1. The monolingual problem of CALL

Sauro (Reference Sauro2016) posits that computer-assisted language learning (CALL) has an “English problem”. I would extend this claim to say that it also has a monolingual problem that arises from a widespread monolingual bias in research and development of CALL (cf. Ortega, Reference Ortega2017: 293–294). In this context, I define monolingual bias as a deficit in considering the full range of pre-existing languages of a learner, of languages used in a setting, or of (plurilingual) skills in the design of and the research on CALL. In other words, monolingual in this context does not deny the presence or relevance of more than one language (a spurious claim in language learning contexts), but points out a deficit in taking other languages besides the target language into consideration, akin to the use of monolingual in Gogolin’s (Reference Gogolin1994) monolingual habitus.

The monolingual problem, in turn, is the consequence of this bias: A lack of knowledge about how CALL products and projects can support learners in using all their linguistic resources to achieve language-learning- and language-using-related goals, and a lack of CALL products and projects that realize this potential, or that support specific plurilingual skill development (cf. Council of Europe, 2018). In other words, I do not wish to imply that CALL, with its focus on (foreign, second, or heritage) language learning, is ever “monolingual” in the sense of being limited to the study of monolingual individuals in monolingual settings. Instead, I use the term “monolingual” to refer to a perspective limited to only one language (i.e. the target language) at a time.Footnote ¹

John Gillespie, in the Reference Gillespie2018 installment of his traditional EUROCALL talk “CALL Research: Where Are We Now?” (Part III), states that CALL “is not just about grammar and basic skills” but “about developing the knowledge and understanding of language within human multilingual culture” (Gillespie, Reference Gillespie2018). At the same time, in his lists of most published, less published, absent, or almost absent topics in CALL research, multilingualism, multilingual practices, or multilingual goals do not even appear. This is also reflected in the structure of key professional organizations within CALL. EUROCALL, for example, has seven special interest groups, and CALICO eight, none of them focusing explicitly on multilingual aspects of CALL. Levy, Hubbard, Stockwell and Colpaert (Reference Levy, Hubbard, Stockwell and Colpaert2015), in their analysis of CALL Research Conference themes mention “video annotation” (a potentially multilingual practice) and “translation”, but merely as examples of “Dedicated tools” within the “Technology” category, a category that was allocated only very low priority by CALL experts in a vote at the end of the conference (Levy et al., Reference Levy, Hubbard, Stockwell and Colpaert2015: 2). Ortega (Reference Ortega2017) reported that she “did not uncover a single published CALL article with ‘multilingual(ism)’ or ‘bilingual(ism)’ in the title” (294). Key contemporary handbooks on CALL (Chapelle & Sauro, Reference Chapelle and Sauro2017; Farr & Murray, Reference Farr and Murray2016; Thomas, Reinders & Warschauer, Reference Thomas, Reinders and Warschauer2013) include no relevant key terms in their tables of contents and only three in their indices.Footnote ² That this monolingual bias is not a new phenomenon is hinted at by the fact that these concepts do not play any role – literally do not appear at all – in historical overviews of CALL (Bax, Reference Bax2003; Davies, Otto & Rüschoff, Reference Davies, Otto, Rüschoff, Thomas, Reinders and Warschauer2013; Hanson-Smith, Reference Hanson-Smith2003; Reiser, Reference Reiser2001a, Reference Reiser2001bFootnote ³ ).

2. Monolingual language teaching and its critique

CALL does not exist in a vacuum. It is impacted both by technological developments as well as by developments within didactics and linguistics. As CALL derives much of its underlying theory from related disciplines such as (instructed) SLA (e.g. Hubbard Reference Hubbard2008; see also Chapelle, Reference Chapelle1997), it has also inherited some of the biases of these disciplines (cf. Cook, Reference Cook1997; Kachru, Reference Kachru1994), which in turn are impacted by the societal and educational context (cf. Gogolin’s, Reference Gogolin1994, monolingual habitus). The purpose of this section is to show how, over the decades, researchers and practitioners have chipped away at a monolingual bias in language instruction more generally. It attempts to show how some concepts and critiques have developed over time, without attempting to give a complete historical overview.

