Language at work in the Big Four: global aspirations and local segmentation

Purpose – The purpose of this paper is to examine how professional service firms (PSFs) manage the linguistic tensions between global Englishization and local multilingualism. It achieves this by analysing the work of Big Four audit firms in Luxembourg, where three official languages co-exist: Luxembourgish, French, and German. In addition, expatriates bring with them their native languages in a corporate environment that uses English as its lingua franca. Design/methodology/approach – The paper combines the institutionalist sociology of the professions with theoretical concepts from sociolinguistics to study the multifaceted role of language in PSFs. Empirically, the paper draws from 25 interviews with current and former audit professionals. Findings – The client orientation of the Big Four segments each firm into language teams based on the client’s language. It is thus the client languages, rather than English as the corporate language, that mediate, define, and structure intraand inter-organizational relationships. While the firms emphasize the benefits of their linguistic adaptability, the paper reveals tensions along language lines, suggesting that language can be a means of creating cohesion and division within the firms. Originality/value – This paper connects research on PSFs with that on the role of language in multinational organizations. In light of the Big Four’s increasingly global workforce, it draws attention to the linguistic divisions within the firms that question the existence of a singular corporate culture. While prior literature has centred on firms’ global–local divide, the paper shows that even single branches of such firm networks are not monolithic constructs, as conflicts and clashes unfold amid a series of “local–local” divides.


Introduction
Professional service firms (PSFs) promise an effortless experience worldwide, as they claim to offer globally integrated services wherever their clients operate (Segal-Horn and Dean, 2009;Faulconbridge and Muzio, 2012;Muzio and Faulconbridge, 2013).Yet, as PSFs aim to coordinate their offices with a common logic in a "one firm" model, conflicts arise in local contexts, where practices might lack legitimacy and clash with local institutional logics (Boussebaa et al., 2012;Muzio and Faulconbridge, 2013).These local barriers are accentuated for professionals, whogiven their autonomydemand a say in their firms' strategy discourse, rendering hierarchical management and global coordination difficult (Faulconbridge and Muzio, 2007).Professions thus remain "nationally contingent" (Faulconbridge and Muzio, 2012, p. 138), and tensions in the global networks have been documented regarding the staffing of projects (Boussebaa, 2009;Boussebaa et al., 2012), the coordination of multinational audits (Barrett et al., 2005), and the local appropriation of accounting standards (Albu et al., 2014) and appraisal processes (Spence et al., 2015(Spence et al., , 2017)).
This study draws attention to a further aspect of PSFs, namely their process of organizational "Englishization" (Boussebaa and Brown, 2017).As English has developed and been adopted as a common language around the world (Neeley, 2012), PSFs have responded to their internationalization by adopting English as a formal corporate language.Yet, although the global expansion of capitalism and the "Englishization" of audit firms promote their increasing employment of an international workforce (Kyriakidou et al., 2016), it conflicts with locally embedded clients that are the main motivation and focus of many auditors (Anderson-Gough et al., 2000;Carter and Spence, 2014).Client-related "structural forces" thus explain both the global expansion of firms and pressures to meet local demands (Roth and Morrison, 1990), and the ways that PSFs manage their organizational language capacities and practices across or within jurisdictions are illustrative of the global-local disparity outlined above.Global audit work entails some difficult navigation through multilingual work, including the translation of specialized concepts and terminology (Evans, 2004;Evans et al., 2015), whereas translation is constructed and negotiated in situ (Kettunen, 2017).Likewise, expressions and demonstrations of professional scepticism rely on effective communication, potentially creating pressure to align with local linguistic practices.While English is the corporate lingua franca in the Big Four firms, individual auditors use different languages depending on the local social and professional context.As PSFs represent a "bundle of practices" (Muzio and Faulconbridge, 2013, p. 918), analysing the effects of a global language policy at the micro level seems most promising to explore the peculiar tension between formal organizational monolingualism and individual multilingualism in the Big Four audit firms.
In addressing these issues, the paper draws from interviews with 25 Luxembourgish audit professionals who are, or have been, employed at the Big Four firms.The choice of context is based on the co-existence of manifold nationalities and languages that allow us to identify phenomena less visible in other contexts.Luxembourg maintains three official languages: Luxembourgish, French, and German.In workplaces, French is the main language in most companies, with Luxembourgish and English following at some distance, and German being spoken in a minority of companies (Gotling-Dudding and Moser, 2011).These languages intermix, making Luxembourg truly multilingual.This status is unusual when appraised through prior literature on multilingual environments, which has often focussed on settings where different languages operate in distinctive geographic areas (Steyaert et al., 2011), or investigated the role of different languages in headquarter-subsidiary relations (Marschan- Piekkari et al., 1999) and cross-border merger and acquisition processes (Piekkari et al., 2005;Piekkari et al., 2013).In contrast to this static notion of multilingualism, the actors in our case are constantly confronted with, and negotiate, changing linguistic realities.The multiple and overlapping divisions between homogenous and heterogeneous language groups, native and non-native speakers, and variances in levels of fluency among actors provide a powerful lens into the role of languages in Big Four audit firms.
The study's theoretical starting point is the institutionalist sociology of the professions that explores the duality between firms' implementation of global management practices and local institutional logics (Kostova, 1999;Kostova and Roth, 2002;Muzio and Faulconbridge, 2013;Spence et al., 2015).Its main tenet is the existence of competing pressures on firms to deploy their resources worldwide, while being sensitive to the local contexts in which they operate (Roth and Morrison, 1990).This viewpoint is enriched by two sociolinguistic concepts, particularly used in the international business literature, to analyse the complex and heterogeneous flow of languages in Big Four audit firms.First, "language absorptive capacity" is used to account for the Big Four's collective ability to absorb communications in different languages (Piekkari et al., 2013;Peltokorpi, 2017), and thus to study their claim of offering consistent services across language domains.Second, the concept of "linguascape" is employed to capture "the discursive space in which an organization or any other actor frames and imagines how it can deal with its (de facto) multilingual composition" (Steyaert et al., 2011, p. 277), thereby revealing a firm's structuring into distinct language teams for each of its clients.We employ these theoretical concepts to understand the dynamic use of multiple languages, and to explore how individuals handle the continuous confrontation with familiar and non-familiar languages.
Building on these insights, this paper seeks to make three contributions.First, by studying the role of languages in Big Four audit firms, it adds to research on the tensions between these firms' global aspirations and the necessities of local contexts (e.g.Barrett et al., 2005;Boussebaa et al., 2012;Muzio and Faulconbridge, 2013;Spence et al., 2017).It is argued that the audit firms' "Englishization" does not permeate through to the interaction with their clients, as it is the client languages that mediate organizational relationships and create, define, and maintain intra-organizational structures and boundaries.As such, the claim of "language absorptive capacity" (Piekkari et al., 2013) is an integrative part of the firms' culture.It begins with the view that linguistic structuration based on client language is a management tool that enables an efficient servicing of clients.The claim is reinforced through organizational discourse and interactions, producing a culture of viewing foreign language resources as a commodity to be activated, whenever needed, and drawn upon as part of the firms' global network.However, because of inappropriately translated terms, the unproblematized adoption of ad hoc translations may have serious implications for the quality of audits (Evans, 2004).The claim of linguistic adaptability thus becomes a discursive device that supports the Big Four's global expansion, while accommodating the needs of local multilingualism.
Second, the paper reveals the tensions emerging from the organizational language policy of the Big Four firms by exposing the dynamic relations between intra-organizational "linguascapes".It is shown that "linguistic capital" (Bourdieu, 1991) can attract by creating social bonds within these discursive spaces, but also repel by introducing an "othering" between the groups.We answer Vaara et al.'s (2005, p. 621) call for further research on the "subjectivity and identity construction" that adopting English as a corporate language might entail.As this "plea remains largely ignored" (Boussebaa and Brown, 2017, p. 7), we argue that the grouping along language lines undermines the firms' homogeneity by nourishing distancing processes between language groups, both discursively and organizationally.Consequently, equality and diversity are only constructed on the outside, whereas the firms are internally divided and riven by different subcultures and linguistic groups.As a result, while prior literature has centred on the global-local divide of PSFs, we show that even single branches of such networks should not be viewed as monolithic constructs; they are full of internal clashes and different practices that unfold amid a series of "local-local" divides.
