Social engineers and myth-busters: A comparative research on Lithuanian, Norwegian and Serbian language experts

This article compares the language ideologies of language experts (both academic and nonacademic) in online news media in Lithuania, Norway and Serbia. The results will reveal that language is understood in diametrically opposed ways amongst Lithuanian and Serbian academic experts on the one, and Norwegian academic experts on the other hand. Lithuanian and Serbian academic experts are influenced by modernist ideas of language as a single, homogenous entity, whose borders ideally match the borders of an ethnic group. Norwegian academic experts function in the public sphere as those who try to deconstruct the modernist notion of language by employing an understanding of language as a cognitive tool that performs communicative and other functions. On the other hand, non-academic experts in all the three countries exhibit a striking similarity in their language ideologies, as the great majority expresses modernist ideals of language.


Introduction
This study seeks to explore the much-neglected role of language experts in the shaping of language ideologies. Authority in language and the construction of "linguistic (in)competence" have been central questions in seminal sociolinguistic studies (Cameron * This paper is largely based on chapter 3.1. of my PhD dissertation "Language serves the people, not the other way around! Ideologies of language in Lithuania, Norway, Serbia and the Baltic, Scandinavian and Ex-Yugoslav regions. A typological perspective on metalinguistic media discourses and language policies" (Vilnius: Lietuvių kalbos institutas, 2019) defended at the Institute of Lithuanian language 31 May 2019. . Social engineers and myth-busters: A comparative research on Lithuanian, Norwegian andSerbian language experts. Taikomoji kalbotyra, 12: 292-344, www.taikomojikalbotyra.lt. 1995, Milroy &Milroy 1999), yet relatively few studies explore the ideologies of the principal voice of authority: the language expert. The discursive construction of a "language expert" is somewhat researched (Milani 2007, Reyes & Bonnin 2016, but the "expert ideologies", have not been of interest to any of the studies I found.
In the contemporary conditions where knowledge is produced and circulated at an increasingly faster pace, experts help lay people orient themselves in the making of practical and ideological choices.
The number of expertsas well as their, sometimes radically different, opinionscan be dazzling. This was pointed out as early as 1966 by Luckman and Berger, who illustrate how one is often in need of "experts on experts" (1991: 60) to make the simplest choice, such as which brand of medicine to buy. When it comes to language experts, their traditional, "modernist" role has been associated with the training in good and correct language use, to set linguistic norms and to preserve a monolingual ideology (Milroy 2001). This view of the language expert is challenged by the increased migration, globalisation, visibility of multilingual practices at workplaces and in the public sphere.
Ideologies of language produced by experts can be shaped by these traditional roles, as well as local histories and local ways of dealing with global trends. This research will take a comparative perspective in order to differentiate between local, regional and possibly universal tendencies. Hence, the discourses of Lithuanian, Norwegian and Serbian experts will be analysed. The three countries have different (geo)political histories, religions, economic success rates, etc., but all the three were created in the 19 th century by new elites that sought independence from larger state formations. Their linguistic histories are specific: the Lithuanian language is the main national symbol of the country thanks to the fascination of Indo-Europeanists with its supposedly archaic features. It was standardised with the basis in the Sudovian speech, which had a written tradition, but with significant purist interventions, such as the erasure of a great number of Slavic and German loanwords. The Norwegian linguistic situation is a textbook example of diglossia: dialects are used in spoken and standard language(s) in written form. There are also two standards, Bokmål and Nynorsk. . Social engineers and myth-busters: A comparative research on Lithuanian, Norwegian and Serbian language experts. Taikomoji kalbotyra, 12: 292-344, www.taikomojikalbotyra.lt. second part of the analysis analyses which beliefs of language go hand-in-hand and form ideological clusters of beliefs. Such ideologically connected beliefs will be considered a notion of language if they appear more than two times in the data. For the sake of convenience, the beliefs and notions identified will be described in most detail in the first country analysed, Lithuania (section 3.1.). Third, all the identified beliefs and notions of language will be compared quantitatively (refer to appendices 1 and 2 for a detailed presentation of the qualitative findings).
However, if all possible beliefs about different elements of language (from sounds and letters to expressions and ways of speaking) were analysed, a comparative research would prove impossible. I will further narrow down the analytical focus to only those aspects of language that the previous research has shown are subject to ideologisation. I have identified three aspects of this type (which will be considered the theoretical model in this research): I Representation. This aspect shows how the relationship between language/variety and the group which speaks that language/variety is idealised. The relationship is considered idealised if there is a process of essentialisation between the two phenomena; for example, (all) Lithuanians are linguistically represented by the Lithuanian language (variation, second language speakers, etc. are erased from the picture).
II Expertise. It is common to assume that some speakers are more competent in language than others according to certain criteria. "Expertise" is the ideologised understanding of these criteria, or what is considered "correct" or "good" language.
III Function, or what function of language is idealised. In most cases it is the communicative or the symbolic function.
The first aspect is historically conditioned by nation-building, during which a modern nationstate was imagined as a perfect match between a territory, language and a people (which will here be called ethnic representation). Even though it is now known that this goal is unreachable, it still remains one of the default ways of thinking about the language-peopleterritory relationship (cf. Gorham 2000, Bauman & Briggs 2003, Moschonas 2004, Spitzmüller 2007, Berthele 2008, Tardy 2009). Such a view of language representation has Vukotić, V. 2019. Social engineers and myth-busters: A comparative research on Lithuanian, Norwegian and Serbian language experts. Taikomoji kalbotyra, 12: 292-344, www.taikomojikalbotyra.lt.
preferable for comparative research, as the findings can be more easily contrasted to each other in-between countries. The shortcoming is the possible exclusion of a number of ideologised elements, which could play a role in the formation of ideologies.
Before moving on to the analysis, I will shortly review the previous research on expert ideologies, with special focus on Lithuania, Norway and Serbia.

Previous research
Three problems are of relevance to the research question at hand. What is the role of the language expert in the public sphere? What counts as a language expert? What ideologies have already been noticed in the discourses of Lithuanian, Norwegian and Serbian language experts?
