Polite impoliteness? How power, gender and language background shape request strategies in English as a Business Lingua Franca (BELF) in corporate email exchanges

1.1 Rationale

Since the turn of the millennium, multinational corporations have increasingly adopted English for their organisation-wide communications (Kankaanranta and Louhiala-Salminen 2013). The dominance of English as a global lingua franca makes it a useful common denominator to fulfil the communicative demands of the ever growing cultural and linguistic diversity of workforce in multinational companies. The adoption of English in professional settings across the globe has given rise to the emergence of Business English as a Lingua Franca (BELF). BELF is a specific form of English as a Lingua Franca (ELF; Seidlhofer 2001, 2011) and involves active use of English in professional and diverse contexts and between speakers for whom English is the first, second, third or additional language (Kankaanranta and Lu 2013; Kankaanranta and Planken 2010). The linguistic and pragmatic features of ELF communication have been of interest to many researchers, and aspects such as morphosyntax and lexicon (Hülmbauer 2007; Pitzl 2009; Pitzl et al. 2008), fluency (Hüttner 2009), and phonology (Jenkins 2000, 2009) have received particular attention. BELF has been less well researched, which partially has to do with the difficulties of obtaining access to corporate, specifically internal communications.

BELF’s quintessential feature is that it is fluid and situationally contingent in a particular context of professional practice. Similar to ELF users, for whom communicative functionality rather than correctness by native-speaker standards is the priority (Hülmbauer 2009), BELF users are driven by pragmatic and situationally relevant needs rather than prescriptive motives. In some cases, it may even be beneficial for interlocutors to flout prescriptive linguistic norms if this leads to a more effective way of getting a message across (Seidlhofer 2011: 124).

This focus on functionality is aligned with the most important aspect of workplace, that of performing work-related goals (Bargiela and Harris 1997; Vine 2020). For the goals to be achieved, requests need to be formulated and responded to (Vine 2020), often in a direct way because this ensures better clarity, transparency and efficiency (Ogiermann 2009). Efficient use of resources – including language resources – saves time and money, which is the modus operandi of business settings (Pullin 2013). Failure to express and/or follow requests might result in miscommunication, unfulfillment of goals and ultimately business losses. For this reason, requests are a key pragmatic function of professional communication performed on a daily basis.

Yet, requests are quintessential instances of Face-Threatening Acts (FTAs) as their semantic content bears great potential for interfering with recipients’ negative ‘face’ demands, i.e. their need to protect their public self-image (Brown and Levinson 1987). Requests in business negotiations are especially problematic because the interference with a business partner’s ‘face’ could lead to inadvertent negotiation outcomes. Previous research on international business communication therefore points to great efforts on the part of participants to establish an ambiance of comity, cooperativeness and harmony in workplace communication (Komori-Glatz 2017; Pullin 2010, 2013; Vine 2020). However, despite the overall need for an amicable communicative environment, the nature of most workplaces with their goal orientation means that it is eventually inevitable for interlocutors to express FTAs such as requests to achieve business aims. Choosing the right strategy to express the FTA can thus mark the ‘make or break’ of the success of a transaction; it can minimise disruptions to comity and establish good relationships between co-workers.

BELF users’ choice of request strategies might be affected by a host of linguistic and non-linguistic contextual variables, including social norms of what constitutes an appropriate request in a specific context, and the power relations between the interlocutors. Culture-specific ways of formulating and interpreting requests add another layer of complexity to an already complex speech act, potentially affecting communication and relationships in linguistically diverse workplaces (e.g. Félix-Brasdefer 2007; Ogiermann 2009; Ogiermann and Bella 2020).