Third, the study shows the power implications of language use that emerge in interpersonal relationships.Such findings have been demonstrated in detail in the international business literature, namely that language affects communication and "imposes its own structure on communication flows and personal networks" (Marschan- Piekkari et al., 1999, p. 421).As such, language skills are a major factor for organizational and professional access and influence in Luxembourgish audit firms.We show how multilingual language fluency is an antecedent to, and also influences, intra-organizational social interactions.Linguistic capital eases client and team communication and broadens the professionals' scope for audit assignments as well as network opportunities.Alternatively, lacking such capital ultimately makes the respective auditor an organizational outsider, closing off the individual from the client and from team discussions.This power of exclusion is accentuated in the Luxembourgish context by the professional accountancy exam being

Language at work in the
Big Four administered in French, making this language a prerequisite to organizational advancement by installing a language-based "glass ceiling" into the profession (Piekkari et al., 2005).The paper proceeds as follows.The next section provides an outline of this study's theoretical background on the role of language in PSFs, which is followed by a description of the research methods employed.Subsequently, the empirical material gathered in the interviews is presented.The concluding section offers a discussion of the findings.
2. The sociology of language in PSFs 2.1 Global PSFs and "language absorptive capacity" Given their emphatic focus on profitability and client satisfaction, PSFs have followed the demands of their multinational clients for globally integrated and holistic services by aligning protocols, processes, and practices across their global networks (Segal-Horn and Dean, 2009;Faulconbridge and Muzio, 2012).The firms are promoting themselves as "one stop shops" of seamless services across jurisdictions wherever their clients are located or operate (Muzio and Faulconbridge, 2013).This has resulted in a "genuinely shared mindset" throughout the networks (Segal-Horn and Dean, 2009, p. 49), to be met and fostered by a "one firm" model that coordinates the ways in which local offices operate as part of a global professional network.Based on a common logic, PSFs organize their management, structures, practices, profit allocation, and remuneration, to align practices, processes, and values across offices (Muzio and Faulconbridge, 2013).
However, an increasing body of research has documented the "fundamental clash between a shared aspiration to be truly 'global' and the reality of local, institutionally conditioned material interests" (Boussebaa et al., 2012, p. 482).While the "one firm" model prescribes a bundle of practices, these do not necessarily correspond to the logics and practices in local settings, and so they lack legitimacy and eventually come undone (Muzio and Faulconbridge, 2013).An important notion for this global-local divide is "institutional duality" (Kostova, 1999;Kostova and Roth, 2002), which juxtaposes firms' striving for globally shared management practices and host countries' distinct institutional contexts.The pressure to conform to local profiles erects a barrier to the effective implementation of firm-wide practices, and only those practices are adopted that are considered appropriate in the national contexts.Local institutions affect global strategies, in that "the specific practices embodied in the 'one firm' model interact with local host-country institutions and their related practices to affect the success of internationalization strategies" (Muzio and Faulconbridge, 2013, p. 898).
In the accounting literature, Barrett et al. (2005) have shown the difficulty of coordinating multinational audits, particularly with regard to how global requirements are interpreted and appropriated in local sites.Such contradictions also relate to the interpretation of global accounting and auditing standards by local actors (Mennicken, 2008;Albu et al., 2014).Based on Boussebaa's (2015) argument that global PSFs (re)produce core periphery relations in the world economy, Belal et al. (2017, p. 145) have shown that the Big Four have "only a tentative presence" in the local market, as local societal factors affect the firms' work practices.Likewise, Big Four promotion processes tend to be informally locally appropriated, so that understanding the local "rules of the game" is vital for career advancement (Spence et al., 2015(Spence et al., , 2017)).
Despite this evidence of a global-local divide, the "one firm" model remains the aspiration of these organizations.As the firms promote a "culture of reciprocity", organizational strategy and rhetoric play a key role in emphasizing the benefits of working as an integrated whole, thereby employing a "language of a unity of purpose" (Greenwood et al., 2010, p. 177).This strategic use of rhetoric and communication corresponds to findings that the audit profession employs language to exercise organizational control and to socialize auditors into pre-defined roles (Dirsmith and Covaleski, 1985).In that spirit, Anderson-Gough et al. (1998) have examined the role of clichés and slogans in audit firms.Kosmala and Herrbach (2006) have argued that auditors' cynical distance from their firms may connote resistance and agency, but transforms them into "compliant" professionals.Again, auditors' client service discourse enables distancing processes that contribute to the sense of auditors' identity (Anderson-Gough et al., 2000;Carmona and Ezzamel, 2016).Discourse thus delineates objectives and affects individuals' ways of thinking, hinting at the performative aspects of language.
The international business literature has long acknowledged the "hybridization of language policies and practices" (Peltokorpi and Vaara, 2012, p. 829).Although language homogeneity promotes knowledge flows in the multinational global networks, the lack of a shared language and limited corporate language proficiency inhibit employees' ability to access this knowledge (Marschan- Piekkari et al., 1999;Mäkelä et al., 2007;Peltokorpi and Vaara, 2012).In multilingual settings, translation becomes a crucial task to mediate between the different languages.Peltokorpi (2017) empirically shows the effect of language-sensitive recruitment and language training to increase firms' linguistic ability to transfer and absorb knowledge.Piekkari et al. (2013) explore how individuals meet these demands as part of their daily work in multilingual environments.Despite the local firm's formal language policies, a central translation department, and technical translation tools, employees solved translation requirements using ad hoc mechanisms, particularly by asking colleagues but also by taking advice from their social networks outside the organization.Piekkari et al. (2013, p. 780) introduce the concept of "language absorptive capacity" as a firm's ability "to absorb and use information and knowledge that cross language boundarieseither entering or moving around the organization".Building on their work, we suggest that it is crucial for the Big Four to manage the "Englishization" of their professional audit work, while meeting the local clients' language demands.Overcoming this divide requires an active management of the language requirements of a global profession in a local context.Viewing this as PSF's "language absorptive capacity", we address ourselves to the following research question:

RQ1. How do Big Four audit firms manage the linguistic tensions between global
Englishization and local multilingualism?
2.2 Linguascapes as intra-organizational arenas and the power effects of language Language may be viewed functionally as a means of communication, but is generally seen as more than that, as being "central to the process of constructing organizational, social, and global realities" (Piekkari and Tietze, 2011, p. 267).Language is an important conduit through which social and organizational realities are constituted, maintained, or dissolved, becoming "an arena where organizational members communicate, while simultaneously providing a space for the processes of organizing upon which organizations depend" (Phillips and Oswick, 2012, p. 441).From this strand of literature, we identify two concepts that play a decisive role in PSFs: linguascapes and language as an empowering mechanism.
The concept of linguascapes echoes the anthropological work of Appadurai (1996), who argued that globalization results from local cultural processes that are characterized by different flows of, for instance, people (ethnoscapes), technologies (technoscapes), or ideas (ideascapes).Rather than an abstract process, globalization is something that is articulated and produced at given moments in a particular local space, "turning the global into a now and here, a globalization from below" (Steyaert et al., 2011, p. 277).Interpreting global processes as the product of local configurations allows us to discuss the identified discursive practices in multilingual PSFs in a broader context of harmonization and globalization.
