Contextual factors , methodological principles and teacher cognition

Teachers in various contexts worldwide are sometimes unfairly criticized for not putting teaching methods developed for the well-resourced classrooms of Western countries into practice. Factors such as the teachers’ “misconceptualizations” of “imported” methods, including Communicative Language Teaching (CLT), are often blamed, though the challenges imposed by “contextual demands,” such as large class sizes, are sometimes recognised. Meanwhile, there is sometimes an assumption that in the West there is a happy congruence between policy supportive of CLT or Task-Based Language Teaching, teacher education and supervision, and curriculum design with teachers’ cognitions and their practices. Our case study of three EFL teachers at a UK adult education college is motivated by a wish to question this assumption. Findings from observational and interview data suggest the practices of two teachers were largely consistent with their methodological principles, relating to stronger and weaker forms of CLT respectively, as well as to more general educational principles, such as a concern for learners; the supportive environment seemed to help. The third teacher appeared to put “difficult” contextual factors, for example, tests, ahead of methodological principles without, however, obviously benefiting. Implications highlight the important role of teacher cognition research in challenging cultural assumptions.


Introduction
Amongst the research community, language teaching methodology is a controversial topic of discussion.Language teaching methods, such as Communicative Language Teaching (CLT), have historically been subject to fierce polarising debate, leading to them being stereotyped and demonized by critics, before being "packaged up" neatly, as if ready for the dustbin (Hunter & Smith, 2012).Attacks on CLT have sometimes focused on the native-speakerist ideology it is said to represent and the unequal power relationships that have helped it colonize the world (Pennycook, 1989), where, Bax (2003) has argued, it is used indiscriminately in diverse socio-cultural learning contexts that do not need it.And yet, despite such criticisms, sometimes supported by documented cases of classroom practitioners rejecting CLT (Kumaravadivelu, 2006), there have been vigorous ministerial attempts in recent years to implement CLT methodology and its off-shoot, Task-Based Language Teaching (TBLT), in contexts such as China, Hong Kong, India, Japan, Libya, Malaysia, the Philippines, Singapore, South Korea, Taiwan, Thailand, Turkey and Vietnam (Butler, 2011;Kırkgöz, 2008;Orafi & Borg, 2009;Wedell, 2008).The huge and ever-growing demand for "good" communication skills in English worldwide and the perception in many countries that students lack such skills, together with CLT's avowed focus on developing communicative competence (Richards, 2005), help explain the attraction.
Despite the popularity of CLT with policy-makers, though, observational studies in many of the contexts where CLT methodology is mandated have suggested limited uptake.Wedell (2008), for example, highlights that what may be seen happening in classrooms in China and Malaysia is very different from the practices anticipated by curriculum designers.Similarly, Butler (2011) focuses on the struggles teachers throughout Asia have faced trying to apply CLT and TBLT methodology in their classrooms, this often resulting in "greatly compromised" adaptations.Such teachers are often criticized, Butler continues, for providing "poorly implemented or lost-in-translation versions of the original" (p. 49).
This gap between the expectations and the reality, between the "intended curriculum and how it is enacted," has led to interest "in understanding the factors which may cause disparity between the two" (Orafi & Borg, 2009, p. 244), which in turn has fuelled research into teacher cognition, the study of "what teachers think, know and believe" (Borg, 2006, p. 1).Once teachers were viewed simply as "mechanical implementers of external prescriptions" but are increasingly seen, Borg continues, as "active, thinking decision-makers who [process and make] sense of a diverse array of information in the course of their work" (p. 7).However, while in recent years research exploring teacher cognitions and practices in relation to CLT has started to emerge (e.g., Sato & Kleinsasser, 1999;Mangubhai, Marland, Dashwood, & Son, 2004;Wyatt, 2009), evidence from such research that accesses teachers' voices and thus provides insights into how they think and act is not always drawn upon by theorists writing about what teachers ought to do.
This situation leads to various imbalances, particularly given the gaps between ELT in BANA (Britain, Australasia and North America) countries and TESEP (tertiary, secondary and primary English language education in the rest of the world); Holliday (1994, p. 3) suggests that the former have exported a "particularly narrow interpretation" of CLT to the latter.In this climate, in TESEP contexts including those in various parts of Asia, teachers' supposed misconceptions of CLT (Butler, 2011), rather than their actual beliefs, have tended to be the common focus of research, this focus in itself perhaps contributing to the "othering" practices of some BANA practitioners (Holliday, 2006), for example, of the type described by Bax (2003), when the non-native speaker "other," presented in terms of "regional or religious cultural stereotypes," is viewed prejudicially as somehow inferior, "traditional," "uncritical and unthinking" (Holliday, 2006, pp. 385-386).This all suggests a need, therefore, for more teacher cognition research worldwide, exploring the complexity of teachers' institutionally-influenced behaviour in TESEP contexts as well as their "actual" beliefs, as this research can then inform curriculum development and educational policies at large in those contexts.
However, an alternative research strategy can also be employed to address the current imbalances described above, namely to further explore teacher cognition in Western contexts in relation to CLT (our strategy in this article).One common assumption, for example, is that in BANA countries "there is considerable freedom to develop classroom methodology as a sophisticated instrument to suit the precise needs of language learners" (Holliday, 1994, p. 4).However, statements such as this need questioning since, despite all the rhetoric on the exporting of CLT methods, apart from several studies conducted in Australia (Sato & Kleinsasser, 1999;Manghubhai et al., 2004) and Canada (Farrell & Bennis, 2013), there has been relatively little exploration of teachers' cognitions and practices regarding CLT in BANA countries.Accordingly, one might ask to what extent teachers in these countries do have the freedom to teach according to methodological principles perfectly aligned to learners' needs, regardless of whether or not these principles relate to CLT.Or, rather than relating to "ideal instructional practices" (Borg, 2006, p. 279), however these teachers conceive these, are their cognitions "situated," reflecting negotiated social practices (Johnson, 2006)?Or do contextual factors intrude more significantly, influencing teachers to take practical decisions seemingly at odds with their declared methodological principles?Addressing a relative gap in the literature, the present study focuses on teacher cognition in relation to CLT in an under-researched UK context.

