Dirtbags, Drunkards and Miniature Mutes: Czech Subjectivity Revealed Through Corpus-Based Discourse Analysis

ABSTRACT

Language may be seen as a tool for constructing and confirming power structures, and a corpus analysis of adjacent words may reveal how individuals or groups of people other than the sender (writer or speaker) are depicted. These depictions frequently reveal a phenomenon known as linguistic othering. The aim of this paper is to present a corpus-based survey of the linguistic othering of Roma, Vietnamese and Ukrainian people in Czech media discourse from 1989 to 2014. The representative result is acquired by comparing neutral, positive and negative adjectives related to the three key lemmata, and a quantitative method is used to answer analytical questions about the query words in this context. Although some previous researchers have used similar methods, it appears that no such study has been performed on such a large body of material for Slavic languages. The outcome reveals how these three groups are differentiated in text, and the source material helps to demonstrate how language usage reflects the discourse of Czech society.

KEYWORDS:

• corpus analysis
• othering
• discourse analysis
• Czech media
• stereotypes
• frequency analysis

0. Introduction

The aim of this article is to consider how language usage may affect recipients’ views of the world. Although different recipients respond to what they read or hear in different ways, a more frequent use of a particular expression often results in acceptance of the expression in question (cf. Lindqvist, Bäck and Gustafsson Sendén 2016, 113–14). The use of corpus analysis makes it possible to discern frequency of usage and thus to test theories regarding the relationship between language and thought (Stubbs 1997, 1).¹

The first goal of the present study is to quantitatively examine the linguistic othering of Roma, Vietnamese and Ukrainians in Czech(oslovak) printed media discourse during the 25 years following the Velvet Revolution in 1989, in order to identify the various noun modifiers that have been used about these minorities. The second goal is to arrive at a comprehensive, representative, corpus-based result by comparing the neutral, positive and negative adjectives that are found adjacent to the query words ‘Rom’, ‘Vietnamec’ and ‘Ukrajinec’ in the data, using the Czech Subjectivity Lexicon (Veselovská, Hajič and Šindlerová 2014, 59). The work focuses on these adjacent adjectives in search of “the above-chance frequent co-occurrence of two words within a pre-determined span” (Baker, Gabrielatos, Khosravinik, Krzyzanowski, McEnery and Wodak 2008, 278).

According to the Czech Statistical Office, the largest minority groups in the Czech Republic are Slovakian, Ukrainian, Vietnamese, Russian² and Romany.³ This paper focuses on three of these groups. After a background description, the following questions will be posed:

1. To what extent are these minority groups described in a stereotypical manner?

2. Are any of the groups (Roma, Vietnamese and Ukrainians) more pejoratively othered in the Czech media discourse than the others?

1. A short socio-political background of Roma, Vietnamese and Ukrainians in the Czech Republic

1.1. Roma

There have been Roma in Czech lands since the year 1416 (Ottův slovník naučný 2892, 366), and possibly even before that. During the National Revival of the 19th century, they emerged in popular books and songs as an exotic and culturally active people, of which the most famous Czech example is Karel Hynek Mácha’s Cikáni. As this paper demonstrates, however, today’s descriptions of Roma are less romantic, proud and colourful.

The number of Roma in the Czech Republic is difficult to assess because many people of Roma background or ancestry try to avoid being registered as such (Doubek, Levínská and Bittnerová 2015, 132–33). It is therefore also difficult to obtain a definitive statistical background regarding their residence in the Czech Republic. However, the web portal Romea.cz has tried to assemble some facts, aided by the Museum for Roma culture and history (Romea 2010). The Czechoslovak census of 1925 counted 62,192 Roma in Slovakia, 1994 in Moravia, 579 in Bohemia and 79 in Silesia. Then, in 1927, a law was passed prohibiting Roma from moving around freely. A decade later, Roma adults were issued with a “Gypsy identification paper,” which they had to carry at all times. During this period schooling was, from the authorities point of view, seen as a way of integrating them into Czechoslovak society, and some Roma even attended higher education – at the same time as they were forced to carry a specific identification paper and could be banned from certain places by law (Lhotka 2010a). After 1945, many Slovak Roma moved to the Czech and Moravian parts of the Republic. During the years 1948–1989, the authorities officially tried to improve the life of the Roma by providing schooling and health information, but also demanded that they ignore their Roma identity in favour of their Czechoslovak identity. After 1989, a Roma political party gained representation in parliament for a few years but even today, not all Czech Roma have acquired an education and many are unemployed and poor, even if the exact numbers are difficult to pin down (Lhotka 2010b).