For this reason, BELF emerges as a natural testing ground for investigating the accomplishment of work-related goals through requests. Yet, little is known about the topic, which is largely due to the confidentiality of internal business communication and difficulties in gaining access to relevant sites. Although some studies on BELF have been based on authentic materials (Cogo 2016; Kankaanranta et al. 2015), many have relied on elicited data generated from interviews with practitioners, discourse completion tasks (Sweeney and Hua 2010), questionnaires (Margić 2017), or simulated transactions (Komori-Glatz 2017). Whilst findings based on elicitations and self-reports provide useful insights into the formal nature of BELF, these types of studies cannot be fully representative of how BELF is actually used in naturalistic settings, and of pragmatic strategies that arise out of real communicative needs.

The current study aims to contribute to the growing body of research on the pragmatics of workplace communications through BELF (e.g. Kankaanranta and Lu 2013; Pullin 2013) by investigating one of the most important speech acts in the context of work, i.e. requests. It is based on a unique corpus of 398 authentic internal emails produced by BELF users who are employees of a large multinational insurance company. Their lingua-cultural backgrounds comprise 10 different first languages (L1s), including L1-speakers of British and Irish English. Given the importance and complexity of requests in the workplace, we specifically investigate what kind of strategies BELF users prefer and how their usage patterns interact with intra- and extra-linguistic variables including email senders’ L1 status, as well as gender and power position within the corporate hierarchy. To describe the nature of the FTA requests, we draw upon the framework of (in)directness of requests by Blum-Kulka et al. (1989).

1.2 Literature review

The term ‘request’ is commonly defined as a speech act in which the message sender wants and/or tries to get the message recipient to do something; that something is beneficial to the former and costly to the latter, threatening their negative face (Ogiermann 2009). In this study, we classify requests based on the different levels of interpretive force required to understand requests as such. Blum-Kulka (1987) offers a useful categorisation of requests into nine distinctive types, ranging along a continuum from the most direct ones (associated with orders), in which the ‘requestive’ force is either marked syntactically or expressed explicitly, to the most indirect ones, which place more interpretative demands on the hearer. The direct category includes mood derivables, performatives, hedged performatives and want statements, whilst the indirect category consists of suggestory formulae, query preparatory, strong and mild hints.

The now vast body of research in cross-cultural pragmatics has revealed different request preferences for different groups of L1 speakers. For example, English and German speakers have been shown to display a preference for conventionally indirect requests (e.g. Ogiermann 2009). Similarly, French and Dutch speakers tend to put more emphasis on negative face-saving strategies and prefer indirect strategies, too (Kerbrat-Orecchioni 2005; Van der Wijst 1995). Conversely, Russian, Polish and Hebrew speakers were shown to use more direct requests (e.g. Blum-Kulka 1987; Ogiermann 2009; Rathmayr 1994; Wierzbicka 1992) though their illocutionary force might be softened with other ‘means’ such as diminutives in Polish (Wierzbicka 1992). BELF can be considered a transcultural communication space (Pennycook 2006) in which speakers with different lingua-cultural backgrounds negotiate emerging pragmatic situations, transgressing cultural or even linguistic boundaries (Cogo 2020). It is therefore relevant to explore how requests are negotiated in professional contexts that are highly inter- or transcultural such as the diverse workplace context investigated in the present study.

Variability has also been observed regarding the perception of politeness of requests. While often the indirect ways of making requests are perceived as more polite, especially in English (Leech 1983), the two dimensions directness and politeness do not necessarily run in parallel. Earlier research by Blum-Kulka (1987) has shown that the most indirect requests (hints) were not always perceived as the most polite by English and Hebrew speakers, and instead the conventionally indirect types such as query preparatory were seen as more polite. Both groups of speakers considered the most direct types such as mood derivables, obligation and want statements as the least polite. Ogiermann (2009) has demonstrated that in Slavic languages such as Polish or Russian direct requests are seen as polite, too, especially when their illocutionary force is modified with downtowners such as ‘please’. The existence of different socio-cultural norms of what constitutes a polite request or not has been shown to create a potential for tensions or misperceptions. A study by Chang and Hsu (1998) found that speakers of American English experienced direct requests expressed by Chinese counterparts (in English) as ‘impolite’. This highlights the importance of understanding the nature of pragma-linguistic transfer of norms and what kind of effects they might trigger.

