Accent stigmatization as a moderator of the relationship between perceived L2 proficiency and L2 use anxiety

1.1 Speaking an L2 with a non-native accent

In second language learning situations, it is common that the intonation and the phonological structure of the first language exert an impact on the pronunciation in the second language. A native-like accent might be seen as desirable but is very difficult, almost impossible, to reach even after many years of language learning (Derwing and Munro 2009; Gluszek et al. 2011). Indeed, those who speak a second language fluently very often do it with a non-native accent.

Although there may be individual differences in the degree of accent strength, a non-native accent is usually recognized immediately by a native speaker. In cases where accent signals a local variety or a dialect of a language, individuals who possess both the variety and the standard form of the language can change their accent to manage social distance between themselves and their interlocutors, a process of converging to or diverging from their speech (e. g. Soliz and Giles 2014).

Yet, accented speech often occurs against the speaker’s intention, and it can lead to adverse psychological consequences (Gluszek and Dovidio 2010a, 2010b). Accent has been shown to spark prejudice (Ng 2007), and speakers using a non-native accent often have to cope with negative attitudes, stigmatization and discrimination of their interlocutors (Derwing 2003; Gluszek and Dovidio 2010b). Decades of research applying matched guise technique (Lambert 1967) summarized in a recent meta-analysis, indicates that listeners generally evaluate a non-native accent less favourably than standard accents (Fuertes et al. 2012; see also Derwing 2003; Munro 2003; Derwing and Munro 2009).

However, communication is an interactive process, and attitudinal outcomes with the listener cannot be regarded as an endpoint. As recent studies by Gluszek and her colleagues have demonstrated (Gluszek and Dovidio 2010a; Gluszek et al. 2011) perception of negative attitudes or prejudice towards a non-native accent can recoil, and lead to serious negative emotional consequences such as communication challenges and feeling less social belonging for the speaker.

When considering the role of accented speech in communication and social settings, one cannot overlook the question of comprehensibility. Indeed, it is possible that strong accents can be tolerated or even celebrated and slight accents may be subjected to prejudice and discrimination. Derwing and Munro stress that “accents do not cause discrimination – the fault is with intolerant, often monolingual interlocutors” (Derwing and Munro 2009: 485). It is the intergroup climate that is most responsible for the types of emotional dispositions that are generated for non-native accents.

The present study aims to connect the stigma of non-native accent to perceived L2 proficiency and L2 use anxiety. More particularly, we will argue that being stigmatized because of a non-native accent can undermine one’s confidence with the second language, evoke anxiety and discourage communication.

1.2 The conceptual model

The conceptual model we test is built on the work of Clément and his associates (Clément et al. 2003) and integrates two perspectives of second language studies, the social context model of second language learning (Clément 1980; see also Clément et al. 2007; Sampasivam and Clément 2014) and the pyramid model of willingness to communicate in the second language (MacIntyre et al. 1998; see also MacIntyre 2007; MacIntyre et al. 2001).

Based on the tenets of the social context model (Clément 1980) and the convergent evidence from prior research (e. g. Clément 1986; Clément and Kruidenier 1985; Clément et al. 2003), we expect that frequent and pleasant contact with members of the second language group will increase L2 confidence. L2 confidence consists of a perception of competence to communicate in a second language and low levels of second language use anxiety (see e. g. Clément, Baker, Josephson and Noels 2005). Although higher levels of competence and lower levels of anxiety can be combined into a conglomerate of “self-confidence,” consistent with the approach taken in the pyramid model of WTC, we aim to assess the impact of contact on perceived L2 proficiency and L2 use anxiety separately. Specifically, we expect that frequent and pleasant contact with members of the second language group will increase perceived L2 proficiency and decrease L2 use anxiety.

Sampasivam and Clément (2014) argue that the relationship between perceived L2 proficiency and L2 use anxiety can vary according to the context (see also, Gardner and MacIntyre 1993). In bilingual settings, where the second language is present in the immediate environment, anxiety appears to be a more important predictor of communication in L2, whilst in monolingual settings, where language learners have little direct contact with speakers of the L2, perceived L2 proficiency more directly determines the willingness to communicate in L2. Perceived L2 proficiency and L2 use anxiety are constructs that interact with each other and their relationship is complex (MacIntyre 1995). Some research shows that anxiety arousal can bias the perception of L2 competence, such that anxious learners tend to underestimate their abilities (MacIntyre et al. 1997). However, when anxiety is defined as an emotional reaction to something gone wrong, or when it is aroused by anticipation of an inability to meet a future L2 challenge, it makes sense to treat anxiety as a consequence of perceived levels of L2 skills (e. g. Zhang 2000; Yim 2014). In line with this view, the present research treats perceived L2 proficiency as a precursor to L2 use anxiety.