Literature review
Teachers wondering how to teach are not short of advice.For example, a recent debate in ELT Journal addressed the question: Which should be more influential, the teaching context or methodological principles?In opening this debate, Penny Ur (2013, p. 470) argues that, while working within external constraints, teachers should adopt "their own situated methodologies, driven directly by the question 'How are my students likely to learn best?'"; this is a focusing question which, at face value, it would seem hard to disagree with.However, Ur also holds the view that since methodological approaches frequently promoted on teacher education programmes, such as CLT or TBLT, can be rejected as unsuitable by teachers in diverse contexts, for example, those working with large classes focused on high stakes exams in TESEP countries, it might be better if theorists stopped promoting these approaches, instead encouraging "localized methodologies" (p.473).In counter-argument, Duncan Hunter (2013) questions firstly whether teachers do conform to theory-driven methods, suggesting instead they might generate new practices autonomously in a principled way.Secondly, he suggests practices should not simply be driven by unfavourable contexts that can inhibit the kind of creativity that might support learning; rather, inspired by principles, teachers might still be able to push the boundaries.
It is increasingly recognized that teachers' beliefs are crucial in shaping their behaviour (Kumaravadivelu, 2001), andHunter (2013, p. 480) criticizes Ur's treatment of these, arguing that rather than being "merely the passive products of a particular context," these beliefs have the potential "to interact autonomously with, and interpret features of [the] environment."This immediately suggests that more attention needs to be paid to these beliefs by theorists such as Ur (2013) and that existing teacher cognition research (e.g., Borg, 2006) could be consulted more closely by any such theorists wishing to argue how context and methodological principles should shape teachers' behaviour.We now turn then to this emerging teacher cognition research area for insights.
One strand of teacher cognition research has focused on the interpretation of teachers' methodological principles.Much of this work has focused on grammar teaching or literacy instruction, as Borg (2006) reports, in citing studies from contexts including the UK, the USA, Hong Kong, Singapore and Colombia.However, Borg also highlights an emerging CLT theme, with several of the most influential of these studies emanating from Australia.These include Sato and Kleinsasser (1999), who interviewed and then observed 10 secondary school teachers of Japanese, exploring their practical understandings of CLT, and Mangubhai et al. (2004), who used observations and stimulated recall interviews to assess a secondary school teacher of German's practical theory of CLT.While the teachers in the Sato and Kleinsasser (1999) study, whose levels of experience and qualifications were mixed, "held varying, even fragmented views" of CLT, which reflected "the challenges these teachers faced" (p.501), greater consistency was found in the Mangubhai et al. (2004) study of one very experienced and highly qualified teacher.Indeed, the authors identified an extensive overlap between the principles that informed her work and "common" CLT features; these included a focus on developing students' communicative competence, engagement in spontaneous interaction, the promotion of fluency and the use of activities that encouraged the negotiation of meaning, as well as a facilitative teacher role.However, besides these CLT features, the teacher's cognitions also contained principles drawn from different strands of education in general, for example, matching lesson content to students' attention spans and encouraging intercultural tolerance.None of these features of general education, though, was considered inconsistent with CLT, while non-CLT features that might include, for example, de-contextualised grammar exercises or grammar translation techniques, were not in evidence.
Other more recent studies have reported a similarly complex picture, for example, Wyatt and Borg (2011), which highlighted how two Omani state school English teachers (thus working in a TESEP context), studying on an inservice teacher education programme that had introduced them to communicative tasks (Cameron, 2001), appeared better able to cope with contextual challenges than a third; they appeared to teach more closely according to their methodological principles and the needs of their learners, as they perceived these.Indeed, the cognitions and practices of these teachers, assessed in terms of how they provided a realistic context for the target language, encouraged meaning-focused interaction through closed pair and group work, and managed learner-centred error correction, for example, through anonymous feedback, elicitation and peer checking, seemed compatible with a "weaker" form of CLT.However, some of their principles can better be classified as relating more closely to supporting child development and education in general, for example, being friendly with learners, patient and motivating, than with CLT in particular (though these "general" principles were not incompatible with this; indeed, they appeared to support it).
Weaker versions of CLT, for example, with lessons structured as presentation -practice -production (PPP) and thus including a pre-planned focus on form, have also been observed in other studies conducted in TESEP contexts, for example, Hong Kong (Carless, 2003).Indeed, there is very little evidence in teacher cognition research, in both BANA and TESEP contexts, of stronger forms of CLT, these typically characterized by TBLT lesson paradigms and the centrality of meaning-focused tasks, with any form focus likely related to emergent needs and provided towards the end of a teaching session (Ellis, 2003).Since stronger versions of CLT are sometimes promoted on teacher education programmes, the apparently limited uptake of this methodology, as captured by the few teacher cognition studies available, might provide some support for Ur's (2013) arguments, if one neglects the likely educational value of learning about a range of methods that might be drawn upon eclectically.Though teacher education might vary in impact, courses that engage in awareness-raising, like the DELTA (Borg, 2011) and the CELTA, which is claimed to provide "skills, knowledge and hands-on practice" (UCLES, 2014), are thought more likely to influence teachers' cognitions than those that do not (Borg, 2006).When teachers trained through more traditional transmission-type courses have been found not to implement CLT, this has sometimes been ascribed to the teachers' lack of understanding of the methodology, for example, by Karavas-Doukas (1996) in Greece.
