Use the following outline structure as you move through the assignment: Thoughts regarding the examination of language teaching beliefs (a few sentences) 1. I believe that… a.Specify and summarize SLA

Keywords:Teacher beliefs; Teacher education; Second language acquisition 1. Introduction The contribution ofsecond language acquisition (SLA) research to English language teaching and to teacher education remains problematic (Ellis, 1997; Lightbown, 1985). While one ofthe inten- tions ofSLA research is to improve language pedagogy (Ellis, 1997), and most SLA researchers have at some time been language teachers (Tarone, Swain, & Fathman, 1976), teacher education programs in general and SLA courses on teachereducation programs in particular are often said to be either excessively theoretical (Brown, 1983; Brumfit, 1983; Lightbown, 1985) or not ‘relevant’ to what goes on in the classroom (Eykin in Markee, 1997). Markee goes so far as to say that ‘‘most potential consumers ofSLA research are frequently repelled by its disregard for real world issues’’(1997, p. 88). Some ofthis criticism might be perceptual; but some ofit could be grounded in the nature ofSLA research itself. In a recent review of50 examples ofSLA research, it emerged that only 15 were actually carried out in authentic language class- rooms (Nunan, 1991). So it is not surprising that Stephen Krashen has concluded that theory is ‘‘rejected by most language teachers’’ (Krashen, 1983, p. 255).

*Corresponding author. Tel.: +44-1786-467933; fax: +44- 1786-466-131.

E-mail addresses:[email protected] (M. MacDonald), [email protected] (R. Badger), [email protected] (G. White).

0742-051X/01/$ - see front matterr2001 Elsevier Science Ltd. All rights reserved.

PII: S0742-051X(01)00042-7 1.1. Theory and practice This division between theory and practice has been echoed by many ofthe undergraduates and postgraduates studying with the authors on B.A.

and M.Sc. programs in teaching English to speakers ofother languages (TESOL). Our student teachers pride themselves on their pragmatism, just wanting to ‘get on with the job’ oflearning the day-to-day practicalities ofhow to teach in the classroom. In keeping with a majority ofsimilar programs (Ellis, 1997), our program contains a course in SLA. We find that our student teachers’ pragmatism conflicts with the theory and research aspects ofSLA. Here are examples ofthe sort of feedback we have been receiving over the years both orally and from the students’ written evalua- tion forms on our courses in SLA:

This course was much more theoretical than I am used to.

This course gave me information overloadFI was not sure how it fitted into the classroom.

There is an imbalance between theory and practice.

Too many theories.

However, there is a paradox here. While many student teachers appear to reject theory, at the same time many ofthem expect from a course ‘‘instant panaceas, rigid rules ofthumb, clear statements ofpractice, and absolute generaliza- tions’’ (Brumfit, 1983, p. 60) or ‘‘definitions, rules, absolutes’’(Brown, 1983, p. 54). Where else can this come but from theory? Krashen states: ‘‘Given a briefworkshop or inservice, the most practical, most valuable information we can provide is a coherent view ofhow language is acquired, a theory ofSLA’’ (1983, p. 281). Perhaps it is not that we are giving our student teacherstoo much theory; but rather that we are not addressing the rightissues (Wright, 1992), or exploring them in therightway.

1.2. Dichotomies in teacher thinking It has been suggested that the social and textual practices ofteachers and researchers actuallyconstitute two different forms of discourseFthat the discourse ofeducational research is either alien to (Wright, 1992, p. 188) or generically different from that of the classroom teacher (Clarke, 1994; Ellis, 1997; Kerlinger, 1977). An important goal, then, ofteacher education should be to establish links between professional discourse and local discourse, both at the level oflanguage and practice (Freeman & Richards, 1993). In order to explore this, we will here consider what constitutes the less clearly defined ‘local’ (Geertz, 1983) or ‘personal’ knowledge (Clandinin & Connelly, 1987; Polanyi, 1958; Winch, 1958) ofclassroom practice which our student teachers contrast with educational research.

Considerable research has been carried out in mapping the cognitive and interpretative frame- works which teachers bring to their professional activities (Freeman, 1994, 1996; Freeman & Richards, 1993; Johnson, 1996; Richards & Nunan, 1990; Woods, 1996). Conventionally, these de- scriptions differentiate areas ofteachers’ cognitive and interpretative frames. Abelson, working in the field ofcognitive science, suggests seven features F‘‘nonconsensuality, existence beliefs, alternative worlds, evaluative components, episodic materials, unboundedness, variable credences’’ (Abelson, 1979, p. 360)Fthat distinguish a beliefsystem from a knowledge system. Within the knowledge system oflanguage education, Johnson (1996) distinguishes between conceptual knowledge [epis- temeFor ‘abstract wisdom’] and perceptual knowledge [phronesisFor ‘practical wisdom’].

She argues for the focus in teacher education to be on perceptual knowledge because the vicissi- tudes ofthe classroom often militate against the application ofa general rule. This distinction between theoretical and practical knowledge is more conventionally framed in the distinction between declarative and procedural knowledge (Woods, 1996). Declarative knowledge is knowl- edgeaboutteachingFknowledge ofsubject areas and the ‘theory’ ofeducation; procedural knowl- edge is knowledge ofhowtoteachFknowledge of instructional routines to be used in the classroom.

Lightbown picks this up when she distinguishes (1985) between ‘teacher education’ (i.e., concep- tual/declarative knowledge) and ‘teacher training’ M. MacDonald et al. / Teaching and Teacher Education 17 (2001) 949–963 950 (i.e., perceptual/procedural knowledge). Richards and Nunan (1990) also distinguish between teacher education and teacher training. They define teacher education as ‘‘characterized by approaches that involve teachers in developing theories of teaching, understanding the nature ofteacher decision making, and strategies for critical self- awareness and self-evaluation’’ and teacher train- ing as ‘‘characterized by approaches that view teacher preparation as familiarizing student tea- chers with techniques and skills to apply in the classroom’’ (in Ellis, 1994, p. 187). Taking a slightly different perspective, Richards relates what is essentially perceptual/procedural knowledge (knowledge relating to curriculum, subject matter and lesson presentation) back to the realm of beliefsFknowledge ‘‘which relates to the teacher’s personal and subjective philosophy ofteaching and the teacher’s view ofwhat constitutes good teaching’’ (Richards, 1996, p. 284). He describes a number of‘maxims’ which constitute not so much the ‘rules’ ofspecific teaching techniques, but rather the set of‘‘beliefs, principles and values’’ (p. 294) underlying more generalized classroom practice.

