2.1 Applying the notions L2 and L3 in research on language teaching

The notion L3 (cf. De Angelis 2007; Hufeisen/Neuner 2004a; Williams/Hammarberg 1998) builds upon the distinctions introduced in the factor model by Hufeisen (1998 and 2000), who suggests that different factors, mainly cognitive, are involved in the learning of a first language (L1), a first foreign language (L2) and any other language thereafter (Lx). The rationale for differentiating L2 and L3 in language teaching is that there is a qualitative difference between the two (cf. Hufeisen 1998: 171; Jessner/Cenoz 2007: 158) and that L3 learners, unlike L2 learners, bring certain knowledge and strategies to the table thanks to their prior experience of learning a foreign language. L3 is conceptually different from L2, since it is “influenced by the degree of bilingualism already attained by the student” (Jessner 2008: 34), which in Swedish L3 classrooms refers to the presence of English as a mutual L2. In Sweden, as in many other contexts, German is introduced in schools after English.. Since the students with Swedish L1 already have the experience of learning a foreign language (English L2), German as a school subject has the status of an L3.

2.2 Teachers’ beliefs and attitudes towards background languages in L3 teaching

Research on teacher cognition has pointed out the indisputable links between teachers’ thinking or beliefs and their decision-making in classroom practices (cf. Borg 2006 and 2019). It is therefore important to take teachers’ stated pedagogical beliefs and attitudes into consideration when languages in educational contexts are discussed. With the multilingual turn in language education from the past decades (cf. Aronin/Jessner 2015; Cenoz/Gorter 2015; Conteh/Meier 2014a; May 2019; Ortega 2013), multilingualism, multilingual approaches and learners’ previous background languages are today seen as assets in language teaching and learning (Conteh/Meier 2014b: 3; Juvonen/Källkvist 2021: 1–6), not least by many L3 teachers (see e.g. Falk/Lindqvist 2022: 161–163; Haukås 2016: 8–10).

Previous research has, however, also noted more conservative viewpoints on this topic, where some teachers believe that background languages can interfere with the learning of a target language (De Angelis 2011: 227; Goldenberg 2013: 9). When it comes to the influence of English, referred to as a “lingua frankensteinia” by Philipson (2008: 251), there are some documented reluctant stances among GL3 teachers in Scandinavia. In the Swedish GL3 teacher discourse, the derogatory term English illness (cf. Falk/Lindqvist 2022: 166; Håkansson/Pienemann/Sayehli 2002: 269) refers to negative transfer from English. Falk/Lindqvist (2022) noted that comparisons with English (unlike those involving Swedish) were not encouraged by the GL3 teachers in their study, where one teacher saw English as both an error and “an insult” (163) and that “comparisons with English [were] ‘forbidden’ in their classrooms when it comes to grammatical structures since this would render ungrammatical sentences” (164). This viewpoint of the GL3 teachers is in clear contrast to teachers of French L3 in the mentioned study, who instead regarded transfer from English as a productive strategy. A similar difference was reported in a Norwegian study (Haukås 2016) on L3 teachers’ pedagogical beliefs about multilingual approaches, where the teachers of German stand out compared to their colleagues of French and Spanish. Although all teachers reported drawing on the learners’ knowledge of both L1 Norwegian and L2 English to enhance the learning of the L3, Haukås’ results indicate that the two German teachers preferred making crosslinguistic comparisons with L1 Norwegian, as illustrated in the following quote:

Jan feared that a focus on L2 English in L3 German class might lead to more language mistakes. He had noticed that some students unconsciously transferred linguistic patterns from L2 English, which led to what he called bad German. ‘Things just get more right if they think in Norwegian’, Jan stated. (Haukås 2016: 10)

In this particular case, the teacher’s viewpoint could be seen as more than a mere belief, but rather a displayed (negative) attitude towards a crosslinguistic approach that would also include explicit references to English, thus disregarding its pedagogical function as a bridging language that could help draw the learners’ attention to the similarities between the L1 and the target L3.

2.3 Multilingual and crosslinguistic pedagogies

The promotion of multi-/plurilingual competence is contained in the European framework of reference for languages (Council of Europe 2020: 30–31), which implies the need for multilingual pedagogical approaches where the core idea is that boundaries between languages should be softened in education (cf. Cenoz/Gorter 2015: 4). The rationale for such pedagogy is that crosslinguistic connections and linking of known languages can help learners increase their language awareness and develop language learning strategies (cf. Cummins 2007: 229; Séror/Gentil 2020: 370–371). It is also suggested that multilingual teaching activities can have a positive effect on students’ motivation (cf. Lyster/Collins/Ballinger 2009: 378). A multilingual pedagogical approach takes the role of all previously learned languages into consideration, acknowledging the “ongoing interactions between the different language stores of the multilingual speaker/listener” (Rast 2010: 162). From an L3 perspective, this implies that features from both L1 and L2, as well as the use of learning strategies and techniques from the experience of learning and mastering the L2 (and possibly other learned L3s), are sources of transfer that can be used as pedagogical tools to promote language teaching and learning (cf. Ballinger et al. 2020: 265–267).

