1.1A research gap

The lack of attention for diverse learner groups has been identified as a gap in LLS research (Cohen et al., 2023Cohen, Andrew D., Yongqi Gu, P., Nyikos, M., Plonsky, L., Harris, V., Gunning, P., Kai-Hui Wang, I., Pawlak, M., Gavriilidou, Z., Mitits, L., Sykes, J. M., & Gao, X. (2023) Tangible insights on the strategizing of language learners and users. Language Teaching, 1–20. ; Harris, 2019Harris, V. (2019) Diversity and integration in language learning strategy instruction. In A. U. Chamot & V. Harris (Eds.), Learning strategy instruction in the language classroom. (pp. 38–52). Multilingual Matters. ; Jones, 2016Jones, A. H. (2016) The discourse of language learning strategies: towards an inclusive approach. International Journal of Inclusive Education, 20 (8), 855–870. ). This qualitative study focuses on the language learning experiences of adult refugees with limited formal education in their home countries, learning a new language in a highly literate society, and learning to read and write — often for the first time — in the new language. I refer to this learner group as ‘LESLLA learners’. LESLLA is an acronym for ‘Literacy Education and Second Language Learning for Adults’. LESLLA learners are a diverse group in terms of countries of origin, first languages, and reasons for migration. They may or may not have some literacy in the home language and they have at the most primary schooling in their native language ( www.leslla.org, 2024 www​.leslla​.org (2024) www​.leslla​.org). Data suggest that the definition applies to a considerable percentage of migrants from refugee backgrounds in Europe. For example, Swedish data from 2018 reported 18% learners with less than six years of schooling in their countries of origin (Norlund Shaswar & Wedin, 2019Norlund Shaswar, A., & Wedin, A. (2019) Language learning strategies and teaching practices in adult L2 education: The case of Swedish for immigrants. Apples Journal of Applied Language Studies, 13 (3), 17–34. ). A study from Belgium found that 16% of asylum-seeking migrants had less than six years of education in their home countries and 27% struggled with most basic home language literacy tasks (Hooft et al., 2020Hooft, H., Vandermeerschen, H., Chakkar, S., Vandommele, G., De Cuyper, P., & Schiepers, M. (2020) Laaggeletterheid bij volwassen nieuwkomers in de Belgische asielopvang. https://​www​.arts​.kuleuven​.be​/cto​/onderzoek​/onderzoek​-geletterdheid​/laaggeletterdheid​-bij​-nieuwkomers​-in​-de​-opvang​/fedasil​-laaggeletterdheid​-bij​-nieuwkomers​-isbn​.pdf).

In migration contexts, the stakes for language learning are high. Many European countries impose language requirements on new citizens as part of their civic integration policy (Rocca, Carlsen, & Deygers, 2020Rocca, L., Carlsen, C. H., & Deygers, B. (2020) Linguistic integration of adult migrants: requirements and learning opportunities. https://​rm​.coe​.int​/linguistic​-integration​-of​-adult​-migrants​-requirements​-and​-learning​-opp​/16809b93cb), with legal consequences associated with language learning (De Waal, 2017De Waal, T. (2017) UvA-DARE (Digital Academic Repository) Conditional belonging. A legal-philosophical inquiry into integration requirements for immigrants in Europe. https://​dare​.uva​.nl). From the learners’ perspective, learning the host country’s language is seen as essential for making friends, finding employment, and for social and psychological well-being (Benseman, 2014Benseman, J. (2014) Adult refugee learners with limited literacy: Needs and effective responses. Refuge, 30 (1), 93–103. ). But for LESLLA learners, learning a new language can be a bumpy road and both learners and educators often struggle (Deygers & Vanbuel, 2022Deygers, B., & Vanbuel, M. (2022) Gauging the impact of literacy and educational background on receptive vocabulary scores. Language Testing, 39 (2), 191–211. ; Housen, 2022Housen, M. (2022) Stagneren in het taalleerproces. Les, 40 (221). https://​www​.tijdschriftles​.nl​/inhoud​/tijdschrift​_artikel​/LE​-40​-221​-11​/Stagneren​-in​-het​-taalleerproces; Strube, 2014Strube, S. (2014) Worstelen met de mondelinge vaardigheden. Les, 32 (192). https://​www​.tijdschriftles​.nl​/inhoud​/tijdschrift​_artikel​/LE​-32​-192​-10​/Worstelen​-met​-de​-mondelinge​-vaardigheden, 2015Strube, S. (2015) Het is makkelijker gezegd dan gedaan. Kopzorgen in de alfaklas. Les, 33 (online only). https://​www​.tijdschriftles​.nl​/inhoud​/tijdschrift​_artikel​/LE​-33​-0​-9​/Het​-is​-makkelijker​-gezegd​-dan​-gedaan​-Deel​-1).

LLS research has focused on effective language learning and, therefore, LESLLA practice might take advantage from this research field. However, in the extensive research base on LLS, participants with LESLLA profiles have rarely been included. From a practitioner’s perspective, this is problematic, since our knowledge of how to support this learner group is still limited. From a researcher’s perspective it is also problematic, since focus on narrow learner populations might cause generalizability issues: we cannot assume that what is true for specific groups of research participants also applies to other groups (Andringa & Godfroid, 2020Andringa, S., & Godfroid, A. (2020) Sampling bias and the problem of generalizability in applied linguistics. Annual Review of Applied Linguistics, 40 , 134–142. ; Bigelow & Tarone, 2004Bigelow, M., & Tarone, E. (2004) The role of literacy level in second language acquisition: Doesn’t who we study determine what we know? TESOL Quarterly, 38 (4), 689–700. ; Ortega, 2005Ortega, L. (2005) For what and for whom is our research? The ethical as transformative lens in instructed SLA. Modern Language Journal, 89 (3), 427–443. ).