Inspired by Appadurai's work, others have developed the notion of linguascape to describe the flow and dynamics of diverse languages that co-exist, such as within a nation ( Jaworski and Piller, 2008) or when entities merge (Steyaert et al., 2011).Following Steyaert et al. (2011, p. 277), we use the term in a socio-political sense "to point at the dynamics of accounting for language use and of their consequences for (im)balances among languages and the respective minorities and majorities these languages represent".Taking this perspective, a linguascape refers to the discursive space in which organizational members make sense of organizational multilingualism.Steyaert et al. (2011) show that the way that languages are prioritized is the result of a negotiation process among individuals concerning how a specific language is adopted, highlighting that language use is anything but "natural" or easily manageable based on language policies.As organizational members argue for and negotiate specific language use, linguascapes connote a degree of reflexivity, which is "inextricably linked with the indexing of specific language forms as ideologically charged with respect to the relative power position and identity of their speakers" ( Jaworski and Piller, 2008, p. 301).
In contrast to Steyaert et al. (2011), who associate different linguascapes with different companies, we conceive the multilingual audit firms as being made up of different linguascapes, which emerge out of their language policies.Such policies may relate to dealing with and coordinating audit engagements, and depend on the firm's client portfolio and the language preferences required to appropriately serve local clients (Anderson-Gough et al., 2000;Carter and Spence, 2014).Organizational language policies affect how language skills are managed and deployed, and may have an effect on hiring, staffing, and promotion decisions.Our intra-organizational perspective allows us to problematize the effects of different yet sometimes overlapping and clashing linguascapes, resulting in our second research question: RQ2.What are the linguascapes that mediate the intra-and inter-organizational relationships?
The second theoretical grid accounts for the implications for individuals that result from the navigation between, and confrontation with, distinct linguascapes.It is based on the idea that, in organizational interactions, language skills are resources that can empower and disempower, and that any linguistic differentials between organizational members hold power implications (Vaara et al., 2005).Having formal and informal exchanges in multilingual settings raises the issue of an individual's ability to navigate through these interactions, and their ability to converse becomes a form of capital."Linguistic capital" can refer to particular (upper class) accents or eloquence (Bourdieu, 1991), but may also cover an individual's ability to speak different languages.Conceptualizing language as linguistic or social capital, language skills become informal sources of power, where individuals with the requisite language fluency are able to create broader organizational networks, while those with confined skills cannot, or need to rely on language mediators (Marschan- Piekkari et al., 1999;Welch and Welch, 2018).This has implications for individuals' status and position in organizations (Neeley, 2013;Neeley and Dumas, 2016), as language skills become associated with competence, leading to a construction of superior and inferior standings among employees (Vaara et al., 2005;Śliwa and Johansson, 2014).These effects on power "emerge from the practices themselves", as such problems "are something inherent to the language policies and practices" (Peltokorpi and Vaara, 2012, p. 827).
The power of language thus accelerates in multilingual contexts, where non-native actors are confronted with native speakers.Berger and Luckmann (1967, pp. 143-144) argue that a foreign language will never take the self-evident reality of one's first language, which one treats with "affective quality".The reality created by a second language is more artificial and vulnerable, because it is "less deeply rooted in consciousness and thus more susceptible to displacement" (p.148).Conversing in a foreign language may close off, or leave inaccessible, some of the meaning-bestowing effects of language.As language is a key aspect of culture, linguistic relativity suggests that people speaking different languages think differently and that one's native language affects one's views of the world (Evans, 2004;Evans et al., 2015).Confirming this affinity with one's own linguistic space, Spence and Brivot (2011) exposed that, in its formative years, the professional accounting association of Montreal aimed to protect the cultural and linguistic habitus of its English-speaking members against Francophone influences by pursuing language-based exclusion strategies.
At the work practice level, the enactment of firms' language policies can be seen in the day-to-day work of audit teams and professionals.It is here where multilingualism is experienced and where work practices, communication, and behaviour in multilingual environments unfold and are enriched with emotions (Guénin-Paracini et al., 2014).While Segal-Horn and Dean (2009) assert that the "one firm" model requires instant communication and trust between organizational members, the international business literature suggests that language diversity and differences in language skills hold power implications that further undermine the vision of a global audit firm.Studying the ways in which individuals cope with translation tasks, Piekkari et al. (2013) also highlight the importance of social networks.While not explicitly discussed, the need for translation assistance may create a further source of, and demand for, social or linguistic capital of organizational actors, which leads us to our third research question: RQ3. Which tensions and power implications emerge from social interactions within the firms' distinct linguascapes?

Research site and research methods
The Grand Duchy of Luxembourg is a landlocked European country with a population of 550,000, neighbouring Belgium, France, and Germany.Notably, half of the country's resident population is not Luxembourgish and half of the country's working population commutes into Luxembourg on a daily basis from the neighbouring countries.Financial services are the main economic pillar, accounting for more than a third of the national income.The Big Four audit firms echo the country's heterogeneity and its status as a leading global financial centre.The firms are among Luxembourg's largest employers, with an overall staff of more than 6,000 people (STATEC, 2014).The intermingling of languages, cultures, and nationalities within a limited geographic space makes the Luxembourgish Big Four an appropriate setting for our research, as the firms are located in and represent themselves as multilingual contexts.One Big Four provided us with an overview of its audit staff, which, at the end of 2016, included 57 nationalities.Most auditors came from France (41 per cent), followed by Germany (18 per cent), and Belgium (17 per cent), with a small fraction from Luxembourg (5 per cent).The remainder came from other European countries (14 per cent) and the rest of the world (5 per cent).While the Big Four differ somewhat in their customer focus and client orientation, this study does not spell out these differences but focusses on the implications of the firms' multilingual work.
Our study is informed by semi-structured interviews with 25 professionals, all of whom are, or have been, employed at the Luxembourgish branches of the Big Four audit firms (see Table AI).Its main goal was to shed light on how these firms manage the multilingual environments in which they operate, which language policies they employ, and how firm members experience these environments.We deemed interviews most suitable to achieve our goal, as we were primarily interested in the ways in which "language use is productive, but also may represent phenomena (practices, meanings) at 'a short distance' from the site of the language use" (Kärreman andAlvesson, 2009, p. 1123).What is expressed in interviews then indicates the perceptions, feelings, and thoughts that interviewees normally experience, such as when discussing work-related topics in conversations with colleagues.
Given the lack of research on the role of language in PSFs, an interview approach further enabled us to pursue our broad objective, which involved recruiting interviewees from different hierarchical levels.To do so, personal networks and snowballing were relied on.The majority of interviews ( 16) were done with either juniors or interns, which reflects the firms' hierarchies.It also recognizes that lower-level auditors experience language-related issues on the firms' work floors more immediately, whereas senior auditors were more informative about the Big Fours' language policies and management of language diversity.To account for the cultural and demographic variety of the work environment, auditors were included from 12 different countries.Interviews were carried out between June 2016 and January 2017, lasting on average for 40 minutes.In total, 11 participants were interviewed face-to-face in Luxembourg and 14 via the phone.One of the co-authors, a German native, used the German or English languages to conduct all the interviews.With the permission of the participants, we recorded and transcribed all but one of the interviews.For the unrecorded interview, the researcher typed out the responses directly after the interview.Interview excerpts in German were translated into English only when writing this paper (marked below with "TR").Quotations from interviews have been edited slightly to improve readability, and interviewee and firm information have been anonymized.
The business context is familiar to one of the co-authors, who lived in Luxembourg for three years, where he supervised and taught some of the interviewees.Apart from the interviews, informal talks during that time informed our understanding and helped to make sense of the data.Knowing some of the interviewees also created a safe atmosphere in the interviews, thereby mitigating the potential reluctance of interviewees to share insights over the phone.As evidence of such a positive atmosphere, we noted that jokes were made, anecdotes related, and detailed information provided, regardless of the interview setting.In fact, interviewees described in rich detail their working experience at the respective audit firm, and particularly related their perceptions of language, translation, and communication in their day-to-day interactions within their audit teams, with superiors, as well as with clients and other parties.Even though some participants were interviewed in the busy season, all interviewees were eager to shed light on their work experiences.Many interviewees aimed to substantiate their observations by recounting stories and examples.Interviewees were encouraged to expand on these, as storytelling and critical incident descriptions were considered most informative about the organizational reality at these audit firms.