From the research above, it is Lightbown who is specifically concerned with the contribution of SLA to teacher training. Unsurprisingly, she places SLA research in the realm ofconceptual/ declarative knowledge, but suggestsFrather ‘‘humbly’’ (Ellis, 1994, p. 175)Fthat it relates to perceptual/procedural knowledge, inasmuch as it enables teachers to ‘‘have much more realistic expectations about what can be accomplished’’ (1985, p. 183–4). In this way, it is suggested that there are connections between these different areas ofteacher cognition. Lightbown is suggesting anFalbeit weakFrelationship between concep- tual/declarative knowledge and perceptual/proce- dural knowledge; and Richards is suggesting a relationship between perceptual/procedural knowledge and teacher beliefs, principles and attitudes. What has not yet been explored is the notion that a relationship might exist between conceptual/declarative knowledge, in which SLA is conventionally seen as playing a role, and the beliefs, principles and attitudes of teachers and student teachers. The central question ofthispaper, then, is to investigate whether the provision ofa course in SLA within the context ofa TESOL program might have an effect on the beliefs, principles and attitudes ofstudent teachers.

1.3. An integrated approach In our discussion so far, we have examined approaches which divide aspects ofteacher cogni- tion into separate categories. A more recent strand ofresearch, however, challenges the categorical distinctions outlined above. Woods (1996) sug- gests that these dichotomies do not accurately reflect the relationship between the beliefs, as- sumptions and knowledge ofteachers and their practices in the classroom.

In order to take appropriate action, people need to understand; and to understand they need knowledge about the world and specifically about the situation they are in (Woods, 1996, p. 59).

Richards and Lockhart (1994) and Johnson (in Richards & Lockhart, 1994) also emphasize the interrelatedness between beliefs and knowledge, and declarative and procedural knowledge:

yESL teachers teach in accordance with their theoretical beliefs and differences in theoretical beliefs may result in differences in the nature ofyinstruction (Johnson in Richards & Lock- hart, 1994, p. 37).

ywhat teachers do is a reflection ofwhat they know and believe, and teacher knowledge and ‘teacher thinking’ provide the underlying frame- work or schema which guides the teacher’s classroom actions (Richards & Lockhart, 1994, p. 29) Woods (1996) goes on to develop a multi- dimensional cycle ofplanning and decision making within teaching. He describes three phases of assessment, planning and implementation which operate recursively to inform different hierarchical levels ofthe teaching processFgoing from the most local level ofdiscrete events in the lesson plan to the most global level ofwhole course planning (p. 139). This is to say, a decision that a teacher M. MacDonald et al. / Teaching and Teacher Education 17 (2001) 949–963951 takes about designing the curriculum can be informed by his/her experience of teaching a particular component ofa lesson the previous semester; and a planning decision that a teacher takes in relation to the curriculum can in turn inform the future staging of that particular lesson.

Woods’s analysis ofinterview data suggests that knowledge structures and beliefsystems ‘‘are not composed ofindependent elements, but [are] rather structured, with certain aspects implying or pre- supposing others’’ (p. 200). Woods proposes a model to signify the evolving system of beliefs, assumptions and knowledge (BAK) that recursively informs/is informed by the context of teaching:

ythe BAK was part ofthe perceiving and organizing ofthe decisions. When a decision was considered, it was considered in the context ofBAK, and when it was remembered later it was also remembered in the context ofBAK (Woods, 1996, p. 247).

Woods’s research suggests, therefore, that class- room practice is not distinguished by a lack of theory, as implied by the commonplace polariza- tion with the research into teaching and learning reviewed above. In fact, Woods has demonstrated that language teachers create and maintain back- ground networks ofbeliefs, assumptions and knowledge which, we would argue, constitute a valid theory ofteaching and learning. These background theoretical networks are grounded in every level ofroutine classroom practice in much the same way that educational theory is grounded in the systematic collection ofempirical data. It would appear, therefore, that what distinguishes the discourse ofclassroom practice from the discourse ofeducational research is not an absence oftheory, but the context in which the theory is constructed and the form in which it is articulated.

1.4. SLA theory and classroom practice As part ofthe discourse ofeducational research, SLA theory has been said to have either a direct or indirect effect on the instructional routines and procedures oflanguage teaching. Early research was sceptical. Tarone et al. suggested that ‘‘hastypedagogical applications should not be made on the basis’’ ofthe current state ofSLA research (1976, p. 29). One example ofthis was Krashen’s attempt to devise a grammatical syllabus based on the Natural Order (1983). 1Lightbown (1985) also suggests that SLA research should play a role in teacher education rather than in teacher training.

However, more recently, a strong claim has been made for the role SLA research has played in one ofthe more recent techniques oflanguage peda- gogy, task-based learning and teaching (Long & Crookes, in Freeman & Richards, 1993). Over the past 20 year SLA research (e.g. Long, 1981) has offered powerful evidence that language intake is facilitated when language learners are engaged in the negotiation ofmeaning, as when attempting to find an outcome ofa problem-solving task.

This has currently led to the widespread design and implementation oftasks in the lan- guage classroom as an approach to language teaching.

However, in the light ofthe recent research into teacher cognition described above (Richards & Lockhart, 1994; Woods, 1996), perhaps a stronger claim can be made for theindirecttransfer of SLA research into classroom practice. Allen suggested a long time ago that:

there is, perhaps, something wrong with the idea that the only way to ‘apply’ the results of research is to write a whole new textbook or a brand new curriculum sequence. Perhaps it is better to see the current applications ofresearch as comprising an influence which indirectly and subtly changes the teacher’s attitude towards what s/he is trying to do in the classroomy (Allen in Tarone et al., 1976, p. 30).