1.2Strategy research in the LESLLA field

Jones (2016)Jones, A. H. (2016) The discourse of language learning strategies: towards an inclusive approach. International Journal of Inclusive Education, 20 (8), 855–870. , based on social learning theories (Lave & Wenger, 1991Lave, J., & Wenger, E. (1991) Situated Learning. Legitimate Peripheral Participation. Cambridge University Press. ), built the argument that repertoires of practice of underprivileged learner groups predominantly draw on social strategies. Empirical research into strategy use of LESLLA learners is scarce (Dalderop, Andringa, & Rispens, 2023Dalderop, K., Andringa, S., & Rispens, J. (2023) Strategy focused instruction in literacy education and second language learning for adults (LESLLA) in the Netherlands. Dutch Journal of Applied Linguistics, 12 . ; Jones, 2016Jones, A. H. (2016) The discourse of language learning strategies: towards an inclusive approach. International Journal of Inclusive Education, 20 (8), 855–870. ). In one of the few published studies, Reimer (2008)Reimer, J. (2008) Learning Strategies and Low-Literacy Hmong Adult Students. https://​conservancy​.umn​.edu​/items​/26bcb760​-dfb1​-4b53​-ae45​-d790fc14d0c9 conducted a qualitative study in which LESLLA learners (N = 11) were trained in a variety of strategies. The intervention was evaluated with the learners but the results were inconclusive, as learners indicated to find a strategy useful but not to use it and the other way around. A study with a quasi-experimental design investigated reading strategy instruction in a group of LESLLA learners (N = 36) and found that learners who had received training outperformed a control group (Huang & Newbern, 2012Huang, J., & Newbern, C. (2012) The effects of meta-cognitive reading strategy instruction on reading performance of adult ESL learners with limited English and literacy skills. Journal of Research and Practice for Adult Literacy, Secondary and Basic Education, 1 (2), 66–78.). Focusing on strategy use, King and Bigelow (2012)King, K. A., & Bigelow, M. (2012) Acquiring English while learning to do school: Resistance and accommodation. In P. Vinogradov & M. Bigelow (Eds.), Low Educated Second Language and Literacy Acquisition (pp. 157–182). University of Minnesota. https://​drive​.google​.com​/file​/d​/1XXGXuEvI9VAmMojouSm5JfHZRNTekN8j​/view described the learning behavior of two young adults learning ‘to do school’. The observed strategies were believed to be coping strategies for doing school rather than ‘real’ LLS that actually enhance learning. Naif and Saad (2017)Naif, A. H., & Saad, N. S. M. (2017) Language learning strategies use and challenges faced by adult arab learners of Finnish as a second language in Finland. English Language Teaching, 10 (4), 111. administered the Strategy Inventory for Language Learning (SILL) (Oxford, 1989Oxford, R. L. (1989) Strategy Inventory for Language Learning Version 7.0.) in a group of adult literacy learners in Finland (N = 30). To facilitate understanding for the participants with emerging literacy, the written statements in the SILL were orally translated. Findings suggested that participants most often used metacognitive strategies, like having clear goals, and social strategies, like asking for help from conversation partners. An ethnographic case study from Sweden (Norlund Shaswar & Wedin, 2019Norlund Shaswar, A., & Wedin, A. (2019) Language learning strategies and teaching practices in adult L2 education: The case of Swedish for immigrants. Apples Journal of Applied Language Studies, 13 (3), 17–34. ) described a critical incident in which a Syrian learner of Swedish indicated to prefer the use of a social strategy while the Swedish teacher insisted on the use of a cognitive strategy. The study pointed to the power relations between the teacher and the learner. In sum, the research base on LLS including LESLLA learners is small and diverse.

1.3Civic integration policy in the Netherlands

The current study is situated in the Netherlands, where for many people new to the country language learning was made mandatory by a civic integration law named Wet inburgering 2013 (Wi2013). When the law was in effect, from 2013-2022, in a final exam learners had to demonstrate language competence in oral and written production and reception at the A2-level of the Common European Framework of Reference (Council of Europe, 2001Council of Europe (2001) Common European Framework of Reference for Languages: Learning, Teaching, Assessment.). Newly arrived refugees were made responsible for organizing their own integration and were given a loan to pay for the course’s costs. If the exams were passed within the given time frame of three years, the loan was converted into a gift. After 600 hours of instruction and three unsuccessful attempts to pass the exams, one could be exempted from civic integration duty (Groenendijk et al., 2021Groenendijk, C. A., de Hart, B., & van Oers, R. (2021) Lessen van 25 jaar inburgering in Nederland. Nederlands Juristenblad, 33 .). In the Netherlands, civic integration provision is a free market activity. The budgets provided to refugees constituted interesting economic opportunities for commercial language schools, which did not always put the learners’ interests first. Evaluation of the Wi2013 brought unintended and illegal practices to light (Algemene Rekenkamer, 2017Algemene Rekenkamer (2017) Inburgering. Eerste resultaten van de Wet inburgering 2013. Algemene Rekenkamer.; Significant, 2018Significant (2018) Inburgering: systeemwereld versus leefwereld: Evaluatie Wet inburgering 2013.) and for people new to the country it was hardly possible to assess the quality of the schools before taking a course. In January 2022, a new law entered into force, named Wet inburgering 2021. However, newcomers must follow a civic integration trajectory according to law that was in force when refugee status was received. Therefore at the moment of data collection for this study many migrants were still affected by the Wi2013. It is in this context of mandatory language learning, with the threat of legal or financial consequences and with language provision of mixed quality, that the experiences of language learners who participated in this study were embedded and must be understood.

