EFL teacher motivation in-situ: Co-adaptive processes, openness and relational motivation over interacting timescales

This paper presents an exploratory case study of the classroom motivational dynamics of an English as a foreign language (EFL) teacher at a Japanese technology college. The article examines how motivation evolved in-context over various timescales through interactions with affect and identity. An introspective research journal generated rich, qualitative data concerning fluctuations in teacher motivation over one academic year. The analysis also drew on student journal data to provide a different perspective on teacher reflections. The study applied a thematic analysis, with “theoretical comparison” (Corbin & Strauss, 2008) to understand teacher motivation from a “person-in-context relational view” (Ushioda, 2009). The article utilises the properties of complex systems to render insight to the evolution of teacher motivation as open to influences “external” to the classroom, yet fundamentally tied to adaptive experiences with a particular class group. A variety of diagrammatic tools are also employed to illuminate the relational development of teacher motivation, affect and identity constantly occurring over interacting timescales.


Introduction
Study into learner motivation is without doubt one of the most prolific areas of second language acquisition (SLA) research.This rich history has involved a move from social psychological constructs, to cognitively-informed situated theorizing, process-oriented perspectives, and more recently to socio-dynamic approaches to conceptualising and exploring the motivation of second language learners (Dörnyei & Ushioda, 2011).Puzzlingly, however, while a good deal of additional language learning worldwide occurs in instructed, classroom settings, theorizing and investigation of the motivation of teachers in these learning spaces is still relatively scarce.What is it that drives second language teachers in their day-to-day enterprise in the classroom?
A small but slowly growing body of research into teacher motivation goes some way to providing insight: There appear to be a variety of contextual factors that impact negatively on teacher motivation in both general education (Pelletier, Seguin-Levesque, & Legault, 2002) and language teaching (Doyle & Kim, 1999;Menyhart, 2008;Pennington, 1995); examples include the influence of the school climate, class sizes, resources, relations with coworkers, school leadership structures and attitudes, workload, salary, student characteristics, and acceptance by the school community of teacher autonomy.Additionally, teacher motivation has been associated in a more positive fashion with intrinsic facets of the teaching process.This motivation stems from teachers' desire to shape future society through working with young people and contributing to individual student growth (Doyle & Kim, 1999;Erkaya, 2013;Kassabgy, Boraie, & Schmidt, 2001;Martin, 2006;Pelletier et al., 2002;Pennington, 1995).
Despite some promising qualitative research in the SLA field (see, e.g.Menyhart, 2008), empirical work to date has however seen a preponderance of large-scale survey studies.Such instruments render one-off, averaged snapshots, removed from distinct classrooms.Moreover, while they provide insights into factors impacting on teacher motivation, they do little to illuminate the fluctuating interplay of these elements as teacher motivation develops over time and in-context.As Kaplan (2014) summarises, "most motivational research has been either correlational research that employed self-report surveys or experimental research conducted under artificial conditions, both of which are quite different from teachers' experiences of the complex and dynamic environments of classrooms" (p.64).
The current article takes a situated approach to understanding language teacher motivation.The paper describes research in which an EFL teacher (the author) was conceptualised as a particular "teacher case."Through re-exploring data obtained in a wider study of classroom language learner motivation (Sampson, 2016) from a different angle through case study, the paper describes the contextualised ebb and flow of teacher motivation in the language classroom.While the article provides insights into the motivational dynamics of only one teacher, as Richardson, Watt and Karabenick (2014) note, "teachers' motivations matter in both the short and the longer term, not only for their own well-being and career satisfaction, but also for how they relate to and interact with students" (p.xv).By bringing to light dynamic issues relevant in one context of teaching practice, the article therefore draws on complex systems theory (CST) to encourage future research to take a situated approach in exploring the relational evolution of teacher motivation in the language classroom.