This suggests that SLA theory and research could be better used to inform the reflexive frameworks which teachers mobilize in the 1As part ofhis Monitor Model, the popular language educationalist, Stephen Krashen (1985) drew on a body of research published throughout the 1970s to support the thesis that learners ofEnglish as a foreign/second language acquired its morphemes in a particular order. This could have implica- tions for the order in which grammatical items were introduced in class. However, the evidence for this ‘‘natural order’’ is still hotly disputed (McLaughlin, 1987). M. MacDonald et al. / Teaching and Teacher Education 17 (2001) 949–963 952 classroom and to enable teachers to refine their interpretative frames so that they can select from the plethora ofpossible teaching approaches available to them.

When we provide theory, we give them (tea- chers) the underlying rationale for methodology in general. This permits adaptation for different situations, evaluations ofnew techniques and evaluations oftheory. Without theory, there is no way to distinguish effective teaching proce- dures from ritual, no way to determine which aspects ofa method are helpful and which are not helpful (Krashen, 1983, p. 261).

To date, there has been surprisingly little published on the relationship between SLA and teacher thinking and how best to exploit SLA in order to enhance it (Ellis, 1997). However, the comments ofour students and ofthose practi- tioners critical ofthe role ofSLA research suggest that some justification is required for the inclusion ofan SLA course in a program preparing people for a career in language education. One approach to this is to examine the relationship between theoretical knowledge and teacher behavior in the classroom (e.g., Freeman & Richards, 1996). The difficulty with this is that the changes in teacher behavior may be a result either oftheir experience ofthe classroom, the theoretical input they received during their initial preparation, or some combination ofthese. Also the way teachers behave in classrooms may be affected by the presence ofobservers or other methods of observation. We would argue that this type of research in the classroom could usefully be augmented by an investigation into the relation- ship between the theoretical knowledge provided on a teacher education program in TESOL and the changes that take place in the beliefs, assump- tions and knowledge ofthe student teachers on it. This paper goes on to describe a small-scale piece ofresearch which was carried out to investigate whether there was any change in the beliefs, assumptions and knowledge of student teachers undergoing undergraduate and postgraduate programs in TESOL in a UK university.2. Procedure The authors teach on B.A. and M.Sc. programs in TESOL at the Center for English Language Teaching, now part ofthe University ofStirling Institute ofEducation in Scotland, UK. Both undergraduate and postgraduate programs include a one-semester course in SLA. The undergraduate SLA course is part ofa three year B.A. degree in English Language Teaching offered specifically for non-native speakers ofEnglish. Students normally take this course in their second year along with courses in Discourse Analysis and Education. A course in micro-teaching and classroom observa- tion is held in the second semester. The post- graduate SLA course is taken in the first semester ofa one year M.Sc. program in TESOL for both native and non-native speakers. In the first semester, postgraduates also take courses in language description (an introduction to func- tional grammar), TESOL methodology (teaching ofskills), and classroom observation. A postgrad- uate micro-teaching course is held in the second semester ofthe program.

Both SLA courses are taught over twelve weeks with three contact hours per week. While the courses are taught separately, by different mem- bers ofstaff, and make use ofdifferent textbooks, there is inevitably considerable overlap in content between the two courses. The style ofteaching in both courses is relatively informal and is discus- sion rather than lecture based, although the postgraduate teaching can be slightly more didac- tic at times. Both courses are assessed by two essays and one three hour examination, although the postgraduates write slightly longer assign- ments. A full description ofthe content ofthe two courses is given in Appendix A.

The students on the two SLA courses described above were the main focus of attention in this paper. Typical undergraduate student numbers are between 15 and 20. The postgraduate group is normally between 20 and 25, mostly non-native speakers with four or five native speakers in each year group. The male : female ratio of both groups is approximately 1 : 4. The largest number of students by far came from Greece, since they made up roughly two-thirds ofthe undergraduate M. MacDonald et al. / Teaching and Teacher Education 17 (2001) 949–963953 group. Details ofthe numbers, gender and nationalities ofthe students are given in Table 1.

Ofthe two groups, only a few ofthe postgraduates had done any previous courses in language learning and teaching or related topics or had previous teaching experience. However, the non- native speakers had all studied English language to a high level and so could draw on considerable experience ofthe language classroom.

The Center for English Language Teaching also runs an undergraduate degree course in English as a Foreign Language for non-native speakers, which does not include components in teacher education; and the Institute ofEducation as a whole also runs undergraduate courses in Initial Teacher Education without any specific focus on foreign language teaching. Since these two groups did not receive any course specific to language teaching pedagogy or SLA, it was decided to use them as controls.

Over two successive years, we administered a questionnaire on language learning to 55 subjects (28 post-graduates and 27 undergraduates) at the beginning and end ofthe semester in which the SLA courses were taught. The same questionnaire was also administered at the beginning and end of the semester to the control group totaling 25 undergraduates drawn from a B.A. course in English as a Foreign Language and an Initial Teacher Education program which did not feature a course in SLA. The questionnaire (based on Lightbown & Spada, 1995, p. xv) contained 12 statements containing key beliefs relating to English language learning. These are included in the order in which they were given to the studentsin Appendix B. Since the statements reflected issues which are central to SLA, the questionnaire was given out in the SLA class. Subjects had to indicate the extent to which they agreed or disagreed on a six point Likert scale where 6 meant strongly disagree and 1 meant strongly agree. The mean responses are shown on the tables that follow. At the end of the semester the subjects again completed the questionnaire and the results were analyzed using a Wilcoxon Signed Ranks Test Asymp. Sig. (2-tailed) (Kinnear & Gray, 1997).

3. Results At the outset ofthe semester there was no significant difference (Table 2) between the key beliefs about language learning (Lightbown & Spada, 1995, p. xv) indicated by the combined cohort ofundergraduates and postgraduates (n¼55) taking the SLA course and those noted by the control group (n¼27).

