Competent non-native users of English ? Requestive behavior of Norwegian EFL teachers

This paper reports the results of a study of English requests produced by Norwegian EFL teachers. The data were collected using a discourse completion task consisting of four scenarios eliciting high and low imposition requests. Head acts of requests and internal and external modifications were analyzed. The findings reveal a complex requestive behavior and sensitivity to the social context of the interaction and suggest that the teachers in the study have developed advanced pragmatic competence.


Introduction
Pragmatic competence, defined as the ability to perform and understand speech acts in a way that appropriately conveys the communicative intent (Celce-Murcia, Dornyei & Thurrell 1995), has been studied extensively by scholars working within the field of interlanguage pragmatics.It entails knowledge of language functions and knowledge of speech act sets, and accounts for abilities such as making and breaking engagements, formulating refusals, agreeing and disagreeing, giving orders, expressing complaints and regrets, and making promises and predictions.
One of the manifestations of pragmatic competence is the performance of speech acts, which often poses a challenge for non-native speakers because, "linguistic, social, and pragmatic knowledge must all be activated and work together in harmony for a speech act to be successful" (Harlow 1990: 328).Even very advanced adult language learners are often unable to perform speech acts in a manner that is deemed appropriate by native speakers (Bardovi-Harlig & Hartford 1990), which can lead to pragmatic failure.Thomas (1983) distinguishes between two types of pragmatic failure, namely sociopragmatic and pragmalinguistic.A language user who fails to perform a speech act expected under given social and linguistic circumstances (e.g., does not apologize for a committed offense) is guilty of sociopragmatic failure, while a language user who performs a speech act using inadequate linguistic means (i.e., by negatively transferring conventional strategies from the first language (L1) to the target language) commits pragmalinguistic failure.Both types of pragmatic failure tend to be judged more harshly than grammar errors because, unlike grammatical mistakes, they are interpreted as demonstrating a lack of politeness and the result of language user's personality rather than the inability to formulate grammatically correct utterances.
In Norway, the new school curriculum introduced in 2006 and known as Knowledge Promotion Reform (LK06) stresses the importance of attaining high levels of pragmatic competence in English.The learning outcomes specifically include the knowledge of various genres, the ability to use formal and informal language styles, and the knowledge of politeness norms and cultural conventions (Utdanningsdirektoratet 2010).Johansen (2008) and Brubaek (2012) have examined whether English language learners in Norway are attaining these goals, focusing on expressions of gratitude and requests, respectively.To date, however, no study has examined the levels of pragmatic competence in English language teachers.Considering the fact that language teachers are expected to act as language models in the classroom, but also that until very recently, English teachers in Norway were not required to obtain any academic level course credits in English, an investigation of Norwegian English teachers' pragmatic competence is warranted.The present study aims to address this gap by examining the requestive behavior of in-service English teachers in Norway.

Theoretical background
Among speech acts, requests constitute a particular challenge for language learners because they require intensive face work.When participating in linguistic interchanges, participants attend to each other's face, i.e., their public self-image (Brown & Levinson 1987) by selecting appropriate strategies.For example, they can seek to agree, manifest optimism, assert common ground and reciprocity, express approval or sympathy with the hearer or joke in order to appeal to the interlocutor's positive face, or a person's wish to be a part of a group and share involvement with others.Another sub-category of strategies appeals to the interlocutor's negative face, or detachment and a need for personal freedom.These strategies include giving options, apologizing, and stressing the importance of one's values.Because performing a request involves high stakes for both interlocutors, requests are classified as face-threatening acts.They are performed because the speaker needs a resource, such as time or material goods that the hearer owns.Performing a request is risky because the speaker infringes upon the hearer's freedom of action, but also because it may meet with a refusal.As such, requests require a heavy employment of face-saving strategies on the part of the speaker (Ellis 2003: 168).
