1. Introduction

1L2 learning learning—i.e. the learning of any “additional language” (The Douglas Fir Group, 2016, p. 21)—is a long process. Therefore, the time variable is inherently associated with language development, and interacts with other learner-internal variables to alter the nature and efficiency of L2 learning across individuals. Adults who enroll in L2 courses for professional development purposes bring along a complex range of personal characteristics, academic achievements, socio-economic and cultural backgrounds, professional experiences, in addition to distinct L2 proficiency levels, experiences and needs. All these individual variables are necessarily time-dependent and model the learning trajectories that keep evolving consistently.

2The role of learner variation in L2 learning cannot be overlooked as attested by the large body of psycholinguistic and SLA research into Individual Differences (IDs). It is indeed misguiding to evaluate the role of learner characteristics without taking time and context into account (Dörnyei & Ryan, 2015). It is even more crucial for adult learners’ given the time scale during which IDs have developed as recently highlighted by Adringa and Dabrowska (2019). As a result, and in the absence of institutional structuring and research-based L2 teaching practices (Brudermann and coll., 2016), designing L2 courses—particularly in current Higher Education (HE) settings that are increasingly challenged by massification and internationalization and employability concerns—represents a major challenge: how is it possible to reconcile the heterogeneity of learner populations while taking the variability of individual learning trajectories for L2 learning and skills development into consideration?

3The language department in the Conservatoire national des arts et métiers (hereafter Cnam) has undertaken a large research project to address these issues and inform the implementation of hybrid (part in‑person, part online) Professional English courses. The present study is part of this project, and it specifically explores the potential links between individual trajectories developed over time by adult learners and the way they structure their L2 learning routes when engaged in a lifelong learning journey. The relevance of the time scale is explored here in its relation to learner variability as a prerequisite to development and therefore a source of information on the language developmental process itself and on efficient instructional designs. In particular, we report on the findings of a one-year pilot study conducted to explore the contribution of non-classroom integrated telecollaboration sessions as an opportunity for L2 learning and skills development which takes into account the variability of individual learning trajectories over time.

2. Theoretical framework

2.1. Individual differences and adult learning routes

4An extensive body of evidence exists on the important role of personal characteristics, referred to as IDs, in L2 classroom learning and achievement (e.g., Adringa & Dabrowska, 2019; Carroll, 1989; Dörnyei & Ryan, 2015). Individual traits such as personality, motivation, language aptitude, and learning strategies and styles have been explored with the goal of identifying the prominent factors that are likely to modulate L2 skill development—thus modifying and personalizing the overall trajectory of the language acquisition processes. If the literature emphasizes the fact that the learning process and IDs are intrinsically linked, the latter are also being revisited through the lens of the time factor within a complex dynamic systems perspective (Lowie & Verspoor, 2019). Dörnyei and Ryan (2015, p. 6) consider that IDs involve “complex constellations made up of different parts that interact with each other and the environment synchronically and diachronically”. This reframing of IDs implies salient temporal and situational variation. IDs, now seen as dynamic and evolving parameters, affect language development and explain the variation in overall language proficiency. This is particularly true in adult learners who generally develop dynamic repertoires of resources from their life-world experiences (The Douglas Fir Group, 2016), and hence varied learning trajectories over time (Adringa & Dabrowska, 2019).

5Adults’ lifelong learning journeys differ. As noted by Toffoli (2020), they may “engage in totally independent journeys, in entirely informal contexts and media, while others choose trajectories marked by institutional constraints, and still others pursue journeys somewhere between these two extremes” (p. 186). Learning trajectories are thus neither linear nor predictable, all the more so as they are likely to evolve over time on an individual basis and interact with other individuals’ trajectories. Recent evidence indicates that even highly similar L2 learners in terms of IDs show undeniably distinct learning trajectories over time: “Apparently, even in a homogeneous group, interactions among all relevant variables are different for different participants at different moments in time.” (Lowie & Verspoor, 2019, p. 202) This necessarily questions the notion of design and more specifically of flexible learning paths in an institutionally time-constrained learning environment.