Broadly speaking, the views ofthe combined cohort taking the SLA course differed significantly before and after the course on issues relating to learning and language, although there was no significant difference on issues relating to learner differences (Table 3). In particular there appeared to be a movement away from the behaviorist views oflearning which the subjects had previously held (statements 1, 2, 6, 9, 12) and from the use of language input which is graded on a strict grammatical basis (statements 7, 8 and 10). The students also significantly lessened their convic- tions that language learning should be introduced early in a school program (statement 5). Although they still basically agreed with the notion, this would suggest that they had also been influenced by approaches that credit the cognitive resources that older learners bring to bear on language learning. However, students were not increasingly persuaded by evidence from SLA research pre- sented on both courses as to the benefits of interaction between non-native speakers on the accuracy oflearners’ utterances (statement 11).

There was also no significant change in their beliefs regarding the relationship between Table 1 Respondents on TESOL training courses 1997–8 UG PG Combined Male 8 7 15 Female 20 20 40 Native Speakers 0 9 9 Greece 17 7 24 Taiwan/HK 4 5 9 Japan 2 3 5 Other 5 3 8 Total 28 27 55M. MacDonald et al. / Teaching and Teacher Education 17 (2001) 949–963 954 Table 2 Comparison ofthe language learning beliefs ofcombined SLA (n¼55) and non-SLA (n¼27) cohorts before an SLA course a The higher the number the more the disagreement Non-SLA SLAP= 1 Languages are learned mainly through imitation 3.04 2.98 0.701 b 2 Parents usually correct young children when they make grammatical errors 3.36 3.02 0.335 b 3 People with high IQs are good language learners 3.68 3.56 0.702 b 4 The most important factor in second language acquisition success is motivation1.96 2.31 0.182 b 5 The earlier a second language is introduced in school programs, the greater the likelihood ofsuccess in learning1.92 1.85 0.664 b 6 Most ofthe mistakes which second language learners make are due to interference from their first language2.76 2.20 0.021 b 7 Teachers should present grammatical rules one at a time, and learners should practice examples ofeach one before going on to another2.44 2.13 0.418 b 8 Teachers should teach simple language structures before complex ones 1.92 1.45 0.088 b 9 Learners’ errors should be corrected as soon as they are made in order to prevent the formation of bad habits2.16 2.11 0.624 b 10 Teachers should use materials that expose students only to those language structures which they have already been taught3.24 3.42 0.605 b 11 When learners are allowed to interact freely (for example in group or pair activities), they learn each others’ mistakes3.08 3.09 0.996 b 12 Students learn what they are taught 3.48 3.36 0.738 b a Adapted from Lightbown and Spada (1995, p. xv).bNo significant differenceP>0:05: Table 3 Language learning beliefs of combined UG/PG cohort (n¼55) before and after taking an SLA course a The higher the number the more the disagreement Pre PostP= 1 Languages are learned mainly through imitation. 2.98 4.15 0.000 d 2 Parents usually correct young children when they make grammatical errors. 3.02 3.75 0.008 d 3 People with high IQs are good language learners. 3.56 3.49 0.664 b 4 The most important factor in second language acquisition success is motivation. 2.31 2.47 0.483 b 5 The earlier a second language is introduced in school programs, the greater the likelihood ofsuccess in learning.1.85 2.27 0.007 d 6 Most ofthe mistakes which second language learners make are due to interference from their first language.2.20 3.29 0.000 d 7 Teachers should present grammatical rules one at a time, and learners should practice examples ofeach one before going on to another.2.13 2.96 0.000 d 8 Teachers should teach simple language structures before complex ones. 1.45 1.89 0.002 d 9 Learners’ errors should be corrected as soon as they are made in order to prevent the formation of bad habits.2.11 3.29 0.000 d 10 Teachers should use materials that expose students only to those language structures which they have already been taught.3.42 4.29 0.000 d 11 When learners are allowed to interact freely (for example in group of pair activities), they learn each others’ mistakes.3.09 3.38 0.185 b 12 Students learn what they are taught. 3.36 4.00 0.005 c a Adapted from Lightbown and Spada (1995, p. xv).bNo significant differenceP>0:05:cA significant differencePo0:05:dA very significant differencePo0:01: M. MacDonald et al. / Teaching and Teacher Education 17 (2001) 949–963955 intelligence (statement 3) or motivation (statement 4) and language learning.

The beliefs of the control group who had not received a course in SLA did not differ significantly at the end ofthe semester from those they had held at the beginning ofthe semester (Table 4).

3.1. Undergraduates and postgraduates combined Ifwe take strong agreement to mean a rating of under two, before the course started the combined undergraduate and postgraduate SLA groups strongly agreed with just two statements (Table 3):

5. The earlier a second language is introduced in school programs, the greater the likelihood of success in learning.

8. Teachers should teach simple language structures before complex ones.

These two notions do not quite add up to a consistent view oflanguage teaching. Statement five suggests a mentalist view oflanguage acquisi- tion as biologically dependent while statementeight is consistent with both behaviorist and cognitivist views. Ifwe take strong disagreement to mean a rating ofover four then the subjects did not strongly disagree with any statements before the course.

Again taking a rating ofunder two as indicating strong agreement, after the course the subjects strongly agreed with one statement (which they had strongly agreed with before the course):

8. Teachers should teach simple language structures before complex ones.

However, the subjects now strongly disagreed with three statements:

1. Languages are learned mainly through imitation.

10. Teachers should use materials that expose students only to those language structures which they have already been taught.

12. Students learn what they are taught.

The strong rejection ofthese three statements suggests that after the course, the subjects had at least taken on board one ofthe few certainties Table 4 Language learning beliefs of non-SLA cohort (n¼25) at the beginning and end ofsemester a The higher the number the more the disagreement Pre PostP= 1 Languages are learned mainly through imitation. 3.04 3.00 0.928 b 2 Parents usually correct young children when they make grammatical errors. 3.36 3.20 0.559 b 3 People with high IQs are good language learners. 3.68 3.64 0.788 b 4 The most important factor in second language acquisition success is motivation. 1.96 2.08 0.499 b 5 The earlier a second language is introduced in school programs, the greater the likelihood ofsuccess in learning.1.92 1.76 0.360 b 6 Most ofthe mistakes which second language learners make are due to interference from their first language.2.76 2.76 0.942 b 7 Teachers should present grammatical rules one at a time, and learners should practice examples ofeach one before going on to another.2.44 2.68 0.343 b 8 Teachers should teach simple language structures before complex ones. 1.92 1.76 0.617 b 9 Learners’ errors should be corrected as soon as they are made in order to prevent the formation of bad habits.2.16 2.20 0.772 b 10 Teachers should use materials that expose students only to those language structures which they have already been taught.3.24 3.68 0.173 b 11 When learners are allowed to interact freely (for example in group of pair activities), they learn each others’ mistakes.3.08 3.00 0.877 b 12 Students learn what they are taught. 3.48 3.56 0.929 b a Adapted from Lightbown and Spada (1995, p. xv).bNo significant differenceP>0:05: M. MacDonald et al. / Teaching and Teacher Education 17 (2001) 949–963 956 afforded by SLA research: a rejection ofthe behaviorist model oflearning. It implies that they now accord more importance to the role of the learner in language learning, which is consistent with the view that learners create their own syllabuses out ofthe language input they receive.