Historically, language teaching was not always conceptualized as monolingual (cf. Howatt & Widdowson, Reference Howatt and Widdowson2004: 18 ff.). The idea that language instruction should be monolingual, and indeed that a language teacher does not need to be able to speak any language except the target language, is closely connected to the “direct method” (Howatt & Widdowson, Reference Howatt and Widdowson2004: 210 ff.) and still influences communicative language teaching, task-based language teaching, and content and language integrated learning instruction today, from the design of textbooks to a widely reported preference for native speaker teachers. The monolingual bias is so ingrained that it often becomes invisible. As Cook observes: “Recent methods do not so much forbid the L1 as ignore its existence altogether. … the only times that the L1 is mentioned is when advice is given on how to minimize its use” (Cook, Reference Cook2001: 404). Of course, a monolingual bias should not be equated with a monolingual classroom. Several studies show that even when teachers aim at a monolingual classroom, non-monolingual practices find their way into their teaching practice (Gierlinger, Reference Gierlinger2015). These are seen, though, as deviations from the ideal, as intrusions into optimal language teaching, which is viewed as monolingual.

In language learning contexts, three lines of critique can be used to challenge the monolingual bias. First, there are negative arguments, or, in other words, a critique of the arguments brought forward in defense of a monolingual bias (e.g. Cook, Reference Cook2001). There are positive arguments, which focus on how a non-monolingually biased approach helps to achieve traditional language learning outcomes (e.g. Butzkamm, Reference Butzkamm1973; Cummins, Reference Cummins, Cummins and Hornberger2008). And third, positive arguments that point out how a non-monolingual approach helps to achieve plurilingual language learning outcomes. In this section, I will focus on the second and third category.

An early critique of the monolingual bias was published by Butzkamm (Reference Butzkamm1973), who criticized the dominant monolingual perspective in foreign language teaching at secondary schools in Germany in the late 60s/early 70s. Butzkamm, drawing on the philosophy of language, psycholinguistics, and practical experience, suggests that a dogmatic monolingualism should be replaced by an enlightened monolingualism that refutes “monolingualism as general teaching principle, as a rule with exceptions only in the case of emergencies” and instead includes specific bilingual learning and teaching approaches (Butzkamm, Reference Butzkamm1973: 175; translation by the author). In other words, instead of aiming at a monolingual classroom and failing, it would be better to design teaching that includes limited amounts of non-target language use. In the English-speaking world, one of the key critiques against a monolingual perspective stems from Cummins (Reference Cummins, Cummins and Hornberger2008). Cummins, focusing on second language learning, especially in bilingual/immersion programs in North America, pointed out that monolingual approaches ignore prior knowledge and interdependency of proficiency across languages (Cummins, Reference Cummins, Cummins and Hornberger2008: 67).

When Butzkamm originally worked on the concept of enlightened monolingualism, he focused exclusively on the role of German as the language of education in Germany. Today, critiques of the monolingual bias incorporate the fact that significant numbers of learners in many regions already possess plurilingual competencies when starting foreign language instruction in their first foreign language and that learners are not limited to studying only one foreign language. The most powerful critiques and concepts, therefore, do not just take into account the role of the first language (L1) or of the dominant school language, but the whole linguistic repertoire a learner brings into the classroom or to the computer. As Hammarberg (Reference Hammarberg2009) points out, “The fundamental theoretical aspect of the study of L3 competence, use and acquisition is the insight that humans are potentially multilingual by nature and that multilingualism is the normal state of linguistic competence” (Hammarberg, Reference Hammarberg2009: 2; italics in the original). The move toward a perspective that both acknowledges multilingualism as a frequent and normal condition of human existence, and that views linguistic heterogeneity as an element of diversity that deserves inclusion in the classroom is referred to as “multilingual turn” (Conteh & Meier, Reference Conteh and Meier2014: 1 ff.; cf. also May, Reference May2014).

