Thrivers and Survivors during Study Abroad: The Individual Cases of Japanese Learners of English

Abstract

Whilst study abroad (SA) periods hold much promise for foreign language development, increasing research suggests sojourners’ experiences are unique, and language development does not always follow a linear trajectory. For some learners, SA has little impact on their language performance despite the affordances of L2 immersion (the Survivors). Other learners maximise the potential of SA, and this has a positive impact on their language development (the Thrivers). This paper examines the selected cases of four Japanese learners of English and their request language performance during a 10-month SA in the UK. Changes in pragmatic knowledge (based on appropriateness ratings) were documented at three equidistant time points. Language contact profile data also provided quantitative insights into the learners’ extracurricular language use and qualitative personal reflections. The selected cases illustrate two learners surviving the SA experience, showing minimal change in their request performance. The other two learners thrived during SA, showing accelerated performance in terms of lexical variation at the production level. This paper reports on the case histories of these learners to better understand these unique experiences and pragmatic discrepancies. Suggestions for how learners might be more pragmatically successful during SA are also offered.

1. Introduction

Pragmatic competence, a key component of communicative competence (e.g., Bachman 1990; Celce-Murcia et al. 1995), concerns speakers’ ability to understand, interpret, or produce language appropriately in a given sociocultural context (Cohen 2010; Taguchi 2012). It encompasses both knowledge about pragmatics and the ability to use it. Aspects of this competence are believed to be assisted by the study abroad (SA) context (Kinginger 2008, 2009; Pérez Vidal and Shively 2019); as compared to traditional classroom settings, the SA context can provide numerous opportunities for L2 learners to communicate in the L2 and observe how language is used in everyday interactions.

Nevertheless, whilst SA periods hold much promise for L2 development, increasing research has suggested that sojourners’ experiences are unique, and language development does not always follow a linear trajectory. A number of empirical studies (e.g., Barron 2003; Economidou-Kogetsidis and Halenko 2022; Halenko and Economidou-Kogetsidis 2022; Li 2014; Schauer 2009) revealed that there can be much variation in the amount and type of pragmatic gains learners receive during their SA sojourns, thus confirming that the relationship between SA contexts and pragmatic development is a complex one. For instance, learners who do not immerse themselves sufficiently into the L2 culture often miss valuable opportunities for interacting with fluent speakers of the target language or observing authentic interactions in the TL culture, therefore preventing them from receiving a sufficient amount of L2 input. This has been seen as sometimes leading to minimal improvement of their pragmatic fluency or comprehension, among other things. Hence, considerable individual differences in L2 gains often stand out (Pérez Vidal and Shively 2019). These might relate to individual factors, such as motivation, L2 proficiency, learning styles, attitudes, aptitude, personality (for a comprehensive read, see Li et al. 2022), and/or context-generated factors that have to do with the length of stay, amount and type of L2 input, amount and type of feedback received (e.g., on the learners’ inappropriacies), and the extracurricular language use of the language (e.g., L1 use over L2), among others. This complex and dynamic picture, which can involve numerous interrelated factors, could serve to explain why some learners see a direct, positive impact on their language development during SA (the Thrivers), while for others, the same sojourn abroad has little impact on their language performance despite the affordances of L2 immersion (the Survivors).

The present investigation is a longitudinal investigation that falls within the fields of interlanguage pragmatics (ILP) and second language acquisition (SLA) and aims to delve deeper into the role of SA learners’ individual differences when it comes to gains relating to their pragmatic knowledge (based on appropriateness ratings). More specifically, it examined the selected cases of four Japanese learners of English and their request language performance during a 10-month SA sojourn in the UK in order to better understand their unique experiences that might be responsible for any pragmatic discrepancies. By using a mixed methods approach, the study analyses the changes in the learners’ pragmatic ability at three equidistant time points. It used data from the learners’ language contact profiles in order to gain insights into their extracurricular language use, amount and type of input exposure, and other context-related factors that might play a role in their pragmatic development (or lack of it). Qualitative data were additionally used to examine personal reflections on SA experiences.

The chapter begins with a review of the relevant research on pragmatic knowledge and the role of individual variation—primarily that of contextual factors in the form of amount of target language contact in the area of pragmatic development, and mainly in SA contexts (Section 2). Section 3 continues with the presentation of the methods and procedures used in the study. Section 4 presents the results, and Section 5 discusses the findings. Section 6 offers the conclusions and limitations of the investigation.

2. Background: Pragmatic Knowledge and Individual Variation

Pragmatic knowledge or ‘appropriateness’ allows learners to select and use appropriate speech act expressions, involving both pragmalinguistic knowledge (i.e., the ability to appropriately select among different linguistic resources for performing language functions) and sociopragmatic knowledge (i.e., the ability to appropriately evaluate contextual factors) (Leech 1983; Thomas 1983). Within the field of L2 pragmatics, studies have measured learners’ knowledge mainly according to the extent to which L2 speech acts approximates or deviates from native speaker performance (e.g., Rose 2000; Trosborg 1995), although a number of studies have also used appropriateness ratings as indicators of pragmatic knowledge (e.g., Taguchi 2007).

Li (2014), for instance, measured the pragmatic knowledge of 31 American learners of Chinese of different proficiency levels through appropriateness ratings carried out at the beginning and end of their SA sojourn. Li found that both of his two proficiency groups made significant improvements in appropriateness as a result of changing their requests production in terms of modification, alerters, and directness. Taguchi’s (2011) cross-sectional study similarly examined the effect of proficiency and SA experience on the appropriateness of 64 Japanese learners’ requests and opinions in L2 English. Unlike Li’s (2014) study, Taguchi found that SA experience alone did not have any impact on the appropriateness and grammaticality of the learners’ speech acts, although there was a significant proficiency effect on appropriateness ratings. Even more recently, Halenko and Economidou-Kogetsidis’ (2022) study examined the impact of SA on the appropriacy of pragmatic production and linguistic processing efficiency in relation to the requests of 10 Japanese learners in the UK. The study found that appropriateness had the most positive change. Yet, the study equally pinpointed the role of individual differences, as results also revealed much individual variation, ‘highlighting the complexity of examining pragmatic performance over time’ (p. 506) and calling for further research on the understudied issue of individual variation in pragmatics research. The ways in which pragmatic development is highly sensitive to individual differences are explored in the following sections.