The broad range of strategies that can be used to perform requests and their cross-linguistic variation only add to the difficulty of the task faced by a foreign language user.The main component of a request is the request proper, or the head act, which can be accompanied by alerters, supportive moves, and internal modifications (Blum-Kulka, House & Kasper 1989).Alerters precede the head act, functioning as attention-getters.They include titles, first and last names, nicknames, expressions of endearment, offensive terms, personal pronouns, and expressions such as 'Excuse me.' Supportive moves serve to mitigate or aggravate requests.They can be situated either before or after the head act.Getting pre-commitment, giving reasons and explanations, promising a reward, and minimizing or maximizing the imposition can serve as supportive moves.Internal modifications, which occur within the head act, include downgraders and upgraders.Downgraders are syntactic, lexical and phrasal devices that lessen the impact of the request, e.g., the use of interrogatives, subjunctives, conditionals, hedges, downtoners and lexical politeness markers such as 'please.'Upgraders, on the other hand, are devices that increase the impact of the request and include strategies such as imperatives or expletives.
The head act itself can be performed using various strategies.In the Cross-Cultural Speech Act Realization Project (CCSARP), Blum-Kulka, et al. (1989) distinguish direct strategies, conventionally indirect strategies and non-conventionally indirect strategies, as well as nine strategy types (Table 1).Direct strategies are statements of need or want, explicitly marked as requests.They are the easiest to interpret because they do not require the hearer to make inferences.Conventionally indirect strategies are linguistic means that are conventionalized as requests in a given language that refer to contextual conditions which are necessary for the performance of a request, e.g., 'would you' or 'could you' in English.Their interpretation depends on language conventions, conversational principles and contextual conventions established in the speech community.Non-conventional strategies are realized through hints about objects or elements needed for the performance of the request and have to be interpreted based on contextual clues.The dog is full of energy.
Mild hints (no reference to object or element that requires the act) It's gorgeous outside!As a part of CCSARP (Blum-Kulka et al. 1989), requests in five languages, including Australian English, were analyzed.The following four scenarios were used to elicit responses through a discourse completion task (DCT): 1) A student asks his roommate to clean up the kitchen the latter had left in a mess the night before.2) A student borrowed the professor's book, which she promised to return that day, but forgot to bring it.3) A student asks another student to lend her some lecture notes.4) An applicant calls for information on a job advertised in a paper.
The study found cross-cultural differences in the level of directness.While all languages under investigation employed all three levels of directness, situational variation in frequency of various strategies was noted.The researchers concluded that requestive behavior is affected by both cultural and situational-contextual factors.
Requests have received more attention than other speech acts in research on interlanguage pragmatics.Several studies have examined the production of requests and the use of strategies by native and non-native speakers.Most have used DCTs or role plays as the elicitation method.
Some studies suggest that non-native speakers tend to transfer requestive strategies from their L1 and may show a preference for direct strategies.Mills (1993) compared English and Russian requests produced by native speakers of English.The analysis revealed a transfer of conventionally indirect strategies from L1 Russian to English, resulting in requests that are highly inappropriate in English.A study by Schmidt (1994) compared the use of oral requests by native and non-native speakers of English.The results of this study indicate that while the most common request type among native speakers appears to be question directives (e.g., 'Do you know if…?' 'Is there…?'), non-native speakers did not show this preference.Instead, the most common request type they used was need statements (e.g., 'I want…,' 'I need…') and mitigated need statements (e.g., 'I would like to…').A similar conclusion was reached by Kankaanranta (2001), who found that her Swedish and Finnish subjects resorted to direct strategies such as imperatives and direct questions when communicating in English, and Kasanga (2006), who noticed a strong preference for direct strategies among native Afrikaans speakers learning English.Likewise, Economidou-Kogetsidis (2011) found that her Greek speaking subjects tended to select direct strategies, mainly imperatives mitigated with 'please,' want statements and interrogatives, and Dendenne (2014) found both positive and negative transfer in requestive strategies produced by Algerian learners of English.
