The Interplay of Contextual Variables and Language Proficiency in Request Realization

The significance of producing socially appropriate utterances, not just linguistically correct ones, has given rise to a plethora of studies on speech acts. In this regard, the performance of second and foreign language learners against the norms of the target language and the factors affecting them have yet to be thoroughly investigated. To contribute to this ongoing research, this study aimed to elicit pragmatic requestive features of Iranian EFL (English as a Foreign Language) learners with respect to (a) request directness, (b) request perspective, (c) internal modification, and (d) external modification in eight social situations. This study also examined whether language proficiency and two contextual factors of social power and social distance affect Iranian EFL learners’ requests in English. Participants were a total of 60 university students or graduates who had taken TOEFL in the past 2 years and were thus categorized into low and high proficiency levels. A Written Discourse Completion Test (WDCT) was employed to elicit requests. To achieve the objectives of this study, both descriptive statistics and two-way ANOVA were employed. The results showed a significant main effect for contextual factors; participants’ choice of requestive features was affected by the two social factors of power and distance. Proficiency, however, did not have a significant effect on the pragmatic performance of participants. The findings of this study showed pragmatic development of participants on the one hand and some evidence of their underdeveloped pragmatic performance on the other hand, which makes explicit instruction a necessity. Plain Language Summary Contextual Variables and Language Proficiency in Requests Purpose: This study aimed to elicit pragmatic requestive features of Iranian EFL (English as a Foreign Language) learners with respect to (a) request directness, (b) request perspective, (c) internal modification, and (d) external modification in eight social situations. This study also examined whether language proficiency and two contextual factors of social power and social distance affect Iranian EFL learners’ requests in English. Methods: Both descriptive statistics and two-way ANOVA were employed in the present study. Conclusions: The findings of this study showed pragmatic development of participants in some aspects of request realization on the one hand and some evidence of their underdeveloped pragmatic performance on the other hand. Implications: Explicit instruction for the development of sociopragmatic competence of EFL learners is essential. Limitations: Because of the lack of access to native speakers, this study employed existing data on their performance in the literature.


Introduction
Being able to communicate effectively is suggested to be the main goal of learning a language.To be a competent speaker, apart from linguistic resources including grammar and vocabulary, one needs to be familiar with the sociocultural principles and norms of the target culture.Ignoring these norms would result in breakdowns and misunderstandings and consequently failure in communication (Bachman & Palmer, 1996;House & Ka´da´r, 2021).Hence, language learners should be able to comprehend and use language correctly and appropriately in given contexts, and that is what pragmatics is concerned with: comprehension, interpretation, and production of communicative acts (utterances) in different social contexts (Huang, 2017).Research has shown that even advanced language learners encounter difficulties when it comes to understanding and producing functions of language in a given context and their performance ''lags far behind native speakers in several respects'' (Bella, 2012(Bella, , p. 1917)).This suggests that proficiency in the target language does not ensure concurrent pragmatic performance (Youn, 2014).This problem has called for an array of studies on speech acts as well as factors that foster or affect the pragmatic performance of learners.Request is the most frequently investigated speech since it is ubiquitous in various settings (Rasheeda, 2020) and it is a ''Face-Threatening Act'' (FTA) (Brown & Levinson, 1987;House & Ka´da´r, 2021).
This study is also among the studies that examine requests not only pragmalinguistically (linguistic choices and forms) but sociopragmatically by examining the effect of contextual variables of social power and distance between the requester and requestee on the choice of request strategies.Furthermore, whether EFL learners' pragmatic competence develops along with their language level has also drawn researchers' attention, though the results have been inconclusive.Hence, the present study intends to identify request patterns of Iranian EFL learners.Identifying patterns, perspectives, and modifiers EFL learners use while requesting in different contexts contributes to the existing knowledge of interlanguage pragmatics, which partly concerns the type of language produced by second and foreign language learners.This study is also a further step toward understanding interlanguage pragmatics by investigating the interplay of contextual factors (social power and social distance) and language proficiency in request realizations and examining whether they foster EFL learners' pragmatic performance.

Request Speech Act
Speech act is one of main concepts in the vast field of pragmatics and has its origin in the work of Austin (1962).Speech acts are everyday activities such as informing, inviting, requesting, threatening, apologizing, and complementing for which we use our language (Austin, 1962).Request, as a fundamental and common human interaction activity (Walker, 2013), is categorized as a face-threatening act in which a speaker imposes force on the hearer (Brown & Levinson, 1987).In their foundational work, Brown and Levinson (1987) proposed their politeness hypothesis which became known as Politeness Theory.They claim that one way to fulfill politeness is by being indirect; ''People tend to choose indirect forms over direct ones to show politeness, since being direct is facethreatening.''Moreover, Brown and Levinson (1987) identified three sociocultural variables, namely, (1) social distance, (2) relative power, and (3) degree of imposition based on which speakers choose different strategies to mitigate or aggravate requests.Social distance is the degree of familiarity and closeness between interlocutors.The second variable concerns whether the hearer has power over the speaker.Finally, the degree of imposition refers to how big the request is.
Request Directness Brown and Levinson (1987) conducted a cross-cultural study comparing requests in English, Tzeltal, and Tamil.They concluded that there is a universal link between indirectness and politeness, that is the more indirect an utterance is, the more polite it is.Leech (1983) also believes that since requests limit the hearer's freedom, using indirect requests increases the degree of politeness.A number of studies conducted globally confirm the above finding.In a pioneering study, an international team of 10 researchers (Blum-Kulka & Olshtain, 1984) conducted ''A Cross-Cultural Study of Speech Act Realization Patterns (CCSARP).''Focusing on request and apology speech acts, their goal was to elicit patterns of speech act realization across languages and to compare and contrast the resulting patterns between native and non-native speakers of English.This large-scale study yielded various findings including the following: three broad categories were identified for requests which reflect three levels of directness, namely, direct requests, conventionally indirect requests, and hints.It was also found that non-native speakers used longer requests than native speakers.They also used more direct strategies compared to native speakers.
In a later study, Blum-Kulka (1987) re-examined the link between indirectness and politeness and argued that ''indirectness does not necessarily imply politeness'' (p.131).Blum-Kulka (1987) reported conventionally indirect requests as the most polite, explaining that they are non-coercive unlike direct requests and they have sufficient pragmatic clarity unlike hints.This finding that conventionally indirect request is the most frequently used strategy among native speakers of English has been repeatedly reported by numerous studies, in particular interlanguage studies in which the performance of second/foreign language learners is compared against that of the native speakers of English (Chen et al., 2013;Economidou-Kogetsidis, 2010;Fe´lix-Brasdefer, 2005, 2007;Flores Salgado, 2011;Hashemian, 2014;Jalilifar, 2009;Ogiermann & Bella, 2020;Ronan, 2022 to name but a few).

