Language practice in the multilingual workplace: A Confucius Institute in Macau Círculo de Lingüística Aplicada a la Comunicación

. This article reports on an ethnographic study that investigated language practice in a multilingual workplace, a Confucius Institute in Macau. In the study we collected multiple data from five staff members through participatory observations, open-ended interviews, and field notes, to examine their language practice in the Institute. The analysis revealed that multiple languages were chosen to fulfill different purposes on different occasions. Specifically, Putonghua served as the working language of the Institute, English emerged as an auxiliary language along with Putonghua, and Cantonese was used as a group language for native-Cantonese speaker staff. This study also identified that the staff members adopted multilingualism (flexibly using different languages) and English as a lingua franca for communicating with learners outside the classroom, as important strategies for dealing with tasks in encounters with language diversity, divergence between spoken communication and written communication, and lack of multilingual competence. These findings suggest that the stakeholders in Confucius Institutes need to pay more attention to the language practice in these multilingual settings, and provide resources and support to enhance the staff’s bilingual/multilingual communication competence.


Introduction
The modern workplace is becoming increasingly globally connected and multilingual in a context of rising labor mobility and linguistic diversity (Angouri, 2014;Cheng & Zhao, 2019). Consequently, organizations and workplaces need to address their language policy and language practice in order to minimize communicative barriers, improve work efficiency and promote collaboration. In the context of the multilingual workplace, language policy relates to explicit and overt, as well as implicit and covert, management of language use, as well as the relevant ideology underlying the language use (Angouri, 2013;Kingsley, 2013), while language practice refers to "choosing to use one language or language variety rather than another in a particular instance" (Goldstein, 2001, p. 80) in multinational 1 organizations. Previous studies on language practice in multilingual workplaces have mostly examined international business communication in European contexts (e.g., Cheng & Zhao, 2019;Gunnarsson, 2013;Lauring & Selmer, 2012), especially language standardization (monolingualism) or language diversity (multilingualism) in multinational companies. Very little research has studied multinational organizations originally from non-English speaking countries-for example, against the changing social, economic, and cultural backdrop of China (Cheng & Zhao, 2019).
Emulating predecessors such as the Instituto Cervantes (Spanish), Goethe-Institut (German), Alliance Française (French), and the British Council (English), in 2002 the Office of Chinese Language Council International (known by its abbreviation Hanban, 國家漢語國際推廣領導小組辦公室, or Confucius Institute Headquarters) launched an initiative to establish Confucius Institutes in collaboration with local institutional partners worldwide. Aimed at promoting the learning and teaching of Chinese internationally and strengthening educational and cultural exchange and cooperation between China and other countries/regions, Confucius Institutes have been set up by the Chinese government in 162 nations/regions around the world since 2004 (Hartig, 2012;Li, 2018;Gong, Gao, & Lyu, 2020;Gong, Ma, Hsiang, & Wang, 2020). In this regard, Confucius Institutes have become a multinational organization, and their staff members are usually from diverse cultural and language backgrounds. At the same time, it may be argued that the global continuity and stability of Confucius Institutes depends largely on the effectiveness of communication between Hanban and the local partner institutions. To be specific, it is dependent on the staff members' language practice as the Confucius Institutes program expands internationally and demands increasingly intensified global communication (Lahtinen, 2015).
However, little research has considered Confucius Institutes as multilingual workplaces and explored the contextual, dynamic, and complex nature of language policy and language practice in these environments. To address these gaps, the present study aims to investigate the language practice at a Confucius Institute in Macau Special Administrative Region (SAR), China. Before describing how we conducted this inquiry, we will provide a succinct review of studies on language practice in multilingual workplaces in general and at Confucius Institutes in particular.

Language policy and language practice
Language policy has long been regarded as a multilayered and complex concept. For instance, Kaplan and Baldauf (1997, ix) suggested that it was in relation to "a body of ideas, laws, regulations, rules and practices intended to achieve the planned language change in the society, group or system". In line with this viewpoint, research on language policy has conventionally concentrated on countries/governmental bodies, as well as institutions and language planning issues in the context of globalization and/or post-colonial discourses (e.g., Pennycook, 1995;Spolsky, 2009;Tollefson & Pérez-Milans, 2018). Also, according to Shohamy (2006) and Spolsky (2004), language policy refers to patterns of language use in a given setting, as well as the ideology (e.g., values, beliefs) underlying the language use, highlighting issues of power and struggles. In their work, both explicit (top-down language management) and implicit (individual language practice) policies were examined in relation to language management and practice. In the same vein, Kingsley (2013) pointed out that many multinational corporations have an explicit policy nominating English as the official working language, but a number of languages are also used flexibly in diverse settings. Language practice is an essential interconnected component of language policy, and generally relates to people's linguistic behaviors and choices (Spolsky, 2004(Spolsky, , 2009). Specific to language practice in the multilingual workplace, it is concerned with staff members' de facto language choice and language use in multinational organizational domains (Angouri, 2014; Cheng & Zhao, 2019).