The first question was addressed in Deborah Cameron"s seminal publication on prescriptivism from 1995. Her book was an intervention into the Anglophone linguistic tradition, which dismissed prescriptivism as an essentially non-linguistic, and therefore irrelevant, research object (1995: 4). Fishman (2006) also mentions that linguists have tended to refrain participation in public debates on language. Nowadays, it is reported that they do participate in the public sphere, but they are often misunderstood and misrepresented by the media (cf. Jaspers 2014). However, these observations may be generalizable to the societies of the former Western bloc but not to the former Eastern Bloc, where a change in the understanding of what a "language expert" is, occurred. In the early Soviet Union, many leading linguists competed to be the first to introduce the best version of a socially engaged "Marxist linguistics", which would oppose the disinterested bourgeois linguistics (Fedorova 2010). Some efforts had a real social dimension to their epistemology, similar to that of Labov"s sociolinguistics from the 1960s, but they did not develop further, as they were killed by Stalin. Stalin"s article on the nature of linguistics (1950) reintroduced purist/prescriptivist attitudes towards language ("speech culture") and gave it hegemonic status (Gorham 2010). In Yugoslavia, a part of the "Non-Aligned Movement" (1961-) of countries that did not align with either of the blocs in the Cold War, the situation was, again, different. Controversially, studies show that the leading language experts in this socialist country were inclined to . Social engineers and myth-busters: A comparative research on Lithuanian, Norwegian and Serbian language experts. Taikomoji kalbotyra, 12: 292-344, www.taikomojikalbotyra.lt. linguistic nationalism. It was explicit and public amongst Croatian linguists (with very prescriptivist and purist elements), which got them into trouble with the central Yugoslav government (Greenberg 2004, Kordić 2010. Linguists in other Yugoslav countries exhibited a much more latent version of nationalism: its influence was noticed in the academic work of dialectologists, as both Bosnian, Croatian and Serbian linguists engaged in drawing ethnic boarders through dialectological research (Greenberg 1996(Greenberg , 1998. It was also noticed in the practice of writing dictionaries, as some dictionaries of Serbian authors often downplayed Croatian varieties and elevated the Serbian ones (Bagdasarov 2018). During the 1980s, leading Serbian and Slovene linguists, too, started expressing their linguistic nationalism more and more openly (Tollefson 2002, Budding 2003. A second question concerns the nature of the expert. Are all experts the same? If no, how are their identities constructed? When studying media discourse, one must admit that the people presented as prominent experts in media debates would rarely be considered experts on the subject at hand in an academic context. Milani (2007) showed how a debate of key issue concerning immigrant language varieties in Sweden came to be dominated by a researcher on comparative literature, because the media provided her with a platform, shunning aside experts with much more academic competence in the field of second language learning, variation and immigration. Media has not only the power to choose who will be considered an "expert" and who a "layman" (cf. Johnson & Milani 2010); it has also historically had the role of constructing the subjects of the modern nation-state, including those who have authority to speak and those who do not (Heller 2011). As my research is focused on the media discourse, I am compelled to treat experts as those people that the media outlet decides to present as experts.
The third question concerns what we already know on language experts in Lithuania, Norway and Serbia.
Lithuanian linguists exhibit strong prescriptivist and nationalist attitudes towards language in their academic writing. The ideological baggage includes a disguise of language regulation as a "theory of language cultivation" (Keršytė 2016, Vaicekauskienė 2017, nationalisation of Vukotić, V. 2019. Social engineers and myth-busters: A comparative research on Lithuanian, Norwegian and Serbian language experts. Taikomoji kalbotyra, 12: 292-344, www.taikomojikalbotyra.lt. language values and a view of standard language as a tool of control of linguistic production in society (Raila & Subačius 2012) and in the educational system in particular (Vaicekauskienė, in print). Other traits include a discursive construction of threats of foreign languages, primarily English , the connection of linguistic correctness and purity to moral decency (Subačius 2016) and a dismissal of urban speech varieties as "bad" or "mixed" language (Čičirkaitė 2019).
Three studies mention ideologies of Norwegian experts. Ims (2014) has researched a TV show about language and found that a Norwegian teacher was constructed as a monolingual authority that is worried over linguistic competences of children with multilingual background, indirectly expressing the idea that their Norwegian has to be "correct" and "pure" (2014: 17). Svendsen and Marzo (2015) analysed media representations of the socalled "Kebab Norwegian", an urban Oslo multiethnolect. They found that academic experts in media expressed neutral or positive attitudes about the variety, while the overall media discourse around it was negative. The experts participating in the debate did not manage to establish their own understanding as dominant and ultimately contributed to the spread of the ideologically laden term "kebab-Norwegian" (Svendsen & Marzo 2015: 75). The multiethnolect is seen as a "homogenous variety" that has a territory (Oslo) and a homogenous group that speaks it (young immigrants), and as an obstacle to social mobilityideologically in line with a purist and nationalist understanding of language. One study was written about the debate on bilingualism, a newspaper-mediated debate between language researchers that took place in November-December 2013 (Saltnes 2016). This MA thesis shows that the debate was framed as "for and against bilingualism in school" (one group stressed the advantages, and the other the disadvantages of it), but in the end, it essentially was "good solutions and methods to help these children that all the researchers were ultimately looking for." (Saltnes 2016: 49). This debate falls in the criteria for data selection; hence, it will be analysed in 3.2.
As mentioned above, many Serbian linguists have expressed nationalist and monolingual ideologies, both period of state socialism and after it (Greenberg 2004). The prominent Serbian linguist Ranko Bugarski, who wrote extensively on linguistic nationalism in Serbia, . Social engineers and myth-busters: A comparative research on Lithuanian, Norwegian and Serbian language experts. Taikomoji kalbotyra, 12: 292-344, www.taikomojikalbotyra.lt. explains that Serbian linguistic nationalism, in its most extreme version, is based on the belief that the only "true" or "original" language in the Balkans is Serbian, while Bosnian, Croatian and Montenegrin are made-up languages of made-up nations (2018: 23). This belief is held by the group of linguists that Robert Greenberg labelled "Neo-Vukovites" (because of their inclination to accept only the authority of Vuk Karadţić, the primary standardiser of Serbian language), whose most prominent figures are the professors of the university of Belgrade Miloš Kovačević and Mihailo Šćepanović. Jovanović (2018) analysed the texts of the same type right-wing linguists and found that the additional elements to their ideology are the Cyrillic script and Orthodox Christianity. He finds that, in the understanding of those linguists, the Cyrillic script, nourished by the Orthodox church, is what makes the Serbian language unique from the neighbouring ones. These linguists then see the Latin script as a foreign one, "occupying" the domestic, Serbian, Cyrillic script.
In conclusion, Lithuanian experts express both prescriptivist, monolingual attitudes and negative views of mixing, while Serbian experts mostly express nationalist, monolingual ideologies. In Norway, the research indicated that a regular language teacher could hold negative attitudes towards "mixing" and multiethnolects (that dominate the media discourse), while academic linguists would see them in a neutral or positive light. It should also be noted that previous research indicates that the discourse on language in Lithuania and Serbia revolves around the majority language, while in Norway it revolves around minority languages and "new" varieties of Norwegian that emerge in contact with migration.
Nevertheless, the ideologies found in public discourse are comparable (see 3.4.).

Data
The data consists of news portal articles that are either written by or contain the voice of the language expert. Language experts are constructed in media discourse by crediting their knowledge as "professional or scientific" (Habermas 2006: 416), for example, by referring to their profession (teacher, university lecturer or researcher) or through affiliations (universities, linguistic academies or societies). As mentioned above, I will not check the credentials attributed to experts by the media, assuming most of the readers do not do so either (including, inevitably, a number of self-proclaimed experts). . Social engineers and myth-busters: A comparative research on Lithuanian, Norwegian and Serbian language experts. Taikomoji kalbotyra, 12: 292-344, www.taikomojikalbotyra.lt.
The articles were collected by searching the most popular news portals. I included all the genres: general news, opinion pieces, columns, interviews, chronicles and other texts. The time frame for the collection was set starting from the year 2008 to (and including) 2016.
This period was chosen because online media became widely used in these three countries in different time periods (in Norway, it happened already in the 1990s, and in Serbia, in the 2000s). 2008 was chosen as an approximate year when online media could definitely be considered a real alternative to print media in all the three countries; many newspapers had by then opened their own portals, where they started mirroring the content from the printed versions.
