Accuracy and Fluency Teaching and the Role of Extramural English: A Tale of Three Countries

: European learners of English are increasingly using this language recreationally, which is referred to as Extramural English (henceforth EE). The level of EE use in a given country might be reﬂected in English Language Teaching (ELT) practices. Yet, no research so far has examined cross-nationally what potential for language learning teachers perceive in their learners’ EE engagement and how this relates to ELT practices. To address this gap, the present study draws on interview data from lower secondary English teachers from Austria, France, and Sweden ( n = 20). They were enquired about (1) their students’ EE engagement and its effects on learning, (2) their accuracy and ﬂuency teaching methods, and (3) the perceived link between EE and ELT. Swedish teachers seemed to have a more positive and ﬁne-grained conceptualization of the impact of EE on learning than Austrian and French participants, especially in terms of grammar acquisition. The implicit learning environment that Swedish students encounter extramurally might extend to the classroom, where the use of explicit grammar rules occurs less dominantly than in the Austrian and French samples. The countries converged in the type of ﬂuency-based instruction they reported. Gaps in language areas not (fully) developed through EE seem to be more intentionally addressed in ELT in Sweden.


Introduction
In Europe and beyond, English has spread into learners' spare time language practices, which Sundqvist (2009) coined as Extramural English (henceforth EE). Until recently, the onset and intensity of learners' EE use in a given country was determined by whether English-language films and series were subtitled, i.e., supplemented by subtitles in the country's majority language, or dubbed, i.e., with original soundtracks being replaced by translated soundtracks. Unsurprisingly, EE engagement happened, and partly still happens, much sooner and more widely in subtitling countries, like Finland, Norway, and Sweden, than in dubbing countries, like Austria, France, and Spain. However, with the surge in online services, including on-demand platforms for music, series, and films, content is more easily available in original language globally. As such, EE is now becoming an increasingly important factor in students' language learning trajectories across countries, and such informal language use may in fact provide a key support for language learning and teaching (Pujadas and Muñoz 2020). However, the relationship between students' engagement in EE and English Language Teaching (henceforth ELT) practices remains under-investigated.
The level of EE use in a given context is likely to be reflected in teaching practices. For instance, previous research has shown that ELT in Sweden seems to rely more strongly on implicit fluency-based instruction than in Austria and in France, where, in contrast, teachers appear to provide more explicit grammar instruction (Schurz and Coumel 2020).
However, given the current stark increase in the use of EE also in dubbing countries, crossnational differences in the way English is learned are partly being neutralized. This might call for the need to find teaching methods that would be most fitting in contexts of high EE usage as well as with increasingly diverse learner groups, owing to different levels and types of EE engagement. In this endeavor, it is especially pertinent to explore what learners already bring to the classroom through EE and in what way ELT can complement those experiences. An exploration of these aspects is particularly informative if done comparatively across countries with varying levels of EE engagement. This study is the first to examine cross-nationally (i.e., in Austria, France, and Sweden) what potential for English language learning teachers perceive in their learners' EE engagement, and how this might relate to ELT practices-notably in terms of grammatical accuracy and fluency teaching.
In the following Section 1.1, we outline the key constructs of extramural English, accuracy, and fluency teaching, while Section 1.2 provides an overview of EE and ELT in Austria, France, and Sweden.

Extramural English (EE)
Extramural English designates English students' contact with English outside of the classroom (Sundqvist 2009, p. 25). This concept has been further defined as learners' primarily voluntary engagement in English activities with the goal of entertainment, with learning happening primarily incidentally rather than intentionally (Sundqvist and Sylvén 2016). The EE research in Europe has been carried out primarily in subtitling countries but is currently also emerging in dubbing countries. These studies clearly show the positive effect of such engagement on language learning, in terms of vocabulary knowledge (e.g., Peters and Webb 2018;Schwarz 2020), listening (e.g., Kuppens 2010;Lefever 2010), reading (e.g., Lefever 2010; Verspoor et al. 2011), and speaking skills (e.g., Lyrigkou 2019; Sundqvist 2009), grammar acquisition (e.g., Muñoz et al. 2021;Pattemore and Muñoz 2020), and learner motivation and/or confidence (e.g., Hannibal Jensen 2019; Sundqvist and Sylvén 2014).
Recently there has been an emerging interest in the effect of EE on ELT. Such research has shown that integrating EE-like material in class, i.e., authentic material that students might also use extramurally, can raise language awareness (e.g., of different registers) (Thorne and Reinhardt 2008) and increase learner motivation (Henry et al. 2018). In Toffoli and Sockett (2015), only about half of the 30 French EFL university professors included in the study felt their learners' EE use influenced their teaching practices. Yet, the vast majority of professors perceived their learners' EE use as beneficial for learning, in particular for their listening skills. Besides these studies, the connection between EE and ELT has received little attention in research.