1.2. Vietnamese

The Vietnamese began migrating in large numbers from the Socialist Republic of Vietnam to the Socialist Republic of Czechoslovakia (ČSSR) in the 1950s, after diplomatic relations had been established between the two countries (Hantonová 2013, 8) and an agreement was drawn up concerning scientific/technical co-operation (Hantonová 2013, 9; Alamgir 2013, 71). From the beginning, they were perceived as quiet and hard-working (Hantonová 2013, 10). For those who arrived during the 1980s, this quietness may have been partly because they were not given as much schooling in the Czech language as their predecessors (Alamgir 2013, 72; Hantonová 2013, 10), though today the Vietnamese are portrayed as having a generally high educational level (Český statistický úřad 2014).

When the Velvet Revolution occurred, there were fewer than 30,000 Vietnamese in the ČSSR (Hantonová 2013, 11), whereas today there are estimated to be some 48,000 permanent residents who still have Vietnamese citizenship (Český statistický úřad 2016), and some 60,000 altogether (Český statistický úřad 2012). After 1989, the new Czechoslovak republic (Česká a Slovenská Federativní Republika, ČSFR) continued to maintain the transnational agreement, and after 1993, the Czech Republic and the Socialist Republic of Vietnam immediately established diplomatic relations (Hantonová 2013, 12).

Approximately half of the estimated 60,000 Vietnamese inhabitants of the Czech Republic have the right to permanent residence. However, few of them have Czech citizenship, which, among other things, would allow them to become more active in political elections and parliamentary politics. The Vietnamese have no parliamentary representation (unlike the Roma did for a while) despite having lived in the country for more than 60 years (Slavíčková and Zvagulis 2014, 157–58).

1.3. Ukrainians

The Ukrainians are a Slavic people from a country close to the Czech Republic (and neighbour to Slovakia) that has been the focus of considerable interest in recent times. Ukrainians have tended to be considered as guest workers, but in recent years they have often stayed for longer periods and some have even settled in the Czech Republic (Leontiyeva 2014, 63). In December 2016, there were some 110,000 Ukrainian citizens in the Czech Republic, of whom 81,000 had permanent residence in the country (Český statistický úřad 2016). However, the history of Ukrainians in the Lands of the Bohemian Crown is much older than this. From a historical perspective, the most famous arrivals may have been the Cossacks, who were part Ukrainian. However, the biggest influx began during the 1920s as Ukrainian academic institutions began to be established in Bohemia (Součková 2015, 67), such as the Ukrainian Free University (Ukrajinská svobodná univerzita), which moved from Vienna to Prague, the Ukrainian Economic Academy (Ukrajinská hospodářská akademie) in Poděbrady, and the Ukrainian Gymnasium (Ukrajinské gymnázium) in the town of Řevnice. Indeed, during the period 1919–1939, the south-westernmost part of present-day Ukraine, Carpathian Ruthenia, was also a part of Czechoslovakia.

2. Theoretical framework

The theoretical starting point for this work is Critical Discourse Analysis (CDA; see inter alia Baker et al. 2008, 280), which in this case focuses on printed media. The analysis is based on the concept that power relations within a society are reflected in that society’s popular media. The language used in these media contributes to the worldview of its recipients (readers or listeners), in some cases even helping to construct it. Hence, the language of popular media (i.e., media directed at a general audience and therefore widely read or listened to) has the power to form general stereotypes, which in turn are reflected in the discourse of society as a whole.

Frequent but widely dispersed stereotypical and negative expressions and collocations are examples of what Fairclough describes as a “hidden power” (Fairclough 2015, 73, 78, 80 and primarily 82), which may not be immediately evident, but which slowly enters general discourse within a society. One example of such expressions is the stereotypical ‘lazy other’ (van Dijk 1987, 57; Rönnbäck 2014). This other may not be explicitly called ‘lazy’, but the connotation of laziness is created by discourse in which representatives of a minority are noted for not making good use of opportunities (van Dijk 1987, 57). Many other such connotations are also implicit and, though not obvious from a cursory view, can influence people’s thoughts (inter alia, van Dijk 2008, viii), thereby assisting in the creation of linguistic othering.

‘Othering’ is a new term for an old concept (e.g. Loomba 2015, 112–13; Said 1995). To define others by separating them from oneself has also been discussed in linguistic research, such as the ‘out-groups’ of van Dijk (1998, 153). But the concept of linguistic othering (whether or not this terminology is used) is defined by looking at the words and phrases used in a certain discourse about specific groups, often a minority of some kind, such as immigrants in a country (Baker, Gabrielatos and McEnery 2013) or women in male-dominated professions (Zasina 2018), or even a nation seen from another nation’s perspective (Bolshakova 2016). In this case, the representation in the media is analysed through the written language, as opposed to pictures or spoken language.