3.2. Undergraduates and postgraduates compared There were only a few areas of difference between the responses ofthe undergraduate and postgraduate cohorts. Tables 5 and 6 compare differences between the undergraduate and post- graduate responses before and after the SLA course. We shall group these into four categories ofstatement.

Statements which reflect a broadly behaviorist viewof language learning(S1, S2, S6, S9, S12).

There was a significant difference between the undergraduate and postgraduate cohorts on only one statement, relating to the immediate correc- tion oferrors (S9), before the course (Table 5); butthere was no significant difference between the two groups in this area after the course (Table 6).

Statements relatingto thegrammatical sequen- cingof language teaching(S7, S8, S10). There was a significant difference between the undergraduate and postgraduate beliefs relating to the idea of staged presentation and practice ofgrammatical rules (S7), a notion which also has some beha- viorist overtones, both before the course (Table 5) and after the course (Table 6).

Statements relatingto learnervariations(S3, S4, S5). There was a significant difference between the undergraduate and postgraduate beliefs relating to the relationship between IQ and language learning (S3) both before the course (Table 5) and after the course (Table 6). Thus it would appear that the SLA course did not minimize the difference between undergraduate and postgraduate beliefs relating to this area.

Statement relatingto learner-learner interaction (S11). There was no significant difference between either the degree or the range ofundergraduate and postgraduate beliefs before and after the courses.

Table 5 Comparison ofthe language learning beliefs ofUG (n¼28) and PG (n¼27) cohorts before an SLA course a The higher the number the more the disagreement UG PGP= 1 Languages are learned mainly through imitation. 3.21 2.74 0.218 b 2 Parents usually correct young children when they make grammatical errors. 2.89 3.15 0.496 b 3 People with high IQs are good language learners. 3.96 3.15 0.048 c 4 The most important factor in second language acquisition success is motivation. 2.57 2.04 0.106 b 5 The earlier a second language is introduced in school programs, the greater the likelihood ofsuccess in learning.1.93 1.78 0.856 b 6 Most ofthe mistakes which second language learners make are due to interference from their first language.2.04 2.37 0.224 b 7 Teachers should present grammatical rules one at a time, and learners should practice examples ofeach one before going on to another.1.79 2.48 0.010 c 8 Teachers should teach simple language structures before complex ones. 1.36 1.56 0.307 b 9 Learners’ errors should be corrected as soon as they are made in order to prevent the formation of bad habits.1.61 2.63 0.007 c 10 Teachers should use materials that expose students only to those language structures which they have already been taught.3.43 3.41 0.952 b 11 When learners are allowed to interact freely (for example in group of pair activities), they learn each others’ mistakes.2.96 3.22 0.526 b 12 Students learn what they are taught. 3.32 3.41 0.810 b a Adapted from Lightbown and Spada (1995, p. xv).bNo significant differenceP>0:05:cA significant differencePo0:05: M. MacDonald et al. / Teaching and Teacher Education 17 (2001) 949–963957 4. Discussion The relatively small numbers ofstudents in- volved in the study limit the generalisability ofthe results. Furthermore, the results may have been skewed by the disproportionate ratio offemales to males and the particular configuration ofdifferent nationalities which we had in our cohort, for example, the preponderance ofGreek learners.

4.1. Changes in student beliefs, assumptions and knowledge We would suggest that, within the context ofthe other courses on our B.A. and M. Sc. programs, our course on SLA research and theoryFas one area ofconceptual/declarative knowledgeFdid have an impact on some ofthe beliefs, assump- tions and knowledge ofour student teachers.

While the control group, who did not take an SLA course, did not register any significant changes in their attitudes towards language learn- ing, our students, who did take the course, didregister significant changes in their attitudes towards certain issues in language learning.

Thus it would seem that, despite the explicit aversion expressed by our students towards the theoretical approach ofthe course, the course did have some measurable effect. Our students ap- peared to have started out with common sense beliefs about language learning which were per- haps closest to a behaviorist model oflanguage learning. It is possible that these reflected didactic classroom practices that they had experienced during their periods oflanguage learning in their home cultures. This could be as true of some ofour more mature British postgraduates as those from Greece and East Asian countries.

The kind ofbeliefs that our students seemed to be moving towards would seem to fit in either with a broadly Krashenite view (Krashen, 1983; Krashen & Terrell, 1995), which sees language learning as a largely unconscious process, or with a broadly cognitive perspective, which emphasizes the potential for conscious language learning.