1.4Research questions

This study aims at contributing towards research on the nexus of LLS and LESLLA, by tapping into LESLLA learners’ experiences, knowledge, and reflections about their language learning and their strategy use. The study aims to answer the following research questions:

The study adopts a Constructivist Grounded Theory (CGT) approach, a data-driven methodology for qualitative research, which will be further explained in Section 2.1. Data-driven approaches typically do not start from a theoretical framework. They start from the data and aim to arrive at a model or a theory (Birks & Mills, 2015Birks, M., & Mills, J. (2015) Grounded Theory (4th ed.). Sage.; Charmaz, 2014Charmaz, K. (2014) Constructing grounded theory. Sage.; Hadley, 2017Hadley, G. (2017) Grounded Theory in SLA (1st ed.). Routledge.). Therefore, this introduction is light on theory. Following CGT approaches, where deemed useful, theory will be discussed and connected to the data in the results section.

2.1Constructivist grounded theory

This study adopted a Constructivist Grounded Theory (CGT) approach (Charmaz, 2014Charmaz, K. (2014) Constructing grounded theory. Sage.). Grounded Theory (GT) is a data-driven methodology for qualitative research which has been widely used across disciplines, although not very often in the field of second language acquisition (Hadley, 2017Hadley, G. (2017) Grounded Theory in SLA (1st ed.). Routledge.). Rather than departing from a theoretical framework, a GT study aims to arrive at a model or a theory that is grounded in the data. The sociologists Strauss and Glaser are recognized as the founders of GT (Birks & Mills, 2015Birks, M., & Mills, J. (2015) Grounded Theory (4th ed.). Sage.). CGT (Charmaz, 2014Charmaz, K. (2014) Constructing grounded theory. Sage.) is a later version of the methodology. A main difference with original GT is that it departs from a different epistemology. Where Strauss and Glaser adopted a positivist view, assuming that factual knowledge was situated in the data and needed to be uncovered, CGT builds on a subjective epistemology, acknowledging the role of the researchers, the participants, and the setting, in data collection as well as analysis. CGT assumes that knowledge is socially constructed rather than discovered (Charmaz, 2014Charmaz, K. (2014) Constructing grounded theory. Sage.). A second important difference concerns the role of prior knowledge. Where original GT claimed that theoretical reasoning should be based on the data alone, suspending one’s background knowledge, CGT promotes to include background knowledge to arrive at theoretical explanations (Hadley, 2017Hadley, G. (2017) Grounded Theory in SLA (1st ed.). Routledge.). The data-driven nature of GT fit with the study’s goal of increasing our understanding of language learning experiences and strategy use in a context that has hardly been researched, by listening to LESLLA learners without too many preconceived notions. The constructivist epistemology aligns with my philosophical stance in research, based on the belief that reality is socially constructed and that knowledge is co-created by both the participants and the researcher. Consequently, CGT was chosen over the traditional GT approach.

The study is based on interview data. Interviews enable gaining understanding of the learners’ perspectives on language learning and strategy use. CGT prescribes that data are collected and analyzed simultaneously, while new data are constantly compared with previous data. Data are analyzed in different rounds and on different levels of coding, starting with initial coding, based on intensive line-by-line analysis, followed by focused coding, which identifies core concepts in the data. Focused codes are the basis of the core categories that build the model or theory. A CGT study aims at understanding experiences of groups of people. Participants are selected based on the emerging analyses, thus filling in gaps that need further exploration, a procedure known as theoretical sampling. Another key element of CGT sampling is saturation. This refers to the point that after new rounds of data collection no new theoretical insights come up in the analysis. Thus, in theory, there is an open end to the amount of data that need to be collected for a study.

2.2Researcher positionality

A CGT study is thought of as a co-construction of meaning, negotiated between an interviewee and a researcher, and in our case an interpreter. Since the researcher’s interest and position is a major influence on both data construction and analysis (Birks & Mills, 2015Birks, M., & Mills, J. (2015) Grounded Theory (4th ed.). Sage.), this section will describe my positionality. Positionality refers to the position that the researcher has chosen to adopt within a given research study (Holmes, 2020Holmes, A. G. D. (2020) Researcher positionality- A consideration of its influence and place in qualitative research- A new researcher guide. International Journal of Educational Research, 8 (4), 1–10. ). Throughout my professional life, I have served the field of Dutch as a second language in a variety of roles: as a volunteer language tutor, subsequently as a teacher, assessment developer, course ware developer, teacher trainer, and advisor. In these roles I was most interested in strengthening learner autonomy, aiming to equip learners with tools that would enhance lifelong language learning. Many projects I was involved in have focused on LESLLA learners, strongly motivated by the notion of education and literacy as a human right. The study is part of my PhD research, focusing on LLS in LESLLA practice. Contributing to improving LESLLA practice is my main drive; I believe that in order to be able to do so, is pivotal to improve our understanding of LESLLA learners’ learning experiences. My background in practice, my role as a researcher, and my aim to contribute to LESLLA practice were explained to participants when inviting them to participate.

2.3Participants

The participants for this study were adult refugees (N = 30), twenty female and ten male, who fled from Syria (n = 16) and Eritrea (n = 14). The Syrian and Eritrean groups were large refugee groups in the Netherlands at the time of study (Vluchtelingenwerk Nederland, 2021Vluchtelingenwerk Nederland (2021) Vluchtelingen in getallen 2021. Vluchtelingenwerk Nederland. https://​www​.vluchtelingenwerk​.nl​/sites​/default​/files​/2021​-12​/Vluchtelingen​/in​/getallen​/2021​_publicatie​.pdf) and well represented in LESLLA classes. Amongst the participants from Syria, nine identified as Syrian, five as Kurdish and one as Palestinian. For the Kurdish participants Arabic was their second language, in which they were fluent according to the interpreter. All Eritrean participants identified as Eritrean and Tigrinya was their first language. Ages ranged from 19 to 65 years. The participants had a maximum of seven years of education in their home countries (overall mean 2.7; Syrian group 4.25; Eritrean group 0.75). Fourteen participants never attended school and 16 participants attended school for three to seven years. I purposefully selected participants with six years of education or less but included one participant who attended school for seven years. Fifteen participants said they could not read in their home language and fifteen said they could, at least to a certain extent. Among the non-readers, 10 were from Eritrea and five from Syria. Among the readers, four were from Eritrea and 11 from Syria.

