Willingness to communicate in a multilingual context: part two, person-context dynamics

ABSTRACT In many contexts of multilingualism, language learners can initiate communication in the target language (TL), or a contact language (such as English). Patterns of use emerging from these choices vary between individuals and affect TL development. Willingness to communicate (WTC) needs to be investigated in ways that capture these variations. So far, WTC has not been studied in multilingual contexts, or using individual-level designs. This case study explores intraindividual variability in the WTC propensities of adult learners of Swedish for whom the TL and English provide viable communication options in community interaction. Carried out over a period where TL skills began to develop, the purpose was to explore the process characteristics of changes in communication-initiation propensities. A person-context dynamics perspective was employed, and analyses of time-serial data were combined with analyses of concurrently generated interview data. Results reveal how changes in WTC could be gradual and nongradual, continuous and discontinuous.


Introduction
Willingness to communicate (WTC) is an emergent, dynamically shifting, psychological state of communicative readiness that fluctuates within and between communication events (MacIntyre 2007(MacIntyre , 2020)).In a multilingual context, communicative readiness involves decisions about speaking or not speaking, and about the language in which communication might be initiated.For the multilingual language learner, these choices involve situated assessments of the 'usability' of different languages; will I do better if I speak in the Lx, the Ly, or the Lz?In a related quantitative, process-oriented study (Henry, Thorsen, and MacIntyre, forthcoming), we used change point analysis (CPA) to examine the WTC behaviours of six learners for whom the target language (Swedish) and a contact language (English) each provided a viable communication option.CPA is a statistical methodology that enables the analysis of changes in time series data where parametric assumptions cannot be made.CPA identifies points or 'thresholds' along a distribution of values on either side of which characteristics can vary (Taylor 2000).This makes it possible to identify states that are stable, and periods where changes towards other stable states take place (Steenbeek, Jansen, and van Geert 2012).CPA is therefore well-suited to investigating developmental processes in a complex dynamic system, such as willingness to communicate (MacIntyre et al. 2017).The data showed that willingness to communicate in both languages underwent change throughout the period, both in absolute and relative WTC propensities.It also showed how developmental patterns differed substantially from one person to the next.WTC was characterised by patterns of stability, gradual change, and points where an abrupt shift could occur.Aligned with conceptualizations of multilingual development as a dynamic process (Aronin 2017;Jessner 2008) where growth and loss are investigated systemically for all languages and over time (De Bot and Jaensch 2015), observed changes in English WTC (EWTC) and Swedish WTC (SWTC) pointed to the non-linearity of evolving communication behaviour, and to patterns of co-evolutionary development.
The present study extends this previous work by drawing on concurrently generated qualitative data from the same participants.This allows us to explore the specifics of communication behaviors at the times when changes were identified.The purpose of the study is thus to explore the nature and the sources of changes in WTC in a context of multilingual development.

An extensively researched, multi-layered construct
As noted in the companion paper (Henry, Thorsen, and MacIntyre, forthcoming), WTC has been defined in two ways.Initially, WTC was regarded as a longer-term individual difference in the propensity to talk.More recently, however, attention has been directed toward rapidly fluctuating patterns, and WTC has been conceptualized as a state of readiness and immediate antecedent to actual communication (MacIntyre 2020).The trait and state approaches to WTC employ (respectively) long versus short timescales.For an individual, the fluctuations in state WTC over time and across situations can influence how they think about their own WTC as a personality trait (MacIntyre, Wang, and Khajavy 2020).Research into WTC has generated extensive bodies of work in both L1 and L2 paradigms, mostly focussed on trait-level measures (Henry, Thorsen, and MacIntyre, forthcoming).However, as Hodis, Bardhan, and Hodis (2010) have argued, there are a number of 'significant problems' with existing WTC research.First, although WTC research began to be published in the late-1980s (McCroskey 1986McCroskey and Richmond 1987), there has not yet been a full description of the developmental processes that underly changes in WTC over time.Second, previous research employing WTC measures has relied on cross-sectional designs to test the network of proposed causal relationships linking WTC with its predictors.Third, previous studies have investigated whether correlates, such as motivation, anxiety, and perceived communication competence measured at one time predict WTC measured at the same point in time.Fourth, research has not identified 'fundamental differences existing between static and dynamic influences of predictors' (Hodis, Bardhan, and Hodis 2010, 251).
Taken together, these criticisms point to a pressing need to understand the complex developmental dynamics of WTC.It can further be noted that the processes by which language learners' communication willingness develops over time have rarely been investigated in the context of authentic, community-based speech events (although see Cameron's (2020) longitudinal study on factors which influence WTC in and out of the classroom).Equally, WTC is yet to be conceptualized beyond interaction in a (single) target language, and development has not been investigated in multilingual contexts.Specifically, little is known about the dynamically shifting patterns of communication initiation propensities in relation to multiple L2s, or about the situated choices that a multilingual language learner makes when communicating in a context where a language (or languages) other than the target language also provide a medium for communication.