2.2. Individual trajectories and institutional learning design

6Several researchers have recently argued in favor of more learner engagement and a rethinking of teacher practices highlighting the necessity of placing learners at the center of their own learning process within a lifelong learning perspective (Mercer & Dörnyei, 2020). The subsequent shift from instructional design to learning design this entails also implies designing meaningful opportunities for learners to model the way they learn in formal and informal settings. Learning centered on the learner’s needs is thus partly shaped by “organizing circumstances” (Spear & Mocker, 1984, p. 5) where time is accounted for as an organizing variable. In the field of adult education in particular, it seems therefore crucial to consider the process of learning as ongoing, being placed within a lifelong learning perspective, and sustained by encounters likely to evolve beyond the formal context provided. Hence the necessity to adapt pedagogical interventions accordingly. Facilitating lifelong learning thus requires providing a flexible rather than a predefined guided environment as suggested by Narcy-Combes (2018, p. 24), taking into account individual needs and characteristics, allowing adults to find their way autonomously and yet with provision of monitoring and support.

7In this respect, telecollaboration, whose contribution to the development of language and cultural skills is well documented (Lewis & O’Dowd, 2016 for instance), can be viewed as a learning circumstance (among others) that takes IDs into consideration well as the variety of adult learning pathways. Telecollaboration is sought here as the added value that allows for flexibility and learner autonomy while meeting individual specific needs in a digitally-supported environment, and which can be carried out in a time frame that is not limited to class time.

8Our particular research objective here is to explore the interaction between adult learners’ individual trajectories and their engagement in the L2 learning process in a longitudinal telecollaboration project based on “agentive action” (Thorne, 2010, p. 144) and situated at the frontier between formal and informal learning—within an environment extending beyond the classroom setting—so as to inform the design of our institutional hybrid L2 courses with a focus on oral skills and in line with a learner-centered approach. Hence, we raise the following research question:

How do adult professionals with varied trajectories structure their L2 autonomy-based telecollaboration experience?

3. Methodology

3.1. Context

9The Cnam-Cardiff Telecollaboration (CCT) project is a partnership between two HE institutions: Cnam Paris (Communication en langues étrangères Department) and Cardiff University (School of Modern Languages). This project is in fact an initiative to attend to the individual needs of adult learners in a professionalizing institutional context confronted with massification. In an early pilot needs analysis survey (N = 60), Cnam learners indicated speaking as the skill to be improved, preferably with native speakers. This study was then meant to evaluate the potential of telecollaboration as a self-directed learning opportunity that meets oral communicative needs, offers flexibility and most importantly accounts for the diversity in individual learning trajectories for a professional adult learner population.

10The CCT pilot project was two-phase and took place between February 2019 and February 2020. The implementation of the project differed from mainstream telecollaboration in hybrid courses that are generally classroom-integrated (Lewis & O’Dowd, 2016). Along with Thorne (2010), a wild semi-structured approach was favored in the sense that the exchange sessions were not integrated into the syllabus of language courses in both institutions: an introductory tutorial session was provided to the volunteers who then worked autonomously during one-on-one meetings with their peers, independently of hybrid organized classroom activity. Minimum tutor guidance was offered differently in the two phases. In the first phase (academic semester from February to mid-July 2019), the guidance included the introductory session, logbooks, emails and counselling sessions upon-request. In the second phase (from August 2019 to February 2020), tutors provided occasional informal follow‑up.

3.2. Participants

11The participant recruitment process started in the Cnam and specifically targeted learners enrolled in three English courses the researchers were in charge of. In the call, participation was presented as an extra opportunity to practice oral interaction in the target language (TL). 10 adult learners specializing in different domains volunteered (age range: 25–62; TL: English). Then, the Cardiff partner sent a call via the Languages for All center looking for 10 participants. The Cardiff volunteers were students majoring in different fields (age range: 18–30; TL: French). All the participants were either graduate or undergraduate, with independent user level, i.e. either B1 or B2 of the CEFR, in their respective TL.