Language practice in multilingual workplaces
In the past two decades language practice has become an increasingly critical issue for many who work in multilingual settings (Cheng & Zhao, 2019;Goldstein, 2001;Spolsky, 2009). A limited but growing number of studies have focused on staff members' language choice and language use in multinational organizations, especially in the international business arena. Drawing on the concept of "top-down" and "bottom-up" language policies, Angouri (2013) investigated language policy and language use in one consortium of three multinational companies in Europe. It was found that the language policy in the multilingual workplace was constructed on flexible language use by both employees and managers. In particular, English served as the working language and was most frequently used alongside other local languages like Greek, Danish, and Japanese; employees constructed linguistic diversity as a given reality and undermined the social function of first language use in the workplace. In a more recent study, Cheng and Zhao (2019) examined language choice in a multilingual company, a branch of a Spanish bank in China, where Spanish and Chinese were used alongside English as common corporate languages. They revealed that although English was emphasized as a lingua franca in the work setting, native speakers were allowed to speak their mother tongues. At the same time, expatriate employees were encouraged to learn local languages in non-English-speaking countries, and employees in the company had no explicit belief in relation to the company's language management. They further noted consistencies and inconsistencies between employees' language choices and beliefs, and identified three levels of factors mediating their language choice, including job responsibility, team context, and individual practice.
From the perspective of managing language problems in multinational companies, Feely and Harzing (2003) discussed different language barriers to and their impacts on international business. These language barriers mainly referred to language diversity (the number of different languages the company has to manage), language penetration (the number of functional areas having to operate cross-lingual communication), and language sophistication (the complexity, refinement, and type of the language skills required). They also proposed a range of strategies to tackle language barriers, including choosing a lingua franca like English, flexibly using a number of languages (multilingualism), employing external language resources like translators, employee language training, nominating a single corporate language, adopting linguistically skilled personnel as language nodes, selectively recruiting linguistically skilled people, using expatriates in subsidiaries to link with corporate headquarters, integrating inpatriate personnel into company management teams, adopting machine translation, and limiting vocabulary and syntax rules to make texts more easily understandable. While Confucius Institutes are different from multinational companies in many regards, they both share a pressing need to achieve responsive integration in their host contexts through effective language policy and language practice, and they also encounter a wide range of language-related issues in complex multilingual and transnational work spaces.

Language practice at Confucius Institutes
From 2004 to 2018 the world has witnessed a dramatic expansion of Chinese as a second or foreign language, and the rapid development of Confucius Institutes are seen as an effort by the Chinese government to enhance interaction with other countries and facilitate participation in globalization (Lien & Oh, 2014;Starr, 2009;Wheeler, 2014). During this period, 57,378 Chinese language teachers and 54,310 volunteer Chinese language teachers were recruited and sent to 548 Confucius Institutes and 1,193 Confucius Classrooms (in primary and secondary schools) in 162 countries/regions across the world (Gong, Lai, & Gao, 2020;Gong, Lyu, & Gao, 2018).
The dramatic growth in the number of Confucius Institutes has generated increasing interest in the language issues they need to deal with. Lo and Pan (2016) even pointed out that the global development of Confucius Institutes might be undermined by a lack of "a high degree of transparency and effective communication with various stakeholders, both internal and external" (p. 520). For instance, informed by the McKinsey 7S framework model (Strategy, Structure, Systems, Style, Staff, Skills and Shared Values) used in management and organization assessment (Peters & Waterman, 1982), Lahtinen (2015) conducted a study on the Confucius Institute at the University of Helsinki and reported a range of challenges that the Institute had to deal with. These included: 1) absence of real, mutually agreed operation strategy among Hanban, the Chinese partner university, and the local university; 2) misunderstandings caused by the partnership structure complexity; 3) language barriers between the Chinese and Finnish systems; 4) work style conflicts; 5) team working problems between Chinese and local staff; 6) lack of teaching, management, and communication skills; and 7) no shared values between Chinese and Finnish staff. Specifically, in terms of language practice, Chinese staff members (director, teacher, and volunteer teachers) "mostly communicated in Chinese" (p. 216), and they had limited communication with local Finish staff members who did not speak Chinese. Consequently, the cooperation between Chinese and Finnish staff became a challenge because of the communication/ language barrier. Likewise, in the UK, Ye and Edwards (2018) found that language barriers were the main obstacle affecting Chinese language teachers who wished to communicate with local staff members. In particular, language barriers impaired their connection with local colleagues, and projected them as someone alien to the Confucius Institute where they worked.
In short, most relevant research has directly or indirectly reported language practice problems as a serious barrier, resulting in ineffective communication between staff members from different cultural and language backgrounds at Confucius Institutes. Therefore, Confucius Institutes worldwide, like multinational corporations, confront the same question: "how best to manage communications across the language barrier" (Feely & Harzing, 2003, p. 38). For this reason, the present study aimed to explore language practice in a Confucius Institute by addressing the following two questions: RQ1: What is the staff members' language choice in the Confucius Institute? RQ2: What is the staff members' language use in the Confucius Institute?