Alternatively, in TESEP countries, context is sometimes blamed for "gaps" between teachers' declared methodological principles and practices, for example, in Richards and Pennington's (1998) study of Hong Kong novices who abandoned CLT to develop survival skills.It seems that, despite apparently valuing CLT, teachers in some TESEP environments, such as the Vietnamese educators in Hiep's (2007) study, can face numerous constraints in employing it: systemic (traditional exams and large class sizes), cultural (beliefs about roles and relationships) and personal (e.g., limited experience in designing activities for monolingual classes with learners who perceive no real need to communicate in the target language).Accordingly, in some of these TESEP contexts, for example, Hong Kong (Carless, 2003), where the adoption of specific CLT methodological practices is mandated by educational authorities and enshrined in materials, there is a danger that teachers might feel themselves reduced to mere "implementors of professional theories" (Kumaravadivelu, 2001, p. 541), although it might be an over-simplification to suggest the teachers themselves would actually accept this position (Hunter, 2013).
A concern that inappropriate methodology might be foisted on students has contributed to the development of a "postmethod" approach centred on supporting learner autonomy, a goal that might only be achieved if teachers themselves are reasonably autonomous, able to draw on context-sensitive methodological principles guided pragmatically by their learners' needs (Kumaravadivelu, 2001).As yet, though, there is still limited teacher cognition research documenting such "postmethod" practices, even though a concern to support learners' needs (a methodological principle drawn from education in general) emerged, according to a corpus study that subjected several decades' of ELT Journal articles to computer analysis to identify trends in choice of lexis (Hunter & Smith, 2012), in the discourse of the 1970s, pre-dating the naming of CLT.
In stimulating principled teacher behaviour, learners' needs can be seen as contributing positively to the context, unlike the negative factors described two paragraphs above.Set in a BANA environment characterized by ongoing professional development and small multilingual groups in suitably-resourced classrooms, conditions which might facilitate a focus on learners' needs, according to Holliday (1994), our own study asks the extent to which the participating teachers' cognitions and practices relate to their methodological principles or seem situated in relation to the context.
As noted above, teacher cognition research in relation to CLT in BANA countries is rare, and of the studies available both Sato and Kleinsasser (1999) and Manghubhai et al. (2004) focus on teachers in secondary schools.Some insights into what can take places in contexts characterized by the "ideal" "private language school ethos" (Holliday, 1994, p. 4) are offered, though, by Farrell and Bennis' ( 2013) study of two CELTA-qualified teachers in Canada (one with 19 years' experience and the other a "novice" with 2 years and 6 months' experience), both teaching small classes in a language centre.Farrell and Bennis found that the beliefs and practices of the more experienced teacher tended to converge (around what appears to be a weak form of CLT), although not entirely; for example, despite a stated preference for "practice activities to be communicative . . .target grammar was only practiced communicatively in one of three [observed] lessons" (p.172).Contextual factors, such as time constraints, "made it difficult [for him] to put his beliefs fully into practice" (p.174).Furthermore, while the teacher emphasised that it was "important to have a variety of [error correction] techniques because different students have different learning styles," he was observed to use one technique predominantly, "possibly [because he] felt that each student could gain from that particular method" (p.174).In the case of the "novice" teacher in the same study, a "divergence rather than convergence between his stated beliefs and classroom practices" was observed, possibly as he was experimenting with a number of approaches and techniques, which suggested to the researchers "that many of his beliefs were not stable at the time of the interviews" (p.174).Interestingly, this novice teacher also seemed to make "instructional decisions based on keeping his students happy," unlike his more experienced colleague, who "focused more closely on needs associated with learning outcomes" (p.175).These findings demonstrate therefore that even in BANA contexts where conditions include small class sizes, well-resourced classrooms and the availability of appropriate books and classroom material, teacher cognition research can reveal a more complex picture than one of a purely methodologically principled teacher focus on learners' needs, of the type Holliday (1994), above, suggests we might find.Our research, set in a similar BANA context, is described below.

Research context and participants
Our research took place at a large adult education college in southern England, which provides EFL courses for international students, who are mostly Arabic, Asian and European.Their classes are of mixed nationality and gender.Class size is typically small (6-15) and classroom layouts tend to be facilitative of interaction.CLT methodology is actively promoted at the college, as is evident in its promotional literature, which also stresses that students will improve their grammar, vocabulary and pronunciation, and skills of reading, writing, speaking and listening through the course.The college takes students from beginner to advanced levels and, while some classes are qualification-focused (e.g., IELTS), most are general EFL, including summer school courses.Students, studying for 25 hours per week, progress from one level to another at intervals generally through successfully completing in-class tests, though this is sometimes left to the discretion of the individual teachers; the tests used, typically drawn from the relevant teachers' books, can be classified as achievement tests (Hughes, 1989) since they assess how well students are able to manipulate the grammatical and lexical forms they have been introduced to.Classes observed in our study were general intermediate-level EFL summer school courses.
