Disagreement Strategies used by Speakers of English as a Lingua Franca in Business Meetings

Numerous studies have attested to the consensus-orientation and cooperative nature of English as a lingua franca (ELF) interactions. However, limited attention has been given to moments of disagreements in ELF communication, with most of the little existing work focusing on disagreements in ELF academic or informal contexts. Consequently, little is known about how ELF users display disagreement in real-life business contexts. For this reason, this study examined disagreement expressions in five ELF business meetings drawn from the VOICE corpus to understand the nature of disagreement in ELF interactions. Following the identification of disagreement instances, the study used Stalpers’s (1995) framework to investigate whether the disagreement was accompanied by a mitigation strategy that reduces its impact. It was found that the examined ELF business speakers express their disagreement in both mitigated and unmitigated forms with a marked preference for using mitigated expressions, indicating that the appropriate linguistic choice for expressing disagreements in a between-company business meeting is a mitigated disagreement form. Another main finding is the frequent use of disagreement utterances, suggesting that ELF speakers do not merely seek consensus, but they also raise objections and state their different opinions. One implication of these findings is that ELF encounters might not be as consensus-seeking and mutually supportive as suggested in previous works. Taken together, the findings of the present study extended the existing body of work on ELF disagreement and, in general, added further to our understanding of ELF interaction.


INTRODUCTION
Numerous studies have attested to the consensus-orientation and cooperative nature of English as a lingua franca (ELF) interactions (Jenkins, Cogo, & Dewey, 2011). Previous research has revealed that ELF speakers employ a variety of interactional cooperative strategies to reach mutual understanding such as repetition (Björkman, 2014;Cogo, 2009), paraphrasing (Alharbi, 2016;Björkman, 2014;Mauranen, 2010), utterance completion (Alharbi, 2016;Björkman, 2014;Cogo & Dewey, 2012) and backchannelling (Alharbi, 2016;Cogo & Dewey, 2012). While much of the early research on ELF interaction mostly focused on such cooperative aspects, more recently, other questions have been raised concerning ELF users' disagreement practices. In particular, ways of expressing conflicting opinions in ELF interactions have been investigated in academic settings (Bjørge, 2012;Komori-Glatz, 2018;Matsumoto, 2018;Toomaneejinda & Harding, 2018), informal settings (Jenks, 2012;Konakahara, 2015Konakahara, , 2017, with little attention to professional business contexts (Marra, 2016). Consequently, little is known about how ELF users display disagreement in real-life business contexts. Thus, the present study examined the disagreement practices used in ELF business meetings in the VOICE corpus (Seidlhofer et al., 2013b) to extend this line of research further. Earlier views of politeness theory maintain that disagreement is a face-threatening act (FTA) since it is the expression of an opposing opinion to that of another speaker (Brown & Stephen, 1987). In contrast, agreement is a more preferred strategy in interaction and better in maintaining a positive face. Consequently, while displaying disagreements is regarded as uncomfortable and unpleasant, showing agreement is a comfortable and even sociable act. To soften this negative impact, speakers can mitigate their disagreements by using token agreements, white lies, or hedges. This implies that disagreement in earlier accounts of politeness theory "has mostly been seen as confrontational and should thus be mitigated or avoided" (Sifianou, 2012(Sifianou, , p. 1554. A more positive approach towards disagreement is found in recent pragmatics research, which has proposed a more nuanced understanding of disagreement by showing that it is not necessarily a sign of conflict (e.g., Angouri, 2012;Sifianou, 2012). The latter perspective is adopted in this paper as it is more consistent with the aim of the present study, i.e., to examine expressions of ELF disagreement and highlight their role in advancing business negotiations.
As the act of expressing disagreement itself does not necessarily result in conflict, such acts should not always be analyzed as an instance of impoliteness (e.g., Konakahara, 2017). In fact, expressing disagreement in a business meeting may foster more sincere and productive discussions among interlocutors and thus be effective in reaching a decision, securing a deal, or convincing a client (e.g., Marra, 2016). In business contexts, in particular, the ability to express disagreement appropriately is an important skill. Working in business sectors requires conducting negotiations and weekly meetings in which expressing disagreement is sometimes unavoidable. In other words, active participation in such business activities necessitates that ELF users can communicate their opinions by going for or against the topic at hand. The problem is in the case of voicing disagreement in such situations as ELF business professionals are expected to maintain a working rapport with the opposing party while doing so.
As expressing disagreement can help in developing conversations in a business context, analyzing how ELF speakers negotiate their disagreement might reveal some of the strategies followed in showing one's disagreement. The aim of this study is to examine disagreement in an ELF business setting in order to show how disagreement expressions are produced by business professionals.