Research context
To address the research questions, we adopted an ethnographic approach to documenting staff members' language practice in the Confucius Institute at the University of Macau, a prestigious English-medium university, over a long period of time (Dörnyei, 2007). The Institute was established through collaboration between Hanban and the university in 2018. It mainly offers part-time elementary, intermediate, pre-advanced, and advanced Putonghua courses for non-ethnic Chinese learners who originally come from different countries and have diverse cultural and language backgrounds. From all the participants' self-reports, we noted that there was no explicit or overt language policy in the Institute.
Macau was a Portuguese colony for more than 100 years before returning to China as a Special Administrative Region in 1999. According to the latest population census in Macau, taken in 2011 (DSEC, 2011), 92.4% of the total population in Macau were ethnic Chinese, and 1.5% consisted of ethnic Portuguese. From 2001 to 2011 the diversity of the population was considerably increased thanks to economic development that drew people to invest, work, or study in Macau. 6.1% of the population was composed of other ethnic groups, such as Filipino, Vietnamese, Indonesian, European, and so forth. While Cantonese (83.3%) and Portuguese (0.7%) are the official languages of Macau, Putonghua (5.0%), English (2.3%), and other Chinese dialects (7.5%) are also spoken as common languages.
The first author worked as an assistant professor in Chinese education at the University of Macau, and had ample opportunities to observe the Institute staff members' language practice in an overt participant role. To conduct the study, we sought ethical approval from the university's Research Ethics Review Committee, and informed consent was obtained from the participants before we undertook observations on a daily basis. All participating staff were assured of confidentiality and anonymity.

Participants
The study was intended to cover staff members at the Confucius Institute who had different work responsibilities/ roles, as they may have diverse repertoires in their language practice. Five participants were voluntarily involved in this study: one program coordinator, three volunteer teachers, and one office secretary. Details of the participants can be found in Table 1.
Ms. Song (program coordinator): As a full-time university member, she was responsible for supervising the daily teaching routines for beginner, intermediate, and advanced courses in the Institute. She also had managerial roles in allocating teaching tasks for volunteer teachers, proposing annual Chinese cultural activities, and training novice teachers and teacher assistants. Ms. Song obtained a Master's degree in the US and had worked there for eight years as a Chinese language teacher at a college. She was highly proficient in English, with Putonghua being her native language.
Ms. Kong, Ms. Zhu, and Ms. Han (volunteer teachers): When the fieldwork was undertaken, three volunteer teachers were teaching at the Institute. They were recruited as full-time staff members through Hanban's screening system (involving paper exams and interviews). They had all obtained Bachelor's degrees in Teaching Chinese as an International Language in mainland China, and Master's degrees from universities in Macau. In addition, they had taught different courses prior to the data collection.
Jean (Office secretary): She was born in Hong Kong and educated in Macau and Canada. She was employed as an office secretary at the Institute under the university recruitment procedure, including English and Chinese (Putonghua or Cantonese) proficiency requirements. Jean managed routine work and handled daily correspondence, such as receptions, learner admission, and junior staff recruitment.