Internal and External Modifications
To mitigate FTAs, the use of internal and external modifications also known as mitigation devices has been suggested.Internal modifiers are defined as ''elements within the request utterance proper, the presence of which is not essential for the utterance to be potentially understood as a request'' (Blum-Kulka et al., 1989, p. 60).Internal modifiers are of two types; they might soften the request which are called downgraders (e.g. using linguistic elements such as ''please,'' ''possibly,'' and ''a bit''), or intensify the force and urgency of a request which are called upgraders (e.g. using really in the following utterance; ''I'm sorry to ask you for this but I really need an extension on this assignment'').Also, Faerch and Kasper (1989) distinguished between two types of downgraders; lexical/phrasal downgraders and syntactic downgraders.The former consists of a number of mitigating devices, such as politeness markers and hedges.Syntactic downgraders is comprised of interrogative, conditional structures, negation, tense, and aspect markings.
In addition to or instead of internal modifiers, speakers might benefit from some external modifiers which are placed in the immediate context and either follow or precede the request.External modification is used to support the request and to prepare the ground for the actual request.Studies on ILP requests have shown that while EFL learners use both internal and external modifications across all proficiency levels, they tend to underuse internal modifiers and overuse external modifiers in comparison with native speakers (Al Masaeed, 2017;Economidou-Kogetsidis, 2010;Faerch & Kasper, 1989;Hassall, 2001Hassall, , 2012;;Najafabadi & Paramasivam, 2012).

Request Perspective
Request perspective has not received as much attention as request directness and modification in the literature.Requests can be speaker-oriented (Can I), heareroriented (Can you), inclusive (Can we), or impersonal (Can one).Since in a request, it is the hearer who is ''under threat''; one way to soften the effect of imposition is to avoid identifying the addressee as the main performer of the act and thereby avoiding hearer perspective (Blum-Kulka & Olshtain, 1984).The existing literature also shows that request in English is characterized by a preference for speaker perspective (Lin, 2009;Ogiermann & Bella, 2020).
In an extensive research project, Ogiermann and Bella (2020) studied request perspective, examining requests produced by 900 participants from five native groups and four groups of English language learners at advanced level.To elicit data, a written discourse completion test was employed in which participants made a request to borrow lecture notes from a fellow student.The results revealed that Polish, Greek, and Russian speakers preferred hearer perspective, while English and German speakers opted for speaker perspective.Moreover, advanced learners of English with different L1 reduced their use of hearer perspective and thus drew more on speaker perspective, which is the norm of the target language, though it was not statistically significant.
Since there is a paucity of data on the requestive perspective of EFL learners, this study aims to fill the void by analyzing this aspect of requests produced by Iranian TOEFL holders, along with in/direct strategies and modifications.The results are then compared against the norms of English language in the literature.

Language Proficiency and Request Speech Act
While some studies on request speech act controlled language proficiency variable by examining the performance of learners at the same proficiency level (Al-Gahtani & Roever, 2013;Daskalovska et al., 2016;Hashemian & Farhang-Ju, 2017;Tan & Farashaiyan, 2015), some other examined whether pragmatic competence of language learners improves along with their general level of proficiency.Theoretical models consider pragmatic development as part of language proficiency (Bachman & Palmer, 1996), and thus it is expected that learners show sociopragmatic development with the increase in their language proficiency.Studies conducted on the relationship between pragmatic competence and L2 proficiency, however, reported mixed results.
Some studies concluded that language proficiency has a relationship with or affects the pragmatic performance of language learners (Abdolrezapour, 2015;Bu, 2012;Flores Salgado, 2011;Haji Maibodi, 2019;Jalilifar, 2009;Kasper & Roever, 2005;Khorshidi et al., 2016;Najafabadi & Paramasivam, 2012;Roever, 2006).In his study, Bu (2012) investigated request strategies employed by Chinese learners of English across two proficiency levels and compared the elicited data with the performance of native speakers of English.The results revealed that Chinese EFL learners performed differently at low and high proficiency levels; advanced learners used fewer direct strategies and instead used more conventionally indirect strategies, indicating ''development tendency toward native English speakers' ' (p. 35).
In yet another study, Flores Salgado (2011) investigated the pragmatic development of English learners of Mexican Spanish at three levels of proficiency (low, intermediate, and advanced).Using an oral production task, Flores Salgado (2011) compared request and apology patterns of Mexican Spanish learners of English with those of the English native speakers.The Results revealed that language proficiency was a determining factor in the performance of language learners as learners at higher levels more approximated native speakers in requesting and apologizing.In a recent study, Haji Maibodi (2019) also found that language proficiency affected EFL learners' performance.EFL learners were aware of the contextual variables as well and considered them while making requests.
Contrary to the above, some studies reported that L2 proficiency does not necessarily result in concurrent pragmatic performance (Allami & Boustani, 2017;Matsumura, 2003;Takahashi, 2005;Youn, 2014).As part of their study, Allami and Boustani (2017) examined politeness strategies in requests by Iranian EFL learners at two proficiency levels.The results showed deficiencies in the performance of EFL learners in comparison with that of the native group.The strategies used deviated from those of the native group.Moreover, EFL learners did not take contextual factors into account while producing speech acts.Neither did language proficiency have any significant effect on the choice of politeness strategies by Iranian EFL learners.This indicated that despite their good command of English, EFL learners were unaware of sociocultural strategies of the target language.
Another study that reports a minimal effect for proficiency is Takahashi's (2005) work on individual differences (language proficiency and motivation) and their correlation with pragmalinguistic awareness of Japanese EFL learners.According to the results, pragmalinguistic awareness and proficiency showed a low correlation.However, motivation subscales, were significantly more correlated with L2 pragmalinguistic awareness of Japanese EFL learners.
These findings lend credence to the notion that L2 proficiency is not necessarily a main factor in predicting learners' L2 pragmatic performance.Therefore, on the one hand, some studies confirm the positive effect of proficiency on pragmatic performance and on the other hand, a rather weak relationship was reported by some other studies.It can be concluded that whether language proficiency affects pragmatic performance remains inconclusive.Such mixed results call for further research to shed light on pragmatic development and the proficiency level of language learners.Hence, as part of this study, the effect of language proficiency of L2 learners on their pragmatic performance was examined.
The research questions of this study are as follows: RQ

