Spanish language teachers’ pedagogical beliefs in Chinese universities

Understanding teachers’ beliefs is critical to improving educational practice. There is a significant need for research on foreign language teachers’ pedagogical beliefs, as these can exert an influence on students’ learning outcomes. This study draws on Kember’s (1997) theorization of teaching beliefs to explore the beliefs of Spanish language teachers regarding their pedagogy in Chinese universities. Twenty teachers from the Spanish departments in eight Chinese universities were recruited as participants, and a Q-methodology analysis with supplementary interview data identified three types of pedagogical beliefs: “integrated capacity-building”, “ELE learning outcome-focused”, and “social interaction-focused”. The findings suggest that ELE teachers in China no longer subscribe to the traditional teaching methodology which places themselves at the class center, and have instead progressed along a continuum from teacher-centeredness to student-centeredness. With a novel focus on Spanish teachers’ beliefs in Chinese universities, this study provides insights into the teaching of languages other than English (LOTEs) in China, and helps to expand the methodological repertoire in the field of linguistic research.


Introduction
Teachers' beliefs are a cognitive dimension referring to "what teachers know, believe and think" during the process of teaching and with regard to students' learning (Borg, 2003: 81). Previous studies have explored the influence of teachers' beliefs on their teaching practice (Kember, 1997;Borg, 2003Borg, , 2006Trigwell, 2012;Gilakjani & Sabouri, 2017) and have found that understanding teachers' beliefs is crucial to improving educational practice and teaching outcomes (Pajares, 1992;Richardson, 2003).
In the field of language teaching, the beliefs of language teachers have also been shown to have a significant impact on their teaching practice, which may lead to different learning outcomes or learning efficiency among students (Li, 2012). A dichotomy has been found in previous studies regarding the beliefs of language teachers, in terms of a confrontation between teacher-centered and student-centered approaches (Postareff & Lindblom-Ylänne, 2008;Amiri & Saberi, 2017). Teachers upholding the former approach tend to lay emphasis on language learning outcomes, while those endorsing the latter approach tend to pay more attention to interacting with their students and to change their perceptions.
In the Chinese context, foreign language teachers' beliefs have been described as dominantly teacher-centered (Hu, 2002;Camacho, 2014;Chen, 2015;Wang & Du, 2016), with a premium on reverence for education and a harmonious hierarchical relationship between teacher and student (Li & Du, 2013;Zhang & Liu, 2014;Poole, 2016). However, since the curriculum reform promoted in the new century, a tendency has been revealed among Chinese teachers to lean toward a more student-centered belief (Moodie, 2016;Yang, 2019).
Despite increasing attention directed towards foreign language teachers' beliefs in China over recent years, studies have mainly focused on English teachers (Lou & Liao, 2005;Qin, 2007;Chen & Chen, 2008;Gao & Qin, 2010), while teachers of languages other than English (LOTEs) remain a largely underexplored group (Guo et al., 2020). For example, the Spanish language has grown popular among learners in China due to increasing recognition of its value in the international community (Zheng, 2015;Lu et al., 2019a, b). Nevertheless, it is still considered as a xiaoyuzhong (a "small" or minority foreign language) in the English-dominant foreign language teaching enterprise (Huang, 2014). Moreover, teaching Spanish as a foreign language (español como lengua extranjera, ELE) faces additional challenges compared to English language teaching in China, among which are the outdated and incomplete curriculum (Sánchez Griñán, 2008) and the saturated labor market for Spanish major graduates (Yi & Zhao, 2016;Tan, 2018). These challenges may have motivated ELE teachers in China to adapt by updating their pedagogical beliefs.
Therefore, studies on ELE teachers' beliefs in China can advance our understanding of language teachers' pedagogical beliefs, particularly the beliefs of LOTE teachers who operate in different ideological and sociocultural spaces than English teachers. As such, the present study aims to answer the following question: What are ELE teachers' beliefs about teaching and learning in Chinese universities?