Masako Fidler and Václav Cvrček (2015) mention how our conception of the world is often formed through words that are frequently collocated, as in the phrase ‘illegal immigrant’. The current study adopts some of the same theoretical framework as Fidler and Cvrček, namely, that language usage to which we are consistently exposed creates a pattern that, in turn, shapes our conceptions and expectations. Thus, if two words are frequently read together, they may become closely linked and eventually transformed into a collocation (Fidler and Cvrček 2015, 198–99). In her 2016 article on othering in the Czech Republic, Masako Fidler writes of her intention to “find a more automatic mental representation” of the “others” whom she studies. Her view of adjacent words and collocations is that they reinforce an idea of the other, even if the receivers do not notice it entering their worldview (Fidler 2016, 38–39). The resulting representation creates a mental image, which is an aspect of othering.

3. Previous linguistic research

In a quantitative study of collocations with the word mluvící (‘speaking’), achieved through keyword analysis of source material from the Czech National Corpus, Masako Fidler concludes that even the language of Roma people, Romani, is othered. This language does not appear as a collocation with the word for ‘speaking’, despite Romani being one of the four officially protected languages of national minorities in the Czech Republic (Fidler 2016).

The most (in)famous Czech collocation with Roma (whether it be with the words Rom or Cikán) is ‘unadaptable’, nepřizpůsobivý. The concept of being ‘unadaptable’ has been discussed in previous critical discourse studies (cf. van Dijk 1987, 58), but in a Czech context, it is almost exclusively used about Roma (Slavíčková 2015, 74). According to one previous study (Elmerot 2016), it is clear that this adjective is more often used adjacent to the query word Cikán (‘Gypsy’) than to Rom (‘Roma’). This is an example of an othering collocation that suggests a perception of a certain other without being too explicit – and which is therefore part of the “hidden power” of language, as noted above. Hence, one of the issues here is whether other minority groups are also described as ‘unadaptable’ in the source material explored in this paper.

In the mid-2000s, six researchers (cf. Baker et al. 2008, in association with Lancaster University) from the University Centre for Computer Corpus Research on Language joined forces in a project to show how corpus analysis could be used in a critical discourse analysis of UK press coverage of refugees and asylum seekers. That project has become something of a template for the methodology used in similar research.

One conclusion drawn from a critical discourse study as early as 1987 was that the “dominant consensus” of opinion on minorities is created in popular media discourse (van Dijk 1987, 40). The book also provides a good, and still relevant, summary of the common traits of minority representation in the European press, no matter which country is studied:

• – minorities are disproportionately under-represented in the press

• – topics associated with minorities are also associated with problems

• – events are described from the majority’s point of view

• – everyday events are rarely described from the minorities’ points of view (unless they have created problems), and

• – racism is rarely reported, and if it is, it is attributed to poor areas or extremists (van Dijk 1987, 44–45).

In 2014, Slavíčková and Zvagulis published an article describing their analysis of anti-minority rhetoric, which was conducted with “the aim of extending existing linguistic studies (that have also focused mainly on Roma) in the Czech Republic” (Slavíčková and Zvagulis 2014, 155). Their conclusions were that the Roma are a special target for such rhetoric (Slavíč ková and Zvagulis 2014, 153), and that while many officials pay lip service to respect for different ethnic groups, this is belied by value-laden language usage. In addition, Slavíčková and Zvagulis point out that the Roma and Vietnamese rarely feature in Czech media except in negative contexts (idem, 160). These findings suggest that the media and society tend to mention minority peoples primarily in relation to social problems, ranging from lack of education to various kinds of abuse. If this is true, it should be borne out in negative adjective usage for all three groups in the current study.

In the work of Homoláč, Karhanová and Nekvapil (2003), the authors attempt to pinpoint which “rhetorical devices and genres” are normally used about minorities in spoken and written media (Homoláč, Karhanová and Nekvapil 2003, 3). One chapter (Homoláč, Karhanová and Nekvapil 2003, 100; 102) argues that when a Romany individual is the victim of a crime, journalists covering the story mention negative stereotypes, even in comparatively neutral texts. Their work does not, however, refer to the focus of the present paper, which is on the precise wording and terms used to describe minorities.

Elmerot (2016) demonstrates that adjectives suggesting prejudice occur more frequently adjacent to the term Cikán (22 per cent) than to the term Rom (9 per cent). For this reason, the latter, more neutral, term will be used in the current study for comparison with terms used to describe the Vietnamese and Ukrainian populations.

In their article comparing oral communication patterns between Ukrainians and Czechs, Podolyan and Nývltová say that recent Ukrainian migrants to the Czech Republic have undergone a “massively quick” assimilation process (Podolyan and Nývltová 2012, 313). These authors argue that the reason for this is the two nations’ shared linguistic Slavic and (for Western Ukraine) political Austro-Hungarian roots.