Participants
Sixty male and female TOEFL holders participated in this study.Aged 19 to 36, participants were either university students or graduates across different levels and fields of education, who studied English mostly for the purpose of higher education abroad.To be precise, out of 60 participants, 10 were BA/BSc holders; 37 participants were at master's level, 9 of whom were students, and 28 graduated; finally, 12 participants were Ph.D. students.With regard to the field of study, 43 participants studied engineering majors; 11 participants studied humanities and the remaining 6 majored in medical sciences.
TOEFL IBT test is one of the most reliable and widely accepted tests of English language proficiency, which measures four main language skills through real-world tasks.Drawing on integrated tasks, the test evaluates the extent of test-takers' ability to effectively communicate in academic settings (Cho & Bridgeman, 2012).Hence, using the TOEFL IBT score is in line with the context of request situations in this study, which are all in institutional discourse.Based on the mean score, participants were grouped into low and high proficiency levels.

Instrument
This study employed Written Discourse Completion Test (WDCT), which has been the most common method of collecting data for pragmatics-based research (Mackey & Gass, 2015).In WDCT, a description of a situation is given to participants.Participants are provided with enough information about the occurrence of a speech act in the situation.A blank space usually follows the description where a response is required (Mackey & Gass, 2015).Research confirms the validity of WDCT as a method for eliciting data (Economidou-Kogetsidis, 2010).To serve the purpose of this study, 20 requestive situations, taken from previous studies, with a focus on institutional discourse, were given to 20 Iranian students and professors.They were asked to rank the situations based on the possibility of encountering them in an academic atmosphere.The ranking was designed on a scale from one (least likely to encounter) to three (most likely to encounter).
The situations were designed based on two sociopragmatic variables, namely, relative power and social distance between interlocutors.According to the ranking, 8 situations were selected, each varying in social factors.Social dominance has two variables: equal status (=P) and high status/hearer dominant ( + P). (There is yet another possible value for social dominance which is low status/speaker dominant (2P).However, all four situations for this variable were among the least probable situations in the eyes of students and professors; therefore, they were excluded.).Similarly, social distance is associated with two variables: negative/familiar interlocutors (2D) and positive/strangers ( + D).The description of situations and their source are presented in Table 1.
The coding scheme developed by Blum-Kulka et al. (1989), used in CCSARP (Cross-Cultural Study of Speech Act Realization Patterns), was the basis for analyzing data elicited from WDCT.In this scheme, requests are categorized based on three levels of directness, modifiers and perspective.As for modification, requests were analyzed in terms of downgraders as internal modification and supportive moves as external modification.It should be mentioned that all the existing taxonomies for coding internal and external modifiers were examined and accordingly an almost comprehensive list was compiled.The basis for all the taxonomies, however, was the CCSARP conducted by Blum-Kulka et al. (1989).The exact definition and examples of the above aspects and classifications are presented in, Table S1 to S5 (see Supplemental Material).

Data Collection and Analysis
It took participants between 30 to 40 minutes to write down their requests in response to eight scenarios designed; there was no time limit.They were instructed to write down in English what they would naturally say in each scenario.
To address research question 1, descriptive statistics were used to identify request patterns and their frequency among participants.Research question 2 was addressed  2015) A student requests assistance from a student s/he does not know to find the dormitory.Byon (2004) using mixed between-within subjects ANOVA to determine whether there were statistically significant differences in requestive behavior under different contextual factors as the within-subjects variable and across two proficiency level groups as the between-subjects variable.

Results
This section presents the study's results for research question 1 and research question 2.

Request Patterns (RQ 1)
Requests were analyzed for directness, request perspective, internal modification, and external modification.There were a total of 480 requests written by 60 participants.

Level of Directness
Requests can be expressed either directly or indirectly; indirect requests are categorized into conventionally indirect and non-conventionally indirect levels.Figure 1 shows the results for this grouping.
As can clearly be seen in Figure 1, participants preferred conventionally indirect strategy in 80% of requests (386 requests).The second most used strategy was direct level; it was used in 18.7% of requests (90).Finally, the least favorite strategy was non-conventionally indirect which was used only in four requests.An example of a request using conventionally indirect form and an example of a direct request are given below (It should be mentioned that the exact sentences written by participants are presented in this section, with no corrections if there were any errors.).
Example 1. ''Unfortunately, I haven't been able to finish the assignment yet.I was wondering if I could hand it in next week.''Example 2. ''Hello Professor.The assignment needed a lot of work and I tried to do it the best possible way.But it's not ready yet.Please give me some more time to do it.''