Three teachers, all British, employed at the college volunteered to participate.Guaranteed confidentiality, anonymity and the right to withdraw at any stage, in accordance with the ethical guidelines required by the university that had authorized this research, they were observed and interviewed once each.Their details are included in Table 1.As the table demonstrates, qualifications and experience varied considerably.However, all had or (in Andrew's case) were working towards advanced qualifications.He was also different in not being "experienced," if we follow definitions used in other teacher cognition studies, for example, Mok (1994) cited in Borg (2006), which set a requirement of at least three years for this.In Farrell and Bennis' (2013) terms, he was in fact a "novice."However, given the way TBLT has become dominant in the last decade (Littlewood, 2007;Ur, 2013), it is likely that, since Andrew's teacher education was more recent, he may have encountered a stronger form of CLT than had Bethany and Caroline earlier.Evidence, for example, that provided by a teacher in Andon and Leung (2014), suggests there may have been a move towards TBLT on teacher education courses such as the DELTA.
Besides drawing on their formal education, the three teachers in our study were able to benefit additionally from a thorough in-house professional development programme at the college involving regular observations, progression targets and encouragement of reflective practice.This included an emphasis on supporting learners' needs, a requirement too of the UK governmental standards agency, OFSTED, which inspected the college.Having introduced the participating teachers and their context, we now describe our research methodology.

Research methodology
To assess the relative importance of context and methodological principles in shaping the teachers' work, "non-participatory observations" (Cohen, Manion, & Morrison, 2007) focused on the participants' teaching practices while their cognitions were elicited through semi-structured interviews (Kvale & Brinkmann, 2008).Interviews followed observations to stimulate recall (Bailey & Nunan, 1996) and explore the rationale for classroom decisions, a procedure which also reduced the risk of the observed teaching being influenced by the observee's knowledge of the research focus.Such risks were further minimised by the nature of the interviewer/interviewee relationship; the observer (the first researcher) was an outsider and had not met the research participants before contacting the college with a request to conduct the research.At the time of engaging in this research, the observer was DELTA qualified, experienced in EFL adult education settings in the UK and overseas, and was studying for a further post-graduate qualification.
To assess practices for evidence of methodological principles, criteria related to differing versions of CLT, including strong and weak, were developed, as in other teacher cognition studies (e.g., Wyatt & Borg, 2011).So, questions such as the following were posed: Did the lesson suggest a TBLT, PPP or other structure?To what extent was language contextualized?How meaning-focused was the interaction?How prominent were pair and group work?How were language errors addressed?
In order to reduce the subjectivity of the observations and the potential for "attribution error" (Kennedy, 2010) and therefore also increase the reliability of subsequent analysis (McDonough & McDonough, 1997), a tool (first piloted) was developed (see Appendix A).This allowed for the recording of timings and participation patterns, felt to be of interest since CLT is characterised by the active ontask participation of students (Thornbury, 2004); this tool would support the narrative recording of events.Lessons were not video or audio-recorded to reduce intrusiveness.Each classroom observation lasted 40-50 minutes.Data analysis commenced immediately afterwards, with participation patterns scrutinised and notes written up.Features of the lesson were assessed against criteria.
As the goal of the interview was to elicit cognitions in an unrestrictive manner, a semi-structured interview format was employed (see Appendix B), allowing participants to contribute freely to the conversation within the guidance of a loose framework (Dörnyei, 2007).Interviews lasted approximately 30 minutes and were tape-recorded; this was to allow for subsequent analysis and to free the interviewer to focus on the conversation rather than on recording what was said (Kvale & Brinkmann, 2008).There were five stages.After the study's broad aims and ethical safeguards were restated, general questions were asked, in a friendly conversational way to help build rapport, about experience and training.Next, to encourage teachers to express their methodological principles in their own words, there were general questions about their practices; these questions avoided potentially ideologically loaded terms such as CLT and TBLT, which might have triggered pre-conceived notions.
Subsequently, a short stimulated recall process was employed, as used in similar teacher cognition studies, for example, Mangubhai et al. (2004), but without the video.This involved presenting detailed observational data, already filtered by the observer, as neutrally as possible in summary form, emphasising staging and participation patterns, exploring the rationale behind classroom decisions and enquiring whether this was typical of the teacher's work.Notwithstanding the lack of video, the process thus employed involved a more precise focus on classroom events to prompt recall than that found in "typical retrospective ethnographic interviews" when informants might be asked to remember only "in the abstract" actions taken and values or strategies linked to these (Dempsey, 2010, p. 350).Follow-up questions explored the possible impact of contextual factors, such as the needs of the particular learners in the class.There was then space for further comments.
Interview data were transcribed and then coded, for example, for descriptions of CLT and non-CLT practices; for instance, if a teacher mentioned believing in meaning focused tasks, this was coded as CLT.Methodological principles and situated cognitions were compared across different interview segments and with reference to observational data.Qualitative descriptive accounts were then developed and are presented below.

Andrew
Andrew's lesson, with 8 young adults (3 Chinese, 2 Arab and 3 European) of intermediate ability, was in a very informal setting for a classroom, with sofas and bean bags instead of chairs and no tables.Regarding the structure of the lesson, there was an initial emphasis on building context and engagement with a real-life situation (described below) before presentation of grammar, itself heavily linked to this context.The lesson began with a hypothetical discussion from picture prompts on what it would be like living in a jungle, followed by reading tasks based upon a woman's true "life in the jungle" story.There was subsequent personalisation of content, leading into meaning-focused activities that indirectly provided exposure to the present perfect simple and continuous tenses, before there was an explicit focus on form.This style of presentation struck the observer as reminiscent of typical DELTA trainee lessons he had witnessed; Andrew of course was completing the DELTA at the time.The lesson was also characterised by considerable student participation and interaction, entirely in English; all tasks were completed in pair work, and feedback was elicited from the students themselves, including the correction of errors such as when a student was given multiple opportunities to pronounce the word bite correctly.Furthermore, the session had an informal, cafe-style feel, partly created by the layout, with casual conversation involving both teacher and students during all stages.