4. Previous ethnic minority research

Many non-linguists (often researchers in journalism or communication) in various Western and Central or Eastern European countries – including Sweden (Brune 2004), the UK (Cheregi 2015), Bulgaria (Bakalova 2013) and the US (Catalano 2014) – have studied minority groups. The current paper, however, focuses on the Czech Republic and therefore the research referred to here also focused on that particular country.

4.1. Research on Roma

As can be deduced from the above, the Roma have a peculiar position in the Czech Republic, being often mentioned in the context of ‘distant others’ like people from other countries and continents (Zalabáková 2012, 83), despite having lived in the region for centuries. Before 1989, the Roma were also othered in public (Czechoslovakian) discourse, as shown for example by Alamgir (2013, 74–75), although the official view from 1948 onwards was that they were indeed able to change, and were not ‘unadaptable’ (nepřizpůsobivý) (cf. section 3.2 above). They were, however, referred to as ‘backward’ (zaostalý). As Alamgir illustrates (2013, 81), they were frequently compared with Vietnamese workers, though often using more pejorative words (Alamgir 2013, 82). It was not until the 1990s, especially when the Czech Republic aimed to become part of the European Union, that oral and written sources began to refer to Roma in a more serious and respectful way, and journalists and other writers began to use the term ‘Roma’ rather than ‘Gypsy’ (Schneeweis 2009, 147).

4.2. Research on the Vietnamese

Hantonová (2013, 67–71) researched the themes that appeared most frequently in coverage of the country Vietnam in four Czech newspapers in 2012. Despite its reliance on a limited range of source material, Hantonová’s study may offer a general picture of the situation. In the newspapers considered, the most common topics associated with the country and its people were the Vietnam war(s), economics, business and politics, although the 60 articles by the daily publication Právo also included seven articles on the Vietnamese minority and a further seven on criminal reports. In addition, Alamgir’s (2013, 79) earlier study of press coverage from the 1980s notes that the Vietnamese were described in an “infantilising” manner.

4.3 Research on Ukrainians

A number of studies, such as those by Zalabáková (2012, 81–82) and Leontiyeva (2006, 91), have suggested that Ukrainians are treated differently from other minorities in the Czech Republic due to the geographical and linguistic connection provided by their origins in a neighbouring, Slavic-speaking country. When they appear in newspaper articles, however, they are frequently portrayed as drunkards, murderers or thieves/robbers (Zalabáková 2012, 84–85; Součková 2015, 68; Leontiyeva 2004, 27). The events of 2013–2014 are barely covered in the data for this paper, so very few references to Ukrainians as refugees are included here. A further factor of interest, though, is that unlike the Romany or Vietnamese populations, the Ukrainians are mentioned in a religious context in previous research on minorities in the Czech Republic (Leontiyeva 2004, 26; Leontiyeva 2006, 37; Součková 2015, 69, 72). This may be something worth comparing with the results of the present study.

The findings above indicate that, following the events of 1989–1990, there appear to be two different depictions of Ukrainians in the Czech media (Leontiyeva 2004, 27), one portraying them as “guest-working” (drinkers and possible criminals) and the other presenting them as “non-problematic.” It will be interesting to observe whether the duality that Leontiyeva claims to see is also evident in the results of the current study.

5. Source material and method

Research for this paper was mainly conducted during February 2017, with additions during the spring and summer of 2018. The source material is the subcorpus SYN, version 4 (Křen et al. 2016),⁴ from the Czech National Corpus, abbreviated as ČNK. This material was chosen because it seemed best suited to the aim of conducting a quantitative study of the adjectives most frequently used before references to the chosen minority groups, thus providing a comprehensive picture of language usage in the most commonly read newspapers and magazines in the Czech Republic. The resulting data could then be used to analyse the extent to which these groups are treated stereotypically and othered. The SYN version 4 is a synchronic, referential corpus consisting of approximately 275 million sentences derived from 14 million texts. The SYN series is a set of monitor corpora (cf. McEnery and Hardie 2012, 6), and therefore well suited to research involving the scanning of large volumes of text, and the application of a statistical method to provide an overview of the everyday written use of collocations and other adjacent expressions (cf. McEnery and Hardie 2012, 7). Using a large corpus, it is also possible to see what is written for the “ideal subject” (Fairclough 2015, 78, 96).

The investigated corpus comprises articles from the national daily newspapers Mladá fronta DNES, Lidové noviny, Právo, Hospodářské noviny, Blesk, and Sport, as well as from regional daily newspapers (chiefly Deníky Bohemia and Moravia) and non-specialised magazines (Reflex, Respekt, Týden), all published during the period 1990–2015. The search was based on the query words ‘Rom’ (Roma), ‘Vietnamec’ (Vietnamese) and ‘Ukrajinec’ (Ukrainian) with an adjective in the position immediately before (to the left of) the query word. It should be pointed out that statistics should not be used to analyse how words are used in their context, but rather as a tool to categorise certain words or expressions within a larger body of source material in order to determine the language usage (Teubert and Čermáková 2007, 82).

