Motivational Dynamics in Language Learning: Change ...

Motivational Dynamics in LanguageLearning: Change, Stability, andContextFREERKIEN WANINGEUniversity of NottinghamSchool of EnglishTrent Building, University ParkNottingham NG7 2RDUnited KingdomEmail: [email protected]

ZOLTAN DORNYEIUniversity of NottinghamSchool of EnglishRoom A52 TrentUniversity ParkNottingham NG7 2RDUnited KingdomEmail: [email protected]

KEES DE BOTRijksuniversiteit GroningenDepartment of Applied LinguisticsFaculteit der LetterenOude Kijk in ’t Jatstraat 269712 EK, GroningenThe NetherlandsEmail: [email protected]

Motivation as a variable in L2 development is no longer seen as the stable individual difference factor itwas once believed to be: Influenced by process-oriented models and principles, and especially by thegrowing understanding of how complex dynamic systems work, researchers have been focusingincreasingly on the dynamic and changeable nature of the motivation process. In this study we micro-map the motivational dynamics of 4 language learners during their language lessons over a period of2 weeks, using a novel instrument—the Motometer—combined with classroom observations and aquestionnaire on motivation and attitude. The article answers three current questions concerningL2 motivation: (a) Can we demonstrate variability in students’ L2 motivation in class; (b) Is there adetectable stable level of students’ in-class motivation; and (c), If both of these are demonstrated, canthey be accounted for by the classroom context? The results affirm that student motivation can besuccessfully explored using a dynamic systems framework. Our findings demonstrate how motivationchanges over time on an individual level, while also being characterised by predictable and stable phases,and how it is inseparable from the learner’s individual learning context. The data also show thatmotivation can be meaningfully studied at different interacting time scales.

Keywords: motivation; emergentism; foreign language (FL) learning; classroom-based research

OVER THE LAST DECADE, APPLIED LINGUIS-tics has seen an intriguing dynamic turn in researchapproach as an increasing number of scholars

started to adopt aspects of a complex dynamicsystems perspective on language development.Whether this was done directly under the headingof the related paradigms of Emergentism,Complexity Theory, or Dynamic Systems Theory(e.g., de Bot, 2008a; Dornyei, 2009b; Jessner,2008; Larsen–Freeman, 2006; Larsen–Freeman &Cameron, 2008; Mercer, 2011a; Verspoor, Lowie,& VanDijk, 2008), or without explicitly adopting a

The Modern Language Journal, 98, 3, (2014)DOI: 10.1111/j.1540-4781.2014.12118.x0026-7902/14/704–723 $1.50/0© 2014 The Modern Language Journal

dynamic systems framework (e.g., Ahmed et al.,2011), the overall emerging picture is that of abroad shift that has departed in several respectsfrom traditionally established viewpoints. One ofthe most radical consequences of this paradigmshift has been the growing recognition thatstraightforward cause–effect relationships are nolonger sufficient in themselves to explain all thecomplex patterns observed in SLA data. Whilethis does not mean that systems are always in astate of unpredictable flux—we can often identifyrelatively stable phases and recurring patternswithin the variation of system behaviour—thelesson of related studies has been that linearpredictability cannot be taken for granted as adefault. This conclusion questions some of themain tenets of statistics-based quantitative re-search methodology, and the situation is furthercomplicated by a second feature of the dynamicconceptualisation of systems, namely that thecontext in which system behaviour occurs is nowseen as part of the developing system rather than adistinct background variable. By definition, thischaracteristic increases the social sensitivity ofresearch conducted in a dynamic systems vein;at the same time it also adds further layers tothe research paradigms necessary to study thecomplex interaction of language, agent, andenvironment. Consequently, as Dornyei (2009b)concludes, “operationalizing this dynamic rela-tionship in specific theoretical and measurementterms takes us into rather uncharted territories,with few specific guidelines or templates currentlyavailable to follow” (p. 244).

One of the seminal papers that put the complexdynamic approach on the SLA research map wasDiane Larsen–Freeman’s (2006) study of the L2development of five Chinese learners of Englishin terms of accuracy, fluency, and complexity.Larsen–Freeman demonstrated convincingly thatan analysis that was based merely on groupaverages was likely to hide important individualpatterns of learner development that came to thesurface only when an individually focused andcontextually sensitive methodology was applied.The present article follows a similar agenda byfocusing on individual learner characteristics,specifically on language learning motivationsituated in a classroom setting. Through thecombined application of novel methodology thattakes online measurements of the learners’motivational disposition on a micro level andquestionnaires that focus on the macro level ofthe learners’ overall motivation and attitudes, weexplore the nature of three key aspects of thedynamics of motivational development: change,

stability, and contextual dependency. In doing so, weintend to demonstrate that some aspects ofmotivational variability can only be made propersense of if we apply a complex dynamic systemsframework.

THEORISING COMPLEX DYNAMIC SYSTEMS

Even a quick scan of the field of appliedlinguistics produces a picture of the growingawareness of the significance of systemic, non-linear patterns in SLA data. After Larsen–Free-man’s (1997) pioneering study on chaos/complexity theory, three special issues of leadinginternational journals highlighted the topic (deBot, 2008b; Ellis & Larsen–Freeman, 2006;Hawkins, 2008), followed by two books (Larsen–Freeman & Cameron, 2008; Verspoor, de Bot, &Lowie, 2011) and several articles on systemsphenomena (for a recent overview, see Larsen–Freeman, 2012). In view of this burgeoningliterature, there can be little doubt that theapplication of a dynamic systems perspective hasresonated with the SLA research community.While the specific paradigms that scholars haveapplied in their investigations show variationaccording to which feeding discipline they originat-ed in (e.g., mathematics, biology, developmentalpsychology, or philosophy), the various rubricsemployed—Chaos Theory, Emergentism, Com-plexity Theory, and Dynamic Systems Theory—represent similar, although not identical, strands,whose overall aim is to view phenomena in aholistic and systemic manner by acknowledgingthe interrelated nature of the components withinthe system and the often nonlinear manner ofself-organisation and emergence by which thesystem evolves and responds to both external andinternal stimuli. For simplicity’s sake, we will referto these strands under the umbrella term ofDynamic Systems Theory (DST) (for a discussionof how the various strands differ from each other,see Dornyei, 2009b). In order to understand someof the main principles and potential benefits of adynamic systems perspective, let us look at threecore characteristics of system dynamics: change,stability, and context.

Change

One of the central features of a dynamic systemis its continuously changeable state. The wordstate refers to the position of the system—that is,the object under study—at a given moment, suchas a student’s level of motivation or the numberof words in a student’s vocabulary. A system is

Freerkien Waninge et al. 705

considered dynamic when it has at least two ormore key elements that are interlinked with eachother but which also change over time. DSTemphasises the significance of this temporalvariation and evolution, highlighting the factthat the observed change can be nonlinear (forrecent discussions, seeMacIntyre&Legatto, 2011;Mercer, 2011a; Spoelman & Verspoor, 2010; VanDijk, Verspoor, & Lowie, 2011). Nonlinearityrefers to the phenomenon whereby a system’schange in output is not proportional to thechange in received input (e.g., an increase inexposure to the L2 does not lead to an equalincrease in proficiency); it follows from this that innonlinear systems there are no automatic andpredictable cause–effect relations governing thesystem’s behaviour: A huge amount of input cansometimes have little or no impact, whereas atother times, to take the other extreme, a minimalincrease in input might have an unexpectedlylarge effect in propelling the system forward, aphenomenon sometimes called the butterfly effect.A consequence of this nonlinearity is thatresearchers should expect to find unpredictablechanges in a system’s state that cannot easily beattributed to any specific single cause.

Stability

With the emphasis on variability and nonlinearchange in DST, one might rightly ask how it ispossible at all to study a dynamic system. Whyformulate a research question when the systemunder study might show nothing but unpredict-able behaviour? This is where a second principleof DST becomes important: Despite the emphasison change and ongoing system dynamics, DSTalso recognises stable states in system behaviour.Dynamic systems are known to self-organise, asa result of which they can settle into preferredstates—referred to as attractor states—during theirdevelopment. Interestingly, some behaviouraloutcomes of the system’s self-organisation processare so stable that they seem to be programmed orhardwired; an example is that most children learnto walk and speak a language up to an advancedlevel (Thelen, 2005). The important point from aresearch perspective is that it is this stability of thesystem that allows salient outcomes and patternsof development to become visible (Mercer,2011b; Van Geert, 1998). While in theory thepossible developmental patterns for a system canbe endless, in reality the system’s self-organisingfunction results in a limited number of outcomes;that is, “even very complex systems tend to arrive

at certain salient outcomes, and although wecannot predict in advance what these outcomesmight be, when we see them we recognise them”(Dornyei, 2014, p. 85). An example of such alimited number of outcomes that most teacherswill be familiar with is the existence of differentstudent types in every classroom (i.e., theemergence of a limited number of studentprofiles that both teachers and students willrecognise as typical).