Request Perspective
A request can be made with the focus being on the hearer (you), the speaker (I), or it can be joint perspective/inclusive (we).It can also be impersonal, that is, without mentioning any specific agent.The results for the 480 requests of this study are presented in Figure 2.
As shown in Figure 2, 77% of requests (369 out of 480 requests) were hearer-oriented, which indicates that participants placed emphasis on the hearer rather than on themselves.Since a request is a face-threatening act inherently, this choice aggravates the situation than softens it.Therefore, participants did not mitigate the effect of imposition in the majority of requests in this regard.Speaker perspective was used in 21% of requests (102).A closer examination of each situation showed that speaker perspective was used the most in situation 2 (58%), which seems to be due to the context of its scenario that is asking a friend for notes of a class to keep up with the class.The use of the other two perspective forms was not noticeable, with only seven requests made with inclusive perspective and two requests with impersonal perspective.One example of a request in situation 2 is as follows: Example 3. ''Since last session I was not present in the class may I borrow your notebook in order to catch up with the material taught in that session?''

Internal Modification
The elicited data were also analyzed for internal modification consisting of syntactic downgraders and lexical/ phrasal downgraders.Internal modifiers can be used to mitigate the force of a request and to get the hearer to comply with it.They modify a request internally, that is, inside the request.A point worth mentioning is that since internal and external modifiers are optional elements, one or more of them can be used simultaneously.Therefore, some requests were classified several times; hence, the total was more than 480, which is the total number of requests.The results for this category are illustrated in Figures 3 and 4. According to Figure 3, 62% of participants (299 requests) preferred interrogative form of request.The second common form was conditional structure (99 requests).However, other downgraders were used infrequently.The examination of the distribution of interrogative form across situations showed that in situation 8, there was a 100% use of question form of request.It is an equal-status and unfamiliar situation in which a student asks a strange person for the location of the dorm.Situations 4 (86%) and 2 (78%) were two other situations for which participants employed a high number of interrogative forms compared to the rest of situations.Moreover, the use of conditional structure was used the Dashti Khavidaki most in situations 1 (45%), 3 (43%), and 5 (28%) respectively, which suggests that participants tended to use this downgrader in hearer-dominant situations (S1: + P, 2D; S3: + P, + D; S5: + P, 2D).Two examples are presented below; one in question form and the other using conditional structure.
Example 4. ''Hey man!Do you know how can I get to the dorm?''Example 5. ''According to a recent sever health condition, I was not able to complete my essay, so I would be so grateful if you could accept my apology and give me one more week to do it.I apologize for any inconvenience caused by this situation.'' Turning to lexical/phrasal downgraders (Figure 4), the first most used modifier in the pool of requests is the marker ''please'' (almost 19%, 90 requests).Consultative devices (13%), understaters/hedgers (11%), and appreciative embedding (11%) are the other most used lexical/ phrasal downgraders, respectively.A closer look into requests showed that the marker ''please'' occurred the most in situations 4 (40%) and 7 (28%).Situation 4 is an equal-status and unfamiliar situation while situation 7 is a hearer-dominant and unfamiliar one, in which direct request was significantly used.So, it seems that when participants used a direct request, they also tried to mitigate the imposition with the marker ''please.''Another point standing out in the figure is the high percentage of zero marking (almost 39%) meaning in 187 requests no lexical/phrasal downgraders were used.In fact, participants failed to modify their requests internally in about 40% of cases and that is significantly more visible in situations 8 (71%) and 2 (63%).Both these situations are equal-status; it seems that participants did not feel the need to modify their request internally where they have the same status as the requestee.Some examples for understaters/ hedgers and appreciative embedding are given below.Example 6. ''Recently I have been invited in a family event, so I could'nt finish my assignment on time.Would you give me some more time to finish it?''Example 7. ''Hi dear, I want to go out for a while, could you watch my books for a few minutes until I come back?''Example 8. ''Hello professor, I'm here to talk to you about the assignment that I supposed to do.I'm so sorry, since I need more time to be done with that I would really appreciate if you extend the dead line for couple of days.''

External Modification
Supportive moves lower the degree of force or imposition involved in a request externally.Figure 5 illustrates the result for this category.As can be seen, the most commonly used supportive move was grounder (almost 84%, 386 requests), which was extensively employed in all situations.In situation 5, there was a 100% use of grounder.Situation 5 has a hearer-dominant and familiar scenario.This modifier was also employed in situations 3 ( + P, + D; 90%) and 1 ( + P, 2D; 86%) more than in other situations.Considering these three situations, in which grounder was used, it can be reported that participants tended to use grounder in hearer-dominant situations more than in other situations.
The second most popular supportive move was sweetener (17%, 81 requests), which was used the most in situations 7 (73%) and 3 (30%) both of which are the same in terms of contextual factors, that is, both are hearer-dominant and unfamiliar situations.To add to the previous results, preparator (16%) and promise (15%) were found to be the third and fourth common supportive moves, respectively.The use of other supportive moves, however, was infrequent and not significant.The results also revealed that in 10% of requests, no external modification was used; zero marking occurred the most in situation 8, which is an equal status and unfamiliar situation.It is probably due to its context in which a student asks a stranger for direction, which is a request for information and is therefore different from others (request for actions).Furthermore, in 32% of requests, participants used a combination of moves.The following first two examples show how participants used grounder as a justification, explanation, or reason for their request, followed by examples for sweetener, preparator, and promise respectively.
Example 9. ''Excuse me, professor, I am really sorry but I have to say that next week's exam has coincided with my brother's wedding.Is it possible I take the exam another time?I really need to be at the wedding!''Example 10. ''Excuse me, professor, I intend to work on a topic but I need some help to develop it.I would appreciate it if we could have a meeting discussing the topic.''Example 11. ''Hello sir.Thank you; it was the greatest lecture I've ever attend on this issue.Could I please have your notes and PowerPoint presentations for further reference?''Example 12. ''Are you ready for the next week exam?Actually, I need a favor.I feel I have many weaknesses in this course, it makes me anxious!May I ask you the favor to help me to prepare for this exam?''Example 13. ''I could not make it to the class, how was it?I am afraid I missed important stuff, don't want to be surprised at the exam!Can I borrow your notes so I could catch up?I promise to return it intact tomorrow.''