Literature review
A multiplicity of definitions and understandings of "belief" can be found in the literature on teachers' beliefs (Pajares, 1992;Borg, 2003;Tondeur et al., 2017). Richardson (2003) defined beliefs as psychological understandings, premises, or "propositions that are accepted as true by the individual holding the belief, but they do not require epistemic warrant" (p.3). According to Fives and Buehl (2012: 487), "teachers hold complex and multifaceted beliefs about a wide range of people and structures", which "have a reciprocal relationship with context and experiences". As such, teachers' beliefs can be seen as a dynamic and changeable system (Woods, 1996;Qin, 2007;Fives & Buehl, 2012). The current study focuses on teachers' pedagogical beliefs, including their conceptions about learning, teaching, students' and teachers' roles, as well as the teaching profession.
Teachers' beliefs about teaching are often classified into teacher-centered and student-centered orientations, which are typically presented as two opposing positions in that a teacher can hold either a teacher-centered or a student-centered belief. However, recent works have indicated that the student-centered approach and the teacher-centered approach are not mutually exclusive, and may coexist in a teacher's belief system (Postareff & Lindblom-Ylänne, 2008;Khan, 2015;Uiboleht et al., 2016;Vosniadou et al., 2020), being placed on a continuum (O'Neill & McMahon, 2005;Radu, 2016). One of the most representative frameworks in this regard is the multi-dimensional categorization by Kember (1997), which we adopt to interpret teachers' pedagogical beliefs in the present study. Kember (1997) sorted teachers' beliefs into five subcategories and placed them on a continuum. The teacher-centered approach features "Imparting Information" and "Transmitting Structured Knowledge", while the student-centered approach features "Facilitating Understanding" and "Conceptual Change/Intellectual Development" (p.264). These two approaches are bridged by "Student-teacher Interaction" as an intermediate category. Based on Kember (1997), Wu (2010) proposed a five-category classification in the Chinese context: Transmit Information, Prepare Them for Work, Develop Skills, Awaken Interests, and Promote Personal Development.
In addition, Kember (1997) suggested the possibility for teachers to simultaneously hold opposite beliefs and to experience changes in their beliefs along the spectrum over time. As the result of a mixture of contextual influences, teachers' beliefs, instead of staying firmly in one place, are more likely to lie at different points along the continuum (Brown et al., 2009). As Belo et al. (2014) pointed out, these beliefs cannot be presented simply as a dichotomy between teacher-centeredness and student-centeredness, since teachers may hold opposing beliefs as a result of their awareness of the students and the features of the course.
However, these studies mostly focus on English teachers, with limited attention given to the beliefs of LOTE teachers (Chen et al., 2020) such as ELE teachers. This neglect possibly results from the dominant position of English, which means that it accounts for most of the research on LOTEs (Dörnyei & Al -Hoorie, 2017;Chen et al., 2020). In the case of the Spanish language, the majority of the studies of ELE teachers' beliefs have been carried out in English-speaking or European countries (Delicado Cantero & Steed, 2015;Rodríguez-Izquierdo et al., 2020), probably due to the popularity of Spanish in these regions. The inadequacy of similar discussions in a number of other countries, including China, has been noted (Delicado Cantero & Steed, 2015;Avilés, 2017;Li, 2018). Furthermore, as suggested in numerous studies, language teachers' beliefs are subject to a number of contextual factors (Borg, 2003;Moodie, 2016;Poole, 2016;Sato & Oyanedel, 2019). Thus, given the differences in teaching environments between English and LOTEs, it is necessary to explore LOTE teachers' pedagogical beliefs. This study on ELE teachers in China will help to enrich our understanding of the beliefs of LOTE teachers.

Research context and participants
Twenty teachers (male=6, female=14) from the Spanish departments in eight Chinese universities were recruited as participants in the study. These eight universities are all prestigious comprehensive research universities in the 985 project (a program by which the Chinese government aims to build world-class universities in the 21 st century) and each has an established Spanish department, which indicates that all the participants were teaching Spanish major students. The demographics of the participants are presented in Table 1.

Data collection and analysis
The Q methodology was employed in the present study, alongside a semi-structured interview to collect supplementary data.