Context

As mentioned earlier, contextual factors canplay such a prominent role in pushing or pulling asystem toward or away from a certain state thatsome of them cannot be meaningfully separatedfrom the dynamics of the whole system and, assuch, form an integral part of the system. Forexample, if we look at the learning environmentof a language classroom and focus on thebehaviour of individual learners, their perfor-mance will be affected by various layers ofcontextual influences such as the behaviour offriends and classmates, the constraints of theclassroom space, or the leadership functionsexercised by the teacher. Thus, as explained byVerspoor et al. (2008), the learner and theenvironment are not independent of one anotherbut rather influence and change each other,leading to systemic variability in development.Our challenge in researching the learner and thecontext is, thus, to “consider simultaneously theongoing multiple influences between environ-mental and learner factors in all their componen-tial complexity, as well as the emerging changes inboth the learner and the environment as a resultof this development” (Dornyei, 2009a, p. 244,original emphasis). This mutual influence oflearner and environment implies that, in orderto understand the development of a dynamicsystem, we need to examine the phenomenonunder study in its naturally occurring context,be it a structured ESL classroom (Larsen–Freeman, 2006), or a more unpredictable kinder-garten playground (Van Geert & Steenbeek,2005).

MOTIVATIONAL DYNAMICS

As well as having influenced the field of SLA as awhole, a complex dynamic systems approach hashad significant influence on the understanding ofL2 motivation. This understanding has under-gone a transformation over the past decade; thedays in which motivation was seen as a static

706 The Modern Language Journal 98 (2014)

individual difference variable within a modularframework—the most famous of which beingRobert Gardner’s (1985) socio-educational model—seem to belong to the past. As soon as motivationcame to be seen as a situated construct duringthe educational shift in the 1990s (e.g., Crookes& Schmidt, 1991; Dornyei, 1994; Williams &Burden, 1997), it was only a matter of time beforeone would arrive at the unavoidable realisationthat motivation also has a prominent temporaldimension (i.e., it displays ongoing change). Thefirst step toward a more dynamic conception ofL2 motivation was the introduction of variousprocess models. Examples of frameworks depictingmotivation as a process include Williams andBurden’s (1997) theory, in which three differentstages of motivation were distinguished: (1) thereason for doing something, (2) deciding todo something, and (3) persisting in doingsomething. Similarly, Dornyei and Otto’s (1998)process model of L2 motivation differentiatedbetween three phases in the development of astudent’s L2 motivation: (1) a pre-actional phase,or choice motivation, in which motivation to initiatean activity is formed, (2) an actional phase, orexecutive motivation, in which the initial level ofmotivation has to be maintained during thelearning activity, and (3) a post-actional phase,or motivational retrospection, in which the process isevaluated and lessons are drawn that affectsubsequent motivation.

In hindsight, these frameworks were still basedon linear cause–effect relationships—althoughthe Dornyei–Otto model also included severalfeedback loops—and it was gradually recognisedthat to account for the real dynamics of theL2 motivational process requires a more radicalreformulation thanmerely designing increasinglycomplicated patchworks of interwoven cause–effect relationships. A highly instructive parallelin this respect has been offered by the evolution ofemotion research, as exemplified by the transfor-mation of the thinking of one of the leadingscholars in the area, Klaus Scherer. Originallyinterested in the componential structure ofemotions (for a review, see Scherer, 2001),Scherer became acutely aware of the changingnature of these components over time and calledfor the abandoning of “static state concepts,”suggesting instead that scholars “move from adomain oriented approach to a process orientedapproach” (Scherer, 1993, p. 5). However, theinterlinking of different functional systems,involving cognitive, affective, and motivationalaspects, pushed his conceptualization one stepfurther:

Unfortunately, neither our conceptual nor ourmethodological tool kits are adapted to dealingwith systems of the degree of complexity exhibitedby emotion processes. There is little hope of‘repairing’ our concepts and methods in a piecemealfashion in order to do justice to the phenomenonunder study. Rather, we need a complete revolutionin our thinking about the nature of emotion,comparable to other paradigm shifts in the historyof science. In particular, we need to move fromthinking in terms of discrete boxes, labels, or evenneural programs to a nonlinear dynamic systemsperspective of emotion. (Scherer, 2000, pp. 77, 80)

In view of the above considerations, a dynamicsystems approach seems attractive in that it is ableto accommodate explanations of both variabilityand stability without relying on cause–effect links.Accounting for fluctuation is a prerequisite toany further advance in our understanding of L2motivation because of the salience of the changeofmotivation over time, the first of the threemainfeatures of a dynamic system as mentioned above.This change appears frequently in recent L2motivation research. Hotho (2000) for examplefound that even when a learner’s overall L2motivational profile remained relatively stableduring a semester, short term motivation wassusceptible to considerable change: “[T]heteachers may sense changes or fluctuations fromweek to week, as a result, perhaps, of a piece ofhomework, a test, or a class that was particularlyinspiring or noninspiring” (p. 326). Similarly,Pawlak (2012) and Poupore (2013) traced thedevelopment of groups of learners for theduration of several classroom hours, and they,too, highlighted the dynamic nature of short-termmotivation. Other investigations have docu-mented long-term tendencies in motivationalevolution, usually evidencing some decline inlevels ofmotivation during the course of extensiveinstitutional engagement (e.g., Chambers, 1999;Dornyei, Csizer, & Nemeth, 2006; Gardner et al.,2004). These studies seem to suggest that motiva-tion may fluctuate at different time scales thatrange from minutes to hours, days, months, oryears. These time scales interact: What happenson the minutes scale has an impact on whathappens on higher time scales and the other wayaround (for a discussion on time scales inlanguage development, see de Bot, 2012).

On the other hand, the developmental varia-tion observed in student motivation has alsoincluded occasional stable periods or predictablepatterns, the second of the three main features ofa dynamic system. For example, Koizumi andMatsuo (1993) found that after a sudden decrease


in motivation following the transition to juniorhigh school, students’ motivation settled intoan equilibrium. In a qualitative study examiningmotivational evolution over the lifespan, Shoaiband Dornyei (2005) elicited a number of recur-ring temporal patterns and key transformationalepisodes affecting motivation, including for exam-ple transitions to new life phases (such as leavingschool and entering the world of work) or theexperience of visiting an English-speaking envi-ronment. To cite a final study that examinedtemporal variation and stability, this time in L2learners’ willingness to communicate (WTC),MacIntyre and Legatto (2011) found not onlyvariability in learners’ willingness to speak, butalso a stable level, or attractor state, when thesurrounding systems supported a learner’s WTC:

WTC shows the properties of a dynamic system. Wesee that there are changes over time wherein eachstate is partially dependent on the previous state. Wealso see the interconnectedness of the linguistic,social, cognitive and emotional systems that produceWTC. When the systems function together tofacilitate communication, we see WTC as an attractorstate. (MacIntyre & Legatto, 2011, p. 169)

Regarding the third of the three main featuresof a dynamic system—the importance of contextin motivation research—advances have gonehand in hand with the more general social turn(Block, 2003) in SLA research that has charac-terised the past fifteen years (for a review, seee.g., Zuengler & Miller, 2006). The best-knownmotivation theory in this respect has beenUshioda’s (2009) Person-in-Context RelationalView of Motivation, which, as the name suggests,highlights the agency of second language learnersconceived as individuals who are located inparticular cultural and historical contexts andwhose motivation and identities shape and areshaped by these contexts. This environmentalimpact has recently been confirmed, for example,by Campbell and Storch’s (2011) study, whichexplored motivational fluctuation and changeamongst university students learning Chinese as aforeign language over a semester. The authorsfound that factors related to the learningenvironment were the most important variablesto impact motivation both in a positive and anegative sense (i.e., both motivating and demo-tivating). Furthermore, the dynamic relationshipbetween learner and context was also confirmedby the finding that when learners had developed aclear future image of themselves as speakers of theL2, they were able to stabilise their motivational

state, even in the face of demotivating contextualfactors, by means of conscious self-motivationstrategies.