Work on bilingualism, historically marked by a deficit perspective, has long turned toward identifying relative advantages of (often: early) bilinguals for further language learning. These studies show that regarding aspects such as language awareness or executive control, individuals with some types of linguistic biographies score higher than students with other kinds of linguistic biographies (e.g. Bialystok, Reference Bialystok2009; Jessner, Reference Jessner2006). Focusing only on early bilingualism as a source of linguistic resources would not result in a full picture, though. A range of studies also demonstrated that any previously learned or acquired languages – including contexts beyond early bilingualism – are connected with benefits for further language learning (e.g. Odlin & Jarvis, Reference Odlin and Jarvis2004; Singleton & Aronin, Reference Singleton and Aronin2007). In addition, many studies have looked at what conditions have to be met to allow learners to profit from pre-existing linguistic resources (Marx, Reference Marx, Gibson, Hufeisen and Personne2008; Singleton & Aronin, Reference Singleton and Aronin2007).

Another vector for a critique of the monolingual bias is to look critically at the assumed purpose of language learning. Individual learners differ in their reasons to study a language. On the institutional and educational policy levels, too, the purpose of language teaching and what constitutes a successful language learner are subject to change. In the history of formal language teaching, changes in teaching method are often associated with changes in the goal of language learning.

In 1996, the New London Group pointed out that literacy today needs to reflect changes in the use of technology as well as enable learners to navigate linguistic plurality, including “the code switching often to be found within a text among different languages, dialects, or registers” (New London Group, 1996: 69). The Common European Framework of Reference for Languages (CEFR) companion volume (Council of Europe, 2018) as well makes explicit that there are specific plurilingual competencies that go beyond a “two solitudes” (Cummins, Reference Cummins, Cummins and Hornberger2008) understanding of language(s): plurilingual comprehension (Council of Europe, 2018), communication building on a plurilingual repertoire (Council of Europe, 2018: 161f.), and some forms of mediation (Council of Europe, 2018: 103 ff.). It acknowledges the fact that some linguistic practices are not possible without at least two sets of linguistic resources interacting: from mediating between two friends without a shared language, to code-switching while telling a joke (also cf. Candelier et al., Reference Candelier, Camilleri-Grima, Castellotti, de Pietro, Lőrincz, Meißner, Noguerol and Schröder-Sura2012). In North America, where the CEFR is not as frequently used, practice based on concepts of translanguaging is comparable in perspective (Canagarajah, Reference Canagarajah2013; García & Kleyn, Reference García and Kleyn2016).

These critiques, discussed for more than half a century, did not remain theoretical considerations only, but directly and indirectly led to a wealth of concepts and materials (e.g. within the Mehrsprachigkeitsdidaktik discourse: Hufeisen, Reference Hufeisen2011; Meissner & Reinfried, Reference Meissner and Reinfried1998; within the L’éveil aux langues discourse: Candelier, Reference Candelier2003; within the tertiary language learning discourse: Cenoz, Hufeisen & Jessner, Reference Cenoz, Hufeisen and Jessner2001; Hammarberg, Reference Hammarberg2009; Hufeisen & Lindemann, Reference Hufeisen and Lindemann1998; Hufeisen & Neuner, Reference Hufeisen and Neuner2004; Jessner, Reference Jessner2006; within intercomprehension discourses: Doyé, Reference Doyé2005; Klein, Reference Klein2004; or within translanguaging discourses: Canagarajah, Reference Canagarajah2013; García & Kleyn, Reference García and Kleyn2016).