Data collection and analysis
The ethnographic work started in September 2019 and lasted half an academic year. The first author introduced the research topic to the participating staff verbally and in writing before data collection. He collected data from multiple sources, including participatory observations, open-ended interviews, field notes, email letters, and relevant university, faculty, as well as Institute documents, "to discern the meaning of conflicts, agreements, disagreements, and strategies" (Cleghorn & Genesee, 1984, p. 604). To avoid intrusion, he followed the normal order of events each week to provide data-gathering situations for three months. On average, general and structured observations were conducted on three days per week, in order to document formal and informal interactions between different people over a variety of administrative, managerial, and pedagogical events at various sites. Examples of observation situations included: staff offices for observing staff communications and interactions between staff and learners, classrooms for observing interactions between teachers and learners, the meeting room for observation of interactions at staff meetings and teacher collective lesson preparations, the corridor for observing brief staff contact and interactions between teachers and learners, and Institute events for observation of staff interactions and communications between staff and learners. A careful log of visits and all the field notes were kept on a daily basis in order to record these observations. In total, eleven episodes of observations and fourteen entries of field notes regarding language choice and language use were collected during the fieldwork in the Institute. Open-ended interviews were conducted with the five participants. As is typical of ethnography, the first interview with the staff members was informal and unstructured (Brown & Holloway, 2008), and two grand tour questions were adopted to build mutual trust and stimulate conversation (Leech, 2002): 1) Could you describe your job in the Institute? 2) What are your observations about the language choice and use in the Institute? The interviews were conducted at different locations within the Institute, and they lasted approximately 15 minutes. After the first interviews conducted in the first week of term, semi-structured interviews were carried out in the privacy of the participants' offices or classrooms. These interviews were framed around several topics and concepts that had emerged from the prior literature and previous interviews. During the semi-structured interviews, the following topics were addressed: 1) language choice and use experiences during the daily work in the Institute, 2) barriers concerning language choice and use in daily work, 3) strategies for or approaches to overcoming language barriers, and 4) achievement in interactions in and outside the Institute. Each interview lasted around one hour. Interview topics/questions were first reviewed and assessed by one expert and one researcher interested in language policy. Then, the topics/questions and the interview technique were pilot-tested with two Chinese teachers working in different Confucius Institutes in Asia. At the same time, relevant documents about the university, the faculty, and the Institute, as well as email correspondence, were scrutinized to make the interview topics/questions more pertinent. Interpretations for each item were checked, suggestions for wording were elicited, and the interview questions were revised accordingly. In total, nine informal interviews and five semi-structured interviews were conducted in the participants' native language (Putonghua or Cantonese) to minimize language barriers. All the interviews were audio recorded, transcribed verbatim into English, and double-checked for accuracy. Due to space limitations, this paper explains the research questions by presenting our interpretation of data derived from observations and open-ended interviews, supplemented with field notes. The field notes were descriptive in nature and captured anything that may augment the observation and interview data, in order to provide a comprehensive and contextualized account of the staff members' language practice (Wolfinger, 2002). The field notes thus yielded a thicker description of the context of language practice, and enriched the explanations of the research questions, as some tacit language practice may have been missed in the interviews or observations. Thematic analysis was adopted to analyze these data using Nvivo 12. The data were first categorized into organizational themes according to the research questions. Two organizational themes were used regarding language practice in this multilingual workplace: 1) language choice in the Confucius Institute, and 2) language use in the Confucius Institute. The excerpts under each organizational theme were then coded and recoded to generate concrete categorizations. Observation episodes, interview transcripts, and field notes were scrutinized reiteratively, and the data that struck the researchers as interesting or important to the theme were first coded using the participants' original words, and then similar codes were aggregated into analytic categories. The initial coding of analytic categories was then compared across the participants to find repeating content, in order to saturate categories with supporting evidence and cross-validate the emerging categories. Annotations and memos were adopted during the analysis to record immediate comments and reflexive thoughts on the data, and the annotations and memos were also used along with the field notes to assist the generation of analytic categories (Maxwell, 2005). Then the analytical categories were examined, with reference to the literature on language practice, to form the major theoretical categories that categorized the connections between the analytical categories (Gong, Guo, Li, Lai, & Wang, 2021). All the supporting quotations and interpretations were translated verbatim from Chinese to English by the first author, and checked with other researchers with expertise in both Chinese and English.

Findings
The analysis of the data identified that in the context of no explicit language policy in the Institute, Putonghua, English, and Cantonese co-existed and were chosen by the staff members for various communication purposes on different occasions. In particular, Putonghua emerged as the working language of the Institute, English was chosen as the auxiliary language alongside Putonghua, and Cantonese served as the group language of native-Cantonese speaker staff. The participants' accounts also revealed that in order to perform work responsibilities, they flexibly used the three languages (multilingualism) and adopted English as a lingua franca in communication with learners outside the classroom, during encounters that were characterized by different language-related factors such as language diversity, divergence between spoken and written communication, and lack of multilingual competence.

Language choice in the Institute
Language choice relates to a strategy of choosing to adopt one language/language variety rather than another during a specific interaction in a multilingual context (Goldstein, 2001). Emerging from the analysis, the data suggested that Putonghua, English, and Cantonese were variously chosen by the staff members and played different roles on different occasions in the Institute. The participants' different language choices will be elaborated using their own accounts and relevant episodes in the following sections.

