“Unacceptable”, “exclusionary” or even “an affront”, the 2008 Miss Belgium pageant didn’t go unnoticed. When several¹ Dutch-medium newspapers questioned the linguistic competences of the laureate after the glamourous crowning ceremony in December 2007, readers’ reactions were not mild: since the French-speaking winner failed to answer questions in Dutch, she could not possibly be an ambassadress of all Belgians.² The French-medium newspaper La Libre Belgique responded to this commotion with an article entitled Miss Belgique 2008 n’est pas tweetalig³ (‘Miss Belgium 2008 is not tweetalig’), with tweetalig (‘bilingual’) in Dutch as derogatory reference to the supposedly exaggerated Flemish language sensitivities. In a list of personal strengths, such as spontaneity and musicality, La Libre further explicitly praised the multilingual repertoire of the newly elected Miss, who in addition to French also had command of Czech and English.

As we will see in this paper, the anecdote captures essential elements of ruling media discourses on the notion of ‘multilingualism’ in the Belgian Dutch- and French-medium press. In the country’s linguistically sensitive context, it is not surprising that language and multilingualism at times cause such heated, emotional and polemical discussions. In the current phase of intense globalisation, after all, different forms of multilingualism become increasingly visible and once again challenge ideals of strict territorial monolingualism by which complex historical tensions between mainly Dutch- and French-speakers are still managed today.

Given this complex, sensitive and emotional context, we therefore consider it meaningful to objectify the ideological thinking and versatility of competing perspectives on multilingualism in Belgian language debates. In order to explore how current valorisations of multilingualism in Belgium interact with perceptions of language typical of the Belgian conflict, this paper aims to map ruling media discourses on the notion of ‘multilingualism’ in the Dutch- and French-medium written press. This way, we seek to explore the extent to which the representation of multilingualism in Belgian media connect to international trends (see i.a. Horner, 2011; Horner & Weber, 2012; Jaffe, 2011; Kelly-Holmes & Milani, 2011; Robichaud & De Schutter, 2012; Vetter, 2013). In doing so, we additionally try to expand existing research of this type on the Belgian/Flemish context (see e.g. Bollen & Baten, 2010; Calis, 2015; Prophète, 2020) with a larger and broader data corpus.

Based on an analysis of 1710 news articles from the pivotal period 1998–2018, we will (1) present a quantitative and qualitative overview of circulating discursive focal points and language orientations regarding multilingualism, and (2) contextualise how these are interwoven with larger (inter)national historical, political, cultural and socio-economic tendencies. Before addressing these objectives, we sketch the role language and media play in the Belgian conflict and explain the compilation and analysis of our news corpus.

Throughout Belgian history, socio-economic, demographic and political developments polarised its two main language communities (e.g. Deneckere et al., 2012; Wils, 2005; Witte et al., 2016). In the late 19th and 20th centuries, tensions between Dutch- and French-speakers have gradually been canalised by, inter alia, adopting principles of territorial monolingualism, politically defining a ‘language border’ and, as such, gradually establishing a federal state structure. Belgium nowadays consists of three linguistic Communities (the Dutch-speaking Vlaamse Gemeenschap in the North; the French-speaking Communauté française de Belgique in the South; the German-speaking Deutschsprachige Gemeinschaft in the East) and three economic Regions (the assumed monolingual Dutch-speaking Vlaams Gewest in the North; the theoretically French-Dutch bilingual Brussels Hoofdstedelijk Gewest/Région de Bruxelles-Capitale in the centre; the Région Wallonne in the South, which is officially monolingual French except for the German-speaking part). Twenty-seven ‘municipalities with facilities’ along the language border and in Brussels’ periphery constitute a touchy exception to these monolingual structures, since they have to offer municipal services in another official national language to members of constitutionally defined national linguistic minority groups. The rigid nature of these facilities and the polemics this generates in times of growing population diversity and urban expansion of Brussels are emblematic of the constant balancing act between the different language communities.

Being based mainly on ideals of territorial monolingualism, these regulations may suggest that the Belgian conflict is in essence language-centred. Nevertheless, as in similar well-known contexts such as Quebec, Switzerland, Catalonia or the Balkans, language is hardly the only pivotal factor. It has rather become an outward, recognisable and easily polarisable marker of deeper political-ideological, socio-economic and ethnic frontiers (see e.g. Blommaert, 2011; Nelde, 1994). Therefore, language discussions have to be understood not only as discussions about language an sich, but also about the ideological role language has in society and about the underlying dynamics between speech communities at various societal levels (Darquennes, 2015).