Some trends and theoretical developments arrive in CALL later than in the rest of foreign language pedagogy, (instructed) SLA, or related disciplines. Egbert and Hanson-Smith (Reference Egbert and Hanson-Smith1999) describe, in early CALL history, a “return in CALL to an earlier pedagogical model” (22), and how CALL “has, interestingly, replicated this 50-year development [within TESOL] in a foreshortened or accelerated manner, retracing the entire pedagogical history of TESOL methods in only about 12 years” (26). Non-adoption of specific theoretical trends might, at specific times in the history of CALL, have been due to “the limitation of early technology; another reason was that computer programmers were not particularly knowledgeable about how language learning worked” (22). The limitations of the technology have lessened, and experts for language learning and acquisition dominate the CALL discourse, yet not all discourses from these related disciplines reach CALL discourses in a timely fashion. If we have inherited the monolingual bias from non-CALL, we can also inherit a critique of this bias. Yet, while the monolingual bias certainly impacts CALL (as argued in more detail below), the growing critique is only slowly reaching CALL (but cf. Ortega, Reference Ortega2017, for a key voice in transporting this critique into CALL).

3. Non-monolingual topics in CALL journals

In this section, I will investigate the posited monolingual bias by (a) quantifying the frequency of certain keywords within top CALL journals, and (b) quantifying articles with those keywords in their titles and subjecting them to a more detailed analysis of key ideas and themes contained therein.

3.2 Results and discussion

Table 1, Figure 1, and Table 2 summarize the findings.

Table 1. Percentage of articles containing specific target words

Figure 1. Overview of keywords per year

Table 2. Percentages of papers including one of the target words (per journal)

Table 1 and Figure 1 both show that the keyword bilingual* (also encompassing, e.g., bilinguals, bilingualism) is surprisingly frequent in this corpus and found in more than a quarter of all papers. This keyword, however, is used in many contexts where the focus is entirely on one source language and one target language, while the learner may still be viewed as monolingual-per-default, for example, bilingual dictionary. In fact, the expression bilingual dictionar* appears in 5.16% of articles in this corpus.

Figure 1 shows that, although there is variance in how many articles per year use these keywords, the proportion of keywords remains fairly similar over time, with bilingual and multilingual making up the majority, and only very small numbers of articles using any of the other terms.

According to Table 2, a reader of ReCALL can expect three to four articles per year that at least mention the words multilingual, plurilingual, third language, tertiary language, L3, translanguaging, or translingual. Naturally, this does not imply that these articles focus on any of these aspects – only that at least one of these words appears at least once.

To put these findings into perspective, all articles were viewed and those articles that contained at least one of the keywords in their title were identified. In the total corpus, only eight articles out of 1,047 (or 0.76%) contain any of the keywords in their title. Table 3 shows their distribution across journals and time, and Table 4 provides key facts and summaries.

Table 3. Number of papers including one of the target words in their title (per journal)

Table 4. Overview of papers containing relevant keywords in their titles

One notable pattern that emerges from Table 4 is the large number of case studies, with five out of eight papers being identified as case studies. Three studies look at formal learning contexts, such as school or university courses (Kulavuz-Onal & Vásquez, Reference Kulavuz-Onal and Vásquez2018; Priego & Liaw, Reference Priego and Liaw2017; van Deusen-Scholl, Reference Van Deusen-Scholl2018), with the rest discussing informal or non-formal social media practices (informal Facebook practices: Chen, Reference Chen2013; Schreiber, Reference Schreiber2015; non-formal blogging: Melo-Pfeifer, Reference Melo-Pfeifer2015; not applicable: Lew & Szarowska, Reference Lew and Szarowska2017; Ortega, Reference Ortega2017). Most studies look at adults, often university students (Chen, Reference Chen2013; Lew & Szarowska, Reference Lew and Szarowska2017; Schreiber, Reference Schreiber2015; van Deusen-Scholl, Reference Van Deusen-Scholl2018; non-university adults: Melo-Pfeifer, Reference Melo-Pfeifer2015), with only two studies looking at any school context (Kulavuz-Onal & Vásquez, Reference Kulavuz-Onal and Vásquez2018; Priego & Liaw, Reference Priego and Liaw2017), and no studies looking specifically at younger learners (primary age or below). Although some papers argue that what has been observed in informal contexts or with adults can and should inform teaching practice at schools (e.g. Schreiber, Reference Schreiber2015: 83 f.), the amount of multilingual CALL studies with a school focus published in these four journals is very limited. This is not to say, of course, that such studies do not exist, just that they are not published in these four journals (cf., e.g., Alvarez, Reference Alvarez, Buendgens-Kosten and Elsner2018; Brunsmeier & Kolb, Reference Brunsmeier, Kolb, Buendgens-Kosten and Elsner2018; Cutrim Schmid, Reference Cutrim Schmid, Buendgens-Kosten and Elsner2018; Dausend, Reference Dausend, Buendgens-Kosten and Elsner2018; Elsner & Buendgens-Kosten, Reference Elsner, Buendgens-Kosten, Buendgens-Kosten and Elsner2018; Kirsch & Bes Izuel, Reference Kirsch and Bes Izuel2019; Lohe, Reference Lohe2018).