WTC in multilingual contexts
When communication is possible in more than one language, the learner's willingness to communicate becomes part of the assembly of the situation.In contexts of multilingual language use, WTC involves assessment of the 'fit' of the different languages in each communication event.In migration situations, a decision can be made about whether to communicate in the host-country language, or in a contact language (a language used by people who do not share a common native language, or who may have different cultural frames of reference) (Canagarajah 2007;Seidlhofer 2005).The processes underlying a person's WTC reflect the integrated push and pull of the different languages, including the implications connected to choosing, or not choosing, to communicate in a particular language (e.g.possibilities for code-switching and/or using nonverbal channels).Thus, WTC reflects a state of communicative readiness that fluctuates with the flexibility of language choice as each communication event unfolds.Through ongoing feedback loops, experiences of the process of making communication choices, and the consequences that these choices have, mean that patterns of WTC will be established, and that these patterns will change over time (Henry, Thorsen, and MacIntyre, forthcoming).
If we are interested in understanding the dynamics of communication-initiation behaviour in contexts of multilingualismthat is, the ways that target language WTC and contact language WTC can change over timeinvestigation is needed at multiple levels, on multiple timescales.In addition to considering changes in the levels of situational WTCas for example can be investigated using observational and idiodynamic methodsit is important to develop an understanding of the influences on the trajectory of a person's longer-term bilingual/multilingual development.It becomes necessary to find out not only whether the person 'is currently going through a moment of language stability or of language change where a language may suddenly acquire new importance whilst another may have less of a role to play' (Grosjean and Byers-Heinlein 2018, 7), but also to identify critical influences or drivers of change.To understand the development of new communication behaviour, it is necessary to focus on what is happening during periods when changes occur that is, points in time where a shift in an observed pattern of WTC takes placeand to explore these shifts in the context of the individual's unfolding personal history.

Investigation of WTC as a process in motion
As language development takes place, there will be periods of continuity in a person's communication propensities, as well as changes that are gradual and those that are abrupt.Emphasis therefore needs to be placed on describing these processes of change.In a companion study, we documented shifts in patterns of WTC over a period of months, identifying periods of stability, gradual change, and sudden transitions (Henry, Thorsen, and MacIntyre, forthcoming).Studies that adopt a process-oriented design seek to shed light on patterns associated with different developmental phenomena.The aim is to make sense of similarities and differences between individual cases, and to generate understandings that move from cases to process characteristics.In short, the goal is to make sense of variability, where both continuity and discontinuity play central roles (Van Dijk and Van Geert 2007).
In process-oriented research, periods of discontinuity or abrupt change will often have particular importance.It is at these points that a system self-organizes into a new stateone qualitatively different from its prior stateand where the emergence of a new skill, ability or disposition can be evidenced by more frequent use.A discontinuous change can signal the arrival of a threshold point, and a shift into a new developmental phase.However, it will rarely signal a wholesale transformation of a previous behaviour.For this reason, an openness to investigating the nature of an observable change, and exploration of the sources of continuous and discontinuous aspects go hand-in-hand (Van Dijk and Van Geert 2007; see also Sternberg and Okaggaki 1989).

Study and purpose
In situations where new opportunities for language use arise, and where linguistic needs and the functions of language constantly change, processes of multilingual development will typically include 'periods of stability, of different durations, and then periods of language reorganization' (Grosjean 2015, 579).In the research immediately preceding the current study, change point analysis (CPA) techniques were used to explore WTC trajectories for a contact language (English) (EWTC) and a target language (Swedish) (SWTC) (Henry, Thorsen, and MacIntyre, forthcoming).In this study, patterns of WTC development revealed differences and consistencies.For all six participants, SWTC increased over the period during which scores were obtained, and EWTC decreased.However, there was substantial inter-individual variation.For several participants, change points were identified.These points can be regarded as indicators of a period of reorganization in a participant's WTC system.
As a next stage in this work, the current study seeks to open a window onto periods where an observable change takes place in a participant's communication-initiation behaviours.Equating a shift in observed values as an indicator of the WTC system's self-organization into a new state, and the emergence of a new behavioural mode, the purpose of the study is to explore the nature and the sources of an observable change (Van Dijk and Van Geert 2007).The following research questions were formulated: RQ1: What features of continuity-discontinuity characterize an observed change?RQ2: What are the sources of this shift in communication behaviours?