Q-methodology
A Q-methodology study can help to reveal a series of shared viewpoints on the topic of interest, serving as an objective measurement of subjectivity (Watts & Stenner, 2012). Moreover, the Q-methodology is suitable for studies involving small numbers of participants (Watts & Stenner, 2012), as in this study. We followed four steps as explained below.
(1) Develop the Q set, which is a sample of statements capturing possible opinions and perspectives related to the topic of interest. Adapted from Kember (1997) and Kember and Kwan (2000), five key themes were identified for analysis: instruction process (planning of teaching, teaching practices, teaching materials, atmosphere); teachers' focus (teachers' role, students' role, interaction, the conception of learning); teachers' assessment; source of experience/knowledge; and professional development. Under each theme, a selection of 5 to 10 items was explored. After piloting and eliminating overlaps and redundancy, these were then reduced to a set of 40 statements (there should be 40 to 80 statements, according to Watts and Stenner, 2012). (2) Develop the P set, which defines the participant group in a Q-methodology study. Watts and Stenner (2012: 73) suggest a minimum ratio of two Q set statements per participant, and therefore 20 ELE teachers were recruited to form the P set. (3) Q sorting. Participants were asked to complete the sorting task online within a recommended time of 15 to 25 minutes. After entering demographic information (name, gender, teaching age, affiliation, academic degree, etc.), they were instructed to perform a forced-choice task in a distribution table as shown in Figure 1, using the list of 40 statements that had been developed. In performing the task, participants were asked to drag each statement into one of 40 grids in a matrix headed with values ranging from -5 to +5, on basis of the degree of their agreement or disagreement (for each statement, there are eleven possibilities of placement on the scale). For example, if one of the participants mostly agreed with Statement 3, he or she may drag it into the grid under the value +5, and the number "3" will appear in the grid. The participant could also adjust their choices during or after each move. A completed set of tables consist of 20 Q sorts named Q1 to Q20.
(4) Factor analysis. The PQ Method (http://schmolck.org/qmethod/#PQMethod) was used to perform the factor analysis. First, the 20 Q sorts collected were entered into the program. After trials of 1 factor, 2 factor, 3 factor, and 4 factor solutions for the centroid extraction, a 3-factor solution was retained with an explained variance of 51%, a reasonable range of which is 35%-40% or above (Kline, 1994). Second, varimax rotation was carried out, in which the factor loading of each Q sort on each factor was presented. Using the equation 2.58 x (1÷√No. of items in Q set), a significant loading value of 0.41 was obtained (p<.01). Any Q sort with a single rotated factor loading above 0.41 might be defined as a member of this particular factor group. The Q sorts that possessed a significant loading for more than one factor were defined as confounding Q sorts.

Semi-structured interview
After the Q methodology analysis, retrospective semi-structured interviews were conducted to enrich the results and present a better explanation of the data collected. The results from the quantitative analysis were used to inform the design of interview protocols (see the appendix). In particular, questions were created to address the most salient beliefs that were identified in the Q methodology analysis. All the questions were open-ended and closely related to the Q set statements, so that the participants could express their opinions or make comments on the beliefs they held, especially in terms of why they held that particular belief. All the participants were invited to communicate with the first two authors individually by telephone in a quiet and private area to ensure their comfort. The interviews were conducted in Chinese and audio recorded with the participants' permission for later transcription. Each interview lasted from 20 to 50 minutes. Analysis of the interview data was conducted following a stepwise procedure (Yang et al., 2013): (1) Transcript browsing: reading the transcripts to get an overview of the content; (2) Annotation: relevant phrases and sections related to the investigated themes (instruction process; teachers' focus; teachers' assessment; source of experience/knowledge; professional development) were labeled and translated into English for data presentation; (3) Conceptualization and segmentation: interview data were categorized into different themes and aligned to the quantitative data collected using the Q methodology; (4) Writing-up: including the appropriate quotes in the corresponding analysis.

Findings
The varimax rotation produced three factors with an overall explained variance of 46%. Two confounding Q sorts were loaded on more than one factor, and one failed to load significantly on any of the three factors. Table 2 shows the classification of the participants. In this study we only focus on the participants loading on the three factors; the confounding and non-significant loading Q sorts are not the objective of the research. Based on the extreme statements shared by each group, the three factors were defined as "Integrated capacity-building" (Factor 1), "ELE learning outcome-focused" (Factor 2), and "Social interaction-focused" (Factor 3). These three groups differed considerably in terms of their attitudes or opinions about their role in language teaching practice, instruction methods, teaching content, and other related aspects. The next section presents and interprets the representative opinions and perspectives in each factor.