Non-monolingualism is conceptualized or becomes visible in those studies in very different ways. The study by Melo-Pfeifer (Reference Melo-Pfeifer2015), for example, discusses blogging in the target language and interactions with other language learners blogging in the same target language. Although she does not discuss any “surface” non-monolingual practices (e.g. code-mixing, use of more than one language for negotiation of meaning, etc.), her analysis is based on the notion of plurilingual competence, understood here as “the interconnection of social and affective repertoires (attitudes, motivations, social representations on languages and the people), linguistic and communicative resources (such as the knowledge on different languages and semiotic resources) and cognitive and verbal abilities (such as the ability to connect diverse verbal data, interpret and relate it)” (Melo-Pfeifer, Reference Melo-Pfeifer2015: 220).

Other studies look at “surface” non-monolingual practices, especially examples of networked multilingualism (Androutsopoulos, Reference Androutsopoulos2015) on Facebook (Chen, Reference Chen2013; Kulavuz-Onal & Vásquez, Reference Kulavuz-Onal and Vásquez2018; Schreiber, Reference Schreiber2015). Networked multilingualism – that is, “multilingual practices that are shaped by two interrelated processes: being networked, i.e. digitally connected to other individuals and groups, and being in the network, i.e. embedded in the global digital mediascape of the web” (Androutsopoulos, Reference Androutsopoulos2015: 188) – is a well-established research area, although often the focus is on informal practices outside of language learning settings.

Two studies look at telecollaboration (Kulavuz-Onal & Vásquez, Reference Kulavuz-Onal and Vásquez2018; Priego & Liaw, Reference Priego and Liaw2017). Non-monolingual aspects are inherent in CALL telecollaboration/virtual exchange, be it in tandem settings (individuals/groups learning each other’s languages) or in lingua franca settings, but these aspects are not necessarily emphasized in research and practice. ^{Footnote 7} Kulavuz-Onal and Vásquez (Reference Kulavuz-Onal and Vásquez2018) analyse language policies and multilingual practices of students and teachers in a virtual exchange using Facebook, while the students in Priego and Liaw’s (Reference Priego and Liaw2017) case study used telecollaboration to create multilingual (= parallel text) digital stories, although the focus of the article is on the creation of the English-language version only.

Three articles cannot easily be grouped with other articles. Lew and Szarowska (Reference Lew and Szarowska2017) discuss a popular form of non-monolingual language learning resource: bilingual online dictionaries. In their discussion, they consider the role of the L1 in the design of bilingual online dictionaries (e.g. including Polish grammatical labels in English-Polish dictionaries). Van Deusen-Scholl (Reference Van Deusen-Scholl2018) discusses the relationship between choice of language course (videoconferencing-based courses of less frequently taught languages) and heritage language background of the learner. She argues for the need of a plurilingual turn in heritage language acquisition. Finally, Ortega (Reference Ortega2017) presents arguments and a research agenda for SLA-CALL research that takes multilingualism and social justice into consideration.