The main criterion for the selection of news portals was their popularity. I initially searched the eight most visited news portals (according to indexing service alexa.com) for articles including statements of language experts. The search was carried out through the websites" own search engine by using the keywords: language, the Lithuanian / Norwegian / Serbian language and linguist / philologist. The search terms in Lithuanian were "kalba", "lietuvių kalba", "kalbininkas / kalbininkė" and "filologas". The Norwegian keywords were "norsk språk", "språk", "språkviter / lingvist", "bokmål" and "nynorsk". In Serbian, the keywords were "jezik", "srpski jezik", "lingvista" and "filolog". When possible, the portals" own subject-tags were used.
As it turned out, only six Lithuanian, four Norwegian and four Serbian news portals contained data relevant for the research. They are, in Lithuania: "delfi.lt", "15min.lt", "bernardinai.lt", "lrt.lt", "lrytas.lt" and "lzinios.lt"; in Norway: "aftenposten.no", "dagbladet.no", "nrk.no" and "vg.no"; in Serbia: "politika.rs", "telegraf.rs", "blic.rs", "danas.rs" and "novosti.net". The rest of the news portals either had no relevant content, or there were articles containing statements of language experts, but without expressing beliefs about language (for example, a new edition of a grammar book is announced, or a short comment is given on a political topic). In some, relevant content was found, but it was unoriginal (fully copied from one of the above-mentioned portals), so such articles were excluded from the data. The portals excluded this way were "respublika.lt" and "tv3.lt" in . Social engineers and myth-busters: A comparative research on Lithuanian, Norwegian and Serbian language experts. Taikomoji kalbotyra, 12: 292-344, www.taikomojikalbotyra.lt.

Lithuanian experts
The discourse of Lithuanian experts is dominated by academics who are or have been involved in official language planning institutions (n=43). Another group of articles are the "alternative" academics who criticize the first, dominant group of academic experts (n=12).
Lastly, non-academic experts voiced their opinions in 10 articles.
The discourses of the first and third group are very similar. Both share a pre-supposition that when one speaks publicly about language, one is supposed to speak about a "single" Lithuanian language and the potential dangers or problems it faces. It belongs to the order of discourse that dictates that the Lithuanian language faces internal (incorrect use of language amongst speakers) and external (English and other languages) dangers (cf. Vaicekauskienė & Šepetys 2016). Minority languages and foreign languages are marginal subjects in the Lithuanian virtual sphere, unless they somehow affect the national (Lithuanian) language. For example, an article about foreign language learning in Lithuanian (LT-4) is entitled "M.
Ramonienė: Learning foreign languages will not harm the Lithuanian language"; or news piece about Polish as a minority language is presented through the lenses of how it will affect the Lithuanian language (LT-40). In other words, the "danger to Lithuanian language" is the criteria that makes a linguistic topic newsworthy. This is due to the fact that linguistic nationalism intensified in the period of the1990s, when it was also institutionalised (Vaicekauskienė & Šepetys 2018).
The first belief frequent in the first group is that language represents an ethnic group. It comes from the historically bound narratives of Lithuanian as "an ancient Indo-European language". I will illustrate this narrative with an article, in which the interviewee, Zigmas Zinkevičius, a professor of Baltistics and former (1996)(1997)(1998) Minister of Culture, says: witnessed the opposite: we were very sedentary for a thousand years, that is why our language Vukotić, V. 2019. Social engineers and myth-busters: A comparative research on Lithuanian, Norwegian and Serbian language experts. Taikomoji kalbotyra, 12: 292-344, www.taikomojikalbotyra.lt. keeps lots of the archaic features, which can be found only back in proto-Indo-European. (…) You have to agree, it is a great victory for a small nation, because of which the great languages of the world discover the Lithuanian languageone of the most archaic Indo-European languages.] (LT-8, italics 1 by me) The diachronic tradition, launched by the discovery of the conservative features of Lithuanian, is still a dominant paradigm in Lithuanian linguistics. What is of interest for this paper is the elevation of this "ancient" character of the language to the status of a national symbol. The words in italics point to the essentialisation of the relationship between a language and a nation. Deictic pronouns are especially powerful in this type of discourse: the first-person plural refers to "all Lithuanians in the past several 1000 years", creating an impression of there being a single language belonging to a single people over a millennium.
This one-nation-one-language ideology erases linguistic variation, changing levels of national and linguistic awareness (etc.) from the picture. In the theoretical model, this is ethnic representation: it is pre-supposed that languages represent ethnic groups.
Ethnic representation is manifested in another narrative, the "aggression of foreign languages". Academic experts use conflict metaphors to talk about the influences of foreign languages on Lithuanian. Two recent former heads of the State Commission of Lithuanian Language (examples 2-3) talk about this in the context of English, while the former head of the Lithuanian Language Commission (from the Soviet times), talks the same way about the influence of the Russian language (example 4).
[Appropriate language protection and regimentation helps evade at least some dangers.] (LT-
Lithuanian and foreign languages are in a battle or conflict, which is illustrated by the words in italics in examples 2-4. These words are about attack, defence ("pushing", "invasion" and "forced"), winning and losing ("threat" and "defence"). This shows that the "battle between languages" is like a battle between two nationsa Lithuanian vs. a non-Lithuanian one. The pre-supposed normal state would be a language made up of only "domestic" material. Hence, language itself is defined through an ethnic criterion.
The only competing belief amongst this group of academic linguists is geographical representation, which comes up in dialect-related topics (example 5).
Mes bandome padėti jas išlaikyti -tai mūsų kultūros paveldas… "Even our Samogitian brothers 2 sometimes avoid speaking in dialect; that is why dialects are indeed very much disappearing. We try to help preserve them -that is our cultural heritage…" (LT-7) In the discourse of non-academic experts, dialects are seen negatively (which is the main difference between the two groups). A prominent Lithuanian language teacher, who served as a head of the linguistic society "Lituanistų sambūris" ("The gathering of Lithuanian philologists") calls local dialects "nothing more than a mash of Russian, Polish, other Slavic and Germanic words" (LT-25). The denial of dialects into the Lithuanian we by pointing out the foreign elements in them is a clear sign that language is understood in exclusively ethnic terms.
When it comes to expertise, the dominant type is external. Language is defined through external authorities, such as institutions, dictionaries, grammar books and linguists" research.
The language that is defined by authorities is considered to be good, pure and correct, while real-life usage of language is considered faulty, impure, bad or even dangerous to the set system.
In the Lithuanian institutional discourse, "language errors" (a frequent topic of many articles) cover virtually all levels of language, divided into 8 categories: "lexical, word structure, use of cases, prefixes, use of forms (morphology), parts of speech and sentence conjunction errors, word order and pronunciation". 3 Journalists themselves initiate interviews by talking about language errors, as the media is the main target of the state-enforced language surveillance (Vaicekauskienė 2012: 88). In one interview, an editor is asked: "How do you manage to correct all stylistic, grammar, punctuation, word stress and other errors?" (LT-58).
Already in the journalist"s question, it is suggested that the correctness of language is a job for external authorities, not users. The expert interviewed accepts this supposition.