In summary, although elements of the lesson were clearly consistent with general education principles, such as an interest in learners, the lesson was also characteristic in many ways of a stronger form of CLT, as outlined by Ellis (2003), above, given the way that it was structured around a meaning-focused task.However, while features of TBLT were dominant, there may have been other influences too.For example, the focus on form towards the end was of a fairly explicit nature.
In the subsequent interview, Andrew suggested his preferred teaching style matched the style observed in his lesson; first, set an activity "to activate schemata . . .so that (the students) are not just going head first into something which they have no idea about" and then present the language in context to encourage learners to notice the form of language for themselves.Furthermore, he asserted that contextualisation was a high priority in his teaching: I think noticing things is a big part of learning . . .learning English and (this kind of activity) puts it in context and it gives it some kind of real world relevance, and I would always, I think, I would always present language in context before and give the students the opportunity to guess what it means before teaching it, definitely . . .He adds that he achieves this predominately through "skills-based activities."Finally, having provided sufficient contextualisation, Andrew teaches the form of the language.
Regarding student participation, Andrew values opportunities to "chat and talk" to create a relaxed "buzzy" environment in which learners, involved as part of "a big community helping each other learn," feel like practising their English.He also prefers involving students in error correction, his reasoning being that, in order to foster genuine learning, he wants "students to work for their answers".
Andrew emphasized, though, that he would vary his teaching style if the contextual factors were different.His limited experience had included a year teaching in secondary schools in Thailand (a TESEP context), and he reported that with children "you might need to be a little bit less openly communicative and a little bit more directive in terms of (achieving some) tangible task output"; this practice might suggest a form of communicative task such as that described by Cameron (2001), allied to a weaker form of CLT, rather than the "deep end" CLT (Thornbury's [2004] term for strong CLT), which Andrew indicated a preference for elsewhere in the interview.Moreover, the context would determine his choice of topic.Choosing appropriately motivating topics was vital with young learners in a TESEP context, he argued, as "you really want to try to engage them." Asked if languages were best learnt through communication, Andrew replied "I think this depends on learners and learning styles . . .some students are naturally better at getting more out of a communicative situation."He also reported that in a monolingual environment, such as Thailand, he "wouldn't be adverse to using translation, if necessary."While this principle suggests no conflict with TBLT (Willis [1996, p. 49] accepts mother tongue use that is "systematic, supportive and relevant"), it does indicate flexibility and willingness to put learners' needs first.
In short, Andrew favoured methodological principles which lean towards strong CLT: the use of tasks to build context and underpin learning with any explicit focus on form coming afterwards.He acknowledged, though, that the specific needs of students and the type of context (BANA/TESEP) influenced how he enacted his methodological principles.

Bethany
Bethany's class consisted of 9 intermediate level young adults (5 Europeans, 3 Asians and 1 Arab) in a standard classroom environment containing chairs and tables in a horseshoe shape.In the lesson observed here, grammar teaching was the central theme.The session observed consisted of three cycles, each beginning with an explanation of a grammar rule (cycle 1: I wish, cycle 2: I wish I had and cycle 3 for other aspects of I wish) and ending with controlled or free practice in using this rule (giving the general impression of a weaker form of CLT).Contextualisation occurred within Bethany's explanations through elicited situations and picture prompts, one example being during the teaching of I wish, which involved two people playing basketball, one short and one tall, with the shorter person wishing he was taller.
Although much of the lesson involved teacher-led explanations, student interaction was maintained throughout.This was achieved through elicitation before presentations of grammar and through both work in mixed-nationality pairs and open class discussions on the topic of students' own wishes, which thus encouraged personalisation.Corrections were elicited, apart from pronunciation errors which the teacher corrected directly.Student participation was greatly encouraged through the informal mood of the session, with students frequently interjecting and Bethany participating in digressions off topic.
When speaking with Bethany afterwards, it was quickly apparent that planning the stages of a lesson was considered very important, a belief born out of training to be a children's teacher, she reported.Her ideal lesson would involve cycles described as "express, correct and repeat": As well as stressing the need for repetition, she emphasised the importance of testing prior knowledge in the first cycle so that new knowledge could be built upon a learner's existing knowledge.When presenting language, Bethany was open to guided learning or teacher-led explanations, the latter more likely within grammatical explanations.Then, in discussing activities in her observed lesson that required students to apply language to their own lives (for example, their own wishes), she commented on the importance of contextualising language learning through personalisation to aid memory, rather than relying on workbook exercises: I keep telling them that if I don't make anything personal, they're never going to remember it, so I did the first activity where it was all my own ideas and I think, they would need a chance to come up with their own experiences, and the textbook, there's . .., we were at the right place in the textbook but a lot of it was with superpowers, being able to see in the dark, being able to see in the future, you know, not really very realistic, and we did it as a revision exercise but I think if they can give a real example of a real situation that's personal to them, they're more likely to remember it . . .On the participation of students, Bethany values an interactive approach and so fosters an informal atmosphere.However, this interactivity is more likely to succeed, she claimed, when controlled by the teacher, as an uncontrolled classroom merely results in confusion, "with people just sitting there, doing nothing." Earlier in her career, Bethany had taught "really big groups" in the Czech Republic (a TESEP context), and felt it had not been "really different.The lesson would be pretty much the same."This suggests she was applying techniques developed with larger groups (children in a UK mainstream school and European students) to the small group teaching in her current context.However, as with Andrew, Bethany stressed that her teaching approach varies according to the students.