In sum, the situated conceptualisation of L2motivation and the subsequent process modelsproposed in the 1990s launched an ongoingprogression of thinking that led to dynamicconceptions of motivational change. As a result,metaphorically speaking, the issue of motivationaldynamics is currently in the air as one of the mainthemes to explore, as attested to by a high-profilecolloquium under the same title at the 2013annual conference of the American Associationfor Applied Linguistics and a subsequentanthology edited by the colloquium organisers(Dornyei, MacIntyre, & Henry, 2014).

The question then remains: How does onestudy motivation from a dynamic perspective?The present study aims to illustrate the threeaforementioned characteristics of a dynamicsystem, that is to say change, stability, and context,in language learners’ motivation. The researchquestion is therefore threefold:

We endeavour to answer these researchquestions by utilising a range of data gatheringmethods and data types in a longitudinalclassroom-oriented investigation that examinedthe motivational variability of young secondarypupils in two different language courses, Germanand Spanish.

METHOD

Study Design

In order to allow for intensive, individual-levelmicroanalysis, four students (two males and twofemales) participated in the study. They wereselected from a class of 28 students in their firstyear of the highest level of Dutch secondaryeducation (referred to as pre-university plus, ages11 to 12). A limit of four students was decided onto facilitate classroom observations of the parti-cipants. The number was thought to be sufficientsince the aim of the study was not to study groupaverages but rather to focus on individual levelanalysis. The motivational variability of these

RQ1. Is there variability to be found in students’in-class motivation?

RQ2. Is there a detectable stable level, orattractor state, in students’ in-classmotivation?

RQ3. If there is indeed variability and stability instudents’ motivation, can this be accountedfor by the classroom context?


students wasmeasured over a period of two weeks.This two-week time span suited the purpose of thestudy, namely micro-inspection of classroommotivation rather than changes over a longerperiod of time, considered the practical con-straints of instruction, such as exams and bankholidays, and was deemed to be broad enough totap into themicro-variation ofmotivation within aclassroom session. Measurements of motivationwere taken at five minute intervals during sixlanguage classes: three sessions of 45–50 minutesin the Spanish class and three sessions of 45–50 minutes in the German class. Classroomactivities were observed and documented by theresearcher present in the classroom for all sixsessions. After the last session, students filled out aquestionnaire on attitude and motivation towardboth language classes and teachers.

The target languages were both compulsoryforeign languages that are not typically taught atthe primary school level. In secondary educationin the Netherlands, German is commonly studiedin this age group, whereas Spanish is normally anoptional module. However, for this particulargroup of students, being taught at pre-universityplus level, both languages were compulsory in thefirst three years of secondary school. Both courseswere aimed at reaching the A1 level in theCommon European Framework of Reference in alittle under a year. The students received twohours of classroom instruction for each of the twolanguages each week, with the lessons typicallylasting about 45–50 minutes.

Students were taught by their regular languageteachers, both experienced teachers in theirforties, who had taught at the school for severalyears. Although their teaching methods aretypical in Dutch language classrooms, the teach-ers differed somewhat in the tasks they employed,the Spanish teacher being more traditional inhis choices whereas the German teacher tookmore risks by introducing games and role plays.Appendix A gives an overview of classroomactivities used across the six lessons.

Instruments

Three instruments were used to measure theparticipants’ motivation and to record ongoingobservations of the classroom environment: a‘Motometer’, a classroom observation form, anda motivation/attitude questionnaire. Followingthe example of a longitudinal classroom studyby Gardner et al. (2004), the Motometer wasintroduced to take a series of ten online (i.e.,real time) measurements of student motivation

during each classroom session of 45–50 minutes.1

Participants received an A4 size sheet of paperwith ten Motometers: thermometer-shaped fig-ures with a ‘0’ at the lowest and a ‘100’ at thehighest point. Students were instructed to indi-cate their level of motivation by drawing ahorizontal line on the next Motometer everytime they were prompted to do so. This prompt-ing was done in time intervals of five minutes bythe researcher present in the classroom usinga prerecorded soft bell sound. Motivation wasdefined as ‘Howmuch effort do I want to put intolearning the material right now’, and ‘How muchdo I enjoy this lesson right now’. On the bottomofeach page, a comments section allowed theparticipants to elaborate on their reported levelsof motivation. At the end of the lessons theMotometers were collected by the researcher and,if necessary, clarifying questions were asked toexplain striking variations in the data. An exampleof the Motometer can be seen in Appendix B.Whereas the original Motometer by Gardner andcolleagues (2004) was used to take a singlemeasure of students’ state motivation duringfour classes throughout the academic year, thecurrent adaptation is designed to tap into thesubtler variation of motivation within a lesson.The tool is straightforward and intuitive and,mostimportantly, can be used in the context of theclassroom without disturbing the natural progres-sion of the lesson. It is similar to Pawlak’s (2012)“motivational grid” (p. 257), which measuredmotivation with time intervals of fiveminutes on a1–7 scale.

In order to account for the contextual influ-ences of the language classroom in which themeasurements took place, an observation formwas used to keep a record of the unfolding lessonplan, the actual actions and tasks performed, andany retrospective comments by both students andteachers. Language activities were describedindicating their starting time and duration (toallow for matching them with the online motiva-tion measures), and the researcher paid specificattention to noting down details concerning thefour participants, such as interaction with theteacher and peers, disruptive behaviour, orvolunteering (raising a hand) to ask or answerquestions.

The four students’ overall L2 motivationalprofile was ascertained through a questionnairecontaining 27 10-point Likert scale items;it measured students’ attitudes toward bothlanguage courses, with scores ranging from‘completely disagree’ (1) to ‘completely agree’(10). Seventeen questions were related to the


students’ attitudes toward the course, includingseven items on enjoying the language classes ingeneral (e.g., ‘Sometimes at the end of Germanclass I would like to continue’), seven items onlinguistic self-confidence (e.g., ‘I feel like I’mmaking progress with German’), and three itemson classroom anxiety (e.g., ‘I’m afraid my class-mates will laugh at me when I make a mistake’).The last ten questionnaire items were relatedto the students’ attitude toward the languageteachers who taught them. To complete thestudents’ learner profiles, teachers reported onthe participants’ overall grades in both courses(German and Spanish) at the end of the semester,which involved composite scores of severalproficiency tests taken during the course. In thisschool, students take several tests varying inlength from 10 to 60 minutes throughout asemester. At the end of each semester, a student’sfinal grade for a course will be the average of thefour to five tests taken. These tests focusmainly onwritten translation from and to the targetlanguage and reading and listening comprehen-sion, skills that are also the main focus of theclassroom sessions and are typical in Dutchlanguage classrooms at this level.

Participants

After four lessons of classroom observation bythe first author and a brief consultation with bothlanguage teachers, four students were asked toparticipate in the study. The main selectionprinciple was to have a mixture representingthe class composition: Student 1, Boris,2 is whatmost teachers would describe as a good pupil: Heusually pays attention and generally knows theanswer to the teacher’s questions. Student 2,Connor, is easily distracted but asks for clarifica-tion when something is not clear to him. Student3, Nina, is actively participating, needs a lot ofconfirmation and therefore asks many questions(unfortunately, during the investigation she had abrief case of the flu that made her miss someclasses). Student 4, Irene, has mastered the art ofdoing other things in class without being noticed.She pays attention when necessary, but does notparticipate unless asked to. Table 1 offers anoverview of each participating student; Figure 2 inthe Results and Discussion section adds to thischaracterisation by presenting summaries of theattitudinal scale scores of the questionnaire data.

Procedures and Data Analysis

The first lessons observed in both subjects weretreated as pilot lessons so that students could

learn to apply the Motometer and the whole classcould get used to hearing the bell sound.Following these initial sessions, data were gath-ered over a period of two weeks using both theclassroom observation form and the Motometerin three lessons of Spanish and three lessons ofGerman. Students were then given two copies ofthe questionnaire—one for their German classand one for their Spanish class—and were askedto fill them out at home.

In the analysis of the data, first the Motometerdata were documented by the first researcher asfollows: The distance between the bottom of theMotometer and the line drawn by the student wasmeasured inmillimetres and converted to a 1–100numeric scale. Using Excel, the data points foreach lesson were listed per participant. For eachof the lessons, the Motometer data were thenentered into a graph, resulting in six separategraphs, three for German and three for Spanish,showing the motivational variability of the fourstudents simultaneously. In these graphs, timewasrepresented on the horizontal axis in steps of fiveminutes. The vertical axis showed the students’level of motivation according to the Motometeron a 1–100 scale.