Contextual Variables and Language Proficiency in Request Realization (RQ 2)
Before presenting the results to address research question 2, descriptive statistics for participants' TOEFL IBT scores and the two proficiency level groups are presented in Table 2.As the table shows, the minimum and maximum scores were 52 and 114 respectively, with the mean being 92.Based on the mean score, participants were divided into two groups; low (those with scores below the mean) and high (those with scores above the mean).Twenty-seven participants fell into the low group and 33 participants in the high proficiency level group.
To address research question 2, a mixed betweenwithin subjects ANOVA was conducted to, on the one hand, examine whether there were statistically significant differences in request realization across low and high proficiency levels and on the other hand, examine whether contextual variables (power and distance) affected participants' requestive behavior.Therefore, between-subjects independent variable is the two groups of participants (low and high) and within-subjects independent variable is contextual factors, the two variables of power ( + , =) and distance ( + , 2) each having two values, which makes a total of four conditions.A mixed betweenwithin subjects ANOVA could also determine whether the interaction between the independent variables makes a difference in pragmatic performance.It needs to be mentioned that in order to meet the assumption of parametric tests, categorical scores were transformed into percentages to obtain interval scores.Using Shapiro-Wilk, the normality of the data was also checked.The results for mixed between-within subjects ANOVA are presented for all the aspects of requestive behavior introduced while addressing research question 1.

Level of Directness
Firstly, frequencies and percentages for request directness across two levels of proficiency are presented in Table 3.As the percentages in Table 3 show, there seems to be little difference between the two groups in their choice of directness.Mixed between-within ANOVA was run for direct level and conventionally indirect level separately.However, non-conventionally indirect level was not a candidate for this test due to insufficient data.
Direct Strategy.The results of mixed between-within ANOVA revealed that contextual factors had a significant main effect on the choice of direct strategy, F (1, 58) = 4.86, p = .01,partial eta squared = .146.Partial Eta squared is a measure for effect size which in this case shows that contextual factor had a small effect, that is, about 15%.The significant effect of contextual factors on participants' choice of direct form is apparent by comparing the mean scores of both groups across the four examined situations, namely hearer-dominant, equal-status, familiar, and unfamiliar situations (see Appendix A, Table A1-A6 for descriptive statistics).Participants preferred direct requests more in familiar situations (mean = 23.33)than in unfamiliar situations (mean = 12.50) and that was statistically significant (p = .003).They also tended to employ direct form more in hearer-dominant situations (mean = 19.58)than in equal-status situations (mean = 16.25).
The means also showed a small difference between low and high groups in the choice of direct strategy in hearer-dominant situations and unfamiliar situations as the low proficiency level group used more direct requests (mean = 23.15 and 16.67) compared to the high proficiency level group (mean = 16.67 and 9.09).Nonetheless, the results of mixed between-within ANOVA, which was performed to ensure these differences were large enough to be considered statistically significant, revealed that unlike contextual factors, there was not a significant main effect for proficiency, F (1, 58) = 0.33, p = .56,partial eta squared = .006.Similarly, no significant effect was found for the interaction between contextual factors and proficiency, F (1, 58) = 2.07, p = .135,partial eta squared = .068.
Conventionally Indirect Strategy.Conventionally indirect strategy was the most commonly used strategy employed in 80% of requests.Similar to direct strategy, the results of mixed between-within ANOVA revealed that contextual factors had a significant main effect on the choice of conventionally indirect strategy, F (1, 58) = 50.08,p = .000,partial eta squared = .637.
According to the mean scores, participants used conventionally indirect requests significantly more in equalstatus situations (mean = 82.92)than in hearerdominant situations (mean = 54.58).This significant difference was also observed for familiar situations (mean = 75.83)compared to unfamiliar situations (mean = 61.67).Therefore, conventionally indirect requests were used mainly in equal-status and familiar situations.Furthermore, as the large effect size (partial eta squared = .637)indicates, contextual factors account for about 64% of the variation in responses.
Despite a significant main effect for contextual factors, such effect was found for neither proficiency, F (1, 58) = 0.368, p = .54,partial eta squared = .006,nor the interaction between contextual factors and proficiency, F (1, 58) = 1.76, p = .18,partial eta squared = .058.In summary, based on the findings, while contextual factors had a significant impact on the strategy chosen by participants, proficiency level did not have a statistically significant effect.