In the last decades, several terms for multilingual approaches have been introduced. A commonly used umbrella term for the acknowledgement and encouragement of background languages and varieties in social, cultural and educational domains is translanguaging (cf. García 2011; García/Li 2014). This term also taps into societal and political discourse about spontaneous language use and the status of regional and minority languages (cf. Fuster/Bardel 2024: 1–5) and is thus not exclusive to language teaching as a craft. More specific to the educational domain is pedagogical translanguaging, which refers to “intentional instructional strategies that integrate two or more languages and aim at the development of the multilingual repertoire as well as metalinguistic and language awareness” (Cenoz/Gorter 2020: 300). A related approach, relevant for the present study, is XLP (cf. Ballinger/Lyster/Sterzuk/Genesee 2017; Ballinger et al. 2020). It is defined as a “pedagogy that supports bridges between languages studied or known by learners, and its impact on learning and teaching” (Ballinger et al. 2020: 265), thereby adding emphasis to the role of formal teaching. Therefore, XLP is applied to the specific instructional context of L3 teaching in Sweden and is used in this article to discuss GL3 teachers’ deliberate inclusion of crosslinguistic references to L2 English when teaching GL3, thereby shedding further light on the notion of English as a bridging language in L3 teaching.

3.1 Participants

The four teachers in this study, henceforth denominated Teachers A-D, were recruited using a purposive non-random convenience sampling (cf. Mackey/Gass 2016: 175). Two inclusion criteria were applied, namely an official national certification for teaching German as a modern language (i.e. GL3) and at least five years’ experience of teaching it at the lower secondary level. Novice teachers were excluded from the sample because experienced subject teachers are more familiar with the professional discourse and can more easily articulate their pedagogical knowledge and beliefs (cf. Gudmundsdottir/Shulman 1987: 61; Basturkmen 2012: 288). The participants, who all work at urban lower secondary schools in the Stockholm area, were contacted by the researcher and accepted the invitation to share their pedagogical thoughts and practices. Some relevant background variables are shown in Table 2:

As can be seen in Table 2, Teacher B is a native speaker of German, unlike the other participants. However, this is not estimated to have any significant influence on the results compared to other individual differences stemming from the professional life stories of the four participants. Most of Teacher B’s long professional experience is gained within the Swedish school system, and this experience is believed to play a decisive role in shaping a teacher’s thinking and practices (cf. Borg 2019: 1153–1155).

3.2 Instruments and procedures

The individual semi-structured qualitative interviews were the first data to be collected. They took place some weeks before the classroom observations. Interviews were chosen because they constitute “an interactive practice in which knowledge, meanings and narratives are jointly produced” (Vähäsanten/Saarinen 2012: 493) and allow for an “inter-change of views […] about a theme of mutual interest” (Brinkmann/Kvale 2015: 4). The procedure was based on an interview guide covering also other areas of GL3 teaching practices than those reported in the present article. Crosslinguistic pedagogy was thus not the main topic of the interviews, but soon emerged as a recurring theme that also appeared frequently in the observed lessons.

The interviews were conducted in Swedish, audio-recorded and transcribed by the researcher. The quotes in this article were later translated into English and written up orthographically using transcription labels (cf. Appendix) inspired by a simplified overview of the HIAT conventions in English (Schmidt 2008). The translations posed a challenge, especially when it comes to everyday and metaphorical language (cf. Yunus/Hartman/Lucassen/Barton/Russell/Altun/Sturgiss 2022: 4), since compromises had to be made regarding both faithfulness to the Swedish original and readability and accessibility of the final report (cf. Nikander 2008: 226). The translations of the quoted parts were therefore assessed by a native speaker of English with a solid knowledge of colloquial Swedish, with the aim of preserving the conversational style of the Swedish original (cf. Aronsson/Cederborg 1997: 85) without risking the intelligibility in English. An overview of question types and examples of answers are presented in Table 3.