The texts of academic experts exhibit the same style of speaking about language. External rules are compared to traffic rules and upbringing.
(…) Of course, we could change normative rules (…). However, the limits of such changes cannot violate the systematic factors. The desire to drive against all rules comes from a poorly developed understanding of the language system in school and the family."  An often-used term in the discourse of academic experts is "semi-language", which is seen as the most dangerous variety. 7. Juk mokslinio pranešimo negalėtume skaityti tarmiškai -tarmė tam nepritaikyta. Jeigu vis tiek bandysime tai padaryti, sukursime tam tikrą fantomą, puskalbę, kuri iš tikrųjų neegzistuoja.
"You simply cannot give a scientific presentation in a dialecta dialect is not fit for it. If we anyway try to do that, we create a phantom, a semi-language, which in reality does not exist."  8. Jei visi ims kalbėti puskalbe, pasikeis sintaksė, niekas nebetaisys klaidų, kalba gali mirti labai greitai. "If everyone stars speaking in semi-language, the syntax will change, no-one will correct errors, language can die very quickly."  Vukotić, V. 2019. Social engineers and myth-busters: A comparative research on Lithuanian, Norwegian and Serbian language experts. Taikomoji kalbotyra, 12: 292-344, www.taikomojikalbotyra.lt.
Semi-language as the catalyser of language death can be interpreted as external expertise in the following way: only the systematic, "grammatical" way of speaking and writing is acceptable, because there are external authorities that prescribe it. External expertise often goes hand in hand with ethnic representation (cf. , forming a notion of language that will be called "monoglot", because it supposes that language is invariable and "that a society is in effect monolingual (…) coupled with a denial of practices that point toward factual multilingualism and linguistic diversity" (Blommaert 2006: 243-44, italics in the original).
The next belief often present is the symbolic function of language as a tool of cultural identification. It is rarely found alone, and is often combined with external expertise, forming a notion of language that I will call "normative". Here, language is seen as a public face, or in the words of the former head of the Lithuanian Language Research Institute, Jolanta Zabarskaitė, "correct language is one"s business card" (LT-24), and, in the words of a leading standard language scholar, Rita Miliūnaitė, "language quickly reveals who you are, how you live" (LT-9). Thus, a strict hierarchy of correct forms and appropriate varieties is formed; their use is seen as a reflection of a person"s social status: 9. (...) iš kalbos nesunku atskirti Gariūnų berniuką nuo studento, mokytoją nuo fūristo ir pan. (…) aukštesniųjų visuomenės sluoksnių kalba bus labiau standartizuota, o iš žemesniųjų niekas to nė nereikalauja. "(...) it is not difficult to separate a market seller from a university student, a teacher from a truck-driver and similar (...) the language of the higher classes will be more standardised, and no-one demands that from the lower classes."  Because of the strict hierarchy of linguistic varieties and social identities expressed in this discourse, this function will be classified as the function of hierarchical social identification.
Another sub-function is that of national identification, observed in those articles where ethnic representation was found. By using his/her language, one is said to express national values, culture and worldviews. In this Humboldtian fashion, as the former head of the Lithuanian Language Research Institute says, "language is a unique product of humanity, allowing us to transfer the culture and the value of different nations from generation to . Social engineers and myth-busters: A comparative research on Lithuanian, Norwegian and Serbian language experts. Taikomoji kalbotyra, 12: 292-344, www.taikomojikalbotyra.lt.
"Now it"s the samean emigrant who left Lithuania around World War II, is much better in Lithuanian than the one who, excuse me, washed dishes for a few years some place in England. When he returns, suddenly he cannot pronounce "t", does not use normal Lithuanian syntax"  This combination of ethnic representation and function of national identification shall be referred to as the "ethnolinguistic" notion of language.
In some cases, we can see a combination of all three aspects. Example 11 illustrates the idea that "correct national language" has the function to identify one"s national and social status (representation: ethnic, expertise: external, function: cultural identificational, both national and hierarchical-social). This will be referred to as the "national-elitist" notion of language.
11. Kaţkur dingo nuostata, kad aš esu lietuvis, rašau taisyklingai, viešai kalbu taisyklingai. "The attitude that I am Lithuanian, I write correctly, I speak correctly in public has somehow disappeared."  An alternative, less frequent, is the understanding of expertise as autonomous (n=8). What is essentially different is the relationship towards language change: unlike in the examples above, where external language regulation is considered necessary and effective, change is seen as inevitable. Language is seen as its own master, separated from the language speakers and authorities alike, developing according to unknown rules. The words "change" and "development" are keywords in this discourse (example 12). It is usually activated when Vukotić, V. 2019. Social engineers and myth-busters: A comparative research on Lithuanian, Norwegian and Serbian language experts. Taikomoji kalbotyra, 12: 292-344, www.taikomojikalbotyra.lt. talking about changes on the level of the whole linguistic community over long periods of time.
12. Nors ţmonės mėgina šnekamąsias kalbas veikti, daţnai jos vis tiek kinta savo nenuspėjamais būdais, tarsi pasijuokdamos iš tų pastangų. "Although people try to influence spoken languages, often they still change in their own unpredictable ways, as if they are making fun of those attempts."  The remaining 12 articles present voices of academics who oppose and criticise the dominant ideas of language. Their position as an "alternative" voice is visible from the titles of the articles, designed to challenge the dominant "danger" narrative; hence, many are titled "The Lithuanian language is not in danger" (LT-5), "Is the youth"s literacy decreasing? Let us not make an idol out of Lithuanian language" (LT-32) or, ironically, "The size of language errors" (LT-41).
The external expertise in the dominant discourse is challenged via an internal understanding of expertise. The experts refer to mother tongue competence as proof that linguistic expertise is internalised in the minds of the speakers, and not in dictionaries, institutions, or authoritative norm-setters.
13. Suaugusiam ţmogui pakeisti gimtosios kalbos elementą (…) reiškia pastangą įrašyti naują ţodį į atmintį vienoje smegenų zonoje (…) Negalima įsivaizduoti tokių keitimų sėkmės visos gimtakalbių bendruomenės mastu. "To change an element of the mother tongue (…) it means an attempt to write a new word in the memory in one zone of the brain (…) We cannot imagine such changes on the scale of a whole linguistic community."  14. Kalbininkas, kaip ir kiekvienas ţmogus, gali sukurti ţodį, bet tai tikrai nėra jo funkcija" (…) kiekvienas iš mūsų yra kalbos kūrėjas "A linguist, like any other person, can create a word, but that really is not his function (…) Any one of us is the creator of language" (LT-36) The keywords "mother tongue" and "language creator" are used to point out that it is the cognitive and social linguistic capabilities as well as experiences what makes one a language expert (internal expertise).
Contrary to the dominant discourse, where language is "more than communication", the counter-discourse stresses that the communicative function is primary. Like in the previous Vukotić, V. 2019. Social engineers and myth-busters: A comparative research on Lithuanian, Norwegian and Serbian language experts. Taikomoji kalbotyra, 12: 292-344, www.taikomojikalbotyra.lt. case, this function is rarely found in isolation from other beliefs; it is most often connected to internal expertise. In example 15, expertise is considered the ability to communicate successfully, rather than to uphold normative rules or avoid non-domestic vocabulary: 15. Mes kalbą naudojame komunikavimui -rašykite kaip patogu. Kai reikia greitai išsiųsti ţinutę, galiu prisiekti, nė vienas tuo metu negalvoja apie lietuvių kalbos kančią -tokio dalyko nėra.