12Pairing was done by the researchers and mainly based on the linguistic level in the TL to facilitate interaction. When possible, the participants’ field of specialization was accounted for to increase their intercultural professional knowledge. Some personal requests—for instance related to genre—were also considered. Once paired, the participants interacted with the same partner for the whole period of the project and were completely in charge of the content and organizational aspects of their telecollaboration practice. The number and duration of the sessions varied from one pair to the other, with an average of 90 minutes per week for most of the 10 pairs.

13A case study approach was adopted given the small size of the French participant sample (N = 10) and the qualitative nature of the collected data which is described in the next section. In addition, and as recently argued by Lowie and Verspoor (2019), studying individual cases is highly reliable for determining factors that structure language learning over time. The two French participants of this study were selected given their very distinct profiles—as identified in the collected data—in terms of general experiential backgrounds and English learning trajectories and objectives as shown in Table 1. In addition, these two learners cumulated the highest number of telecollaboration sessions and still exchange with their partners. We were hoping that these particular learner experiences would help comprehend the key features that sustain fruitful telecollaboration extended over time when L2 learners work on their own (setting their own learning goals and strategies) so as to ultimately enlighten its potential embedding in future blended L2 courses at the Cnam.

3.3. Data collection and analysis

14A three-phase protocol was designed to gather data from the French participants: introductory tutorial, telecollaboration, mid and end of project feedback. During the introductory tutorial session (February 2019), the research project was presented to the volunteers. Guidelines were given regarding the organisation of the telecollaboration in terms of content and task types, the TL allotted time, the running of the first session, and feedback on partner’s production. The participants were also given the opportunity to reflect on their telecollaboration learning objectives as well as their previous and ongoing experiences with the English language and culture via an initial questionnaire (Appendix 1). Documents to complete (logbooks and final report) were discussed in relation to autonomous practice and learning how to learn. Finally, a tandem guide was provided and participants were told they could benefit from counselling sessions upon-request.

15For each telecollaboration one-on-one session, participants completed a logbook in which they reported on the dedicated time to French and English, any potential preparation for the session and its content, as well as any corrective feedback proposed by the partner. They also reflected on what they particularly liked in the session and what they would like to work on and improve in the following meeting. Some of these learners’ exchange sessions were audio recorded: learners were asked to ideally record one at the beginning and one at the end of the telecollaboration experience. Participants were asked to send their logbooks (and recordings) right after each session. Regarding our study subjects, Learner 1 recorded 2 sessions and Learner 2 recorded one—each lasting about two hours.

16At the end of phase 1 of the CCT project (July 2019), the participants completed a feedback report (Appendix 2) probing their general satisfaction in relation to their (pre-defined and revised) goals, perceived post-experience level in English, as well as recommendations to tutors and future telecollaboration participants. These aspects were further investigated at the end of the project in February 2020 by means of semi-structured interviews, with an additional focus on organization, provided guidance, encountered difficulties and ways of overcoming them. The interviews lasted about one hour, and were conducted face-to-face for Learner 1 and by phone for Learner 2.

17Case studies are descriptive-interpretive in nature. In this paper, we sought to account for the interplay between individual learner trajectories and self-guided L2 (English) learning occurring in telecollaboration via detailed descriptions and comparison of two positive experiences. To answer the research question, the two researchers conducted an inductive thematic analysis based on Braun and Clarke (2006). First, the researchers separately analyzed the verbatim transcriptions of the interviews looking at occurrences related to factors structuring the telecollaboration as well as learners’ perceptions of potential learning benefits. The data was coded manually by each researcher who later compared and discussed their coding. The final retained categories that emerged out of this thematic analysis were later cross-checked with data reported in the logbooks and feedback reports. Illustrative participants’ verbatim comments—translated from French to English—are provided when relevant.

4. Findings

18The data enabled us to identify two main categories of time-related individual factors interacting to structure the telecollaboration experience: life trajectory and organization.

4.1. Life trajectory

19The life trajectory involves educational background as well as professional experiences and objectives. These relate to past and on-going individual experiences that model the telecollaboration, affecting participants’ initial English learning goals—and hence potential subsequent experiences—and how they go about achieving them.