In accordance with the studies reviewed above, we contend that whether the perception of high L2 proficiency results in relatively lower L2 use anxiety or not may depend on various factors. One of these factors is the feedback individuals receive from native speakers of L2 when communicating with them. In broad terms, feedback can be more or less supportive and may be related to various aspects of L2 communication (Marton and Vincze 2014). Clearly, positive feedback can reduce L2 use anxiety even in the case of the perception of low L2 proficiency, whereas negative feedback may make individuals more anxious also when they perceive themselves to be proficient in L2. Response to one’s accent in L2, especially if it is negative, may be an important aspect of feedback. This line of thinking generates a key hypothesis to be tested in the present study: L2 speakers’ feedback on the accent of their interlocutor can moderate the effect of perceived L2 proficiency on L2 use anxiety. Specifically, individuals who tend to receive more negative feedback on their accent from L2 speakers likely will be more anxious to use L2 even if they perceive themselves to be proficient in L2, whilst those who receive less negative feedback on their accent are likely to be less anxious even if their skills in L2 are lacking. Taken together, we expect that perceived L2 proficiency will be more closely related to L2 use anxiety in the case of little accent stigmatization than in the case of more accent stigmatization.

[Correction added after online publication 28 June 2016: Reference “(author, 2013)” was updated to “(Marton and Vincze 2014)” and the detailed reference was added to the reference list.]

Finally, as perceived L2 proficiency and L2 use anxiety are the most immediate predictors of willingness to communicate in L2 (MacIntyre et al. 1998; see also Clément et al. 2007), we anticipate that higher levels of perceived L2 proficiency and lower levels of L2 use anxiety will result in greater willingness to communicate in L2. The proposed model is depicted in Figure 1.

Figure 1:

The conceptual model. The moderating effect (i. e., the interaction of accent stigmatization with perceived L2 proficiency) is depicted as an arrow from interaction of accent stigmatization to the path from perceived L2 proficiency to L2 use anxiety.

1.3 The setting

Giles et al. (1977) offered a framework for describing the strength, or ethnolinguistic vitality, of language groups along three dimensions: demography, status and institutional support (for a recent review, see Bourhis et al. 2012).

The Hungarian language group got into a minority position after the territory of the present Slovakia was torn from the Austro-Hungarian Empire as a result of the peace treaties following the World War I.

At the time of the 1921 census, ethnic Hungarians constituted 635,000 or 21 % of the population. However, since that time, the size of the Slovak majority more than doubled, the number of ethnic Hungarians has been continuously shrinking (Kocsis and Hodosi 1998). The Hungarian minority has suffered the most substantial decrease in the last 20 years: while in 1991 there were 567,000 ethnic Hungarians in the country (10.8 %), in 2011 only 458,000 (8.5 %); this is a drop of almost 20 %, or 100,000 persons, during 20 years (Gyurgyík 2013).

The Hungarian language has no official status in Slovakia. Although the Act on the Use of Languages of National Minorities (2011) allows the use of Hungarian in settlements where the proportion of the Hungarian population reaches 15 %, the State Language Act strictly sanctions any divergence from the use of Slovak. Notably, as it is stated in the Act on the Use of Languages of the National Minorities “The public administration body and employees thereof are not obligated to have command of the minority language.” (Act on the Use of Minority Language, 1999/2011, Section 7, Paragraph 3). As a consequence, the official use of Hungarian is rather incidental and contradictory (e. g. Vass 2013).

There is a separate school system for the Hungarian minority, where the language of instruction is Hungarian. However, a quarter of Hungarian families regularly send their children to schools where the language of instruction is Slovak. In mixed language (i. e. Hungarian-Slovak) families this is the norm (Vincze 2010), and this tendency is seen usually as the most decisive factor in assimilation. The Slovak language is taught as a compulsory second language from first grade of primary school; in Slovak language schools it is not possible to learn Hungarian even as an optional subject.

Since 2004 the Hungarian minority has had a higher education institution, the Selye János University in the city of Komárom/Komárno. The university has three faculties (economy, education, and theology) and the language of instruction is Hungarian. Yet, a higher education degree is still far less common among ethnic Hungarians than among Slovaks. There is one Hungarian daily newspaper, “Új Szó”, and also Slovak public television and radio channels have a small amount Hungarian language programming (see Szabómihály 2003).

When it comes to intergroup relations, it is important to note that the relationship between Slovaks and Hungarians is burdened with several historical and political conflicts, which manifest not only with respect to the majority-minority context, but often also in the relationship between Slovakia and Hungary (e. g. Lanstyák and Szabómihály 2005, 2009).

All in all, the ethnolinguistic vitality of Hungarian can be considered low with respect to all three dimensions, and considerable assimilative tendencies are present.