Factor 1: Integrated capacity-building
The distinguishing statements identified by z-scores (significance at p < .01; the z-score is "a weighted average of the values that the Q-sorts most closely related to the factor give to a statement" (Zabala & Pascual, 2016: 5)) for Factor 1 are shown as follows: Table 3. Distinguishing statements for Factor 1: Integrated capacity-building.
Although they were all teachers based in Spanish departments, this group strongly believed that proficiency in the language by itself could not determine students' future development, and that integrated capacity should play a decisive role (Statement 32). Some explained their choice as follows: Q12: Learning Spanish is important for a Spanish major student, but you cannot stop there, because sometimes language learning will force you to obey rules. I don't want to train groups of parrots. Being a teacher in a comprehensive research university, I must develop students' integrated capacity.
Moreover, participants belonging to this group emphasized the importance of students' self-planning and independent learning. They believed that students should be at the center of teaching, and the teachers' role was to help them only in complementary ways (Statements 17 and 8). These beliefs were confirmed by the interview data: Q10: University students are not children anymore and they should realize that they are in charge of their learning. Teachers and students should work together cooperatively. Students should have their study plans and learn how to solve problems independently. But, of course, if they do have some difficulties, I am here to help.
Apart from seeing autonomous learning as an important element in teaching, participants in this group also perceived that after-school activities and extracurricular practices were key to ensuring the effectiveness of language teaching. They did not consider "classroom knowledge transference" as the only or the most important path for students' language acquisition, and neither did they value language proficiency tests highly (Statement 30), as explained below: Q17: We have only a few hours in class, which is insufficient for developing the comprehensive ability to use a language. The best way to understand and overcome cultural differences and different thought patterns is through practice in real-time communication… Our Spanish proficiency test EEE focuses too much on the grammatical details while ignoring the role of the language as a communicative tool in daily life.
In addition, it is worth mentioning that the participants in Factor 1 did not subscribe to the idea that Chinese students experience favorable conditions and a conducive atmosphere to learn Spanish (Statements 27 and 40), and most suggested the necessity of improving the curriculum system and training programs for Spanish majors in the Chinese context (Statement 31). Some of them explained: Q14: The current job market is different from that back 10 or 20 years ago. Now the market is almost saturated and comprehensive talents are in demand, instead of those who can only speak the language. As a result, university training programs and curriculum systems should be adapted to meet the needs of society.
Q12: For now, our curriculum system is relatively 'thin and weak' compared with other majors in China or Spanish majors in the Hispanic world. We fail to offer students a 'rich menu'. For one thing, there is a lack of competent teachers; for another, our students start to learn Spanish from scratch… In terms of learning conditions and resources, I also believe that there is still room for improvement.
To summarize, participants in Factor 1 were defined as the "integrated capacity-building" type, since they had switched their focus from imparting knowledge in class to encouraging students to learn independently and to practice outside class. Moreover, these ELE teachers recognized the value of integrated capacity over language skills in their students' future development. Table 4 presents the statements identified by significant Q-sort value and Z-score, showcasing the representative beliefs of teachers sorted into Factor 2.