Complex systems theory and relational motivation
CST is a theoretical and philosophical position that has been gaining increasing attention in the field of applied linguistics (see, e.g. the seminal work of Larsen-Freeman & Cameron, 2008).Mitchell (2009) defines a complex system as one in which "large networks of components with no central control and simple rules of operation give rise to complex collective behaviour, sophisticated information processing, and adaptation via learning or evolution" (p.13).The metaphoric tools and understandings of CST are argued to be facilitative to social science research through encouraging the exploration of possibilities in interpretive forms of describing "rather than prescribing [emphasis added] relationships and processes" (Kuhn, 2007, p. 299).Indeed, offering intriguing possibilities for understanding teacher motivation, a review of the literature turns up the following commonly discussed properties of complex systems: • made up of multiple agents with considerable diversity and redundancy; • distributed control; • open (both offer and receive energy in interaction with other systems); • nonlinear interactions (introduced change may have varying effects); • co-adaptive interactions that alter both agent(s) and context; • constant dynamic change across different timescales; • phase shifts (sharp whole-system change); • self-organisation and emergence (gradual whole-system change) (Cilliers, 1998;Davis & Sumara, 2006;Larsen-Freeman & Cameron, 2008).While the recent years have seen an incremental growth in the number of studies utilising CST to conceptualise language learner motivation (see, e.g., chapters in Dörnyei, MacIntyre, & Henry, 2015), research into language teacher motivation from a CST perspective is still extremely scarce.One exception is Kimura's (2014) investigation of the motivation of two middle-school teachers in China.Kimura conducted video-taped observation of lessons and semistructured interviews with the teachers and seven of their students on two occasions spaced around three years apart.He used CST to organise understandings of how the teachers' motivation changed over micro (observed lessons) and macro (the three year interval) timescales.Kimura's (2014, pp. 323-324) study reveals a number of key influences on these teachers' motivation that differed by context and stage of professional career.Unfortunately however, because of the rather wide gaps between data-collection points and interviews, the study does not allow a detailed understanding of the ways in which these influences interact and motivation develops in-situ.
The work of Hiver (2015;Hiver & Dörnyei, 2015) takes a more contextualised approach.Hiver (2015) investigated what he classified as "teacher immunity," that is, how teachers cope with stress to remain motivated.He conducted a number of one-to-one interviews with four language teachers in South Korea.Hiver came to a realisation that the CST concept of self-organisation offered a useful metaphor for understanding insights from the data.It was found that motivation-sustaining teacher immunity developed through a four-stage process: (a) an initial perturbation that triggers a response; (b) a coupling process whereby an adaptive response is implemented; (c) a reformation of higher-order patterns in the system based on these adaptations, and; (d) consolidation of the new pattern of stability (pp. 220-225).Hiver (2015) concluded that motivation and teacher immunity emerge as teachers come to terms with adapting to the situated accumulation of "events, concerns or realities . . . in their daily practice" (p.220) and "accept and solidify this residue of experience as a new aspect of their identity" (p.224).Hiver's (2015) work draws attention to a fundamental process in complex systems, that of co-adaptation.Such processes involve a "kind of mutual causality, in which change in one system leads to change in another system connected to it, and this mutual influencing continues over time" (Larsen-Freeman & Cameron, 2008, p. 233).Co-adaptation involves positive and negative feedback.When an agent acts in an environment and perceives negative feedback indicating that that this behaviour is inappropriate for the circumstances, the likelihood of the adoption of a similar behaviour in the future is reduced.On the other hand, positive feedback occurs when an agent acts and perceives that this behaviour is appropriate in the environment.The behaviour is reinforced, and this may lead to an increased likelihood of its adoption in the future.Crucially, proponents of CST understand that any form of behaviour alters the environment at the same time as it alters the agent (Larsen-Freeman & Cameron, 2008).That is, complex systems are open in the sense that they interact with other systems to both receive and offer energy in interaction with their context.This said, it is important to remember that the "boundaries" of these systems are merely conceptually defined and are more akin to permeable and shifting interfaces determined by the observer (Cilliers, 1998).
The previous discussion may perhaps give an unintentionally straightforward, linear impression of the development of teacher motivation.Proponents of CST (and indeed Hiver) however argue that change is nonlinear.Rather than an effect being directly attributable to a specific cause, or development occurring in regimented stages, the present state of a system (such as motivation) is more the accumulation of change in various interrelated elements up to that point in time (Cilliers, 1998).A key property of complex systems is that some change may occur suddenly over a relatively shorter span of time, while other change takes place at a slower rate involving interactions over longer spans of time ( de Bot, 2015).As de Bot (2015) notes, a CST approach holds that "development on one scale is influenced by what happens on smaller and larger scales" (p.32).As Figure 1 depicts, classroom activity may involve a timescale of change in terms of minutes, whereas a lesson or homeroom period entails a timescale of hours.These timescales interact with each other in such a way that perceptions of teacher motivation at the timescale of a semester or academic year are an emergent outcome of experiences across different timescales.One motivation theory that seems to mesh well with CST understandings is Ushioda's (2009) "person-in-context relational view of motivation."This conceptualisation holds that motivation emerges through the processes by which people make meaning in their social context.In terms of classroom motivation, Ushioda (2011) asserts that "it is through social participation in opportunities, negotiations and activities that people's motivations and identities develop and emerge as dynamically co-constructed processes [emphasis added]" (pp. 21-22).Psychological systems, such as motivation, affect and identity are in dynamic interaction with each other as well as the evolving contextual system.Rather than treating the context as a background variable, individual sense-making and action is constrained and made possible by context, while also acting to change the context.In contrast to linear cause-effect understandings, such an approach necessitates a relational view of multiple interacting elements whereby motivation is "an organic process that emerges through the complex system of interrelations" (Ushioda, 2009, p. 220).
Taken together, the CST literature in general, Hiver's (2015) focus on teacher adaptations, and Ushioda's (2009) relational conceptualisation of motivation suggest that research examining the situated, dynamic evolution of teacher motivation through experiences and adaptations in interaction with particular class groups could make an additional contribution to the existing body of research.In order to investigate how my own contextualised teacher motivation evolved in a relational fashion, I developed the following two research questions: 1. What understandings about the emergent educational context appear to affect the development of the motivation of an EFL teacher with one class group?(RQ1) 2. In what ways does this motivation evolve over time?(RQ2) Details of the particular setting from which the data for this study were drawn are provided in the following section.

