUAGE = aa rather limited when compared to the role assigned to internal or mental processes in descriptions of Cognitive theory given by Ausubel (1968), Ausubel, Novak, and Hanestan (1978), Ells (1985, 1990), and McLaughlin (1987, 1990), For this re son, the theory has been placed toward the rationalist end of the continuum in Tl- lustration 2.2, ‘We have seen that Universal Grammar theory considers the role of innate lin gulstic universals in language acquisition and claims that there is @ specltic in ‘sulstic capacity that fs unique to the human species. Cognitive theory, by ‘contrast, derives from the field of cognitive psychology and focuses on the role of ‘more general cognitive processes involved in language acquisition, such as trans- fer, simplification, generallzation, and restructuring (McLaughlin 1987), Like ‘Universal Grammar, Cognitive theory is in direct opposition to Behaviosist the- ‘ory because, from a cognitive perspective, learning is believed to result from in- temal mental activity tather than from something imposed irom outside the leaner (Filis 1990). McLaughlin (1990) characterizes the cognitive approach to second-language acquisition as follows: 1. Cognitive psychology emphasizes knowing rather than responding and is concerned with studying mental processes involved in the acquisition and use of knowledge, "The focus is not stimulus-response bonds, but mental events" (p. 113). 2.'The cognitive approach emphasizes mental structure or organization. Cognitive psychology assumes human knovrledge is organized and that anything new that is learned is integrated into this structute 3. Cognitive theory, as oppased to Behaviorist theory, views the learner es one who acts, constructs, and plans rather than simply receives stimull from the environment, Therefore, a complete understanding of human cognition ‘would require an analysis of strategies used for thinking, understancin, ‘temembering, and producing language. According to Cognitive theory, second-language learing is seen as “the acqu}- sition of a complex cognitive skill” (McLaughlin 1987, p. 133). For a language leatner to become proficient, subskills ofthis complex task must be practiced, au~ tomatized, integrated, and organized into internal representations, or rule sys- tems, that are constantly restructured as proficiency develops. “Automatization refers to the process of making a skill routine through practice MeLaughlin (1987) explains the way this is thought to occur using an informa tion processing model developed by Shifrin and Schneider (1977). n this model, ‘memory is thought to consist of a large number of “nodes” that become assocl- ated with one another and activated in sequence through learning, In automatic processing, certain nodes are activated almost every time a certain input Is pre- sented. This activation pattem has been built up through consistent practice so that it becomes a learned response over time. Once such an automatic response 1s learned, it occurs quite quickly and is difficult to suppress ar change (Shiffrin and Schneider 1977, pp. 185-56)66 TEACHING LANGUAGE IN CONTEXT# 304 edition In contvoed processing, memory nodes are activated in a given sequence on a temporary basis—that i, the response has not yet been “learned” or automatized, For the response to happen, the learner has to give the process his full attention, It's difficult, therefore, to do “controlled” tasks if there is any distraction or inter- ference, Shifftin and Schnetder speculate that for the development of “complex infor- mation-processing skills,” such as learning to read, learners would use controlied processing first, aying down “stepping stones of automatic processing” as they ‘move from Jower to higher levels of learning: In short, the staged development of skilled automatic performance can be interpreted asa sequence of transitions fiom controlled to automatic processing -170). Schmidt (1992) points out that although “automatic" and “controlled” process {ing were originally thought of in terms ofa dichotomy, mote recent discussions of these concepts suggest that they really should be viewed as ends of a continuum. He emphasizes the role of practice in moving new material along this continuurn, affirming the earlier speculations of Shiffrin and Schneider, cited above: ‘The development of skilled! behavior Involves a shift with practice fiom controled to artomatic processing. Novices of ail kinds, including beginning L2 learners, ‘must pay careful attention to every slep of the procedure, whereas experts do not (chia 1992, p. 360). In discussing the development of speaking fluency, he suggests that various levels of processing may actually be used simultaneously, a point that Shitfrin and Schneider (1977) also make when referting to complex processing such as reading (see p. 161). Schmidt argues that, rather than thinking of the processing of speech as sequential in