Design and methodology
The current study is designed as an individual case study, and a companion to the aforementioned study by Henry, Thorsen, and MacIntyre (Forthcoming).The previous study examined trajectories of development in WTC for English and Swedish from a quantitative, individual-level perspective, and change point analysis was used to identify significant changes in WTC levels over time.The present study takes a person-context dynamics perspective, and employs a person-and process-oriented methodology (Van Der Steen, Steenbeek, and Van Geert 2012;van Geert 2019).In a person-context dynamics perspective, psychological systems are understood to be intertwined with the social environment.The social context is not viewed separately (as an outside influence), but as part of a continuous person-environment loop (Van Der Steen, Steenbeek, and Van Geert 2012).In a person-and processoriented methodology, the person is viewed as an adaptive agent (van Geert 2019).In combination, the person-context dynamics perspective and person-and process-oriented methodology mean that WTC can be investigated and understood as a changing state that unfolds within an active process.

Setting
The study took place in Sweden.In Sweden, community proficiency in English is generally high.English provides an effective communication option in nearly all situations of everyday life.For immigrants who arrive in Sweden with prior knowledge of English, social interaction can be greatly facilitated.English is a valued asset in daily transactions, in classroom meta-communication, and in communication with representatives of state agencies (Henry 2016).
All participants were recruited at the beginning of a fast-track program for migrants with higher education and/or professional backgrounds offered by an educational provider in the west of Sweden 1 .The inclusion criteria for the study were that the participant: (i) had not been living in Sweden for more than 18 months, (ii) possessed functional communication skills in English at CEFR-B1 or above, and (iii) was not an English native-speaker.

Ethics
The research was conducted in accordance with the ethical guidelines of the Swedish Research Council.Participants were informed orally and in writing about (i) the purposes and procedures of the research, (ii) that participation was voluntary, (iii) that withdrawal was possible at any time and without need for explanation, and (iv) that confidentiality would be ensured.

Instrument
To measure the participants' communication orientations, the Willingness to Communicate (WTC) scale (McCroskey 1992(McCroskey , 1997) ) was used.The WTC scale contains 20 items, 12 focal items (that combine communicating with friends, acquaintances and strangers in each of dyads, small groups, large groups, and in public), and 8 items that describe specific communication partners (see Scoring below).Mirror-imaged scales were used.One scale measured English WTC (EWTC) and the other measured Swedish WTC (SWTC).Items were identically worded.For detailed information on instrumentation, see Henry, Thorsen, and MacIntyre (Forthcoming) and the Supporting Information.
The online questionnaire also contained an open-ended item.Participants were invited to describe situations during the preceding period when they had experienced the sense of having made a choice about communicating in English or Swedish.The responses to these open-ended questions were monitored by the second author.When a described event seemed to shed light on a possible change in communication-initiation behaviours, an interview was arranged.Such events could be, for example, a meeting at an employment office where a participant had begun an interview using Swedish, or a situation where they had spoken with a stranger on the street in Swedish.Additionally, and particularly for participants who did not often provide event descriptions, interviews were arranged at strategically spaced points.The purpose of the interviews was to explore the nature and sources of possible changes in WTC.

Quantitative data and analyses
Analyses of the questionnaire data were carried out using change point analysis (CPA) techniques.CPA is a method appropriate for studying dynamics in contexts of language development (MacIntyre et al. 2017;Taylor 2000).CPA enables the analysis of changes in time series data where parametric assumptions cannot be made.CPA identifies points or 'thresholds' along a distribution of values on either side of which characteristics can vary.CPA generates cumulative sum (CUSUM) charts as a means by which significant changes can be identified with an established probability (e.g.p < .05).It uses bootstrapping to provide confidence levels and confidence intervals for changes that are detected (Taylor 2000).CPA makes it possible to identify 'stable states and changes towards other stable states' (Steenbeek,Jansen,& van Geert,201,68,emphasis added).This makes it particularly valuable in understanding developmental processes in open systems.

Participants
In the original study, data from six participants was analysed.The present research focusses on two of these participants, hereafter referred to by the pseudonyms 'Betty' and 'Clara' (respectively P2 and P1 in Henry, Thorsen, and MacIntyre, forthcoming).These participants were selected for the following reasons: . They were the participants who generated data over the longest time period (12 months).
. Both took part in in-depth interviews throughout the period (3 with Betty, 6 with Clara).
. The CPA analyses revealed very different WTC trajectories.Both participants had the same overall pattern of increasing SWTC and decreasing EWTC.Both started with substantially higher EWTC at the beginning of the period.However, for Clara SWTC increased substantially (and, for a time, was higher than EWTC).For Betty, EWTC remained substantially higher than SWTC.
. The sequencing of the observed changes differed between the participants.While for Clara, changes in SWTC and EWTC were sequential (an upwards shift in SWTC substantially preceded a subsequent downwards shift in EWTC), for Betty the changes in SWTC and EWTC occurred in close proximity.
In accordance with the person-context dynamics perspective, and the person-and processoriented methodology (Van Der Steen, Steenbeek, and Van Geert 2012;van Geert 2019), we offer a close examination of shifts in SWTC and EWTC for these two individuals.The WTC trajectories and change points for both participants are provided in Figure 1.