A Categorization of the Type of Instruction
ELT tasks can be subdivided into accuracy-oriented and fluency-oriented activities (e.g., Brumfit 1984). While the former are dedicated to teaching accurate, norm-conforming language forms, such as in terms of orthography, phonology, or morphosyntax, the latter focus on improving spontaneous and seemingly effortless, fluent L2 production. In Communicative Language Teaching (henceforth CLT), i.e., the approach underlying ELT curricula and syllabi across Europe and beyond, accuracy is generally conceptualized in terms of grammar. According to CLT, accuracy and fluency teaching should ideally be combined in a balanced and meaningful way. This section provides the definition of the most important constructs within accuracy and fluency teaching, though we acknowledge that such categorizations may over-simplify actual teaching practices. Note that in this study, we focused on spoken rather than written fluency.

Accuracy Teaching
In order to investigate the relative effectiveness of specific methodological choices to support the acquisition of grammatical accuracy, researchers have typically classified teaching approaches in terms of explicit vs. implicit and inductive vs. deductive instruction (e.g., Ellis 2001Ellis , 2016Richards 2006).
In the case of explicit grammar teaching, teachers themselves bring the rules underlying target grammatical features to students' awareness. They may do so with or without using metalinguistic terminology. By contrast, in implicit instruction, learners are expected to grasp and acquire structures subconsciously. Within explicit instruction, teachers may present grammar inductively or deductively. The first technique implies that learners infer grammatical rules from text, which typically is enriched with the target feature, thus increasing its saliency. In deductive instruction, teachers present grammatical rules and in turn ask students to practice and produce applications thereof ('PPP' method). Involving students in working out rules themselves is more in line with the learner-centred approach of CLT and may increase learning outcomes (Richards 2006).

Fluency Teaching
While (grammatical) accuracy is a fairly well-defined construct, the complementary concept of fluency is much less clearly described. Lennon (1990) distinguishes between fluency in a broad sense, i.e., as a near-synonym to proficiency, and in a narrow sense, as referring to "the psycholinguistic processes of speech planning and speech production ( . . . ) functioning easily and efficiently" (Lennon 1990, p. 391). Research on fluency has mostly focused on the narrow view of fluency. Key predictors of proficient fluency are speakers' speech rate (usually measured in syllables per minute), mean length of runs (i.e., syllables produced between pauses), and numbers of dysfluencies and (filled and unfilled) pauses (Lennon 1990;Préfontaine and Kormos 2015). In terms of teaching practice, this body of research suggests positive effects of explicit instructions in the use of formulaic sequences (Wray 2008), increased planning time for tasks (e.g., Tavakoli and Skehan 2005), and repetition, possibly with increased time pressure (e.g., Sample and Michel 2014). Tavakoli and Hunter (2018) established, however, that teachers in the UK, firstly, generally interpret fluency in the broad sense, with a focus on speaking only and additionally on conveying clear meaning. Secondly, these research-based suggestions found hardly any uptake in teachers, whereas communicative free production activities and general language proficiency activities (e.g., listening to Native Speakers, vocabulary work, etc.) were mentioned by 53.6% and 13.5% of respondents, respectively (ibid., p. 339). This is in line with experiences in other countries, like Austria (Hüttner 2013), and with teacher guidance found in popular resources (e.g., Richards 2015;Hedge 2008)  Through today's broad access to online games, social media apps, and streaming platforms, EE use is skyrocketing globally. In Sweden, 10-to 11-year-olds (n = 76) were found to spend a weekly average of about 7 h on EE activities (Sundqvist and Sylvén 2014), compared to 35 h among 16-to 19-year-olds (n = 230) (Olsson and Sylvén 2015). These numbers are expected to be even higher today, but no recent data are available. In Austria, about 40% of 13-to 14-year-old students watch films and series in English at least multiple times a month (Federal Institute for Education Research, Innovation and Development 2020), and in the capital city of Vienna, 15-to 16-year-olds reported a weekly average of 28 h of EE use (Schwarz 2020). Although no report on weekly hours are available for France, 34% of 14-year-olds reported watching subtitled English-language audio-visuals at least once a week, and 50% of students responded using English on the internet or in video games often or very often (National Research Center of School Systems 2019). Unfortunately, the available data do not allow for a direct comparison between countries. However, despite the global surge in EE, we expect differences in levels of EE use between subtitling and dubbing countries to subsist. For instance, on streaming platforms such as Amazon Prime and Netflix, speakers of languages that are not as widely spoken (e.g., Finnish, Norwegian, and Swedish) generally do not have access to dubbed foreignlanguage content, whereas speakers of more widely spoken languages (e.g., German, French, and Spanish) do. Likewise, the opportunity to engage in social media content and co-gaming practices in one's first language may be more available to speakers of languages with a larger speech community.