In their 2008 article, Baker et al. explain that the combination of statistics, corpus linguistics and discourse analysis can help researchers avoid missing important facts, and that the inclusion of these three approaches “aims to do justice” to the material (Baker et al. 2008, 276). The method used here inevitably means that a number of search hits involving nouns instead of adjectives, which may be highly relevant for a socio-linguistic study or qualitative discourse analysis, have been ignored. Non-adjacent adjectives are also excluded from this research, even if they might have been of interest. The aim here is not to examine every relevant phrase, but rather to obtain a representative picture, while the statistics provide a clearer sense of detail.

The statistical processing of the material began with sorting the adjectives into two groups:

1. positive/neutral/negative adjectives.

2. stereotypical/non-stereotypical adjectives derived from the categories identified in point 1.

In the first step, the terms ‘positive’ and ‘negative’ are not used as gradable adjectives (Sassoon 2010, 141–42); instead, the terms are used as a measure of how the reader is likely to perceive them as conveying a positive or negative image of the people in question. In the second step, the adjectives were each classified as stereotypical or not, based on previous research. Since Roma, Ukrainians and Vietnamese are perceived differently, as noted above in the background sections, there are different stereotypes for each group, creating a need to make the categorisation more detailed. For example, according to previous research mentioned earlier in this article (e.g. Leontiyeva 2004, 27; Zalabáková 2012, 84–85; Hantonová 2013, 67; Homoláč, Karhanová and Nekvapil 2003), all three groups are – albeit to different degrees – considered to be both violent and frequent victims of violence, so adjectives like ‘beaten’ or ‘attacked’ were considered stereotypically negative, whereas any adjectives related to wealth were considered to be stereotypically positive, most notably in relation to Ukrainians. In the case where the Czech adjectives are already in the Czech Subjectivity Lexicon (Veselovská, Hajič and Šindlerová, 2014, 59), they are classified in the same way, and in other cases, the adjectives are individually classified partly by intuition and partly according to previous research mentioned here.

One methodological problem of this type of study is that some positive or neutral adjectives (such as přizpůsobivý = ‘adaptable’) placed before the targeted query words are written with a negation, such as nepřizpůsobivý, ‘unadaptable’, a word with a negative connotation (cf. Slavíčková 2015). The corpus filters do not (yet) distinguish between negated and non-negated adjectives, but instead list all words as non-negated in the frequency results.

Adjectives that are used sarcastically constitute another methodological problem. One example is the adjective slušný, ‘decent’, when placed before Rom, since there is a stereotypical view of Roma as ‘indecent’ (whatever definition may be used for that word). These kinds of adjectives are, therefore, more difficult to categorise and sort from a purely statistical point of view, and thus make the work more time-consuming. Both of these problems – negation and irony or sarcasm – were solved by manual analysis and then classification depending on the actual context.

During the sorting process, it became clear that Roma are frequently categorised by where they come from, so geographical adjectives such as ‘Czech’ or ‘living in Kolín’ were placed in a neutral category (non-stereotypical neutral). Ukrainians and Vietnamese, on the other hand, are often categorised by their age, and therefore age-related adjectives were placed in the same non-stereotypical neutral category. The latter categorisation, however, may be subject to change in a future study based on examples of age-related discourse when it comes to criminals or suspects (cf. the case of Uruguay in Julios-Costa 2017).

6. Results

6.1. The most frequent adjectives

Three separate searches were conducted, one for Rom, one for Ukrajinec and one for Vietnamec. Of all the hundreds (in the case of Roma, thousands) of adjectives found in the lemma search, the twenty most frequent adjectives were identified first. They are displayed here as absolute frequencies (Baker et al. 2008, 285; McEnery and Hardie 2012, 28) to indicate the words that could possibly be regarded as collocations with Rom, Vietnamec or Ukrajinec, respectively. The translations are my own, but verified through several Czech-English and Czech-Swedish dictionaries, and the parallel corpus application Treq (treq.korpus.cz).