Qualitative data and analyses
In research exploring psychological and educational phenomena from CDST perspectives that emphasize the complex and changing interactions among factors, it is increasingly common to find studies that adopt person-and process-oriented methodologies, and which draw on qualitative and quantitative data (van Geert 2019).While in some studies, analyses of quantitative and qualitative data offer differently nuanced perspectives on an observed phenomenon, in others quantitative data is used to measure outcomes at the group level, and qualitative data is used to generate insights into underlying processes (see e.g.Dalenberg et al. 2016;Kupers et al. 2015).Here, as in the study by Van Vondel and colleagues (2016), quantitative data was used to identify variability, and qualitative data was used to explore the nature and sources of observed changes.
Analyses were carried out in the following steps.First, close readings of the transcripts of the 3 interviews with Betty and the 6 interviews with Clara were independently carried out by the first and second authors 2 .Next, the events described in the interviews were calibrated with the change points identified in the participants' SWTC and EWTC trajectories.Thereafter, and to ensure that these calibrations and the understandings of the events were accurate, additional follow-up interviews were carried out with both participants.Finally, from the events described in the interviews at the observed points of change (and which had been checked with and confirmed by the participants in the follow-up interviews), a narrative account of each observed transformation was created.Narrative provides an important of means of understanding the dynamics of development.It is a tool that can be used by researchers and study participants to make sense of events, patterns, and the underlying sources of change (van Geert 2006).

Results and discussion
With a focus on identified change points, in the sections that follow we offer a narrative of relevant events and situations in the lives of the two women.Both women were motivated in their language learning and shared the goal of becoming established in Swedish society.They saw TL competence as facilitating social interaction and providing employment opportunities.They described enjoying the learning process, and how they actively sought opportunities for TL use that could help them achieve their goals.
At the start of the period, Betty had been living in Sweden for around 18 months.Betty was married to a Swedish citizen, whom she had met while he was on business in South East Asia.In her home country, Betty had worked in an import-export company.She was used to speaking English in work situations, and her goal was to continue her career in Sweden.Betty's two teenaged children had followed her to Sweden, and the family lived in Smalltown (a small rural community).Betty's husband was often away on business.Her children were enrolled in international English-speaking schools in larger towns some distance away.This involved hour-long each-way commutes.Being a social person, Betty said she could often experience feeling lonely and socially isolated.
Clara was an asylum-seeker from a north east African country and had arrived in Sweden through a UNHCR program.At the start of the project, Clara had been in Sweden for just 8 months.During the project period, Clara lived alone.In her home country, Clara had been an upper secondary English teacher.Her goal was to continue her teaching career in Sweden.In the period before she could enrol on a Swedish language program, Clara took it upon herself to learn the language using available resources (primarily YouTube videos).These efforts meant that once she was able to start her education, she progressed rapidly through a series of basic-and intermediatelevel courses.

What features of continuity-discontinuity characterize the change?
In a dynamic system, change can be both gradual and nongradual.In some cases, a discontinuity can be very noticeable.In other cases, shifts from one mode of behaviour to another may be less clear (Van Dijk and Van Geert 2007).Studies that use a person-and process-oriented methodology make it possible to cast light on these variabilities (van Geert 2019).

Betty: gradual change
For Betty, a significant upward change in SWTC (at point 17) is separated by just one measurement point from a significant downward change in EWTC (at point 18) (figure 1).These changes point to a process of self-organization taking place within the WTC system.The temporal proximity of these two shifts, the relatively small increase in values, and the closely-ranged values of preceding and subsequent trajectories, indicate a gradual shift in the WTC system from one relatively stable state to another, and change that can be described as gradual or continuous.
Clara: nongradual change While Betty's SWTC and EWTC trajectories indicate a pattern of gradual self-organization, change in a dynamic system can also be nongradual.For Clara, changes in the WTC system demonstrate a patten of discontinuous development in several ways.First, the periods of change for SWTC and EWTC differ in time.While a significant (upwards) change in SWTC occurs at measurement point 8, for EWTC a corresponding (downwards) change is not observed until measurement point 13.Second, when the first of these changes (SWTC) takes place, the jump in values is substantial (over 30 scale points).Finally, looking at the trajectories for SWTC and EWTC before and after each change point, the patten of development appears generally variable.
What are the sources of shifts in communication behaviours?