English Language Teaching
It is worth reviewing previous research on the type of instruction in ELT in Austria, France, and Sweden against the background of the countries' national ELT curricula. Since this study focuses on ELT in lower secondary education, we first need to briefly describe each country's school system. At the beginning of lower secondary education in Austria, the student population is split into Middle School (Ger.: Mittelschule) and the more selective Academic High School (Ger.: Allgemeinbildende Höhere Schule) at the age of 10-11 years (Austrian Federal Ministry of Education (AME) 2018a). Comparatively, in France, students follow the same educational trajectory and enter Middle School (Fr.: Collège) at the age of 11 (Ministère de l'Education Nationale (MEN) 2020). In Sweden, the comprehensive school system encompasses years 1-9 (Swedish National Agency for Education 2020).
Even though the countries' respective ELT curricula all purport CLT, there exist some differences in their references to grammar teaching or the lack thereof. The Middle School curricula of Austria (Austrian Federal Ministry of Education (AME) 2018b) and France (Ministère de l'Education Nationale (MEN) 2015, 2016) seem prescriptive in specifying which grammar features teachers should introduce in each grade level. Such references are absent in the curricula of the Swedish Comprehensive School (Swedish National Agency for Education 2017) and the Austrian Academic High School (Austrian Federal Ministry of Education (AME) 2020). The Austrian Middle School and Academic High School curricula (Austrian Federal Ministry of Education (AME) 2018b, 2020) additionally recommend specific grammar teaching methods, namely the use of implicit and inductive techniques. The French curriculum (Ministère de l'Education Nationale (MEN) 2015) very vaguely refers to implicit instruction when listing possible activities in reading tasks, namely the memorization and transfer of occurring structures. Such more or less specific methodological recommendations are absent in the Swedish curriculum (Swedish National Agency for Education 2017). Whereas neither the Austrian nor the French curricula refer to the students' spare time use of English, the Swedish curriculum proposes considering learners' interests, experiences, and needs in instruction (Swedish National Agency for Education 2017), which arguably includes the consideration of students' EE. Overall, these data suggest that in Sweden, teachers may have greater leeway as to what to teach, when, and how.
Focussing on teacher-reported practices, Schurz and Coumel (2020) found that lower secondary school teachers (n = 615) in Sweden provide significantly more implicit-fluencybased and significantly less explicit instruction than practitioners in Austria and France. These findings were corroborated by a qualitative follow-up study using teacher interviews (n = 20) (Schurz and Coumel 2021). The two studies further revealed that incidental (rather than systematic) grammar teaching, as based on current student needs and interests, was particularly dominant in Sweden. The authors hypothesized that this might be due to the greater proficiency diversity in Swedish classrooms, given the rare occurrence of grade repetition and a school system that is comprehensive rather than selective up to grade 9. Finally, Schurz and Coumel (2020) found that French teachers agreed more strongly to introduce grammar rules inductively (vs. deductively) than teachers from the other countries.

Research Questions and Hypotheses
Although the above-mentioned research suggests that ELT varies between Austria, France, and Sweden Coumel 2020, 2021), no study, to the best of our knowledge, has cross-nationally investigated teachers' perceptions of their learners' EE use, and how this might explain apparent differences in the type of instruction they apply. We addressed these research gaps with five research questions. We first examined lower secondary school learners' EE use (RQ1) and the impact their teachers thought such EE use had on the students' learning (RQ2). Then, to be able to assess the link between EE and ELT, we first deemed necessary to investigate which teaching practices the teachers would report using: specifically, we focused on accuracy (RQ3) and fluency teaching (RQ4). Finally, to better understand the relationship between the learners' EE engagement and English instruction, we analyzed data in which teachers themselves drew explicit links between EE and ELT (RQ5).

RQ1
What are the EE practices of lower secondary school learners of English as reported by teachers in Austria, France, and Sweden? RQ2 What is the effect of EE on learning as estimated by teachers from the three countries? RQ3 How is grammatical accuracy taught-e.g., implicitly vs. explicitly, inductively vs. deductively-as reported by the teachers from the three countries? RQ4 How is fluency taught as reported by the teachers from the three countries? RQ5 What is the impact of learners' engagement in EE on ELT, according to the teachers' views?
In response to these research questions, we formulated five hypotheses. First, we assumed that lower secondary school students in Sweden would engage more extensively in EE than students in Austria and France (H1). This was because we expected countries with subtitling practices and with a majority language that counts relatively few speakers to show higher levels of EE (see Section 1.2.1). Second, given this pattern, we also expected teachers in Sweden, a country with a longer experience in considerable EE use, to believe more strongly in the acquisitional benefits of EE as compared to teachers in Austria and France (H2). Third, based on Schurz and Coumel (2020), we hypothesized that teaching would be more explicit and rule-based in Austria and France than in Sweden (H3.1), and more inductive in France than in the other countries (H3.2). Fourth, we predicted that instruction targeting the promotion of fluency rather than accuracy would prevail in Swedish rather than Austrian and French classrooms (H4) (Schurz and Coumel 2020). Finally, while acknowledging that the data of this study would not be sufficient evidence to determine a causal relationship, we expected the teachers' methodological choices to reflect their learners' level of engagement in EE to some extent (H5), since traces thereof had previously been observed Coumel 2020, 2021).