The second data set consists of 12 lesson observations documented as ethnographic fieldnotes (Walford 2009: 117–120), mainly as bullet points, written down by the researcher from a position at the very back of the classroom. For ethical and practical reasons, no recording was done in the classroom, which poses challenges since it is impossible for the observer to catch everything that happens in the room and there is no recording of the events to review afterwards. However, since the lessons were typically no longer than 50 minutes and the researcher was familiar with the context, content and materials used, it was a manageable task to collect relevant qualitative data under these circumstances. Some photos were taken of the whiteboard and artefacts used and these were later added to the fieldnote documents and grouped with the corresponding event that was observed. This helped recognizing and remembering the context and was a useful supplement to the notes written down under time constraints by a single researcher. Added to the fieldnotes are also summaries of the teachers’ oral accounts and evaluations shared with the researcher immediately after each lesson (see Table 1).

The third data set consists of outlined plans for lesson cycles, submitted to the researcher in the time frame (2-4 weeks) between the interview and the first observed lesson of each participant, as well as other teaching artefacts used during the lessons. The participants were given a material from a textbook chapter on morning routines, based on which they were asked to plan for a cycle of three consecutive lessons with their year 7 groups. It was made explicit that the application of the V2 rule in adverbial-fronted main clauses was the grammatical focus of the unit and the teachers were free to design the lesson cycle in any fashion preferred. The degree of detail varied across the four submitted documents, but all teachers clearly indicated which activities they had outlined for each individual lesson. The other artefacts were documented by photos taken in the classrooms or mailed as a file to the researcher after the lessons.

3.3 Analysis

Qualitative content analysis (Graneheim/Lundman 2004; Mayring 2022) was employed because it is a pragmatic method to seek answers to the research questions (Downe-Wamboldt 1992: 320) with an openness to new themes that might emerge during the coding process (White/Marsh 2006: 39). The interview data were the first to be analyzed. As units of analysis, the answers to each question (including follow-up questions) from the interview guide were used, since they would be large enough to provide a coherent context for the identified meaning units³ (cf. Graneheim/Lundman 2004: 106). Hence, passages from all four participants’ transcripts were aggregated into one text per interview question, in which meaning units were identified and condensed, i.e. shortened in a way that preserves the core of the content and context (ibid.). The meaning units were labeled with one or several codes and the codes were later grouped in clusters and categories. The codes relevant for the two research questions are presented in Table 4. As can be seen in the table, the analyzed content could be assigned to two overarching domains: 1) pedagogical beliefs about the influence of English in GL3, and 2) pedagogical knowledge (stemming from experience) about the role crosslinguistic links between L1, L2 and L3 might play in GL3 teaching.

With the categories and codes obtained from the interview data as an analytical starting point, the content analysis was extended to the second data set. The fieldnotes with bullet points from the lessons (each considered a unit of analysis) were first processed into a coherent running text, chronologically describing the unfolding events from the observed lessons (i.e. an interpretation by the researcher). Meaning units were then identified, labeled and categorized across these interpretive texts and the summaries of the post-observational reflections. The appropriateness of the coding was finally checked against the bullet points to ensure that the codes assigned corresponded to what was originally documented during the observed lessons. An example of an analyzed passage from the second data set is presented in Table 5:

Finally, the third data set was screened to determine whether there were contents in the lesson plans and other collected artefacts that were related to the categories and codes obtained from the content analysis of the previous two sets and that could, hence, provide a better understanding of the phenomena addressed by the research questions. The face validity of the categorization of the obtained codes was also checked in a deliberation with a teacher and researcher colleague who suggested some clarifications and adjustments. In addition, a member-checking procedure was employed to assess the global trustworthiness of the analysis by sending an early version of this text to the participants themselves and ask them to point out possible misinterpretations or other errors.

4.1 Pedagogical views on English as a bridging language in German L3 teaching

Overall, the teachers in this study do not display negative attitudes towards the use of other languages in teaching situations, and several of them express positive beliefs about multilingualism. As an example, the positive attitude of Teacher C, who even mentions translanguaging explicitly in the interview, can clearly be sensed in the following excerpt:

And we are living in an amazing time. I mean, [the students] will probably learn more languages. The alternative is everyone speaking English only. (Teacher C, interview)

She concludes that language learners should be encouraged to use all linguistic resources available, not the least in classrooms, and shares positive memories of previous students who had other mother tongues. Consequently, she also has positive attitudes towards using English as a bridging language and does not think that other teachers of GL3 nowadays have negative viewpoints in this regard. The following interview passage illustrates this well:

Interviewer: And do you think that our German teacher colleagues would have a different opinion on this topic [the English influence on the students’ German]?