English as a Lingua Franca in Business Communication
In an international business context, the use of English as a shared communication code both within and across companies has become common. This is referred to as ELF, which, in a narrow sense, describes the use of English in communication by two or more speakers who do not share a primary language (Jenkins, 2007). A typical example is when business people from China and Saudi use English to communicate with each other in their meetings rather than using either Chinese or Arabic. However, the definition of ELF does not strictly apply to communication that takes place among non-native English speakers. A broader view of ELF includes interactions between native speakers of English and non-natives (Nickerson & Planken, 2016). This occurs when some participants in a business meeting speak English as a first language (L1) and others that do not (e.g., L1 speakers of Korean, Spanish, Japanese…etc.). The examined sub-corpus in the present study contains spoken interactions between ELF speakers who have different L1s with no native speakers of English, suggesting that it meets the conditions set by the narrower view of ELF interaction. Another defining characteristic of ELF as a language variety, in both a narrow and broad sense, is that its speakers concentrate on communicating the message rather than on grammatical accuracy (Jenkins et al., 2011). In other words, ELF interactants focus more on the accuracy of the content instead of the linguistic accuracy of the utterance.
Studies examining business discourse further developed the concept of ELF by introducing the term Business English Lingua Franca (BELF) (Kankaanranta & Louhiala-Salminen, 2013;Louhiala-Salminen, Charles, & Kankaanranta, 2005). The term BELF has been proposed to highlight the domain of use, giving more attention to the business context of ELF interactions and its impact on communication rather than the ELF variety itself. Similar to ELF, the language of BELF is characterized as a simplified, hybrid, and highly dynamic one. However, unlike ELF, being a competent user of BELF requires both (a) delivering accurate messages and (b) understanding/using business-specific vocabulary and genre conventions (Kankaanranta & Planken, 2010). Table 1 below clarifies some of the main aspects of BELF communication by contrasting it with the concept of English as a Foreign Language (EFL). The main idea behind this comparison is to demonstrate the fact that BELF speakers are not learners of English, and therefore they do not aim to speak like English natives but rather to get the message across and "get the job done". As such, BELF speakers are "communicators in their own right" rather than non-native speakers nor learners of English.
Based on this table (1), we can see that competence in BELF depends on several factors, including having (a) adequate business communication skills, (b) accomplishing the job while maintaining a working relationship. This suggests that experienced BELF users are not only expected to be knowledgeable in their field, but they also need to manage their interpersonal relationships, indicating that one integral aspect of BELF competence is maintaining a good relationship and managing rapport. Another aspect of BELF communication is that English is seen as a contact language that is "neutral and no longer associated with any of the nations that speak it as a first language" (Nickerson & Planken, 2016, p. 16). Put differently, BELF is a language that belongs to everybody and is not owned by native speakers of English.
In addition to the ELF/BELF characteristics mentioned above, another emphasized feature of BELF/ELF oral communication is their cooperative, consensus-seeking, and mutually supportive nature (Firth, 1996;Jenkins et al., 2011;Seidlhofer, 2001). Examining international business telephone calls, Firth (1996) coined the let-it-pass principle to describe the tendency of ELF interlocutors to ignore grammatical inaccuracies and other non-standard features of English. Over time, ELF speakers have been described as tolerant as they ignore communication problems, which will be either resolved later as the conversation progresses or is not important enough to be addressed (Jenks, 2018). However, although the majority of previous works on ELF interaction have demonstrated the highly cooperative nature of ELF communication (see the Introduction), focusing primarily on the communicative strategies employed by ELF speakers might capture a distorted picture of this type of interaction. Indeed, this research focus has raised some concerns in relation to its depiction of ELF speakers as one-dimensional individuals who mostly tend to cooperate and rarely confront and object to some points in their communication (Jenks, 2012(Jenks, , 2018Matsumoto, 2018). Following this line of research, the present study investigated disagreement expressions in ELF business meetings to explore how ELF speakers in such a setting voice their objections, a topic that has been relatively underexplored. The next section reviews some of the works that examined ELF disagreement in different settings to show that ELF speakers may not always be supportive and that they sometimes have to deal with confrontational moments.