Factor 2: ELE learning outcome-focused
Contrary to Factor 1, participants in Factor 2 believed that Spanish learners in China have favorable conditions and convenient access to rich learning resources (Statements 27 and 40). Moreover, they did not feel there was any need for improvement in either the training program or the curriculum system for Spanish majors in China (Statement 34). Rather, they considered the overall environment in China relatively mature for learning Spanish, after years of development. The interview data illustrated this point: Q8: Recently, through my communication with Spanish teachers from different colleges and universities, I found that the training program at each institution has its characteristics, with different curriculum systems. I think we would be able to cultivate Spanish-speaking talents who are competent for different market posi-tions…I see it as very easy to find learning resources online, not only the translated publicity information of the Chinese government, as well as original materials from the Hispanic world. With regard to teaching practice, Factor 2 teachers highly valued knowledge transmission in the classroom, believing that classroom instructions and activities guarantee students' language acquisition (Statement 9). They considered the result more important than the process of teaching (Statement 6), and saw language level tests as a reflection of the effectiveness of their teaching (Statement 30). Moreover, compared with language fluency, they placed a premium on accuracy and reported that they would correct students' errors promptly (Statement 18). Beyond that, they strongly endorsed the use of national unified textbooks in their teaching (Statements 13 and 1), which differs considerably from Factor 1 beliefs (see Statement 13 in Table 3). Some of the Factor 2 teachers expressed their beliefs as follows: Q6: I always spend a lot of time preparing lessons to ensure that students can acquire knowledge in class, so classroom instruction is the most important thing to me.
Q8: The national unified textbook is compiled by the most authoritative Spanish teachers in China. The articles they selected have a high quality and match our national standard. Of course, I will also adjust teaching contents according to the characteristics and needs of different students. It is very important to have a national standard for an academic discipline, just like the EEE 4 or EEE 8 tests: if you don't have the certificate, who knows your Spanish level? Moreover, students' test scores are also a reflection of the effect of our teaching, and we need to prepare students to live up to the national standard.
Furthermore, participants in Factor 2 underlined the centrality of students' academic study, in our case ELE learning, and believed that Spanish major students should concentrate on Spanish learning instead of being distracted by other university activities (Statement 26). As noted by one participant: Q6: Nowadays, many of our students are not satisfied with learning Spanish only. They put a lot of time and energy into other courses or activities, which I don't think is very appropriate. Developing other skills and comprehensive abilities is very important. But as a Spanish major student, if you cannot speak Spanish well, how can you prove that you have a strong learning ability?
To recap, Factor 2 teachers' beliefs were found to be more traditional than those of Factor 1 in several aspects: choice of teaching materials, decisions on teaching content, attitudes towards exams and academic records, as well as views on language use. Factor 2 teachers seemed to be more inclined to adopt the teacher-centered approach (Kember, 1997), but in the interviews they insisted that they saw students as the center of their teaching and would address students' needs by making adjustments accordingly.

Factor 3: Social interaction-focused
Factor 3 was defined as social interaction-focused. Distinguishing statements shared by teachers of this group are summarized in Table 5: Table 5. Distinguishing statements for Factor 3: social interaction-focused.
In contrast with the results of teaching, participants in Factor 3 prioritized the teaching process and attached special importance to interaction, not only between teachers and students but also among the students themselves (Statements 36, 33, and 16). They preferred to conduct their class so as to share learning experiences or methods, rather than merely imparting information (Statements 38 and 23). The interview data shows that these teachers insisted on maintaining interaction with students, either online or face-to-face, in the classroom or outside class hours, because they saw interaction as the core of their teaching practice. For example: Q3: Maintaining interaction with students may help narrow the distance between teachers and students, which will probably make teaching more effective. I also see interactions between students themselves as very important and helpful. After all, they will have to walk out of college and step into society, and interaction is critical for social life.
Q7: I learned Spanish abroad and I believe that it is more important to teach students the learning method and to share with them my own experience, as well as to cultivate their cross-cultural thinking ability, than only helping them to master the language itself. Furthermore, participants who loaded on Factor 3 also expressed the view that students are at the center of the class and play a decisive role in the teaching process, and they disagreed with the view that teachers should act as conductors during their interaction with students (Statement 5). This was confirmed by the interview data: Q18: There is no doubt that students are at the center of teaching. I don't think I'm their leader or conductor; we teachers should have empathy and show care towards students' behaviors.
Consistent with Factor 1 and opposite to Factor 2, this group of participants did not favor the use of national unified textbooks, preferring to choose materials that meet the needs of society and adapt to changing requirements (Statement 1). As one of them said: Q16: To be honest, I rarely use Español Moderno now. I combine several domestic and original foreign textbooks, which are more useful for the students. Español Moderno pays too much attention to grammar. I hope that students can focus more on developing communicative skills.
In conclusion, teachers in Factor 3 gave more attention to improving the interaction and communicative skills of their students than knowledge transmission in class. They greatly valued cultivating students' interests and sharing their own learning experiences and methods. This group of teachers showcased a clear student-centered teaching approach, according to Kember's (1997) classification of teachers' beliefs, with the greatest importance attached to guiding students to realize their interactive value in society.