4. Non-monolingual design in CALL products and projects

When we talk about the monolingual problem in CALL, it does not suffice to talk about research contexts only. Another form this monolingual problem takes is related to a bias in CALL product development. Quantifying a monolingual bias in product development is more difficult than doing this for research, as there are no generally recognized repositories for CALL products and projects. A survey of CALL products and projects would ideally not only include whether more than one language is included in the product/project, but also how it is included and which function it serves. This paper does not attempt such a systematic survey, but merely presents some general observations regarding CALL products and projects.

Several popular CALL products today are bilingual: Duolingo is, as is Rosetta Stone and similar products. A myriad of (language) learning apps (e.g. Anki, Quizlet, Memrise) and drill-and-kill games support bilingual vocabulary memorization techniques. Indeed, these products use more than one language, but they do this by following a historic understanding of language learning (e.g. grammar-translation, behaviorism), rather than a contemporary understanding of language learning pedagogy. It is the author’s subjective impression that while products and projects with an explicit multilingual or plurilingual perspective exist (cf. Table 5 for examples), these are usually non-commercial products developed through public funding (often: EU/EC funding) to explicitly address an existing gap in product development. Plurilingual design does not happen accidentally. Plurilingual design, when it occurs, is intentional and often the key agenda of a project. Only rarely are, for example, affordances for multilingual practices included outside such, usually non-commercial, projects.Footnote ⁸

Table 5. Overview of CALL projects and products with an explicit multilingual or plurilingual perspective

It is important to note that looking at a CALL product as it “comes from the box” only tells half the story. An app, a piece of software, a media bundle with reading pen are used by learners, often cooperatively or collaboratively, leading to the potential for enrichment with additional languages and multilingual practices even for a monolingual product or project (Brunsmeier & Kolb, Reference Brunsmeier, Kolb, Buendgens-Kosten and Elsner2018; Cutrim Schmid, Reference Cutrim Schmid, Buendgens-Kosten and Elsner2018; cf. also Gee’s (Reference Gee and Salen2008: 24) distinction between “game” and “Game”). The monolingual bias may limit attention to these factors, though, and risks that the potential of this remains untapped by product developers and understudied by researchers.

5. Multilingual CALL

Above, I have argued that work on non-monolingual CALL may exist, but does not constitute a visible trend or a coherent approach within CALL. Perhaps one of the contributing factors to this relative invisibility is the fact that studies on, for example, multilingual computer-mediated communication (CMC), or subtitling, or bilingual dictionary use, are perceived primarily as studies on CMC, or film use, or dictionaries. What connects these studies – viewing the presence and use of more than one language as prerequisite for language learning, goal of language learning, or tool/resource in language learning – remains in the margins.

The CALL community, like many academic communities, has a penchant for puns and acronyms. On a certain level, it seems obvious to joke about the proliferation of –ALL acronyms like MALL, BALL, or ICALL. Yet these and other labels serve an essential function: They provide a focusing lens that allows us to view research on seemingly disparate areas as being related. One way to tackle the monolingual bias is to create a label that subsumes work that does not subscribe to this bias, or that actively counters this bias.

Just as Garret argues that CALL is “not shorthand for ‘the use of technology’ but designates a dynamic complex in which technology, theory, and pedagogy are inseparably interwoven” (Garrett, Reference Garrett2009: 719–720), it can be argued that multilingual CALL (MCALL) is more than the mere existence of more than one language in a CALL. Buendgens-Kosten and Elsner (Reference Buendgens-Kosten, Elsner, Buendgens-Kosten and Elsner2018a) tentatively define MCALL as

This definition is intentionally broad, encompassing a wide range of practices already part of standard language teaching and learning practices, such as subtitling or digital bilingual dictionaries. The intention of this term is not to single out the most cutting-edge (plurilingualism/translanguaging-inspired) work on MCALL, but to provide a focusing lens for “work that actively and consciously works on the interstices of multilingualism and CALL” (Buendgens-Kosten & Elsner, Reference Buendgens-Kosten, Elsner, Buendgens-Kosten and Elsner2018a: xvii).