Setting and participants
The current article concerns the author's experiences whilst working at a kosen, a 5-year college of technology in Japan.These colleges endeavour to foster the creativity and practical skills of future technology workers through a mix of vocational engineering education with general learning.Kosen are a combination of the usual three years of senior high school in Japan with the first two years of undergraduate study.Students range from around 15 to 20 years of age.Further to educational responsibilities, teachers at these colleges are expected to be research active.
In this context, I had two educational roles, as a classroom teacher and homeroom advisor, with one first-grade class of 40 students aged 15-16 years, referred to herein as Class E. These students were of equivalent age to senior high school first-grade students in the regular Japanese educational system.The college context defined that this group and my roles within it would naturally occur for only one academic year (in 2011-2012), before participants (myself included) were split into different classes.The paper examines my teacher motivation with Class E over this year of 32 lesson-weeks.The analysis concerns primarily my role as a classroom teacher for one of the three weekly 90-minute English lessons for these students.In addition to these lessons, the paper will occasionally refer to experiences in "homeroom periods," which were regular, 45-minute sessions with Class E once every week involving general administrative tasks and life and career guidance.
As this paper centres on my teacher motivation, it is necessary to provide some brief information about my background.At the time of the study I was 35 years old, a Caucasian Australian national.I studied education as an undergraduate in Australia and shifted to Japan for employment purposes following graduation.I had lived in Japan for over 10 years at the time of data collection, and I operate comfortably in English and Japanese.In my time in Japan, I have worked in a range of education settings and continued my own education through postgraduate studies.I had only been employed at the kosen for one year prior to the study, and the year of data-collection coincided with my first experience as a homeroom teacher.

Research approach and data collection
In the study on which this paper is based (Sampson, 2016), I had used action research to introduce activities that would assist students to reflect on their motivation to study English.However, after the study was completed, I became intrigued to investigate how my own motivation had evolved over the year with Class E. As such, I employ an autoethnographic case study approach here to focus on "the particularity and complexity of a single case" through which we might come "to understand its activity within important circumstances" (Stake, 1995, p. xi).Case studies typically propose a theoretical boundedness to the case, endeavouring to provide an in-depth, contextualised interpretation of the phenomenon of interest (Duff, 2008).An autoethnographic case study approach meshed well with my intention to embark on an exploratory investigation that might suggest directions and questions for future research and theorising about teacher motivation.
In the initial study, prior to the commencement of data collection, a participant information session was held.Informed consent was received from student participants and their caregivers.Data were collected about both regular curriculum activities and change-action activities through introspective journals (see Appendix A for instructions to students).Journals were chosen in order to afford insights into participants' experience of the learning environments yet also allow an efficient way of gathering regular data without obstructing curriculum content.Students wrote in Japanese, which was translated into English before analysis.Along with learners, I had also kept a research journal.Although there were additional data-collection tools (see Appendix B), the extracts presented in this article are drawn in the main from these journals.I made every effort to write in the research journal immediately after lessons or homeroom periods.I also occasionally wrote observations when I could find time during sessions.The length of my entries varied as certain incidents or reflections led to deeper understandings, resulting in 64 A4 pages totalling around 27,000 words.