Beginning with Taguchi’s (2012) non-SA longitudinal investigation, the pragmatic development of 48 Japanese students of English in an English-medium university in Japan was traced over one academic year. Great variation in the patterns and rate of pragmatic development in eight individual case histories showed that learner characteristics such as motivation, personality, learning styles, strategies, affect, and sociocultural experiences all accounted for the unique trajectories of pragmatic change (or lack of it). Her study confirmed that there was ‘a great deal of intra- and intervariability’ that had to do with ‘immense variation within and among individuals both in terms of when and how they move through key milestones of pragmatic development’ (p. 252). Schauer (2006, 2009) and Woodfield (2012) also attributed mixed successes during SA to individual variation. For instance, Schauer’s (2009) study with SA German learners of English found that the ‘most striking development with regard to SA learners’ request strategy use appear[ed] to be on an individual level, highlighting the importance of individual learner differences’ (p. 196). For example, two of her learners, who initially made use of imperatives, stopped doing so and started using more appropriate indirect strategies instead. This was not something evidenced in the performance of other learners. Along similar lines, Woodfield’s (2012) SA study with eight graduate students in a UK university found evidence of important individual variation despite the fact that all the learners acquired new forms of internal modification. Woodfield’s retrospective interviews with the learners identified the frequency of interaction in the TL community as one of the factors responsible for this variation (Woodfield 2012, p. 41), which is further explored in the next section.

Examining individual variation and L2 contact through the lens of social networks is also a growing trend in SA research. Social networks are the strength of ties between members of social circles (Milroy 1987), and increasing studies have been offering strong links between sustained interpersonal exchanges through social networks and language gains. For instance, McManus (2019) examined the range and sophistication (lexical complexity) of L2 French vocabulary via interviews with 29 SA participants from a British university. Among other findings, higher lexical complexity scores were linked to greater numbers of L2-using contacts reported on a social network questionnaire (SNQ), whilst lower lexical complexity scores were found amongst those reporting more L1-using contacts. More recently, Kennedy Terry (2022) reported on the ‘critical importance of social networks‘ (p. 106) for developing target-like patterns of phonological variation also in L2 French. Findings showed that if learners experienced a full academic year in the L2 context and interacted in French for 2–3 h per day within a social network of L2 users, they were able to more successfully produce phonological variation. A time-limited one-semester stay appeared insufficient for creating the social networks needed to be able to profit from L2 exchanges and develop stylistic variation. On a much larger scale, Baker-Smemoe et al. (2014) explored proficiency gains of 100 English-speaking SA students across six different SA locations (Mexico, Spain, France, Egypt, Russia, and China). Of all the variables examined, which included learner attributes (age, gender, and personality) and pre-SA proficiency, social network variables were the strongest predictor of language gains across all languages. The authors highlighted the high L2 language proficiency of friends and the strengthening of fewer, more meaningful relationships over time to be statistically significant contributors to language gains made. What makes this study particularly important is that regardless of the SA locations, languages, circumstances, experiences, and cultures featured in the data, the benefits of social networks for language development are strikingly consistent. Finally, by linking social networks and pragmatic development during SA in China, Li et al. (2020) found that learners’ pragmatic choices were closely connected to the composition and structure of their social networks. ‘Loose network’ contact clusters, which presumably allow for greater exposure to more varied pragmatic input and feedback, seem to offer greater scope for pragmatic development than ‘dense networks’.

The significance of target language contact has also been addressed by numerous L2 pragmatics studies, especially in relation to SA contexts (e.g., Bardovi-Harlig and Bastos 2011; Matsumura 2003; Taguchi 2008; Taguchi et al. 2013, 2016). Taguchi (2008), for instance, using a language contact survey, examined the correlation between pragmatic gains and the time her Japanese ESL learners spent using the L2 outside the class during their SA semester. Her results showed that learners made significant improvements in comprehension speed and that the amount of speaking and reading outside class significantly correlated with these gains. Similarly, Matsumura (2003), who also examined the pragmatic development of Japanese L2 English learners during a sojourn abroad but in relation to advice-giving, also found that the learners’ pragmatic gains were facilitated more by the amount of their self-reported exposure to English than by their language proficiency.

More recent support for the important role of social contact has been offered by Taguchi et al. (2016) and Sanchez-Hernandez and Alcon-Soler (2019). Taguchi et al. (2016) examined the effects of intercultural competence and the amount of social contact in the development of pragmatic knowledge of American learners of Chinese in SA. The study, which analysed the time the learners spent on different social activities in the L2, revealed that intercultural competence ‘had indirect impact on pragmatic gains through social contact’ (p. 789) and that social contact and cross-cultural adaptability together ‘explained 26% of pragmatic gains’ (p. 775). Along similar lines, Sanchez-Hernandez and Alcon-Soler’s (2019) SA study with 31 Brazilian students in a US university showed that the significant gains the students made in the recognition of pragmatic routines were the result of their sociocultural adaptation and, more specifically, the frequency of interaction with L2 users. The authors conclude that ‘learners’ willingness to acculturate in the SA environment and exposure to recurrent situations outside of the classroom are determinant aspects for routine recognition’ (p. 54).