A finding supported by a number of studies, commonly referred to as the 'waffle phenomenon' (Ellis 2003: 72) is that non-native speakers tend to produce longer requests than native speakers.Blum-Kulka (1982), who examined requests in Hebrew, concluded that the non-native speakers in the study displayed more supportive moves than the native speakers did.House and Kasper (1987), who compared native English, German and Danish requests with non-native German and Danish requests, found interesting differences in the overall length of utterances.Non-native speakers used more supportive moves such as pre-requests or reasons for requesting, which resulted in the increased length of their utterances.Blum-Kulka and Olshtain (1986), Faerch and Kasper (1989), Edmondson and House (1991) and Hassal (2001) obtained similar results.
Finally, findings from some studies suggest that non-native speakers are unable to differentiate their choice of strategies depending on the context and the interlocutor.Brunak and Scarcella (1979) found that their non-native speaker subjects who were native speakers of Arabic produced requests using a limited range of politeness features and showed a minimal ability to differentiate strategies based on the social context of the interaction.Tanaka (1988) concluded that Japanese learners of English were unable to differentiate levels of formality depending on the social distance from the interlocutor.A similar conclusion was also reached by Rose (2000), who investigated various speech acts in English produced by native speakers of Cantonese, and by Brubaek (2012), who examined English requests written by native speakers of Norwegian.Brubaek's study, specifically, used four different DCT scenarios to elicit requests of varying levels of formality.The data were analyzed using Brown and Levinson's model of FTAs (Brown & Levinson 1987).The study concluded that while the participants were able to rely on their L1 pragmatic knowledge to select appropriate requestive strategies in situations in which both interlocutors share the same social status, they failed to make native-like choices when the social status of the interlocutor was higher than their own.They overused well-known fixed expressions such as 'could I'-formulations, and did not use sufficient hedging.
Overall, research on requests suggests that native and non-native speakers differ in the choice of the strategies to perform the head act as well as in how they mitigate or reinforce their requests using internal and external modifications.Non-native speakers also tend to use more modifications and thus produce requestive utterances that are longer than those produced by native speakers, and display less situational variation in their choice of strategies.

The present study
To date, studies on non-native requests have examined a range of languages and subjects with varied first language backgrounds.So far, however, only one study has examined English requests produced by native speakers of Norwegian (Brubaek 2012).Brubaek's study, like most studies of speech acts, examined requests produced by second or foreign language learners rather than competent bilingual speakers.The present study aims to make a contribution to the body of research on requests by examining data from participants whose native language was Norwegian and who, as English language teachers, were not language learners, but advanced language users.
More specifically, the study aims to analyze requests elicited using four different scenarios with varied social contexts.The analysis focuses on both the head act and internal and external modifications.The research questions of the study are as follows: 1) What strategies are used by Norwegian users of English to formulate the head acts of requests?2) What internal and external modifications of requests are employed by Norwegian users of English (e.g., lexical or syntactic downgraders, upgraders, mitigating and aggravating supportive moves)?3) Is there variation in the employment of strategies in different social contexts, depending on the social distance between the interlocutors?
The results are discussed in the light of the findings presented in Blum-Kulka and House (1989), a study which was conducted within the CCSARP project and elicited native-speaker requests in English using the same DCT scenarios as the present study.

Methodology 4.1 Participants
The present study investigates the performance of requests by advanced users of English who were native speakers of Norwegian.All participants were elementary or middle school teachers of English who enrolled in an in-service teacher course in 2013/14 or 2014/15.Advanced proficiency in English was a prerequisite for the course.The course description specifies that English is the language of instruction, and that students must be sufficiently proficient in the language to participate in course activities such as lectures and group work, read academic texts, complete oral and written academic assignments, and communicate with instructors and classmates orally and in writing.Forty-one teachers, four males and 37 females from 39 different schools participated in the study.Forty participants responded to four scenarios, and one participant responded to three scenarios, giving a total of 163 elicited requests.In addition, control data from L1 Norwegian were collected to help explain possible cases of positive or negative transfer of requestive strategies from the first language.These data were obtained from 25 participants, native speakers of Norwegian, who were faculty or staff at Sør-Trøndelag University College.