Second, the classroom observation data werecoded and placed into two categories by the firstresearcher: the more general ‘classroom activity’category on the one hand, and ‘episodic instan-ces’ on the other. The observations in this firstcategory, classroom activity, focused on theoverall structure of the lesson and activities suchas ‘teacher explains the use of past tense’, and‘students work in pairs on grammar exercise’. Theobservations in the second category, episodicinstances, were shorter in length and includedinstances such as ‘teacher tells class off for beingtoo noisy’. Episodes involving one of the fourparticipants were also included in this category,such as ‘teacher helps out Connor with theexercise’ and ‘Boris is chatting with his friend’.

Next, the first researcher combinedMotometerdata and observational data in so- called compositecharts. In these charts, the Motometer data of allfour participants were plotted against the obser-vational data, which were organized underneaththe horizontal axis. Furthermore, the writtencomments some participants provided on theirMotometers, elaborating on their motivation at agiven moment, were added underneath theobservational data. The combined observationaldata and students’ comments provided an onlinecontext for the Motometer data. An exampleof a composite chart is provided in the resultssection.


Finally, the first researcher used written obser-vations of the participants’ behaviour in class andtheir average grades on both subjects to draw up astudent profile for each of the four participants asshown in the Participants section of this article.The results of the motivation/attitude question-naires were analysed in Excel and used to draw upan attitude profile for each of the participants.This will be presented in the following section.

RESULTS AND DISCUSSION

Composite Charts

The various data gathering procedures gener-ated a considerable amount of information overthe six observed lessons. As previously described,six large composite charts were prepared (one foreach class) tomap theMotometer data against thevarious pieces of contextual information. Figure 1,also reproduced on the Web page in the onlineversion of the journal, offers an illustration takenfrom the Spanish course. As can be seen, thehorizontal axis represented time in steps of fiveminutes. Underneath the axis, the first level ofactivity blocks indicated the events making up theclass content (e.g., ‘teacher explains the use ofpast tense’, ‘students work on their assignment inpairs’). The second level presented classroomobservational comments (e.g., ‘student 1 ischatting with his friend’, ‘teacher helps Ireneindividually’, and ‘teacher gets annoyed anddemands silence’); these comments referred toepisodic instances rather than blocks of activities,and can be linked to specific points in the lesson.The rest of the boxes contained commentsprovided by the students on the Motometer at

specific measurement points. Unfortunately, dueto space limitations, we can only illustrate thiscombined data source in one chart in Figure 1,and even that cannot contain all the originalformatting and colour-coding features. There-fore, in the following discussion, subsets of theoverall data display charts will be presented insmaller, separate graphs. (The full-colour versionof Figure 1 is available in the online version of thisarticle.)

Student Profiles

The questionnaire data combined with class-room observation notes were used to constructindividual profiles for the four participants. Themain function of these profiles was to help toexplain or clarify some aspects of themotivationaldynamics observed. An overview of the learners’salient characteristics, overall classroom behav-iour, and learning achievements was alreadyoffered in Table 1 in the Participants section.This characterisation is further augmented byFigure 2, which presents questionnaire scalescores on the students’ reported enjoyment inclass, their confidence in their language learningabilities, classroom anxiety, and attitudes towardthe language teacher. Eachof these is representedon a 1–10 scale; a higher score indicates moreenjoyment, confidence, and anxiety, and a morepositive attitude toward the teacher. The profileshighlight several differences in the students’overall attitudes toward the German and Spanishcourses. We should note here that although thedata displayed in Figure 2 come from a quantita-tive measure, because of the low participantnumber it has been interpreted in a qualitative

TABLE 1Student Profiles

Student details General characteristics Learning characteristics Grades (x/100)

Boris One of the ‘cooler’ kids in class.Serious about learning.

Learns easily, high grades formost subjects if he likes them.

Spanish: 82Gender: M German: 69Age: 12Connor Behaves relatively young for his

age. Enthusiastic, but easily dis-tracted.

Has to work hard to keep up.Sometimes struggles to under-stand the teacher’s explanation.Low to moderate grades.

Spanish: 51Gender: M German: 61Age: 11

Nina Serious about her learning.Quite uncertain: She asks forclarification whenever some-thing is not clear and often asksfor the teacher’s confirmation.

Despite being uncertain, learnseasily and generally has veryhigh grades.

Spanish: 90Gender: F German: 80Age: 11

Irene Smart, but easily distracted andoften appears uninterested.

Does not put much effort intoclassoom tasks, moderate togood grades.

Spanish: 67Gender: F German: 76Age: 12


FIG

URE1

Composite

DataDisplayChart

0102030405060708090100

5404

5303

5202

5101

50

Mo�va�on 1-100

Tim

e in

step

s of 5

min

utes

All S

tude

nts,

Spa

nish

Cla

ss 3

Boris

Conn

or

Nin

a

Irene

Teac

her a

nnou

nces

test

at

the

end

of th

e le

sson

and

ex

plai

ns th

e st

ruct

ure

of th

e te

st.

Teac

her

misi

nter

pret

s a

ques

�on

by o

ne

of th

e st

uden

ts.

Teac

her w

arns

stud

ent

3 th

at sh

e ha

s to

keep

qu

iet o

r will

hav

e to

cl

ean

the

clas

sroo

m

a�er

the

less

on.Te

ache

r rep

eats

som

e of

the

basic

s of c

olou

rs in

Spa

nish

,and

an

noun

ces a

n ex

erci

se to

pr

ac�c

e th

em.

Teac

her a

sks

ques

�on

of

stud

ent 3

.

Teac

her e

xpla

ins

the

exer

cise

be

fore

the

SO.

Teac

her e

xpla

ins

the

exer

cise

, but

cau

ses

conf

usio

n by

giv

ing

oppo

site

answ

ers

to

stud

ents

' que

s�on

s.Stud

ents

do

the

assig

nmen

ts (m

akin

g up

sent

ence

s, d

escr

ibin

g an

obj

ect,

and

givi

ng it

a c

olou

r).

Stud

ent 4

is a

sked

to

dist

ribut

e di

c�on

arie

s.

Oth

erst

uden

t st

eals

dic�

onar

y

Teac

her g

ives

som

e ex

tra

a�en

�on

to st

uden

t 2 a

�er

repe

ated

que

s�on

s by

the

stud

ent.

Disr

up�o

n of

the

exer

cise

: One

of t

he

stud

ents

doe

sn't

have

a d

ic�o

nary

.

Teac

her g

ives

stu

dent

next

to st

uden

t 2 so

me

help

.

Teac

her g

ives

ex

tra

help

to

stud

ent 2

.

Oth

er st

uden

t ask

s st

uden

t 3 fo

r hel

p.

Clas

s sta

rts g

e�ng

noi

sier.

One

of

the

stud

ents

star

ts h

umm

ing

a so

ng.

Teac

her

dem

ands

sil

ence

and

en

ds

exer

cise

.

Stud

ents

and

teac

her c

heck

the

assig

nmen

ts. S

tude

nts r

ead

out t

heir

sent

ence

san

d as

k qu

es�o

ns

Teac

her t

hrea

tens

with

'sw

eepi

ng th

e cl

assr

oom

' ag

ain

to si

lenc

e th

e cl

ass.

Stud

ent 3

talk

s, te

ache

r sile

nces

he

r and

says

the

next

one

to ta

lk

has t

o 'st

ay in

' a�e

r cla

ss.

Frie

nd o

f stu

dent

2

talk

s, h

as to

stay

a�

er cl

ass.

Stud

ent 2

raise

s hi

s ha

nd to

ask

a q

ues�

on;

teac

her d

oesn

't se

e hi

m.Te

ache

r rea

lizes

stu

dent

2 w

ants

to a

sk a

qu

es�o

n an

d tu

rns t

o hi

mTe

ache

r m

isund

erst

ands

que

s�on

and

hal

f ans

wer

s it.

Stud

ent 2

ask

s an

othe

r que

s�on

and

gi

ves a

cor

rect

ans

wer

. Tea

cher

tells

hi

m 'y

ou're

in e

xcel

lent

form

toda

y'.

Oth

er st

uden

tdisr

upts

less

on, a

nd is

pla

ced

in th

e fr

ont o

f the

cla

ssro

om d

irect

ly in

fron

t of t

each

er.

Anot

her

ques

�on

by

Que

s�on

by

stud

ent 3Te

st

Stud

ent 2

seem

s to

be

sad

and/

or

cryi

ng, s

ome

com

mo�

on a

bout

bo

oks i

n th

at

corn

er.