Analysis
I was all too aware of criticisms of autoethnographic approaches as being the realm of "self-absorbed narcissists who do not fulfil scholarly obligations of hypothesising, analysing and theorising" (Ellis, Adams, & Bochner, 2010, para. 37).In conducting a re-analysis, I wanted to start afresh and build a detailed picture of teacher motivation based on a sound understanding of the data.In order to remain analytically open-minded as I read through the research journal, I applied a basic definition of motivation as "a cumulative arousal, or want [emphasis added], that we are aware of" (Dörnyei, 2009, p. 209).Initial analysis attempted to uncover instances of such "want" towards teaching practice.I used NVivo for Mac to look for repetitions and regularities in the text as I emphasised ideas that held significance for myself over the course of the teaching year (Ryan & Bernard, 2003).This first pass through the data engendered an expanded exploration as interactions between what I conceptualised as systems-motivation, affect, identity (often in the form of beliefs), and contextual and temporal elements-took on increasing importance.These categories were then microcoded to uncover vital qualities apparent from the data.Table 1 presents a selection of codes developed for each category.Theoretical comparison (Corbin & Strauss, 2008) of the ongoing analysis with concepts from the existing literature moreover suggested that a "person-incontext relational" perspective on motivation (Ushioda, 2009) and certain properties of complex systems offered a useful way of structuring the dynamic interactions between themes.Theoretical comparison is a tool with which "properties and dimensions that are derived from the 'outside' . . .give us ideas of what to look for in the data, making us sensitive to things we might have overlooked before" (Corbin & Strauss, 2008, p. 76).Through this process I became aware of interactions between the categories that seemed to "sit together" in a conceptual sense.I used the Boolean query feature of NVivo to examine the interrelations (or lack thereof) between categories and particular instances in the data and grouped related items into sets.The data items collected together into these sets were then compared with similarly grouped items a number of times, progressive iterations refining understandings at both "higher" and "lower" hierarchical levels.
Proponents of case study also urge the use of multiple data sources to gain different perspectives on the phenomenon of interest (Duff, 2008).The developing conceptual framework based on CST and realisations about particular events pushed me to revisit student data to examine it for ways that the interactions I was finding were illuminated.Due to the large amount of student data, a number of methods were employed to hone in on pertinent extracts.For instance, I used a simple text search for keywords such as teacher in the bank of student data.The analysis of the research journal had also suggested certain events as particularly influential on my motivation at points across the year.By examining corresponding student texts, I searched for ways in which the existing analysis could be enlightened and refined to further support or question my understandings.

Results and interpretations
In order to give a sense of the relational dynamics of teacher motivation/affect/identity, the following discussion presents a connected series of extracts.These extracts are representative of the macrothemes from analysis and allow a flavour of the uncovered "lower-level" themes adding qualitative details to those "higher-level" themes, whilst also providing a taste of the lived, narrative experience of teacher motivation in the classroom.Considering the research questions for this study, the first two sections below illuminate the contextualised nature of my teacher motivation (RQ1) through drawing on the CST properties of co-adaptation and openness.This said, nonlinear fluctuations in my motivation/affect/identity over different timescales are moreover evident throughout these first two sections (RQ2).The third section then more specifically addresses RQ2, drawing the discussion together by offering a number of different representations of the contextually-based yet relational development of motivation/affect/identity through nested timescales.In what follows, I mark extracts by week, type of session (English lesson or homeroom session), and source.Research journal is abbreviated to RJ, student learning journals to LJ.All student names are pseudonyms.