Participants’ length of stay in the Netherlands varied from one to seven years. Thirteen participants were still enrolled in civic integration programs and one was on maternity leave. Fifteen were exempted from civic integration duty after having followed 600 hours of language instruction. One participant had passed the civic integration exam and was enrolled in vocational training. Participants were invited through the interpreters’ as well as the researcher’s networks. Often educators and key persons from Syrian and Eritrean communities served as mediators. In order to communicate our interview request with potential participants, the researcher and the interpreters recorded short videos in Arabic and in Tigrinya, briefly explaining the study’s goals and inviting participation to the study. Participants were found in various places in the Netherlands, including cities, smaller towns and villages. They took their mandatory civic integration classes in a variety of language schools.

Ethics approval for the study was obtained from the University of Amsterdam ethics committee. Researching LESLLA learners comes with a number of concerns regarding research ethics (Fox et al., 2020Fox, A., Baker, S., Charitonos, K., Jack, V., & Moser-Mercer, B. (2020) Ethics-in-practice in fragile contexts: Research in education for displaced persons, refugees and asylum seekers. British Educational Research Journal, 46 (4), 829–847. ; Michaud et al., 2022Michaud, A. H., Fortier, V., & Amireault, V. (2022) “Do I have to sign my real name?” Ethical and methodological challenges in multilingual research with adult SLIFE learning French as a second language. Languages, 7 (126), 1–16. ). Standard ethics procedure are often inappropriate as they heavily rely on literacy and discuss abstract topics such as risks and insurance. In consultation with the ethics committee an adapted procedure was applied, including oral explanation of the study’s goal, explanation of participants’ rights to not answer questions or stop the interview, explanation of the fact that their information would be kept anonymous and that only the researcher would have access to the recording. A four-eye principle was introduced by having the interpreters sign a form confirming that they had explained the elements mentioned above to the participant, that the participant was given the space to ask questions, and that the participant had confirmed that the procedure was clear and had consented to participate.

2.4Interpreters

All interviews were conducted with the help of an interpreter. Following the recommendations of Kosny et al., (2014)Kosny, A., MacEachen, E., Lifshen, M., & Smith, P. (2014) Another person in the room: Using interpreters during interviews with immigrant workers. Qualitative Health Research, 24 (6), 837–845. , the interpreters served on the project for the entire duration of the study, in intensive collaboration with the researcher. The Arabic speaking interpreter was a student from Syria who assisted in the project as a research apprentice. The Eritrean interpreter was a registered professional cultural mediator with extensive experience in interpreting. In most interviews the conversation was fully dependent on the support of the interpreter. Three participants chose to share their experiences in Dutch and some Syrian participants mixed some Dutch into their Arabic. The conversations with the Eritrean speaking participants all heavily relied on interpretation.

2.5Interviews

The dataset is comprised of thirty interviews, varying in length from about 30 to 60 minutes. The interviews were conducted between June 2022 and March 2023 and each participant was interviewed once. A semi-structured interview guideline was constructed (Gerson & Damaske, 2021Gerson, K., & Damaske, S. (2021) The science and art of interviewing. In The Science and Art of Interviewing. Oxford University Press. ), tapping into language learning experiences and strategy use in and out of school. The strategy focused interview questions addressed learning goals, resources used for learning, cognitive strategies, social environment and social strategies, as well as affect and affective strategies. The interview guideline is available as Appendix A. The guideline was used flexibly, dedicating more time to topics that seemed of special interest to the participant, or to those topics that had remained underdeveloped in previous interviews, in line with the guidelines for conducting CGT studies (Charmaz, 2014Charmaz, K. (2014) Constructing grounded theory. Sage.). Participants could choose the interview location. Participants were interviewed in their houses, community centers, and language schools.

In line with the concept of theoretical sampling, I only planned a few interviews ahead, continuously reflecting on whose voices could add most to the understandings gained so far. For example, the first few participants discussed disappointing experiences with their language classes. This led me to look for new participants who had joined classes at schools I believed to be trustworthy. In doing so, I acquired a richer, more diverse understanding of the perceptions of learners regarding the language instruction they had received.

2.6Data analyses and coding

All interviews were audio recorded and I transcribed all Dutch translations and Dutch participant contributions verbatim, using MS Word. Subsequently, I analyzed the data according to Charmaz’s (2014)Charmaz, K. (2014) Constructing grounded theory. Sage. guidelines, described in Section 2.1. Starting right after the first interview, I constructed, initial codes through a detailed process of line-by-line analysis of the data. Following Charmaz’ recommendations, I phrased codes as actions. Table 1 shows an example of initial coding.

After having conducted and coded seven interviews, based on patterns in the data focused codes for both helping and hindering conditions were developed and these were refined after subsequent interviews. Focused codes are listed in Table 2.

Based on the emerging analyses, after about 25 interviews a model was drafted and subsequently further refined. This will be further explained in the results section.