Participants
The data collected for the present study originate from the same interviews as reported in Schurz and Coumel (2021). A total of 20 English teachers-six from Austria, seven from France, and seven from Sweden-were interviewed in November 2019-2020. We used a convenience sample and recruited teachers via the researchers' existing social and professional networks. Invitations to participate were also posted on social media, i.e., Facebook, and participants took part voluntarily. All of them teach at lower secondary level and more specifically in Austrian Middle School and Academic High School, French Middle School, and Swedish Comprehensive School. While the interviews conducted with teachers from Austria and France focused on students at the age of 13-14 years, teachers from France responded more broadly in terms of learners aged 11-15 years. The mean age of participants was 40.2, 48.4, and 43.7 years and the mean experience in teaching 16.8, 22.7, and 14.9 years for Austria, France, and Sweden, respectively. The majority of teachers taught in a public school. For the teachers' pseudonyms and their age, number of years of teaching experience, the age group they reported on, and class sizes, see Table 1. The study was approved by the Ethics Committee of the University of Vienna and the teachers provided informed consent prior to participation.

The Interviews
The semi-structured interviews (see Table 2) contained full-length overarching questions as well as prompts for sub-topics (cf. bullet points); the former were used as such during the interviews, while the latter were used as starting points or examples to extend the discussion. The questions targeted (1) the learners' level and type of EE engagement; methods applied in (2) accuracy teaching and (3) fluency teaching and the estimated impact of EE on learners' accuracy and fluency, and (4) the link between EE and ELT, e.g., as seen in the teachers' encouragement of learner engagement in EE, the teachers' perceived impact of EE on learning, and teachers' adaptations of ELT based on the learners' EE use. Authors 1 and 2 conducted the interviews, which took place on-site (for Austria and Sweden) or online (for France), and in the participant's first language. The 10-25-min interviews were audio recorded.

Data Analysis
Following orthographic interview transcription, we first used MAXQDA to code the data according to the broad categories (1)-(4) listed in the previous paragraph for each teacher. We then refined our analysis with a more specific coding system. For instance, we examined the specific EE activities that the teachers reported their students to engage in (e.g., blogging, reading books, gaming, travelling, watching TV/series/films, using video apps, listening to music); teachers' use of implicit, explicit, inductive, and deductive grammar instruction; monologic and dialogic fluency-based activities and learner difficulties in such practice; the language areas that teachers believed EE would benefit; and the extent to which teachers reported accounting for EE in their language teaching practices. Finally, we summarized the findings per country, e.g., by assessing the number of teachers describing a given practice in each country.

EE Practices (RQ1)
As predicted (H1), based on teachers' reports, lower secondary students appear to use EE most extensively in Sweden. There, students seem to "have English around themselves everywhere" (Christine), "all the time" (Pia), and already early on (Sara). It seems to be used recreationally by all students, albeit to different degrees (Magnus). In contrast, in Austria and in France, regular EE use at that age-apart from listening to music-seems to be the exception rather than the rule. It typically concerns just a few individuals per class, as expressed by two Middle School and two Academic High School teachers. Elena pointed out that in her class, EE is used "not so much . . . of the 16 people . . . of [two] I can say with certainty that they definitely do something in their free time".
While the French teachers did not quantify the amount of EE users per class, their illustrations evoked the same picture as in Austria.
The types of EE activities seemed comparable across countries, with series, videos, and gaming being favored-disregarding today's omnipresence of anglophone music. Each Austrian teacher reported there to be gamers among their eighth graders, many of whom use English to communicate with co-players. Only very few children read books in English, while three teachers affirmed that in each of their 4th grade classes some students regularly watch audio-visuals or use TikTok. Likewise, in France, series and gaming emerged as the most frequent EE activities, pointed out by four teachers. Other types of activities were conversing with others when travelling and communicating with pen pals, listed by one French teacher each. Swedish students are also reported to most heavily engage in audiovisual media and gaming, though other EE activities, like blogging (Pia) and karaoke (Karin), were mentioned as well. Gaming and the oral or written interactions it often entails can happen quite extensively, with Pia reporting a striking 4-5 h a day for some students. Previous research, too, has shown watching audiovisuals, and, among boys, gaming to be the most popular EE activities across countries (e.g., Hahn 2018; National Research Center of School Systems 2019; Olsson and Sylvén 2015;Sundqvist and Sylvén 2012).
While in Sweden, students are to a certain degree necessarily immersed in English due to the country's subtitling practices, three Austrian and two French teachers reported that students mainly used this language recreationally out of personal interest. For instance, bloggers (or, rather, vloggers), described as "ultra-trendy" (Julia), often create their content in English, which leaves students with little alternatives than to use this language. Yet, one Austrian teacher, Elena, said that EE is mostly used by upper secondary students, to improve their language skills for tests.
In sum, the data allowed us to confirm H1, according to which students would engage in EE most extensively in Sweden (see Section 1.2.1). Nevertheless, EE undoubtedly has reached the dubbing countries, where sometimes already pre-teenage learners start engaging in it regularly. In general, however, the lower levels of EE reported in our Austrian sample may illustrate that countries with a majority language that is more globally used, such as German in Austria and French in France, have greater access to original or dubbed content in the given language; in contrast, countries with a majority language that counts relatively few speakers, such as Sweden, show higher levels of EE (see Section 1.2.1).