The present investigation is a follow-up of Halenko and Economidou-Kogetsidis’ (2022) study and an attempt to respond to the call for further research on the overlooked issue of individual variation and the role of social contact and L2 exposure in the learners’ pragmatic development during SA. It, therefore, zooms in on the individual cases of four of the Japanese learners who participated in the 2022 study (two high achievers—the Thrivers—and two low achievers—the Survivors) in order to examine how their pragmatic knowledge in the form of appropriateness develops and how their trajectories of pragmatic change are affected by contextual factors, such as their extracurricular language use.

The research questions guiding the study are the following:

• RQ 1. What are some of the contextual factors (e.g., L2 input exposure, extracurricular L2 language use) that shape the pragmatic developmental trajectories of individual learners during SA?
• RQ 2. Do such factors have an effect on these individual learners’ pragmatic knowledge in terms of appropriateness ratings during SA?

3. Materials and Methods

3.1. The CAPT

Data concerning the learners’ pragmatic development were collected using a computer-animated production task (CAPT), which engaged the learners in virtual role plays and allowed for the elicitation of oral responses. The CAPT included six virtual request scenarios, which took place in the familiar context of a university campus and included four animated interlocutors (i.e., a librarian, a campus security guard, an accommodation officer, and a class tutor). The power relationship between the students and these interlocutors could be considered as high (+P), while the level of social distance (±SD) and degree of imposition (±R) of the requests presented varied. The request scenarios on the CAPT, along with the social distance and imposition constellations, can be seen in

Table 1. Request scenarios on the CAPT.

Scenario	Interlocutor
1. Student requests information on how to book a study room at the library.	Librarian (+SD) (−R)
2. Student requests an intervention with badly behaved students.	Campus security guard (−SD) (−R)
3. Student requests worksheets after a missed class.	Class tutor (−SD) (−R)
4. Student requests permission to change an unsuitable room in accommodation.	Accommodation officer (+SD) (+R)
5. Student requests an extension for an assignment.	Class tutor (−SD) (+R)
6. Student requests a library loan extension beyond the due date.	Librarian (+SD) (+R)

(for a more detailed description of the CAPT, see Halenko and Economidou-Kogetsidis 2022).

The CAPT was administered in a computer lab at three equidistant points in time during an SA sojourn: month 1 (T1: October), month 5 (T2: February), and month 8 (T3: May). Even though participants were given unlimited time to complete the questionnaire, the average time needed for its completion was fifteen minutes. To reduce test effects, the identical scenarios were ordered differently each time.

3.2. Appropriacy Ratings

Appropriateness ratings were the focus of the present study for three main reasons. Firstly, space limitations did not allow for an examination of L2 contact with all appropriateness and fluency measures obtained in Halenko and Economidou-Kogetsidis’ (2022). Secondly, appropriateness ratings were specifically selected because the data showed this measure to have the greatest range of standard deviation within the group, indicating considerable learner differences. Finally, since this study was interested in the effect of L2 contact on learners’ interlanguage, the appropriateness rating was deemed the most suitable of all the measures.

A perception Likert scale provided the means to rate appropriateness. Two EFL tutors (one male and one female) were recruited to rate all the oral requests following the final test stage at T3. These were native speakers of English, had a similar teaching experience of 10–15 years, and both participated in a standardisation session prior to evaluating the data.

The perception scale utilised a six-point Likert scale (0–5), which was drawn from Taguchi (2011) and was further adapted in Halenko (2021). In order to gain a holistic assessment of the request responses, the rating scale included aspects of both pragmalinguistic and sociopragmatic features. The appropriacy scale with the criteria and the rating scores are presented in

Table 2. Appropriacy rating criteria.

Rating Score	Description
5	I would feel completely satisfied with this response because the levels of directness, politeness and formality are almost entirely appropriate and effective for the situation.
4	I would feel very satisfied with this response because the levels of directness, politeness and formality are appropriate and effective for the situation. Where there are non-target-like features, these are minor and unlikely to affect a positive outcome.
3	I would feel satisfied with this response because the levels of directness, politeness and formality are generally appropriate and effective. The expressions may contain several non-target-like features, but the expression would be regarded as achieving minimal levels of appropriateness for a positive outcome, nevertheless.
2	I would not feel very satisfied with this response because the levels of directness, politeness or formality are not sufficiently appropriate or effective for the situation. Features are more inappropriate than appropriate and fail to achieve the satisfactory levels of indirectness, or the expressions contain insufficient mitigation for a positive outcome.
1	I would not feel satisfied at all with this response because the levels of directness, politeness and formality are entirely inappropriate and ineffective for the situation. It is difficult to imagine a positive outcome could be offered.
0	No response provided.

. Raters were asked to read through the transcribed requests and provide a rating based on this scale. Responses that differed by more than one point within the scale were reviewed and consensually adjusted. High levels of interrater reliability were achieved (0.85).

3.3. The Four Case Studies—Background and Language Contact Profile

Four individual learners—two females and two males—provided the case studies for the present investigation. Case selection was systematic and purposeful (Paltridge and Phakiti 2010) as we opted for a sample of four contrastive cases (Duff 2007), which represented low and high developmental trajectories to demonstrate diversity. This was achieved by calculating T1–T3 gain scores on the measure of appropriateness. The two students who featured at the top (i.e., the Thrivers) and another two who featured at the bottom (i.e., the Survivors) were selected as the focal participants for this multiple-case study.

Hikari and Chiaki (pseudonyms) were the two selected high performers (Thrivers), while Taisei and Yui (pseudonyms) were the two selected low performers (Survivors). All four were Japanese undergraduate exchange students who were spending a ten-month sojourn abroad at a UK university, enrolled on various teacher training and business communication programmes. They all met the CEFR B2 level requirement of the exchange programme and received an additional six hours per week of English language classes. This multiple-case study intended to be both relational (examining how variables are related to one another) and explanatory (answering how and why questions) (Duff 2011) to yield insights into the observed individual trends in pragmatic growth.

The informants’ appropriateness scores (Halenko and Economidou-Kogetsidis 2022) were explored alongside the original data included in this paper. Firstly, background information on each informant was compiled.