Data collection
The data were collected using a written DCT that consisted of four request scenarios designed to elicit the speech act targeted by the study and accessible through an online survey link.To enable an analysis in the light of comparable native-speaker data, the scenarios used were the same as the request scenarios used in the CCSARP (Blum-Kulka et al. 1989).The participants were instructed to read each of the prompts and imagine they were the person described in it.The DCT items were presented as openended questions, i.e., there was no length limit to the responses.Four distractors were also used, and the scenarios were presented in a random order.All responses were anonymous and participants' IP addresses were not stored.The scenarios were as follows: -Situation 1 (Kitchen): You are a student.Ask your roommate to clean up the kitchen, which he left in a mess.-Situation 2 (Notes): You are a student.Ask another student to lend you her notes.-Situation 3 (Job): You want to apply for a job.Call the company and ask for information about a job they advertised in a paper.-Situation 4 (Presentation): You are a university teacher.Ask a student to give his presentation a week earlier than scheduled.
The power dimension between the speaker and the addressee, the social context, and the level of imposition varied among the scenarios.In the first two situations, the imagined interaction takes place between two equals.In the third scenario, the addressee is in the position of power and has a weak obligation to provide the requested information.In the last scenario, on the contrary, the requester is in the position of power and the addressee has a strong obligation to comply.This was to ensure a collection of both strong and weak imposition requests and a broad sample of a range of requestive strategies.
The data were analyzed using the framework from the CCSARP (Blum- Kulka et al. 1989).The framework was slightly simplified as it excluded the analysis of request perspective, which was included in the coding manual in the Blum-Kulka et al. study.The head acts were classified according to the level of directness (direct, conventionally direct and non-conventionally indirect) and the type of strategy used (see Table 1).External and internal modifications were classified using the following categories: -Downgraders -lexical (e.g., 'please'), phrasal (e.g., 'will you?'), and syntactic (e.g., conditionals, tenses) modifications that reduce the illocutionary force of the request -Upgraders -modifications that intensify the illocutionary force of the request (e.g., 'terrible,' 'bloody') -Alerters -titles, names, terms of endearment, personal pronouns -Supportive moves -external modifications that either mitigate or aggravate the request (e.g., 'Could you do me a favor?' or promise of a reward)

Head acts
The responses were coded using the coding manual from the CCSARP (Blum-Kulka et al. 1989).The distribution of requestive strategy types in the head act in all four scenarios was as follows.Conventionally indirect strategies were the most frequently used (69.94%),followed by direct strategies (28.22%).Very few non-conventionally indirect strategies were used (1.84%).
A closer look at the head acts indicates that overall the most preferred strategy type was query preparatory (67.48% of all requests).The typical query preparatory construction used was 'can/could + I/you.'The second most widely used strategy was hedged performatives, e.g., 'I would like to+performative verb' which was employed in 15.34% of requests.Both are examples of highly scripted, conventionalized requestive behavior.The remaining strategies were used sparingly (see Table 2).The examples below illustrate the use of query preparatory strategies (1) and hedged performatives (2).
[1] Scenario 1: Could you please clean up your mess?Scenario 2: Can I borrow your notes please?Scenario 3: Can you please send me some information?Scenario 4: Could you deliver your lecture a week earlier, please?
[2] Scenario 1: Will you please clean up the mess in the kitchen?Scenario 3: I would like some more information about the job at your company.Scenario 4: I need you to deliver your lecture a week earlier than scheduled.