A summary of the beliefs of ELE teachers in China
The above analysis suggests that ELE teachers present different beliefs with regard to aspects of language teaching. In terms of the instruction process, participants in Factor 1 and Factor 3 argued against the use of national unified textbooks in the classroom, while participants in Factor 2 were supportive of their usage. Factor 1 teachers highlighted the importance of students' independent learning, while those in Factor 2 prioritized an immediate addressing of students' questions or errors. In terms of the role played by teachers in conducting teaching, participants in Factor 1 and Factor 3 considered their teaching as a facilitator for students' learning, while those in Factor 2 saw their instruction as providing guidance for students to acquire knowledge. Moreover, participants in Factor 3, in comparison with those in Factor 1 and Factor 2, were more concerned about teacher-student interaction and students' cognitive progress and conceptual change in teaching.
Despite these differences, a number of similarities could also be observed. First, the participants all expressed strong agreement with the global rise of multilingualism (Statement 22, Factor 1: +3, Factor 2: +5, Factor 3: +5); second, they strongly endorsed the student-centered approach, believing that the acquisition of Spanish largely depends on students themselves (Statement 11, Factor 1: +4, Factor 2: +4, Factor 3: +4); third, they strongly denied the advantages of native teachers over local ones in China (Statement 24, Factor 1: -4, Factor 2: -3; Factor 3: -5 ); and finally they disagreed with the view that the Spanish language, aside from being the students' major, would play a definitive role in shaping their future profession or life (Statement 19, Factor 1: -4, Factor 2: -3, Factor 3: -3). All these ELE teachers, while recognizing their main task as imparting knowledge of the Spanish language and its culture, shared a recognition of a multilingualism, and as such they recognized the value of English as a global language and some even encouraged students to learn multiple foreign languages (Factor 1), while teachers of Factor 2 insisted on the priority of ELE learning among their students.

Discussion
Under the framework proposed by Kember (1997) and Wu (2010), the classification of teachers' beliefs in the study can be presented as follows: Belief dimensions (Kember,1997) Belief types (Wu, 2010) Teacher-centered Student-centered