Betty
For Betty, the observed changes in SWTC and EWTC occur in close succession, at point 17 (SWTC) and point 18 (EWTC).These points fall at beginning of the summer, the former in middle of May, and the latter at beginning of June.In the interviews prior to this time, Betty talked about the isolation she experienced living in Smalltown.She described long hours of being on her own after classes had finished, her husband away on business, and her children not having returned from school.She recognised how this situation affected her mood and limited her opportunities to communicate with speakers of Swedish.Being a person who liked to learn languages through social interaction -'I mostly like to study Swedish, mostly like with talking in conversation' -Betty viewed social isolation as an obstacle standing in the way of progress.Indeed, it seemed almost an insoluble problem; during the winter people in Smalltown remained indoors, while in the summer they moved away to their summer homes: EXCERPT ONE (February) Yes, all Swedish, yes, they are friendly, friendly you see.But they have no have time for social [interaction], because they are busy working.In the morning they already go to work and then [home] in afternoon.Also, especially in winter they, in winter after working they go home, just sitting in the house.Maybe they do not want to go out.So not lot of social [interaction].And also, during the summer, they have a house, a second house, a summerhouse, they go to the summerhouse and then Smalltown is empty.No, Smalltown is difficult for social [interaction].
In the following interview, which took place at the end of the summer, Betty was more positive about opportunities to speak Swedish.At the beginning of the summer, two important changes had taken place in her life.First, while buying goods in the local store, she met a woman from another South East Asian country, and they had exchanged greetings.A few weeks later she chanced upon this woman again.Following this second encounter, the women began to meet more regularly.Because they did not share an L1 or any other language, their conversations took place in Swedish.The second change that occurred around the start of the summer involved the family's decision to become a foster home for unaccompanied minors.In July, Betty and her family received two boys, one from Afghanistan and another from Albania.In the interview at the beginning of September, Betty described how the summer's events had meant that she had become more confident speaking Swedish, and that learning Swedish had become more enjoyable:

EXCERPT TWO I:
When you go to the employment office, or the civil registration office, do you speak Swedish?Betty: Yes Yes, I also speak English, because I have two refugee children living with me.One from Afghanistan, he can't speak Swedish, but he can speak English.So, with him I speak English.
And with the boy from Albania, he too can't speak Swedish, well, he speaks [some] Swedish.So, we try to speak, we try to speak Swedish with them.
From a person-context dynamics perspective (Van Der Steen, Steenbeek, and Van Geert 2012), Betty's family and social situations can be understood to function as control parameters for the WTC system.A control parameter is a parameter to which the behaviour of a complex dynamic system is sensitive (Larsen-Freeman and Cameron 2008;van Geert 2008).To explain how the trajectory of a complex dynamic system is influenced by its control parameters, Steenbeek and van Geert (2008) use the example of a child's capacity for play that maybe be sensitive to the control parameter: 'experiences with peers'.As peer experiences accumulate (and the control parameter changes), shifts take place within the system.
Here, the arrival of the boys in the family's home increases the opportunities for Betty to speak Swedish.So does the forging of a new friendship.These changed social circumstances also appear to have a positive effect on Betty's well-being.Changes in an important control parameter can trigger a change to the system.Here, increased opportunities for social interaction and growing confidence to use Swedish in social situations bring about a change in the WTC system, where SWTC and EWTC shift into different modes.As Betty describes it in the September interview, she had become more willing to communicate in Swedish in social interactions beyond immediate social relationships (with her friend and the boys).For example, she said that she felt more confident about using Swedish when visiting the employment and social security offices, and that she was less reliant on English: 'If meet my case-worker, for example, I speak Swedish.I try to speak Swedish.If they don't understand, I use English.Sometimes.'.
While CPA is a methodology designed to pinpoint shifts and stabilities (Steenbeek, Jansen, and van Geert 2012), the significant (upwards) shift in SWTC values and corresponding (downwards) shift in EWTC values do not signal any major or discontinuous change.Rather, they draw attention to a period where a gradual change takes place.Accumulated experiences of communication in Swedish with her new-found friend, and with the boys placed in her home, mean that the control parameters for the Betty's WTC system shift.This change has the effect that her propensity to use English in the types of situation made focal in the Willingness to Communicate scale weakens somewhat, while the likelihood of initiating communication in Swedish increases in a similar way.