Moreover, the request perspective, i.e. whether a request is speaker or hearer-orientated, plays an important role in the interpretation of requests, too (Blum-Kulka 1989). The perspective can be a form of mitigation and may trigger different perceptions of politeness (Ogiermann and Bella 2020). Speaker-orientated requests (Can I) foreground the agency of the message sender, and in some languages, such as English, are seen as more polite, while in others, such as Spanish, as more formal. Hearer-orientated requests on the contrary (Can you) emphasise the role of the message recipient, and as such are seen as more polite in languages such as Russian or Japanese (Ogiermann and Bella 2020).

The multitude of ways of expressing requests, the level of directness or indirectness involved, the request perspective and diverging socio-cultural norms that link some types of requests with politeness while others with impoliteness make requests highly complex speech acts. Additional complexities arise when requests are performed in more formal professional contexts in which speakers of different lingua-cultural backgrounds come together and interact through the medium of BELF.

Linguistically and culturally diverse workplace settings have often been seen as sites ripe for tensions and misunderstandings obstructing work progress. A possible source of miscommunication can arise from differences in politeness strategies. Past research has often highlighted the difference between L2 and L1 speakers of English as a crucial differentiating factor. L2 speakers have for instance been suggested to use more direct and different politeness strategies than L1 speakers (e.g. Maier 1992). Yet, differences in politeness strategies do not need to hinder the achievement of communicative goals. In this study we take the perspective that in BELF contexts these types of phenomena will be mitigated by the accommodation strategies within BELF communities of practice. Existing studies have for instance described a range of strategic accommodation strategies, e.g. used by L1 speakers to adapt to L2 speakers (Sweeney and Hua 2010). This has been confirmed in a study by Incelli (2013), which compared business emails written by English-L1 speakers and English-L2 speakers with Italian as an L1. They found that English-L1 speakers used a range of accommodation strategies to adapt to the BELF produced by Italian colleagues. Overall, grammatical and lexical errors seemed interactionally irrelevant as long as the purpose and technical aspects of communication were conveyed and understood.

The accommodative nature of ELF communication has been previously highlighted by Seidlhofer (2011), who pointed to the strategic relevance of communicative functionality in BELF users’ communications over compliance with linguistic or stylistic standards by L1 standards. Kankaanranta and Louhiala-Salminen (2010) conducted a study based on authentic internal emails between L1 Swedish and L1 Finnish employees. They investigated differences in strategies for a wide range of discourse features, such as salutations, closings, signatures, and requests. L1 Finnish speakers were found to use more direct request strategies than L1 Swedish speakers. Pullin (2013) analysed the complex and subtle usage of facework to achieve comity in the context of potentially conflictual business meetings conducted through BELF. Their study highlights the importance of mitigating devices, such as hedges, but they also found that such strategies may be modulated not just by the interlocutors’ lingua-cultural background but also by their power status within the organisation.

Indeed, status and power potentially play a more considerable role in organisational settings than in informal encounters. In most instances, a workplace is organised in a hierarchical system, resulting in asymmetrical power relationships, which, in turn, influences workplace communication. Low-power employees are conventionally expected to use more negative politeness strategies associated with more indirect types of requests than high-power employees (Scollon et al. 2011). Yet, the power dimension might be affected by the different linguistic realisations and norms around requests that are likely to come into contact with each other in linguistically and culturally diverse workplaces. Previous research in a multinational company has indeed shown that in BELF communication power relationships are at play and divergent linguistic behaviour, such as impoliteness, may be exploited to get the upper hand in competitive corporate environments (Ehrenreich 2011).