Interestingly, I did a similar lesson with another class and they didn't know anything whereas there was a girl in that class who seemed to know the grammar quite well.So a lot of the first bit was just testing to see how much they already knew.
Furthermore, Bethany reported eliciting continually in the observed class due to its vocal nature, but for less vocal classes she would elicit less, accordingly.She also reported sensitivity to individual needs: "every learner's different . . .while some do just want general communication skills for the outside world, others actually want to learn a language, all the rules and grammatical structures."So Bethany's preferred methods resembled weak CLT to some degree, with a focus on form being mixed-in with personalisation, interaction and other forms of contextualisation.As with Andrew, she believed that learner characteristics and needs impacted the way she enacted her methodological principles.

Caroline
For Caroline's lesson, 9 intermediate level young adults (3 Asians, 2 Arabs and 4 Europeans) were present, sitting in a standard classroom with tables and chairs in a horseshoe shape.As in Bethany's lesson, there was an emphasis on the mastering of grammatical structures, though the approach seemed different in some ways.
After beginning with a teacher-led activity revising collocations learnt the previous day, students completed a task in which they invented questions to ask Royal Family members, hardly very realistic!Questions had to begin with who, what or where, a requirement which could have stifled spontaneity but was strictly enforced.Caroline then used or adapted students' utterances to teach the difference between object questions ("Where are you going on holiday?") and subject questions ("Who loves Mary?").The session ended with two activities providing practice in using subject and object questions, the first involving the rearrangement of jumbled words and the second requiring students to write subject or object questions in response to teacher prompts.
The lesson contained some features which would not typically be considered CLT; for instance, contextualisation was limited largely to the invention of questions early in the lesson, before subsequent explanations and exercises focused mainly on the structure being taught.Students participated through inventing questions in pairs and responding to teacher prompts.However, as was the case in Andrew's and Bethany's classes, students were placed in mixed-nationality pairs and consequently used English to communicate.On the whole, though, less student participation was observed than in Andrew's and Bethany's classes; much lesson time involved Caroline explaining, instructing and providing corrections, although some instances of self-correction were encouraged.Overall, the atmosphere was quite formal, with students often quiet, not contributing more than necessary.
The subsequent interview with Caroline revealed cognitions which indicated responding to "student factors" was more important to her than following any methodology.In her interview, Caroline did outline a preference for methods comparable with PPP.This was evident firstly from frequent uses of the word build to describe how activities should follow one upon another: we always kind of build up to an activity where they are hopefully going to be able to express their own opinions or putting in their own language, giving their own ideas in their own way based on what the input has been . . .Caroline emphasized the importance of "building up their confidence" through encouraging speaking in pairs and small groups, in which they "feel secure": a lot of them do enjoy just answering questions and just discussing things you've led up to . . .I think you just have to give them the opportunity to feel confident and to try to express themselves, as and when they want.
Furthermore, it was evident Caroline valued PPP; she claims that she usually begins with explanations of grammar, which lead in to activities providing contextualised practice and allowing self-expression.However, this reported practice was in contrast to the largely non-CLT methodology observed.Caroline also said the self-discovery of grammar would normally permeate her lessons; again, this was not observed.
During stimulated recall, Caroline claimed she limited self-discovery, student participation and contextualisation in the observed lesson deliberately to address the needs and wants of students.Meeting these was more important to her than following an approach; for example, although Caroline reported valuing communicative tasks, she was conscious of her students' preference for accuracy-focused activities, a preference she consequently needed to respect.Student-stated preference had also played a part in her decision to favour teacher-led correction: you also get students who say 'it's all very well doing these communicative activities and speaking in pairs and groups but, you know, we really want you to correct us, both grammatically and pronunciation wise,' and I think they're quite right really . . .Caroline acknowledged that "what everyone wants to do (is) to be able to speak" and confessed she was "perhaps a bit of a stickler for accuracy."Nevertheless, the learners observed were retaking a level previously failed; Caroline prioritised accuracy and emphasised correct forms to help them avoid failing again: I do think that structures at this level, at intermediate level are actually important because if they're not going to learn the grammar at this stage there are going to be fossilised errors and they're going to struggle throughout the next level and so on . . .To summarise, although CLT methods were apparently appealing to Caroline, when she was reminded of what actually happened in her lesson, situated cognitions came to the fore.One interpretation might be that her own methodological preferences were of less importance to her than were student-led contextual factors she felt she needed to respond to, such as those outlined above.However, an additional possible interpretation is that, when confronted with contextual realities, "politically correct" espousals of CLT methodology that she may have been exposed to throughout her career and drawn upon, given the very pervasiveness of this political correctness (Waters, 2013), fell away, and methodological principles, important to her but nevertheless suppressed initially in interview, which relate more closely to non-CLT practices, for example, prioritizing decontextualized grammatical accuracy practice activities, emerged.

Discussion
We now consider these cases together and in relation to research questions and the literature.Firstly, the study uncovered two clear examples (in Andrew and Bethany) of harmony between methodological principles, situated cognitions and teaching context.Both teachers combined a concern with learners' needs, which, as we have said, relate primarily to general education (Mangubhai et al., 2004), with methodological principles that relate clearly to CLT.However, while Bethany's observed lesson and interview afterwards revealed a preference for a weaker version of CLT, in Andrew's case, preference for a stronger version was evident.So, in his lesson, tasks preceded focus on form, which was heavily contextualised; students were communicatively involved throughout.Then, in interview, Andrew's comments indicated a compatible preference for collaborative skills-based activities which provide contextualisation and opportunities for the noticing of linguistic features.Like Bethany, who varied levels of elicitation and explicit instruction according to the learners' needs, Andrew felt he was flexible in adjusting his methodology according to the students, though these adjustments still seemed compatible with his principles.Therefore, there was no need to compromise methodological principles, as Ur (2013) argues might be necessary, in order to accommodate contextual concerns.