Data analysis made use of elements of both content and discourse analysis (Hardy et al., 2004;Phillips and Oswick, 2012).Both authors listened to the audio files separately and read the transcripts individually, before discussing them jointly.Taking an iterative approach to this analysis meant that insights from the interviews were reflected on and discussed throughout the data gathering process, both to enrich further interviews and to develop and augment theoretical viewpoints.Based on the extant literature, themes were identified in the transcripts that were subsequently interpreted through the organizational discourse and international business literature.Going back and forth between transcripts and theory, also as advised by fellow researchers and reviewers, led to the theoretical and empirical exposition as presented in this paper.That is, once our theoretical views as outlined above became crystallized, we went through all transcripts again to analyse interviewees' views on the role of language within the Big Four; between the main language groups or linguascapes; and for individuals' positions in Luxembourgish audit firms.The excerpts reproduced are representative of the themes and issues identified in the interviews, and were structured according to the research questions.The interviews were supplemented by archival material collected from public sources, such as the Luxembourgish websites of audit firms and contacts at the firms' human resources departments, publications by the professional institute of auditors (Institut des Réviseurs d'Entreprises (IRE)), and the local business press.
It is acknowledged that the research raises issues similar to the ones examined in this paper (Evans, 2018).First, interviewing participants in their non-native language might weaken their incident descriptions.While some interviewees were indeed less fluent in the interview language, they all work in a context where English is the corporate language.Using "communicative effectiveness" as a benchmark, "the primary purpose of language is to enable communication" (Ehrenreich, 2010, p. 410).In that spirit, all participants were able to express themselves in English, although one or two interviewees needed somewhat more time to do so.To the extent needed, we ensured that they were given full opportunity to relate their views and perceptions, occasionally asking them to clarify statements.Overall, we have no indication from the actual interviews, audio files, or transcripts that things remained unsaid or were skipped because of a lack of language skills.Instead, the individuals who were not as fluent in English seemed most eager to explain their views on issues with which they were confronted frequently.Second, interviewees may have been somewhat reluctant to talk to a German interviewer about stereotypes of German colleagues.Vaara et al. (2005, p. 621) label the French-German relationship as "problematic", and we acknowledge that we may have been unable to fully access the views of the French language group, being the dominant one in Luxembourg, with regard to their German colleagues.Yet, as this paper does not aspire to provide an inventory of stereotypes in Luxembourgish audit firms, this issue may only be reflected in the kind of examples provided, rather than preventing insights into relations between language groups.Third, translation of interview extracts adds an interpretive layer to the quotes and may change the social reality conveyed (Xian, 2008).We minimized such bias by reviewing each other's translations and, to the extent these differed, discussing their interpretations.While language issues inevitably arise in studies of multilingualism, we aimed to appropriately convey interviewees' experiences at Luxembourgish Big Four audit firms.
4. Navigating the global and the local: language absorptive capacity, linguascapes, and the power of language The empirical discussion begins by outlining the firms' strategies to cope with local language diversity.Next, the research turns to the firms' structuring into linguascapes, highlighting how language adaptability converts from an organizational capacity that facilitates client interaction into a force that creates and shapes organizational reality.Finally, the study illustrates how, through the dynamic, sometimes fluid, interactions between the linguascapes, language affects social interactions by having a range of power implications.

Language absorptive capacity: client focus and translation
The corporate lingua franca at Luxembourgish audit firms is English, having been installed at all Big Four by the mid-2000s (I No. 4, former director, Luxembourgish/French/ Italian).From that point on, it was decreed that "we speak English, and we work in English" (I No. 7,manager,French).The adoption of "the global language of business" (Neeley, 2012, p. 117) pertained to all communication, documentation, and working papers.English became the formal language in e-mails and professional trainings, and the individual's ability to communicate in English was added as a criterion to the firms' feedback and performance systems (I No. 10, senior, US).At the same time, PSFs remain bound to practices, norms, and cultural legacies at the local level (Muzio and Faulconbridge, 2013).Specifically in terms of language, the Big Four are embedded in a mash of linguistic plurality, resulting from their clients' preference for their respective native language, as "just because the auditor speaks English, they will not change their documentation" (I No. 2a, senior manager, German).
To manage the tensions between global Englishization and a local "client as king ethos" (Carter and Spence, 2014, p. 950), the audit firms have advanced a specific form of "language absorptive capacity", which is cultivated in organizational and individual responses to the Big Four's multilingual environment (Piekkari et al., 2013).The firms' ability to absorb communications in different languages is attained, first, by a strategy of structuring the firms into language-based groups and, second, by cultivating an organizational linguistic mindset that encourages individuals to draw on other firm members' language resources.These two managerial responses activate and reinforce the firms' capacity to cross language boundaries.
First, the firms' linguistic adaptability aligns internal structures with clients' linguistic profiles, because it is "the firm's policy to address the client in their main language" (I No. 2b, senior, Luxembourgish).As a result, this paper argues that the audit firms' pursuit of "Englishization" and globalization ends with their clients: When something is important, or we really want to make sure that the client is happy, the director will email them in French (I No. 10, senior, US).
The firms' overarching client focus goes beyond that shown by prior research (Anderson-Gough et al., 2000), in that it manifests itself in three distinct types of audit teams that exist at all Big Four firms: the majority of teams are French, consisting mostly of French and Frenchspeaking Belgian auditors, followed by a smaller number of German teams, as well as some international teams.Affiliation with these teams is based on mother tongue or language skills.The international teams consist of members with different nationalities and mother tongues, where English is indeed the working language in the audit engagements.In this way, the "imperative to be client-friendly" (Anderson-Gough et al., 2000, p. 1156) creates language-homogenous teams based on the client's language, which is echoed and reinforced in the firms' team staffing and recruitment procedures [1]: When you schedule [that is, request team members], […] you can pick if you want English, French, or German-speaking, so there is the option to pick [team members based on their language] (I No. 10, senior, US).
Organizing audit teams in line with the respective client language may be a natural response to local multilingualism, but it fragments the firms into distinct local units and introduces a "shadow structure" based on languages (Marschan- Piekkari et al., 1999).As local practices undermine the "one firm" model with its "language of a unity of purpose" (Greenwood et al., 2010, p. 177), relating to the client was said to work through language and "brings [with it], I think, this specialization in a certain language or certain nationality" (I No. 2a, senior manager, German).This was considered most important at a senior level, where speaking the client's language entails an understanding on a more profound level: The client interaction of course gets higher and higher, gets more and more, the more, let's say, you grow in the company.You start getting a client-facing position, where you also start to make marketing, for instance, in the respective country.And then, of course, it is relatively clear that, if you have a visit, let's say, for instance, in Germany or France and you meet your client, then it is most likely that you need to adapt to their language (I No. 2a, senior manager, German).
While such a client focus follows from the firms' language absorptive capacity as part of their local practices, this paper argues that it results in the client languages de facto creating, defining, and maintaining intra-organizational structures and boundaries.Language use thus reverberates through the firms and affects the management practices and processes.