A related factor that has been shown to interact with power in workplace communication is that of gender (Holmes and Stubbe 2003). Widely held views suggest that women and men talk differently in work settings, and that these differences have to do with their distinctive patterns of socialisations and gender norms disseminated in society. The now large body of research on gender and workplace discourse (e.g. Baxter 2010; Holmes and Stubbe 2003; Mullany 2007) has challenged the essentialist assumptions about the existence of women’s and men’s language showing that there are rather masculine and feminine communicative repertoires that are available to both genders and that the use of features associated with a female or male repertoire depend not so much on the gender of the person but on the status within the organisation and the situational context (Baxter 2010). Although direct requests have often been associated with male language, Holmes and Stubbe (2003) have shown that women in higher power positions make greater use of directives and other features associated with a ‘masculine repertoire’. How the interaction of gender and power play out in the context of workplace communication where speakers of different L1s come together and communicate through BELF is still unknown.

Request strategies may also be modulated by the modality of the communication channel. BELF research has so far focused predominantly on spoken face-to-face discourse, as can be seen from the available corpora (Vienna Oxford International Corpus of English [VOICE 2013], English as a Lingua Franca in Academic Settings [ELFA 2008]). This is in contrast to the fact that in the international ‘ELF space’ face-to-face communication is not always practical due to differences in time zones and travel restrictions. Since 2020, electronic forms of communication have become even more prevalent as a result of the global COVID pandemic. A key tool of international BELF is therefore the email (Darics 2015). However, the characteristics of written ELF, specifically BELF, are poorly understood because of issues with access to data resulting in a lack of written ELF corpora (Mauranen 2012). Initiatives such as the WrELFA (2015) aim to address this gap in research by making available a corpus of written ELF academic dissertations (www.helsinki.fi/elfa/). Yet, WrELFA focuses on ‘classically written’ and formal academic genres that are different to the forms and formats of communication used in business settings. For example, much of BELF writing involves computer-mediated business communication that consists of chats or embedded email chains which could be likened to spoken conversations (Jenks 2012; Kaankanranta and Louhiala-Salminen 2010). A growing body of research is investigating computer-mediated ELF (cf. chapters in Martin-Rubió 2018), and existing studies of ELF in the electronic domain have highlighted its fluid and transcultural character (Baker and Sangiamchit 2019), in line with previous observations about spoken ELF. In a similar vein, and in order to gain a better understanding of the ways in which BELF is used, we need to explore what kind of strategies BELF users employ in their actual online workplace communications.

1.3 Research questions of this study

This study attempts to contribute to the growing body of research on authentic computer-mediated workplace communication through BELF by exploring how requests are performed by employees with different lingua-cultural backgrounds, organisational status and gender. Specifically, we explore what types of requests BELF users prefer, and to what extent their preferences are affected by their status within the organisational hierarchy, their gender, L1 and cultural background. We do so by analysing a unique corpus of 398 internal emails produced by BELF users with a great variety of L1 backgrounds, including Spanish, Italian, Dutch, Hebrew, Chinese, German, French, Swedish and Greek, as well as L1 speakers of English. At the micro-level, we investigate the influence of each individual L1 background on request strategies. At the macro-level, we also compare L1 and L2 speakers of English. Whilst previous research has highlighted the differences between these groups, we predict that within the transcultural communities of practice in BELF contexts, L1 and L2 speakers will converge in their politeness strategies due to mutual accommodation strategies, although these tendencies may be modulated by prevalent extra-linguistic factors, such as power and gender. To summarise, we aim to address the following research questions:

RQ1: What kind of request types are predominantly used in business emails in a multilingual organisational setting?

RQ2: To what extent is the directness of requests modulated by BELF users’ lingua-cultural background, both at the macro-level (comparison English-L1 and English-L2 speakers) and at the micro-level (diversity of lingua-cultural backgrounds)?

RQ3: To what extent is the directness of requests affected by power, and how do any effects of power interact with language background?

RQ4: To what extent is the directness of requests affected by gender, and how do any effects of gender interact with language background?