Previous Studies on ELF Disagreement
As was mentioned in the Introduction, much of the existing research on disagreement in ELF interactions is confined to academic and informal contexts with little focus on business ELF speakers. A common finding in these studies is the frequent use of disagreement strategies in ELF interactions with a general preference to use less direct ways to disagree (e.g., hedging, gaze, gestures) than explicit ones (e.g., negation), indicating that using indirect means for expressing one's disagreement is the more appropriate choice in most ELF settings.
The majority of studies on ELF disagreement acts were conducted in an academic setting (e.g., Bjørge, 2012;Komori-Glatz, 2018;Matsumoto, 2018;Toomaneejinda & Harding, 2018). One main study is by Bjørge (2012), who followed a comparative approach by comparing the linguistic forms for disagreeing found in business English textbooks and the ones used by international students in simulated negotiations. It was found that the student group rarely used the one-sentence expressions of disagreement recommended by their textbooks, suggesting a mismatch between the actual conversational data and that presented in the textbooks. Another finding is that the student ELF speakers used more mitigated disagreement by employing strategies such as "delay (e.g., of course, yeah I know well) in addition to added support (e.g.., because … Sundays)" than unmitigated disagreement which involves "blunt contradictions frequently preceded by but (e.g., but that is not the issue here)" (Bjørge, 2012, pp. 15-17). Similarly, Toomaneejinda and Harding (2018) found that academic group discussions among graduate students were rich in disagreement acts which were carefully expressed in a non-oppositional manner.
Other works looked at less formal settings by examining how ELF interactants voice their disagreement in casual conversations and online chat rooms (Jenks, 2012;Konakahara, 2015Konakahara, , 2017. For example, Jenks (2012) analyzed disagreement in multi-participant voice-based chat rooms and reported that ELF speakers raise their objection through laughter, joking, and ridicule. Meanwhile, Konakahara's (2017) study examined how two friends living in the UK as international students manage adversarial moments in casual conversations. The results show that the examined ELF speakers use implicit linguistic expressions in a pragmatically sensitive manner.
Although several works have examined academic and informal ELF registers, only one study investigated disagreement in ELF professional business interactions. Marra (2016) analyzed four hours of business ELF interaction recorded at a large-scale international business event in Germany to examine how business professionals express and handle disagreement. It was found that most instances of disagreement are mitigated through the use of mitigation devices such as delay and added support rather than unmitigated expressions in the form of blunt contradictions. A further observation is that disagreement in business seems to have a dual purpose that of (1) increasing clarity and (2) indicating the speaker's stance in the argument.

The Present Study
As the literature review has shown, several works have looked at ELF speakers' disagreement practices in different contexts, including academic discussions, informal chatting, and business meetings. The growing interest in disagreement acts in ELF interaction has demonstrated that such interactions are rich in disagreement expressions, suggesting that ELF speakers do not always seek consensus. However, this interest has been limited to academic and informal contexts, leaving the business register relatively underexplored. To further advance this line of research on ELF disagreement, the present study followed a corpus-based approach to examine ELF business meetings in the VOICE corpus (2013). This approach is expected to reveal disagreement patterns in ELF business interactions that are most likely representative of different business contexts rather than limited to one company/country context. In other words, analyzing a spoken corpus of natural business ELF interactions recorded in various countries allows capturing a general picture of how ELF speakers disagree with their interlocutors in a business meeting. With this in mind, the following research questions were addressed: 1) How do ELF speakers express their disagreement in business meetings? 2) What are the mitigation strategies used in expressing disagreement in this context?

METHODOLOGY
Following previous works (Bjørge, 2012;Marra, 2016), a corpus-based conversation analytic approach was followed in the present study. This approach allows the examination of disagreement expressions in a relatively large data pool, capturing patterns of their use that otherwise would not be easily obtained. An analytic conversation method was also employed in examining the corpus-based data to take into account the context of the interaction and what goes before and after the disagreement expression. A further reason for employing this method is that it puts more emphasis on analyzing the data from the perspective of the interactants, allowing the researcher to detect both direct and indirect disagreement expressions (Firth, 1996).
To do so, this study analyzed ELF meetings drawn from the VOICE corpus (POS XML 2.0 Version) which contains 151 audio recordings of "naturally-occurring, non-scripted, face-to-face interactions" (Seidlhofer et al., 2013a, para. 5) spoken by ELF speakers with 49 first languages backgrounds. The recordings took place in a variety of settings, and as such they cover five different discourse domains (e.g., educational, leisure, professional business) and ten speech types (e.g., conversation, interview, meeting). What makes this corpus particularly valuable is that it describes the participants and topics of the interactions, making it possible to interpret data in relation to the meeting context. More specifically, the number of speakers, their relationship (e.g., fairly acquainted, predominantly acquainted), their job roles (e.g., manager, assistant), and the topics discussed in the interaction are described for each recording.

Corpus Construction
To analyze disagreements expressions in ELF business meetings, the first step involved creating a sub-corpus of five meetings from the professional business domain found in the VOICE corpus (Seidlhofer et al., 2013b), making a sub-corpus of 119, 142 words. The number of different speakers involved, the duration, and the total words of each recording in the constructed sub-corpus is shown in Table 2. The size of the collected data was restricted by the fact that the manual analysis of context-based disagreement expressions requires a close examination of every word in the sub-corpus (119, 142 words). Although only five professional business meetings were examined, a sufficient amount of data was obtained in the analysis as 87 disagreement acts were found and were revealed to have some patterns in terms of their use as the results section would show.