Although the linguistic history of Belgium has been fairly turbulent, language conflicts seem to be generally dealt with at the discursive level (Darquennes, 2015; Rindler Schjerve, 2003). In their role as socio-cultural gatekeepers, the media have considerable importance in (re)producing such discursive language-ideological debates (Bell, 1991; Blommaert, 1999; Fairclough, 1989). Influenced by what is considered newsworthy, they create discursive spaces in terms of news and discussions, in which different arguments and viewpoints can be produced, reproduced and challenged in order to impact on public opinion and official policies (Johnson & Milani, 2010). This makes news and media debates analytical angles par excellence to study circulating discourses, argumentation and ideologies (e.g. Heller, 2009; Horner, 2011).

Moreover, media discourses can mirror and at the same time reinforce a country’s deeper societal tensions. In Belgium, the media landscape is language-divided, making the Dutch- and French-medium systems operate separately, each depending on their own societal context and market dynamics (De Bens & Raeymaeckers, 2010). A national, Belgian, approach to news is thus almost non-existent and information about the other language communities, including in Brussels, is scarce or otherwise often stereotyped (Potvin et al., 2004; Sinardet, 2007).

Nevertheless, news remains a delicate issue for analysis. Baker (2006, p. 73) argues: “newspaper discourses, where they exist, should not be taken out of context or always viewed as mainstream or hegemonic just because they have occurred within a newspaper”. Hence, we sought to establish a diverse and representative corpus based on different text types from various national and regional written media, that may reflect multiple sides of public opinion in different time frames, and in which it would be possible to find recurrent quantitative and qualitative patterns.

The self-compiled corpus (n = 1710) contains both online and printed media articles from the GoPress Academic archive⁶ and is split into a Dutch-medium (n = 1085; henceforth NL-press) and a French-medium subcorpus (n = 625; henceforth FR-press). In order to obtain generalisable yet manually manageable data, articles were selected in accordance with the following criteria:

This led to 1362 hits for the Dutch- and 879 for the French-medium subcorpus, giving a total corpus size of 2241 analysable keyword occurrences.⁹

Using TextSTAT (Hüning, 2015), concordance lists were created embedding the keywords in their broader textual context, hence “highlighting the evaluative polarity, be it good or bad, of the item’s collocational environment” (Mautner, 2016, p. 161; see also Baker, 2006). These word lines, together with close reading of the whole text, served as basis for further manual coding using thematic analysis (Braun & Clarke, 2006). Initial codes that meaningfully interrelated were merged into new, bigger, codes allowing us to identify recurrent quantitative and qualitative patterns concerning the following two aspects:

In emotional discussions it is useful to zoom out and map the discursive focal points at play: what are we actually talking about when talking about ‘multilingualism’? From a quantitative and comparative perspective, this section starts by describing the thematic conceptualisations (themes and topics) of the notion ‘multilingualism’, before giving an overview of its ideological valuations (language orientations). We end this section with a description and illustration of the eight recurrent argumentative patterns for evaluating multilingualism that were identified in our qualitative analysis.

What strikes in the abovementioned quantitative and qualitative patterns, is the fundamental versatility in perceptions, conceptualisations and valuations of ‘multilingualism’ (see e.g. Blommaert, 1999; Gal, 1998). The notion is ubiquitous, arises in diverse themes and topics, both big and trivial, and is valued or problematised with a variety of mutually contestable arguments. Yet this multiplicity is not random. Quite the contrary, as we will discuss in the present section. Positive or negative valuations of multilingualism once again prove to be socially, linguistically and geographically context-dependent, situationally conditioned and inextricably anchored in deeper (inter)national historical, political and socio-economic tendencies.