Separate analyses of the strategies used in each of the scenarios indicate a certain degree of variation.The general trend seen above was reflected in the first scenario (Kitchen), with query preparatory strategies being the most preferred, followed by hedged performatives.Imperatives, obligation statements and suggestory formulae were used to a lesser degree, and want statements and hints were not employed at all (see Table 3).In scenario 2 (Notes), only two types of requests were used, namely query preparatory and suggestory formulae, with query preparatory being by far the most preferred (97.56%) (see Table 4).In comparison to scenarios 1 and 2, the head act strategies employed in scenarios 3 and 4 were quite different (Table 5).In scenario 3 (Job), four different strategy types were used, namely hedged performatives, want statements, query preparatory and hints, with hedged performatives and query preparatory being the two most preferred strategies.This is the only scenario in which hints were used, albeit not to a great extent.Nevertheless, this suggests that some of the participants displayed a lesser degree of perceived entitlement to make the request.The strategies used in scenario 4 (Presentation) were fairly diversified and more evenly distributed than in the other scenarios, although query preparatory requests were employed the most frequently (Table 6).Want statements were the second most preferred strategy in this scenario, followed by performatives and obligations statements.Hedged performatives were used the least frequently.

External and internal modifications
The following external and internal modifications were identified in the data: downgraders (including lexical, syntactic, and phrasal), upgraders, alerters and supportive moves (Table 7).Downgraders, found in 85.27% of the responses, were the most common modification.A lexical marker (e.g., 'please,' 'possibly') was the most preferred mitigator.The requests in all scenarios also contained syntactic modifications (e.g., verb modality).In fact, 36.8% of the requests were modified syntactically.In comparison, only 11.04% of the requests contained phrasal mitigators (e.g., 'will you' or 'it's ok').Very few requests overall were modified with upgraders (i.e., intensifiers of the head act).Importantly, however, upgraders were only found in scenario 1 (Kitchen) responses.The examples below (3) illustrate the use of upgraders (highlighted).
[3] Could you please clean up your mess?Can you please clear the kitchen?It looks like a bomb went off… Can you please clean up the kitchen?You can't leave it like this! Can you please clean up the mess in the kitchen?Remember that you do not live here alone and your mother does not work here.
Alerters and supportive moves were used extensively in all scenarios.The types of alerters in the data include first names, 'hello' 'excuse me' or 'I'm sorry.' Supportive moves include the reason for the request, an expression of gratitude, an acknowledgement of the imposition, an encouragement, and an apology.The examples below (4) illustrate supportive moves found in the data.
[4] Scenario 1: I will appreciate it.Scenario 2: I was sick yesterday.Scenario 3: I'm afraid I don't quite understand the job description.Scenario 4: I'm so sorry about this and I hope it won't be a big problem for you.
It is important to note that some of the requests contained more than one supportive move or downgrader.For example, some of the responses included an acknowledgement of the imposition (e.g., 'I understand if this leads to a problem for you') and encouragement (e.g., 'I am sure you will do your best').Thus, the numerical values in Table 7 reflect the total number of each type of modification made by the subjects in each scenario.However, the percentages listed pertain to the number of individual responses that contained each type of modification.
Differences were noted in the use of query preparatory strategies and supportive moves in the scenarios with no social distance between the interlocutors (scenarios 1 and 2) and the scenarios in which the social status of the interlocutors was different (scenarios 3 and 4).Namely, scenarios 1 and 2 contained more query preparatory strategies but fewer supportive moves than scenarios 3 and 4 (Table 8).The Chi-square test of independence was conducted to find out if there was a significant difference between the use of query preparatory strategies in the scenarios with no social distance between the interlocutors (scenarios 1 and 2) and the scenarios in which the requester and the addressee had a different social status (scenarios 3 and 4).The statistical results indicate that the participants used a significantly higher number of query preparatory strategies in the scenarios where the interlocutors had the same social status.These results were significant at a p<0.001 level (Table 9).The Chi-square test of independence was also conducted to examine if there was a significant difference between the use of supportive moves in the scenarios where the social status of the interlocutors was the same (scenarios 1 and 2) and those where it was different (scenarios 3 and 4).The results indicate a statistically significant difference at p<0.00 level (Table 9).