Putonghua: the working language of the Institute
The analysis found that Putonghua was the most important working language in the Institute. It was used as the basic medium of instruction in the classroom, during Chinese cultural activities, and for daily communication among the staff. This finding was expected as the choice of Putonghua was primarily mandated by the nature and mission of the Confucius Institutes, one of whose most important aims is to promote the teaching and learning of Chinese language and culture globally (Ma, Gong, Gao, & Xiang, 2017). In addition, languages are usually chosen to signal interpersonal functions, such as identity, rapport, and solidarity, in the multinational workplace (Kingsley, 2013). In the Institute, Putonghua was also the language that all the participants could use, and for this reason it was a shared language that brought them together, as Ms. Han commented: [1] In our classroom, over 95% of the language we use is Putonghua. … Regarding the communication among staff, apart from our secretary, who is a local from Macau, other colleagues are all from Mainland China, so basically we communicate in Putonghua.
In her mind, in terms of interpersonal goals, Putonghua had a special status in the Institute as the most widely shared language, and it had emerged as essential because no other single language could connect all the staff. Consequently, Putonghua was regarded as the working language of the Institute by all the participants. At the same time, Putonghua constituted a gate-keeping mechanism for both Hanban and the Institute to selectively recruit Chinese language teachers. As the teacher participants unanimously noted, Putonghua Level 2, Grade A (普通話二級甲等) is a basic requirement for employment, and a trial lecture in Putonghua served as a necessary part of their recruitment interview.
The comment above was confirmed by Jean, the office secretary, who had not studied Chinese since graduating from primary school in Macau. Compared to her Cantonese and English, Jean's Putonghua was relatively basic, but she often chose to use Putonghua to communicate with other colleagues in the Institute. Because of the high numbers of native-Putonghua staff in the Institute, her language choice seemed to be a socialization process through which Jean integrated herself into the multilingual workplace (King, 2018;Kingsley, 2013). This language choice socialization could be seen as consisting of institutional, professional, and social/personal aspects (Roberts, 2010), and facilitated Jean to become a full participant in the Institute and even the university, as she noted: [2] In my daily work, I mostly use Putonghua. Because the colleagues I work with in the Confucius Institute are all originally from Mainland China and they are all native Putonghua speakers, I communicate with them in Putonghua here. We use English very little in face-to-face interactions. … Sometimes I even feel in this university all people's mother tongues seem to be Putonghua.
While it may be an illusion for Jean that "in the university all people's mother tongues seem to be Putonghua", it was true that the staff in the Institute and the university were generally bilingual or multilingual, and Putonghua normally served as the predominant language in the Institute. This was also illustrated in an observation episode involving the Institute Dean and learners who had completed their Chinese courses: [3] The Dean entered the Institute hall and delivered a welcome speech. We all know he obtains his PhD from the US and speaks English very fluently. On this occasion, he spoke Putonghua the whole time. The words and expressions he used were very simple, and all learners here could understand. The learners who had just completed beginner courses could understand his meaning.
This episode was not unusual, and it showed the official status and authority of Putonghua in the Institute. This was a purposeful example to send a clear message to the learners: on occasions of language contact, especially when important people are present, Putonghua is chosen.

English: the auxiliary language alongside Putonghua
The participating staff all reported that English was chosen as an additional or auxiliary language along with Putonghua in the Institute. In particular, the teachers highlighted the explanatory role that English played in their classroom teaching. Ms. Kong taught intermediate and advanced Chinese courses and commented: [4] English is usually used to assist learners in understanding, especially for the low-level learners. …When I teach some grammar points, I certainly tend to use English to explain the grammar rules, like grammatical positions, grammatical collocations and relevant issues of how to use the grammars.
Her account revealed that mixing English as an auxiliary language along with Putonghua could facilitate rapid and effective teaching and learning processes (Wang, 2019). In the same vein, Ms. Song called the English she used in Chinese classrooms "教學式英語" (literally, teaching-style English); her utterances were very short and simple, but this style worked effectively in her mind. In practice, the English she used was "controlled language" (Feely & Harzing, 2003), without conveying unnecessary instructional detail. During teachers' collective lesson preparations, she usually reminded other teachers to adopt this controlled English and emphasized its practicality, as noted: [5] I tell them (teachers) to master some simple English for teaching. For example, when teaching Pinyin b, p, m, and f, you need to tell learners how to pronounce these Pinyin correctly. All our learners can understand English. We don't tell them tongue position details, or they will be confused. We just tell them "round the lip" and "tongue back". Ok, and two or three words are enough and can work well. We don't need to say much English in the classroom. This will waste our teaching time and make learners confused. They (learners) don't need professional words.
Ms. Song's remark suggested that the auxiliary status of English had practical implications for both teaching and learning in the Chinese classroom.
In addition, English was chosen for its managerial function, in order to reduce communication barriers in the Institute. The participants unanimously noted that on most occasions, formal documents regarding course descriptions, Institute administrative rules, and cultural activity arrangements for the learners needed to be written in English. According to their accounts, English was chosen as the document language because it could ensure learners' understanding and improve the work efficiency of the staff in this multilingual context. Ms. Song felt it was necessary to choose English because "the learners with the low (Chinese) proficiency cannot understand our regulations", and also it "can provide a consistent guideline". Jean generally communicated with other colleagues in Putonghua. However, unlike the other staff members, as the only person whose working language was solely English when communicating with learners, she held a strong belief that English was the official language of the university: [6] Because they (learners) are all non-native Chinese speakers, I only speak English with them. … You must know English in this Institute if you want to support this program well. The learners all speak English. It (English) is the official language of the university, and you must have the competence in using English.
According to her account, English was mandated by the language policy of the university and the legitimacy of her language choice was thus enhanced by the top-down policy. Nevertheless, she reported that English was only used when interacting with learners, and Putonghua to a large extent promoted her professional participation in the Institute.

