Investigating foreign language anxiety of Chinese Dongxiang ethnic group: a language conflict perspective

2.1 Foreign language anxiety

Foreign language anxiety has been defined as “learners’ distress at their inability to be themselves and to connect authentically with other people through the limitation of the new language” (Horwitz 2017, p. 41). As “the most widely studied emotion in second language acquisition in the past four decades” (MacIntyre 2017, p. 11), FLA and its contributing factors have been studied by scholars with various focuses. For instance, the correlation between FLA and learner-internal variables such as learners’ gender (Dewaele et al. 2016), personality (Horwitz 1986), language proficiency (Tao and He 2021), and other emotional aspects (Dewaele and Alfawzan 2018; Resnik and Dewaele 2020) have been investigated by scholars with the help of quantitative analysis.

Since the dynamic turn of FLA research around 2010 (Dewaele 2017), studies on the influencing factors of FLA have extended beyond learner-internal variables to language-related sociocultural factors. For instance, Dewaele et al. (2008) examined the communicative anxiety in the first language and FLA in the second, third, and fourth languages of 464 multilingual individuals, and found that participants suffered increasingly higher levels of FLA in languages acquired later in life. Based on the findings, the researchers concluded that the knowledge of languages and its frequency of usage, the socialization in a language, and the network of interlocutors were linked to learners’ FLA level. Thompson and Lee (2014) took Korean EFL (English as a foreign language) learners as the participants and found that the experience abroad could help to improve participants’ language proficiency and hence reduce their FLA. Similarly, according to Dewaele et al.’s (2015) study, experiencing study abroad, even for a few weeks, could significantly reduce participants’ FLA level in the classroom. Additionally, with 190 Latin American immigrants living in Australia as the participants, Blakeley et al. (2017) found that participants’ foreign language anxiety varied significantly across social contexts, with speaking with friends, with strangers, at work, on the phone, and in public rated as progressively more anxiety provoking. Shirvan and Taherian’s (2021) study had 367 undergraduates in Iran as the participants and found that the interacting elements, such as teachers’ feedback, peers’ reactions, performance of classroom activities, etc., were factors that affect learners’ FLA.

It is worth noting that the effect of being multilingual on learners’ FLA is not always clear-cut. On the one hand, some scholars have proposed that being multilingual was an advantage when learning new languages (e.g., Bensalem and Thompson 2021; Cenoz 2003; Dewaele 2010). For instance, with 106 adult students in the UK as participants, Dewaele’s (2007) study indicated that leaner’s knowledge of more languages was linked to lower levels of FLA in L2. With 953 French users as the participants, his research (Dewaele 2010) also found that multilingualism could enhance learners’ self-perceived communicative competence, and further led to less communicative anxiety in some challenging communicative situations. Thompson and Khawaja’s (2016) study used 156 EFL undergraduates in Turkey as the participants, and found that multilingual undergraduates got both higher levels of self-confidence and lower levels of FLA than the bilingual undergraduates. In a different context, with 354 students in Saudi universities as the participants, Bensalem and Thompson’s (2021) study also revealed that bilingual students experienced a higher level of FLA than their multilingual peers.

On the other hand, some studies have shown that there was no significant correlation between learners’ knowledge of more languages and FLA, and other studies even indicated that being multilingual might aggravate learners’ FLA levels. For instance, with 1,031 Chinese university students as the participants, Jiang and Dewaele (2020) investigated the correlation between sociobiographical variables (gender, ethnic group affiliation, geographical background and experience abroad) and participants’ FLA, and concluded that there was no significant difference in FLA between Han^[2] students and ethnic minorities across all the situations in the study (ethnic minorities in China are generally multilingual, learning English as their third language). Gong and Gao (2009) measured the FLA of 197 Tibetan students and 418 Han students and found that Tibetan students, as the typical trilingual learners in China, had higher levels of anxiety compared with Han, which was mainly reflected as communication apprehension in a foreign language classroom. Santos et al. (2017) investigated 532 adult participants in Spain and found that the communicative anxiety was higher in L3 (English) contexts for Spanish L1 speakers and Basque L1 speakers than in L2 (Basque or Spanish).