Teacher: No way. I doubt that. I think everybody understands what is going on here. Also, we do know that translanguaging, that, there is nothing negative about mentioning: I am, you are, he/she/it is [when teaching German grammar]. We do have _some_ verb conjugations in English, but German has _even more_ forms.

Interviewer: I mean, that is exactly what some teachers apparently have, at least traditionally, been opposing. That you should kind of not even mention English when teaching German.

Teacher: Oh no, I do that all the time. I mean, I just think that is translanguaging. We have to use all the linguistic resources we have. […] Even all languages. I think it is super important to compare languages, all the time. (Teacher C, interview)

None of the four teachers talk about the English influence on their students’ German as something negative. They conclude that transfer from English generally does not hinder the intelligibility of the students’ utterances, although they state the importance of making the students aware that some transferred sentences might be regarded as errors in the target language. On this topic, Teacher A makes a comparison to non-native-like Swedish, where an utterance is perfectly comprehensible although it contains grammatical errors that to some extent might come across as a bit annoying to an interlocutor.

When it comes to transfer and crosslinguistic influence, Teacher B reports having observed several English traces in her students’ German. Examples brought up are the transfer of –ing forms and verb constructions with an auxiliary verb, such as *sie ist schwimmen. To her knowledge, none of her students have another L1 than Swedish. Therefore, Teacher B concludes that these constructions are probably influenced by the L2, since “many students have so much contact with English”. Teacher C also reports the general observation among her colleagues that the influence of English is so strong that some students nowadays write better compositions in English than in their L1 Swedish.

Regarding the non-application of the V2 rule in German, several participants explicitly refer to it as “the English word order”. In the interviews, they report noticing how their students produce V3 constructions even in fairly basic sentences in German and they all somehow link this phenomenon to the influence of English. Teacher A’s answers below constitute a good example:

((smiling)) Oh yes, it is this, a bit of this English, English word order. That they… /Ah, you know what I mean. (Teacher A, interview)

After a few turns of clarification, it becomes clear that Teacher A is referring to sentences where the V2 rule is violated. When asked when she usually notices this error type in her students’ production, the answer comprises another reference to English:

((3 sec)) It depends. And it is also, well, it differs between groups. And it depends on how much English they get influence from and, like… /But generally, I guess it is in grade 7. (Teacher A, interview)

When hearing the ungrammatical example “und dann ich spiele Fußball” from the interview guide, she finally exclaims:

Yes, that’s it. _Exactly_ what I meant. _Ex-act-ly_ that word order. It is very common. (Teacher A, interview)

The other three teachers also report coming across this ungrammatical word order in their students’ German. Some of them express a surprise that the students transfer the English verb order rather than the Swedish one. On one occasion, Teacher D calls this quite “peculiar”. In conclusion, all four teachers clearly report and display pedagogical beliefs that the influence from the background L2 English and its word order is the most salient factor that causes ungrammatical V3 sentences in the students’ production in L3 German.

Some of the teachers’ pedagogical views on crosslinguistic influences are also reflected in comments made while they are making explicit references to English. A rich example is found in Teacher C’s introduction of the lesson cycle, where she caters for a thorough discussion on the role of English L2 influence. The students are first asked the rhetorical question which language they hear the most in their spare time. This is followed by a point made by the teacher that the students cannot remain unaffected by all the English surrounding them in their everyday lives. During the second lesson, when she is about to introduce verb placement as the main content of that day’s classroom work, Teacher C contextualizes the content by telling the students that it is important to pay attention to verb order when narrating consecutive events in German, “so that one does not get caught in the English trap”. A similar reference is made by Teacher A when she points out an alleged English influence on an ungrammatical non-V2 sentence in German, which she concludes with a personal comment: “I tend to opine, that you should be more influenced by Swedish here [interpretation: not by English]” (Teacher A, lesson 3).

According to the teachers, they generally do bring up and make comments on the English influence they believe is related to the breaking of the V2 rule in German. For example, Teacher D reports that he usually tells his students: “do you understand, look here, it [the word order] is exactly the same as it is in Swedish, it is _not_ as it is in English”. In his experience, students sometimes state that the ungrammatical V3 order in German “just sounds better in their minds” and that they have a tendency to forget his explanations about the necessity to invert the verb. Therefore, Teacher D stresses the importance of constantly reminding the students about the crosslinguistic differences when it comes to word order.