Perhaps such harmony is not surprising.Andrew's lesson, after all, was centred on a "task" (Ellis, 2003), and, as Hunter (2013, p. 477) suggests, these can attract teachers due to "the creative and humanistic possibilities they offer; [they] are adaptable for use with topics that interest learners: they absorb students in meaning and involve them in intense social activity."Teachers who believe "in the essentially humanistic and communicative nature of language" may be drawn to CLT methodology (Harmer, 2003, p. 291) they can enact more easily if promoted on their teacher education programmes and supported by their contexts.Unlike teachers in Richards and Pennington (1998), Andrew and Bethany benefited from a BANA context characterized by a multilingual environment, small class sizes, and continuing in-service professional development that emphasized appropriate response to learners' needs allied to a CLT approach.This approach was realized in relatively relaxed general EFL summer school courses.Furthermore, while Andrew had no more experience than the CELTA-qualified "novice" in Farrell and Bennis' (2013) study, this was clearly varied and he was already continuing his education through the DELTA, which, as Borg (2011) explains, is an advanced course.This continuing education might partially explain the relative congruence between Andrew's cognitions and practices.
Such harmony was not apparent in the work of the third teacher, Caroline, who, despite espousing support for a weak version of CLT, did not teach in a way congruous with her stated beliefs; non-CLT practices were more in evidence.Nonetheless, in interview she offered a clear contextual reason why.Indeed, Caroline underlined that the needs and wants of the students were her primary consideration in all her teaching choices, regardless of her personally held preference for a PPP delivery style.Her students had failed end of level achievement tests of their grammar and vocabulary and had also specifically requested direct teacher instruction to support linguistic accuracy.Caroline was sympathetic to this, and there are different possible interpretations.One is that Caroline had suppressed non-CLT methodological principles early in the interview, as these may not have seemed politically correct to her.Alternatively, though, if we take her stated beliefs at face value (remembering that, rather than being elicited directly, these beliefs were inferred from her choice of lexis and descriptions of how activities ideally related to one another, a methodological procedure we followed with a view to accessing more deeply-held principles), Caroline was adopting her own "situated methodology," driven by her interpretation of learners' needs, an approach which Ur (2013, p. 470) argues is likely to result in "a substantial improvement in learning for students." A more critical perspective, though, might be that Caroline was positioning herself as a "client-satisfier" (Harmer, 2003, p. 288), in so doing abandoning methodological principles that many researchers (e.g., Ellis, 2003) believe may lead to deeper learning.The extent to which this was a risk, though, is difficult to assess.Only one lesson was observed and, while the learners did not appear to be engaged to any great extent, such engagement in itself is of course only one possible indicator of learning.These learners were not interviewed subsequently and nor was their progress monitored as part of our research.Caution in interpretation is thus essential here, particularly since Caroline may have been channelling her observations and reflections into her teaching.
Nevertheless, Caroline did employ non-CLT practices.While she explained clearly that she had other priorities, Hunter (2013, p. 479) warns that worrying excessively about "difficult" contextual factors, such as learners' tests, can lead to teachers abandoning creativity in favour of "unhelpful conservatism."This may possibly have been the case here, although without substantially more observational data any such conclusion would be far too premature.
This suggestion does allow the insight, though, that even in such a BANA teaching context, where the concern on EFL summer courses was generally primarily perhaps with "immersing learners in Anglo Saxon society" (Hiep, 2007, p. 195), "difficult" contextual factors, for example, test pressures, could still seem to exert "undue" influence.Even though the college appeared to be encouraging CLT, were its end of level achievement tests in fact inducing negative washback, for example, by "focusing heavily on surface features of grammar" (Hunter, 2013, p. 479)?While a detailed study of these tests was outside the scope of our investigation, that they were achievement rather than proficiency tests (Hughes, 1989), assessing discrete grammatical and lexical items introduced in the course book rather than communication skills more holistically, is telling.Caroline's situated cognitions and practices suggest her interpretation of these tests may have been that accuracy was more important than other features of communication.
In Caroline's BANA environment, there was an emphasis on supporting learners' needs, partly driven by OFSTED, and, as Ur (2013) recommends, her practices seemed governed by her situated cognitions.However, when considering contextual factors, an important distinction can be made.On one hand, there are learner needs, in terms of characteristics and learning styles, for example, confidence of the learner in a communicative situation, which the "postmethod" teacher would probably be only too willing to draw upon in fine-tuning their teaching in a principled way, while on the other, there are "difficult" contextual factors.This second set of factors, for example, the high stakes exams, particularly found in TESEP contexts, that might encourage various forms of undesirable student behaviour such as surface approaches to learning (Chan, Spratt, & Humphreys, 2002), may make it harder for a teacher to teach in a way consistent with their methodological principles.
Ur's (2013) discussion of "situated methodologies" does not make this distinction clearly enough.While Andrew and Bethany were drawing on methodological principles in meeting the individual styles of their learners, as Holliday (1994) in fact suggests can happen in BANA contexts, Caroline was also swayed by apparently "difficult" contextual factors, namely the learners' end of level achievement tests.Ur (2013) acknowledges that such tests are invariably "paper and pencil" since "the testing of oral proficiency is relatively expensive and time consuming" (p.471) and that this will impact teaching, as will marking that "involves substantial subtraction of points for grammatical and spelling mistakes" (p.472).She nevertheless argues that such contextual factors should still drive teachers' practices.Hunter (2013), however, disputes this, pointing out that "many teachers, thankfully, refuse to accept the notion that classroom teaching and learning should be directed solely towards assessment" (p.479).