"We use language for communicationwrite as you please [ref. to texting without specific Lithuanian characters]. When you need to promptly text a message, I can swear that no one is thinking about a language that is sufferingthere is no such thing."  The combination of these two beliefs will be referred to as the pragmatic notion of language.
The symbolic function is also found in the data. Unlike in the dominant discourse, it is stressed here that we can be "different" through language: 16. Pagal socialinę nuojautą sprendţiame, kaip ir kada tinkama kalbėti, o ne visada vienodai.
Kalba yra labai ryški tapatybės dalis ir, laimei, leidţia mums būti skirtingiems. "According to the social instinct, we choose when and how we should speak, and not always the same way.
Language is a very clear marker of identity and, luckily, it allows us to be different."  This view of language does not offer a fixed group that is represented through language (i.e. "Lithuanians"), nor a social class / status in society (compare example 16 with examples 9 and 10), but rather different social identities can be expressed by an individual. Thus, this horizontal social-identificational function relies on the belief that social identity is what is represented through language. This combination of beliefs (the social and the linguistic) comes from sociolinguistic research, so it will be referred to as variationist notion of language.
Alternative academic experts, who criticise the above-described discourse, tend to distance themselves from the belief that language represents an ethnic group, for example, a Lithuanian professor from the University of Illinois states: "In America, there are Lithuanians who do not speak Lithuanian. You will not tell him, that he is not Lithuanian." (LT-62). It should be noted that, even though no competing belief regarding representations is offered, the ethnic belief is intentionally rejected by pointing out that there is not one-on-one Vukotić, V. 2019. Social engineers and myth-busters: A comparative research on Lithuanian, Norwegian and Serbian language experts. Taikomoji kalbotyra, 12: 292-344, www.taikomojikalbotyra.lt. relationship between language and ethnicity. This will be coded as anti-ethnic representation, as it is also an important element in the construction of ideology.
To summarize, the beliefs and notions of mainstream academics (ethnic representation, external expertise and social-identificational function) can be categorised as monolingual, as they bear similarity to the ideology of language origination in European modernity, a synthesis of linguistic normativism and nationalism (Bauman & Briggs 2003). The beliefs that come as a critique of monolingual beliefs (social and individual representation, internal expertise, horizontal social-identificational and the communicative function) seem to take their ideological starting point in the idea that language originates within human consciousness, not historical narratives, and will thus be categorised as cognitivist beliefs and notions of language. Autonomous expertise does not fit into either of these categories, so it will be classified as a historical belief, as the explanation for the nature of language is found in its own history.
In the discourse of most academic and virtually all non-academic experts, language is understood as the main feature of national identity, and one"s public face and speaking correctly are connected to one"s moral values and patriotism. In the analytical model, the aspect of representation is ethnic; the expertise is external. The geographical representation is present amongst some academic linguists as secondary, but not amongst non-linguists. The function is mostly identificational: in some articles, language is a symbolic act of showing a person"s social identity (hierarchical social-identificational), and in others, one"s national identity (national identification), depending on whether ethnic representation or the external expertise are also present in the article.
The opposing notions of language are found in the counter-discourse of some academics, here labelled alternative academic experts. The aspect of representation largely remains unrealised in their discourse (see appendix 2; in a few cases, anti-ethnic and social representation were found): the expertise is internal, and the function is communicative or horizontalidentificational. . Social engineers and myth-busters: A comparative research on Lithuanian, Norwegian and Serbian language experts. Taikomoji kalbotyra, 12: 292-344, www.taikomojikalbotyra.lt.

Norwegian experts
The 44 articles gathered encompass a broad range of topics. A series of five articles was a debate about bilingualism (mentioned in 2.2.); another five are more about youth language and emerging linguistic varieties in urban areas; nine articles are about dialects and written languages; five about language change; four about language and politics (two about language policy, and two about the language of politicians).
The largest number of articles collected was written by academic experts (n=8). The first group of academic experts are labelled researchers, because they are presented through research institution affiliations, talk about the results of their research and comment on the linguistic issues from that perspective. The other group are popular experts, linguistic celebrities (n=8) that regularly comment on language-related subjects: Helene Uri, Sylfest Lomheim and the journalist Per Egil Hegge. The first two do hold a PhD in linguistics, but they are not presented as researchers in the news article because of their general fame. Helene Uri is famous primarily as a best-selling novelist, and Sylfest Lomheim is a former head of the Language Council, a politician and a permanent commentator on a weekly radio show that deals with language. The third celebrity, Per Egil Hegge, is a columnist popular for his column on language correctness. Non-academic experts (n=8) include language teachers, heads of linguistic societies and journalists.
The first subject to be analysed is the 2013 "bilingualism debate". The debate was initiated by researchers from the Department of Education (University of Oslo), who pointed out that immigrant children need more Norwegian language training, noting that their bilingualism is an impairment rather than an advantage. As a response, researchers from the University of There is much research that shows that a child who learns several languages develops cognitive capabilities and specific resources."  To this, the first group answered back that: Vukotić, V. 2019. Social engineers and myth-busters: A comparative research on Lithuanian, Norwegian and Serbian language experts. Taikomoji kalbotyra, 12: 292-344, www.taikomojikalbotyra.lt. In the discussions about multiethnolects, a subject close to that of bilingualism and migration, identity becomes a topic inevitably connected to language. In such discussions, beliefs connected to representation are often expressed. One article explores the phenomenon that ethnic Norwegian young people are using the same language as many children of immigrants (often called "Kebab-Norwegian"). An academic expert comments that this situation is typical and natural: 19. Norskfødt ungdom snakker gebrokkent fordi de vil markere tilhørighet til sine venner som har innvandrerbakgrunn. Dette handler om nye måter å vaere norsk på. Norge er mer enn Kari og Ola. "A Norwegian-born youth speaks with this accent because they want to mark the affiliation with their friends who have immigrant background. This is about new ways of being Norwegian. Norway is more than Kari and Ola." (NO-8) "Kari and Ola" is a culturally bound phrase meaning "typically Norwegian", as Kari and Ola are typical Norwegian names. By using this phrase, the ethnic representation is deconstructed and replaced by the belief that language represents a social group. Also, language is seen as a tool of identification, and as mixed urban varieties are seen as natural, language functions as an identifier of a group/social identity. . Social engineers and myth-busters: A comparative research on Lithuanian, Norwegian and Serbian language experts. Taikomoji kalbotyra, 12: 292-344, www.taikomojikalbotyra.lt.