Discussion and conclusion
This study was undertaken to examine requestive strategies used by Norwegian teachers of English.The analyses of the head acts, and external and internal modifications show some interesting patterns and suggest the following tentative conclusions.Conventionally indirect strategies were the most preferred strategy type, which is similar to requestive behavior of native speakers of English.Blum-Kulka and House (1989), who used the same DCT scenarios, concluded that this strategy was used in 82% of their native speaker data.Similarly, Cenoz and Valencia (1996) found this strategy in 85% of the native requests and Eslami-Rasekh (1993) observed it in 75% of the requests produced by native speakers of English.In comparison with native speakers of English, the Norwegian subjects in the present study used more direct strategies and fewer non-conventionally indirect strategies.Previous findings suggest that native speakers of English use direct strategies in 10% of cases and non-conventionally indirect strategies in 4.8-8% of cases (Cenoz & Valencia 1996;Blum-Kulka & House 1989).Similar to the results of the Blum- Kulka and Kasper (1989), which provided a design model for the present study, the participants in the present study showed a preference for routinized requests.Even though the participants in this study used, in comparison, more direct strategies than native speakers, the ranking of strategy preference among Norwegian speakers in this study and native speakers of English is the same, with the majority of requests being conventionally indirect, followed by direct and with very few non-conventionally indirect strategies.This suggests that the Norwegian teachers' of English requestive behavior is near-native.
The distribution of the requestive strategies in the control Norwegian data displayed similar patterns.Only 10% of the Norwegian requests were phrased using direct strategies; 72% were phrased as conventionally indirect, and 8% were hints.The percentage of conventionally indirect requests in the English responses and in the control data was thus nearly identical, whereas a slightly higher percentage of hints suggests preference for non-conventional indirectness in native-Norwegian responses.Similar patterns can be noted when responses to individual scenarios are analyzed.In fact, in responses to scenario 2, all Norwegian requests were formulated using query preparatory strategies, as compared to 97.56% in the English data, whereas in scenario 4, performatives were used to a relatively high degree by both groups (17.5% in the English data and 40% in the Norwegian data).The most notable difference is observed concerning hints, which were used in all scenarios in the Norwegian responses, but were only employed in scenario 3 in the English data.In sum, the nearnativeness of head requests found in the English responses could have partially resulted from cross-linguistic interaction.
Overall, the participants in this study used a variety of head act strategies including imperatives, performatives, hedged performatives, want statements, suggestory formulae, query preparatory and hints.They also varied their use of strategies depending on the status difference between the interlocutors, e.g., they used significantly more query preparatory strategies in low social distance scenarios than in high social distance scenarios, which corresponds to native-speaker performance on the same tasks (Blum-Kulka & House 1989).The use of supportive moves was also significantly higher in the scenarios with a high social distance between the interlocutors.These findings suggest that non-native speakers of English can develop sensitivity to social context and vary their use of requestive strategies in English depending on the situation, which is contrary to the conclusions reached by Brubaek (2012), who examined the use of requestive strategies in English by Norwegian high school students.She reported an underuse of conventionally indirect strategies and request modifications aiming at minimizing the imposition, in particular in formal contexts.She concluded that her participants' pragmatic competence was underdeveloped and called for a need to devote more attention to pragmatic development in English education in Norway.The findings from the present study are optimistic as they suggest that English teachers in Norway can act as pragmatics role models in the classroom, which can be seen as the first step in meeting the learning outcomes for the attainment of pragmatic competence outlined in LK06 (Utdanningsdirektoratet 2010).
It is worth noticing the difference between the responses to scenario 1 and scenario 2, in both of which the social status of the interlocutors was absent.Upgraders, or those strategies that increase the level of imposition, were used in scenario 1, but not in scenario 2. In scenario1, the requester and the addressee have the same social status (students) and the addressee is strongly complied to oblige with the request.In scenario 2, however, the addressee, who is asked to share his/her class notes, does not have such an obligation.The systematic variation in the modification of requests in these two scenarios once again suggests that the subjects have attained contextual sensitivity in English.