In sum, it appears that the influencing factors of FLA are not only limited to learner-internal variables, but also to some learners’ language-related sociocultural factors, such as learners’ geographical background and language variables, types of setting in which learners are interacting, etc. (Dewaele 2017; Jiang and Dewaele 2020). Meanwhile, the correlation between learners’ being multilingual and their FLA varies depending on their geographical background, and learners’ FLA levels are influenced by both outside and inside classroom factors (Jiang and Dewaele 2020). Accordingly, in this study, we assume that learners’ language environment plays an important role in determining whether being multilingual can aggravate or relieve learners’ FLA level, and attempt to interpret FLA of Chinese Dongxiang undergraduates from a language conflict perspective.

2.2 Language conflict

According to Dahrendorf (1959), conflicts, including some milder forms, are normal aspects of social life rather than dysfunctional occurrences, and social conflicts are multifaceted rather than congeal around one single issue. Among them, language conflict is a state of confrontation between two or more language groups in a bilingual or multilingual environment (Nelde 1998). Swaan (2001) proposed that language conflict is a disagreement generated among groups within a people, nation, or other political entities. It can occur not only between different countries, but also within different ethnic groups in the same country, or even among different branches of the same nation. He and Zhou (2010) put forward that in a broad sense, language conflict includes not only the fierce fights or wars caused by language problems, but also the discordant phenomena such as language competition or disputes in speech. Drawing on the descriptions of language conflict provided above, the present study proposes that language conflict can refer more broadly to implicit language competition or disharmony among and within certain groups due to any irreconcilable goals.

The past decades have witnessed the development of language conflict research from theory to practice. Scholars have paid close attention to language conflict in different countries and regions, ranging from Europe (Darquennes 2010), America (Merritt 2011), Asia (Al-Haq 1985), to Africa (Hashim 2006). Most of the studies explored the issue of language conflict within a certain group by analyzing their history (Cecillon 2007), culture (Merritt 2011), religion (Hashim 2006), education and language policy (Arutyunova and Zamyatin 2020; Li 2015). For instance, taking autochthonous language minorities in the European Union as an example, Darquennes (2010) examined the close relation between language contact and language conflict, and proposed that a language very often develops into a manifest significant symbol of social conflict in minority settings, even when it may not be the direct cause of the conflict. Davies and Dubinsky’s (2018) study presented detailed cases of language conflicts in Africa, Asia, Europe, and North America, and addressed the role language competitions, language ideologies and language shifts played in language conflicts. Arutyunova and Zamyatin’s (2020) study investigated both the official and public discourses in Russia, and elicited a conflict between the state languages’ mandatory instruction in schools and the daily language usage of Russia’s titular nationalities.

In China, documented literature tends to study language conflict from perspectives of language education or language policy and regard language conflict as a medium for promoting multicultural development. For instance, Xue and Liu (1999) pointed out that when cross-cultural communication occurs, although language conflict plays a negative role in social inclusion, it is beneficial to the identification of national cultural history and the promotion of national cultural reconstruction. He and Zhou (2010) proposed that as a ubiquitous social phenomenon, language conflict has both negative and positive functions. For the latter, it includes increasing the cohesion of a specific language group and enhancing their language awareness, maintaining or developing local linguistic and cultural diversity, reducing other types of conflicts and serving as a “social safety valve” (Li 2015). By comparing the phonetic, lexical and syntactic differences between the Jingpo language (a branch of the Sino-Tibetan language family) and English, Yin (2014) proposed that with a multilingual and multicultural background, the foreign language would compete with the learners’ mother tongue and local culture. Accordingly, some suggestions on local vocational English education were provided in the study to alleviate the negative influences arising from language conflict.