Clara
Temporal processes of self-organization can take a variety of forms (van Geert 2019).For example, there might be a gradual change that is characterised by a pattern of intrinsic fluctuations that shift the system in the direction of a new, and more generally stable state.In such circumstances, change can be manifested in nondramatic accelerations or decelerations.However, self-organization can also be discontinuous.For example, a distinctive pattern of behaviour can rapidly transform into a very different pattern.Equally, a situation can arise where a pattern of gradual change results in the system becoming critically unstable.In such a state, a single external event or perturbation can result in a sudden shift into a different behavioural pattern.For Clara, the first of the identified change points involves a dramatic (upwards) shift in SWTC.Here, a difference in values of some 30 points before and after measurement point 8 is observed.For EWTC, there is also a sudden shift, this time at measurement point 13.Unrelated in time, and of some magnitude, these shifts in the system's behaviour provide evidence of discontinuous self-organization.

SWTC: reaching a point of self-organized criticality
While Clara made rapid progress in her studies, and was uniformly praised by her teachers, she could often find the classroom environment intimidating.She recognised that she was seen by others as a 'quiet student'.A language teacher herself, Clara was aware of the psychological and social factors constraining her opportunities for development.In her first interview, she identified both her apprehensiveness about speaking Swedish, and her fear of making errors: 'afraid that 'oh if I say this' it will be wrong'.Like Betty, she spoke of limited everyday opportunities for communication, identifying possibilities when shopping and on public transport: 'I can't say it's a very big opportunity, but it could be'.
It is against this background that the shift in SWTC values can be understood.In an interview which took place immediately after measurement point 8, Clara's account of events at this time sheds light on the changes in SWTC that were taking place: EXCERPT THREE Clara: Actually, I've learned here very well.And all the teachers, they are good, their system, so I am planning again to take another test on Monday.I: Yes, good.Clara: So that I can soon start [next-level program] and then finish.I: Yes, right, Clara: I hope so.I: Hopefully.That's the goal, right?Clara: Yes, that's the goal I: Right.When you, because I wanted to interview you because when you answered me, one of your emails when we were corresponding, I saw that you answered me in Swedish, while I wrote in English.
The recognition that she is successfully progressing, and that she profits from teaching methods that suit her needs, boosts Clara's self-confidence.From this point onwards, all her correspondence with the second author was in Swedish.This new-found self-confidence also gains expression in her communication behaviour: EXCERPT FOUR Clara: I really appreciate that all Swedish people, they can understand they can speak they can write [in English].I mean, they can do a lot of things in English, so I don't have any problem when I go anyplace, whether the employment office, whether the shop, whatever.Because if I talk to them in English, then they will try to speak with me in English.But sometimes I also think that if I keep on talking in English, then, if I keep on asking in English, if I keep on writing in English, maybe I will end up not learning Swedish.I: So, when you go to meetings, like the employment office, do you try to use Swedish?Clara: Yeah.Like last time, I wanted to ask about study grants, its money that you can lend, you can borrow, then I went to the employment office but there is one girl who works there, she's from my country so we speak the same mother language.Then I talked with her, and she told me 'Ok wait', because my case-worker was not there.Then another lady, she came, and she starts to speak with me in Swedish.And she asks me why I want to meet my case-worker.Then I told her, 'I want to ask her about a study loan'.Then she said, 'OK I, I should book a time for you'.Then she asks me, 'Do you want an interpreter, a translator do you want a translator?'I told her, 'No I don't want'.But maybe she, she was, 'Just to make sure that she has given all the information correctly'.Then she called an interpreter.Then I told the interpreter, I told him, 'I don't need an interpreter'.
And he said, 'Why?'I told him, 'I want to try my Swedish'.EXCERPT FIVE Clara: I feel a little bit, eh, careful when I speak with the people in employment or other office.But when I speak with people who, standing for buses, or in the shops, I mean it's Ok if I make mistakes.It's Ok.I: Or if you speak in English?Clara: Yeah.And sometimes when I say it, I feel like, 'Was it, was that correct what I said, or was it a mistake?' Three days before, I think, I was waiting for the bus then some people were standing near the bus station, so there was a woman and she dropped something from her, I don't … I: Say it in Swedish Clara: bag / … / She dropped something, but she didn't see it.Then I saw it.So, I was thinking 'can I tell her that you have dropped something?' 'Or can I tell her in Swedish?'And then I was not sure whether, how, 'Have you dropped something?' I don't know.I was not sure of the word for drop.But anyway, I said it, but she understood it.I: She understood?Clara: She understood, and she said 'Thanks' EXCERPT SIX Sometimes I feel like maybe I'm making mistakes, so people might not understand it well.But sometimes I also think that if … .For example, yesterday, one man he was standing next to me in the bus station, so he asked me in Swedish, but he spoke very fast, he asked me, 'Which bus are you taking?'I guess.So, then I just heard some In these three excerpts, Clara tells the story of communication events where she had made a conscious decision to communicate in Swedish.In excerpt four, she describes a recent visit to the employment office.In this communication event, negotiations revolve around the setting up of a meeting with an employment counsellor.The assistant who arranges the meeting is insistent that Clara use the services of an interpreter.Because she thinks Clara has not understood this message, she calls an interpreter to come and explain.When the interpreter arrives, Clara tells him that she does not want an interpreter at the scheduled meeting because she wants to 'try my Swedish'.
Excerpts five and six involve communication events that occur while Clara was waiting for a bus, three days prior to the interview (excerpt five), and on the day immediately preceding the interview (excerpt six).In these communication events, Clara initiates the communication in Swedish (excerpt five) and responds to a question in Swedish (excerpt six).In these situations, communication in Swedish also 'works'.Clara can successfully conclude the transactions, explaining to the woman that she has dropped something (excerpt five), and telling the man the number of the bus (excerpt six).
At certain times, a complex dynamic system can move into a state of critical instability.In such a state, a single external event or perturbation can result in the system's sudden shift into a different behavioural pattern.As van Geert (2019, 160) explains, when a system is in a state of critical instability, 'a particular experience or challenge that might be totally insignificant in the great majority of cases, could have a cascading pattern of serious consequences leading to very rapid irreversible change'.This is what appears to take place in Clara's WTC system.While it is not possible to point to any of these events (or any other event) as being the event, or to exactly identify the point where the shift in system behaviour is triggered, a person-context dynamics lens does enable us to understand how this point of self-organized criticality is reached, and how any of the events that Clara describes (or any other similar event) could have tipped the balance and caused the system to self-organize into a newly emergent state.