Data Analysis
Each text of the recorded meetings was read, and then a manual count of disagreement instances in the data was conducted.
Then, to answer the first research question, the identified dis-agreement expressions were labeled as mitigated or unmitigated. The second research question was answered by using Stalpers's (1995) categorization of mitigated disagreement acts which groups mitigation strategies into three types; (1) delay, (2) added support, and (3) modulation/indirectness. However, the data analysis in the present study revealed that not all indirect disagreement acts are modulated. Therefore, the present study considered the two mitigation strategies, indirectness, and modulation, as two separate categories rather than one as was done in earlier research. Furthermore, not all disagreements include a mitigation strategy. As such, the present study qualified such disagreement instances as unmitigated. Table 3 below shows the categorization of disagreement acts along with a brief explanation and examples for each act. This slightly changed categorization was used in analyzing the constructed sub-corpus.

RESULTS
Overall, the five analyzed meetings contained 87 disagreement acts, averaging around 17 (SD = 5.5) for the whole dataset, as shown in Table 4. Although each examined meeting (A, B, C, D, E) contained different frequencies of disagreement acts, the following analysis shows that there is an overall tendency in the sub-corpus to use certain types of disagreement strategies, suggesting relative data homogeneity.

First Research Question: Mitigated or Unmitigated
One observed pattern in the present sub-corpus is that ELF speakers tend to express their disagreement in a mitigated manner. The analysis of the business meeting sup-corpus showed that there is a preference for using mitigated, less direct disagreement expressions (69%) than unmitigated ones (31%). Table 5 demonstrates the higher frequency/percentage of mitigated disagreement acts compared to unmitigated ones, indicating that mitigating one's disagreement is more pragmatically appropriate in a business meeting context. In other words, it is more socially acceptable for EFL users in the business sector to express an opposing view in a mitigated manner to soften the impact of the disagreement rather than bluntly stating it without giving space for further negotiation.
The difference between mitigated and unmitigated disagreement acts is clearly visible in the following two extracts where more attention would be given to how unmitigated disagreements are expressed as the remaining of this section would focus on mitigated strategies. Note that the disagreement act in all the data extracts in this paper is in bold, and the strategy used to express it is enclosed in brackets. Also, the formatting and spelling of the extracts were mostly not altered; they were presented in this paper as they were shown in the VOICE corpus.
Starting with the mitigated disagreement act in extract A, the interaction in this extract is between two speakers who are managers from two different companies and were talking about another employee that is not present in the meeting. Speaker 1, a manager in a Korean distribution company, was informing speaker 4, a manager in a German food company, about the job positions taken by his coworker. In line 1, The symbol # means number.
3. S1: that was his last position. 4. S4: mHM 5. S1: <soft> okay </soft> e:r 6. S4: from cosmetics to sweets. 7. S1: but we w-actually we: HANDLING so many [mitigated disagreement] 8. S4: right 9. S1: different products. [mitigated disagreement] The mitigation in extract A was accomplished using two techniques. The first is using mitigating linguistic items such as 'but' and 'actually', serving to convey the speaker's divergent perspective and orienting the listener to that perspective. The second technique used to mitigate the disagreement is "focus shift" which occurs when a speaker raises an alternative claim to oppose an earlier proposition, introducing new dimensions to the topic discussed (e.g., Georgakopoulou, 2001;Osvaldsson, 2004;Toomaneejinda & Harding, 2018). Here, we can see that speaker 1 shifted the focus in line 7 and 9 from speaker's 4 observation that the company buys and sells "sweets" to the fact that it handles "different products," including but not limited to sweets. Focus shifting allowed speaker 1 to express his disagreement in an indirect manner. Speaker 1 could have directly disagreed by denying speaker's 4 claim that the company only distributes sweet food by responding, for example, "no we do not handle only sweets but also many different products". Rather, speaker 1 avoided directly addressing the point raised by speaker 4, preferring to change the focus to the fact that the company is a distributor of various products.
Unlike extract A, the disagreement act in extract B is unmitigated, directly stating the speaker's opposing view without the use of any mitigation device. In extract B, speaker 4 was suggesting in line 1 that company 1 and company 4 do the same work. While at first speaker 1 agreed with this (line 2), he later completely disagreed (line 4) shortly after speaker 3 announced his disagreement in line 3. Speaker's 3 response to speaker 4 was simply "no", indicating that the Table 3. Classification of disagreement along with illustrative examples from the examined data (adapted from Stalpers (1995, p. 278) and Bjørge (2012, p. 5