Request Perspective
Frequencies and percentages for the choice of request perspective across the two groups are presented in Table 4.
As mentioned earlier in the request pattern section, the most frequently used perspective was hearer-oriented (77%) and the second most common perspective was found to be speaker-oriented (21%).Participants did not use the other two perspectives as much.Similar to the strategy type, mixed between-within ANOVA was thus performed for the first two perspectives separately.
Hearer Perspective.Based on the mean scores of the two proficiency level groups across the four situations, hearer perspective was mainly used in equal-status situations (mean = 77.92)than in hearer-dominant situations (mean = 52.50)and it was found to be statistically significant (p = .000).The results of ANOVA confirmed a significant main effect for contextual factors in participant's choice of hearer perspective, F (1, 58) = 30.65,p = .000,partial eta squared = .518.A close look at the effect size (518) shows that contextual factors accounted for 50% of the variation of participants' choice in responses.Furthermore, the means also showed that hearer perspective was used equally in both familiar (mean = 65.42) and unfamiliar (mean = 65) situations as it was statistically confirmed by pairwise comparisons (p = .99)that there was no significant difference between the means of these two factors.
Speaker Perspective.Participants were also significantly influenced by contextual factors in their choice of speaker perspective as it was statistically observed, F (1, 58) = 23.22,p = .000,partial eta squared = .449.Based on the effect size, contextual variables accounted for approximately 45% of the variation in responses.A look at the means of both groups shows that participants tended to use this perspective significantly (p = .000)more in familiar situations (mean = 31.67)than in unfamiliar ones (mean = 9.58).Moreover, the use of speaker perspective was not significantly different (p = .54)in hearer-dominant situations (mean = 21.67) and equalstatus situations (mean = 19.58).
Further results showed no significant main effect for proficiency, F (1, 58) = .512,p = .47,partial eta squared = .009.Likewise, there was not a significant main effect for the interaction between contextual factors and proficiency, F (1, 58) = .865,p = .42,partial eta squared = .029.In brief, similar to the results for the directness aspect, participants were affected by contextual factors in their choice of perspective and proficiency did not have a significant main effect.

Modification
To analyze the data obtained for internal and external modification, the number of modifiers each participant used was added up and then converted to percentages to satisfy the assumption of ANOVA regarding interval scores.Therefore, internal and external modification was each analyzed as a whole and not in terms of different types of modifiers.Frequencies and percentages for internal and external modification across the two groups are presented in Table 5.
Both syntactic downgraders and lexical/phrasal downgraders are a sub-category of internal modification, so, they were calculated together.In both internal and external modifications, all modifiers were taken into account except for combination and zero marking categories.
Internal Modification.Internal modifiers reduce the imposition involved in a request using syntactic and lexical/phrasal downgraders inside the head act.A mixed between-within ANOVA found a significant main effect for contextual factors, F (1, 58) = 23.22,p = .000,partial eta squared = .449.The effect size value indicates the accountability of contextual factors for variation in responses is approximately 45%.
According to the mean scores of both low and high proficiency level groups, there was a difference in the use of internal modifiers between the mean score of heareroriented situations (mean = 21.67) and equal-status situations (mean = 19.58),indicating that participants were more likely to use internal modifiers in hearer-dominant situations than in equal-status situations.Moreover, participants tended to use internal modifiers significantly (p = .000)more in familiar situations (mean = 31.67)than in unfamiliar situations (mean = 9.58).
The analysis of mixed design revealed that proficiency, F (1, 58) = 0.512, p = .47,partial eta squared = .009did not significantly influence participants' use of internal modifiers.It was also the case for the interaction between contextual factor and proficiency, F (1, 58) = 0.865, p = .42,partial eta squared = .029.In brief, it can be reported that contrary to the significant effect of contextual factors on the participants' use of internal modifiers, proficiency did not play a significant role.
External Modification.External modifiers mitigate the force of a request through the use of one or some optional elements before or after the head act.To probe the significant differences among the participants and across contextual factors, mixed between-within ANOVA was conducted and the results revealed that there was a significant main effect for contextual factors, F (1, 58) = 14, p = .000,partial eta squared = .329.The effect size value indicates that despite reaching statistical significance, the accountability of contextual factors is rather small, that is, approximately 33%.
External modifiers were used significantly (p = .000)more in hearer-dominant situations (mean = 16.10)than in equal-status situations (mean = 12.84).It was also noted that participants employed external modifiers significantly (p = .007)more in familiar situations (mean = 15.38)than in unfamiliar situations (mean = 13.56).
Further results showed that similar to the previously analyzed aspects, no significant main effect was observed for proficiency, F (1, 58) = 0.046, p = .83,partial eta squared = .001,nor was it observed for the interaction between contextual factors and proficiency, F (1, 58) = 0.047, p = .95,partial eta squared = .002,meaning that participants, regardless of their proficiency level, took only contextual factors into account in their use of external modifiers.
In sum, the analysis of data revealed that participants considered contextual factors when using internal and external modifiers.However, their proficiency level was found to be a non-significant variable.

Summary of the Results
Beginning with request patterns, conventionally indirect strategy was the most popular strategy among Iranian EFL learners.A great number of requests were heareroriented, that is, the emphasis was on the hearer rather than the speaker.In terms of internal modifiers, interrogative form and conditional structure were used the most, respectively.Furthermore, the marker ''Please,'' Consultative devices, understaters/hedgers, and appreciative embedding were the most commonly used lexical/ phrasal downgraders.Finally, grounder was the most popular external modifier.The results further showed that proficiency did not have a significant main effect on any aspects of participants' pragmatic performance.Instead, participants were affected by contextual factors which, however, did not account the same for the variation in responses in the analyzed aspects.It went from 15% accountability in the choice of direct requests to almost 64% in the choice of conventionally indirect requests.

Discussion
This study intended to shed light on the pragmatic performance of Iranian learners of English on requests and to determine their requesting behavior regarding the level of directness, request perspective, internal modification and external modification.It also aimed to explore whether participants were affected by language proficiency and contextual variables when requesting in English.