Cantonese: group language of native-Cantonese speaker staff
Compared to Putonghua and English, the analysis showed that Cantonese was not often chosen in formal or informal communication in the Institute. While Jean generally spoke Cantonese with other administrative colleagues in the general office of the faculty, she had very few opportunities to use Cantonese in the Institute. Its peripheral position was demonstrated in an observation episode involving Jean and the other two teachers: [7] Ms. Zhu entered the Institute office and said "Zou san" ("Good morning" in Cantonese) to Jean. Jean and Ms. Zhu happily talked in Cantonese for a while. Ms. Han came into the office when they were talking. Jean saw her coming and immediately asked "You shi ma?" ("May I help you?" in Putonghua). Ms. Zhu saw Jean's face and froze for a moment. Later, they all chatted in Putonghua.
On this occasion, Jean and Ms. Zhu originally talked in Cantonese and then switched to Putonghua when Ms. Han joined the conversation. This episode of language choice switching showed that Cantonese was rarely chosen when it was not a language common to all interlocutors. At the same time, it revealed that language choice sometimes served relational purposes, such as establishing rapport and enhancing relationships between staff members (Kingsley, 2013).

Language use in the Institute
Language use refers to people's actual linguistic behaviors in a particular situation in a multilingual context (Angouri, 2014). The participants' accounts showed that they undertook different patterns of language use to promote their work performances, construct professional identities, and build co-membership and rapport with the other staff members in the Institute, and different factors mediated their language use.

Flexibly use different languages
The participants unanimously reported that there was no explicit language policy or overt requirement for language practice in the Institute. In terms of language choice, Putonghua surfaced as the predominant language on occasions such as classroom teaching, Chinese cultural activities, collective lesson preparations, and staff daily interactions. English worked as an auxiliary language alongside Putonghua in classroom teaching and Chinese cultural activities, while Cantonese served as the group language for communication among native-Cantonese-speaking staff. In practice, language use seems to be more complex and needs to follow "a decision-making 'algorithm' which is grounded on the language that fits the interactional context as well as the linguistic skills and competences" (Angouri, 2013, p. 570). Consequently, individual staff members tended to negotiate according to common sense on the various occasions in which they participated, in order to establish the best match between language use and the specific situation.
The results indicated that all the staff members flexibly used different languages to deal with different language issues and to address the multifaceted reality in the Institute. The linguistic repertoire (and competence) of the target group and group members was normally the most important factor influencing language use in this multilingual context (Chen & Zhao, 2019;Kingsley, 2013). For instance, while Ms. Kong "had never sent English emails or used English in emails before", and had "no idea about the rules about writing English emails", she mostly replied to Jean's (the office secretary) emails in English because she thought English was Jean's and also the university's working language. However, after interacting with Jean face-to-face she realized that Jean could read Chinese, and after that she flexibly used Chinese or English subject to the email content. Similarly, Jean noted that she had never worked in a setting where Putonghua was widely and predominantly used. "To fit in with different interlocutors" on different occasions, she used Putonghua with colleagues in the Institute, English with learners, and Cantonese with colleagues from the general office of the faculty. Unlike the other staff, Ms. Song (a program coordinator) had worked in a Chinese program in the US for eight years, and she had rich work experience of communicating with people with diverse language and cultural backgrounds. This multilingual communication experience benefited her equal levels of English competence in reading, writing, speaking, and listening, and enabled her to engage learners deeply in and outside the classroom. She could freely switch to English if she realized that the learners felt uncomfortable with speaking Putonghua. These cases suggest that the participants' flexible language use often connected with their professional identity construction and relation maintenance with both colleagues and learners. Kingsley (2013) and Chew (2005) indicated that languages were used differently in written communication (e.g. reports, emails, slides) than they were in spoken communication (e.g. meetings, presentations). This divergence between spoken and written communication was frequently reported in the participants' accounts, and it also influenced staff members' language use, as Ms. Zhu mentioned: [8] My only struggle and concern is how to reply to the Institute secretary's emails. English or Chinese? Which should I use?
Different from spoken communication, English was predominantly used in written communication in the Institute, such as the above-mentioned documents written in English. Emails were mostly sent, received, and forwarded in English. While Ms. Zhu could speak Putonghua or Cantonese to the office secretary on face-to-face occasions, the working environment, where two kinds of communication patterns divided, increased her difficulty with language use. Also, having had "little chance to use English during pursuing my Master's degree", Ms. Zhu tended to switch to Chinese for content she could not express accurately, although she often wrote emails in English. Her language use was mediated by both the divergence in the institutional context and her own language competence.
In short, the absence of an explicit language policy in the Institute offered freedom for individual staff members to flexibly use the three languages. Multilingualism was a crucial strategy for the staff members to achieve transactional or interpersonal goals on diverse occasions. At the same time, both language diversity and linguistic repertoire mediated the participants' language use.