To sum up, in view of the controversy over the effect of learners’ being multilingual on FLA (e.g., Cenoz 2003; Dewaele 2010; Jiang and Dewaele 2020), and the negative impact of language conflict on learners’ foreign language learning in multilingual areas (e.g., Gong and Gao 2009; Yin 2014) as presented above, this study assumes that the implicit language antagonism or competition in learners’ linguistic repertoires and their language environments play crucial roles in determining whether being multilingual can aggravate learners’ FLA or not. In addition, as many studies on FLA were quantitative, they tended to apply questionnaires or scales to measure the correlation between learners’ FLA and other variables. In this study, by highlighting the interaction between learners’ FLA and their language environment (such as language conflict in their language learning environment), we take trilingual learners – the Chinese Dongxiang ethnic group – as the subject, and adopt a mixed methods approach to explore how language environment affects learners’ FLA from a language conflict perspective.

3.1 Participants

Participants of this study were non-English-major Dongxiang undergraduates. By using a purposive sampling approach,^[3] we recruited participants in four universities with similar admission scores in Lanzhou, an educational and economic center of Gansu province where most of the Dongxiang ethnic group lives. Given that the four universities in Lanzhou only offer college English courses in the first and second years, a major selection criterion is that participants be in their first or second year of university study.

In addition, according to our pre-survey, all undergraduates at each university were assigned to different English classes based on their English scores in the College Entrance Examination. Since the admission scores of these four universities were comparable, undergraduates in the same level of class were considered to have the same level of English proficiency. Meanwhile, the majority of Dongxiang undergraduates at each university took the “third” or “C” level English classes with similar English scores on College Entrance Examination. As a result, in the quantitative section of this study, both Dongxiang undergraduates at the “third” or “C” level and Han undergraduates in identical English classes were selected for an independent-samples t-test to investigate whether there were significant differences in FLA between Dongxiang and Han undergraduates.

Finally, we selected 194 Dongxiang undergraduates and 239 Han undergraduates in total, with their age ranging from 17 to 21 (mean = 18.9) (see Table 1).

In the second phase, namely the qualitative section, we took participants’ year in college as the variable and selected a higher FLA group according to the t-test in Phase One (see Table 3). Then we chose students who scored in the top 10 % of the high FLA group, taking account of their family location, language learning experience, and willingness to cooperate. Finally, 3 of them were selected as the cases in the second phase. With the consent of the participants, pseudonyms “student J”, “student L” and “student M” were used in the study. More information is presented in Table 2.

3.2 Instruments

The classic Foreign Language Classroom Anxiety Scale (FLCAS) from Horwitz et al. (1986) was adopted to find out if there was a significant difference between the FLA of Dongxiang undergraduates and that of Han undergraduates with the same English proficiency level, namely to answer RQ (1). The FLCAS consists of 33 items of a five-point Likert Scale from 1 (strongly disagree) to 5 (strongly agree). The total scores of FLCAS range from 33 to 165, with a higher score indicating a higher level of FLA. In the first phase of this study, we distributed the Chinese version of FLCAS (Wang 2003) to the participants. Internal consistency for this scale was measured and a Cronbach’s alpha of 0.89 was obtained.

In the second phase, semi-structured interviews (face to face and online) and observations were adopted to explore the specificities of the Dongxiang undergraduates’ language environment and how these specificities might affect their FLA (RQ 2 & RQ 3). For the semi-structured interviews, we mainly focused on understanding participants’ educational background, language environment, challenges, and motivations in their English learning process. Guiding questions (the original version is Chinese) such as “what are the major challenges in your English learning”, “do you have a feeling of burden or anxiety in English learning”, “what activities do you usually take part in outside school”, “what role do you think the Dongxiang language, Mandarin Chinese and English play in your life”, etc. were raised to the interviewees. The Aicloud intelligent transliteration software was adopted to transfer interview recordings into the written form. For the observations, with the help of notes taking and a digital camera, we investigated participants’ English classroom learning and the Linguistic Landscapes (Landry and Bourhis 1997; Shang and Zhao 2014) in their local communities.

3.3 Data collection and analysis

In the quantitative phase, a total of 456 questionnaires were collected, of which 433 were valid (194 from Dongxiang undergraduates and 239 from Han undergraduates). Twenty-three incomplete questionnaires were discarded. Descriptive data analysis and independent samples t-test were then carried out with the help of SPSS 27.