Request Directness
Based on the results of this study, conventionally indirect strategy was the most preferred strategy, regardless of proficiency level.Conventionally indirect strategies are considered less direct and therefore, more polite as Brown and Levinson (1987) argue since these forms allow the hearer to decide whether to comply with the request or not.Therefore, this finding seems to be congruent with Brown & Levinson's politeness theory and its universality.
It also coincides with the results of a large number of cross-cultural and interlanguage studies suggesting that English native speakers use conventionally indirect strategy as a polite way of requesting (Blum-Kulka et al., 1989;Bu, 2012;Chen et al., 2013;Economidou-Kogetsidis, 2010;Ogiermann & Bella, 2020;Ronan, 2022).Hence, it can be concluded that, in this study, Iranians as foreign language learners of English were sociopragmatically aware of the norm of the target language regarding the level of directness in requests.Previous studies conducted on Iranian EFL learners also confirm participants' pragmatic competence in choosing the level of directness (Hashemian, 2014;Hashemian & Farhang-Ju, 2017;Jalilifar, 2009;Panahzadeh & Asadi, 2019); this is particularly of importance because Iranians as native speakers of Persian use direct strategies the most (Hashemian, 2014;Kalantari Khandani, 2017), while at the same time accompanying direct requests with several internal and external modifiers (Kalantari Khandani, 2017).Contrary to this study, however, Bu (2012) and Jalilifar (2009) reported pragmatic development with the increase in proficiency; the use of conventionally indirect requests grew with proficiency.
Unlike language proficiency, participants took contextual factors into account in their choice of request directness.While the research by Flores Salgado (2011) and Fe´lix-Brasdefer (2005, 2007) reported that native speakers of English used conventionally indirect requests in hearer-dominant situations ( + P) and unfamiliar situations ( + distance), both being pragmatically demanding and requiring mitigation, participants in this study used conventionally indirect requests significantly more in equal-status (=P) and familiar (2D) situations.This shows that participants followed different patterns than those of the native speakers in the literature and thus require explicit L2 pragmatic instruction on considering contextual factors when requesting in English.Instead, participants were inclined to use direct requests in hearerdominant situations; though deeper analysis showed that they tended to accompany direct requests with one or several internal and external modifiers.One explanation for the use of direct strategy might be the simplicity of its structure.However, since both low and high proficiency level groups used the direct form to a similar extent and mainly in hearer-dominant situations, it cannot be attributed to grammar knowledge and linguistic repertoire of learners but to their unawareness of conventional rules of a given context.

Request Perspective
According to the results, participants strongly favored hearer perspective, putting the emphasis on the hearer rather than on themselves in their requests, which is in contrast with the performance of native speakers of English (Blum-Kulka & Olshtain, 1984;Fe´lix-Brasdefer, 2007;Lin, 2009;Ogiermann & Bella, 2020;Woodfield & Economidou-Kogetsidis, 2010).To elaborate, a request is considered a face-threatening act per se (Brown & Levinson, 1987) in which a speaker threatens the hearer's face; ''any avoidance in naming the addressee as the principal performer of the act serves to soften the impact of the imposition'' (Blum-Kulka & Olshtain, 1984, p. 203).It seems that participants, even those at an advanced level, failed to soften the effect of imposition in most requests in this regard.It is thus assumed that participants of this study were not pragmatically aware of such a requirement to lower the imposition of requests.
The results also showed little evidence of the use of joint and impersonal perspectives.Similar results were reported by Woodfield andEconomidou-Kogetsidis (2010), Fe´lix-Brasdefer (2007) and Ogiermann and Bella (2020).Moreover, it was noted that hearer perspective was used significantly more in equal-status situations than in hearer-dominant situations.This finding is explainable given that having the same social status could result in more comfort and consequently more willingness to put the imposition on the interlocutor.Familiarity, nonetheless, was not taken into account in this regard and hearer-oriented requests were almost equally used in both familiar and unfamiliar situations.
It was also noted that there was not a statistically significant difference between low and high proficiency learners with regard to the choice of request perspective.This finding is in line with that of Fe´lix-Brasdefer (2007) in that advanced learners opted for hearer perspective, while native speakers of English preferred speaker perspective.In contrast to the finding of this study, Woodfield and Economidou-Kogetsidis (2010) found that advanced learners chose speaker perspective significantly more than other perspectives and thus approximating native speakers of English.

Internal Modification
Examining the elicited requests for internal modifiers showed interrogative form as the most preferred syntactic downgrader.Flores Salgado (2011) reported the same result with the question form being ''the most important syntactic mitigator'' (p.100).She further adds that the use of question form as the most preferred syntactic downgrader is due to employing a high number of conventionally indirect requests since these strategies are realized in question form.It appears that the finding of this study, too, could be explained in light of this fact.Also, conditional structure was found as the second most used syntactic mitigator.Regarding lexical/phrasal downgraders, the marker ''please'' was the most preferred strategy used significantly more in hearer-dominant situations.That was in line with Flores Salgado's (2011) study.
Furthermore, the results showed a significant main effect for social factors of power and distance in the use of internal modifiers; participants used them significantly more in hearer-dominant situations and also in familiar situations.Proficiency, nonetheless, did not have a significant impact on the choice of internal modification; instead, participants were affected by contextual factors.Unlike this finding, Otcu and Zeyrek (2008) and Go¨y et al. (2012) found that higher proficiency levels used internal downgraders significantly more than lower proficiency levels.Go¨y et al. (2012) further reported that upper intermediate levels used internal modifiers significantly fewer than native speakers.
A considerable proportion of zero marking was also noted among other findings.It was conspicuous, particularly in equal-status situations.Given that, it seems that participants of this study displayed an underuse of downgraders.This is in agreement with the findings of Economidou-Kogetsidis (2009) and Woodfield (2008).Using a narrow range of downgraders from both types was another finding of this study.EFL learners restricted their use to some of the modifiers and forewent the rest.It is argued that these findings might be due to ''a rather restricted pragmatic repertoire in learner's production'' (Economidou-Kogetsidis, 2012, p. 169).It seems that both low and high proficiency EFL learners in this study lacked sufficient pragmatic repertoire.