3.1Main helping and hindering conditions

Table 2 presented focused codes for conditions helping and hindering learning. During the coding process, subcodes were developed for many of the focused codes. Table 3 displays codes and subcodes for the most important helping and hindering conditions, ‘most important’ defined quantitatively as occurring in ten or more interviews. For example, for the metacognitive strategy ‘selecting resources’, subcodes included: ‘watching TV’, ‘using digital tools’, ‘visiting nonformal learning activities’, and ‘using textbooks’, among others. Only the strategies ‘watching TV’ and ‘using digital tools’ were used by ten or more interviewees, thus earning a place in Table 3. The last column of Table 3 displays the number of participants who addressed the specific topic. This information must be interpreted with caution: a CGT study aims at understanding experiences of a group, and principles of theoretical sampling recommend to explore in subsequent interviews especially those topics that so far had remained less well understood. This means that not all interviews may have touched on the exact same topics. Nevertheless, topics that reoccurred many times must be seen as relevant.

3.2The learners’ voices

The previous section outlined the main helping and hindering conditions affecting LESLLA learners’ learning experiences. This section aims to deepen the picture by including the participants’ voices. It will illustrate the helping conditions ‘motivation’ and ‘hopes and aspirations’, followed by LLS, before ending with the conditions that hindered learning. Some remarks must be made about the excerpts presented in this section. First, for the sake of this article excerpts were translated into English. Excerpts were translated by the researcher, adopting a process of translation and back-translation, assisted by translation software and dictionaries. Secondly, both interpreters occasionally switched between first and third person, which explains why excerpts may vary in style. Further, numbers were used to refer to participants to protect their identities. Needless to say that I see participants as individuals with names, biographies and stories to tell.

Motivation, hopes and aspiration

The participants’ narratives clearly showed their motivation for language learning. Most participants said that they had been happy to go to school and learn the Dutch language, and many stressed how hard they had worked.

[I have] not been to school in my country and when I heard I would go to language school I was very happy, I was excited, and I worked hard. I can now write my name and I also can read a little and I am happy about that [P30, female, Eritrea].

You need to take charge and work hard [P08, female, Syria].

A majority of participants experienced support from their teacher. Participant 04 [female, Syria] said:

[-] The teacher was very funny and has given much support and the teacher showed that she wanted us to learn the language well. So that helped me with learning Dutch.

Many participants reported that valuable topics had been addressed in class, such as communicating with their doctor and traveling. A participant who learned to read the numbers explained:

If I watch the numbers on the metro, I can know where it goes to [P03, female, Syria].

However, a supportive teacher and useful content did not always lead to perceived language learning.

I did not understand. She was completely committed to me, but I did not understand a thing. The problem was me, I think [P31, female, Eritrea].

She did her best, but I could not understand my teacher, it was hard for me [P01, female, Syria].

Without exception, all participants spoke about their hopes and wishes related to improved language proficiency. Most importantly, they wished to be able to independently organize their lives.

I get very frustrated because of not knowing Dutch, since there are so many things I cannot do. For example, I cannot travel since I would get lost [-]. I cannot travel on my own, I need support. And that does not make me happy [P21, female, Eritrea].

Participants also wished to be able to express themselves and make new friends. Participant 11 [male, Syria], articulated it as follows:

I want to meet more Dutch people, I just want to go to a place where Dutch people are, I want to order something, I want to tell what I think, I just want to be able to talk.

Many participants aspired to work or study:

I would like to work in a home for the elderly or in childcare. I would like to help people [P04, female, Syria].

Participants realized that working means: meeting people, which could enhance language learning:

I have worked in construction and I would be really confident to do that work here. But I don’t know the language. But if you do not work, you cannot learn the language [P20, male, Eritrea].

The learners’ narratives thus reflect their motivation for language learning, while being motivated did not necessarily go hand in hand with perceived success in learning.

Language learning strategies

All participants reported on the use of at least some LLS. The metacognitive strategy “obtaining resources” (Oxford, 2017Oxford, R. L. (2017) Teaching and researching language learning strategies. Routledge., p.155) was mentioned often. Mostly, participants reported on watching Dutch programs on TV and using digital tools for learning, Like Participant 01 [female, Syria], who said:

I watched TV and I watched programs meant for children, I did everything to improve my language.

The most frequently mentioned cognitive strategy was rehearsing (Oxford, 2017Oxford, R. L. (2017) Teaching and researching language learning strategies. Routledge., p. 182). Some learners also used tools to help them rehearse, like flashcards, vocabulary notebooks or sticky notes. Note how rehearsing relies on literacy:

If I learn a new word, I will write it down in a notebook, so I can rehearse. And [-] I write them on sticky notes and I stick them in my room so I can read them all the time [P08, female, Syria].

Social strategies, the strategies that draw on the sociocultural context to facilitate communication as well as learning (Oxford, 2017Oxford, R. L. (2017) Teaching and researching language learning strategies. Routledge., p. 200), were used frequently, for learning as well as to overcome communicative obstacles. The participants’ children most often supported communication with offices. But participants relied on their help unwillingly and expressed a strong wish to be independent:

I want to be able to do things without depending on other people, I am sick of being dependent, I am sick of it [P06, male, Syria].

For learning too, the help from family members was not always appreciated. Children were pictured as absent, impatient, or not knowledgeable. The social strategy most appreciated and most effective was practicing Dutch with Dutch-speaking people. This required a new social network, and participants who had been able to build such networks strongly benefitted from it. They were remarkably more happy about their learning and expressed confidence in further learning:

Yes I certainly have people, my neighbors, [-], they are not neighbors, they are family to me. Not that they teach me the language, to read and write, but we talk together, we meet, we drink coffee together and then we also talk. I use everything, hands, feet, and yes, it is getting better [P29, female, Eritrea].

If I hear a word, on the radio, I ask my colleague, what does this word mean? Or, my boss, what does this word mean? I want to learn! [P14, male, Syria].

The excerpts illustrate how learners use a variety of strategies. Social strategies are particularly effective and highly contributing to language learning, for those learners who can rely on a new social network of Dutch-speaking people.