Disagreements Sub-types Explanation Examples
Mitigated Delay stalling the disagreement act by a pause, a discourse marker, a qualifier, token agreement, appreciation or apology, hesitation, or by being displaced over more than one speaking turn.  Total (SD) 87 SD = 5.5) speaker 1 mentioned that his coworker had worked for a company that distributes cosmetic products before joining their company. In line 4, speaker 4 comments on this topic in one clause "from cosmetic to sweets", indicating that speaker's 1 coworker seemed to have moved from the cosmetic industry to a sweets distribution company, the place at which speaker 1 currently works. Speaker 1 disagrees with speaker's 4 comment by explaining in line 7 and 9 that he works for a company that sells "different products" rather than only sweets, as was suggested earlier by speaker 4. A. Extract from meeting A (PBmtg3) 1. S1: he worked for [org12] [org13] and e:r some local er importing company selling DISTRIBU-TION company selling er er cosmetics. 2. S4: mhm two companies do not perform the same work. Speaker's 3 abrupt, blunt disagreement to speaker 4 is unmitigated as no mitigation strategy was used and the opposing stance was explicitly stated. This was partially accepted by speaker 4 as he responded in line 5 and 6 with the statement that in some cases, the two companies provide different services rather than entirely identical ones as was suggested by him earlier.
Thus, while extract A included some mitigation devices, the disagreement in extract B includes none, closing the space for negotiation about the discussed topic. B. Extract from meeting B (PBmtg269) 1. S4: and they are paid with a: monthly fee for [org1]? and now they will have a fee for [org4] but it is the same work? 2. S1: this is the same 3. S3: <loud> no </loud> [unmitigated disagreement] 4. S1: no it's not t's not the SAME work 5. S4: in some 6. S4: in some in some cases Interestingly, using an unmitigated disagreement act in the examined data seems to be done by speakers from different organizational levels, indicating that simply saying 'no' to disagree with an interlocutor's proposition is not restricted to those in higher and more powerful positions. For instance, in extract (C), speaker 6, a junior worker in a German forwarding agency, directly disagreed with the sales representative of a Dutch airline company, speaker 2. Perhaps the humorous atmosphere of this interaction between speaker 2 and 6 as indicated by speaker's 6 laughs in line 2 made expressing this type of disagreement possible despite the unequal relationship between the two speakers. C. Extract from meeting C (PBmtg300) 1. S2: yeah well it's not a good scenario and by THIS if you had this opportunity to work TOGETHER it will give you more strength to do d-so. and it's er well it's good because SHE (referring to S6) maybe can explain to you for hey [S1] you're doing things like this maybe WRONG 2. S6: @@@@@ [laughter] 3. S2: no but it gives YOU the motivation no? er as a as an 4. S6: no just to be fair [unmitigated disagreement] This finding may imply that ELF speakers in a business context occupying different organizational positions can bluntly disagree with each other if the atmosphere is friendly and humorous. Business ELF interactants with equal or unequal status might use unmitigated forms when expressing an opposing stance if the atmosphere tends to be friendly. This observation is further supported by meeting D data, which was full of laughter and humorous remarks and had higher instances of unmitigated disagreements (n= 5) than mitigated ones (n = 4). This might suggest that the friendlier the interlocutors, the more direct the disagreement form.

Second Research Question: The Mitigation Strategies Used
Further, the use of disagreement mitigation strategies in the present data was analyzed to understand the nature of disagreement in ELF business meeting. The analysis in Table 6 reveals that the ELF speakers in the five business meetings were more likely to use indirect (67%), and delay (20%) strategies when expressing their opposing opinion, while minimally using added support (12%) and modulation strategies (1%). The following subsections examine each strategy in an illustrative example to show how it was used by ELF speakers in the present study.

Indirectness
The most used mitigation strategy in the analyzed sub-corpus is indirectness, constituting 67% of all identified mitigated disagreement acts. In the literature, this strategy is sometimes called focus shifting which occurs when a speaker takes a neutral stance by not rejecting the controversial statement and attempts instead to change the focus in his response to a different point that was not addressed in that statement (e.g., Georgakopoulou, 2001;Osvaldsson, 2004;Toomaneejinda & Harding, 2018). Thus, rather than explicitly negating the statement, the speaker would introduce a new topic in his reply to express his disagreement. This was demonstrated in extract A above, and this seems to apply to extract D as well.
Extract D starts with speaker's 1 statement, a representative of a distribution company, that his team cannot give a food distribution plan to speaker's 4 company as was requested. Speaker 2, speaker's 1 coworker, pleaded that the other party should wait a "little more" to get the requested plan. However, speaker 4 disagreed with speaker's 1 suggestion, maintaining in line 5 that her team's work depends on having this plan. This disagreement was done in an indirect way as speaker 5 avoided expressing her rejection of the idea explicitly by saying, for example, "no that is not the ideal thing for us, we need to receive plans from your side". Instead, she preferred to highlight the importance of plans for her work, to introduce a new dimension to the topic being negotiated and send at the same time a down-toned disagreement message to the second party. This Introduction of a new perspective on the role of plans in the work of speaker's 5 company serves as a focus shift. The topical focus is no longer on the creation of plans but rather on the key role they play in the work progress in speaker' 5 company. D. Extract from meeting E (PBmtg463) 1. S1: if it's (.) we we wou-what would be the ideal thing for us it would be (.) not to give any plan (.) just to 2. SX-3: yeah but 3. S2: to wait little more 4. S1: you canno:t Added support 7 12 Modulation 1 1 5. S4: yeah but we have to work with plans because we have to [mitigated disagreement: indirectness] 6. S1: yes i know The high frequency of indirect disagreements in the present study may be explained by the observation that they serve two purposes. This form of mitigation can be used when interlocutors wish to express their disagreement in a serious business negotiation setting and simultaneously maintain the working relationship between the two opposing parties, posing little threat to this relationship (Brown & Stephen, 1987). Another observed function of this mitigation strategy is that it helps in informing the second party about a new fact or point that could not be otherwise known if the first party did not shift the focus in her/his disagreement. Thus, mitigating a disagreement using the indirect strategy in a professional business context allows the speaker to disagree with the addressee's proposition while maintaining their working rapport and making the addressee take into account a new dimension of the topic being negotiated.