20Learner 1 and Learner 2 are quite different as far as the life trajectory is concerned (see Table 1). As a high school dropout, Learner 1 has always been aware of her academic weaknesses. She explained during the semi-structured interview that she has been trying to overcome them and enhance her general knowledge and personal development through learning English mainly in informal environments:

• 1 Symbol indicating deletion of a short irrelevant part of the verbatim comments.

I am self-taught […]1 I have been improving my personal development since 2010. […] But there were no resources in French […] The frustration of not having access to what I wanted made me curious, and English was the medium I needed to gain access to other things including science […] as in many Ted Talks.

21As a declared autodidact, Learner 1 came to value not only the English language but also organization and planning, acknowledging their critical role in the learning process. To compensate for a lack of institutionalized methodology, she took a mind-mapping course that helped her become a structured learner—a distinctive personality trait of this participant that naturally influenced the way she organized her telecollaboration experience as will be highlighted below.

22In contrast, Learner 2 has a more conventional educational trajectory with completion of formal secondary and higher education. She completed a degree in communication studies, but did not have the opportunity to work in related professional domains. She thus had to adjust to the reality of the job market and started retraining in Computer Science and Information Technology in the Cnam. At the time the CCT project started, Learner 2 had just been recruited as a Software Developer in a Norwegian financial services company where English is the common corporate language for internal and external communication. Her new workplace language needs prevailed as she adopted a more professionalizing approach to telecollaboration than Learner 1: her main initial objective was in fact to improve her English communicative abilities in the IT (Information Technology) and Finance fields—in particular writing skills. Learner 1, on the other hand, was unemployed at the time of study—an informed personal decision to stay on unemployment benefits after being made redundant and enrolling in (Cnam) courses. Her telecollaboration objectives included meeting both general and broad professional English needs with priority given to developing spoken fluency.

4.2. Organization

23The second factor corresponds to self-organization and collaborative organization as shaped by the native partner (and initial guidance provided by the tutors). As the participants were completely in charge of their own tandem, this category refers to their decisions about the linguistic content as well as the methodological and practical aspects of the telecollaboration taken over time and influenced by the participants’ environment—including evolving needs and objectives.

4.2.1. Self-organization

24It is noteworthy to mention that explicit instances of organized and regulated individual approaches were found mainly in Learner 1’s data. She had in fact established a set routine for her telecollaboration sessions, which involved mental and material preparation, note taking during the exchange, immediate completion of the logbook as well as classification and storage of resource files:

I set an alarm on my phone half an hour before the Skype session […] I set my mind to the exchange and prepared my environment: I have my logbook, resources, notepad, and pen ready.

I continuously took notes during the sessions which I used to complete the logbook, and I immediately sent all documents to the tutor. I also classified all my files properly. I anticipated each subsequent session in the same way.

25Learner 1 also reported on the benefits of such an organization in terms of motivation and information retention:

To me, the session includes both the Skype interaction and the logbook. The cognitive effort is reduced if the task [of completing the logbook] is done immediately otherwise I would forget important information […] When I postponed these procedures, I quickly realized that I was leaning towards procrastination.

4.2.2. Tandem organization

26Analysis of the French participants’ interview data showed that the general structure of telecollaboration experiences depended mostly on the native partner’s personal trajectory and on changes occurring fortuitously in the environments wherein the learners and their partners lived. Thus, adult learners seem to continuously make learning choices as they encounter external organizing factors that include other individuals. Our two learners adopted a flexible and reflective approach to their telecollaboration. All along the project, they had to adapt to different circumstances which, in the case of Learner 2, affected significantly the content and format of her exchanges.

27The American partner of Learner 2 was enrolled in an exchange program at Cardiff University and preparing a degree in International Relations—a domain quite different from the finance sector wherein Learner 2 worked. Learner 2 also had to switch tools at the beginning of the telecollaboration because she encountered computer and internet connection problems, and her partner was traveling across Europe. These circumstances compelled her to revise her initial objectives focusing rather on oral fluency, and in particular pronunciation, and postpone developing writing skills:

She wanted to develop her oral fluency in French and I wanted to improve my English writing skills. But at the end, we focused more on speaking and texted sometimes. Given my internet and PC problems, and the fact that I am expected to communicate well orally at work, I am quite satisfied with what we did.