The Effect of EE on Learning (RQ2)
While the general attitude of lower secondary teachers towards the effect of EE on learning was rather positive across the three groups, the starkest cross-country differences emerged in terms of whether teachers thought that EE supported grammar learning. The Swedish teachers indeed believed more strongly in the benefits of EE on grammar learning than teachers from the other countries, which provides partial support for H2 (see below).
First of all, the teachers from the three countries most often reported that EE could benefit vocabulary learning. This is also the prime area of language learning targeted in EE research (see Schwarz 2020), which could have influenced teachers' perceptions. Other language areas addressed were improved speaking skills in general and pronunciation and fluency in particular, enhanced receptive skills, and increased learner self-confidence and motivation. Likewise, as discussed in Section 1.1.1, substantial research has demonstrated the very positive effects EE has on vocabulary learning, aural and written comprehension, speaking skills, and learner confidence and motivation (e.g., Hannibal Jensen 2019; Kuppens 2010; Schwarz 2020; Sundqvist 2009).
In contrast, only Swedish teachers and one French teacher mentioned the effect of EE on language registers and writing. Three Swedish teachers stated that EE did not help students improve their writing and/or spelling skills, with autocorrect seeming a major issue (Eva). Another Swedish teacher pointed to the fact that in EE use, students are almost exclusively exposed to informal language. Only according to one teacher are writing skills being developed through EE, though they did not specify which language registers students would acquire this way. Among Austrian and French teachers, only Anne from France referred to registers and indicated that EE increases learners' awareness of different accents and registers. The otherwise suggested lack of support of EE in the development of writing skills and formal language might call for a focus on these aspects with learner groups where EE engagement is high (Thorne and Reinhardt 2008; see 'bridging activities', Section 4.5).
Importantly, the countries mostly differed in the extent to which the teachers believed EE could benefit grammar acquisition. In Austria, the three Academic High School teachers generally believed in grammar being acquired through EE, although Andrea conceded that in series, for instance, conditionals and passive voice are not always used "correctly". This comment aside, Julia described a student who uses EE extensively as producing very native-like English. In contrast, two Middle School teachers were less optimistic. For example, Barbara felt that the speed of fluent speech in authentic input did not give learners the opportunity to use language accurately, thereby disregarding the possibility of learning through receptive language use. Similarly, Lukas excluded the possibility of learners at that level acquiring grammar through EE use and notably gaming. He explained that "when you work on TeamSpeak with other gamers that's-it's about very fast reactions, not only in playing but also in speaking. And for that they are not experienced enough with the language." Likewise, in France, Lucie excluded the possibility of grammar being learned extramurally, with the remainder of French teachers not referring to this possibility at all.
In Sweden, in contrast, all teachers indicated that at least to some extent, EE supports grammar acquisition. Yet, the degree to which this is possible may hinge on the type and intensity of EE use and on learner talent (Eva). Other research similarly suggests that learner talent, or more specifically, implicit aptitude, could determine levels of uptake through implicit learning conditions (e.g., Godfroid and Kim 2021)-although more research is needed to better understand the nature of this relationship. Sara contended quite illustratively that grammar acquisition through EE "can work to a certain level, and then one could start learning the rules. Like why should it say this and why not that . . . But I think you can learn it quite well like for example with irregular verbs." In this statement, Sara makes explicit the relative support provided by EE and ELT in constructing grammatical knowledge. Another aspect emanating from the quote is the target feature's saliency. Salient and high frequency features, such as irregular verbs, might be acquired relatively easily implicitly and thus through EE; in contrast, low-saliency features, such as third person -s, perhaps require more explicit teaching (e.g., Kang et al. 2019).
Overall, we found stark cross-national differences in teachers' perceptions of the effect of EE only in terms of grammar acquisition: Swedish teachers were more positive in that regard than teachers from the other countries. Since teacher perceptions otherwise did not differ greatly across countries, we can only partly confirm H2 (see Section 1.1.1).