Table 3. Informants’ background information.

	Hikari (High)	Chiaki (High)	Taisei (Low)	Yui (Low)
Gender	Female	Female	Male	Male
Age	20	21	20	20
No. of years of formal lang. learning	6	6	6	7
No. of years formal English lang. learning	6	6	6	7
No. of years of informal English lang. learning	6	3	0	7
Time spent abroad (purpose)	0	0	1 month (study)	2 weeks (study)

displays information in relation to the four informants’ backgrounds, i.e., their gender, age, number of years of formal language learning (any foreign language) before university, number of years of formal English language learning (e.g., classroom-based instruction) and informal English language learning (e.g., socially with L1 speakers) before university, and amount of previous time spent in an English-speaking country.

Secondly, a ten-item Language Contact Profile (LCP) was administered at the same three time points as the language tests (T1—October (month 1); T2—Feb (month 5); T3—May (month 8)). The LCP employed here is an adapted version of Freed et al.’s (2004) survey, which captures the amount of L2 use across different situations. The instrument relies on eliciting perceptions of L2 contact, which, although risks learners under- or over-estimating actual behaviour, has been used in a range of studies of this kind to measure language contact (Bardovi-Harlig and Bastos 2011; Matsumura 2003; Taguchi 2008; Taguchi et al. 2013; Taguchi et al. 2016).

The LCP used for this study recorded how many hours per day and how many days per week the learners spent using English outside of class across three areas: (a) study-related activities, e.g., completing homework in English, (b) non-study-related activities, e.g., communicating with other students in English, or (c) English-based activities, e.g., watching English movies. In this way, the LCP provided information on both the types and frequency of L2 contact. Learners were further encouraged to provide open comments on their experiences for each category. Data on L1 Japanese use (covering all modes and formats of L1 use) were also elicited for comparison.

4. Results

This section first provides an overview of the LCP data gathered T1–T3 for the entire SA cohort. Presenting group-level data helps to illustrate the extent of individual variation.

Table 4. SA cohort’s total hours of L2 English and L1 Japanese use per week (T1–T3).

	English Use (Hours) (Study-Related)			English Use (Hours) (Non-Study-Related)			Out-of-Class Activities Using English (Hours)			L1 Japanese Use (Hours)
	T1	T2	T3	T1	T2	T3	T1	T2	T3	T1	T2	T3
Chiaki *	6	15.5	2	66	69	40.75	27	36.5	31	1.5	15	17.5
Chika	21.5	14.5	29.5	11.5	18.5	8.75	7	8	10	17.5	12.5	10.5
Hikari *	54	18	43.5	72	71.5	25.5	42.5	28.5	19	2.5	2.5	3.5
Koki	45	11.5	9	51	29	4.75	16	21	23	35	24.5	17.5
Madoka	31	8.5	26	43.5	24	8.5	65	52.5	42	35	17.5	3
Maiko	7	14	16.5	26	19.5	12.25	15	18	17.5	24.5	7.5	7
Masafumi	20.5	18	16.5	24	39.5	18.75	30.5	30.5	30.5	31.5	17.5	15
Shota	20	25	22.5	43	72	50.25	24.5	31.5	31.5	17.5	17.5	17.5
Taisei *	13.5	15	9	38.5	37	2.75	21	6.5	6	17.5	17.5	10.5
Yui *	5.5	20.5	20.5	6	8	2.75	6.5	11	6.5	35	35	35
Average	22.4	16.1	19.5	38.2	38.8	17.5	25.5	24.4	21.7	21.8	16.7	13.7

Notes: Learners are listed alphabetically. ‘Study-related use’ = completing homework, communicating with teachers, and general study-related activities. ‘Non-study-related use’ = communicating with other students, communicating with service personnel, and general non-study-related activities. ‘Out of class activities’ = reading in English, watching English TV/movies, using English online. * = The case study informants.

illustrates the self-reported data based on each participant’s total time per week spent using English in each of the three categories at each time interval (T1, T2, and T3). The table summarises the frequency of L2 interaction and types of L2 contact. The scores were achieved by calculating the product of the hours per day multiplied by days per week. L1 Japanese use is also documented here. The case study informants are marked as *.

By beginning with brief observations at the group level, average English use firstly indicates that, across all time points, English was most commonly used for communicating with peers/service personnel and other non-study-related activities. This shows that English was frequently used in a productive capacity, with most learners actively engaging with L2. English was also used widely for a range of out-of-class activities, such as reading and online use. ‘Study-related purposes’ displayed the lowest frequency. Secondly, with the exception of ‘English for study’, the data tended to show a peak in L2 use at T2 (mid-way point of the SA), followed by a decrease at T3, lower than that reported at T1. Finally, a notable T1–T3 decrease of 8.1 average weekly hours of L1 Japanese use was recorded across the SA period. Typically, the group generally reported using L1 Japanese less frequently than L2 English on a weekly basis.

Though the average weekly hours of L2 use do not fluctuate to a large degree across time for any category (the T3 dip of ‘non-study-related’ English use is an exception), individually reported language contact within each of the three categories shows considerable variation. T1 and T3 ‘study-related English’ use ranged from 5.5 h per week to 54 h per week and 2 h per week to 43.5 h per week, respectively. Similar dramatic T1 and T3 variations also emerged within ‘non-study-related use’ (T1—6 h to 66 h; T3—2.75 h to 50.25) and within ‘out of class activities’ (T1—6.5 h to 42.5 h; T3—6 h to 42 h). These marked discrepancies are the primary interest of this paper and are explored further using the case histories of the Thrivers and Survivors in the next sections. First, L2 contact data were examined, followed by a consideration of how L2 contact mediates request development in terms of appropriateness ratings.

4.1. The Thrivers

Chiaki and Hikari were identified as some of the most successful study abroaders within the group in terms of their performance on appropriacy tests. We begin by examining each of the learners’ individual journeys.