Data collection and analysis in the qualitative study follows the framework of “Learning lives,” which originates from the Phenomenology theory (Husserl 1992). As Van Manen (2003) noted, “phenomenon” embodies a specific “situation” humans experience in education and life. Education phenomenon focuses on educational behaviors or ideas that intersubjective produce in such “situation.” Therefore, education should be rooted in learners’ life-world. Accordingly, by combining the researchers’ voice, participants’ voice, and local school English teachers’ voice, we elicited Dongxiang students’ physical spaces in their learning-lives and divided them into four aspects: the school field, the after-class activity field, the community field and the family field.

The school field is concerned with participants’ classroom learning period, including their L1, L2 and L3 learning in the class; the after-class activity field is concerned with participants’ activities after the class, such as self-study, recreation, etc.; the community field addresses the language environment in participants’ hometown Dongxiang and their language behaviours there; the family field focuses its attention on participants’ language behaviours in family and their family members’ attitudes toward L1, L2 and L3. Subsequently, the observation and interview data were coded and analysed from the four physical spaces to explore the contributing factors to Dongxiang undergraduates’ FLA (see Figure 1).

Figure 1:

Analytical framework of Dongxiang undergraduates’ learning-lives.

On the one hand, we observed each case study participant three times in the school field (6 hours of classroom observation to get the participants’ learning behaviors and strategies). Observations in the community and family fields mainly focused on the language behaviors of participants, their friends and family members, as well as the Linguistic Landscape in their communities.

On the other hand, the semi-structured interviews were conducted in Mandarin Chinese face-to-face. Each interview lasted around 40 minutes and was audio recorded before being transcribed for analysis. Then, following the guidelines proposed by Braun and Clarke (2006), and Corbin and Strauss (2015), the data were analyzed inductively and recursively. In this process, both the open coding and axial coding were performed. The former was performed to generate a cluster of codes, which were then converted into different categories; the latter was performed to compare and integrate the codes to elicit tensions or conflicts in participants’ language environments. An inter-reliability test was conducted with two independent coders, yielding a Cohen’s kappa of 0.83; all transcriptions were given to the participants for member checking (Birt et al. 2016) to increase validity. Finally, with three to four rounds of interviews and the coding procedures above, we gradually enriched our data and finally formed case analysis based on the four fields in Figure 1.

4.1 Phase One: the FLCAS

Phase One is designed to answer RQ (1). In this stage, 433 valid responses were collected, including 194 from Dongxiang undergraduates and 239 from Han undergraduates. As we can see from Table 3, the results of the FLCAS showed that the Han participants averaged 94.43 while the Dongxiang participants averaged 111.11. When all Dongxiang participants were compared to all Han participants, the FLA of the Dongxiang participants was significantly higher (P = 0.000 < 0.05).

In addition, we also took the year-of-study of the Dongxiang participants as a variable to conduct another t-test, with the aim of finding the most “informative participants” (Patton 1990) for Phase Two. Further analysis in Table 3 showed that the first-year Dongxiang undergraduates (n = 101) averaged 107.12 while the second-year Dongxiang undergraduates (n = 93) averaged 115.44 on the FLCAS, and the second-year Dongxiang undergraduates had significantly higher FLA than the first-year Dongxiang undergraduates (P = 0.002 < 0.05).

4.2 Phase Two: the case study

Phase Two is designed to analyze Dongxiang undergraduates’ language environments in different fields (see Figure 1) and further to elicit the contributing factors to their FLA.

According to the quantitative analysis above, the FLA of sophomore Dongxiang was significantly higher than that of freshman Dongxiang. As a result, we used the former (n = 93) as a main study group and chose sophomore Dongxiang whose FLA score was in the top 10 % (n = 9). Furthermore, we also considered their family locations, English learning experiences, and willingness to cooperate, thus student L, student J, and student M (pseudonyms) were selected as the cases in the qualitative part of this study (see Table 2).