End

of te

st, h

andi

ng in

of

pape

rs, w

rapp

ing

up o

f le

sson

Stud

ent 3

writ

es o

n th

e ta

ble,

teac

her c

omm

ents

on

it.

Que

s�on

by

stud

ent 2

0102030405060708090100

5404

5303

5202

5101

50

Mo�va�on 1-100

Tim

e in

step

s of 5

min

utes

All S

tude

nts,

Spa

nish

Cla

ss 3

Boris

Conn

or

Nin

a

Irene

Teac

her a

nnou

nces

test

at

the

end

of th

e le

sson

and

ex

plai

ns th

e st

ruct

ure

of th

e te

st.

Teac

her

misi

nter

pret

s a

ques

�on

by o

ne

of th

e st

uden

ts.

Teac

her w

arns

stud

ent

3 th

at sh

e ha

s to

keep

qu

iet o

r will

hav

e to

cl

ean

the

clas

sroo

m

a�er

the

less

on.Te

ache

r rep

eats

som

e of

the

basic

s of c

olou

rs in

Spa

nish

,and

an

noun

ces a

n ex

erci

se to

pr

ac�c

e th

em.

Teac

her a

sks

ques

�on

of

stud

ent 3

.

Teac

her e

xpla

ins

the

exer

cise

be

fore

the

SO.

Teac

her e

xpla

ins

the

exer

cise

, but

cau

ses

conf

usio

n by

giv

ing

oppo

site

answ

ers

to

stud

ents

' que

s�on

s.Stud

ents

do

the

assig

nmen

ts (m

akin

g up

sent

ence

s, d

escr

ibin

g an

obj

ect,

and

givi

ng it

a c

olou

r).

Stud

ent 4

is a

sked

to

dist

ribut

e di

c�on

arie

s.

Oth

erst

uden

t st

eals

dic�

onar

y

Teac

her g

ives

som

e ex

tra

a�en

�on

to st

uden

t 2 a

�er

repe

ated

que

s�on

s by

the

stud

ent.

Disr

up�o

n of

the

exer

cise

: One

of t

he

stud

ents

doe

sn't

have

a d

ic�o

nary

.

Teac

her g

ives

stu

dent

next

to st

uden

t 2 so

me

help

.

Teac

her g

ives

ex

tra

help

to

stud

ent 2

.

Oth

er st

uden

t ask

s st

uden

t 3 fo

r hel

p.

Clas

s sta

rts g

e�ng

noi

sier.

One

of

the

stud

ents

star

ts h

umm

ing

a so

ng.

Teac

her

dem

ands

sil

ence

and

en

ds

exer

cise

.

Stud

ents

and

teac

her c

heck

the

assig

nmen

ts. S

tude

nts r

ead

out t

heir

sent

ence

san

d as

k qu

es�o

ns

Teac

her t

hrea

tens

with

'sw

eepi

ng th

e cl

assr

oom

' ag

ain

to si

lenc

e th

e cl

ass.

Stud

ent 3

talk

s, te

ache

r sile

nces

he

r and

says

the

next

one

to ta

lk

has t

o 'st

ay in

' a�e

r cla

ss.

Frie

nd o

f stu

dent

2

talk

s, h

as to

stay

a�

er cl

ass.

Stud

ent 2

raise

s hi

s ha

nd to

ask

a q

ues�

on;

teac

her d

oesn

't se

e hi

m.Te

ache

r rea

lizes

stu

dent

2 w

ants

to a

sk a

qu

es�o

n an

d tu

rns t

o hi

mTe

ache

r m

isund

erst

ands

que

s�on

and

hal

f ans

wer

s it.

Stud

ent 2

ask

s an

othe

r que

s�on

and

gi

ves a

cor

rect

ans

wer

. Tea

cher

tells

hi

m 'y

ou're

in e

xcel

lent

form

toda

y'.

Oth

er st

uden

tdisr

upts

less

on, a

nd is

pla

ced

in th

e fr

ont o

f the

cla

ssro

om d

irect

ly in

fron

t of t

each

er.

Anot

her

ques

�on

by

Que

s�on

by

stud

ent 3Te

st

Stud

ent 2

seem

s to

be

sad

and/

or

cryi

ng, s

ome

com

mo�

on a

bout

bo

oks i

n th

at

corn

er.

End

of te

st, h

andi

ng in

of

pape

rs, w

rapp

ing

up o

f le

sson

Stud

ent 3

writ

es o

n th

e ta

ble,

teac

her c

omm

ents

on

it.

Que

s�on

by

stud

ent 2

Conn

or: W

e ca

n s�

ll pr

ac�c

e(fo

r the

te

st)

Conn

or: I

like

this,

m

akin

g th

ese

sent

ence

s. Ju

st

tran

sla�n

g se

nten

ces

is no

fun

at a

ll.

Conn

or: S

omeo

neth

rew

a

dic�

onar

y to

my

head

.Co

nnor

:The

test

w

ent w

ell.

Nin

a: O

nly

one

mor

e ho

ur a

nd th

is sc

hool

-day

is ov

er!

Irene

: Thi

s le

sson

is o

ver.

Onl

y on

e m

ore

hour

and

we'

re fr

ee fo

r to

day.


manner, as part of the students’ profiles. Tohighlight the most salient patterns, three of thefour students show a preference for Spanish overGerman, and in one student’s case (Boris) there isa striking disparity between his disposition towardSpanish and German. Only Connor has a morepositive attitude toward German than Spanish.

Change and Variability

The findings reveal a considerable amount ofindividual variability in the four students’motiva-tion. To illustrate this, Figures 3 and 4 present themotivational development of the participantsduring a German lesson (Lesson 3), with thedata charted both at the group and the individuallevels. Figure 3 displays a relatively straightforwardpattern: The overall group motivation shows asteady increase from a relatively low level (30/100) to a moderately high level (60/100), with asmall final decline. However, the individual leveltrajectories in Figure 4 do not reflect this patternin any of the four students’ motivational progres-sion: Here we can see dramatic ups and downs,

and only two of the participants, Connor andNina, actually gain motivation during the lesson.Boris seems to be rather unmotivated duringthe whole lesson, whereas the progression ofIrene’s motivation shows an increase followedby a decrease and another increase, suddenlyending in a decrease. These findings are in full

FIGURE 2Student Attitude Profiles for German and Spanish on a 1–10 Scale, Obtained From Questionnaire Scores

FIGURE 3German Lesson 3, Group Average of MotivationalProgression


accordancewithLarsen–Freeman’s (2006) results,whereby none of the individual students exam-ined displayed the pattern indicated by the groupaverage.

Some of the variability observed in this lessoncan be explained by contextual factors: Irene(with a partner) was asked to act out a written‘restaurant scene’ between Minutes 17 and 22,which was followed by playing a vocabulary game,starting at Minute 27. Irene’s motivation beforethe game showed variability because she was askedto perform in front of her classmates, duringwhich time some of her friends were gigglingrepeatedly. The subsequent game seeminglyhelped to increase her motivation, along withConnor’s and Nina’s; however, this positiveimpetus did not seem to affect Boris, who hadthe lowest level of overall attitudes towardGerman of the four students (see Figure 2).Interestingly, Irene’s motivational profile is thesecond lowest regarding German and Connor’s isthe highest, which seems to be in accord with theirmotivational state in this class. These observationspoint toward a tentative conclusion that some-one’s generalised attitude toward a language mayact as a higher order attractor (i.e., functioningon a higher time scale) that affects the situatedfluctuation in classroom motivation, but thiseffect is far from uniform and fully predictable:In Boris’s case it resulted in a general disinterestand indifference, while in Irene’s and Connor’scases it affected the general level of theirmotivational intensity without altering the direc-tions and the scope of the fluctuation. Thus, wecan say that, while Irene and Connor respondedto the stimuli in a proportionate manner withintheir overall motivational range, Boris’s responsewas nonlinear as it did not reflect the positiveimpact of the game.