nature, it should be seen asa type of “parallel” processing, He cites, Levelt’ (1989) assertion that if It did not involve parallel processing, “speaking ‘would be more like playing chess: an overt move now and then, but mostly silent processing” (Levelt 1989, p. 27, cited in Schinldt, p. 376). Schmict adds that for novice speakers, itis indeed the case that “speaking sometimes does seem to re- Quire as much thought and effort as planning a chess move" (p. 376), Those of us who have taught beginning language learners can testify to the truth of this ob- servation; it should also lead us to consider the possibility of giving leamers more time to plan their discourse when asking them to express theit own meaning in the foreign language In beginning and intermediate classes, ‘The distinction between controlled and automatic processing can be useful as one considers the various tasks involved in second-language learning. Tarone (1982, 1983) describes a whole range of language “styles” that learners produce hen engaged in various kinds of tasks, The vernacular style, vepresented by infor. ‘mal use of the language with litle attention to form, 1s produced when language \s being processed automatically. The careful style, on the other hand, is elicited. ‘when learners engage in heavy monitoring and/or attention to the form of thei production. This monitoring represents a more controlled processing of the laON.AEARNING.A.LANGUAGE guage noeded to accomplish the task Tasks that demand such monttoring include grammeticality judgments or form-focused production activities of various kinds Farone (1982) explains that the learners’ interlanguage system should be thought of ava continacien, ranging from the vernacular to the careful style, and does not, fs Krashen (1982) has claimed, consist of two discrete systems differentiated on the basis of whether attention to form is conscious or subconscious. “The “variability” of learner language is evident when students at diferent pro- ficiency levels engage in tasks of different sypes. Teachers may have noticed this phenemenon of variability when thetrstudents perform differently whlte doing a ‘iserete-point grammar task on a test or for an assignment than they do when using the language more naturally ot informally in conversation or in free com position. Tarone (1987) adds that other factors, such as the kdentty or te role of the learner's conversational partner, the topic of conversation, the mode of clis- ‘course (Le, the functions that are being performed, such as giving directions, de- scription, narration, argumentation, and the Iike), and other task oF situational variables can have on effect on the accuracy of the language produced. Rather than feeling frustrated and confused by this phenomenon, teachers and students night be encouraged by a view of language learning such a5 ths that accounts for such differences in performance. ‘While Shiffrin and Schnelder contrast controlled and automatic processing, Fils (1990) adds Anderson's (1980, 1995) distinction between declarative and pro- cedural knowledge as another Way t0 100k at how information is processed and Stored, Declarative knowledge is explicit and conscious, and can be articulated by the learnes, It involves "knowing that” (¢.g,, definitions of words, facts, rules) Pro- ‘cedural knowledge, on the other hand, is “knowing how” (ef, how to produce lan guage as one performs Hingulstically), This type of Knowledge might be more or Jess implicit or explicit, conscious or unconscious, or relatively controlled or tela tively automatic in natuse (ee Anderson 1995 for a thorough discussion of these ‘two types of knowledge). Andersons model of skill acquisition consists of three Stages: (1) the cognitive stage, where learners use conscious declarative knowl: ‘edge; (2) the associative stage, where they start to proceduralize this knowledge: and (2) the autonomous stage, where performance becomes more or less auto- matic and errors dlsappear (Andetson 1995, pp. 273-275), Allof these models at- tempt to explain the processes by which learning becomes internalized and eventually “automatic,” but each looks atthe processes involved in somewhat dif: ferent ways. Cognitive theory further maintains that there fs more {o developing a complex skill han automatizing the sub-skills of which itis comprised (McLaughiin 1987). “The learner also has to impose an organizational structure on the new informa- tion that is constantly being added to the system. As new information is learned, the organization of the existing information might have to be changed, of *te- structured," to accommodate what is new. That is why both autoratization and re- Structuring ace Key concepts in this view of language learning (pp. 134-136) “The idea of the development of Internal "structures" or organized cognitive systems and networks is central to views of learning that derive from Cognitive