EWTC: continuous and discontinuous development
In the case of the significant (downwards) shift in values for EWTC at point 13, there is no indication in Clara's narratives around this time that the WTC system has entered a period of critical instability, similar to that when the shift in SWTC took place at point 8. Non-gradual change does not presuppose an external event or chain of events that trigger a shift.Rather, non-gradual change will often be brought about by the system's intrinsic dynamics, and the numerous 'interactions and couplings between the system's components ' (van Geert 2019, 160).Here, the observed shift in values at point 13 appears to stem from Clara's growing recognition that Swedish can adequately fulfil her basic communicative needs.
The following except is from an interview carried out two days prior to measurement point 13.At this point in time, Clara had successfully completed her intermediate-level language education, and she was due to start a work experience placement at a preschool.She had also begun the process of looking for jobs.Clara described how English no longer played a prominent role in everyday communication: Occurring some months after the upwards shift in SWTC at measurement point 8, the downward shift in EWTC at point 13 can be understood as an effect of intrinsic dynamics within the WTC system.On the one hand, it might be expected that a change in one of the system's components, SWTC, would trigger a change in the other component, EWTC.Indeed, this was the case when change took place in Betty's WTC system.However, when a complex dynamic system shifts into a new mode, it does not automatically follow that there will be a total transformation of behaviour.The emergence of a new pattern of behaviour will not necessarily mean that a previously pertaining pattern will immediately cease, or totally disappear.Rather, following a discontinuous shift, the newly observed mode of behaviour and the old mode can co-exist for a time, development being 'simultaneously continuous and discontinuous' (Van Dijk and Van Geert 2007, 29).
In excerpt seven, Clara reflects on the role that English currently plays in her life.She recalls how, a few days previously, she had told a course-mate that she felt that she was forgetting her English.She tells how she no longer used English when shopping, and that in meetings with agency staff, English was now mostly a fallback.Recalling the words of a former teacher, she reflected on the idea that English can be put aside ('you can park it'), to be used only when needed.
As becomes clear in Clara's narrative, although Swedish may have become the language mainly used for communication in everyday situations, English remains a strong background presence, and an important resource in more demanding interactions.At a point where development is simultaneously continuous and discontinuous, the dynamics exhibited by Clara's WTC system map onto patterns of activation and deactivation conceptualized in Grosjean's (2001) language mode model.During a period of change, the relative importance of the languages in bilingual/multilingual person's repertoire is likely to shift (Dewaele 2001;Grosjean 2001).In a multilingual context, the increasing influence that an emerging L2 has in structuring interactions can mean that the influence of another, previously dominant L2, will weaken, the language being less likely to function as the base language (Grosjean 2015; Grosjean and Byers-Heinlein 2018).In the period in focus, Clara's data sheds light on how, as a base language, the influence of English is in decline.