Although the moderating effect of a higherorder attractor can make the data more predict-able, some of the motivational variability amongthe participants remains complex. This is wellillustrated by the fact that students who showsimilar motivational patterns in one lesson can beeach other’s opposites in another. An example ofthis is given in Figures 5, 6, and 7, which compareBoris’s and Connor’s motivational progressionin three lessons. Figure 5 shows a contrastingpattern: Boris was losing while Connor wasgaining motivation throughout a Spanish lesson,which is in fact contradictory to their Spanishmotivational profile (Boris is superior to Connorby a large margin). It appears that here thestarting point—that is, the initial condition—ofsystem behaviour played a significant role: Boris’sinitial motivation at the beginning of the lessonwas close to maximum, and even after the declineit still exceeded Connor’s, whose upward trend inthis class started from a rather low motivationallevel. Then, in Figure 6, we find a virtuallyidentical motivational development for the twoboys in a German lesson, whereas when we look atFigure 7, taken from another Spanish lesson, it ishard to imagine that the two boys were attendingthe same class. Another telling illustration ofbetween-participant variability is a comparison ofNina’s and Irene’s cases. According to Figure 2,they share fairly similar motivational profiles inboth languages, but this similarity is not reflectedin corresponding motivational reactions in theactual language lessons. As can be seen in Figure8, they can react in the opposite way: Recall thatthis was the lesson in which Irene was asked toperform a little act in front of everyone, whichexplains partly, but not fully, why the two girls hada completely different classroom experience.

FIGURE 5Boris’s and Connor’s Motivational Development inSpanish Lesson 1

FIGURE 4German Lesson 3, Individual MotivationalProgression


Finally, in addition to variation among thedifferent participants, the data also showedwithin-participant variation. A good example of the lattertype of variation is provided by a closer inspectionof Irene’s motivational development. Looking atFigures 9 and 10, which chart her motivationalchange in Spanish and German, respectively, thedominant pattern appears to be a moderate andgradual increase in motivation, without anysudden increases or decreases. However, duringthe last lesson of German and the last lesson ofSpanish, her motivation deviated considerablyfrom this pattern shown in previous lessons.Regarding the German lesson, the plummetingmotivation may still be related to the acting fiascodescribed earlier; as for the sharp rise in theSpanish class, it might be relevant to note that thisparticular lesson took place on a Friday afternoonand Irene commented at the endof the lesson (seeFigure 1): “This lesson is over. Only one morehour and we are free for today!”

Stability

One important lesson that we candraw from thedata is that change is not always random orunpredictable and even group averages may notalways be misleading. First of all, in several lessonphases the group-level data and the individualtrajectories showed remarkable similarities, ascan be seen, for example, in Figures 11 and 12,representing motivational development in themiddle section of a Spanish lesson (Minutes 10–30). The pattern of decline! increase! increase!decline is uniform among all four participantsand thus coincides with the group-level trajectory.According to DST, such a commonality within amulticomponential system such as a language classindicates the influence of a strong attractor; in thisparticular case the accompanying contextualinformation reveals that the increase in studentmotivation can be linked to a change in classroomactivity: After a lengthy explanation related to the

FIGURE 8Nina’s and Irene’s Motivational Development inGerman Lesson 3

FIGURE 9Irene’s Motivational Development in the SpanishLessons

FIGURE 6Boris’s and Connor’s Motivational Development inGerman Lesson 2

FIGURE 7Boris’s and Connor’s Motivational Development inSpanish Lesson 3


next test (which dampened motivation), theteacher introduced an engaging listening exercisein Minute 22, resulting in increased motivation,which was also confirmed by student comments(e.g., the very short but clear note by Connor:“Listening-exercise: nice!”).

We can gain further insights into the relation-ship between stable/predictable and variable/unpredictable phases of classroom learning whenwe look at the motivational progress that preced-ed and succeeded the attractor-governed centralphase of the Spanish lesson discussed above.Figure 13 presents the four participants’ motiva-tional development at the beginning and at theend of the Spanish lesson described in Figures 11and 12. As we can see, in these framing phases,students experienced highly varied motivationalstates: In the beginning, two of them displayedincreasing, the other two declining trajectories,which were then unified by the listening activitythat acted as a strong attractor. When this activity

came to an end, in the absence of any powerfulattractor-based regulation (in this case, thelistening activity), the students once again startedto show disparate patterns of motivational devel-opment. Figure 13 shows four very differentpatterns for the four students during the last 10minutes of the lesson (rise, fall, rise–fall, andfall–rise). One reason for this before-and-afterdiversity is likely to be related to the influence ofpre-lesson events at the beginning of class—goodteachers are usually aware of the need to pull the

FIGURE 13Motivational Development During the Beginningand End of Spanish Lesson 2

FIGURE 12Motivational Development Individual ProgressionDuring Spanish Lesson 2

FIGURE 10Irene’s Motivational Development in the GermanLessons

FIGURE 11Motivational Development Group ProgressionDuring Spanish Lesson 2


students together with some ‘warmers’—as well asthe anticipation of lessons or activities that were tofollow the class. This highlights the importance ofthe temporal context of language lessons.

Apart from any observable uniformity acrossparticipants, we also find a certain amount ofstability within the students’ own dispositions.The most striking example of this is Boris’smotivation during Spanish classes. As can beseen in his profile in Figure 2, his attitude towardSpanish is extremely positive. This echoes in allhis motivational data gathered through theMotometer: While in Figures 14, 15, and 16 wecan see how his motivation is still susceptible tovariation, this happens at a different level fromthe fluctuation observed in the other students.That is, in spite of his very high generalmotivationto learn Spanish, Boris still has motivational upsand downs, but the overall picture shows a muchhigher level of motivational engagement than forConnor, Nina, and Irene. One of his commentswritten on the Motometer sheds important lighton this issue: “I like learning Spanish, also becauseof Mr. Seran [the teacher]. That’s why theMotometer-score is so high all the time” (Spanishlesson 1). This comment confirms a key tenet ofclassroom reality: Teachers have a salient class-room role by definition, allowing them to becomepowerful motivational attractors (or repellents).

Context

As we have seen in the previous examples,information about the classroom context isessential in uncovering factors that affect motiva-tional variability. Some of the more pronouncedcontextual influences, such as those pertaining tothe language learning tasks or the students’generalised attitudes toward the language (po-tentially caused by their teachers), have been

mentioned earlier. Other influences are not soobvious or easy to identify, as shown in thefollowing examples. To start with, in Figure 17 wecan see themotivation of Nina and Irene during alesson of Spanish. Both girls reported a substan-tial increase in their motivation toward the endof the lesson which did not seem to be related tothe lesson content. When asked to explain thissudden change they confirmed—separately fromeach other—that the reason they reported more

FIGURE 14Motivational Development During Spanish Lesson 1



FIGURE 17Motivational Development ofNina and IreneDuringSpanish Lesson 3


motivation was: “Only one more hour and thisschool-day is over!” (Nina, Spanish lesson 3). Asmentioned earlier, the fact that the class inquestion was on a Friday afternoon will most likelyhave been a factor in this shared enthusiasm,which reiterates the significance of the temporalcontext and also opens up the classroom spacebeyond the actual classroom walls.

In Figure 18 we see another example of theimportance of context, in this case for Connor.At first his motivation is rising steadily to about70/100, only to suddenly drop to 40/100. At firstglance the decreasing motivation could beattributed to anxiety caused by an oncomingtest, starting just after Minute 40. However,additional observational data—Connor seemedto be sad and crying—and his subsequent com-ments show a different picture. What happenedwas not language-related at all: While distributingdictionaries, one of the students threw a dictio-nary to the table of Connor and his friend.However, his friend didn’t catch the book in timeand it hit Connor’s head instead, causing himto cry and lose motivation. After the test hecommented, “I did well on the test!” whichexplains his surge in motivation.

CONCLUSION

This study utilised real-time assessment ofclassroom motivation, charting the progressionof four selected learners’ dispositions at five-minute intervals. To answer our first researchquestion of whether there is variability to befound in students’ in-class motivation, the resultsprovided a clear illustration that student motiva-tion is susceptible to variation even on a rathershort time scale: We have observed considerableups and downs and shifts within the learners’motivational state within a single classroom

session of 45–50 minutes, even if the time frameof the study only spanned two weeks. Confirmingthe findings of Larsen–Freeman’s (2006) investi-gation, our study showed that averaged groupdata can often be insufficient to capture thevariation in individual learners’ development: Insome of the lessons examined, the trajectory ofthe group data did not coincide with any of thestudent’s individual trends, which was bestillustrated by a case when the learners as a groupshowed a steady increase in their motivationwhereas half of the students actually experienceda decline. In other cases, however, a grand sweepof the data did offer a fair representation of whatwas going on. Our findings showed that theinstances when all the participants reacted in asimilar manner were associated with powerfulattractor states: Certain contextual factors consti-tuted a strong enough force to regulate disparatesystem behaviour. To recognise these attractorstates and identify the most frequent sourceswould be a helpful tool for teachers to enhanceand maintain learners’ in-class motivation.