Remarkably, in the written press of a country built around strict structures of territorial monolingualism, the notion of multilingualism is nowadays primarily perceived as an asset. Zooming in on those positive visions of multilingualism in both corpora, however, they seem to take shape within largely two argumentative lines (see Robichaud & De Schutter, 2012 for comparable patterns): an economically motivated discourse on the one hand and a more classical rhetoric of diversity and plurality on the other, the former being more prevalent than the latter. As illustrated by the examples above, the economic and market-related approach to multilingualism is generally upward-pointing and conceives of language as human capital or commodity that can ensure greater efficiency and higher esteem on the job market (see Heller, 2003, 2010). In this sense, our data suggests that when multilingualism is referred to as an economic asset, we are dealing with what can be called ‘elite’ or ‘prestigious multilingualism’ (Barakos & Selleck, 2019; Jaspers, 2009). This concerns additive, well-commanded, individual forms of multilingualism in terms of internationally acclaimed and strictly demarcated standard languages that have little impact on personal and social life, so that this can continue in the region’s official (standard) language(s). In our dataset, i.a. English, Spanish, French and Chinese are explicitly mentioned in that sense.

A second argumentative line within the language-as-resource orientation in our corpora fosters linguistic diversity, language equality and cultural plurality. This is the case in articles that celebrate Brussels’ and Europe’s multiculturalism and diversity, in which both individual and societal multilingualism are represented as enhancing openness, broader worldviews, tolerance and mutual understanding between different (inter)national communities and minorities. Note, however, that Blommaert and Verschueren (1991) pointed to possible sloganism in these kinds of arguments, remarking that Europe ironically is the least linguistically diversified of all continents. Moreover, arguments on the promotion and valuation of linguistic diversity in our corpora are often based on the essentialisation of language with culture, cultural identity, worldview and equality. When such arguments are indeed sloganesque or superficially developed, overestimating the culturally broadening abilities of language an sich (one could refer to Example 5 above), they may have an opposite effect: in overdoing the premises of linguistic relativity, they actually accentuate the otherness of a cultural ‘other’ and, as a consequence, possibly reinforce perceptions of cultural difference (see also McWorther, 2014). A rhetoric of multilingualism as vehicle for cultural openness, furthermore, is also found in articles on rather prestigious projects of additive individual multilingualism. A recurrent argument (see Example 6) in favour of the CLIL programme in Dutch-medium education, for instance, is that it provides cultural enrichment and therefore better personal development.

As such, our data seem to support tendencies described by Ricento (2005) and Kubota (2011), namely that the orientation of multilingualism-as-resource shows an evolution from ideals of cosmopolitanism and cultural diversity towards a greater focus on economic utility and the role of prestigious standard languages. The argumentative trends within the multilingualism-as-resource orientation in both corpora, furthermore, resonate broader international tendencies and are anchored in specific national developments on the economic level. Regarding larger EU language policy, for instance, studies indicate that the proclaimed linguistic diversity of Europe is in fact often framed by neoliberal arguments emphasising primarily the importance of individual proficiency in national standard languages in a Knowledge Based Economy (e.g. Horner, 2011; Krzyżanowski & Wodak, 2010; Sayers & Lea Láncos, 2017; Vetter, 2013). In the FR-press, multilingualism is frequently associated with such economic (and touristic) topics and arguments, this way echoing the much-needed economic revitalisation of Wallonia.

The forms of multilingualism presented as problematic, then, seem to be mainly of societal kind (cf. Cenoz, 2013, pp. 5–6). In our corpus, this concerns the presence of more than one language in institutions and geographical areas where existing monolingual structures and language laws may be compromised: administration and schools in Flanders or Brussels and the linguistic landscape in ‘municipalities with facilities’ or the wider Flemish periphery around Brussels. In this historically linguistically sensitive context around the language border, the press mainly appears to pick up symbolic issues linked to the Belgian conflict between speakers of Dutch and French, such as multilingual road signs, election campaigns or language use in sports clubs. In addition, where positive arguments mainly referred to prestigious and additive multilingualism, negative arguments are also found in contexts of subtractive and plebeian multilingualism. This is the everyday language repertoire of multi-ethnic and poorly educated urban working class families, which are expected to learn the official community language and replace their mother tongues that are seen as less useful for emancipation and integration (Jaspers, 2009).