Delay
The second most used mitigation strategy in the data is delay (20%) in which the speaker partially agrees with the idea initially and then introduces his disagreement. This tones down the explicitness of the disagreement expression as it signals that the upcoming utterance will bring a different view to the discussed topic, rendering delayed disagreement acts as "a rapport-management device" (Bjørge, 2012, p. 14). In extract E, speaker 4 seems to do so by using four delaying devices in his utterance in line 1, including prefaces "the only thing here, I mean" a comment expressing appreciation "it's a very nice display", and the conjunction "but". Speaker 4 introduced various mitigating elements to reduce the impact of his actual disagreement on the possibility of getting approval from 'the licenser of hello kitty'. E. Extract from meeting A (PBmtg3) 1. S4: the only thing HERE i mean it's a very nice display but [mitigated disagreement: delay] 2. S1: mhm (.

Added support
This mitigation strategy was minimally used, accounting for 12% of mitigated disagreement use. Using this strategy, the speaker can justify the disagreement to show the validity of his view and ultimately persuade the addressee of the point being raised. An example of a disagreement supported by an explanation is shown in extract F. In this extract, speaker 1 and 2 are employees representing two different companies and are discussing the increased demands for air cargo in speaker's 2 airline company. While speaker 1 sees that the current situation is beneficial for the airline company and does not suspect that the situation would turn bad, speaker 2 does not agree with this comment. In line 8, speaker 2 initially concedes to speaker's 2 point but shortly after he introduces a 'but' to argue that if the situation continued, it would lead only to 'danger'. The added support for this disagreement is laid out in the next utterance (line 9) in which speaker 1 continues to explain why the airline company might be facing a problem soon. F. Extract from meeting C (PBmtg300) 1. S1:so at the end of the day y:ou play (.) you pay your round trip (.) 2. S2: yeah. 3. S1: with import cargo (.) 4. S2: yeah. 5. S1: have already PROFIT on there and whatever you put on here is er (.) 6. S2: yeah. 7. S1:as long as your (admin) is is covered e:r (.) 8. S2: yeah yeah but the danger the danger 9. S2: [S1] the DANGER is quite near(er). (.) look (1) the westbound is now okay? (.) but the eastbound e:r as we all know rates are whhhh @@ (.) and (.) you see this IMBALANCE (.) is getting closer to the the break EVEN more and more because the (.) [mitigated disagreement: added support] It should be noted that the added support strategy mostly occurs in combination with another mitigation device in the analyzed sup-corpus. Out of the seven instances of added support, six of them occurred following the delay strategy (e.g., yeah yeah), suggesting that disagreement statements adding support are rarely present in the business context. A combination of delay followed by added support might be more communicatively suitable for the examined ELF speakers who aim to maintain their working relationship by delaying the disagreement and support their points by using adding support elements.

Modulation
The most rarely used mitigation strategy in the data is modulation, occurring only once in the sub-corpus (1%). Modulation describes when a speaker expresses her/his disagreement using use modal verbs and is mostly included with indirectness under the same mitigation category, i.e., modulation/indirectness category (Bjørge, 2012;Stalpers, 1995). However, the data examined in this study showed that the two categories are not mutually inclusive; an instance of indirect disagreement can occur without the use of modal verbs. To account for this observation, the present study considered the two mitigation strategies, indirectness, and modulation, as two independent categories. The only instance of modulation in a disagreement statement in the present study is provided in extract E.
In the following extract, speakers 3 and 2 work for different companies and are discussing the color design of a product's label that would be distributed by speaker's 2 company. It starts with speaker 3 trying to inform speaker 2 about the color of the packaging (line 1) and then trying to convince her about the need to use the color green for the product's packaging (line 3). In response, speaker 3 indirectly disagrees with this idea by using a modal verb 'would' to indicate her dislike of having the package colored in this way.
The use of modal verbs helps in mitigating the disagreement and presenting it as one possible view out of many options, giving space for more negotiation among business ELF interactants. G. Extract from meeting E (PBmtg463) 1. S3: <to S2> but it is the main packaging will be green. cos you need er </to S2> 2. SX-7:(decorate) 3. S2: <to S3> back back color </to S3> 4. S3: <to S2> you need a decoration you need you need a color that you can (read) the decoration </to S2> 5. S2: <to S3> but this would be better not to be the same color like the the this </to S3> [mitigated disagreement: modulation] As modulation had only one occurrence in the examined business meetings, its use might not be preferred by ELF speakers in such meetings. One possible reason for this limited use is the somewhat formal nature of modal verbs (e.g., could, may, might, shall, should, must, would) which might not be appropriate for all business meeting settings as they have different formality levels depending on the context in which a meeting takes place. For instance, the speakers in the examined meetings were mostly in familiar terms; the majority of them have met each other at least once before the meeting. Due to this familiarity, the examined ELF speakers might not have found it appropriate to use more formal wordings, including modal verbs, while expressing their disagreement.