The next big topic is language change. Here, most of the academics function as "mythbusters", deconstructing the notion that standard language is a guarantee of stability (external expertise). A frequent belief that comes up is autonomous expertise, meaning that language is "its own master". The example below personifies language through the metaphor LANGUAGE IS AN ORGANISM; it is given attributes of a living being, such as "will": 20. Man kan lage regler, man kan bruke rødblyanten, man kan forklare og argumentere. Men språkets iboende krefter er sterkere enn alle foreldre og laerere til sammen. (…) Språket lever sitt eget liv og har en sterk egenvilje. "One can make rules, one can use the red pen, one can give explanations and arguments. But the internal powers of language are stronger than all parents and teachers put together. (…) Language lives its own life and has its own strong will."  Academic experts also refer to the "purity myth", engaging in a discussion about representation and detaching it from the notions of ethnicity (anti-ethnic representation): 21. Norsk er likevel ikke et truet språk. Men det endrer seg, slik det alltid har gjort. Et språk består i stor grad av lånord. På den måten er alle språk hybrider. "Norwegian is, nevertheless, not an endangered language. But it changes, as it always has done. A language consists of a high degree of borrowings. In this way, all languages are hybrids."  The discourse of popular experts is structured so that one (Helene Uri) is constructed as the "liberal" and the other (Sylfest Lomheim) as the "conservative" voice on language issues. The ideology seen in Lomheim"s articles is based on the ethnic notion of representation and external expertise.

22.
vi savner bevissthet på å ta vare på det språket vi er best på, norsk "we do not have the consciousness to take care of the language that we know best, Norwegian" (NO-3) Many linguists in Norway see Lomheim to be a lone voice in these views, yet they acknowledge his popularity amongst non-linguists. It is important to notice that Lomheim was the head of the Norwegian Language Council from 2003 to 2010 even though his ideas were met with strong resistance from academics (cf. NO-4). On the other hand, Helene Uri is constructed in the media as the liberal expert, who is "pro English words" and "pro youth language". However, the notions of language found in her articles vary greatly. In most cases the function is communicative and expertise internal, like in example 23, where good language is the question of the speaker and communicative situations. Vukotić, V. 2019. Social engineers and myth-busters: A comparative research on Lithuanian, Norwegian and Serbian language experts. Taikomoji kalbotyra, 12: 292-344, www.taikomojikalbotyra.lt.

23.
Det er ikke noe som er lov og ikke lov, det er opp til språkbrukeren selv. Det kommer også an på hvem avsenderen er. "There is no allowed and not allowed, that is up to language users themselves. It also depends on who the sender is"  The notion of representation is sometimes activated in Uri"s articles alongside a variety of other approaches. In example 23, it appears that language is observed on an individual level.
With the communicative function of language, it constructs an atomic notion of language (language is a personal possession, used for intrapersonal communication). However, in other articles, it varies between anti-ethnic and ethnic: 24. Og norsk forblir norsk likevel -selv om det skulle komme inn enorme mengder lånord og oversettelseslån fra engelsk "Norwegian will be Norwegian anywaydespite the fact that enormous amounts of loanwords and translation loans will come from English."  25. Den største trusselen er at man bruker engelsk der det er fullt mulig å bruke norsk. Den strategiske språkpolitikken i Norge må dermed legge opp til å vise at norsk duger. "The biggest threat is that one uses the English language [in situations] where it is totally possible to use Norwegian. The strategic language policy in Norway, therefore, has to make sure that Norwegian becomes good enough."  These discrepancies could have emerged due to several reasons: either because the subject varies between the corpus to status of language; it could also be a product of journalistic practice (as only one of Uri"s articles in my data was authored by Uri herself); or they could also be a result of a changed opinion. This question is not that relevant for the present research, but future studies should note the discrepancy.
Articles of non-academics mostly exhibit language monolingual beliefs: external expertise and ethnic representation (5 out of 6 articles). In an opinion article, a language teacher asks: "Is the Norwegian Language Council asleep?", referring to a lack of normative and regulative practices of the institution. He writes that the Council should protect the language, referring to a linguist who, he claims, used to function as "language police": I thought that the Council should protect the Norwegian language (…) Finn-Erik Vinje, we need the language police!"  The notion of function was rarely found in the Norwegian data, except for the debate on bilingualism, where both sides employed the instrumentalist notion of language. The communicative function was found in 5 articles, and the symbolic one appeared in those discussing multiethnolects.
To summarize, the distribution of the beliefs about representation is either role-or topicdependant. The ethnic representation is found amongst non-academics, while the academic linguists either distance themselves from the ethnic notion of language or employ the notion that language represents social groups. Norwegian academic experts exhibit only internal or autonomous expertise, in no case external. External expertise is found amongst nonacademics and in the texts of the popular linguist Sylfest Lomheim. This clear difference between academic and non-academic experts might come from the fact that Norwegian experts build their identity as "myth busters": they seek to deconstruct a certain belief about language (usually ethnic representation or external expertise). The topic of youth language and identity provokes the belief in (horizontal) social identification as a central function of language.

Serbian experts
Unlike in Norway and Lithuania, four out of five news-portals seem to have a clear linguistic ideology. The more right-wing "politika." (n=44), "novosti.rs" (n=6) and "telegraf.rs" (n=2) provide space for the voice of many experts from Serbian academic institutions (both in Serbia and the parts of Bosnia and Montenegro where Serbs live) and well-known language editors. Their ideology, as I will illustrate below, most often includes beliefs on the monolingual side. The liberal-left "danas.rs" (n=17) endorses linguists that criticise the Vukotić, V. 2019. Social engineers and myth-busters: A comparative research on Lithuanian, Norwegian and Serbian language experts. Taikomoji kalbotyra, 12: 292-344, www.taikomojikalbotyra.lt.
ideologies of language of the first group, expressing cognitivist beliefs. Their views are shared by like-minded experts from Bosnia, Croatia and Montenegro, which is why their opponents label them as "pro-Yugoslav" (because they also express positive attitudes towards Serbo-Croatian). The news portal "Blic" (n=9) does not seem to represent any specific language ideology.
The ideology of the experts of the first group is largely based on ethnic representation, external expertise and the symbolic function of language. The narratives revolving around ethnic representation connect, much like in the Lithuanian context (cf. example 1), Serbian identity to a presumably homogenous Serbian language over space and time: 27. bili smo pod okriljem Pećke patrijaršije (…). I to je bilo dovoljno da sačuvamo srpstvo, da ne zaboravimo sopstveno srpsko ime, svoj srpski jezik i njegovo pismo. "we were under the Patriarchate of Peć (…). And that was enough to preserve Serbianhood, not to forget our own Serbian name, our Serbian language and its [Cyrillic] script."  The first part of the quote, which refers to the Patriarchate of Peć, the first autocephalous (Serbian) Orthodox church established in 1346, is used to signal the unity of Serbian people and its language over many centuries.
Serbian experts additionally voice the views of ethnic representation. Some linguists represented on "politika.rs" and "novosti.rs" (n=8) belong to an ideological group whose discourse features the belief that language represents a primordial ethnic group. Since the beginning of the 1990s, these linguists have been employing a "Greater Serbian narrative" in their texts. "Greater Serbia" is an extreme-right wing version of Serbian nationalism, which seeks to unite all the lands in which Serbs have presumably lived. What they see as Serbs are, however, not limited to self-identified Serbs, but include all who speak a language mutually intelligible with Serbian (including Bosnian, Croatian and Montenegrin). In this discourse, Croats who speak dialects close to standard Serbian (that is, most of them) are "catholicised Serbs", Bosnians are "Islamised Serbs", while Orthodoxy (the dominant religion in Serbia) is seen as the original religion (changed through the missionary work of the Ottoman and the Austro-Hungarian empires). As the Serbian (Orthodox) are the original people, the Serbian language is also seen as the original language in the region, while Bosnian, Croatian or Vukotić, V. 2019. Social engineers and myth-busters: A comparative research on Lithuanian, Norwegian and Serbian language experts. Taikomoji kalbotyra, 12: 292-344, www.taikomojikalbotyra.lt.