Accuracy Teaching (RQ3)
As concluded above, lower secondary teachers in Sweden seem to believe more strongly that EE supports grammar acquisition than the other teacher groups. This may explain why, as predicted in H3.1, Austrian and French teachers teach grammar more explicitly than Swedish teachers.
Indeed, our results suggest that in Austria, the presentation of grammar features typically happens explicitly, either inductively or deductively. Julia and Lukas report using an inductive approach, first exposing students to a text or sentences and guiding them towards constructing the underlying rule. Others contended that the inductive approach only works well with certain features, such as regular past (Elena), indefinite articles, and plurals with -ies (Barbara). Such references to the inductive approach in the Austrian sample may reflect the recommendation to resort to inductive teaching in the Austrian curricula (Austrian Federal Ministry of Education (AME) 2018b, 2020) (see Section 1.2.2).
Similarly, all teachers in France prefer an explicit approach and provide students with grammar rules. Three teachers reported introducing grammatical features through an inductive approach. For example, Lucie explained having her students "discover [a grammar feature] with sentences written on the blackboard, from a previously studied text. Students then think, I help them think . . . and then give them the grammar rule".
Likewise, three teachers preceded the presentation of a new grammatical rule with elements of what can be referred to as 'discovery learning' (see Gollin 1998). For example, Laure makes use of situations in which students themselves direct attention to form based on their immediate need of a specific structure to perform a task. This allows her to subsequently introduce the new form.
Following the introduction of a grammatical feature, teachers in Austria and France reported having their students practice it, with many participants referring to traditional gap-fill activities. Three Austrian and five French teachers illustrated having students practice grammar in spoken and/or written production. The production tasks thus seem to be implemented to practice grammar rather than skills (e.g., 'writing to learn', (Manchón and de Larios 2011), although in a more implicit than explicit way. Implicit grammar practice is another suggestion put forward by the Austrian (Austrian Federal Ministry of Education (AME) 2018b, 2020) and vaguely also in the French (Ministère de l'Education Nationale (MEN) 2015) curricula (see Section 1.2.2). The otherwise apparent important role teachers attributed to explicit instruction is, however, not a reflection of curricular recommendations. The overall pattern of grammar teaching in Austria and France reported here neatly follows the traditional presentation, practice, and production formula (PPP method, Ellis 2001; Hedge 2008, see Section 1.1.2.1).
In Sweden, as discussed in Section 4.2, all teachers believe that grammar is acquired implicitly. One teacher, for instance, believed more in learning grammar through listening and reading rather than through explicit instruction. In contrast to the other two contexts, students usually encounter grammatical features extramurally first. Only after such encounters do students seem to be receptive to and interested in learning the corresponding underlying rules. Like Sara (see quote in Section 4.2), Karin affirmed that "letting them . . . speak and . . . when they have reached a certain language level this is when grammar becomes interesting. Because this is when they realize that there has to be a rule".
Yet, two Swedish teachers emphasized that it is not important for learners to know the rules. Eva points to individual differences in whether rule-based learning is beneficial, with some students already having a feeling for how a structure should sound. Similarly, Christine argued that it is students with a lower language aptitude who require more explicit instruction. Teachers also mentioned the idea of directing attention to form in response to learner interests or needs. Three Swedish teachers for instance draw on learners' mistakes in class based on their written production and have them discuss possible corrections in groups (Pia). This learner-centered approach is also reflected in the common procedure of individualized grammar practice, as for example hinted at by Emma: "If you haven't understood at all, then you should practice on kids' level but if you understood a lot, why practice grammar if you master it already?" To summarize, teachers from all three countries at least partly rely on explicit teaching. However, Austrian and French teachers seem to apply the PPP method more consistently and might attach a greater role to rule-based knowledge. Thus, our results confirm H3.1, according to which Austrian and French teachers resort more to explicit instruction than Swedish teachers. Swedish teachers appear more open to cater to individual needs for instance in terms of whether grammar rules aid learning, and grammar practice perhaps happens on a more learner-centered, individual basis. This is in line with previous reports Coumel 2020, 2021) and reflects CLT (see Section 1.1.2). The cross-national differences might be linked to specific educational traditions, such as ideologies of education and resulting selection policies. As reported in Schurz and Coumel (2021), the Swedish Comprehensive School necessitates greater individualized instruction, whereas the more selective school systems of Austria and France allow for highly teacher-led instruction. Likewise, while the Swedish curriculum (Swedish National Agency for Education 2017) provides teachers with considerable leeway as to what to teach and when, the Austrian and French curricula (Austrian Federal Ministry of Education (AME) 2018b, 2020; Ministère de l'Education Nationale (MEN) 2015, 2016) appear more prescriptive. Yet, the Austrian and French curricula do not advocate explicit instruction, although this appeared to be so commonly used by our participants. Another important factor determining didactic choices seems to be learners' EE practices. Learners' amount of and type of EE use is highly individual, enhancing proficiency diversity within classes. In Austria and in France, learning English still takes place primarily in the classroom, allowing for all students in a given class to move on simultaneously and as determined by the teacher. As the distinction of inductive vs. deductive instruction presupposes a teacher-centered introduction of a new target feature, such a categorization seems not applicable to the Swedish context, where high EE use typically implies incidental encounters with a given grammatical feature already before it is referred to in class. Therefore, our results can support H3.2 only for Austria and France, with French teachers indeed apparently teaching grammar more often inductively than Austrian teachers.