DISCUSSION
The analysis showed that ELF business speakers express their disagreement in both mitigated and unmitigated forms, with a marked preference for using mitigated expressions. This finding was also reported in two previous studies by Bjørge (2012) and Marra (2016). Another pattern found in the current data is that there are specific mitigation strategies drawn on in ELF business meetings, indirectness being the most common way of expressing disagreements followed by the delay strategy, while little use was made of added support and modulation strategies. These findings are further discussed below in relation to previous studies. Nevertheless, the present findings might suggest that the appropriate linguistic choice for expressing disagreements in a between-company business meeting is the mitigated disagreement form.
Additionally, the finding that a total of 87 disagreement instances were used by ELF speakers in a small sub-corpus of 119,142 words may indicate the frequency of such expressions in professional business spoken communication, constituting 7% of all utterances in the sub-corpus. Expressions with a disagreement function are considered frequent in the examined sub-corpus because in corpus-based research a sequence of words with a discourse-level function is regarded as frequent if it reaches the threshold of 10/million, i.e., comprising 0.5% and more of the whole corpus (D. Biber, Conrad, & Cortes, 2004;Douglas Biber, Johansson, Leech, Conrad, & Finegan., 1999;. Based on this, it is more likely that ELF speakers voice their disagree-ment rather than avoiding stating their different opinions in business meetings, suggesting that ELF encounters are not always consensus-seeking and mutually supportive. This finding supports earlier ELF research on disagreement acts in classroom discourse (Matsumoto, 2018), academic group discussion (Toomaneejinda & Harding, 2018), and online voice-based chat rooms (Jenks, 2012), reporting that ELF interactions contain a variety of disagreement expressions. Similarly, the results of this study reveal that ELF speakers in a business context frequently express their opposing views, suggesting that encountering and expressing different views is expected in a business meeting setting. One possible explanation for this is that disagreeing with someone in such a context is necessary for successful negotiation because it builds a background for clarifying details about the topic discussed. For example, a disagreement act might trigger the addressee to discuss in great detail some aspects that were not clear for the opposing party, leading both parties to reach a better understanding of the topic. Also, business ELF speakers' frequent use of disagreement helps in reaching a decision on aspects related to on-going projects, suggesting that expressing disagreement and challenging the interlocutor's proposition is pragmatically appropriate in ELF professional business meetings and an important skill for the participants taking part in these meetings. The results of the present study highlighted moments of disagreement in professional ELF business meetings to "problematize the tendency to describe of ELF interactions as collaborative" (Matsumoto, 2018, p. 36), further extending ELF research on disagreement by investigating an underexplored context that of business meetings.
Although the analysis showed the frequent use of disagreement expressions, it was observed that most of them were mitigated using several linguistic strategies that reduce the force of the expression. It seems that the examined ELF speakers were simultaneously performing two tasks while expressing disagreement, disagreeing with the addressee and maintaining the relationship between the two parties. In other words, expressing disagreements in business ELF communication requires using pragmatically sensitive ways that do not pose a threat to the working relationship. This may explain the higher number of mitigated disagreements in the data as the use of such mitigation elements helps in managing rapport, a required skill for ELF speakers in the business world (Kankaanranta & Louhiala-Salminen, 2013).
Out of all the four mitigation strategies, only one (i.e., indirectness) accounted for more than half of the identified mitigated disagreement instances. Indirectness or focus shifting was the most commonly used strategy in the present study whereby ELF speakers do not explicitly communicate their different stance on the topic but rather highlight in their response a new dimension that was not mentioned in the addressee's proposition, introducing a slight change to the discussed topic. While this strategy was the most preferred in graduate students' group discussions (Toomaneejinda & Harding, 2018), it was the least used in business students' simulated negotiations (Bjørge, 2012) and business professionals' meetings in a convention (Marra, 2016). One possible reason for this difference is that Bjørge (2012) explored disagreement acts of students in a hypothetical business scenario while the examined ELF speakers in the present study are professional business people and are interacting in a real meeting. Also, the fact that indirectness as a mitigation strategy was commonly used in the present study might be explained by the long duration of the examined five meetings. The interactants in the current study had no less than one hour for each meeting, while the average length of Marra's (2016) recorded meetings were around 33 minutes. This indicates that the ELF speaker in the present corpus had sufficient time to negotiate their disagreement and use indirect ways to express it, unlike those in Marra (2016).
Delay was the most second frequently used mitigation strategy in which a linguistic element is presented in a clause-initial position to stall the disagreement and orient the addressee to the different coming stance. Likewise, Bjørge (2012) and Marra (2016) revealed the common use of this strategy by both ELF business students and professionals while voicing their divergent opinions in spoken face-toface negotiations. This is also noticed in the analysis of the present corpus, showing that the examined ELF speakers favor delaying their disagreement to reduce its impact on the addressee and maintain rapport among the speakers.
The results also showed that the examined ELF speaker minimally included added support and modulation elements in their disagreement statements. Added support involves giving arguments to strengthen the speaker's disagreement, while modulation describes the use of modal verbs in expressing an opposing perspective. Previous research reported that added support was frequent, while modulation was not present in their data (Bjørge, 2012;Marra, 2016).
It is possible that there were few instances of added support in the present study because the analyzed corpus is different from the data examined in Bjørge (2012) and Marra (2016). Unlike the student business meetings in Bjørge (2012), the interactants in the present study's corpus are far more experienced; they have been working for years in the industry, making them knowledgeable about the ins and out of their job. Knowing that the addressee has knowledge in the field might explain why the current study's ELF speakers who initiated disagreement did little work to support it. However, the added support strategy was commonly used in Marra's (2016) study, which examined meetings from a large international business event setting, a similar corpus to the present study, i.e., ELF meetings in an international company context. Nevertheless, the meetings in Marra's (2016) work and this study are held in somewhat different situations, one is convention halls (Marra), and the other is company rooms (this study). It might be interesting for future research to investigate differences of ELF business negotiations in these two situations.
As for modulation, similar to what was reported in previous works (Bjørge, 2012;Marra, 2016), this study found only one instance of its use in the analyzed corpus. This may be explained by the observation that modulation requires the use of more formal verbs such as could, would which might not match the level of formality in the examined meetings, exhibiting largely more relaxed atmosphere due to the small number of participants ranging from 5 to 7 people for each meeting. It should be noted, however, that the present study considered modulation and indirectness as two distinct mitigation strategies, unlike earlier works lumping the two under one category (Bjørge, 2012;Marra, 2016). This more granular analysis allowed the current study to spot disagreement instances that use indirectness without modulated elements, indicating that the two should be treated as two related but autonomous mitigation strategies.