Thus, we hope to provide insights into the interaction of a range of factors, such as gender, power and language background in shaping BELF communicative practices to increase our understanding of workplace communication in linguistically and culturally diverse settings. In a broader sense, we explore the extent to which BELF users develop and follow communicative norms specific to their community of practice. We predict that the observed phenomena will transgress traditional dichotomies, such as L1 versus L2 differences. Moreover and contrary to previous assumptions, we assume that tensions and miscommunications will be rare due to the accommodation strategies employed by experienced business practitioners and BELF users.

3.1 Types of requests

In the 398 emails analysed, we identified 223 instances of requests, suggesting a high relevance of this speech act in business communication. The email senders used significantly more direct (62%) than indirect requests (38%) (χ² = 232, p < 0.000), so the level of directness of their FTAs was generally high across all email senders in the corpus. When further sub-dividing the request strategies by perspective, we found the majority of emails to take an HO perspective (χ² = 147, p < 0.0000). The majority (69%) of all direct requests and the majority of all indirect requests (90%) were examples of HO requests.

Identified requests were further categorised into types applying Blum-Kulka et al.’s (1989) framework of request strategies. As seen in Figure 1, the majority of direct requests in the data set were MDs (68%).

Figure 1:

Categories of direct requests in the data set.

A common pattern identified in MDs was the combination of the politeness marker ‘please’ with a verb in the imperative creating a formulaic sequence please + verb in the imperative. Yet, while the use of imperative with ‘please’ dominates in the requests, the high level of imposition that the structure involves is not necessarily mitigated with this politeness marker. Research suggests that the use of ‘please’ in written workplace communication has become a conventional requesting routine, especially in British English, rather than a serious ‘softener’ of imposition (Leech 1983; Murphy and De Felice 2019). Using Watts’ (2003) distinction between ‘politic’ and ‘polite’ please, in the context of the MDs studied here, ‘please’ is rather politic in meaning in that its presence does not have much impact on mitigating the FTA, though its absence could lead to the requests being interpreted as impolite (Murphy and De Felice 2019). This is evidenced by the formulaic use of ‘please’ almost exclusively in the initial position. Interestingly, examining the verbs used in the MDs has shown that the semantic choices had to do with the power status of the sender within the organisational hierarchy. In emails sent from high to low-power employees, the email sender often chose verbs that convey telling or ordering the recipient to do something. Below are some indicative examples:

In emails sent from low to high-power employees, the email senders frequently asked for help or sought advice:

It could be argued that asking for help or advice involves a lesser level of imposition than the act of ordering the addressee to do something. In this case, the choice of verbs and not the politeness marker were indicative of mitigation.

After MDs, the next most common direct request strategy (12%) were EPs. The predominant EP pattern was we + (kindly/please) request/ask, as illustrated by the following examples:

Most of these types of requests were produced by high-power employees who tended to choose the SO perspective and used the first person plural ‘we’. The pronoun ‘we’ is quite versatile in terms of the membership that it can index. In the context of this study, it seems to represent the corporate ‘we’; here the requester uses the inclusive pronoun to speak on behalf of the organisation, which increases the force of the request. The addressees are not asked to do something for the requester but for the organisation as a whole and which have employed them. As in the above examples, the use of the politeness markers ‘please’ and ‘kindly’ acts as a conventional requesting routine and not necessarily a mitigation device.

As far as the indirect requests are concerned, the vast majority were instances of QPs (98%). There were no examples of SFs and MHs. As can be seen in Figure 2, these QPs followed the pattern of modal verb + please + verb. Interestingly, we observed an effect that mirrored the semantic differences in verb choices observed for direct request strategies. High-power employees tended to ask the low-power employees to do something, also using the HO perspective:

Figure 2:

Categories of indirect request strategies in the data set.