4.2 Applications of crosslinguistic pedagogy in German L3 teaching practices

In the lesson plans submitted prior to the lesson cycle (see 3.2), only Teacher C explicitly defined an activity built on crosslinguistic comparisons with English, which was to be introduced in the first lesson within a context of defining the clause elements. As Teacher C carefully explained in the submitted document, the overarching aim of this part of the lesson cycle was to activate the students’ metalinguistic understanding of the verb and its placement. Figure 1 shows the key to the activity, which consisted of a short text about a first-person narrator’s morning written in Swedish, English and German:

Figure 1: Teacher C’s text in Swedish, English and German with some clause elements color-coded

As the activity was introduced according to the plan, the students’ first task was to analyze the clause elements in the three versions of the text by coding them with markers in different colors. They were also instructed to discuss if they can spot any difference in the color sequences and put words to the patterns. Once the activity was completed and reviewed, Teacher C finished the exercise by asking the students to summarize what happened to the adverbial-fronted sentences in the English version of the text. A student was quick to provide the correct answer: “the verb came in the third position”. This highlighting and drawing the students’ attention to the English sentences with V3 order created a clear contrast to the German sentences with V2 order and related in a clear way to explanations given previously during the same lesson.

In the second lesson, Teacher C revisited the crosslinguistic differences in verb order in adverbial-fronted sentences as an opening activity. Drawing on an example sentence from the previous lesson, she first wrote the sentence in Swedish and asked the students for a German translation. Then she asked the class to provide an English translation as well. As she wrote the English sentence on the board, she asked the students to pay attention to what was happening with the sentence. Referring to the teacher’s explanation from the previous lesson, a student provided a quick and adequate account of the V2 rule and the fact that English is not a V2 language, whereby the verb in this case can come in the third position. Teacher C praised the student for the explanation, but suggested a crosslinguistic clause analysis on the whiteboard to verify this. The ensuing analysis was visualized with circles around the verb in all three versions of the sentence and an added number “2” or “3” next to the verb, as shown in Figure 2:

Figure 2: Trilingual arrangement of sentences with highlighted verb elements

Another crosslinguistic pedagogical take was observed during the second lesson with Teacher C when she showed a Youtube video for learners of German in which the content creator⁵ mixes German and English. Teacher C commented the fact that the Youtuber is mixing the languages “wildly” in a seemingly positive fashion, confirming thereby her own positive beliefs about a multilingual approach to language teaching (see 4.1). Also in the third lesson, Teacher C made a crosslinguistic reference to English when she concluded one of the aims of that day’s lesson: “to retell several events and becoming aware of the fact that the word order in German is similar to the Swedish one, _not_ the English one”. Similarly, the “not” was strongly stressed by Teacher A (lesson 3) when she explained that, unlike Swedish and German, English is not considered a V2 language that requires the verb inversion.

In a similar fashion, Teacher D introduced an adverbial-fronted German sentence during his first lesson and stressed that the word order would have been different in English. Teacher D repeated this crosslinguistic reference several times over the course of the lesson cycle. On those occasions, he made sure to draw the students’ attention to the fact that the verb element can occupy the third clause position in English but that this does not hold true for German or Swedish. Teacher D also made a spontaneous crosslinguistic reference during a recapitulation activity (lesson 3), in which he prompted the students with four adverbial-fronted sentences in German, one of which was written with the verb in the third position. He clarified that the word order is the same in German and Swedish, but that English differs in this kind of sentence. These careful reminders were frequently stressed by Teacher D in the interview (see 4.1), suggesting an alignment of his stated pedagogical beliefs and observed teaching practices when it comes to these crosslinguistic features.

Furthermore, Teacher B was observed making recurring crosslinguistic references to English when explaining syntax. During lesson 2, she introduced a teaching activity by turning to the students with the words: “and sometimes I talk about the English word order, right, do you remember?”, to which some students nodded affirmatively. She then wrote three sentences, two in English and one in German, on the board (see Figure 3) and commented on them carefully by referring to the constituent clause elements:

Figure 3: Teacher B’s example sentences in English and German

Once the German sentence was added, she commented on the order of the clause elements and made a point about the difference between the second English and the German sentence. As the two sentences were now clearly contrasted to each other, Teacher B pointed to the board and summarized her point by saying:

So, this is now our difficulty. We are so used to… /And sometimes you think in English, and sometimes you get the wrong word order. But actually, it is very easy. (Teacher B, lesson 2)

During the same lesson, Teacher B provided another activity in which the students had to tell apart sentences in German with correct word order from those written with an ungrammatical V3 order. As she drew arrows indicating necessary verb inversions (see example in Figure 4), she commented: “and here you can clearly _feel_ the English”. At the wrapping up, she also called upon the students to double-check whether some German sentences they had produced had taken an English detour. Teacher B finally concluded the activity by stating: “And the best help for you here is: translate each sentence into Swedish”.

Figure 4: Arrows pointing out the need for inversion in an incorrect V3 sentence