From these twenty most frequent examples per group, it is easy to see that geography is more commonly used as a descriptor for Roma than for Vietnamese or Ukrainians, and that there are fewer other adjectives which are frequently used for Roma (‘Czech’ being ten per cent of the total). Even when the news articles concern crimes and police reports, there are several positive words to describe Vietnamese and Ukrainians, which makes for a more balanced report – not all members of these groups are considered criminals. However, the idea that Ukrainians drink more than average is shown by the fact that ‘drunk’ (opilý) is the third most frequent adjective used for Ukrainians. On the other hand, ‘working’ (pracující) and ‘employed’ (zaměstnaný, which may in some cases be translated as ‘working’) are also among the twenty most frequent adjectives used for Ukrainians, albeit ranked in fourteenth and seventeenth place, respectively. It is worth noting that the word for ‘rich’, bohatý, was discovered, when checked manually, to be used mainly about Ukrainians not living in the Czech Republic. However, since Ukrainians living outside the Czech Republic are also part of the linguistic othering process, these usages were retained as part of the analysis.

The patterns of othering found in previous research are confirmed by this study, although the number of geographical (in the case of Roma) and age-related (in the case of Ukrainians and Vietnamese) adjectives was more abundant than expected. When these neutral adjectives are ignored, mostly stereotypical negative or pejorative adjectives remain, from articles on criminality (‘injured’, ‘detained’, etc.) and intoxication (‘drunk’, and a few instances of ‘stoned’ in all three groups). Although nothing was found in the results concerning the religiosity of Ukrainians, there were many instances of words relating to success, from bohatý to favorizovaný. No mention of this latter depiction of Ukrainian immigrants has been found in previous research on the Czech Republic, unless that is what Leontiyeva (2004, 27) means by her dual categories of “guest-working” and “non-problematic” Ukrainians. Even if the Ukrainians characterised as successful live in Ukraine, these positive stereotypes still contribute to the general representation of Ukrainians.

Since the Roma have been living in the Czech lands for so long, where they are from seems to be of more interest than anything else. In contrast, the Ukrainians and Vietnamese are not as integrated into Czech society, and therefore their adjacent adjectives are much more likely to be references to their age. Nevertheless, the Vietnamese have ‘Czech’ as their third most frequent adjective, místní (‘local’) as their ninth most frequent, and zdejší (‘local, from here’) as their nineteenth most frequent. This is of course because these particular Vietnamese have been living in the country for a long period of time, whilst the Ukrainians are too proximate as an ‘other’ to ever be called ‘Czech’.

6.2. Adjectives before Roma

The above-mentioned pejorative adjective nepřizpůsobivý (noted as non-negated in place 19 in Table 1) appears in this material with a frequency of 0.7 per cent before Rom, compared to 0.07 per cent before Vietnamese and not at all before Ukrainian. (Percentages are used here to make the comparison easier than with absolute numbers.) This figure may also be compared to 1.23 per cent before the query word Cikán in previous research (Elmerot 2016, 18). The seemingly positive zaměstnaný (‘employed’ or ‘working’) is negated in the Roma context, except when it is used in the context of ‘black’, i.e. illegal, labour (načerno). Among the neutral adjectives, there are just two among the top ten that are not geographical (malý and další).

Table 1. The 20 most frequent adjacent adjectives for each term (absolute frequencies per term).able 2

	Rom (29,657)	Vietnamec (4,470)	Ukrajinec (5,335)
1	český Czech (2,937)	mladý young (380)	mladý young (294)
2	mladý young (2,217)	další other (141)	bohatý rich (168)
3	slovenský Slovak (2,159)	český Czech (128)	opilý drunk (135)
4	místní local (2,042)	jiný other (100)	další other (128)
5	olašský Vlax (1,582)	malý small (92)	třicetiletý 30-year-old (97)
6	malý small (704)	třiačtyřicetiletý 43-year-old (88)	mnohý many (93)
7	rumunský Romanian (542)	zadržený detained (71)	zraněný injured (87)
8	další other (540)	obžalovaný accused, culprit (60)	šestadvacetiletý 26-year-old (72)
9	ostravský Ostravian (537)	místní local (59)	pětadvacetiletý 25-year-old (66)
10	vsetínský from Vsetín (453)	mrtvý dead (53)	čtyřiadvacetiletý 24-year-old (65)
11	zdejší local, from here (449)	mnohý many (52)	známý famous (63)
12	samotný alone or the before­mentioned (433)	čtyřicetiletý 40-year-old (51)	starý old (60)
13	zaměstnaný employed (344)	samotný alone (48)	dvaatřicetiletý 32-year-old (59)
14	ostatní other, remaining (300)	pětatřicetiletý 35-year-old (48)	pracující working (59)
15	tamní local, from there (295)	ubitý beaten to death (48)	dvaadvacetiletý 22-year-old (58)
16	mnohý many (246)	známý well-known (47)	devětadvacetiletý 29-year-old (57)
17	jiný other (231)	zraněný injured (46)	zaměstnaný employed (56)
18	brněnský from Brno (226)	dvacetiletý 20-year-old (39)	sedmadvacetiletý 27-year-old (55)
19	přizpůsobivý adaptable (224)	zdejší local, from here (38)	třiadvacetiletý 23-year-old (55)
20	starý old (209)	sedmadvacetiletý 27-year-old (38)	osmadvacetiletý 28-year-old (55)
Share of total (per group)	56.2%	36.4%	33.4%