Conversely, low-power employees tended to seek help or advice from high-power employees:

3.2 The impact of macro-level language differences (L1 vs. English-L2) on request strategies

In order to investigate whether English-L1 speakers and English-L2 speakers manage FTAs differently, we counted the types of requests sent by L1 and English-L2 speakers. The general trend of using more direct FTAs was observed across both groups, the interaction between L1 versus L2 English speaker status and request strategy was not significant (χ² = 0.00, p = 0.99). Both L1 and L2 speakers used 62% direct request strategies (English-L2 speakers: n = 101; English-L1 speakers: n = 38) and 38% indirect request strategies (English-L2 speakers: n = 61; English-L1 speakers: n = 23). Language background (English-L1 vs. English-L2 speakers) on its own was therefore not the determining factor of the directness of requests (Figure 3).

Figure 3:

Types of request strategies used by English-L1 and English-L2 speakers.

3.3 The impact of power status on request strategies in BELF

Our next research question explored the influence of the hierarchical face system framework (Scollon et al. 2011), i.e. the prediction that individuals with higher levels of power within the organisational hierarchy will feel more empowered to use direct strategies than low-power employees. Our analyses revealed that both high and low-power employees used more direct (High-power: n = 63, Low-power: n = 76) than indirect request strategies (High-power: n = 41, Low-power: n = 43), and that power did not interact significantly with the directness of the request strategy (χ² = 2.61, p = 0.61).

To explore whether macro-level language differences (L1 versus L2) would make a difference in this respect, we analysed the effects of power on request strategies in the English-L1 and L2 groups separately. In the English-L1 group, the interaction between power and the directness of the request strategy was significant (χ² = 10.99, p = 0.001): high-power employees produced significantly more direct (n = 38) than indirect requests (n = 17), whereas low-power employees produced exclusively indirect requests (Figure 4). The sample contained a low number of low-power L1 speakers, so we identified only six requests produced by low-power employees, and these were all indirect in nature. In the English-L2 group, the interaction between power and the directness of request strategies was only marginally significant (χ² = 3.84, p = 0.05). Crucially, the trend went in the opposite direction: high-power employees produced nearly equal numbers of direct requests (n = 25) and indirect requests (n = 24), whilst low-power employees produced a greater number of direct requests (n = 76) than indirect requests (n = 37) (see Figure 5).

Figure 4:

Request strategies by power in the English-L1 group.

Figure 5:

Request strategies by power status in the English-L2 group.

In addition, we assessed whether the power status of the email recipients would influence request strategies. When looking at the total corpus, the pattern of preference for direct over indirect request strategies prevailed, regardless of whether the email request had been addressed to a low or high employee (χ² = 0.17, p = 0.68). However, when assessing the impact of recipient power status in the English-L1 and L2 groups separately, interesting differences emerged. Whilst for English-L2 speakers the interaction between recipient power status and request strategy was not significant (χ² = 2.16, p = 0.14), in the English-L1 group the interaction was significant (χ² = 10.00, p = 0.001). Requests emailed to high-power employees (n = 6) were exclusively formulated using indirect strategies, whilst requests occurring in emails addressed to low-power employees were more often direct (n = 38) than indirect (n = 17). This means that in the English-L1 group the patterns observed for both the sender and recipient power were in line with predictions based on the hierarchical face system (Scollon et al. 2011), whilst in the English-L2 group the pattern diverged from what is to be expected based on common socio-pragmatic conventions.

3.4 The impact of gender on request strategies in BELF

The gender comparison revealed a trend towards direct FTAs amongst both male and female employees. At a descriptive level, male employees used more direct strategies (n = 90) than indirect strategies (n = 45), whilst female employees used a similar number of direct strategies (n = 49) and indirect strategies (n = 39). However, the gender variable did not interact significantly with the directness variable (χ² = 2.74, p = 0.098). The same pattern prevailed in the English-L1 and English-L2 groups, in which the interaction between gender and the directness of the request strategy was not significant either (English-L2: χ² = 1.68, p = 0.196; English-L1: χ² = 1.11, p = 0.29).