Figure 2. Dimensions of ELE teachers' beliefs in China.
Different from the findings of some previous studies (Hu, 2002;Li & Du, 2013;Zhang & Liu, 2014), it was found that a substantial number of ELE teachers in China (Factor 1+ Factor 3, 65%) no longer subscribed to the traditional teaching methodology which places the teacher at the center of the class (Huang, 2014;Sánchez Griñán, 2008;Camacho, 2014) and were beginning to incorporate new teaching methods and a student-centered approach. A similar tendency has also been observed in studies by Moodie (2016) and Yang (2019). Even with teachers in Factor 2, whose distinguishing opinions remained relatively traditional compared with those of teachers in Factor 1 and Factor 3, the student-centered belief was starting to take hold. Their own experiences of studying abroad and cross-cultural communication maybe encouraging these teachers to abandon the "prevalent hierarchical, authoritarian and pastoral role" to a certain extent (Wang & Du, 2016: 9).
The characteristics presented by participants in Factor 1-for example, insistence on the teachers' facilitator role in the teaching process, focusing on students' integrated capacity-building, after-school practices, and independent learning-showcased that these teachers likely correspond to the "facilitating understanding" dimension (Kember, 1997) and the "capacity-type" (Wu, 2010). Meanwhile, the beliefs held by participants in Factor 3 suggested a strong tendency toward the student-centered approach, as they exhibited mixed traits from the "interest-type" and the "value-type" (Wu, 2010). This group of teachers believed that teaching is not limited to imparting factual knowledge; instead, it should be a useful way to share experience and cultivate abilities, and they expected that their students would achieve interactive value in society through language learning. These pedagogical beliefs maybe related to the shrinking demand for language professionals in the saturated market (Yi & Zhao, 2016;Tan, 2018), where the type of talents required has changed from language-focused individuals to integrated and innovative teachers (Cai, 2020).
Regarding the participants in Factor 2, instead of staying firmly on the "transmitting knowledge" dimension (Kember, 1997), they showed an evolving process along the continuum from teacher-centeredness to student-centeredness, with an inclination to move towards the latter. This group of teachers, despite an emphasis on the importance of ELE learning, did not completely abandon the value of building integrated capacity in students. Previous studies have indicated that teachers may have conflicting beliefs (Tavakoli & Baniasad-Azad, 2017) and that student-centered and teacher-centered approaches can coexist in teachers' belief systems (Postareff & Lindblom-Ylänne, 2008;Khan, 2015;Uiboleht et al., 2016;Vosniadou et al., 2020). The contradictions within Factor 2 may also be regarded as a pervasive tendency on the constantly evolving continuum (O'Neill & McMahon, 2005;Radu, 2016) representing a shift from teacher-centered to student-centered teaching (Woods, 1996;Kember, 1997;Vosniadou et al., 2020). This change may result from the complexity of teachers' responsibilities, and pressure from the school educational polices they are subject to (Brown et al., 2009;Belo et al., 2014).
Although we used Kember's framework to explain the typical features of different types of teachers' beliefs, we have noticed that the pedagogical beliefs of ELE teachers in China are not found to feature a teacher-centered versus student-centered dichotomy. Their ideas and attitudes can be located at various points on the continuum between these two extremes as proposed by Kember, demonstrating a dynamic development-although Kember himself confirmed that "development or progression is unlikely to be a rapid or easy process" (Kember, 1997: 263). As teachers' beliefs tend to guide their teaching behaviors, they will ultimately produce students who are enabled to achieve desirable learning outcomes. Therefore, no matter what typical idealistic characteristics these teachers present and which cognitive dimensions they incline to at certain stages, their beliefs will inevitably change and develop (Chan, 2014). As long as these beliefs are conducive to students' learning, they can be regarded as effective means to facilitate the teaching process (Belo et al., 2014). In addition, viewed as a complex system (Borg, 2003(Borg, , 2006, beliefs are shaped by factors at different levels. In other words, the updating of teachers' beliefs will always accord with changes in their socio-cultural contexts, school environments, and personal experience. Only by keeping pace with students' needs and social changes will it be possible to achieve their teaching goals (Belo et al., 2014).

Conclusion
On the basis of the Q-methodology and the supplementary interview data, this study has investigated ELE teachers' beliefs about their teaching and their students' learning in the Chinese university context. We identified three belief types based on their different viewpoints on the instruction process, the teachers' focus, teacher assessment, their sources of experience/knowledge, and their professional development. These include the "Integrated capacity-building" type (Factor 1), the "ELE learning outcome-focused" type (Factor 2), and the "Social interaction-focused" type (Factor 3). A student-centered approach was adopted by teachers in Factor 1 and Factor 3, with an emphasis on the development of students' comprehensive abilities among the former and a focus on interactions with students and their communication and practices in society among the latter. Teachers in Factor 2, although influenced by some relatively traditional beliefs, demonstrated a dynamic evolution along the continuum from teacher-centered to student-centered teaching, with an inclination towards the latter. Furthermore, opposing beliefs from different dimensions can coexist in individuals' belief systems, and pedagogical beliefs are found to constitute a changeable and developing system that is subject to a mixture of contextual factors.
One potential limitation of the study is its small number of participants. This has been compensated by the adoption of the Q-methodology, the validity of which has been recognized in other studies involving subjectivity in linguistic research. Also, the surprising finding that teachers in China are no longer confined to traditional beliefs and teaching methods but are embracing the student-centered approach, which may apply to the teaching practices of many teachers, indicates significant room for further investigation. More research is needed to examine the impact of socioeconomic changes and the ensuing changes in talent demand on foreign language teachers' viewpoints on how to cultivate potential talents in university. The interaction between social context and teachers' beliefs, and how these beliefs change and develop over time, are possible directions for future research.