2040
AAAJ 31,7 As a second aspect of firms' linguistic adaptability, we identify a culture of ad-hoc translation as an integrative part of the firms' local practices.This issue becomes relevant when the shadow structure of language reaches its limits, that is, if audit engagements unexpectedly require other language skills, such as team members being faced with material in a language other than English or the team language.Translation comes into play and, in line with Kettunen (2017), interviewees confirmed that it is almost exclusively constructed and negotiated at the work practice level, where individuals resort to informal and "ad-hoc coping mechanisms to solve translation requirements" (Piekkari et al., 2013, p. 772).As a result, translation is perceived as a temporary problem to be overcome by pragmatic means, such as the use of external online translation tools (e.g.Google Translate), whereas professional translation services are used only when reports to the regulator or the ministry are produced.As individuals struggle with audit material in foreign languages, "you reach out to your colleagues" (I No. 2a, senior manager, German) who are better versed in the respective language.Interviewees emphasized that, if clients send documents in the "wrong" language, "you nevertheless can do your audit, because you can then, the few documents you receive, you can translate" (I No. 2a, senior manager, German).Through this culture and attitude of "we will make it work" (I No. 12, senior manager, Filipino), auditors search for and activate foreign language resources within their organizations.For example: It was my first day in [audit firm] office and they just caught somebody speaking Chinese and I spent one week translating the long-form report from Chinese to English.But still it's not word-by-word translation.It's key information, key sentence, key phrase translation.[…] In the long-form report, sometimes they mention a specific transaction or practice of the bank, of which I have no idea.I have not been working in a bank, so I have to search online, to be honest (I No. 11,intern,Chinese) Similarly, a US senior (I No. 10) reported that she gave up on correcting English grammar mistakes and typos in work files, because "it's not taking away from the audit, like if I change this one word to make the tense correct, it doesn't change the audit, you know, like, which conclusions we reached".Translation tasks are thus broken down into single words or "key information", while further details and context are disregarded.This approach does not overcome language barriers, but circumvents them by partitioning translation tasks into what is perceived as accessible portions.For these, advice from colleagues is sought (Piekkari et al., 2013).It is this unproblematized understanding of language and translation through which language diversity is cultivated into a competitive advantage of the Luxembourgish Big Four: I didn't see a lot of problems with translation between French-German-English, because those are the languages you can easily find someone within the close proximity to check something, you know.You can ask your teammate or someone you work [with] (I No. 1, intern, Turkish).
Language absorptive capacity is thus not only a resource that the firms hold because of their multilingual workforce, but de facto makes language and translation something that is worked around or treated as a non-issue.Conceptualizing language absorptive capacity as an organizational mindset connects with the uneasiness and discomfort that newcomers report when facing translation tasks.These activities are forced on junior auditors regardless of whether they possess adequate language skills or had sufficient access to formal language training.As a Luxembourgish junior (I No. 5) joked in German that his employer did not know how good his German is, he did so using a grammatically awkward expression.Several other junior interviewees recalled translation tasks where they did not feel at ease, such as when translating legal contracts, either due to lacking education in the foreign business system (I No. 8, senior, German) or due to having insufficient knowledge of the foreign language, such as a Chinese junior (I No. 15), who considered herself "the only, kind of, German-speaking junior" on an audit of a German real estate client.It may be posited that these individuals have not yet embraced the firms' mindset on language and translation.

Language at work in the Big Four
However, adopting ad-hoc translations without cross-checks or further reflection may have serious implications for the quality of audits, as problems might result from the use of inappropriately translated terms in the engagements (Evans, 2004).Recognizing that expressions are anchored to a specific language complicates translation, which entails more than finding an equivalent term.As an example of such a "discursive void" (Tietze et al., 2017), an interviewee (I No. 4, former director, Luxembourgish/French/Italian) referred to the impairment of assets, which under Luxembourgish generally accepted accounting principles (GAAP) need to be written down "if you have a durable loss in value".The English version uses the term "permanent", where "the problem is that 'permanent' and 'durable', at least in French, is clearly not the same thing.'Durable' is something that is going to last, but that is not definitive.'Permanent' is something definitive[3]".Awareness of such issues requires a level of fluency beyond the "communicative effectiveness in English"or other languagesthat is propagated and prevalent in corporate contexts (Ehrenreich, 2010, p. 408).While non-native languages are often used in local audit work, the issue is largely dismissed given the firms' mindset of language absorptive capacity.
This understanding of language and translation suggests that language absorptive capacity is not just the sum of the linguistic skills of organizational members.Instead, it is an integrative part of the organizations' culture, cultivated and reinforced through organizational discourse and interactions, and drawn upon as part of the firms' global network.While this symbolizes and enables their status as global audit firms, it reflects an unproblematized attitude that ignores the subtleties of language and translation.

Linguascapes: discursive and organizational boundaries
Local language practices structure the Big Four into distinct audit teamsor "linguascapes", in which a specific language prevails (Steyaert et al., 2011).These linguistic landscapes are discursively legitimized "in the name of the client" (Anderson-Gough et al., 2000) and are different from the firms' branches or industry groups.This section argues that the demarcation of these linguascapes is reinforced due to the stereotyping by the team members of their own and the "other" team.This makes the group-homogeneous languages more than informal working languages, and discursive and organizational boundaries are created around the linguascapes.
In line with Steyaert et al. (2011, p. 273), the linguistic spaces emerge from auditors' "need to adjust to [their] conversational partner", as interviewees agreed unanimously that linguistic adaptability, practiced at all levels, to the requirements of the client is essential.This attitude is institutionalized in all firms, as underlined by interviewees' frequent use of the term "of course", and it was confirmed that "generally, you really try to service clients, so, of course, whatever they need, that needs to be met" (I No. 12, senior manager, Filipino).Consequently, and despite English being the corporate language, the firms' client focus offers French and German as parallel informal languages in the Big Four firms.Besides facilitating discussions with the client, internal interactions in French or German were described as more efficient and precise than those in English.The two informal languages are thus not only constitutive of temporary audit teams, but create distinct intra-organizational groups, which stand in sharp contrast to the proclaimed diversity of the Big Four's workforce [4].
The French and German linguascapes possess their own group identities, fabricated and maintained through "the usual stereotypes, which you have between the two cultures" (I No. 2b, senior, Luxembourgish).The Germans are said to "come early", "are very organized and structured", and "more serious", whereas the French "are more flexible", "cool", "curse all the time", and "enjoy drinking after work".As these clichés float through the firms, they tend to be dismissed as "jokes", because auditors "still should do their job" (I No. 12, senior manager, Filipino) and because "we are so busy during the busy season that we don't have time to build up [such] differences [and] to allow people to build up, you know, bad rumours about each other" (I No. 2b, senior, Luxembourgish).Yet discourses give rise to processes of identification (Kärreman and Alvesson, 2009;Phillips and Oswick, 2012), and it is generally agreed that "you see a little of that" in the audit teams and the firms more widely (I No. 10, senior, US).
Stereotypes about working hours and behaviour at work quickly extend to discourses about other work habits, which enhance the boundaries of the linguascapes.Such discourses cover the core product of the Big Four, namely auditing.Although systems and procedures, such as audit methodology and standards, define the audit process in line with the "concept of effortless experience" (Segal-Horn and Dean, 2009, p. 49), the language groups insist that they find room to customize the process according to their work preferences (I No. 14, senior, French).In line with the stereotype of being structured, German teams are said to often use ready-made templates to run their testing procedures.From the perspective of the "other" team, such automation introduces the risk that "some junior people will tend to [or] don't put in perspective what they're doing and don't think much of what they're doing" (I No. 14, senior, French).As the French team is "more disorganized", but "flexible", newcomers fit in more easily.French teams might require more hours to complete an audit, but have more fun and are more social.They would also be "more autonomous in that they do not accept that you give them an order" (I No. 7,manager,French).This contrasting of work habits reveals that the separation of audit staff by client language constitutes organizational reality and introduces "othering" and distancing processes into the firms (Carmona and Ezzamel, 2016).
The resulting boundaries between the groups are rigidified by organizational rumours (Michelson and Suchitra, 2000).It was suggested that French auditors came from all over France, including graduates of elite institutions, whereas German graduates were typically recruited only from the border region.This caused a low supply of German recruits, the consequence of which triggered the suggestion: "Any German that will apply here will get hired, for sure" (I No. 3, manager, Lebanese).German-speaking recruits would thus draw higher starting salaries than French-speaking recruits and "the difference is not just 500 Euros" (I No. 8, senior, German, TR).The interviewee reports that this perceived unequal treatment enhances the distancing between the two language groups, not only because higher pay for the same work is considered unfair but also because it raises expectations as to individual performance.Thus, what begins as a teasing discourse creates and exposes more fundamental rivalries.Individuals intricately mix language with culture and nationality, as they infuse their narratives with assumptions about, for instance, the politics of promotion processes: This is something that is much talked about, that a partner prefers promoting someone of the same nationality.[…] But what I have seen, for example, is that there is a German partner for whom I have worked a lot […] and I was therewe were having drinkswhen he said to a senior: "Yes, I would like to make you a manager, because we have only two German managers and we need some more again."This is due to operational reasons and social, cultural reasons.That is, of course, he very much likes to work with a German, because he knows that person, because he knows him well, and he knows that they have the same quality standards.It is of course also about the client, because a German client prefers to have a German manager.That is, it can be brought down to two things, but, yet, you never quite know.And the loser of this race will always say: "Yes, this is a culture issue" (I No. 5, junior, Luxembourgish, TR).