Fluency Teaching (RQ4)
Our results do not clearly support H4, as we found fewer cross-national differences in terms of fluency instruction. In all countries, lower secondary teachers appear to engage students in both monologic and interactive tasks, with pair or group work being a very common interaction format. Across countries, teachers mostly listed activities such as summarizing a text, presenting a book, and discussing random topics. While teachers from all samples referred to learner difficulties in fluency and speaking, the examples teachers gave were very diverse.
The Austrian Middle School teachers reported difficulties students face in speaking even for a short amount of time, such as three minutes (Veronika). This is why the teachers often let learners prepare speaking tasks in advance at home. Academic High school teachers, on the other hand, did not mention learner difficulties in speaking, with Julia pointing out that her students are very much used to talking in pairs.
In France, according to three teachers, students' main difficulty in fluency tasks arises from a lack of confidence, while Lucie stated that learners try to rely on word-by-word translation from French. Teachers report using scaffolding to help students prepare for fluency-based tasks, which can consist of basic structures, such as for instance question types (Laure), vocabulary or keywords (Laure, Sophie), or short scripts with underlined words that students need to change later (Anne). Other techniques were making sure that her students know that they are allowed to make mistakes (Sophie) and giving students activities that are adapted to their abilities (Anne).
Similarly, three Swedish teachers reported that their learners often felt uncomfortable to speak in class. Another issue raised was the one of students switching back to Swedish during English speaking tasks (Sara). Related to the proficiency diversity observed in Section 4.3, Eva pointed to great differences in the learners' perceived difficulty of such activities. To give students the opportunity to speak English in a comfortable environment, Christine reported frequently having students talk and record their speech in small groups in class, in a 'group room' attached to the classroom, or at home. Giving students the chance to practice speaking at home also occurred in the Austrian Middle School sample.
Finally, and perhaps very much illustrating the Swedish context, Eva and Magnus explained the issue of the learners' lack of motivation in class, which may be related to extramural English. According to Magnus, students often "feel that [speaking in class] is just another task and that it is just school, and that this is why they do it." This is a straightforward manifestation of what Henry (2013) termed the 'authenticity-gap' between students' high engagement in 'real' language through EE and in comparatively 'inauthentic' and 'boring' lessons. To try to keep up the learners' motivation and fill this gap, Magnus suggested engaging learners in activities and topics they can relate to, while Eva likes resorting to speaking games. Similarly, a study conducted in Sweden showed that teachers who successfully connect students' EE with ELT often integrate authentic materials and digital technologies in class (Henry et al. 2018).
In sum, in line with the findings of Tavakoli and Hunter (2018), teachers generally interpreted fluency in the broad sense, with a focus on speaking (see Section 1.1.2.2). Both monologic and dialogic speaking activities occurred in all countries, and teachers in France, Sweden, and Austrian Middle School reported that their students often struggled when performing speaking tasks. Therefore, it was not possible to confirm H4, which proposed that English classrooms would be more fluency-based in Swedish rather than Austrian and French lower secondary school (e.g., Schurz and Coumel 2020). Yet, peculiarities found in the Swedish sample were individual references to the previously observed great ability diversity in classes and the importance attributed to the incorporation of (authentic) tasks that students find interesting and 'real'. Especially the last observation was found to be quite clearly linked to learners' extramural English.