4.2.1 School field

Based on our semi-structured interviews and observations, we found that in Dongxiang County, students lacked adequate Chinese input and output in their early stages. Most of them were exposed to Chinese till entering elementary schools through the Min-Chinese bilingual education,^[4] and once the students learned some basic Chinese, they began to learn English with the help of Chinese. As a result, Chinese is both the teaching objective of Min-Chinese bilingual education and the medium of English-Chinese bilingual education, and it serves as a major instrument in the two bilingual education models (see Figure 2).

Figure 2:

The role Chinese plays in two different bilingual educations.

However, in the bilingual education mode depicted in Figure 2, Chinese and English compete with each other for a certain period of time in most primary schools of Dongxiang County. For instance, in the first 3 years, the Dongxiang language is employed as a medium of instruction. Subsequently, students will use Chinese as the medium to learn other school subjects once they have reached a certain level of Chinese proficiency (usually in the third year). Meanwhile, to comply with the local education policy, each primary school begins to offer English courses in the third year. The transfer of classroom language occurs practically simultaneously with the introduction of English courses, making Dongxiang students resort to their inadequate Chinese to learn English (see Excerpt 1).

Excerpt 1

When I was in primary school, I learned both English and Chinese. … Learning Chinese was OK, at least we used it sometimes in our daily lives. While learning English was challenging for me, we did not use it in everyday life. More importantly, we learned English through Chinese, despite our limited Chinese skills at the time.

(Interview with Student L, 20/07/2022)

Furthermore, one of our interviewees claimed that they had begun learning Chinese in primary school with Chinese pinyin. Chinese pinyin and English letters seem the same, but their pronunciation is different. In many cases, the interviewee found it hard to even distinguish Chinese Pinyin and English letters (see Excerpt 2).

Excerpt 2

After the third year, we began to have English courses. However, the Chinese Pinyin looked the same as English but sounded different. I was really confused, sometimes I couldn’t tell the difference between Chinese Pinyin and English letters. … I was under a lot of strain in English lessons, I was scared that if I mispronounced a word and my classmates would laugh at me.

(Interview with student J, 04/07/2022)

4.2.3 Community field

In the community field, we found that both the Dongxiang language and Arabic played crucial roles in the three participants’ daily lives. On the one hand, the population in Dongxiang County is homogeneous. Located in a mountainous area, most people there are Dongxiang minorities and the Dongxiang language is the only language being used in their daily lives. As a result, children in Dongxiang are less likely to acquire other languages. According to our interviews, the structural disparities between the Dongxiang language and English, as well as the asymmetry of cultural information between the two languages, have caused a negative transfer of Dongxiang students to learn English (see Excerpt 5).

Excerpt 5

I didn’t even know what the word ‘coffee’ was until many years later. The word ‘coffee’ does not exist in the Dongxiang language and we had never seen coffee before … There was also a problem of word order. The word order of English is the same as Chinese, for example, ‘I take meal’, but in the Dongxiang language it is ‘I meal take’, At that time, I felt the more I learnt, the more confused I was.

(Interview with student L, 06/09/2022)

On the other hand, Dongxiang people are Islam believers. Consequently, Arabic exerts a great impact upon Dongxiang communities. We observed that the most important activity in Dongxiang communities on weekends was to chant Qur’an and worship at the local mosque. In addition, commercial signs with Arabic elements were found everywhere in local economic and cultural centers of Dongxiang County and Linxia City (see Figure 3).

Figure 3:

Advertising billboards written in Arabic and Chinese in Dongxiang County and Linxia City.

The Linguistic Landscape not only enables readers to know the composition or the characteristics of language usage within a certain community, but also can express power and solidarity through language choices in a multilingual area (Shang and Zhao 2014). Therefore, the semiotic signs in Figure 3 show that in Dongxiang ethnic community, although Chinese holds the most sway, Arabic also has a place there.

Excerpt 6

In our hometown, Arabic is widely used in mosques and even shop signs on the street. These Arabic words might express our Islamic culture … As for English, we hardly see it in our daily lives, with the exception of school textbooks.

(Interview with Student M, 12/07/2022)