4.1.1. Chiaki

Chiaki’s case history of L2 contact is an encouraging one. Following the group trend, she spent most of her SA time using English to communicate with other students, service personnel, or other non-study-related activities (using English online tops these activities). Weekly reported frequency was lowest at T1 (66 h), peaked at T2 (69 h), and decreased at T3 (40.75 h). Still, the rate was (or was almost) double the group average at each time point. This shows her consistent commitment to using L2 English as a social activity ‘with others’ as opposed to her less frequent use of English for other out-of-class activities, which, for this category, was not discernibly higher than the group average. With the exception of T2, Chiaki reportedly spent well below average time using English for study-related purposes at T1 (6 h) and T3 (2 h). Meanwhile, she reported typically using high levels of L1 Japanese, too, which, at T2 and T3, were either slightly above or below group average use.

The qualitative data from the LCP reveal that she embraced opportunities to participate in extracurricular activities. She reported that she ‘enjoys meeting other people’ and singles out activities, such as ‘eating British food’ and ‘going to the pub’, as particular highlights of her social calendar. She also commented on enjoying ‘shopping with friends’. Taken together with the L2 contact data, she seemed to be making the most out of her first overseas experience. Chiaki is the eldest student of the group (by one year) but reported fewer years of informal English language learning (half of the number of years recorded by her peers).

Turning to Chiaki’s performance on appropriacy tests, Figure 1 presents Chiaki’s change in pragmatic measures in relation to group means. Figure 1 illustrates that, beyond T1, she consistently performs above the group mean the longer she is active in the SA environment. In other words, her request responses are deemed to be increasingly effective and situation-appropriate over time in terms of directness, politeness, and formality (see Table 5).

Chiaki’s biggest improvement was made between T1 and T2 (October–February), with a further slight improvement in the latter stages of her SA stay (February–May). From a linguistic perspective, Chiaki was one of two students regularly producing mitigating expressions using bi-clausal structures (e.g., ‘I was wondering if’) at T2 and T3. Her levels of indirect requests appropriate to each situation were consistent from the outset, but the choice of strategy changed. At T1, Chiaki relied on permission statements such as ‘Can I’, ‘Could I’, and ‘May I’, but by T2, bi-clausals became a regular and noticeable feature of her interlanguage. The range of bi-clausals employed was also impressive. She accurately reproduced the formulaic tokens ‘I was wondering if’, ‘is it possible to’, ‘would it be possible’, and ‘would you mind’ interchangeably in her request output. These expressions clearly had a positive impact on the raters, and she was rewarded accordingly. To a lesser extent, there is also evidence of external modification, which was absent in the early SA period. Though not consistently applied, reasons (Grounders) and Apologies also started to appear at T2 and were maintained at T3. All these modifications are likely to have contributed to Chiaki’s improved appropriateness scores. A representative example of the changes is illustrated below:

Table 5. Example T1–T3 request output (Chiaki). Scenario 4: Student requests changing an unsuitable room in the accommodation building.

T1	T2	T3
Hello, excuse me. I have, er, some problems with my accommodation. I’d like move, er, to move to another accommodation. Could you err, could you suggest another accommodation for me?	Hi. I am a resident of this accommodation. I’m sorry but I am having some problems so, would you mind if I ask you to solvey my problems? umm … is it possible to … err, ask you to help me to find the other accommodation which I can … move to?	Excuse me. I’m a student living in this accommodation. I have some problems with the other people, so, I want to move to the other accommodation. I’m wondering if, err…, I could ask you to help me to find a new place to live?

Table 5. Example T1–T3 request output (Chiaki). Scenario 4: Student requests changing an unsuitable room in the accommodation building.

T1	T2	T3
Hello, excuse me. I have, er, some problems with my accommodation. I’d like move, er, to move to another accommodation. Could you err, could you suggest another accommodation for me?	Hi. I am a resident of this accommodation. I’m sorry but I am having some problems so, would you mind if I ask you to solvey my problems? umm … is it possible to … err, ask you to help me to find the other accommodation which I can … move to?	Excuse me. I’m a student living in this accommodation. I have some problems with the other people, so, I want to move to the other accommodation. I’m wondering if, err…, I could ask you to help me to find a new place to live?

4.1.2. Hikari

Whilst also performing as a Thriver in terms of pragmatic development, Hikari’s journey contrasts with that of Chiaki in several ways. Similar to Chiaki, Hikari’s preference for L2 use was for social activities outside of her studies. In fact, Hikari reported the highest levels of English use in this category amongst the entire group at T1 and T2 (T1—72 h; T2—71.5 h). Despite a sharp drop in L2 use at T3 (25.5 h), the rate was still higher than the group average of 17.5 h at that stage. Although a decrease over time is noted, Hikari also spent considerable time using English for activities outside of class, independent of others. She spent most of her time ‘reading’, particularly at T1 (42.5 h against the group average of 25.5 h). Overall, unlike Chiaki, she demonstrated high levels of T1 and T2 English use, as much as with others as on her own. Unpredictably, these high levels decreased by T3, whilst Chiaki maintained her rate of L2 use in both areas. Again, in contrast to Chiaki, Hikari also spent a large proportion of time using English for study purposes. Despite an uncharacteristic dip at T2, Hikari again spent the most weekly hours using English for study reasons out of the entire group (T1—54 h; T2—18 h; T3—43.5 h). A final feature to note is Hikari’s minimal levels of L1 Japanese use during the SA stay. Her weekly reported hours of first language use represent the lowest across the entire group at every time point—her highest being 3.5 h at T3, against a group average of 13.7 h. This aspect stands in contrast to Chiaki’s reported L1 use, which remained consistently high throughout.