Linking EE and ELT (RQ5)
After these tentative evaluations of the relationship between the learners' EE engagement and accuracy and fluency teaching, we here look at data in which teachers themselves explicitly drew the link between EE and ELT. To do so, we asked the teachers to report whether they encouraged their learners to use EE and whether they adapted their classroom practices to better complement EE use. Teachers from all samples affirmed that they recommend learners to engage in EE. However, as expected in H5, it emanated from the reports that learners' level of EE use in a given country impacts teaching practices.
In the Austrian and French samples, the teachers generally did not appear to take EE greatly into account in their teaching practices. Three teachers in Austria and five teachers in France said that they do not purposely consider their learners' EE use in their pedagogical choices, likely owing to the limited EE use in this context. For instance, Charlotte reported "if I had students who were not exposed [to EE], I would use the same teaching material".
As to the remaining teachers, Veronika in Austria explained that she had her students sometimes analyze song texts, and further recalled having tried watching films in English with her fourth graders in class, which turned out too difficult for them without German subtitles. Andrea, on the other hand, was convinced that EE use nowadays forms an intricate component of learning the language, being equally important as instructed learning. In France, only Lise exemplified that she asks high EE users to do presentations on specific topics or explain grammar rules to their peers.
In contrast to Austria and France, the overall higher EE use among Swedish learners seems to affect classroom practices to a greater extent. For this phenomenon, different teachers provided different examples. Emma and Karin referred to major proficiency differences within classrooms emerging because of the varying levels of EE engagement. As discussed in Schurz and Coumel (2021), proficiency gaps within classes could make teachers opt for incidental rather than systematic grammar teaching. In line with the finding of Swedish teachers generally believing in EE aiding grammar acquisition, Emma also reported that in her class, since many students do not use English at home, grammar required more work. However, as previously discussed, grammar learning is perceived to be promoted through EE up to a certain point, after which teaching rules becomes necessary (see Section 4.3). Regarding the issue of EE involving primarily informal, both aural and oral language use, Magnus mentioned the idea of having students write about their EE activities in order for them to practice writing. In an attempt to raise learners' awareness of different registers, Christine explained that "you have to show them different levels-like the language they use with their friends at home, that is very good, but then one needs to give them the other side as well".
For instance, she reported drawing learners' attention to the difference between formal and informal letters. This represents another manifestation of 'bridging activities' (Thorne and Reinhardt 2008), used to complement for areas of language competence that are not promoted through recreational engagement (see also Section 4.2 and Henry et al. 2018). The observation of four Swedish teachers readily illustrating examples of where the connection between EE and ELT becomes visible may reflect the Swedish curriculum and its recommendation to take into account learners' own experiences in the classroom (Swedish National Agency for Education 2017) (see Section 1.2.2).
Taking the findings from Sections 4.3-4.5 together, we can confirm H5, suggesting that teaching practices in the three countries would reflect the learners' level of EE engagement. While this conclusion must be more tentative in terms of the reported pedagogical choices in accuracy and fluency teaching, we could spot a number of instances where teachers themselves drew a direct link between ELT and EE. Such a link emerged primarily in the case of Sweden-as hypothesized, the country with the longest experience in learners' EE usage.

Conclusions
Today's lifeworlds of young people integrate an online, typically English-speaking, element. This revolves heavily around watching series and videos in the English original, as well as gaming with other users, frequently employing English as a lingua franca. The challenge for English teachers is now to ensure that their classrooms relate in a positive way with these lifeworlds.
This study addressed lower secondary school teachers' responses to this challenge in three European countries with diverse profiles of EE: Sweden, with a long tradition of using recreational resources in English directly or only with subtitles, and Austria and France, both countries where films and series are traditionally dubbed into the national languages, but where the rise of Netflix and Gaming has increased the use of EE over the last few years.
Findings suggest a growing awareness among teachers in all countries of the role played by EE in students' lives and on the potential positive benefits of EE on general English language proficiency. Teachers from all samples see benefits for receptive skills, learners' vocabulary range, as well as self-confidence and motivation. Interesting differences exist with regard to grammar, where teachers in Sweden view EE as having a clearly positive learning effect on grammar; a view not shared by French teachers and seemingly under debate among Austrian teachers. These differences seem to correlate with grammar teaching practices, where Swedish teachers more heavily rely on implicit grammar learning, but French and Austrian teachers consider explicit teaching (whether deductive or inductive) essential for their learners. Whether this is a direct effect of high EE use, or an inter-relation of high levels of EE and specific educational traditions cannot be unambiguously answered. Moreover, given that we only assessed these issues indirectly, via teacher reports, further research ideally should integrate student reports and classroom observations.
Overall, however, we can observe that the rise of EE challenges the role of English as a purely foreign language for school learners. English increasingly takes on the status of a second language, present in the young learners' digital lifeworlds (see Andersson 2013). In order to keep ELT relevant and not just "another task" to be done at school, teachers in these three European countries are responding to varying degrees to the more complex learning environment of their students. The future will show when and how educational policies will more uniformly seek to adapt ELT to the changing environment.

Data Availability Statement:
The interview transcripts of Austrian and Swedish teachers will appear in the first author's dissertation in Fall 2021. The interview transcripts of French teachers are not available because per-mission to publish them in their entirety was not obtained from the teachers.