3.5 The modulatory impact of power and gender on the effects of L1 versus L2 status

To investigate the modulatory impact of the power and gender variables on the effects of L1 versus L2 status, we investigated the interaction of this variable with request strategies in each of the power and gender combinations (high-power female, low-power female, high-power male, low-power male). These analyses revealed that in the high-power female group, the interaction between language status and request strategy was not significant (χ² = 0.18, p = 0.67). The English-L2 females converged with the English-L1 females in displaying the same preference for direct requests (English-L1: n = 13, English-L2: n = 15) over indirect requests (English-L1: n = 9, English-L2: n = 8) (Figure 6).

Figure 6:

Request strategies by English-L1 versus English-L2 in the high-power female group.

In the high-power male group, the picture was different because the interaction between language status and request strategies was strongly significant (χ² = 8.38, p = 0.004) (see Figure 7). This was due to the fact that the male English L2 speakers displayed a pattern opposite to that of the English-L1 speakers. Whilst the English-L1 speakers produced more direct (n = 25) than indirect requests (n = 8), the English-L2 speakers produced more indirect (n = 16) than direct requests (n = 10). Thus, the high-power male English-L1 speakers behaved in accordance with the norms predicted by the hierarchical face system, and in line with the female high-power employees (English-L1 and L2 speakers) in the corpus. The male English-L2 speakers, on the other hand, diverged from the general pragma-linguistic trends observed in this BELF corpus.

Figure 7:

Request strategies by English-L1 versus English-L2 speakers in the high-power male group.

In the low-power groups, the validity of statistical analyses is limited because the English-L1 speaker groups (both female and male) produced exclusively indirect request strategies, which means that no instances of direct requests were generated by low-power English-L1 speakers. Nevertheless, the results for the interactions between language status and request strategies differed in the male and female groups. In the low-power female group, there was no significant interaction between language status and request strategies (χ² = 2.002, p = 0.157). English-L1 speaker females used similar (low) numbers of direct and indirect requests (direct: n = 0, indirect: n = 2), and English-L2 speaker females also used comparable numbers of direct (n = 20) and indirect (n = 21) requests (see Figure 8).

Figure 8:

Request strategies by English-L1 versus English-L2 speakers in the low-power female group.

In the male group, on the other hand, the interaction between language status and the directness of the request strategies was highly significant (χ² = 11.058, p = 0.001). While low-power English-L1 speakers used only indirect strategies when making requests (n = 4), the English-L2 speakers diverged from that pattern by choosing more direct (n = 55) than indirect (n = 17) requests (see Figure 9). As was the case in the high-power group, the female English-L2 speakers thus converged with the English-L1 speakers, whilst the male English-L2 speakers displayed patterns that diverged from male English-L1 speakers by choosing request strategies opposite to those predicted by the Hierarchical Face System approach.

Figure 9:

Request strategies by English-L1 versus L2 speakers in the low-power male group.

3.6 The impact of L1 and cultural background on request strategies

We investigated whether more micro-level fine-grained distinctions between email senders’ lingua-cultural backgrounds (beyond the broader L2-L1 speaker distinction) could explain differences in request strategies. Our analyses suggested that there was a significant interaction between the lingua-cultural background of email senders and the directness of their request strategy (χ² = 29.31, p = 0.001), generating a complex picture (Figure 10). The majority of email senders followed the pattern observed across the whole corpus, displaying a direct request proportion of roughly 60%. This applied to English-L1 speakers (64.4% direct) and English-L2 speakers with L1 German (62.5% direct), Italian (60.5% direct), Chinese (60%) and Spanish (57.1%). However, requests produced by individuals from three lingua-cultural backgrounds stood out as diverging from the general pattern. Firstly, L1-Hebrew speakers produced predominantly direct requests (96.8%). Example (20) from an email exchange below illustrates the pattern found frequently in the corpus:

Figure 10:

Porportion of requests (direct versus indirect) by L1 of the email senders.

Secondly, L1-speakers of both French and Dutch produced below-average proportions of direct requests (French: 33.3%, Dutch: 39.1%) for this corpus. Extracts (21) and (22) are examples of email exchanges