Hindrances for learning

Participants experienced learning Dutch as a tough job and the Dutch language was perceived as difficult. Many also attributed their difficulties to themselves as learners and identified ‘forgetting’ as their main problem. Participants believed forgetting to result from age, from limited educational experience, or from both. Having no L1 literacy made learners struggle with rehearsing, as they could not rely on notes taken in class and thus depended on their memory.

But yes, since I cannot read and write, I could not remember much [P31, female, Eritrea].

Another important hindrance for learning was stress. Stress can easily be interpreted as arising from participants’ background as a refugee. Indeed, some participants reported worries about family members in their home countries. However, stress mainly seemed to arise from feeling unable to organize life and thus protect themselves and their families in their new society, and from feeling dependent on others as a result (Also see Hollifeld et al., 2018Hollifeld, M., Warner, T. B., Krakow, B., & Westermeyer, J. (2018) Mental health effects of stress over the life span of refugees. Journal of Clinical Medicine, 7 (25), 1–11. ).

We have so many problems in our home, due to language deficiency, and I can’t solve it myself, I depend on others, and I always have to wait for my daughter, she is 21 years old, but she works and she goes to school, she is very busy. And then I wait until she has time to call or until she has time to go, so yes, I depend on her, and I hate it [P23, female, Eritrea].

Although practicing Dutch within a new social network was felt to be important, not all participants had been able to build such networks and many participants felt extremely isolated. Especially those living in cities talked about having no one to talk to, to help them learn the language or to help them out with practical problems.

Three times I asked my neighbor, come drink coffee together, come, but [she said]: another time [P08, female, Syria].

I live here since two years in my house, and never one of the neighbors came by to ask, hi, how are you, I am your neighbor. Dutch people don’t do that, unfortunately [P05, male, Syria].

Isolation was directly connected to experiencing stress:

[If I could improve my Dutch] I wouldn’t have any mental problems. Now I do, I have many mental problems. I would like to go to the doctor, I would like to tell everything, be able to tell the doctor, but I can’t do that, I can’t put my thoughts into words. I like to get to know other people, talk to other people. My neighbor, he doesn’t know me at all, he really doesn’t know whether or not I’m a good man, because we can’t communicate. From the first moment I’ve been here I’ve had psychological problems, because of the language [P06, male, Syria].

The excerpts thus show how language learning is challenging for LESLLA learners, with forgetting, stress, and isolation hindering learning.

3.3Motivation

In a Grounded Theory study, topics may emerge that were not anticipated in advance. The analysis identified LLS as well as motivation as helping conditions, enhancing learning. LLS have been introduced earlier but since participants discussed motivation extensively, the topic deserves attention too. This section briefly introduces a theoretical perspective on motivation.

SLA research has understood motivation in various ways, most often related to outcome expectations. According to Dörnyei and Ushioda (2021)Dörnyei, Z., & Ushioda, E. (2021) Teaching and Researching Motivation (3rd ed.). Routledge/Taylor & Francis. motivation guides the direction and magnitude of human behavior and accounts for choice of action, persistence, and effort. The most influential theory in L2 motivation is Dörnyei’s L2 Motivational Self System (L2MSS) (Dörnyei, 2009Dörnyei, Z. (2009) The L2 Motivational Self System. In Z. Dörnyei & E. Ushioda (Eds.), Motivation, Language Identity and the L2 Self (pp. 9–42). Multilingual Matters. ; Dörnyei & Ushioda, 2021Dörnyei, Z., & Ushioda, E. (2021) Teaching and Researching Motivation (3rd ed.). Routledge/Taylor & Francis. ). The L2MSS includes three components: (1) The Ideal L2 self, which drives the desire to reduce the gap between an individual’s actual and ideal self; (2) The Ought-to L2 self, which concerns motivation to meet expectations of others or to avoid negative consequences; (3) The L2 learning experience, which concerns motives related to the learning environment and learning experience. At first glance, the L2MSS seems a good fit for our data. In this study’s analysis, the Ideal L2 self is reflected in the category ‘Hopes and aspirations’. Dörnyei’s L2 learning experience is reflected as ‘Initial Motivation’ and ‘Positive School Experiences’, and it can be argued that the participants’ wish to be able to organize their lives can be driven by the Ought-to self as much as by an Ideal self image. What I did not see, however, was a relationship between motivation and learning outcomes: many participants who were motivated for learning perceived their learning as failed. Hence, motivation did no seem to predict learning outcomes. This lead me to explore an alternative theoretical perspective.

Where the L2MSS has adopted a cognitive perspective on motivation, others have studied motivation from a social perspective. Norton (2013)Norton, B. (2013) Identity and Language Learning (2nd ed.). Multilingual Matters. and Darvin and Norton (2015Darvin, R., & Norton, B. (2015) Identity and a model of investment in applied linguistics. Annual Review of Applied Linguistics, 35 , 36–56. , 2021Darvin, Ron, & Norton, B. (2021) Investment and motivation in language learning: What’s the difference? Language Teaching. ) introduced a theory of investment in language learning. Investment is related but not synonymous to motivation: “While motivation is a psychological construct that focuses on conscious and unconscious factors, investment is primarily sociological and focuses on how histories, lived experiences and social practices shape language learning” (Darvin & Norton, 2021Darvin, Ron, & Norton, B. (2021) Investment and motivation in language learning: What’s the difference? Language Teaching. , p.1). Investment theory explains how people can be motivated for language learning while not or no longer being invested in it.