The contextualised nature of teacher motivation: Co-adaptive processes
A group of understandings about the emergent educational context was that continual processes of co-adaptation were evident in the development of my teacher motivation with Class E, as illustrated by the following series of extracts.In a homeroom session in Week 1 of the academic year, I introduced an activity for students to share ideas about their futures.This was my first opportunity to stand back and watch students interacting: Most students, from appearances, seemed to enjoy the chance to interact with others, and, even though they must hardly know each other, try to imagine about the future of other students.There were, of course, some students who hesitated to find a partner to talk to . . .but I was also pleasantly able to notice that in most cases other students came and tried to include these students.(Week 1, homeroom, RJ) At the timescale of this activity (a matter of minutes), my writing clearly evinces my own positive affect as I observed students engaging each other.This affect is also interacting at a longer timescale with my teacher identity in the form of expectations: Over the course of many "first lessons" with new groups of students, I appear to have built up certain expectations about hesitancy from Japanese students.This belief meets with negative feedback in my perceptions of student behaviour in Class E, interacting positively with affect.In turn, these impressions about the newlyformed class group interacted with my motivation, prompting me to adapt an activity in the English lesson the following week: "After how they took to the homeroom activity, I decided to make this into an activity where they mingled . . .Students looked really interested comparing experiences. . . .There was quite a lot of animated conversation" (Week 2, English lesson, RJ).
These two extracts reveal processes of positive feedback connected with student responses to my planned activities.Whilst it may be tempting to conceive that there is a causal chain between affect and motivation in the moment, I again believe that issues of identity are at play: My preference as a language teacher for communicative activities that encourage students to share their ideas in interaction has built up over a still longer timescale through my postgraduate studies.Lemke's (2000) use of the concept of heterochrony in CST is facilitative here, whereby "a long timescale process produces an effect in a much shorter timescale activity" (p.280).An extract focusing on a more negative experience the very next lesson also uncovers the way in which the co-adaptation of my teacher motivation continued to emerge across a lesson series: Who knows how things will turn out now? My class were just on another planet today.In the end I had to stop them, and have a 'stern word' or two.So now, of course, I feel terrible . . .Who knows what dynamics came together to create that, what, mess?Looking around as they were making their business cards, only about half of them seemed to actually be getting into the spirit of things . . .Then there was a mingling activity, using these business cards to meet 'co-workers.'Once again, it seemed as if precious few students were really getting into things. . .(Week 3, English lesson, RJ) This event appears to have a critical impact on my motivation.I begin the entry by focusing not on the activity of students, but on the outcome of having "to stop them, and have a 'stern word' or two."The action of chastising the students for their behaviour seems to run against my identity as a teacher, making me "feel terrible."At the timescale of the lesson, I define the whole session as a "mess."The experience has a crushing effect on my ideas of future lessons with Class E, leading me to worry about "how things will turn out now." My concern appears to come from the perception that the kind of communicative activity that had worked so well up to this point was met by "precious few students . . .getting into things."In CST terms, the positive feedback I had been receiving up to this point in my interactions with the student group (over a timescale of weeks) was abruptly replaced by negative feedback that something about this activity on this day with Class E did not foster the kind of learning environment I had been anticipating.In fact, despite the negative affect I attached to these experiences at the time, examining student reflections about the same event allowed a different perspective on my action of "scolding" the students: Today everyone wasn't concentrating on studying seriously, so we made the teacher angry.I'd been talking about some completely unrelated topic with a friend, so I need to think about my attitude as well. . . .I want to concentrate more seriously in lessons from now on.(Hide, Week 3, English lesson, LJ) Hide's writing reveals his understanding that perhaps a behavioural line had been crossed by a good number of students: "Everyone wasn't concentrating on studying seriously."Of the 40 students, 17 made reference in some form to this event and my reactions as a teacher in their journal writing.As evident in Hide's reflection, many students clearly recognized my affect in being "angry," with some students alternately interpreting that I "looked sad" when I felt it necessary to interrupt the lesson.However, far from the worries that I expressed for future lessons with Class E, student writing suggests my response to have had a critical dampening function.It allowed students a chance to reflect on their actions and retrain their own motivation towards future lessons.Another prominent theme related to students' drive to coform a more positive classroom environment: "I want to work together to make the class atmosphere a good one"; "I want to create a classroom environment where if one of us thinks, 'Hey, that's playing around a bit too much,' then we can tell each other."Taking the student perspective gives key insight into the ongoing processes of co-adaptation in class groups: The classroom I would walk into the following week would be a different space.
An entry in the research journal provides evidence of my recognition of this change.
Well, that was great.I don't know whether it was the students that changed, or me (probably both), but this week's lesson was so incredibly better than last week's. . . .The activities (for example, an info-gap to find out and imagine about a product; and getting put into new teams -new seating -and introducing themselves to other team members) seemed to go well, and many more students were trying to use English this week . . .I'm also getting a feel for the types of activities that might go well, and the need to repeat phrases over and over, and review things they've learnt, so that it sticks.I'll get to work on revising some upcoming lessons, where I've realised some things I had planned probably wouldn't go well.(Week 4, English lesson, RJ) My reflection in some ways provides a conclusion to the motivational stanza that continued over this timescale of a lesson series.As discussed in the review of literature, Hiver's (2015) study revealed that the way in which teachers adapt to experiences in their daily practice and accept these adaptations as part of their identity influences their accumulated levels of negative affect and stress.As I came to terms with the (English) ability level of students in Class E, I adapted my teaching approach to "repeat phrases over and over, and review things they've learnt, so that it sticks."In alignment with research finding an intrinsic nature to teacher motivation (see, e.g.Erkaya, 2013;Pelletier et al., 2002), my reflections give the impression that my confidence as a teacher was renewed through my perception that "the activities . . .seemed to go well, and many more students were trying to use English this week."In fact, research by Martin (2006) into teacher affect revealed a strong connection between teacher enjoyment and teachers' perceptions of students proactively working to improve their ability and knowledge.My writing similarly suggests the crucial importance of my perceptions of students putting in effort to improve their English ability.On a still longer timescale, I understand the extract as also suggesting processes through which my identity as a competent teacher has likely built up through repetition of numerous similar classroom experiences.I link my teaching adaptations to the outcome that the lesson was "so incredibly better than last week's."However, rather than presuming this to be a linear result of my own adjustments, an implicit recognition of relational co-adaptation is also revealed in my puzzlement as to "whether it was the students that changed, or me (probably both)."