Hikari’s comments on the LCP reveal a strong commitment to the SA experience. She mentioned feeling ‘lucky’ to have the opportunity to be in the UK, suggesting that she wanted to make her SA experience meaningful from the outset. Her profile information showed that this was her first time visiting overseas. Hikari’s enthusiasm seemed to be reflected in her all-around engagement in the different study- and non-study-related activities, in addition to her infrequent use of L1 Japanese in comparison to her peers. Her memorable activities were ‘travelling’ and ‘meeting new friends’.

Examining Hikari’s journey in terms of appropriacy scores, she demonstrated progress at a slower rate than Chiaki. Her appropriacy scores evidence smaller improvements across time. As can be seen in Figure 2, her T2 progress took a leap forward, but only marginal gains were made beyond this point by T3.

Hikari made fewer marked linguistic changes. Her request for modifications over time was more subtle in contrast to Chiaki’s use of impactful mitigated expressions. Hikari developed the skill of including more external modifications, which appeared for the first time at T2 and was maintained at T3. Several of her responses started to include initial Preparators (‘could I ask you a favour’, ‘could you do me a favour’) and specific reasons for the request (Grounders). Within the head act, there were also increased (but limited) instances of internal modification from T2. For instance, Hikari acquired the adverb ‘kindly’ at some stage of her SA experience and often used this as a go-to hedging mitigator to soften the requestive force. Overall, the changes in her request output (see

Table 6. Example T1–T3 request output (Hikari). Scenario 5: Student requests an extension to an assignment.

T1	T2	T3
Hello … I couldn’t finish my homework so I have a paper I couldn’t finish. I couldn’t finish my homework so would you give me more time?	Hello. Can I ask a favour about our class? Err … I think I need more time to finish my homework. I’d like to think about the topic of my homework deeply, so, could you kindly give me more time?	Hello. Er.. Do you kindly have time to speak me please? Err … I … you gave my homework yesterday, but I think I need more time to finish it, so, can I … can I submit it tomorrow please?

for examples ) are evaluated as being ‘more satisfactory’, but the nuanced modifications may account for the more modest gains observed in her appropriateness scores.

In summary, the Thrivers exhibit both active participation in the SA experience and positive, pragmatic development in their request behaviour. Chiaki could be characterised as the ‘socialite’, as her enthusiasm for integrating into the local context over other activities appeared to positively impact the appropriateness of her requests over time. Hikari could be described as the ‘all-rounder’. She appeared to engage in study- and non-study-related activities in equal measure, which also appears to have some performance-enhancing effect.

4.2. The Survivors

Taisei and Yui were identified as low-performing students in terms of overall pragmatic growth. Here, their case histories are reviewed against their reported L2 use and appropriateness of requests over time.

4.2.1. Taisei

An initial feature to note in Table 4 is Taisei’s reported lower-than-average L2 use across all categories at each time point. The early months of SA, however, tend to evidence only marginal differences between the group average (T2 English for study—1.1 h; T2 English for non-study—1.8 h; T1 English for out-of-class activities—4.5 h). In fact, in line with the group trend, Taisei showed a clear T1 and T2 preference for using English ‘with others’ for non-study purposes, generally maintaining group average levels. In this area, at least, he showed a steady commitment to English use. By T3, however, there was a dramatic downward shift. Reported weekly L2 use decreases dramatically in all three categories compared with the group average (English for study—10.5 h; English for non-study—14.75 h; English out of class—15.7 h). By the end of the SA stay, there was a distinct withdrawal from L2 engagement. Taisei’s reported L2 use across all categories was considerably less than half the group average. In terms of L1 Japanese use, Taisei maintained slightly higher (T2) or slightly lower (T1 and T3) levels than the group average. Overall, he tended to use L2 English and L1 Japanese in similar measures.

The qualitative aspect of the LCP does not reveal much insight into Taisei’s SA experience. He refrained from sharing specific examples but rather offered general comments that English is ‘sometimes difficult’. He also observed that there is ‘little opportunity to speak English’, which may reveal some reluctance to seek out opportunities to interact. Interestingly, Taisei had previous overseas experience in the UK of one month in a study capacity (the longest of the group) and a similar history of prior English study. Still, this seems to have had little lasting effect on his present engagement or pragmatic development.

The evaluation of appropriateness for Taisei’s request responses seems to link to the trend observed in his L2 contact. As can be seen in Figure 3, in the early months, Taisei’s requests were assessed on a par with the group average (T1) or slightly above it (T2). By T3, there was a slight T3 dip in appropriacy scores, which seemed to mirror his (partial) withdrawal from L2 engagement towards the end of SA (see L2 contact section).

Unlike the ‘Thrivers’, there are no discernible changes in Taisei’s request behaviour over time. Some T2 (and even T3) requests included one-off attempts at mitigation, which possibly accounted for the slight increases in appropriateness scores (e.g., Confirmations—‘is it possible?’; Grounders—‘I had a bad cold’; Attention-getters—‘Can I ask you something?’). However, since these were not replicated beyond these isolated incidents, the changes cannot be considered fully acquired in his interlanguage. The inconsistency of these situationally appropriate features is likely to be reflected in his lower-than-average appropriateness scores since otherwise, Taisei’s T1–T3 requests typically relied on a mix of Preparatory Questions (e.g., ‘Could you’) and Want Statements (‘I want’, ‘I need’) (see

Table 7. Example T1–T3 request output (Taisei). Scenario 6: Student requests extending a library loan beyond the due date.

T1	T2	T3
Excuse me … I borrowed this book but I want to borrow until next week. Could you do it?	Excuse me. Can I ask you something? Now, I have … library book … at home and I want to … borrow it until next week to finish my work. Is it possible?	Excuse me. I have a book in my house and now … I want to ask you … I want to borrow the book until next week because I need to finish my work and … I need to have some time to read it, until next week … so it is possible?

for examples).