To answer our second research questionconcerning detectable stable levels, or attractorstates, in students’ classroom motivation, theresults showed that even over a time span of justtwo weeks we could indeed detect a rather stablelevel of overall motivation, of which Boris’s overallhighly motivated state in Spanish class is theclearest example. We also presented evidence ofa likely regulating influence, namely a student’soverall liking of the subject matter, such as, inBoris’s case, a very positive attitude towardlearning Spanish. For Boris, this predilectionwas traced back to his positive attitudes toward theSpanish teacher, and it resulted in a generallymore positive engagement with classroom activi-ties. However, it was noteworthy that the trajectoryof system behaviour was also dependent on theinitial condition of the system: One tellingexample was when Boris’s negative and Connor’spositive motivational trajectories in a lesson werecontradictory to their overall language preferen-ces and could be explained by the initialmotivational level with which they entered theparticular language lesson. Taking into accountwhat was shown earlier on in Figure 15, where theactivity following the current lesson influencesstudents’ motivation in class, we suggest that theevents prior to the lesson can similarly affectstudents’ (initial) motivation during their lesson.Given that the initial level of motivation can be ofconsiderable importance, findings like thesereiterate that it is vital for teachers to invest atthe beginning of a classroom session, either using

FIGURE 18Motivational Development of Connor DuringSpanish Lesson 3


a warm-up activity, or simply by making a point ofhaving everyone’s attention at the beginning ofthe lesson, something that was not always done inthe sessions observed in this study.

So far, in line with the principles of DST, ourmain finding was that regulated and seeminglystable phases of the students’ behaviour alternatedwith seemingly erratic reactions. With respect toour third research question on the importance ofcontext, sometimes these latter reactions could beattributed to specific classroom episodes (e.g., adictionary hitting a student’s head) or could beexplained by retrospective comments such as anunexpected increase in two students’ level ofmotivation which turned out to be related to thejoy of realising that it was Friday afternoon andfreedom was near. These instances served asevidence for the close, inextricable link betweencontext (both spatial and temporal) and systembehaviour, confirming the fundamental DSTtenet that the immediate context is best perceivedas part of the overall system. On the other hand,in some cases evenpowerful regulating forces suchas the introduction of an engaging listeningexercise did not have a proportionate impact onall the students, which was in accordance withthe DST principles of nonlinearity in systembehaviour.

Thus, the overall picture we obtained from ourmicro-inspection of classroom motivation is thatof a mixture of dynamic stability, governed byattractor states, and individual variability, causedby a combination of multiple issues. The variabil-ity can sometimes—but not always—be accountedfor by specifics in the situation if we consider awide enough range of potentially interrelatedfactors, but in some cases no obvious reason maybe forthcoming. Furthermore, even if we canidentify certain powerful contextual influences,they may have a disproportionate—that is,nonlinear—impact on the learners’ behaviour.While all this may sound rather confusing, such amixed picture of partial order and partial chaos isin fact not too far from many language teachers’perceptions of what classroom reality is reallylike. Therefore, we would suggest that a dynamicsystem perspective offers a suitable lens forexamining classroom phenomena in theirrichness.

Implications for language teachers include, asmentioned before, the importance of the start of alesson and pulling the students together withsome ‘warmers’. Another piece of advice that canbe drawn from this study is for teachers to becomeaware of the forces in their classrooms that canfunction as a push and pull strong enough to

create an attractor state in the students’ motiva-tion, either negative (such as a long grammarexplanation) or positive (such as the introductionof a vocabulary game). A final more general pointto take away for both teachers and learners is thearticle’s overarching argument that in-class moti-vation is by no means a stable trait. As mostlanguage teachers are generally aware, while acertain group of learners or an individual learnermay display a higher or lower level of motivationoverall, their motivation in the language class-room is still susceptible to change. The mostmotivated student can lose interest after half anhour of grammar explanation and be in need of amotivational boost, while a student who generallyseems to be unmotivated can suddenly becomeengaged in a new task. These two levels ofmotivation most likely interact with one anotherconstantly. Indeed, this motivational variabilitywithin a classroom session may help shapestudents’ overall L2 motivation. Therefore, thisvariability itself and the extent to which theteacher has influence over it most certainlymatters.

Our study has its limitations. First of all, ithas been conducted with a limited number ofparticipants: Data from a larger sample ofstudents for a longer period of time undoubtedlywould provide more insights into the existence ofpossible attractor states and into the relationshipof motivational development at an individual anda class/group level. We must also bear in mindthat the learners in this study were relativelyyoung; older teenagers will have developedfurther meta-motivational maturity to regulatethe incidental ups and down of their situateddisposition. Finally, the data in this study weregathered by a single researcher, the first author,and the study’s reliability would likely have beenenhanced by another pair of observing eyes.Despite these limitations, we hope to have shownthat using a relatively straightforward and com-mon-sense methodology that is based on theprinciples of DST can uncover layers of thedynamic processes influencing student motiva-tion that otherwise would have remained hidden.The present article was not intended to offera comprehensive treatment of all these layers.Instead, its aim was to offer convincing illustra-tions of the fact that documenting variousmotivational pushes and pulls in a dynamicsystems vein can identify relatively stable—andthus potentially researchable and manipulable—attractor and repellent states. This results in thepossibility of modelling student behaviour at anenhanced level of reality.


NOTES

1 Although lessons officially lasted 50 minutes, sincestudents had to move from one classroom to another inbetween lessons, the teaching time ranged from 40 to45 minutes, depending on how long the teacher took tostart the lesson.

2 All participant and teacher names have beenchanged.

REFERENCES

Ahmed, W., Van der Werf, G., Minnaert, A., & Kuyper,H. (2011). Students’ daily emotions in theclassroom: Intra-individual variability and ap-praisal correlates. British Journal of EducationalPsychology, 80, 583–597.

Block, D. (2003). The social turn in second languageacquisition. Edinburgh, UK: Edinburgh UniversityPress.

Campbell, E., & Storch, N. (2011). The changing face ofmotivation: A study of second language learners’motivation over time. Australian Review of AppliedLinguistics, 34, 166–192.

Chambers, G. N. (1999). Motivating language learners.Clevedon, UK: Multilingual Matters.

Crookes, G., & Schmidt, R. (1991). Motivation: Reopen-ing the research agenda. Language Learning, 41,469–512.

de Bot, K. (2008a). Introduction: Second languagedevelopment as a dynamic process. ModernLanguage Journal, 92, 166–178.

de Bot, K. (2008b). Second language development as adynamic process [Special issue]. Modern LanguageJournal, 92, 166–283.

de Bot, K. (2012). Time scales in second languagedevelopment.Dutch Journal of Applied Linguistics, 1,143–149.

Dornyei, Z. (1994). Motivation and motivating in theforeign language classroom. Modern LanguageJournal, 78, 273–284.

Dornyei, Z. (2009a). Individual differences: Interplay oflearner characteristics and learning environment.Language Learning, 59, 230–248.

Dornyei, Z. (2009b). The psychology of second languageacquisition. Oxford: Oxford University Press.

Dornyei, Z. (2014). Researching complex dynamicsystems: ‘Retrodictive qualitative modelling’ inthe language classroom. Language Teaching, 47,80–91.

Dornyei, Z., Csizer, K., & Nemeth, N. (2006).Motivation,language attitudes and globalisation: A Hungarianperspective. Clevedon, UK: Mutlilingual Matters.

Dornyei, Z., MacIntyre, P., & Henry, A. (2014).Motivational dynamics. Bristol, UK: MultilingualMatters.

Dornyei, Z., & Otto, I. (1998). Motivation in action: Aprocess model of L2 motivation. Working Papers

in Applied Linguistics (Thames Valley University,London), 4, 43–69.

Ellis, N., & Larsen–Freeman, D. (2006). Languageemergence [Special issue]. Applied Linguistics, 27,558–740.

Gardner, R. C. (1985). Social psychology and secondlanguage learning: The role of attitudes and motivation.London: Edward Arnold.

Gardner, R. C., Masgoret, A. M., Tennant, J., & Mihic, L.(2004). Integrative motivation: Changes during ayear long intermediate level language course.Language Learning, 54, 1–34.

Hawkins, R. (2008). Current emergentist and nativistperspectives on second language acquisition[Special issue]. Lingua, 118, 445–641.

Hotho, S. (2000). “Same” or “different”? A comparativeexamination of classroom factors in secondlanguage settings. Foreign Language Annals, 33,320–329.

Jessner, U. (2008). A DSTmodel of multilingualism andthe role of metalinguistic awareness. ModernLanguage Journal, 92, 270–283.