CONCLUSION
This study aimed to examine how ELF business professionals express their disagreement in company meetings and investigate whether they used certain strategies while doing so. The current study analyzed expressions of disagreements in five ELF business meetings drawn from the VOICE corpus to extend our knowledge about the nature of disagreement in ELF interaction by investigating an underexplored setting within this research framework, i.e., business meetings. The analysis of these meetings revealed two major findings. One is that mitigated disagreement forms are more widely used than unmitigated ones, possibly indicating that the appropriate linguistic choice for expressing disagreements in between-company business meetings includes an expression with a mitigated device. It seems that the examined ELF speakers chose to mitigate their disagreement to reduce its impact and ultimately avoid posing a threat to the working relationship, a requirement for maintaining and building connections in business communication (Kankaanranta & Louhiala-Salminen, 2013). Another main finding is the frequent use of disagreement utterances in the present sub-corpus, a finding that aligns with previous research on ELF disagreement/uncooperative acts. The frequent occurrence of such utterances in the current study might show that ELF speakers do not always seek consensus, but they also raise objections and state their different opinions. Taken together, the findings of the present study extended the existing body of work on ELF disagreement and, in general, added further to our understanding of ELF interaction.
Several limitations should be noted, and possible areas for future research can be suggested. Although the study aimed at exploring ELF speakers' disagreement acts in company meetings, the examined ELF speakers are mainly native speakers of Indo-European languages (e.g., Dutch, German, French), making the present sub-corpus less representative of ELF speakers of other native languages such as Mandarin, Tamil, Indonesian, Tagalog, Arabic, among others. Future research may examine disagreement expressions in a well-representative corpus of ELF speakers to ensure the coverage of a wide range of first languages. Another limitation is that the identification and coding of disagreement utterances were carried out by the researcher with no additional rater. This poses a problem as coding disagreement is subjective and may vary from one rater to another due to the interpretive nature of this type of analysis. Thus, future research should ensure that the coding is reliable by using two raters and conducting an inter-reliability test to check for coding consistency and reliability. Possible future areas of research include (a) comparing between disagreement expressions in spoken and written ELF business professional registers (e.g., emails versus meetings), (b) investigating the difference between disagreement expressions in different types of ELF business meetings (meeting in companies versus conventions), and (c) analyzing both non-linguistic (e.g., gaze, gestures) and linguistic disagreement acts in ELF intra-company business meetings.