Investment theory has pointed to power relationships that may negatively impact learning and has drawn attention to the fact that people are not always able to choose the conditions of interaction. “While learners can speak from multiple conditions, and perform diverse identities, they can also be positioned by these powerful others in ways that can limit their opportunities to speak and be heard” (Darvin & Norton, 2021Darvin, Ron, & Norton, B. (2021) Investment and motivation in language learning: What’s the difference? Language Teaching. , p.2). For participants in this study, investment in learning was reduced by policy, which does not allow refugees to take on jobs while they are in an integration trajectory or to continue joining classes after being exempted from civic integration duty. Isolation seemed to be stronger in big cities compared to smaller places, while in the Netherlands refugees cannot choose where to live; they are referred to a municipality that becomes responsible for their housing. Investment was also reduced by neighbors who did not reach out, or declined invitations, thus limiting the learners’ opportunities to build networks and engage in use of the Dutch language. Also teachers could affect learners’ investment, as is evident from the words of participant 15 [male, Syria], who explained how he tried to ask a classmate for clarification, […] but then I heard that it is not permitted to use my own language to ask questions. That made me a bit demotivated.

3.4Self-efficacy

Participants’ narratives revealed a variety of perceptions regarding perceived language learning success and confidence in further learning. All participants stressed the importance of improving their Dutch in relationship to their aims and ambitions. Some expressed contentment about their achievements and were confident that they could further improve their proficiency in Dutch, like participant 29 cited earlier, who felt that communication with her new chosen family was getting better, or as speaks from this excerpt:

I did not know a thing, I could do nothing, but now I can travel, and I can do so many things without help. So I notice a big difference between how it was and how it is now [P33, male, Eritrea].

Others, however, qualified their language learning as ‘failed’. The dataset revealed 11 participants who literally said ‘I did not succeed’. Participant 07 said that the 600 hours of classes he had attended had been a waste of time ‘since I learned nothing’ and participant 04 believed there to be ‘no hope for me anymore’. Based on these success notions from the learners’ perspectives, ‘self-efficacy’ was chosen as a label covering the elements ‘perceived success’ and ‘confidence in further learning’. Self-efficacy refers to learners’ believe in their ability to succeed in learning (Bandura, 1997Bandura, A. (1997) Self-efficacy. The exercise of control. W.H. Freeman and Company.). This construct also plays a crucial role in LLS theory: Oxford (2017)Oxford, R. L. (2017) Teaching and researching language learning strategies. Routledge. has pointed to self-efficacy as an element that constitutes what she has called ‘the soul of strategies’. Thus, our analysis identified conditions positively or negatively impacting self-efficacy.

3.5Summarizing the findings

This study identified conditions positively or negatively impacting learning. The hindering conditions included forgetting, stress, and isolation. The conditions coded as helping included motivation and LLS. Helping conditions increase and hindering conditions decrease learners’ self-efficacy. However, it would too simple an understanding of the data to assume that helping conditions automatically lead to language learning and hindering conditions impede learning, with equal impact of each. Figure 1 summarizes the findings. The size of the blocks symbolize the explanatory power of each of the conditions, as will be further discussed.

The hindering conditions forgetting, stress, and isolation all seriously affected learning. While many participants used memory strategies to try and help them learn, forgetting remained a main issue. Isolation was a two-sided sword, as it left people with no one to practice the Dutch language and it also caused stress. Stress made people feel ‘their head was full’ and nothing new fitted in, with forgetting as a result. Isolation, stress, and forgetting are thus strongly related.

Motivation as well as LLS were identified as helping conditions. Participants emphasized their motivation for language learning, their efforts, as well as their vivid hopes and aspirations related to improved proficiency in Dutch. Nevertheless, for many this ideal-self image coexisted with a perception of their learning as failed. Therefore, I believe the explanatory power of motivation in this study to be limited.

All participants mentioned the use of at least some LLS. Their main strategies were selection and use of resources, memory strategies, and a variety of social strategies. The impact of memory strategies was limited, as learners using memory strategies still identified forgetting as a main problem. The same applied to obtaining resources for learning. Many participants watched Dutch TV programs and used digital tools for learning, but this was not related to increased self-efficacy or perceived learning outcomes. Hence, the explanatory power of these strategies was also limited. The impact of social strategies was slightly more complex. Cohen (2011)Cohen, A. D. (2011) Strategies in learning and using a second language. (2nd ed.). Routledge. distinguished between language learning and language use strategies. For language use, more specific to overcome communicative obstacles in formal situations, participants used the social strategy of seeking help from family members. However, this contributed to stress, as it interfered with the participants’ wish to be independent. For language learning, the family was generally not appreciated: family members were qualified as absent or impatient. However, social strategies were very effective for those participants who could rely on a new network of Dutch-speaking neighbors, friends, and colleagues. Participants who had succeeded in building such new social networks felt considerably better about their progress and showed greater confidence in further learning. This means that social strategies is the condition the study found to contribute most to LESLLA learners’ self-efficacy, with ‘new social networks’ as a prerequisite for social strategies to be effective.

The study thus provided a nuanced understanding of conditions that influence LESLLA learners’ language learning. It highlighted the complexity and interconnectedness of these conditions and showed that not all helping conditions have straightforward positive effects on language learning (for instance, memory strategies might not effectively reduce forgetting). The study pointed to interconnectedness of hindering conditions, creating a negative impact on learning. This suggests that it is important to address these conditions holistically rather than in isolation. The study emphasized that social strategies in new social networks have a significant positive impact on learners’ self-efficacy and language learning. This implies that social support in LESLLA learners’ language learning is critical.