Table 6 Comparison ofthe language learning beliefs ofUG (n¼28) and PG (n¼27) cohorts after an SLA course a The higher the number the more the disagreement U/G P/GP= 1 Languages are learned mainly through imitation. 4.39 3.89 0.123 b 2 Parents usually correct young children when they make grammatical errors. 3.50 4.00 0.129 b 3 People with high IQs are good language learners. 3.89 3.07 0.013 c 4 The most important factor in second language acquisition success is motivation. 2.50 2.44 0.943 b 5 The earlier a second language is introduced in school programs, the greater the likelihood ofsuccess in learning.2.50 2.04 0.142 b 6 Most ofthe mistakes which second language learners make are due to interference from their first language.3.32 3.26 0.766 b 7 Teachers should present grammatical rules one at a time, and learners should practice examples ofeach one before going on to another.2.61 3.33 0.048 c 8 Teachers should teach simple language structures before complex ones. 1.71 2.07 0.394 b 9 Learners’ errors should be corrected as soon as they are made in order to prevent the formation of bad habits.3.21 3.37 0.617 b 10 Teachers should use materials that expose students only to those language structures which they have already been taught.4.43 4.15 0.319 b 11 When learners are allowed to interact freely (for example in group of pair activities), they learn each others’ mistakes.3.46 3.30 0.815 b 12 Students learn what they are taught. 4.04 3.96 0.816 b a Adapted from Lightbown and Spada (1995, p. xv).bNo significant differenceP>0:05:cA significant differencePo0:05: M. MacDonald et al. / Teaching and Teacher Education 17 (2001) 949–963 958 4.1.1. Cohort overall Differences in significance levels ofthe changes in the beliefs of the cohort taken as a whole might reflect differences in the degrees ofcertainty with which different areas ofknowledge within the emerging paradigm ofSLA research are viewed.

Research such as Chomsky’s (1959) critique of behaviorism and Dulay and Burt’s (1972) account oflanguage learning processes, which militate strongly against behaviorist notions oflanguage learning and downplay the influence ofthe first language on the second language, have by now- Fdespite the latter’s limited database (Kachru in Liu, 1998)Fbecome an unproblematic part ofthe SLA canon and a central plank ofmany founda- tion courses at undergraduate and Master’s level, including our own. It may also be the case that the debate over behaviorismFhowever wellworn in Anglo-American pedagogyFrelates most strongly to many ofour own students’ re-evaluations of their experience ofteaching and learning lan- guages. However, the SLA research into learner characteristics, such as the relationship between language learning and motivation (Gardner, 1985) and intelligence (Genessee, 1976) has tended to be less conclusive and has also been carried out relatively more recently. This may explain the fact that the SLA cohort as a whole did not register significant degrees ofchange in their beliefs regarding these areas. Our combined cohort also remained largely unpersuaded about the positive impact ofinteraction between non-native speakers on language learning despite aspects ofthis being supported by extensive current research into Task- Based Learning (Long & Crookes, 1992). This is particularly ironic, since it is precisely in this area ofSLA research that the strength ofthe interface between theory and practice has been noted (Markee, 1997). One can only speculate that our students’ lack ofconviction might have been affected by two factors: again, the comparative recency ofthe research in this area (Markee, 1997); and the fact that cultural influences were still proving more powerful for them than empirical research. It is perhaps unsurprising that a one semester course did not entirely persuade our students ofan alternative view oflanguage learning, sinceFfor perfectly valid reasonsFmany ofthem were probably only still emerging from a lengthy period of being positioned as passive learners.

4.1.2. Undergraduatesvs. postgraduates There are two areas ofinterest in which the undergraduate and postgraduate cohorts differed both before and after the course: error correction and the relationship between language teaching and language learning.

First, it is paradoxical that the undergraduate cohort agreed more than the postgraduate cohort with immediate classroom error correction at the beginning ofthe course but their degree ofchange was greater by the end ofthe course. Many ofthe undergraduates and postgraduates had come from teaching cultures where there is a low tolerance of error in the classroom. However, our under- graduates had been studying in an ethos of communicative language teaching for one year prior to the study, whereas most ofour post- graduates had come directly to the program from their home countries. This may account for the fact that the undergraduate beliefs changed more by end ofthe course; but it does not account for the fact that they agreed more strongly with the idea ofimmediate error correction at the beginning ofthe course. Furthermore, the converse is the case with regard to parental error correction, where the postgraduates became much more assured ofa non-behaviorist position by the end ofthe course.

Secondly, the undergraduate cohort agreed significantly more than the postgraduates with the idea ofstaged presentation and practice of grammatical rules both before and after the course. In this respect, our SLA course would seem to have had little impression on them.

Moreover, this suggests that the undergraduates viewed language learning as more teacher-directed than the postgraduate cohort. This may be due to the fact that the undergraduate program was by its very nature more teacher-directed than the post- graduate program. It is possible that the under- graduates were responding here not so much to what they were taught as to the way in which they were taught. M. MacDonald et al. / Teaching and Teacher Education 17 (2001) 949–963959 4.2. Theory into practice It would hardly be adequate ifall that happened in a TESOL program was a change in beliefs and assumptions. Central to Woods’s (1996) concep- tualization ofbeliefs, assumptions, and knowledge (BAK) is the idea that changes in teachers’ belief systems can also lead to changes in their percep- tual/procedural systems, which in turn produces outcomes in terms ofclassroom practice. Although it clearly is a limitation ofour study that we have no data from classroom practice relating to the changes in beliefs and attitudes that appeared to take place, we would argue on the basis ofWoods’s (1996) research that SLA theory actuallydoesplay a part in the development ofthe everyday class- room routines ofprospective teachers.

However, from their comments, both orally and on their course evaluation forms, it would appear that our students were either unaware of, or undervalued, the changes that were taking place.

Three factors might contribute towards this. We have already commented upon the differences between the discourse ofeducational research and classroom practice. Markee (1997) notes how inefficient SLA research is in particular when it comes to communicating with its clients, teachers. Secondly, our SLA course is not unusual in being clearly distinguished from other courses on classroom pedagogy on the B.A. and M.Sc.

programs. As is the case elsewhere, our students are also products ofan educational system which sustains the classification ofknowledge into clearly bounded units (Bernstein, 1975). Both the framing ofthe teacher education curriculum and the corresponding socialization ofstudents militate against the ultimate re-integration oftheories of language learning with their pedagogic application by their users. Finally, given that the teacher education curriculum is conventionally classified into research and theory courses on the one hand, and practical teaching courses on the other, it will tend to be the theory that becomes devalued in the current pragmatic intellectual climate where there is a ‘‘desire for immediate application of research to general education’’ (Kerlinger in Lightbown, 1985, p. 180).