The contextualised nature of teacher motivation: Openness
Already evident in the discussion of co-adaptation, the open and blended nature of my teacher motivation/affect/identity interacting with context emerged as a strongly recurrent theme throughout the research journal.While focusing on my experiences in the classroom with Class E, my motivation/affect/identity was also revealed to be open to interactions with contextual and temporal elements "external" to the present classroom system.In an entry a couple of weeks later, I begin by introducing my experiences in the lesson not by writing about Class E, but another class: Actually, I taught the same lesson the day before, and it was just terrible -probably a combination of it being the very first day back after Golden Week, and it not being a very good lesson plan (as it turned out).Even I came away from the lesson feeling that students probably hadn't gotten much from it, which made me really sorry for them (and I wonder what they will think of me as a teacher after that experience!).Anyway, I spent a restless night.In the end it got pretty thoroughly revised.(Week 6,English lesson,RJ) This extract suggests that my motivational system is open in a number of ways: Firstly, I ascribe the teaching failure in part to the effects of a break due to the extended Golden Week holiday period in Japan. 1 In fact, such temporally-based negative influences on my motivation, for example from the timing of other subject lessons or breaks imposed because of national holidays or college events, were a recurring theme in the analysis.I perceived that such perturbations impeded student progress and ability to concentrate on material and lesson activities, which in turn affected my motivation as a teacher.Analysis of student data also frequently turned up comments about the negative effects of similar occurrences, suggesting my fears to be well-founded.Secondly, as in this example, I often referred to experiences in other classes.In the extract above, I attribute a further reason for the poor lesson with the other class as due to the lesson plan.A study by Menyhart (2008) into the motivation of EFL teachers at the university level in Hungary revealed high levels of stress associated with what these teachers perceived as unsuccessful lessons.My writing clearly reveals strongly negative affect as the "terrible lesson" which made me "sorry for [those students]" resulted in the physical response of a "restless night."These experiences and the resultant affect however also motivated me to create a "thoroughly revised" plan for the lesson with Class E. Another area of interest is my concern as to "what [the students] will think of me after that experience!"This statement again reveals my identity as a teacher, built up over years, to also be highly dependent on my perceptions of momentary experiences with students and affective understandings of lessons.It further implies that one motivation to work to make what I hoped would be a more successful lesson with Class E was an effort to move away from this image of failure towards a more ideal teacher-self (Dörnyei & Kubanyiova, 2014).
In the end, one of the changes that I made was to increase the opportunities for students to interact and discuss during activities.As I continued in my entry about the lesson with Class E: Most students were using Japanese to discuss their opinions, and writing in English.However, Kazu tried everything in English! . . .Furthermore, for a while, the feeling spread to some of his other teammates.Initially they were speaking Japanese, and looking a bit oddly at him, but then some tried also in English.I wish this could happen in other classes! . . .Anyway, it all worked a lot better than had the previous day's lesson!(Week 6, English lesson, RJ) I understand hints of the CST property of nonlinearity in this extract.While I begin by focusing on "most students," I then remark on the actions of one particular student.My writing reveals a clear picture of affect interacting with motivation as this student and some other students took advantage of the more interactive nature of the revised lesson, allowing me to feel a sense of contentment that "it all worked a lot better than had the previous day's lesson!"The quantitatively small influence of this one student had a qualitatively critical effect on my impression of the lesson.Also evident is the way in which I make a connection this time between my experiences in Class E with a desire for similar outcomes in other classes that I was teaching: "I wish this could happen in other classes!"All in all, these extracts give a sense that my teacher identity and motivation, negatively influenced by experiences in another class, recover through the positive feedback I receive from Class E about the adaptations to teaching that I implemented.
In conjunction to these regular curriculum activities, I was moreover conducting action research to assist students to reflect on their English language learning.Particularly near the start of the year, I felt frustration in trying to introduce change-action: As things are going, there is just so much time taken with boring surveys or menial information I have to pass on to students . . .And so this makes me frustrated for my research, and for the possibilities of doing something worthwhile with these students -I kind of feel like chances are slipping away . . .(Week 7,general reflection,RJ) This extract starkly evinces the way in which I held strongly negative affect towards contextual responsibilities that I had not accepted as part of my professional identity.There was a clash between interacting and merged teacher and researcher identity systems.This disdain also however illuminates my motivation to conduct research beneficial to students to be a developing yet essential aspect of my professional identity as I refer to "chances . . .slipping away" for "doing something worthwhile [emphasis added] with these students."Frustration arises through the impediments I perceive to working towards this ideal professional identity to such an extent that I belittle what were in fact vital administrative duties in my role as a homeroom teacher.
Nevertheless, while these concerns peaked on occasion, I gradually became more capable in making time to incorporate change-action.I made use of my identity as a fellow additional language learner a couple of weeks later.I showed students a series of entries from a Japanese diary I had kept: I'd been doubtful that they would be interested, but I think the entries I showed them expressed my frustrations with studying Japanese, yet also showed a clear progression (and had a running theme of me hating boring meetings, which many laughed about), so overall there seemed to be a really good atmosphere after I'd shown this -like they had gotten a bit closer to the teacher. . . .(I'm really looking forward to reading the comments in the LJs about this activity.Are things like this useful to them?).(Week 11, English lesson, RJ) Ushioda (2011) argues convincingly for the motivational benefits of interacting with students "as 'people' rather than as simply 'language learners'" (p.17).While I initially express trepidation about how students would greet this activity, my writing suggests that my positive perceptions of learner engagement with this aspect of my identity foster motivation towards creating similar activities.Student reflections on this activity uncover the fact that my impressions were not far off the mark: The teacher's story was deeply interesting.I thought, even something that you don't understand or can't do at all, if you repeat and keep trying little by little, you'll be able to do it in the end.I thought I want to study English not just during lessons, but outside of class too.(Kosuke, Week 11, English lesson, LJ) These extracts show my motivational system to be open through the way in which genuine interactions with students utilising dimensions of my identity from "outside" of the classroom-in this case as a fellow additional language learner-met with positive responses from students.Such instances of me drawing in my "transportable identities" (Zimmerman, 1998) to interact with students held somewhat of a risk, by exposing aspects of my self that were in many ways private and not a regular part of my teacher identity.Additional instances also suggest the way in which teacher motivation is both situationallygrounded yet also relationally dynamic over longer timescales: My impressions are of student responses to my transportable identities during an activity; these perceptions gain (positive) feedback a couple of weeks later through student journal writing.Yet, the dimensions of my identity upon which I drew for these activities (such as my identity as a second language learner or past dreams I held for my future) had evolved over my time in Japan or even longer timescales.