Conclusion, limitations and future research
L2 WTC has generated unprecedented interest in recent years.Despite this growth, blank spaces in the multi-layered conceptualization remain.There is need for studies outside of language classrooms (as Freiermuth and Ito 2020 have called for) and, in migration contexts, research that takes consideration of a diversity of learner experiences, such as 'language-based rejection sensitivity' (Lou and Noels 2019) and the effects of prevailing language ideologies (Sevinç and Backus 2019).Moreover, and reflecting the situation in language learning psychology generally (Serafini 2017), there is a lack of research exploring developmental processes, and intra-individual complexity.In a time where the world is becoming increasingly multilingual, and where language learning can be distributed across a range of intersecting environments (The Douglas Fir Group 2016), WTC needs to be investigated in settings beyond the classroom, and in contexts where communication can take place in the TL or in other L2s.When researched in these settings, WTC needs to be conceptualized as an integral part of dynamic processes of multilingual growth and loss.
Using a design where individual-level, time-serial data from a WTC questionnaire was combined with narrative data from strategically conducted participant interviews, the current study explored WTC development in a context were a TL and a contact language provided viable communication options.The person-and process-oriented methodology (van Geert 2019) adopted for this study meant that we could cast light on the shape of developmental transformations in WTC, and on the sources of continuous and discontinuous changes.In explaining the findings, we argued that patterns of TL and contact language WTC emerge from system-internal and person-context interactions.
Together with its companion (Henry, Thorsen, and MacIntyre, forthcoming), the study differs from most prior WTC research in both L1 and L2 paradigms.While in previous research, focus has been on trait stabilities, in the current work focus is on change over a period of weeks and months.While most studies have used cross-sectional designs and inferential statistics, we have taken a longitudinal approach and examined intraindividual variability in time-serial data.While previous work has examined outcomes and sought generalizability, we have chosen to focus on dynamic processes, and on the generation of contextualised understandings.
From a traditionalist perspective, then, the current research suffers from a number of limitations.Being individual-level work, findings do not generalize to the group-level (although we would point out that the reverse is similarly true, and that it is not possible to generalize form groups to individual persons).It could also be argued that the value of a study such as this is limited to the description of a particular case(s).However, we would argue that generalizability does indeed become possible.When several time-serial analyses are combined, and when the research objective involves the exploration of processes, rather than tendencies and outcomes, it becomes possible to make generalizations about process characteristics.With the conduct of further individually-focused, and process-oriented studies that explore WTC in multilingual contexts, we believe that it is possible to develop generalizable understandings of the process characteristics of WTC development.With regard to reproducibilityanother criticism that could be levelled at this workvan Geert (2019,176) argues that in individual-level research, the important issue is whether the results make sense in the research context, and whether the methodology is usable by other researchers.In relation to the current findings, and the research design, we believe this is certainly the case.
In future research where WTC is explored using person-focused, context-sensitive designs, we can identify several important objectives.These include (i) more closely exploring interactions among system components (e.g.L2WTC and L3WTC), (ii) exploring the effects that continuous or discontinuous change in one subsystem can have on the other subsystem, and on the WTC system as a whole, and (iii) exploring the effects on the other social and linguistic systems within which the WTC system is nested.Our hope is that the current study may not only provide inspiration for such research, but also a methodological blueprint.

Notes
1. Students enrolled on the program could also study basic-and intermediate-level Swedish for Immigrants (SFI) courses at municipally-operated centres.2. The interview dataset totalled 34 000 words.While initial interviews were carried out in English, by the end of the period interviews with both participants were carried out in Swedish.In the excerpts, translations from Swedish are in italics.

Figure 1 .
Figure 1.SWTC and EWTC trajectories.Note:Each measurement point corresponds to the fortnightly basis on which participants responded to the questionnaire.
. If meet my case-worker, for example, I speak Swedish.I try to speak Swedish.
Yes.For me, it's fun.Because if I can speak Swedish, I can feel better.So, I'm practicing speaking Swedish, by being social and speaking with my friend, because I like to talk, and I don't like just studying from books [laughs] /../ I: Do you also speak English sometimes?Betty: … , 'Which bus are you taking?' something like that.Then I thought, I guessed maybe he's asking me, 'Which bus are you taking?' Then I said, 'Number 2'.Again, he asked another question, but I don't fully understand what he said.Then I say, 'What did you say?' Again, he said it, then I said, 'I don't understand'.Then he tried to explain to me, 'It's a place.Are you going to this place?' 'No', I said, 'No, I am going to Winter Street'.
Maybe.Maybe.I concentrate a lot on Swedish, to practice it.I never speak English when I go to the shops.When I go to the employment office, I sometimes speak Swedish.When I think that I'm not understanding things, then I can switch to English.Sometimes.Not often.I don't.[X], a teacher here, has taught me that … I said to her, 'maybe we will forget English?'.And she said, 'No, you can't forget English, but maybe you can park it, and when you need it, you can use it'.