6.3 Adjectives before Vietnamese

Vietnamese people living in the Czech Republic are often considered to keep to themselves, which is represented in Table 1 by the (stereotypical) word samotný (either ‘alone’ or ‘him-/herself/themselves’). Words meaning ‘small’ are also classified here as stereotypical, due to the “infantilisation” of Vietnamese in the Czech media (see above, Alamgir 2013, 79). They are also considered neutral, partly because they are not included in the Czech Subjectivity Lexicon, partly due to previous research not mentioning anything about this being either a negative or positive factor in the representation of Vietnamese. (It must be noted that one specific word for ‘small’, malý, can also mean ‘child’, and indeed, approximately two-thirds of the contexts for the word ‘small’ in relation to Vietnamese were about children and education.) Of the various adjectives for “small” that do not mean child, the word drobný, ‘small in stature’, scored 3.5 per cent of the total number of adjectives found adjacent to Vietnamec, and there were also several instances of the adjective 45kilový (‘a 45-kilo-person’), which is an expression for a small adult, regardless of the person’s actual weight.

Since Vietnamese people are also mentioned frequently in articles about crime, words which clearly suggest they are the culprit (e.g. zadržený, obžalovaný) are classified as stereotypical, based on previous research – provided that they are not taken from an article on the Vietnam war – whereas ‘dead’ (mrtvý) is not, since it cannot be considered stereotypical to be described as dead per se. All these adjectives are however taken as negative, in line with the Czech Subjectivity Lexicon. Of the adjectives appearing in the table, známý (‘known’, ‘familiar’) is often found negated, neznámý, particularly in articles on crime, and this has been classified here as purely neutral. The seemingly positive zaměstnaný is also often negated, or used in the context of illegal labour, but this is not always the case, so it is again classified as neutral here.

6.4 Adjectives before Ukrainian

The most surprising finding was that bohatý, ‘rich’, appears so frequently. To put this in context, this is partly due to reports of the murders of “one of the richest Ukrainians” in 1996 and two “rich” Ukrainians in 1998, as well as some news on an oligarch. In 0.1 per cent of the cases, this adjective is exchanged for the synonym zámožný.

The fact that Ukrainians often succeed in sports events is obvious from the adjectives slavný (‘famous’), favorizovaný (‘favoured’ or ‘fancied’) and vítězný (‘winning’), which occur frequently in articles about sports and games. Finally, the adjective zaměstnaný is more often used in a positive context, and is less frequently negated (‘unemployed’) than it is with the Roma and Vietnamese.

7. Classification of the adjectives

Adjectives representing at least one per mille of the total frequency in the material were classified, and the classification of these as stereotypical was, as mentioned above, based on previous research in line with what had become evident from the results themselves. Once the adjectives had been sorted according to their classification, a subtotal was calculated for both frequency and percentage, as well as a ratio to demonstrate more clearly how the numbers compared with one another. The comparative results are shown in the following tables.

The word ‘young’ is in the top three for all three groups. Where geographical adjectives have a high frequency for the term Rom, age-related adjectives seem to be of the greatest interest in relation to Vietnamec and Ukrajinec, although the Vietnamese are sometimes noted as living in the Czech Republic. There are comparatively fewer stereotypical negative adjectives before Roma, which may signify the more neutral connotations of this term in comparison with Cikán. For the Ukrainians, 15 per cent of the total number of adjectives applied to them in the material are stereotypical and 65 per cent non-stereotypical. Both Vietnamese and Ukrainians are often described as ‘rich’, but the context is often negative when that descriptor is used for the former, and the usage is less frequent with regard to the latter.

Tables 3 and 4 reveal a considerable difference in the language used in relation to these minority groups in the Czech Republic. The ratio in Table 4 is the share of negative adjectives divided by the share of positive adjectives. The higher the number, the greater the negativity, indicating that Roma are depicted negatively five times more frequently than Ukrainians. The difference is statistically significant when calculated from the absolute values (Pearson chi2(10) = 4.0e + 03; P value = 0.000).

Table 2. Comparison of the most frequent adjective in each category, per group.