Use English as a lingua franca
In multilingual workplaces, English is generally adopted "to include and involve" (Kingsley, 2013, p. 545). Language diversity in multilingual workplaces has generally created a difficult managerial situation (Feely & Harzing, 2003;Tange & Lauring, 2009). In the Institute, even though Putonghua was used as the most widely shared language among staff and the working language for most transactional purposes, it was not a must for learners to speak Putonghua. Thus, the learners' different language backgrounds and Putonghua proficiency levels to some extent augmented the language diversity, and became a further factor influencing teachers' language use. Different from predominantly using Putonghua in the classroom, most teacher participants stated that when they communicated with learners in interpersonal settings outside the classroom, English often emerged as an essential lingua franca. For example, com-paring her experience of interacting with learners in the US to the situation in the Institute, Ms. Song felt that the language use for meeting with learners was very challenging: [9] I spoke much more Putonghua to learners in the US than I do here. In the US, I generally spoke Putonghua to them because they always tried their best to speak Putonghua to me. However, here I often ask myself which language I should use. Putonghua or English? … They (learners) don't speak Putonghua when they feel it is difficult. They tend to use English because they feel more comfortable. … I usually start the conversation with them in Putonghua. If they insist on speaking English, I have to speak English then. No choice! In her mind, Putonghua as the medium of instruction should be ideally used outside the classroom to increase learners' target language exposure. However, there was no explicit language use requirement for the learners in the Institute, and they tended to use the language they felt comfortable with. Accordingly, English was specifically used as a lingua franca between teachers and learners outside the classroom, and this lingua franca discourse between Ms. Song and the learners was typically characterized by a cooperative spirit to achieve successful communication (Kankaanranta, 2005).