As such, the grouping that originates with the client is fortified by stereotypes and discursive practices, and ultimately becomes performative, as "in the end, the structure of the company will always drive the behaviour of people" (I No. 14, senior, French).The resulting segmentation of audit staff interacts with individuals' preference to gather in groups in which they are able to speak their native language.In this way, grouping exerts forces beyond the immediate work practices, and continues, for instance, during lunch breaks in line with the work habits stereotypes: "The Germans go [to lunch] at 11:30, something like this, because they came early in the morning, [so] they eat earlier.And the French [go] more at one" (I No. 13,junior,French).Likewise, we identify open-space offices and after-work social events as reinforcing the firms' fragmentation.Being organized as an open space, the firms' offices are meant to enable more interaction among auditors and, along with a flat hierarchy, make everyone accessible (EY, 2016).Interviewees argued that open-space offices emphasized the "little bubbles" where the respective languages are spoken (I No. 2c, senior, Italian).Although they report that this structure enables them to get to know more people, they tend to separate themselves and their teams in "cocoons" (I No. 16, junior, Russian, TR), possibly even on separate floors (I No. 9, intern, German).Groups of desks become known as "little Germany" (I No. 5, junior, Luxembourgish) or where "the wild ones" work (I No. 8, senior, German, TR).
Similarly, social events do not per se enhance intergroup communication.These events are often structured around themes or activities, such as movies, food, or cocktails, and are to allow particularly younger staff to create an organizational network.Yet again, two types of people were said to exist, namely "those that are herding together" being the majority and a few who "talk to people who do not share the same mother tongue" (I No. 16, junior, Russian, TR): People mingle, people work across, talk across, but oftentimes the people who are closest are all from the same country and speak the same language, and they'll speak that language (I No. 10, senior, US).
While these organizational practices are meant to counter intra-organizational segmentation and foster broader relationships and identities, they end up exposing and rigidifying the social divisions in the firms.As a result, the linguascapes erect a "shadow structure" and undermine the firms' striving for a singular corporate culture.
This structure exists primarily for the French and German linguascapes, which possess strong in-group identities.The international linguascape is more heterogeneous and has a different degree of perceived closeness.As international team members move back and forth between their audit teams and respective native language group, there are "no very strict boundaries" between the groups (I No. 1, intern, Turkish).Yet we also identify the beginnings of such grouping.As "networks of communication and interaction are established based on language skills" (Vaara et al., 2005, p. 618), fellow nationals were seen as vital to foster the corporate experience of new joiners.This group represents a "mini-family" that helps recruits to integrate (I No. 12, senior manager, Filipino), and individuals seek out fellow nationals for their advice, because "they may want to help me [more] easily" (I No. 23,intern,Cameroonian).International team members also tend to have lunch with fellow nationals, "even if they are from different departments or different teams" (I No. 18, junior, Chinese).As "you will have maybe more affinity with someone from your country" (I No. 7, manager, French), such grouping fuels individuals with common assumptions and beliefs that "express, normalize, and reproduce organizational practices" (Anderson-Gough et al., 1998, p. 587).
In sum, the process of "cultural matching" (Rivera, 2012) reverberates through the firms and teams, with language, or closeness of backgrounds, as the principal reference points.The performativity of the language division is enhanced by similarity between individuals creating "a powerful emotional glue" (Rivera, 2012(Rivera, , p. 1001) that is a stronger force than exerted by the firms' efforts to homogenize their workforce.The linguascapes thus represent dominant language groups that define and impose their own rules for social interactions and events.As the groups tend to reproduce "their respective fields of social relations" through hegemonic practices (Tietze and Dick, 2013, p. 123), such behaviour makes it difficult for others to join and may act as social closure mechanisms.

Language skills as sources of power: competence assessment, and power of exclusion
In light of the dominance of the French and German linguascapes, their hegemonic practices may have consequences for other individuals.This section outlines how the identified linguascapes may have implications for human resources (Piekkari et al., 2005).First, it is argued that they may affect an individual's position in the organization through the assessment of one's competence (Śliwa and Johansson, 2014).Second, the linguascapes may exert strong exclusionary power (Vaara et al., 2005).As language introduces a powerful structure into the firms, we argue that it may construct or dissolve individual quality and performance.
First, language has the capacity to produce personal credibility and authority within the firms.As such, membership of the dominant linguascapes is seen as a salient characteristic to form expectations and pass judgement on others: When I work with somebody […] I cannot see his nationality.But I know in advance that he has some advantages at some levels, because when he is German or Belgian I know his strong points are discipline, good organizational skills, and communication (I No. 17, junior, Greek).
While the interviewee began by suggesting that nationality and culture do not play a role, indirectly he demonstrates a strong association of language, culture, and skills.Another interviewee asserted that "you work with people that you rely on and many times it is people from the same nationality" (I No. 2a, senior manager, German), demonstrating how quality, or reliability, is attributed to language group membership.Language reflects and reproduces unproblematized and taken-for-granted understandings of the organization, but also of the world in general: "He's Polish, but more a German type" (I No. 15, junior, Chinese).As a result, language becomes a proxy, if not a direct criterion itself, for how individuals are performing.In other words: language reveals and constructs quality.
This aspect becomes most apparent when considering the language versatility of local audit staff, which transforms the Luxembourgish into valuable assets.While French, German, Luxembourgish, and English have a "natural" status at the Big Four, individual auditors that master these four languages have an advantage in the firms: You have these internal profiles, where it is shown what you specialize in, which other professional qualifications you have, which languages you speak.And, as a Luxembourger, you immediately have the four languages flagged up (I No. 5, junior, Luxembourgish, TR).
Consequently, "people say Luxembourgish people are above everyone" (I No. 10, senior, US).Outside the firms they can interact with local clients and authorities, are knowledgeable in local GAAP, and, on a senior level, can obtain and manage local audit engagements.For example, a Luxembourgish interviewee (I No. 20, junior) recognizes from the client's welcoming or accent which language they prefer and adapts accordingly.Inside the organizations, they possess multiple group identities through which they are able to move freely between the linguascapes.As language fluency qualifies auditors and language skills are an indicator of competence (Śliwa and Johansson, 2014), local auditors are "very smart people" that are needed in this market (I No. 17,junior,Greek).
This special status is nurtured by a very low number of Luxembourgish audit staff, as natives tend to earn considerably more, while working less, at state institutions.One interviewee recalls his superior telling a colleague: "Don't give him [the interviewee] too much work, don't stress him, we need him in the future" (I No. 20,junior,Luxembourgish).While commitment and dedication to auditing are equally required, the example of local audit staff emphasizes the importance of "linguistic capital" (Bourdieu, 1991).Language skills thus translate into "unearned status gain", that is an elevated standing "attained not through individual effort or achievement but from a shift in organizationally valued characteristics" (Neeley and Dumas, 2016, p. 14): We have a Luxembourgish guy, who performed really, really badly in our team, at our client.And, […] of course, he got a very bad evaluation, but normally these people should be fired and cannot pass the trial period.[…] But since he's Luxembourgish, the company gave him an exception, yes, and then he stayed, till now.He didn't get fired.But his evaluation is so bad, that normally he should get fired (I No. 15, junior, Chinese).