Koizumi, R., & Matsuo, K. (1993). A longitudinal studyof attitudes and motivation in learning Englishamong Japanese seventh-grade students. JapanesePsychologuy Research, 35, 1–11.

Larsen–Freeman, D. (1997). Chaos/complexity scienceand second language acquisition. Applied Linguis-tics, 18, 140–165.

Larsen–Freeman, D. (2006). The emergence of com-plexity, fluency and accuracy in the oral andwritten production of five Chinese learners ofEnglish. Applied Linguistics, 27, 590–619.

Larsen–Freeman, D. (2012). Complex, dynamic sys-tems: A new transdisciplinary theme for appliedlinguistics? Language Teaching, 45, 202–214.

Larsen–Freeman, D., & Cameron, L. (2008). Complexsystems and applied linguistics. Oxford: OxfordUniversity Press.

MacIntyre, P., & Legatto, J. J. (2011). A dynamicsystems approach to willingness to communicate:Developing an idiodynamic method to capturerapidly changing affect. Applied Linguistics, 32,149–171.

Mercer, S. (2011a). Language learner self-concept:Complexity, continuity and change. System, 39,335–346.

Mercer, S. (2011b). The self as a complex dynamicsystem. Studies in Second Language Learning andTeaching, 1, 57–82.

Pawlak, M. (2012). The dynamic nature of motivation inlanguage learning: A classroom perspective. Stud-ies in Second Language Learning and Teaching, 2,249–278.

Poupore, G. (2013). Task motivation in process: Acomplex systems perspective. Canadian ModernLanguage Review, 69, 1–26.

Scherer, K. R. (1993). Plato’s legacy: Relationships betweencognition, emotion, and motivation. Paper presentedat the Societe Psychologique de Quebec, Quebec,Canada.


Scherer, K. R. (2000). Emotions as episodes of subsystemsynchronization driven by nonlinear appraisalprocesses. In M. D. Lewis & I. Granic (Eds.),Emotion, development, and self-organization: Dynamicsystems approaches to emotional development (pp. 70–99). Cambridge: Cambridge University Press.

Scherer, K. R. (2001). Psychological structureof emotions.In N. J. Smelser & P. B. Baltes (Eds.), Internationalencyclopedia of the social & behavioral sciences (pp.4472–4477). Oxford, UK: Pergamon Press.

Shoaib, A., & Dornyei, Z. (2005). Affect in lifelonglearning: Exploring motivation as a dynamicprocess. In P. Benson&D. Nunan (Eds.), Learners’stories: Difference and diversity in language learning(pp. 22–41). Cambridge: Cambridge UniversityPress.

Spoelman,M., &Verspoor,M. (2010). Dynamic patternsin development of accuracy and complexity: Alongitudinal case study in the acquisition ofFinnish. Applied Linguistics, 31, 532–553.

Thelen, E. (2005). Dynamic systems theory and thecomplexity of change. Psychoanalytic Dialogues, 15,255–283.

Ushioda, E. (2009). A person-in-context relational viewof emergent motivation, self and identity. In Z.Dornyei & E. Ushioda (Eds.), Motivation, languageidentity and the L2 self (pp. 215–228). Bristol, UK:Multilingual Matters.

Van Dijk, M., Verspoor, M., & Lowie, W. (2011).Variability and DST. In M. Verspoor, K. de Bot,& W. Lowie (Eds.), A dynamic approach to secondlanguage development, methods and techniques. Phila-delphia/Amsterdam: John Benjamins.

Van Geert, P. (1998). A dynamic systems model of basicdevelopmental mechanisms: Piaget, Vygotsky, andbeyond. Psychological Review, 105, 634–677.

Van Geert, P., & Steenbeek, H. (2005). Explaining afterby before: Basic aspects of a dynamic systemsapproach to the study of development. Develop-mental Review, 25, 408–442.

Verspoor, M., de Bot, K., & Lowie, W. (Eds.). (2011). Adynamic approach to second language development,methods and techniques. Philadelphia/Amsterdam:John Benjamins.

Verspoor, M., Lowie, W., & Van Dijk, M. (2008).Variability in second language developmentfrom a dynamic systems perspective. ModernLanguage Journal, 92, 214–231.

Williams, M., & Burden, R. (1997). Psychology for languageteachers: A social constructivist approach. Cambridge:Cambridge University Press.

Zuengler, J., & Miller, E. R. (2006). Cognitive andsociocultural perspectives: Two parallel SLAworlds? TESOL Quarterly, 40, 35–58.

APPENDIX A

Classroom Activities in the Observed LanguageLessons

TABLE A1German Lesson 1

Time(minutes) Activity

0–7 The teacher explains the schedule forthe next few weeks and students andteacher decide on the dates for twoupcoming tests. The teacher talks aboutthe books students will read and a filmthey will watch.

7–17 Students receive the results of theirlatest proficiency test and are given theopportunity to reflect on their answersand ask questions.

17–27 The teacher and students together goover the homework, checking students’answers and adding explanations wherenecessary.

27–37 New exercise: Students are given a list ofitems for which they don’t know theGerman translation, and practise de-scribing the items using the targetlanguage, gestures, and sounds. Afterstudents have described all the items inpairs, the teacher asks some students togive their description to the class.

37–39 Students start working on a second taskindividually. They translate sentencessimilar to the ones used in the previoustask, describing common objects.

40 End of lesson, teacher gives homework.


Time(minutes)

Activity

0–8 Students are given three minutes tocomplete the unfinished task from thelast lesson. Afterward, students andteacher correct the task together.

8–10 The teacher returns an assignment thestudents had handed in a week earlier: arestaurant-themed play they had writtenin pairs. Two students are asked toperform their play next week.

10–41 The teacher introduces a new vocabu-lary game: Each student receives eightempty flashcards. On one side they writeDutch words, on the other side theGerman translation; they win a cardevery time they translate it correctly. Thegame is slightly chaotic; students havefun but don’t always follow the rules.

41–42 Teacher ends the game and asks stu-dents for feedback. She gives homeworkfor next week and dismisses the class.




0–5 Teacher checks students’ understandingof the text they had to read as home-work. She explains the use of cases inGerman to answer a student’s question.

5–13 The teacher answers grammar questionsthat students have concerning the use of‘wen’ and ‘wer’ and the use of gender inGerman.

13–18 Two students act out their restaurantscene as agreed upon in the last lesson.The teacher gives them some feedback.

19–21 Teacher explains a new vocabulary gamesimilar to the one they did in the lastlesson. Students are asked to vote foreither playing the game or practicingtheir restaurant scenes: The majoritychooses the game.

21–40 The teacher distributes flashcards withnew vocabulary, students form groups ofthree or four to play the game. The rulesare similar to that of ‘Quartet’, a Dutchcard game.

40 End of class, teacher collects flashcardsand gives homework.

TABLE A4Spanish Lesson 1


0–6 The teacher starts with a grammarexplanation of the use of ‘muy’ and‘mucho’.

6–17 The teacher dictates examples,students copy them in their books.

17–21 Students ask another question about thegrammar explanation, the teacher ex-plains.

21–32 The teacher dictates sentences for thestudents to translate into the targetlanguage. Students copy them in theirbooks and start translating in silence.

32–37 Teacher and students check andcorrect students’ translations together

37–40 Teacher refers back to homework fromlast week and asks students to namecolours in Spanish.

40 Teacher gives homework and dismissesclass.



0–14 Teacher announces and sets date for theupcoming test. He goes through thevocabulary and the grammar that will betested.

14–18 Teacher answers questions about thematerial covered in the test.

18–25 Teacher asks students to name coloursin Spanish.

25–39 Listening exercise: Students listen toseveral short recordings and answerpre-set questions.

39–42 Teacher explains the grammar of ‘muy’and ‘mucho’.

42 Teacher gives homework, end of class.



0–4 The teacher announces a short test thatwill take place at the end of the lessonand explains the structure of the test(eight sentences to be translated).

4–23 Students work on a task individually;they have to make up their ownsentences describing an object andgiving it a colour. They are allowed to usea dictionary.

23–36 Teacher and students check and correctstudents’ answers to the previous task.Students are given the time to ask somelast questions before the test.

36–45 Proficiency test.45 End of test, teacher wraps up lesson and

class is dismissed.


APPENDIX B

MotometerRate your motivation, considering

� How much effort do I want to put intolearning the material right now?

� How much do I enjoy this lesson right now?

Comments:

� . . .� . . .� . . .� . . .� . . .


Motivational Dynamics in Language Learning: Change ...

Documents