Motivational development: Relational motivation over interacting timescales
Analysis found motivation/affect/identity to be continuously reforming in the space between context and the individual.It is relational: I was not merely acting and reacting but acting and reacting in co-adaptive interrelation with others in the class group.The evolving interrelations altered ideas of what was possible for me as a teacher in the social context of the particular class group, as well as my ongoing "identity project" (Lemke, 2000) as teacher and as a person in general.
As evident in the previous sections, these interactions occurred over different timescales.How might a CST approach facilitate understandings of the dynamic, relational nature of my motivation/affect/identity apparent in the analysis?One promising method is revealed in Yashima and Arano (2015).These researchers took a sociocultural approach to interpreting the motivation of language learners.They represented coded experiences from student interviews over three sociocultural domains: contingent experiences, personal meaning, and deeply internalised values or thought patterns.Adapting their representative tool, in Figure 2 I display a selection of coded references to teacher motivation/affect/identity emergent over the year of data collection.Rather than the sociocultural "levels" employed by Yashima and Arano (2015), through analysis I understood my reflections as referring to different interacting timescales.The analysis further allowed me to discern valence and intensity from motivation/affect references (see Appendix C for examples of valenced references).On a 5-point scale, Figure 2 shows these fluctuations of motivation/affect in interaction with the different timescales of experience.Looking vertically at Figure 2 fosters a recognition that motivation/affect/identity, for example at the timescale of a classroom activity, interacts with contextual events and is further interrelated with longer timescales such as impressions of a lesson in general or motivation to adapt over a lesson series.Although perhaps difficult to take in at a glance, the horizontal axis also allows a glimpse of some varying dynamics of my motivation/affect/identity over the time window of the study.For instance, in strong alignment with the past findings of Martin (2006), my motivation/affect/identity was in constant interaction with perceptions of student attempts to improve ability, their proactive planning and persistence in learning.Similarly, occasional incidents involving my role as a homeroom teacher when there was some kind of disciplinary trouble remained a constant source of strongly negative affect (evidenced by the negatively-valenced spikes on the intensity graph) as I struggled to incorporate these responsibilities into my image of professional identity.Conversely, it is also possible to observe (through absence) the way in which frustrations with the research process decreased.I became more capable of making adaptations to teaching that allowed the introduction of change-action, but furthermore, the context simply allowed more freedom as time passed.
What is apparent from the previous discussion and Figure 2 is that the analysis suggested my motivation/affect/identity to be highly contingent on perceptions of student action at both an individual level and as a whole class group.In this respect, the CST literature also suggests a useful tool for visualising the evolving, relational form of my teacher motivation/affect/identity: multiple threading (Davis & Sumara, 2006).Multiple threading is a way of representing the "many strands" involved in a single group narrative, whilst recognising that "some may be only brief phrases or single images that punctuate the text, and strands may overlap or interlink at times" (Davis & Sumara, 2006, p. 162).By searching my research journal for mentions of the human focus of the development of my motivation/affect/identity, I was able to build a picture of its spread across perceptions of experiences with the members of Class E, which is shown in Figure 3.  (Davis & Sumara, 2006) representation of interactions between perceptions of student actions and teacher motivation/affect/identity across a year of study As discernable in Figure 3, there is positive and negative valence connected to individual students and the class group as a whole ("They" in the bottom row) across the timescale of the academic year.In terms of individual students, only 32 of the 40 students in Class E were mentioned explicitly in my research journal.As far as can be discerned from the data, some of these students appear only briefly to impact on my motivation/affect/identity, whilst others provide a more sustained contribution to the narrative over a longer timescale.Similarly, my perceptions of some students appear to have a relatively consistently valenced pattern (e.g., positive: Kazu, Yusuke; negative: Tomoe, Kanata) or more fluctuating relationship (e.g., Koji, Haruki) with my motivation/affect/identity.It can also be seen that at the level of the class group at the timescale of the year of study my motivation/affect/identity settles into an increasingly positive pattern, even while at times I have negative motivation/affect/identity connected with individual students.Such multiple threading is one useful tool to begin to understand the way in which teacher motivation is intricately intertwined with experiences in particular class groups, and how perceptions in interaction with students build up relationally over nested timescales.