The second performative aspect of language relates to its power of exclusion (Vaara et al., 2005).This is evidenced by the French and German linguascapes strengthening their group boundaries through their native languages in contexts where English should be used: It's not quite easy, I mean, they would keep talking until, I don't know, at one point, they start to explain and then they would switch back to French again (I No. 18, junior, Chinese).
Interviewees expressed anxiety and reported tensions arising from being assigned to "wrong" teams, such as to a French audit team without speaking French.This occurs in "situations where you just need to staff your team" regardless of language (I No. 4, former director, Luxembourgish/French/Italian), and especially junior audit staff tend to be assigned to teams randomly (I No. 2a, senior manager, German).A Chinese intern recalled her first assignment: That is hard.I felt a lot of pressure at the time.But the people in the team were nice as well.They also explained things to me in English.But, still, for example, when the senior manager came, we all had a question about the testing.Actually, I was doing the testing, [and] I raised the question to the senior and he didn't know how to do it as well.So, when the senior manager passed by he asked the question.The senior manager answered the question very well, but in French.And I didn't understand all of that, so only after the senior manager left, the senior explained it to me […] in English.So, I mean, it could be a problem, but as long as that is the case that everybody else in the team speaks French as mother tongue, while I am not, then they would of course prefer to speak their mother tongue (I No. 11,intern,Chinese).
Not only was the intern shut out of the conversation, she received an explanation only after the senior manager left by a "language mediator" (Marschan- Piekkari et al., 1999)[5].What seemed to be a valid question became associated with the senior, thereby raising his visibility and leaving the intern in an inferior position (Vaara et al., 2005).As interviewees emphasized communication skills as vital for career advancement, interacting with superiors and building an organizational network were seen as essential factors to show intellectual curiosity and commitment.However, these interactions strongly rely on the individual's language proficiency (Piekkari et al., 2005;Welch and Welch, 2018).Otherwise, it is difficult to be "a good team player" and "help others out" (I No. 9, intern, German, TR).Hence, "you need to communicate.If you don't, you will not survive" (I No. 13,junior,French).One interviewee argued that her lack of French language skills affected her post-engagement evaluation: Afterwards, I didn't get good feedback on [the engagement], because they said: "Hey, you were not motivated".[And I said:] "Well, I sat there the whole day and didn't understand a thing" (I No. 8, senior, German, TR).
This situation shows the status loss of individuals unable to communicate in their team's language (Neeley, 2013).It also reinforces hegemony, which is achieved when "the dominant group successfully projects its own particular ways of seeing the world, human and social relationships, such that those who are actually subordinated by these views come to accept them as being 'common sense' or 'natural'" (Tietze and Dick, 2013, p. 123).Indeed, despite the awareness that others' language use excludes them, interviewees agreed that it was "normal", "human nature", and "more efficient" if individuals conversed in their "comfortable language", as "they don't have a language barrier" and it was easier to discuss audit-related technicalities, thereby conforming to the discursive practices of the linguascapes.Interviewees rather perceived it as "a nice bonus" to understand the respective team language and listen in to the discussions of fellow team members, as these potentially relate to one's tasks and team members might "forget to translate" (I No. 22,junior,Russian).Notably, superiors also do not fully recognize the exclusionary power of language, and interviewees reported that it often depended on seniors how inclusive the language teams were with foreigners.Some interviewees thus perceived the need to learn one of the two major languages -French or Germanto join in social interactions, as "people can feel like robots" if exchanges are only about work (I No. 1,intern,Turkish).Yet this primarily pertains to staff-level auditors: We have directors and partners who would be from Canada or from England, you know.But then, they would be there, not speaking French or German.But then, they would have connections.They would bring in clients.And for me, it would be hard because I don't think that I would have those kinds of connections to bring in clients, and I wouldn't speak the languages (I No. 22,junior,Russian).
The lack of language skills can thus be compensated by other skills or social capital, such as client networks, but it generally diminishes individuals' roles in the firms (Piekkari et al., 2005).Concurring, individuals who do not speak their team's language find themselves unable to fully access the organizational experience and, as a result, run the risk of remaining organizational outsiders.In that spirit, language skills have extensive implications for one's job satisfaction, but importantly also for individuals' career advancement and success.
The exclusionary power of language is also manifested in the national audit credential, which in Luxembourg is the Réviseur d'Entreprises Agréé.Credentials from EU countries may be transferred upon obtaining a complementary certificate in local rules and regulations.Despite the country's three official languages this examination is set in the French language, while, upon request, answers may be given in Luxembourgish, German, or English.Expatriates seeking the local audit credential thus need to have a solid working knowledge of the French language.To bypass this barrier, international staff often pursue a foreign credential, mostly the UK's Chartered Certified Accountant (ACCA).Although this does not allow them to sign local financial statements, it qualifies them for audit work elsewhere or to take the complementary training exam at a later point.The prevailing view on the policy to have questions only in French was perceived as "discrimination" against non-French speakers: It's a problem for a lot of people.[…] In general, most people coming from abroad, you get many people coming with an ACCA and they cannot sign accounts, because they have to pass the IRE [exam] and they struggle to understand the questions.They are really qualified, but they're blocked because of that.It's a big issue.There is much discussion about that.[…] The strategy I see [is this]: people are a bit upset, but they try to pass it anyway.They try to understand the French questions by memorizing what it means, but they don't learn French to pass it.They won't start French just to pass the exam.They will try to understand the questions and try [to pass] the exam (I No. 14, senior, French).
As "you cannot go beyond senior manager if you don't have the qualification" (I No. 12, senior manager, Filipino), the credential becomes essential to become partner, turning the professional exam into a language-based "glass ceiling" (Piekkari et al., 2005, p. 340).Indeed, a closer look at the Luxembourgish Big Four reveals that the mother tongue of half of the partners is French, followed by English (14 per cent), German (13 per cent), and Luxembourgish (10 per cent), while audit partners with a different mother tongue make up less than 3 per cent [6].
The professional exam's language requirements thus emphasize the performative nature of language, relating to both the assessment of one's competence and its exclusionary power (Vaara et al., 2005;Śliwa and Johansson, 2014).Language skills allow auditors to navigate the multilingual context, and obtain social capital that is vital for career success (Welch and Welch, 2018).Yet, despite the firms' international workforce, the French and German

2047
Language at work in the Big Four linguascapes introduce a language-based hegemony into the firms (Tietze and Dick, 2013).Individuals with relevant language skills find themselves in better positions, even though such status gain may be unearned (Neeley and Dumas, 2016).At a senior level, where client management is a core task, language diversity is further reduced, and the local languages become almost indispensable prerequisites to career advancement.

Discussion and conclusion
This paper has demonstrated the multifaceted role of language in multilingual Big Four audit firms as being visible and subtle, as well as functional and performative.Although English is the corporate lingua franca at the Luxembourgish Big Four firms, the study suggests that it is other languages that mediate the intra-and inter-organizational relationships.The Big Four organize their audit teams primarily based on the language of their clients, resulting in predominantly French and German teams as well as international ones.This client alignment introduces a grouping in the Big Four that is fostered and primarily expressed discursively.Yet, as social identities and identification processes offer comfort and security, the grouping "creates a bond between the individual and the collective" (Kärreman andAlvesson, 2009, p. 1127).As a result, the firms are internally divided and different subcultures characterize the Big Four, rather than a singular corporate culture of a global audit firm.Testifying to the power implications of language, the paper also showed that language skills endow individual auditors with "linguistic capital" (Bourdieu, 1991).Organizational actors with relevant language skills are situated in more powerful positions.Language thus turns into a channel of influence through which intra-organizational power is exerted, as it imposes a powerful "shadow structure" on communication patterns, hierarchies, and informal networks that lie behind the firms' formal organization charts (Marschan- Piekkari et al., 1999).