(…) all of them abolished the Serbian name of Vuk"s language, renaming it into the so-called Croatian, Bosnian/Bosniak and Montenegrin languages."  On the other hand, the experts in the liberal-left news portal "Danas.rs" (labelled "alternative experts" in appendix 2) speak directly to this notion of representation and offer a competing, anti-ethnic representation. In the following quote, the one-nation-one-language ideal is taken from the discourse of the linguists described above, deconstructed as "Nazi ideology": 30. "Jedna drţava, jedan narod, jedan jezik" bio je Hitlerov ideal (…) Nauka je tokom prošlog veka napustila romantičarsku ideju o jedinstvu jezika i nacije, aktuelnu u vreme kad su se one tek formirale, uvideći da nacije ne nastaju na osnovi jezika. Današnje drţave i jezici koji se u njima govore jasno to pokazuju (Austrijanci nisu Nemci, ali govore nemački…). "One-stateone-people-one-language was Hitler"s ideal (…) Science has abandoned the romantic idea of the unity of the language and the nation during the last century, relevant in the time of their formation, realizing that nations do not appear on a linguistic basis. The today"s states and language that are spoken there clearly show that (Austrians are not Germans, but speak German…)"  When these experts criticise the Greater Serbian discourse by claiming that languages belong to "speakers" (example 31), but do not closely define what they mean by that, it can also be treated as either anti-ethnic or individual representation: 31. Jezici pripadaju svojim govornicima ma kojem narodu oni pripadali "Languages belong to their speakers, no matter what people they belong to"  The prescriptivist notion of language (and its components, external expertise and social identificational function) is activated in the articles of the first group of academics by a related subject, the question of literacy and correctness. A professional linguist is replying to a worried mother (example 32), who noticed that her son talks to her using a lot of slang words.
Жаргон не би смео да продире у књижевни језик. Опасност не представљају поједине речи већ конструкције које нарушавају језичку структуру. "Even though modern life is mostly democratised, and the border between the allowed and the not allowed has moved, we have to have moderation when it comes to language. Slang should not penetrate into literary language. The danger are not single words, but constructions that disrupt the linguistic structure."  Later in the text, the linguist (otherwise an author of handbooks on correct language) rephrases the use of incorrect grammar as "sinning against the norm", a sign that bad use of language is connected to something morally wrong. Through language, one can measure how much of the existing social norms and social value systems the person upholds or does not.
The function of language as a tool of social identification is here seen in a hierarchical way, in which "lower" varieties are seen as faulty and dangerous, and "higher" ones as morally superior and valuable.
In the discourse of the second group of academics, the internal expertise is highlighted by referring to the function of language to produce meaningful utterances (communicative), as in example 33, or by connecting it to varieties, as in example (34). 33. Ima tekstova koji su bez pravopisnih i gramatičkih grešaka, ali nije jasno šta je pisac hteo da kaţe. Sa druge strane, ima tekstova i sa jednim i sa drugim greškama, ali je potpuno jasno šta pisac govori. "There are texts with no written or grammatical errors, but it is unclear what the The reference to non-standard varieties signalizes both geographical representation, but also (in this context) that expertise is internal, as it is used as an argument against the Vukotić, V. 2019. Social engineers and myth-busters: A comparative research on Lithuanian, Norwegian and Serbian language experts. Taikomoji kalbotyra, 12: 292-344, www.taikomojikalbotyra.lt.
"correct/incorrect" narrative (external expertise). Also, the inclusion of non-standard varieties into a legitimate place of language points to the social identificational function of language, but unlike in the example above, it is horizonal, as all varieties are seen as valid.
To summarise, the unique feature of the Serbian experts" is primordial ethnic representation, while the other aspects shift between mutually exclusive beliefs such as internal vs. external expertise, the function of hierarchical social identification vs.

Quantitative and comparative results, discussion
In this section, I will present a comparison of the beliefs, the types of experts and will supplement this information with quantitative data (all numbers and percentages are presented in appendices 1 and 2).

Types of beliefs:
As stressed in the summary of 3.1., beliefs about language can be classified into two main categories, the monolingual and the cognitivist, and one small category: historical. The first category is close to what has been described as the modernist (Bauman & Briggs 2003) or monoglot (Silverstein 1996) ideology of language: real language is seen as the one that is "ethnically pure" and "correct" and is thus used to measure one"s loyalty to a nation-state and personal social status. In the theoretical model, it takes the form of ethnic (sometimes also geographical) representation, external expertise and culturalidentificational functions, as well as the complex notions formed by those beliefs (nationalelitist, monoglot, ethnolinguistic, normativist and prescriptivist). Cognitivist beliefs come as a criticism of the monolingual beliefs; they are social or individual (or without) representation, internal expertise and the communicative function, and the notions formed by those beliefs (pragmatic, variationist and atomic). Lastly, the historical ideology consists of only one belief: the autonomous expertise. This is so because this belief usually erases the aspect of representation and function from the picture: language is talked about in its own terms, . Social engineers and myth-busters: A comparative research on Lithuanian,Norwegian and Serbian language experts. Taikomoji kalbotyra,www.taikomojikalbotyra.lt. separated from the speakers that use it. These overall ideologies will be discussed more in the end, but first, a quantitative comparison is provided.

Notions of language:
The national-elitist notion was found 4 Lithuanian, 1 Norwegian and 4 Serbian articles. More frequent were the monoglot (LT: 17.2%, SR: 13.9%) and ethnolinguistic (LT: 12.9%, SR: 8.9%) notions of language (appendix 1). Most of these were found amongst academic experts in both Lithuania and Serbia, and a smaller part amongst NO n=1,SR n=9,ethnolinguistic: LT n=8,NO n=1,SR n=9). In the Norwegian data, the most prominent notion of language was pragmatic (n=8), . Social engineers and myth-busters: A comparative research on Lithuanian, Norwegian and Serbian language experts. Taikomoji kalbotyra, 12: 292-344, www.taikomojikalbotyra.lt. and it was exclusively academic experts who expressed this notion in Norway. In Lithuania and Serbia, the distribution is upside-down if compared to Norway; the pragmatic notion is found only amongst "alternative" academic experts (LT n=5, SR n=5), and none amongst (mainstream) academic and non-academic experts. In other words, Lithuanian and Serbian mainstream academic experts express monolingual, while Norwegian ones express cognitivist beliefs and notions of language.
As shown, the most similar group in the three countries are non-academic experts (predominantly monolingual). On the other hand, Lithuanian and Serbian alternative experts are similar to Norwegian mainstream academic experts (predominantly cognitivist beliefs, but also more historical in the Norwegian case). Lithuanian and Serbian "dominant" experts exhibit also a great deal of similarity (monolingual), except that the Serbian ones also express primordial ethnic representation, not just ethnic.