Discussion
Based on participatory observations, open-ended interviews and field notes from ethnographic work, the present study examined a group of staff members' language practice in a multilingual workplace, a Confucius Institute in Macau. The findings emerging from the analysis indicated that communication is increasingly becoming a complex multifaceted issue in multilingual workplaces, with language practice being dynamically negotiated between staff members and the specific settings they were situated in, and also mediated by different factors. Overall, regarding language choice, the study found that Putonghua was chosen as the official working language of the Institute. It emerged as a predominant language on most occasions, such as classroom teaching, Chinese cultural activities, collective lesson preparations, and staff daily communications. English served as an additional or auxiliary language alongside Putonghua in classroom teaching and Chinese cultural activities, while Cantonese emerged as a group interpersonal language among native-Cantonese-speaker staff. This finding differed from the results reported in most previous studies on language choice in multilingual companies (e.g. Angouri, 2013; Cheng & Zhao, 2019;Kingsley, 2013), which unanimously identified that English was chosen as the official working language or an essential lingua franca.
Adding to what has been found in prior research on language practice in multilingual workplaces, the present study echoed Nickerson's (2005, pp. 370-371) note that "the communication event is often considerably more complex than the label of English as a lingua franca would suggest". The finding that Putonghua was used as the predominant language concurred with Lahtinen's (2017) study. Language practice is normally influenced by internal and external factors (Cheng & Zhao, 2019). The nature and mission of the Confucius Institute primarily defined the languages it chose and used to fulfill the needs of international learners learning Chinese language and culture, especially in a region where Putonghua is not used as an official language.
To further reveal the dynamic interplay in the internal environment, factors can be categorized into macro-, meso-, and micro-levels (Guo, Gao, Shao, & Zhu, 2020;Zheng & Mei, 2020;Zheng, Lu, & Ren, 2020). The Institute environment involved macro-level factors like work responsibilities. As required by the Institute structure, staff members often needed to work in a team, which formed a smaller meso-level working community influenced by participants, topics, and context. Micro-level factors refer to individual linguistic repertoires, experience, and action, such as native language backgrounds, work and education experience, and learning goals. It also should be noted that although English is regarded as the official working language of the university, the negotiation of language practice in the Institute was rarely influenced by this top-down language policy. Thus, the dynamics of language choice and use in this multilingual setting, and their negotiation in the different communities and contexts, still warrant further research. More attention needs to be paid to multilingual organizations originally from non-English speaking or non-European countries/regions, and relevant research could provide a more comprehensive picture of language practice in the multilingual context. Moreover, rather than simply reveal a different reality of language choice in the multilingual workplace, we went beyond the scope of existing research to further identify that, in terms of language use, multilingualism and English as a lingua franca for communicating with learners outside the classroom were adopted as important strategies to address language-related factors in the Institute, such as lack of multilingual competence, language diversity, and divergence between spoken and written communication. This finding was consistent with the study by Feely and Harzing (2003), which reported that functional multilingualism and lingua franca are two commonly-used strategies to address multilingual problems in multinational companies. Also, in line with Angouri (2013), it seemed that multilingualism was an ambiguous strategy (strategic ambiguity; see Eisenberg, 1984) that was unanimously accepted and adopted by the staff in the Institute, allowing the co-existence of multiple languages and the freedom for individuals to alter their choice and use on different occasions. In accordance with prior research (e.g. Cheng & Zhao, 2019; Kingsley, 2013), in most multilingual settings English, as "a default lingua franca" (Kingsley, 2013, p. 545), was adopted to include and involve, as well as to achieve transactional and interpersonal goals. In this regard, English and other local languages are needed for staff members' linguistic repertoire for many reasons, and lacking access to global or local languages can lead to possible exclusion between groups, especially in non-English speaking countries/regions.
In terms of factors mediating language practice in the multilingual workplace, this research not only showed that staff members needed to deal with language diversity and the divergence between spoken communication and written communication, as reported by previous research (e.g. Feely & Harzing, 2003;Kingsley, 2013;Ye & Edwards, 2018). Further, the study suggested that they should effectively address the lack of multilingual competence in the workplace. According to the participants' accounts, their educational and prior work experience in multilingual and multicultural workplaces may affect their professional and linguistic performances. In this regard, experience exchange between staff members is crucial, and staff with rich multilingual communication experience should play a mentoring role in facilitating novice colleagues. In addition, it was particularly noteworthy in this research that when facing issues related to language practice in the Institute, the participants usually adopted different strategic responses and employed social and personal resources to improve their linguistic competence, whether they were bilingually or multilingually skilled. For instance, Ms. Zhu enrolled in various online paid courses to improve her spoken English and enlarge her vocabulary. In order to overcome her "only struggle" regarding writing English emails, Ms. Kong repeatedly scrutinized the office secretary's emails and expected to learn "how to write (English emails) in the local style". She also "purposefully" searched for and learned from relevant online resources to gain systematic knowledge, in order to deal with issues related to email communication. In the same vein, Jean attended a part-time training course provided by the university to improve her spoken and written Chinese communication capability. In summary, the staff members made individual efforts and drew on linguistic resources they possessed to improve their own communication competence and further fulfill their transactional and interpersonal goals (Angouri, 2014). In this regard, individual agency seemed to play a critical role in language practice, and the participants demonstrated their participation in the professional community through different forms of investment (e.g. monetary and time). Thus, future studies need to take a closer ethnographic look at the dynamic interplay between staff's enactment of agency and their language practice (Eteläpelto, Vähäsantanen, Hökkä, & Paloniemi, 2013;Tao & Gao, 2017). At the same time, the stakeholders in the Institute should consider providing training resources and relevant support for staff to enhance their language skills (Gong, Hu, & Lai, 2018).

Conclusion
This study has explored five staff members' language practice in a multilingual workplace. Analysis of observations, interviews and field notes from the ethnographic research suggests that they chose multiple languages, Putonghua, English, and Cantonese, to fulfill different goals on different occasions in the Confucius Institute. In order to deal with language-related factors such as lack of multilingual competence, language diversity, and divergence between spoken communication and written communication, the staff adopted multilingualism and English as a lingua franca as crucial strategies for interacting with learners outside the classroom (Gong, Gao, Li, & Lai, forthcoming).
It must be noted that this research was conducted at a locally-run Confucius Institute in a university in Macau, and generalization of the findings to language practice at Confucius Institutes in other countries/regions and operated in other modes should be undertaken with caution. Our research drew its fieldwork data from one program coordinator, three volunteer teachers, and one office secretary, so the scope of our participant pool is rather small. Therefore, it would be helpful to carry out additional studies involving more participants with different language and cultural backgrounds, in order to map more comprehensive patterns of Confucius Institutes' language practices. Despite these limitations, however, we believe that the findings of this study show the significance of understanding the language practice in a multilingual organization originally from China. The present research also suggests that further studies are needed to explore how to help staff members communicate more effectively in multinational workplaces.