4.2.2. Yui

In comparison to Taisei, who showed early bursts of L2 engagement with others, Table 4 illustrates that Yui’s L2 use was consistently low in this aspect. Yui showed little active L2 participation, whether with others (highest 8 h; lowest 2.75 h) or using English to engage in independent activities outside of class (highest 11 h; lowest 6.5 h). In most categories and time points, he reported the lowest frequency of weekly L2 use within the group, with his weekly average rarely reaching double figures. An interesting feature that seems to buck this trend, however, lies in his use of English for study purposes. At the start of SA in October, Yui reportedly dedicated 5.5 per week on average. By Feb (T2) and May (T3), this had increased almost fourfold to 20.5 h per week, which was higher than the group average. Yui was also one of the students who used L1 Japanese the most and at consistently high levels across the entire SA period. His weekly average was almost double that of the overall group at each time point, revealing a higher frequency of L1 Japanese use than L2 English. No additional qualitative data appeared on the LCP, offering personal reflections on his SA experience.

As far as Yui’s appropriacy scores are concerned, his strategies for executing his requests maintained a similar basic pattern throughout T1–T3. As can be observed in Figure 4, at the beginning of SA, his responses were evaluated on a level with the group average. As time passed, however, there was a steady T2 decline, followed by a sharp decrease in appropriateness at T3.

Few meaningful observations can be made about Yui’s requests since there are no discernible differences between his request production at the beginning, middle, or end of SA. Overall, the requests were brief and typically loaded with hedged/unhedged imperatives (e.g., ‘Please give me/give me’), Want Statements (e.g., ‘I want/I need’) and Preparatory Questions (e.g., ‘Could you’). There is little, if any, evidence of internal or external modification at any time point. Yui depended on routines, which were already familiar, and showed no signs of creativity or experimentation. Unsurprisingly, his appropriacy scores remained consistently low (below the group average) throughout. He mostly achieves scores that equated to either ‘unsatisfactory’ or ‘of minimal satisfaction’ on the appropriacy scale (see

Table 8. Example T1–T3 request output (Yui). Scenario 3: Student requests worksheets from a tutor after a missed class.

T1	T2	T3
Hello. I missed my English class this week so I want the worksheets. Can you give this sheet?	Hello. I’m sorry not to be in classes this week. Could I have the worksheets from the classes please?	Hello, Professor. I’m sorry not to attend your class. Err, I would like the worksheets of the class.

for examples).

In summary, the Survivors’ SA experience is not an encouraging one. Their turbulent journeys seem to negatively contribute to the quality and quantity of their pragmatic development. Taisei could be described as a ‘shape-shifter’ since his early L2 engagement and request behaviour show promise, but this takes an unpredictable downturn towards the end of his stay. Yui, on the other hand, exhibits consistently low L2 engagement and interaction. He may be considered more of a reluctant ‘homebody’ since much of his time is reportedly spent studying language rather than engaging in meaningful L2 practice.

5. Discussion

This paper examined potential relationships between the frequency and types of L2 contact during a university-based SA (RQ1) and the possible effects on appropriateness ratings of request language produced over time (RQ2). In this section, both research questions are addressed simultaneously since the data suggest that they are closely connected.

Commenting firstly on the frequency of L2 contact, the variable levels of L2 use seem to correspond with high and low assessments of request appropriateness over the ten-month SA period. Put simply, the greater the interaction in the L2 community, the more positive the impact on request production. This finding adds further weight to evidence suggesting that the frequency of L2 interaction, more than the length of stay, for instance, is a likely condition for pragmatic development (Bardovi-Harlig and Bastos 2011; Sanchez-Hernandez and Alcon-Soler 2019; Taguchi et al. 2016). Chiaki’s story offers the best example of this link. Chiaki’s frequent and extended interaction across a range of social situations and time points is likely to have contributed to her improved situationally appropriate requests. Chiaki also reported spending most of her out-of-class time using English online, furthering the potential for interactive opportunities. As with learners in Woodfield’s (2012) request study, who were also monitored three times over a similar eight-month period, Chiaki may be able to attribute her linguistic gains to the social contact she experienced. Unusually, Chiaki’s growth was most apparent in her use of the formal register, which is a little perplexing since such linguistic forms might be considered atypical in the informal situations she reported on the LCP, such as going to the pub and shopping with friends. Still, she managed to internalise the formulaic tokens and recall them in a productive way from T2.

Whilst Chiaki was highly functional in L2 English, Yui, on the other hand, was not. Yui primarily used L1 Japanese (the highest level in the group) to function during SA. Though he was not new to an SA experience, he appeared to fail to connect with the target language community. Such levels of L1 use (almost double the group average at each time point) suggest he was very sociable (and perhaps felt more comfortable) with his Japanese peer group, yet he had no-to-limited contact with target language speakers where he would be exposed to authentic language and practice. These defining features of his SA are probable explanations for why he made no progress in his request language, receiving low appropriateness scores throughout. Hikari and Taisei’s stories further support the evidence presented so far. Their case histories offered a mixed picture of the learner experience in that when they interacted in the L2, good things happened to their pragmatic development. When they did not interact in the L2, intentionally or not, this severely limited their potential for pragmatic growth, as evidenced by their appropriateness scores. Related to these findings of high and low networking, both Baker-Smemoe et al. (2014) and McManus (2019) reported gradual decreases in sojourn social networking over time but highlighted the positive impacts on the language development of having fewer, stronger, and deeper social relationships towards the end of a SA period. This could also possibly explain why, at both the group and individual levels, learners in this study self-reported a decrease in their overall English use between T2 and T3 as their network size reduced over time but became more meaningful and (pragmatically) influential for learners such as Chiaki and Hikari. Unlike language contact studies, which are commonly limited to semester-long SA programmes and one or two data points (Sanchez-Hernandez and Alcon-Soler 2019; Taguchi 2008; Taguchi et al. 2013, 2016), this longitudinal investigation enabled data capture at three time points to showcase the changeability of learners’ experiences. It is recommended that future studies also adopt this approach to gain a more nuanced picture of how SA unfolds for different learners at different stages.