Non-academic experts, along with mainstream Lithuanian and Serbian academic experts can be grouped into the modernist category of experts: real language is seen as the one that is "ethnically pure" and "correct" and is thus used to measure one"s loyalty to a nation-state and personal social status (cf. Bauman & Briggs 2003). The Norwegian mainstream and Lithuanian and Serbian alternative academics could be grouped as experts critical of modernity. Their role is sometimes constructed as that of a "myth buster", the one who speaks against supposedly widespread misconceptions about language. The fact that Lithuanian and Serbian academic experts express beliefs that would be considered "myths" in the Norwegian discourse points to the fact that a common state-socialist past might have played a role here.
The role of the intellectual in a state-socialist system entailed that scientists and researchers contribute to the building of a communist future, and they are often needed (regardless of what they were actually doing) to present their work as compatible with the ideology of the centralist state. Lithuanian and Serbian experts could also see themselves as "social engineers", an idea that gained momentum in the early Soviet Union. The articles in the news can be interpreted as an attempt to, through an appropriate ideological tone, raise public awareness of the importance of their profession. The material conditions could also play a role here: most workplaces were allocated by centralised work planning committees, and one Vukotić, V. 2019. Social engineers and myth-busters: A comparative research on Lithuanian, Norwegian and Serbian language experts. Taikomoji kalbotyra, 12: 292-344, www.taikomojikalbotyra.lt. of the ways of getting a job would be to publicly show that the system needs your profession.
Both Lithuanian and Serbian experts seem to be trying to raise "moral panics" to justify the need for their profession: language control and/or purification.
The difference in the representation (largely absent amongst Norwegian experts, and dominantly ethnic in Lithuania and Serbia) might come from the fact that Lithuania and Serbia are countries that are in the process of "nation-re-building", and cultural elites are engaged in re-engineering the culture and traditions they feel were damaged by the five decades of state socialism, and language is one of them (the perceived fear of Russification and the Serbo-Croatian language policy amongst others).
The Norwegian academic experts seem to see their role in public life in the tradition of the Enlightenment to battle prejudices and "folk" understandings of language. It is also possible that Norwegian academics take their understanding of expertise from various schools of linguistics. Autonomous expertise comes from the school of historical linguistics, which sees language change as inevitable and independent of the language users" efforts, while internal expertise comes from the study of language variation and change: language is formed and created by its users through acts of communication. It also suggests that the observations made in the Norwegian studies (Ims 2014) are generalizable: non-academic experts express beliefs on the monolingual side, while academic experts express beliefs on the cognitivist side, deconstructing the monolingual beliefs caused by perceived mixing and "bad" language (cf. Svendsen & Marzo 2015).
A separate piece of research would need to investigate how successfully an expert or article transmits their ideology. This could be, for example, an analysis of comments to articles on social networks, the inclusion of the number of reads per article or the numbers of republications on other news portals by adopting the "uses and gratifications" approach. The present study is only an overview of the types of experts, the beliefs they express and their overall dominance in public space.

Conclusions
Vukotić, V. 2019. Social engineers and myth-busters: A comparative research on Lithuanian, Norwegian and Serbian language experts. Taikomoji kalbotyra, 12: 292-344, www.taikomojikalbotyra.lt. This paper aimed to compare the ideologies of language experts in Lithuania, Norway and Serbia. At first, two main types of ideological beliefs were identified. The first set of beliefs is connected to the monolingual modernist understanding of language, where language is understood in terms of its symbolic value, ethnic and/or geographical origin and a system based on external rules. The second set of beliefs expresses a cognitivist view of language as a product of the human brain, interaction and identity-making. One belief could not be categorised as either, and that is the belief that language is its own master, separate from both individuals and groups, authorities and functions (called historical in this paper).
Second, the types of experts participating in the virtual sphere were identified. In all the three countries, these were academic (presented as affiliated with a research institution) and nonacademic experts (teachers, editors, journalists, writers, and others). In Lithuania and Serbia, academic experts are subdivided into a (larger) mainstream and a (smaller) alternative group because of their opposed ideologies. In Norway, academic experts are categorised into researchers, who only comment on issues from the perspective of their research, and "popular experts", who are famous for their non-academic work and are default commentators on virtually all linguistic issues.
Monolingual beliefs dominate the Lithuanian and Serbian public sphere, while in Norway, there are slightly more cognitivist beliefs expressed than monolingual ones (appendix 1). The most surprising finding is the ideology of academic experts. In Lithuania and Serbia, the dominant group of academic experts exhibited almost exclusively monolingual beliefs, while the alternative group exhibited almost only historical and cognitivist beliefs. Norwegian academic experts are ideologically very similar to the Lithuanian and Serbian alternative group, expressing only cognitive and autonomous beliefs. On the other hand, non-academic experts shared very similar ideological views in the three countries, i.e. monolingual views (appendix 2).
Despite that, some general differences emerge: representation is activated significantly more often in the discourse of Lithuanian (75.8%) and Serbian (75.9%) experts than Norwegian (38.6%), while expertise and function are similar. This could be the result of the data Vukotić, V. 2019. Social engineers and myth-busters: A comparative research on Lithuanian, Norwegian and Serbian language experts. Taikomoji kalbotyra, 12: 292-344, www.taikomojikalbotyra.lt. collection method and the years covered in this research, but it could also come from the post-state socialist nation-re-building efforts, in which concerns over the "national" and "ethnic" are made a public priority (subjects seen as damaging to the nation such as emigration are high on the national agenda). Ethnic representation is dominant in Lithuania and Serbia, but in Serbia the primordial ethnic one can also be found. It could be due to the fact that mutually intelligible / almost identical standard languages exist around Serbia; thus, experts have to prove the uniqueness of the Serbian language and nation by reaching for primordial arguments.
The distribution of ideologies can be explained in a number of ways. The similarity between Lithuania and Serbia logically entails that the period of 1945-1990 plays an important role.
An expert as a "social engineer" was an idea dominant in the Soviet Union, which could be one source of explanation. Language experts were not supposed to just explain language, but they also had to create and/or sustain it. In the Norwegian case, the role of academic and some popular experts in fighting language myths and "folk" understanding of language reflects the idea popularised by language planning scholars and sociolinguists (such as Joshua Fishman and Deborah Cameron) that linguists should take active part in the debates about language and not leave language to prescriptivists. are identified using a discourse analytical approach, looking for implicit beliefs about language and their frequency. The comparison is based on a pre-constructed theoretical model of a language ideology, encompassing three ideological aspects: representation, expertise and function. The first aspect refers to the ideologised relationship between a language / variety and the imagined group that speaks that variety. The second describes what is considered "good" language; it can be the forms of language defined by authorities or the language of the speakers themselves. The third is about which function of language is considered the main one, most often it is the communicative or the symbolic. The results show that various beliefs could be grouped into two main (mutually exclusive) categories, the monolingual and the cognitivist. The first category includes those beliefs that point to the ideology that one ethnic group is represented by one (homogenous) language, its content is defined by authorities, and its function is to identify one"s patriotism and social status. The second group of beliefs come as a reaction to the first one: language is separated from national identity and connected to personal identities, the common language users are considered the experts, and the function is primarily communicative. Lithuanian