The fact that such arguments are more common in the NL-press shows how rooted they are in historically grown language-ideological sensitivities and discourses of minorisation and endangerment (see Duchêne & Heller, 2009 for an analysis of the prominence of such discourses in language discussions in general). Although the current socio-economic situation in Flanders is completely different, ideas grown throughout the 19th and 20th centuries regarding the societal status and emancipatory power of Dutch are prominent in our data up to today. In other words, and as Bollen & Baten (2010) also conclude for the media coverage of bilingual education projects in Flanders, the NL-press is predominantly positive towards multilingualism and applauds its prestigious individual forms as an asset, but where- and whenever the hard-fought position of Dutch in society (hence the nation state) is ‘at risk’, the orientations seem to invert. This tendency is reminiscent of what Heller et al. (2015) describe as the ideological tension between “language as skill” and “language as marker of belonging”, i.e. of nationhood and nationality: the fact that the promotion of (individual) multilingualism is accompanied by concerns for implications on the position and knowledge of the national/community language.

To summarise, our analyses mainly reveal three frames of multilingualism: that of prestigious individual multilingualism, that of societal multilingualism linked to the Belgian conflict and that of plebeian multilingualism linked to migration, integration and multiculturalism. Other forms of the multilingual reality, interestingly, remain largely out of the picture: hits on German, the third official national language (that is less controversial in the general Belgian conflict), hardly appear in our corpora. The notion of ‘multilingualism’ is also very scarcely linked to, from a Belgian point of view, ‘less canonical’ European standard or migrant languages (as compared to i.a. Arabic, English, French, Spanish, Turkish, …). The same applies to dialects, regiolects, Flemish tussentaal (lit. ‘in-between-language’, see e.g. De Caluwe et al., 2013) and other forms on the spectrum between dialects and standard language; all these forms, as such, constituting instances of so-called ‘hidden multilingualism’ (see Vogl, 2012; Tamburelli & Tosco, 2021). Hence, drawing on Blommaert et al. (2005), it is meaningful to interpret the notion of ‘multilingualism’ in our corpus not as something fixed, as something an individual or a society is said to have or lack, but rather as something an environment, with all its layered complexities, enables and disables. What Belgian Dutch- and French-language media discourses, rooted in their societal contexts, mainly enable is an understanding and valuation of ‘multilingualism’ as the economically useful and countable sum of demarcated prestigious standard languages, and not as someone’s repertoire of varieties, styles, registers and genres, as ‘linguistic competence’ in general.

Back to 2008, back to Miss Belgium. The ‘language fuss’ and conflicting reactions in the media after the election gala have shown to be exemplary for the ways the Belgian Dutch- and French-medium written press project their linguistic historiography on current multilingual practices: age-old Belgian linguistic frictions still seem to leave fresh traces.

At first sight, however, the quantitative and qualitative similarities between the Dutch- and French-medium press outnumber the differences. In similar themes and topics in both corpora, elite multilingualism – involving individual proficiency in high-status standard languages – is generally embraced and promoted as economic or cultural asset in a neoliberal and diverse world. The broad and consensual positive perception of multilingualism seems to go against strict monolingual thinking, which might then, one could argue, transcend or leave aside the historical focus on the ‘own’ community language in Belgian linguistic quarrels between Flemish and francophones.

Nevertheless, looking deeper, multilingualism appears to be appropriate particularly if it can strengthen the socio-economic position of the own speech community, internationally but also in the national power relationship, without threatening the position of and proficiency in the (standard variant of the) community language. The major focus on additive forms of individual multilingualism and the negative valuation of societal and plebeian multilingualism attest to this, as do the many symbolic topics taken up by the Dutch- and French-medium press. In such articles, ‘multilingualism’ regularly refers to the Dutch-French opposition. In the NL-press, this is accompanied by explicit protective reflexes towards the ‘hard-fought’ monolingual structures of Flanders, which is then on the French-speaking side recuperated in a discursively conscious, but at times also derogatory, way of being (stereo)typically Flemish.

As such, our data underline that language debates are always about more than only language and that contemporary discussions on multilingualism, in an intensely globalising context, cannot be separated from historical sensitivities characterising the complex Belgian ‘language’ conflict. The notion of ‘multilingualism’ in our corpora is versatile, ubiquitous and multilayered, a mirror of deeper societal tendencies. It emerges in a wide range of major and more or less anecdotal topics, in which we observe a dynamic of different arguments that propagate diverse context-dependent valuations and ideological perspectives, but that, in turn, are anchored in (inter)national socio-economic and politico-historical developments transcending language. On the one hand, this deep societal stratification explains the recurrent emotionality of Belgian language discussions. On the other hand, highlighting and contextualising this – at school through research and in the media – may at the same time provide more nuance and reflexivity in public debate, something, eventually, Miss Belgium 2008 might have hoped for.