Conclusion
This exploratory study focused on the situated motivation of myself as a teacher in the specific context of one language learning class group.As such, it is certainly limited in its focus on only one teacher and its self-reflective nature.There was no use of peer checking by other researchers to re-analyse data.The themes are restricted to being the subjective understandings and interpretations of one researcher.Moreover, the initial research from which data were drawn was not set up to study teacher motivation.However, by employing a case study approach to examine in depth the experiences and perceptions of myself as a teacher and triangulating where possible with student data, I was able to build a picture of EFL teacher motivation arguably more detailed and dynamic than previous studies.The study also suggests profitable directions for future research: A more refined design could specifically focus teachers on writing reflectively about their motivation/affect/identity while students are also encouraged to write perceptions of the teacher each lesson.Video recording of lessons would contribute an additional perspective on these introspective accounts.
The study does however indicate that CST concepts might provide a much more satisfying range of possibilities to explore and understand the "messiness" of a person-in-context relational view of motivation.By focusing on a particular context and examining data longitudinally, the analysis revealed the very situated and dynamic nature of teacher motivation.It suggests that while teacher motivational trajectories develop their own character in interaction with members and experiences in classrooms, motivation, affect and identity also interact with elements conceptualised "outside" of any particular class group.Perceptions of these interactions build up to form a consolidated motivational/affective/identity trajectory that is contextually bound to a class group.My study intimates that further intriguing insights into the motivation of teachers could be forthcoming from similar inquiries that examine the lived experience of teaching over time.The contextually-based yet open nature of teacher motivation also hints at the fascinating possibilities of investigating the development of motivation/affect of the same teacher with different class groups.Finally, building on Hiver's (2015) research into teacher immunity, the study suggests the usefulness of encouraging teachers to reflect on and become conscious of the ways in which teacher motivation/affect/identity adapts and evolves in-situ over still longer timescales.

Figure 1
Figure 1 Some of the interacting timescales relevant to teacher motivational development with particular class groups (adapted fromDavis & Sumara, 2006)

Figure 2
Figure 2Longitudinal fluctuations in valence and intensity of motivation/affect in interaction with different timescales evident from analysis (* The college summer vacation period continued from late July to early September; ** the college winter vacation period continued from late December to early January.)

Figure 3
Figure 3Multiple threading(Davis & Sumara, 2006) representation of interactions between perceptions of student actions and teacher motivation/affect/identity across a year of study

Table 1
Category and code examples