We would like to conclude by pointing up some possible ways forward for the reframing of SLAresearch on teacher education programs so that student teachers might become more aware ofthe changes that are taking place in their beliefs, assumptions and knowledge and recognize their inseparability from pedagogic practice. Under- lying this is the notion that the way SLA is taught, and in particular its contextualisation within a teacher education program, is at least as important as the content ofthe course. With regard to the exclusivity ofeducational research, Freeman (1996) argues that there is a need for a change in the genre ofresearch to enable the teacher’s voice to be heard. He goes on to say (1996, p. 10) that narrative accounts ofclassroom experience which reveal the identity ofthe teller should be accorded the same validity as more objective research reports. The use ofnarratives could enable student teachers to identify more powerfully with their reading and might serve to compensate for some of the deficiencies in communication apparent in the conventional SLA literature (Markee, 1997).

Student teachers could also be encouraged within the seminar to generate both oral and written narratives oftheir own experiences as language teachers or learners. Ifthe voice ofthe lan- guage teacher is only partially heard within the discourse oflanguage education, it is surely the voice ofthe language learner that is truly absent. In this way, student teachers might ‘‘develop their own systematic ways ofcommunicating their own experi- entially derived understandings ofwhat will chal- lenge our preconceptions, suggest falsifications of some ofour hypotheses, and enable teaching itselfto develop more openly’’ (Brumfit, 1983, p. 71).

Teacher education could also establish stronger links between the divergent discourses ofeduca- tional research and classroom practice by educat- ing prospective teachers in classroom-based forms ofresearch and embedding teacher education in the classroom. Rather than students being posi- tioned as the passive recipients ofresearch and theory, they could be positioned as potential users.

This would also help prepare them for more active future engagement in research as teachers (Sten- house, 1975; Markee, 1997; Wright, 1992). This research could be contextualised within case-based methods (Richert & Shulman in Johnson, 1996) and portfolio assessment (Johnson, 1996). M. MacDonald et al. / Teaching and Teacher Education 17 (2001) 949–963 960 The compartmentalization ofeducational knowledge within the curriculum ofteacher education programs in TESOL could also be addressed, although this would be a major under- taking for any program. Rather than having discrete specialisms running ‘horizontally’ in par- allel through the program, it might be possible to experiment with the idea ofexploring ‘vertical’ topics in an integrated way. One unit would be devoted to ‘errors’, another to ‘strategies’ and so on. Each unit would provide a focus through which aspects ofclassroom practice could be explored from the most theoretical to the most practical level. For example, a module on errors would address psycholinguistic reasons for errors (often dealt with separately on an SLA course), linguistic descriptions oferrors (often dealt with separately in a language description course) and techniques ofhandling errors (often dealt with separately on a methodology course).

5. Conclusion In this paper we have observed that, despite student teachers’ avowed aversion to SLA theory and research, certain changes in key beliefs in their own attitudes and beliefs towards language learn- ing did take place during programs which included an SLA course. Taking on board Woods’s (1996) argument that teachers’ beliefs and attitudes are inextricably linked with classroom practice, we would argue that the theory and research compo- nent ofthese programs does have an effect on the evolving classroom routines ofour students.

However, it has to be conceded that neither the discourse ofeducational research in which SLA is grounded nor the way in which SLA courses are framed within the curriculum of teacher education programs in TESOL make it any easier for the prospective teacher to come to terms with this often arcane body of knowledge. However, if some ofthe changes which have been suggested were to be implementedFofnecessity over some consider- able period oftimeFwe might see SLA research and theory begin to become ‘‘more explicitly grounded in the real world’’ (Markee, 1997, p.

88), and in so doing, be placed ‘‘back into itsoriginating context’’ (Buchmann, 1984, p. 434).

Further research needs to be carried out to correlate changes in the beliefs, assumptions and knowledge ofstudent teachers in TESOL with observable data ofchanges in their classroom routines in the context ofmicro-teaching or teaching practice.

Acknowledgements We would like to thank our B.A. ELT and M.Sc. TESOL cohorts (1997 and 1998) for their good humor and persistence; also Patsy Light- bown for permission to use the questionnaire.

Appendix A. Content ofundergraduate and post- graduate SLA courses Undergraduate course Learning a First Language (behaviorism, ment- alism, interactionism) Learner Language (contrastive analysis, error analysis, natural order) Social Aspects ofInterlanguage (acculturation and social identity) Discourse Aspects ofInterlanguage (input, out- put and interaction) Psycholinguistic Aspects ofInterlanguage (transfer, consciousness, communication strate- gies) Linguistic Aspects ofInterlanguage (Chomsky, TG and CPH) Individual Differences in L2 Acquisition (apti- tude, motivation, the affective filter and learning strategies) Interaction in Language Learning (groupwork and feedback) Instruction and L2 Acquisition (form focused instruction and the natural approach) Postgraduate course First Language Acquisition (behaviorism, ment- alism, interactionism) M. MacDonald et al. / Teaching and Teacher Education 17 (2001) 949–963961 Error Analysis and Language Transfer (psycho- linguistic causes oferrors; influence ofL1 on L2) Interlanguage ( idiosyncratic dialects, approx- imate systems, interlanguage) Monitor Model (Krashen’s approach to lan- guage learning) Learner Differences (age, motivation, field dependence/independence) Input in Language Learning (caretaker talk, teacher talk, foreigner talk) Interaction in Language Learning (groupwork and feedback) Learner Strategies (metacognitive and cognitive strategies, strategy training) Teaching and Language Learning (form-focused instruction) Appendix B. Questionnaire statements Languages are learned mainly through imita- tion.

Parents usually correct young children when they make grammatical errors.

People with high IQs are good language learners.

The most important factor in second language acquisition success is motivation.

The earlier a second language is introduced in school programs, the greater the likelihood of success in learning.

Most ofthe mistakes which second language learners make are due to interference from their first language.

Teachers should present grammatical rules one at a time, and learners should practice examples of each one before going on to another.

Teachers should teach simple language struc- tures before complex ones.

Learners’ errors should be corrected as soon as they are made in order to prevent the formation of bad habits.

Teachers should use materials that expose students only to those language structures which they have already been taught.

When learners are allowed to interact freely (for example in group or pair activities), they learn each others’ mistakes.

Students learn what they are taught.References Abelson, R. P. (1979). Difference between beliefand knowledge systems.Cognitive Science,3, 355–366.