The focus on the types of L2 interaction in the data also makes an innovative contribution to SA research. Moyer’s (2005) useful categorisation of interactive and non-interactive domains helps differentiate which types of L2 contact have the best potential for negotiation of meaning, furthering language development, but few studies investigate this distinction. The self-reported data in this study empirically support activities in the interactive domain, involving face-to-face communication with others, to be superior to those in the non-interactive domain for the case histories described. Each of the learner experiences offers evidence for this by way of the presence or absence of L2 social contact and the relative effects this has on the appropriateness of the requests. According to their LCPs, Hikari and particularly Yui adopted a passive, form-oriented approach to language learning through academic study, which did not seem to pay off from an appropriateness perspective. This finding was similar to that of some of Taguchi’s (2012) learners whose preference for good study habits (non-interactive domain) did not necessarily facilitate improved pragmatic development or compensate for the benefits afforded to those who had frequent and extended multi-party interactions (interactive domain). In this and the present study, occasional and brief interactive episodes, for example, with class tutors, were insufficient for effectuating a positive change in performance. In contrast, exploiting opportunities for sustained speaking practice, as demonstrated in Chiaki’s case, was critical for development. That said, her firm control of the pragmalinguistic aspects of bi-clausal structures did not always match her sociopragmatic expertise. Sometimes, she displayed an overdependence on bi-clausals as politeness markers, unable to differentiate between high and low requests. This struggle may support the argument that modelling and feedback (Sanchez-Hernandez and Alcon-Soler 2019; Taguchi et al. 2016) and/or communicative class instruction (Moyer 2005) are equally influential in refining and modifying the pragmatic output.

Related to this observation is the notion that ‘availability of practice does not guarantee accessibility’ (Taguchi et al. 2016, p. 789). As observed with the informants in this study, learners’ agency to access opportunities may be closely tied to individual personalities and attitudes. Those who were socially proactive and exhibited outgoing personalities, i.e., Chiaki (and to some extent Hikari and Taisei), seemed to profit the most and were rewarded with incidental opportunities for pragmatic learning. Similarly, Isabelli-García (2006) also noted that high levels of motivation seemed to lead to the development of more extensive social networks. Those who retracted from social contact, such as Yui, perhaps due to having a more introverted personality and/or relying on academic study as a main means of language development, were unable to exploit the learning opportunities presented before them. As a result, this may have led to other limiting factors, such as a lack of self-confidence (to meet others socially or express ideas in English). Drawing on Bandura’s (1977) early work on self-efficacy, which is broadly defined as one’s self-belief in managing tasks successfully, Petersdotter et al. (2017) found that high numbers of social contacts seem to ‘play a decisive role in developing higher [perceived] self-efficacy while sojourning’ (p. 177). It is reasonable to conclude that the readiness to engage in the L2 (driven by an outgoing personality and positive attitude) led to both affective and linguistic gains for the informants in this study, too.

6. Conclusions

This paper was motivated by a curiosity to understand intriguing levels of pragmatic variance recorded during SA (Halenko and Economidou-Kogetsidis 2022). The variables of L2 language contact and L2 frequency of interaction were examined alongside linguistic measures of appropriateness to uncover possible links to explain the diversity of pragmatic growth. Two high-performing (the Thrivers) and two low-performing (the Survivors) informants provided case study histories to help illustrate this variability.

The findings generally indicated a relationship between L2 contact and pragmatic growth in the form of request production. In doing so, our study adds some weight to the debate that frequency of interaction can predict L2 pragmatic learning. Furthermore, the findings highlight that investigating variables characterising the stay (types of L2 contact) can offer a more fine-grained analysis than simply capturing the frequency of contact and length of overall stay. Since not all types of L2 contact are equal in their potential to facilitate active engagement or allow learners to apply their knowledge (interactive vs. non-interactive), understanding the detail of how learners occupy their time during SA can provide better insights into pragmatic variability—an often-cited feature of SA research. The Survivors’ disengagement in the L2 and under-achievement in terms of appropriateness scores are as characteristic and revealing as the Thrivers’ success stories. Both the Thrivers’ and Survivors’ case histories point to why frequent and sustained L2 social interaction is critical. The informants’ stories also suggest a range of other critical features, such as social skills, learner agency, and personality traits, to be influential. Though not the specific focus of this study, performance based on gender may also be an issue for future exploration since our high-achievers were female and our low-achievers were male.

We must acknowledge that the findings have limited interpretation. The study is small in scope, extending existing data to focus on the L2 contact experiences of one subset of intermediate learners from Japan studying at a British university. Generalisations to other populations are, therefore, challenging. Additional qualitative data, eliciting descriptive accounts of the amount and nature of interaction, could have provided explanations for some of the sudden, uncharacteristic changes apparent in the data. For example, the underlying causes of the Survivors’ linguistic and social decline at the end of SA could have been usefully captured through qualitative means. It is also important to remember that the suggested relationship between L2 contact and appropriateness can only be indirect since the self-report LCP elicits perceived exposure rather than actual behaviour (Taguchi 2012).

Considering the pedagogical implications, we are left wondering how to facilitate more L2 contact during SA, given the cumulative evidence of positive effects on pragmatic comprehension and production. One recent endeavour (Ngai et al. 2020) offers a promising way of developing intercultural and communicative competencies by extending existing engagement with social networking sites (SNSs) for educational purposes. The authors found active use of SNSs for social interactions was key to developing communicative competence and, therefore, predicted that SNS-based learning assignments have the potential to create and maintain international relationships between learners and develop skills required for effective intercultural interactions. Additionally, since social relationships are typically slow to develop, Kennedy Terry (2022) proposed facilitating meaningful relationships with target language speakers in the form of offering homestays, organising social events, or creating language exchange or buddy initiatives. Such ideas for increasing L2 interaction and cultural awareness are to be applauded and should serve as a basis for future empirical study.