External Modification
Based on the results of this study, grounder which provides reasons and justification for a request was the most preferred modifier.This finding coincides with several previous studies (Blum-Kulka et al., 1989;Economidou-Kogetsidis, 2009;Fe´lix-Brasdefer, 2005;Flores Salgado, 2011;Hassall, 2001).In these studies, grounder was considered an efficient and practical mitigating device that triggers a positive attitude toward the speaker on the part of the hearer.
It was also found that participants' preference for the use of supportive moves was driven by contextual factors as they benefited from supportive moves significantly more in hearer-dominant situations and also, in familiar situations.That seems to concur with Flores Salgado's (2011) work in which supportive moves were mainly used in situations that power rests on the hearer.This finding is explainable given that hearer-dominant situations ''tend to be more pragmatically demanding and complex and they generally require greater pragmatic skill'' (Economidou-Kogetsidis, 2009, p. 81).More importantly, such situations need ''more mitigation and less directness.''Unlike Flores Salgado (2011), nonetheless, means of external modification were used more in familiar situations than in unfamiliar situations.It is argued that similar to hearer-dominant situations, unfamiliar situations require more mitigation and reduction of the force.
Another finding of this study was that unlike contextual factors, neither proficiency nor interaction between proficiency and contextual factors showed a significant effect on the participants' use of external modifiers.Contrary to this finding, Flores Salgado (2011) reported a significant main effect for proficiency, with native speakers and advanced learners using the most supportive moves for each request and basic language learners the fewest.Many other studies controlled the effect of proficiency by choosing one proficiency level, which in most cases has been the advanced level (Economidou-Kogetsidis, 2009;Woodfield & Economidou-Kogetsidis, 2010).
This study also found that participants used external modifiers more than internal modifiers.This tendency of learners to opt for external modifiers rather than internal ones is reported in the literature (Economidou-Kogetsidis, 2009, 2012;Hassall, 2001Hassall, , 2012)).To explain the preference of learners for external modification over internal modification, Hassall (2001) argues that external mitigators tend to be ''syntactically less demanding and pragmalinguistically less complex'' (p.274).Therefore, learners are ''more able to use external modifiers'' (p.273).

Conclusion
Based on the findings of this study, it can be reported that on the one hand, participants showed pragmatic awareness and competence by using conventionally indirect requests as the most preferred strategy.On the other hand, there was some evidence of their underdeveloped pragmatic performance and some ''non-native like'' request realizations including the choice of speaker perspective, underuse of internal modifiers, and the fact that their performance showed no sign of development with the increase in proficiency.To add to the above, even though participants took contextual factors of social power and social distance into account when choosing request strategies and modifiers, their performance deviated from those of the native speakers as reported in the literature.The most salient example is the use of direct strategy in hearer-dominant situations, while native speakers of English tend to choose conventionally indirect requests in such situations.This can be attributed to participants' sociopragmatic unawareness of the culture and norms of English as the target language.All this underscores the need for explicit instruction; when those who score high in TOEFL, which is widely accepted throughout the world as the criterion for the ability to effectively communicate, seem to be, at best, partly aware of the sociocultural requirements of the target language, explicit instruction seems a necessity.
Language instructors should empower students in their pragmatic performance by informing them about the socio-cultural conventions and norms of the target language.Learners need to be warned that it is not just enough to familiarize themselves with a variety of words and structures; it is also of prominent importance that they know where, when, and how to use their linguistic repertoire.Otherwise, it would result in confusion and misunderstanding in interaction, ultimately leading to failure in communication.Teachers can discuss the universal pragmatic elements, and also, the elements that tend to be specific to the target culture.A comparative study between the mother tongue and the target language would suit this aim.
It should be mentioned that while most interlanguage pragmatic studies compared speech act realizations of language learners with those of native speakers, this study drew on the existing literature and in fact, it is one limitation of the study.Also, in this study, data were analyzed with regard to the level of directness, perspective, internal and external modification.It is recommended that further research analyze request realizations in terms of syntactic complexity, including frequency counts and morphosyntactic analysis among others.The contribution of language proficiency can be investigated in this regard.Further research is needed to determine whether proficiency level affects the syntactic complexity of requests.
Moreover, in the present study, grammatical errors that participants made in their requests were not taken into account, though they have been noticed.It is of importance to determine the extent of the grammatical inaccuracy of requests and to what extent it is related to individual variations examined in the study.Thus, grammatical accuracy of participants' production needs to be considered and measured in future studies.Lastly, in this study, proficiency did not show any effect on learners' performance and contextual factors did not account for all the variation in responses.There might be other factors at work one of which might be the effect of L1 transfer, which calls for further research.

Table 1 .
Description of Eight DCT Situations.
S7 (Presentation) [ + P, + D]Student-professor, A student requests that a professor s/he does not know email him/her a presentation file.Eslami et al. (

Table 2 .
Descriptive Statistics for TOEFL IBT and Low and High Proficiency Level Groups (N = 60).

Table 3 .
Percentages and Frequencies for Request Directness Across Two Proficiency Levels (N = 60).

Table 4 .
Percentages and Frequencies for Request Perspectives Across Two Proficiency Levels (N = 60).

Table 5 .
Percentages and Frequencies for the Use of Modification Across Two Proficiency Levels (N = 60).

Table A2 .
Descriptive Statistics for Participants' Use of Conventionally Indirect Request Under Different Contextual Factors Across Two Proficiency Levels (N = 60).

Table A3 .
Descriptive Statistics for Participants' Use of Hearer Perspective Under Different Contextual Factors Across Two Proficiency Levels (N = 60).

Table A4 .
Descriptive Statistics for Participants' Use of Speaker Perspective Under Different Contextual Factors Across Two Proficiency Levels (N = 60).

Table A5 .
Descriptive Statistics for Participants' Use of Internal Modifiers Under Different Contextual Factors Across Two Proficiency Levels (N = 60).

Table A6 .
Descriptive Statistics for Participants' Use of External Modifiers Under Different Contextual Factors Across Two Proficiency Levels (N = 60).