Recall that supportive moves, e.g., giving reasons and explanations, promising a reward, and minimizing or maximizing the imposition, were found in the high social distance scenarios but not in low social distance scenarios.While this systematic difference in the use of supportive moves may suggest an attainment of native-like pragmatic competence, it has also caused a noticeable difference in the length of the large social difference requests in comparison with the low social difference requests.Because the requests in scenarios 3 and 4 contained extensive supportive moves, they were typically much longer (2-4 sentences) than the requests in scenarios 1 and 2 (1-2 sentences).Several of the responses contained more than one supportive move, for instance a combination of an acknowledgement of imposition and a justification of a request.Such extensive and complex use of supportive moves was not present in the Norwegian control data.Although similar to English responses, 48% of the Norwegian requests were modified by supportive moves, only 11% contained syntactic modifications, 28% contained lexical modifications, and very few contained more than one type of modification.Overuse of supportive moves has been reported in former studies on interlanguage requests and is referred to as the 'waffle phenomenon' (House 1989;Faerch & Kasper 1989;House & Kasper 1987).Production of extensively long requests can be perceived by native speakers as a violation of the Gricean maxim of manner which states 'Be brief' (Grice 1975: 79) and thus lead to the misunderstanding of the interlocutor's intentions (Blum-Kulka & Olshtain 1986), or, to use the term coined by Thomas (1983), cause 'pragmatic failure.' It is important to acknowledge that this study has some limitations, in particular the sample size and the limitations inherent in the use of DCT as a data collection method.Data used in this study were obtained from a relatively small group of participants who shared the same professional background.The data were collected using a DCT, and thus were only representations of what participants think they would produce under the described conditions.Whereas DCT enables collection of large amounts of data in a range of socially controlled situations, there is no guarantee that the responses provided by the participants correspond to their language use in natural communication settings.This is because the information provided in the scenarios does not provide the same amount of detail a natural setting does, and because responses tend to be idealized.Future research should be conducted using natural, ethnographic data obtained from larger and more heterogeneous groups of participants.
Notwithstanding its limitations, the findings of this study suggest that the development of near-native pragmatic competence is possible.Advanced non-native speakers of English can draw upon a wide range of requestive strategies and vary strategy selection depending on the social context.While in certain regards, the choice of strategies can be a result of L1 transfer and an overuse of supportive moves may be perceived as non-native, several similarities to native performance and sensitivity to the social context are also present.
The comparison of native and non-native speaker performance has characterized the majority of studies on interlanguage pragmatics.In general, as Cenoz and Gorter (2014) point out, "the communicative skills of multilingual speakers have traditionally been measured from a monolingual perspective against the yardstick of the ideal native speaker of each of the languages involved" (243).However, in contrast to monolingual language users, bilingual or multilingual speakers use the languages available to them in different contexts and for different purposes (Cenoz & Gorter 2014;Cook 2010;Grosjean 1985).Considering the complexity of the requestive behavior revealed in the analysis above, I would like to argue that the non-native users of English who were the participants in this study are bilingual "language users who really mean what they say" (Harlow 1990: 348).
Considerable research is needed to confirm these findings.In particular, research of speech acts produced by advanced multilingual speakers of a variety of L1 backgrounds could help us understand to what degree the attainment of sociopragmatic competence is possible.Rather than focusing on pragmatic failure (Thomas 1983), future studies should examine to what extent bi-and multilingual language users can draw on the linguistic resources available to them to successfully communicate in an additional language.

Table 1 :
Summary of different levels of directness used to realize request proper.Adapted from Blum-Kulka,House and Kasper (1989: 18)

Table 8 :
Percentage of responses containing query preparatory strategies and supportive moves in weak and strong imposition scenarios

Table 9 :
Chi square results: association between social distance and supportive moves and query strategies