Returning to our research questions, we first asked about the ways in which the Big Four manage the linguistic tensions between global Englishization and local multilingualism.The study showed that the firms emphasize their linguistic adaptability that Piekkari et al. (2013) conceptualized as "language absorptive capacity".This claim is an integrative part of the culture of the Luxembourgish Big Four branches and stresses their ability to provide effective services to any client.The firms' client orientation (Anderson-Gough et al., 2000) structures the firms into three main language teams that cater to the language requirements of local clients.In addition, other languages are covered by the firms' absorptive capacity and we argue that language resources become a commodity within the Big Four that is freely available and activated whenever needed.Such a view of language and translation fits with the global aspirations of PSFs.However, it also exhibits an unproblematized perspective in that the firms' linguistic diversity can be activated as competitive advantage only through an organizational culture that disregards issues of translation.It confirms the tensions recognized in prior research between potential problems resulting from technical translations (Evans, 2004) and those being negotiated in situ (Kettunen, 2017).More broadly, the paper adds language as a further factor that undermines the "one firm" model (Barrett et al., 2005;Boussebaa et al., 2012;Muzio and Faulconbridge, 2013;Spence et al., 2017).Echoing the notion of institutional duality (Kostova, 1999;Kostova and Roth, 2002), it has demonstrated that local contexts require local language policies and practices.While bearing in mind that our empirical site represents an extreme case of local multilingualism, it appears only as a matter of degree that PSFs in other non-English contexts privilege local language skills, as "Englishization" does not encompass the interaction with local clients.
The second question explored the relations between the firms' linguascapes (Steyaert et al., 2011).Responding to Vaara et al.'s (2005, p. 621) call for research on the "subjectivity and identity construction" of English as a corporate language, we argue that such identities are created and maintained along language lines, as the firms' linguascapes engage in distancing processes, both discursively and organizationally.To be precise, the underlying nationality narratives and stereotypes on the language groups are continuously constructed and de-constructed as they float through the Luxembourgish Big Four.While interviewees consistently linked the same stories to specific groups, they emphasized that auditors are "too busy to be affected by [such] talk" (I No. 2b, senior, Luxembourgish).Despite this affirmation, stereotypical discourse enforces the grouping within firms and interviews suggest that association with a particular language team runs deeper than the discursive level.Language group membership is the basis for communication and interaction networks (Vaara et al., 2005) and eventually becomes a criterion for a process of "cultural matching" that bonds individuals together (Rivera, 2012).The grouping is enhanced organizationally, primarily in the open office and at social events, indicating that the firms' segmentation creates a strong sense of belonging to a particular linguascape.As such, tensions emerge from the organizational language policy of the Luxembourgish Big Four, as the homogeneity of the firms' workforce is undermined.Language becomes an identifier for individuals and creates social bonds within the firms' linguascapes.Between the language groups, it repels, as it introduces an "othering" into the firms (Carmona and Ezzamel, 2016).Although diversity is a strong part of the image of the Big Four in Luxembourg, this paper reveals that the internal structures reinforce conformity and homogeneity.Yet, as the multilingual environment in fact produces monolingual teams, equality and diversity are unmasked as linguistic adaptability.Abstracting beyond the case of Luxembourg, where language seems most important, it is apparent that other contexts may be divided by one or a combination of other factors, such as class, gender, or ethnicity.Thus, despite attempts to increase the profession's inclusiveness (Kyriakidou et al., 2016), the firms may still be internally divided and riven by different subcultures that put into question the firms' homogeneity, in terms of language, culture, and working practices.As the different subcultures employ different practices and engage in internal clashes, this paper has shown that the resulting series of "local-local" divides challenges the view of single branches of a PSF network as monolithic constructs; in other words, the "one firm" model remains an elusive ideal even for single branch offices.
Our third and final question asked about the tensions and power implications that emerge from the co-existence of distinct intra-organizational linguascapes that reveal individuals' uneven language skills.In line with the international business literature (e.g.Marschan-Piekkari et al., 1999;Vaara et al., 2005;Śliwa and Johansson, 2014), we find that language and communication affect the visibility of auditors on engagements, where those without relevant language skills run the risk of being excluded and remaining organizational outsiders.As language skills are a major factor for organizational and professional access and influence in Luxembourgish audit firms, language proficiency endows the speaker of that language with "linguistic capital" (Bourdieu, 1991) and helps auditors to connect better with their teams and clients.It thus enlarges their scope of project assignments and enables them to create intra-organizational links and relationships, thereby obtaining important social capital (Welch and Welch, 2018).Vice versa, the formal and informal hierarchies introduced by language limit an individual's career advancement, unless auditors speak at least one of the dominant languages of French and German.First, an informal barrier is set by the need for these languages in team communication and client management.To the extent that a foreign language is needed for a particular assignment, auditors may feel shut out of the team.Since audit engagements are temporary assignments of varying lengths, the linguistic capital also seems temporary.That is, unless there is a client portfolio of, for instance, Italian clients, a person finding their Italian language skills are needed in one engagement may find these skills of less use in a subsequent audit of a French client.Thus, the hierarchies introduced by "linguistic capital" seem to be ephemeral, particularly for junior staff.Nonetheless, a more permanent shift to a 2049 Language at work in the Big Four certain client portfolio, such as due to increasing numbers of Chinese firms in Luxembourg, increases the demand for Chinese auditors and possibly endows this group with more permanent "linguistic capital".A second and more formal barrier is erected when attempting to pass the Luxemburgish audit credential, which requires a reading proficiency in French to understand accounting questions.The resulting barrier may be likened to a language-based "glass ceiling" that prevents "promising individuals with management talent from advancing in the organization" (Piekkari et al., 2005, p. 340).While it can be argued that auditors in Luxembourg need to have a working knowledge of the local language(s), the Big Four firms claim to be international environments that harbour a diverse set of backgrounds.This claim is undermined by the environment that possibly prevents international employees from advancing in the profession.Future research is invited to examine in more detail the issues that have come to light here, primarily relating to three areas.First, this study has only begun to explore the language policies and practices of audit firms, and further efforts are needed to explore how language skills and versatility affect professional audit work.This relates to the audit work in multilingual teams and the management of these teams, as well as the substantive issues on which the language groups differ in their work and the power dynamics between the linguascapes.Second, studies of the corporate experience of expatriates might corroborate our findings and shed further light on the difficulties that an increasingly international workforce experiences in monolingual, foreign audit contexts.Such studies could also explore the interaction between language and communication in the socialization of newcomers, as well as implications thereof for professional identity in global audit firms and how individuals experience and cope with conflicting practices across audit teams.Third, this study can only suggest, rather than fully examine, that interviewees' perceptions of the role of language differed depending on their hierarchical positions.Juniors placed much more emphasis on language issues than (senior) managers.Perhaps foreign managers have found ways to adapt to the language requirements or have found a niche by working with international clients.Future work may explore how language demands change as auditors rise through the ranks and how they discursively and linguistically meet these demands or find ways around them.Beyond these more detailed issues, this research has revealed that the role of language in professional work contexts provides ample opportunities for further research.
3. While the interviewee refers to Luxembourgish GAAP, the example actually stems from the EU's accounting directives, from which Luxembourgish GAAP is derived.In both the Fourth Directive of 1978 (Article 35) and the more recent Accounting Directive of 2013 (Article 12), the English version requires the value of fixed assets to be adjusted "if it is expected that the reduction in their value will be permanent", which in the French version is translated as "si l'on prévoit que la dépréciation sera durable" (emphasis supplied).
4. Indeed, all Big Four firms emphasize their global workforce on their career websites, seeking "people who are good at working with people from different backgrounds to their own" (EY), stressing that "you will be using different languages at work" (KPMG), that "cultural diversity is a reality" (PwC), and that "sound language skills are fundamental if we are to excel in our professional activities" (Deloitte).
5.She noted that "later it came to my revenge" when she was staffed on a Chinese engagement.Since translations are typically not reviewed, translators obtain crucial power in their teams because of their language skills.
6.The list of audit partners was taken from the professional institute (IRE).Nationality and mother tongue were assigned based on LinkedIn profiles.For 10 per cent of the partners we could not find the relevant information.