The extent to which Caroline and her colleagues had a choice in the way they set and marked these achievements tests on their summer school courses is unclear.However, Caroline's apparent acceptance of a perhaps difficult contextual factor, of the type that Hunter (2013) argues needs to be guarded against, may partially explain the relative lack of harmony identified between her declared methodological principles and situated cognitions.Much greater harmony was observed, as we have noted, in Andrew and Bethany.
There are various implications arising from this study.First, to help all concerned, this particular college could re-evaluate the communicativeness of its tests.In such a typical BANA context, characterized by small classes on summer courses that have been set up to aim primarily at improving oral communication skills (though of course Caroline's learners may have been more interested in accuracy even if they had not failed a level and been asked to repeat), it would seem that other means of testing would be possible.For example, carefully constructed role plays, assessed according to criteria that relate to communicative competence (Bachman & Palmer, 1996), could be incorporated to supplement teachers' book tests.The inclusion of such role-plays would better match the college's stated goals to develop oral communication skills, but of course such an innovation would need careful management and monitoring.Second, though we would not wish to generalize from such a small-scale study, Caroline's case does illustrate the concern voiced by Hunter (2013) that if contextual factors, including those open to challenge since they might seem to have a negative impact on learning, such as assessment that is ill-matched to the aims of a course, are placed ahead of methodology, as Ur (2013) recommends, this may not necessarily result in the positive outcomes she anticipates, for example, "more professional satisfaction for the teacher" (p.470).Indeed, if there is a lack of harmony between methodological principles and situated cognitions, such a positive outcome intuitively seems to us less likely.Though our own study could ideally have probed this particular issue more deeply through follow-up interviews and observations, this insight does suggest that any advice given to teachers on meeting needs and handling context should be informed more fully by teacher cognition research.
Furthermore, Caroline's case, taken together with the findings of Farrell and Bennis (2013) in Canada, demonstrates that even in contexts characterized by small, multilingual groups of students studying in well-resourced school environments in English-speaking countries that are apparently favourable to convergence between methodological principles and practices (Holliday, 1994), divergence may nevertheless occur.Concerns about assessment, pressures of time, and, perhaps in the case of the novice in Farrell and Bennis (2013), undeveloped belief systems due to inexperience, can all lead to methodological principles, which in these cases seemed to relate to a weaker form of CLT, not being put into practice.For example, while Caroline and the two teachers in Farrell and Bennis (2013) affirmed in interviews that language should be contextualized, in their practice this contextualization was largely absent.
This finding then, admittedly emerging from two small scale BANA studies, should help to put into perspective criticisms of CLT "failing" in monolingual TESEP contexts characterized by large classes and limited resources.As Holliday (2006) has implied, there is too often an "othering" in public criticisms of teachers' "misconceptions" of and "inability" to apply CLT in TESEP contexts, which seems unjustified.Teachers in BANA contexts, too, might not find perfect congruence between their methodological principles and practices, as we have demonstrated.A modest implication of this study then is that Robinson Crusoe nativespeakerist types looking to enlighten the rest of the world with their superior methodological practices (Holliday, 2006) should look first to home and the complexities evident in their classrooms there.
On a more positive note, the findings here relating to Andrew and Bethany demonstrate that, in a BANA context, it is indeed possible, perhaps with a "postmethod" stance (Kumaravadivelu, 2001), to consider learner characteristics pragmatically while teaching according to methodological principles, including those relating to different forms of CLT.Indeed, our findings suggest that creating a facilitative environment, characterized by small classes, sensitivity to the learners, and ongoing professional support, can help teachers such as these achieve this balance.

Conclusions
In this study, we set out to explore the congruence between methodological principles and practices of three teachers in a UK adult education college.In two cases, the teachers' personally held methodological principles, regarding CLT and education more generally, were highly evident in their practices.Various features of the BANA context in which the study took place appear to have been helpful in allowing them to teach harmoniously in accordance with both principles and learners' needs.The case of the third teacher, though, reminds us of how contextual factors, such as a preoccupation with tests, can lead a teacher away from classroom practices that match declared beliefs.Even though it did not appear to affect the other two, the assessment strategy employed at the college appears to have influenced this third teacher, contributing to an observed lack of fit between stated principles and practices, and a gap between the intended curriculum and how it was enacted.However, this gap could have been reduced through the adoption and careful monitoring by the college of assessment methodology that encourages communicative competence.In this BANA context, then, it seems there may have been a need for clearer leadership in this regard, with the assessment strategy in place subjected to careful review and subsequent revision.This insight, in itself, demonstrates a value of teacher cognition research such as this, which, through being small scale and local, can address local issues.Furthermore, if research of this nature is shared more widely, it can support the questioning of broader assumptions often made all too readily, for example, regarding the presence of ideal conditions in all BANA contexts.Such studies can only be produced, though, through the willing participation of teachers, and if the results are to be meaningful it is important that their contributions are unfiltered as much as possible by affective concerns.Notwithstanding its limitations, a notable strength of the current study was that the observer, while being an outsider, was able to build a degree of trust.To retain this trust, so that teachers feel they can continue to contribute to the development of context-sensitive knowledge of classrooms in under-researched parts of the world, it is vital that teacher cognition studies such as this are underpinned by research methodology that is rigorously ethical.

I
think providing opportunities to express themselves . . .and then correcting what they produce I suppose.And then repeating . . .a lot of repetition, in different ways, you know, recycling language in a different way but making sure it's repeated and repeated and repeated.

Table 1
The participating teachers