1.4 The Hungarian accent

It is typical in both Hungarian and Slovak that stress is placed on the first syllable, therefore the two languages follow similar “rhythm”. Although the phonetic system of Slovak and Hungarian overlap to a great extent, there are a few sounds in the Slovak language that are not found in Hungarian. In addition, while consonant clusters are not rare in Hungarian, more combinations of consonants occur in Slovak than in Hungarian, and it is sometimes difficult for Hungarian speakers to pronounce these. As Slovaks and Hungarians do not differ much in externally visible ways, the Hungarian accent is a salient factor that makes the ethnolinguistic boundary perceptible and defines who is “in” and who is “out”.

Views on speaking Slovak with a Hungarian accent are rather negative within the Slovak majority and the Hungarian accent is typically stigmatized (Kontra 2013). This phenomenon is present at various layers of the society. For instance, in a debate in the national council of Slovakia, Anna Belousovová, a politician of the Slovak National Party criticised in her speech a politician of the Hungarian minority party because of his accent: “You are speaking with a terrible Hungarian accent; I was hardly able to translate some of your words” (Žemlová 2007).

The Hungarian accent is also often an object of humour in the mass media. The popular sitcom, Susedia (Neighbours) tells the story of two neighbouring couples, one of which is a Hungarian couple speaking with distinct Hungarian accent, which is the basis of many jokes (Nedelcu et al. 2010). However, Susedia is not the only TV show in which the Hungarian accent is mocked. As Kostovčík notes (2009: 178), “There are programs like SOS or Uragán, which build their humour on exaggerated archetypically-constructed characters in costume, with over-the-top accents mimicking the Hungarian or Roma language or various Slovak regional dialects, featuring in sketches with simple points, usually playing, again, on stereotypical traits of the social or cultural groups being mimicked” (Kostovčík 2009: 178).

The Hungarian accent is also a prevalent topic in everyday life. A present google search yields over 3,000 hits on the expression “Hungarian accent” in Slovak; and it appears to be a common topic in online discussion fora.

1.5 The present study

The present study is designed to test a model of the relationships among contact, anxiety, perceived L2 proficiency and WTC in the context of L1-Hungarian-speaking secondary school participants in Slovakia using the Slovak language as an L2. Our starting point is the theoretical model depicted in Figure 1, which proposes that more frequent and higher quality contact tends to be associated with lower L2 use anxiety and perceptions of higher L2 proficiency, in turn anxiety and perceive proficiency predict levels of WTC. In addition, we suggest that the relationship between anxiety and perceived L2 proficiency is expected to be significantly moderated by L1 (Hungarian) accent stigmatization.

2.1 Participants

Self-report questionnaire data were collected in Hungarian secondary schools in November 2014 in Galánta/Galanta, Nagykapos/Veľké Kapušany and Királyelmec/Kráľovský Chlmec (N=310). 62 % of the respondents were boys, 38 % girls. The average age of the participants was slightly less than 17 years (M=16.71, SD=1.22).

2.2 Measures

Contact. The quality of contact was measured with six 5-point items (e. g. “My contact with Slovaks is usually enriching.”) adapted from Clément and Baker (2001). The scale had a good reliability, α=0.84. The frequency of contact was measured with three items based on Islam and Hewstone (1993); respondents were asked how much contact they had with Slovaks within their family and friends, how often they were engaged in informal conversations with Slovaks, and how often they visited the home of Slovaks. The reliability of the scale was good, α=0.83. Higher values indicate more frequent contact and better quality of contact.

Perceived L2 proficiency. Respondents were asked to indicate on how they evaluate their proficiency on Slovak with respect to the four skills (speaking, listening, writing, and reading). A response format from 1 (poor) to 5 (excellent) was used. The measure had a good reliability, α=0.78. Higher values indicate better proficiency.

L2 use anxiety. L2 use anxiety was measured with eight 5-point items (e. g. “When I make a telephone call, I get mixed up if I have to speak English”) adapted from Clément and Baker (2001) and Gardner (2010). The compound scale had a good reliability, α=0.88. Higher values indicate greater anxiety.

Accent stigmatization. Accent stigmatization was measured with five 5-point items from the accent stigmatization scale (Gluszek and Dovidio 2010). The scale included sample items such as “I think that in certain situations I am being discriminated against because I have an accent” and “I feel like an outsider because of my accent”. The measure had a good reliability, α=0.71. Higher values indicate being more stigmatized by Slovaks because of a Hungarian accent.

Willingness to communicate in L2. Willingness to communicate was measured with seven 4-point items guided by MacIntyre et al. (2001). Respondents were asked to indicate how willing they would be to communicate in Slovak in certain situations (e. g. “Speaking in a group about your summer vacation”, “With a German/Swedish speaking acquaintance while standing in line”). The scale had a good reliability, α=0.89. Higher values indicate greater willingness to communicate in the L2.