• 2 In reference to her frequent exposure to Welsh through her life partner.

I was lucky to be paired with [partner’s name]. At work, I use English with people of different origins so my accent should not be Welsh2 (laughs). She helped me a lot with my pronunciation.

28Learner 2 reported some examples of pronunciation correction provided by the partner in her logbook; for instance law/low (logbook 10, 3rd June 2019) and whole/hole (logbook 12, 17th June 2019).

29In addition, these particular events—conditioning the experience—did not allow Learner 2 to conduct the telecollaboration as she initially planned—which undoubtedly explains the lack of occurrences showing structured self-regulation in her data as compared to Learner 1. The telecollaboration sessions were not held at a fixed time slot each week and were very rarely prepared in advance:

We realized that we could not function as we initially hoped and planned. At the beginning, we tried to exchange documents beforehand. But we could not keep on because of the technical issues and I honestly didn’t have much time as I just started my new job. So, we sometimes looked for resources during the sessions.

30Compared to Learner 2, it seems that Learner 1 did not encounter constraining external factors. As she showed evidence that may justify her being categorized as a structured person, one might also argue that she could regulate her environment. Just like her self-organization, her tandem organization was quite structured. With her partner, she kept a precise telecollaboration schedule, and thoroughly planned each session according to pre-defined learning objectives. They always used the same videoconference tool (Skype), varied the resources, sent each other documents in advance, and organized their correction mode:

To me, it is just pragmatism, i.e. I like to organize myself and to set rules. We are obliged to define things together because it is important to know where we are starting from. However, I revise the rules if they become rigid or a brake on productivity and well-being.

31Learner 1 occasionally adjusted the content and format of the sessions to meet her partner’s needs changing over time:

There was one session where we only spoke French because my partner had to prepare for her oral exams.

We were always on time for our sessions. We postponed once only because she had to go to a farewell party.

My partner is only 21. Even if she can help with my CV as far as translating to English is concerned, she cannot give me methodological tips. She is not equipped at her age as she barely has work experience.

5. Discussion

5.2. Individual variation and L2 learning design

33Our longitudinal case study has highlighted the importance of individual trajectories for language development. In answering our research question, we found evidence of links that exist between on the one hand adults’ personal characteristics, their past and ongoing experiences and their specific needs, and on the other hand their attitudes toward the learning project (here telecollaboration practice) and the way they design and organize their L2 learning. The revealed structuring factors are definitely subject to change over time. Individual variation and the fact that IDs themselves fluctuate—as underlined by Lowie and Verspoor (2019, p. 188) who also stressed that not all ID factors are equally variable, and that variability depends on the time scale—which implies introducing flexibility in L2 learning design in terms of time and approach. This is particularly relevant for adult learners engaged in a lifelong learning process, given their various L2 learning trajectories and more broadly their previous experiential backgrounds. Taking into account individual trajectories and needs, and their variability therefore seems to be a necessary condition, considering the limitations of the one-size-fits-all approach. As underscored by Brudermann and coll. (2016), the challenge for course designers and instructors when dealing with large groups of heterogeneous student populations is all the more significant in the case of adult learners as the diversity of their learning pathways, needs, and expectations call for even more variety.

34The study participants grasped the opportunity that was presented to them to tailor their learning to their needs and according to the environment—both evolving continuously. In this respect, the self-directed approach adopted in our study seems to respond to the flexibility needed by adult learners, and to recent calls in the field of SLA for more learner engagement which better suits contemporary language classrooms (Mercer & Dörnyei, 2020). A learner-centered design which includes (longitudinal) autonomous telecollaboration and minimum tutor guidance can therefore account for IDs. Tutor monitoring was kept to a minimum in this classroom-independent autonomy-based telecollaboration experience. The possibility to benefit from counselling sessions upon request was offered to the participants. Interestingly, however, none of the 10 volunteers asked for such sessions. The amount of tutor guidance provided in the CCT project was perceived as appropriate by Learners 1 and 2:

Initially, I feared that the tutor would be too demanding and that there would be too much pressure. (Learner 2)

I am autonomous […] what you provided [as guidance] was enough […] and then I like to be free. (Learner 1)

5.3. Current and future learning gains

35Our adult professionals with distinct individual trajectories succeeded in their telecollaboration experiences. This study adds to the extensive body of research on the contribution of telecollaboration to L2 linguistic and cultural development though a wild autonomous approach was adopted—as opposed to more common classroom-integrated telecollaboration. It is beyond the scope of this paper to provide a detailed description of our participants’ learning gains. We will here briefly report on their perception of these benefits during this one-year project—acknowledging that real gains will have to be measured and then compared with the perceived ones in another study.