The history of CALL is closely intertwined with the technological options available for realizing CALL. This fact can hold CALL back, but it can also inspire CALL. In the remainder of this section, I will try to suggest how we can use emergent technologies to change how we think not only about CALL but also about multilingual CALL.

Digital tools can function as an enabler in MCALL contexts by bringing diverse L1s into the classroom: Even if the teacher does not have a good command of all their students’ L1s, or if a speaker lacks literacy skills in (one of) their L1s, multimedia tools can support learning activities involving a wide range of languages (cf. Buendgens-Kosten & Elsner, Reference Buendgens-Kosten, Elsner, Buendgens-Kosten and Elsner2018a: xv). Today, specialized tools exist that include text in multiple language versions and allow learners to practice receptive code-switching (Bündgens-Kosten & Elsner, Reference Bündgens-Kosten and Elsner2014); for example, the MuViT storybook, bilingual story apps (e.g. Beelinguapp), or projects utilizing reading pens to add acoustic information in multiple languages to a picture book (Dube & Gürsoy, Reference Dube, Gürsoy, Kutzelmann and Massler2018). In all of these, the languages available have to be prepared ahead of time, either by the teachers themselves, or by the app producers (possibly with the help of translation software). Some teachers already utilize the fact that online platforms like Wikipedia, although not intended for language learning, provide articles on the same topics in multiple languages (Alvarez, Reference Alvarez, Buendgens-Kosten and Elsner2018). In the future, even more flexibility might be provided by on-the-fly addition of languages, for example, through tools such as Google Translate, which allow learners to photograph a segment of a book or worksheet and to be provided a translation of this into a language of their choice. The technology necessary to do so already exists and is moving into broad use (cf. auto-translations on social media platforms), and existing quality issues will most likely diminish over time.

Adaptivity, including learning analytics, might be another area that is promising for CALL and MCALL alike. Adaptivity should include adapting to all the linguistic resources a learner brings to the table. A rather basic example for this might be the EU·DO·IT platform (Buendgens-Kosten, Lohe & Elsner, Reference Buendgens-Kosten, Lohe and Elsner2019; http://eudoit.eu), where learners of a range of target languages can choose a support language, and the material then includes an increased number of cognates, loan words, internationalisms, as well as shared structures, between the target language and the support language. In the case of EU·DO·IT, though, this is still very basic, allowing only three support languages from three language families, with all translanguaging support hardcoded in the textual materials. The next logical step would be to include more support languages, to allow one learner to use several support languages, and react flexibly to what the learner finds and does not find useful for their learning.

A further potential avenue might be virtual reality. Virtual reality, like more traditional virtual worlds, is not bound to copy reality. Sometimes, though, this is the intention – for example, to help a learner of the local variety of French to prepare to visit a local coffee shop (see, e.g., Papin, Reference Papin, Taalas, Jalkanen, Bradley and Thouësny2018). In those instances, the same effort that is used to replicate the environment could be extended to the linguistic environment, by replicating – to a certain degree – the linguistic landscape and soundscape of the actual place. Ideally, even linguistic practices such as code-switching, code-choice, linguistic accommodation, etc. – always in the ways typical of a specific location – might be included. Within such sociolinguistics-informed virtual reality, learners indeed could develop their “knowledge and understanding of language within human multilingual culture” (Gillespie, Reference Gillespie2018).

CALL product development is essential, as it allows us to push the limits of our understanding of what is conceptually possible, and to empirically test our assumptions about how changes in design (e.g. making a design more multilingual) impact learning. This way, CALL is in the position to not only inherit a critique of monolingual approaches from its related disciplines, but also contribute to theory building in these areas through developing and studying tools and tasks using multilingual and multilingual-sensitive digital environments.