4.1Motivation and investment

The study found LESLLA learners are motivated to learn Dutch and have heartfelt wishes related to being able to communicate in Dutch. Although research has suggested that motivation is an important predictor for language learning success (Dörnyei, 2009Dörnyei, Z. (2009) The L2 Motivational Self System. In Z. Dörnyei & E. Ushioda (Eds.), Motivation, Language Identity and the L2 Self (pp. 9–42). Multilingual Matters. ; Ushioda, 2008), in our data this relationship is less clear. Despite their motivation not all participants had learned Dutch sufficiently to engage in the communicative encounters they pictured themselves in and this affected their self-efficacy: many lost confidence in further language learning (Bandura, 1997Bandura, A. (1997) Self-efficacy. The exercise of control. W.H. Freeman and Company.). Dörnyei (2009Dörnyei, Z. (2009) The L2 Motivational Self System. In Z. Dörnyei & E. Ushioda (Eds.), Motivation, Language Identity and the L2 Self (pp. 9–42). Multilingual Matters. , p. 21) has argued that future self-guides must be accompanied by relevant action plans and effective procedural strategies, which suggests that for motivation to enhance learning, metacognitive strategies are required. Dörnyei’s pointing to ‘relevant’ action plan reflects a belief in individual responsibility and learners’ self-regulatory capacity, a perspective that has also been noticed in LLS theory (Thomas & Rose, 2019Thomas, N., & Rose, H. (2019) Do language learning strategies need to be self-directed? Disentangling strategies from self-regulated learning. TESOL Quarterly, 53 (1), 248–257. ). Following Dörnyei’s line of thought, it could be concluded that LESLLA learners are not effectively regulating their learning or have action plans that are not effective. However, an alternative explanation for the limited impact of motivation on LESLLA learners’ language learning can be found in Investment Theory (Darvin & Norton, 2015Darvin, R., & Norton, B. (2015) Identity and a model of investment in applied linguistics. Annual Review of Applied Linguistics, 35 , 36–56. ; Darvin & Norton, 2021Darvin, Ron, & Norton, B. (2021) Investment and motivation in language learning: What’s the difference? Language Teaching. ; Norton, 2013Norton, B. (2013) Identity and Language Learning (2nd ed.). Multilingual Matters.), which includes the social environment and explains how power relationships may impact language learning. As described in Section 3.3, LESLLA learners’ investment in language learning may be reduced by civic integration policy, which determines who must and may not attend school, and who is allowed to work. It is also impacted by housing policy: people with refugee status can not choose where to live, while becoming isolated is more likely when housed in a big city. Investment may be reduced by unwelcoming neighbors and communities. Thus, the social world and the power relations involved in it strongly affected investment in language learning, while motivation remained basically unaffected. Hence, compared to Dörnyei’s (2009)Dörnyei, Z. (2009) The L2 Motivational Self System. In Z. Dörnyei & E. Ushioda (Eds.), Motivation, Language Identity and the L2 Self (pp. 9–42). Multilingual Matters. concept of motivation, Investment Theory (Darvin & Norton, 2015Darvin, R., & Norton, B. (2015) Identity and a model of investment in applied linguistics. Annual Review of Applied Linguistics, 35 , 36–56. , 2021Darvin, Ron, & Norton, B. (2021) Investment and motivation in language learning: What’s the difference? Language Teaching. ; Norton, 2013Norton, B. (2013) Identity and Language Learning (2nd ed.). Multilingual Matters.) explains the learning of the participants in our study better.

4.2Language learning strategies

The study aimed at better understanding LESLLA learners’ use of LLS. With no exception, learners reported on the use of at least some, predominantly metacognitive, cognitive and social strategies, to help them learn. Many strategies did not seem to be LESLLA-specific, although an impact of L1 literacy on the use of memory strategies was found: learners with limited L1 literacy struggled with rehearsing. The use of cognitive and metacognitive strategies were not decisive for learning, as many learners who perceived their learning as ‘failed’ nevertheless revealed the use of those learning strategies.

Griffiths and Cansiz (2015)Griffiths, C., & Cansiz, G. (2015) Language Learning strategies: An holistic view. Studies in Second Language Learning and Teaching, 5 (3), 473–493. have argued that successful strategy use “may be related to a complex amalgamation of how many strategies are employed, how often, and how well they are orchestrated” (p. 497). This argument shows similarities with Dörnyei’s (2009)Dörnyei, Z. (2009) The L2 Motivational Self System. In Z. Dörnyei & E. Ushioda (Eds.), Motivation, Language Identity and the L2 Self (pp. 9–42). Multilingual Matters. notion of ‘relevant’ action plans and ‘effective’ procedural strategies for motivation to predict learning. It suggests that LLS are effective under the condition that they are effectively used. This is a somewhat circular argument. Griffiths and Cansiz’ (2015)Griffiths, C., & Cansiz, G. (2015) Language Learning strategies: An holistic view. Studies in Second Language Learning and Teaching, 5 (3), 473–493. notion of ‘orchestrating strategies’ implies individual responsibility and assumes learners’ self-regulating capacity. Hence, it depicts LLS as an individual’s asset. Our findings show that for social strategies to be effective, a supportive environment is needed, whereas building networks is strongly affected by the learners’ social context. Hence, social strategies are not purely an individual’s asset and it is not solely in the hands of individuals to orchestrate their use. Rather, social strategies are distributed in a social system, they are enabled by influential others as well as by circumstances. Unfortunately LESLLA learners are largely not in a position of power to shape those circumstances.

4.3Limitations and suggestions for further research

Aiming to shed light on their learning experience and LLS use, this study presented a CGT based on interviews with 30 LESLLA learners. There are certainly limitations to the study. First, the study draws on interviews that relied on interpretation. This somewhat disturbed the flow of the conversations and nuances of meaning may get lost in translation. Replications by researchers sharing the participants’ first language can be useful. Secondly, I interviewed participants once, which means that I could not tap into participants’ or researcher’s reflections that occurred after the interview. Also I could not capture development over time. Adopting a longitudinal, ethnographic approach could be a way forward, as it would enable an in-depth study of strategy use and motivation within contexts. Thirdly, participants were interviewed in an out of school context. Contextualizing interviews in classroom settings may elicit more detailed reflections on LESLLA learners’ strategy use and strategy needs. Studies in the classroom may therefore also provide insights in how LESLLA learners can be supported to expand their strategic repertoire.