36In the feedback report, both learners indicated that their English had improved. They rated their level in the different language skills between C1 and C2 compared to between B1 and C1 as reported in their initial questionnaire at the beginning of the project. Perceived benefits related not only to linguistic skills but also to cultural aspects and their potential impact in professional practice:

It really helped me a lot for pronunciation. I learnt a lot about culture too since as a computer developer, when we develop software, we are supposed to integrate local cultural elements, for example you do not write the date similarly in the US or in France. So I paid more attention to that because she pointed it out to me. Same thing with kilometers, miles, etc. I was not used to paying attention to that before, but now I do. (Learner 2)

To me, academic success means professional success, because my objective is to find a job in France, like anyone else, without being told that I have no diploma. So everything I do is job-oriented and now, my level in English has improved so much that even if I want to work for a living like with Eurostar for example, I can. (Learner 1)

37General satisfaction and other instances of gains were also expressed. Learner 1 for instance reported on having a sense of gratification when seeing her partner applying her corrections during their conversations. The very nature of the telecollaboration evolved over time at a more personal level as both participants even met their partner in-person in Paris. It is also worth noting that Learner 1 and Learner 2 continue to exchange with their partners as of today. Lately, Learner 1 even widened the scope of the experience and included another person in the exchanges—a French friend. This certainly places us into a lifelong learning perspective that is not restricted by institutional constraints such as the content to be covered or the limited time devoted to L2 classroom learning.

6. Conclusion and directions for future research

38Our results are based on the two most successful learners in the CCT project who happen to be of an independent B level in English. This could be viewed as a limitation. Indeed, a contrastive analysis involving learners of other levels (particularly basic A levels) could yield interesting data regarding linguistic and cultural development over time, as well as the effects of the collaboration dynamics (control over speech flow, rephrasing, code-switching, body-language, etc.). In addition, analysis of the less successful telecollaboration experiences could further inform the implementation of independent exchanges by identifying the factors that hinder telecollaboration. In this project, the other eight French participants mentioned mainly lack of time and personality differences with partners as major difficulties. The impact of personality traits on telecollaboration practice in institutionalised contexts should be empirically studied as it certainly affects the type of relationship the pairs can establish over time.

39Learners 1 and 2 represent distinct profiles, and our data revealed that individual variability over time is undoubtedly a prerequisite to L2 development. Introducing flexibility in course design is necessary as it takes into account individual differences (including in terms of personality). Autonomy-based telecollaboration is a less common approach that seems to be more appealing and suitable to adult professionals—though a larger scale study is required to confirm. Unlike results in mainstream literature focusing on university students, our research has shown that telecollaboration can certainly happen independently without being integrated in classroom set‑ups (with varying degrees of success)—and as such can be seen as a response to the challenge course designers face when dealing with large groups of heterogeneous populations. Telecollaboration with minimum tutor guidance has proved to be the added value that offers time and content flexibility in an institutional setting dedicated to lifelong L2 learning adults who have different personal trajectories, specific needs, and professional constraints. This self-guided learning approach places individual variation and learner-centered design at the core of a dynamic complex learning system evolving over time.

40The present research certainly informs the future implementation of autonomy-based telecollaboration in blended EFL courses at the Cnam. However, following up on this experience, we believe that it would be worth giving a stronger professionalizing dimension to pairing, within a community of practice—which is further in line with the avenue for research designed by Thorne (2010) in terms of ‘agentive action’ as learners are engaged in specialized online interest communities, assuming complete responsibility for their linguistic and intercultural (professional) development.