Bernstein, B. (Ed.) (1975). On the classification and framing of educational knowledge.Class codes and control, Vol. 3 (pp.

85–116). London: Routledge and Kegan Paul.

Brown, H. D. (1983). From ivory tower to real world: A search for relevance. In J. E. Alatis, H. H. Stern, & P. Strevens (Eds.),Georgetown university round table in languages and linguistics(pp. 53–59). Washington DC: Georgetown University Press.

Brumfit, C. (1983). The integration oftheory and practice. In J. E. Alatis, H. H. Stern, & P. Strevens (Eds.),Georgetown university round table in languages and linguistics(pp.

59–74). Washington DC: Georgetown University Press.

Buchmann, M. (1984). The use ofresearch knowledge in teacher education and teaching.American Journal of Education,92, 421–439.

Chomsky, N. (1959). Review of‘‘verbal behavior’’ by B.F.

Skinner.Language,35, 26–58.

Clandinin, D. J., & Connelly, F. M. (1987). Teachers’ personal knowledge: What counts as ‘personal’ in studies ofthe personal.Journal of Curriculum Studies,19(6), 487–500.

Clarke, M. A. (1994). The dysfunctions of the theory/practice discourse.TESOL Quarterly,28(1), 9–25.

Dulay, H., & Burt, M. (1972). Goofing, an indicator of children’s second language learning strategies.Language Learning,22, 234–252.

Ellis, R. (1994). Second language acquisition research and teacher development: The case ofteachers’ questions. In D. C. S. Li, D. Mahoney, & J. C. Richards (Eds.),Exploring second language teacher development(pp. 175–193). Hong Kong: City Polytechnic ofHong Kong.

Ellis, R. (1997). SLA and language pedagogy: An educational perspective.SSLA,19, 69–92.

Freeman, D. (1994). Knowing into doing: Teacher education and the problem oftransfer. In D. C. S. Li, D. Mahoney, & J. C. Richards (Eds.),Exploringsecond language teacher development(pp. 1–20). Hong Kong: City Polytechnic of Hong Kong.

Freeman, D. (1996). Redefining the relationship between research and what teachers know. In K. M. Bailey, & D.

Nunan (Eds.),Voices from the language classroom: Quali- tative research in second language education(pp. 88–115).

Cambridge: CUP Press.

Freeman, D., & Richards,, J. C. (1993). Conceptions of teaching and the education ofsecond language teachers.

TESOL Quarterly, 27(2), 193–216.

Freeman, D., Richards, J. C. (Eds.) (1996).Teacher learningin language teaching.Cambridge: CUP.

Gardner, R. (1985).Social psychology and second language learning: The role of attitude and motivation. London:

Edward Arnold.

Geertz, C. (Ed.) (1983). Common sense as a cultural system.

Local knowledge(pp. 73–94).London: Basic Books. M. MacDonald et al. / Teaching and Teacher Education 17 (2001) 949–963 962 Genessee, F. (1976). The role ofintelligence in second language learning.Language Learning,26, 267–280.

Johnson, K. (1996). The role oftheory in L2 teacher education.

TESOL Quarterly,30(4), 765–771.

Kerlinger, F. H. (1977). The influence ofresearch on education practice.Educational Research,6, 5–12.

Kinnear, P. R., & Gray, C. D. (1997).SPSS for windows.

London: Psychology Press.

Krashen, S. D. (1983). Second language acquisition theory and the preparation ofteachers: Towards a rationale. In J. E.

Alatis, H. H. Stern, & P. Strevens (Eds.),Georgetown university round table in languages and linguistics(pp.

255–264). Washington DC: Georgetown University Press.

Krashen, S. D. (1985).The input hypothesis. London: Longman.

Krashen, S. D., & Terrell, T. D. (1995).The natural approach.

Hemel Hempstead: Prentice Hall.

Lightbown, P. (1985). Great expectations: Second-language acquisition research and classroom teaching.Applied Linguistics,6, 173–189.

Lightbown, P., & Spada, N. (1995).Howlanguages are learned.

Oxford: OUP.

Liu, D. (1998). Ethnocentrism in TESOL: Teacher education and the neglected needs ofinternational TESOL students.

English Language TeachingJournal,52(1), 3–10.

Long, M. (1981). Input, interaction, and SLA.Annals of the NewYork Academy of Sciences,37(9), 259–278.

Long, M., & Crookes, G. (1992). Three approaches to task- based syllabus design.TESOL Quarterly,26(1), 27–56.

Markee, N. (1997). Second language acquisition research: A resource for changing teachers’ professional cultures?The Modern Language Journal,81, 80–93.McLaughlin, B. (1987).Theories of second-language learning.

London: Edward Arnold.

Nunan, D. (1991). Second language acquisition research in the language classroom. In E. Sadtono (Ed.),Language acquisition and the second/foreign language classroom (pp. 1–16). Singapore: SEAMO Regional English Language Centre.

Polanyi, M. (1958).Personal knowledge: Towards a post-critical philosophy. London: Routledge and Keegan Paul.

Richards, J. C. (1996). Teachers’ maxims in language teaching.

TESOL Quarterly,30(2), 281–296.

Richards, J. C., & Lockhart, C. (1994).Reflective teachingin second language classrooms. Cambridge: CUP.

Richards, J. C., Nunan, D. (Eds.), (1990).Second language teacher education.Cambridge: CUP.

Stenhouse, L. (1975). Research as a basis for teaching. In Authority, education and emancipation(pp. 177–195).

London: Heineman.

Tarone, E., Swain, M., & Fathman, A. (1976). Some limi- tations to the classroom applications ofcurrent second language acquisition research.TESOL Quarterly,10(1), 19–32.

Winch, P. (1958).The idea of a social science. London:

Routledge and Keegan Paul.

Woods, D. (1996).Teacher cognition in language teaching.

Cambridge: CUP.

Wright, T. (1992). L2 classroom research and L2 teacher education: Towards a collaborative approach. In J. Flower- dew, M. Brock, & S. Hsia (Eds.),Perspectives on second language acquisition(pp. 187–223). Hong Kong: City Polytechnic ofHong Kong. M. MacDonald et al. / Teaching and Teacher Education 17 (2001) 949–963963