	Negative adjectives		Neutral adjectives		Positive adjectives
	Non-stereo­typi­cal neg.	Stereo­typi­cal neg.	Non-stereo­typical neu.	Stereo­typi­cal neu.	Non-stereo­typical pos.	Stereo­typi­cal pos.
Rom	zavražděný murdered (0.2%)	(ne)zaměstnaný (un)employed (1.2%)	český Czech (9.9%)	kočovný vagabond (0.3%)	slušný decent (0.6%)	vzdělaný educated (0.5%)
Vietnamec	mrtvý dead (1.2%)	zadržený detained (1.6%)	mladý young (8.5%)	malý small (2.0%)	zaměstnaný employed (0.5%)	bohatý rich (0.5%)
Ukrajinec	mrtvý dead (0.8%)	opilý drunk (2.5%)	mladý young (5.5%)	kontrolovaný monitored (0.2%)	pracující working (1.1%)	bohatý rich (3.1%)

Table 3: Percentage of each classification and total percentage of adjectives used less frequently than one per mille.

	Negative adjectives		Neutral adjectives		Positive adjectives		Less than 1‰
	Non-stereo­typi­cal neg.	Stereo­­typi­cal neg.	Non-stereo­­typical neu.	Stereo­typical neu.	Non-stereo­typical pos.	Stereo­typical pos.
Rom	1.6%	5.7%	72%	0.6	0.4	1.0%	18.5%
Viet­namec	12%	7.2%	64%	2.9%	2.5%	3.0%	8.0%
Ukra­jinec	1.3%	9.6%	56%	0.2%	7.7%	4.3%	19%

Table 4. Totals and ratios.

	Negative total	Positive total	Ratio negative/positive
Rom	7.3	1.4	5.2
Vietnamec	19.2	5.5	3.5
Ukrajinec	10.9	12.0	0.9

8. Concluding discussion

The research questions for the present study were:

1. To what extent are these people described in a stereotypical manner?

2. Are any of the groups (Roma, Vietnamese and Ukrainians) more pejoratively othered in the Czech media discourse than the others?

In light of the research findings, these questions can be answered as follows.

Previous research has indicated that these minority groups are often described using adjectives with negative connotations, and the results of the current study confirm this. For example, the Roma are perceived as ‘unadaptable’ (place 19 in frequency), ‘unemployed’ and victims of violence, whereas the Vietnamese are seen as victims, occasionally ‘unadaptable’ (235^th in frequency), but more frequently ‘tiny’ or just different. The Ukrainians are never represented as ‘unadaptable’, but are often described as ‘drunk’ or ‘armed’; on the other hand, they also appear as ‘rich’ and famous, or at least as celebrated sportspeople.

It is clear that the Ukrainians are more frequently mentioned in a positive light. As Podolyan and Nývltová (2012) argue, this may well be due to their relative kinship with the Czechs compared with the Vietnamese and Roma. It also seems clear that the Vietnamese are represented almost as negatively as the Roma in this source material, the ratio being 5.21 for the Roma and 3.49 for the Vietnamese. In contrast, the Ukrainians’ negative to positive ratio is 0.92, meaning that there are almost as many positive as negative adjectives for Ukrainians (had the ratio been 1.0, there would have been an equal number in both categories). The religiosity of the Ukrainians mentioned in previous studies does not feature among these results, however. All three groups have the adjective ‘young’ among the top three most frequent descriptors, which may be due to the bias of the articles and the tendency to depict young people as offenders (cf. again, Julios-Costa 2017, 6).

In a study like this, based on a large corpus, it is possible to see how power relations are reflected when the media address their “ideal subject” (Fairclough 2015, 78, 96), in this case the Czech general public. This study is based on millions of words, among which those actually placed next to one another were picked out and categorised; it could therefore be argued that the most frequently occurring combinations are possible collocations – ‘drunk Ukrainian’, for example. This method, which reveals the adjectives that occur most frequently next to the query words, illustrates the patterns arising in printed media discourse. The results are in line with previous research that aims to provide a deeper perspective and offer an alternative approach to analysing the language that a majority of the general public are exposed to on a daily basis.

Acknowledgments

I would like to thank the reviewers and editors of this article for valuable comments, as well as my supervisor Tora Hedin for guiding me through the mazes to get here, and finally professor Klas Rönnbäck for invaluable help with the statistics and structure.

ORCID

Irene Elmerot http://orcid.org/0000-0002-9809-8207

Notes

1 This article is a revised and updated version of a Stockholm university thesis from 2017 (see Elmerot 2017). Translations of Czech words into English are by the author.

2 https://www.czso.cz/documents/11292/27914491/1512_c01t01.pdf/7e9ea54d-13c8-4a41-bffa-bfe518d66ddc?version=1.0

3 The actual number of Roma persons in the Czech Republic is not clear. http://www.romea.cz/en/news/czech/patrik-banga-i-m-for-collecting-ethnic-data-in-all-areas-but-the-majority-can-t-tell-who-is-romani-2.

4 http://wiki.korpus.cz/doku.php/en:cnk:syn:verze4 – there is now a newer version available, but version 4 was the newest during the research for this article.