Studies in Second Language Learning and Teaching

This paper synthesizes cross-sectional studies of the effect of proficiency on second language (L2) pragmatics to answer the synthesis question: Does proficiency affect adult learners’ pragmatic competence? Findings have revealed an overall positive proficiency effect on pragmatic competence, and in most cases higher proficiency learners have higher pragmatic competence. However, increased proficiency does not guarantee a native-like pragmatic performance because proficiency effect varies depending on the nature of target pragmatic features such as types of speech acts (degrees of directness and conventionality) (e.g., Cook & Liddicoat, 2002; Félix-Brasdefer, 2007), modalities of pragmatic performance (comprehension and production) (e.g., Bradovi-Harlig, 2008, 2009), social variables involved in task situations, such as social status (e.g., Allami & Naeimi, 2011), social distance (e.g., Maeshiba, Yoshinaga, Kasper, & Ross, 1996), and power relationship (e.g., Al-Gahtani & Roever, 2012). Moreover, proficiency effect is mediated by contextual variables such as length of stay in the target language community (e.g., Shardakova, 2005; Taguchi, 2011, 2013; Xu, Case, & Wang, 2009).


Introduction
Originated in Hymes (1972), the ability to use language appropriately in communication is regarded as important as knowledge of grammatical rules in all theoretical models of communicative competence (e.g., Bachman & Palmer, 1996, 2010;Canale & Swain, 1980).Pragmatic competence, namely the ability to understand and use linguistic forms appropriately according to different situations, is accepted as a vital component of language ability (e.g., Bachman & Palmer, 1996, 2010).Along with the emergence of theoretical models of communicative competence, pragmatic competence and development have captured a growing interest in L2 pragmatics research (for a review, see Kasper & Rose, 1999, 2002;Kasper & Roever, 2005).
One focus of these studies is to investigate factors affecting pragmatic competence, and general language proficiency is the single most widely examined factor in previous research on L2 pragmatics.The interest in proficiency effect partially stems from the assumption that general proficiency is a precondition of pragmatic competence.In other words, L2 pragmatic acquisition requires learners to achieve a threshold level of proficiency, suggesting a positive proficiency effect on L2 pragmatic competence (for a review, see Bardovi-Harlig, 1999, 2001, 2013;Kasper & Rose, 1999, 2002).This assumption has been supported by many empirical studies with a cross-sectional design across different proficiency levels or a comparison between L2 learners and native speakers (NSs) (e.g., Al-Gahtani& Roever, 2012;Bardovi-Harlig & Dörnyei, 1998;Dalmau & Gotor, 2007;Garcia, 2004;Geyer, 2007;Maeshilba, Yoshinaga, Kasper, & Ross, 1996).

Terminology
Since the present synthesis paper focuses on proficiency effect on L2 pragmatic competence, key terms, L2 proficiency and pragmatic competence, are defined in this section.

L2 proficiency
In the present study, L2 proficiency is defined as overall L2 competence, which includes organizational knowledge and pragmatic knowledge.The former refers to the knowledge of organizing utterances or sentences and texts such as lexical and grammatical knowledge, while the latter refers to the knowledge of using sentences and texts appropriately in situations (Bachman & Palmer, 1996, 2010).In previous studies, levels of L2 proficiency were determined by standardized test scores (e.g., Allami & Naeimi, 2011;Garcia, 2004;Nguyen, 2008;Taguchi, 2005Taguchi, , 2006;;Takahashi, 1996;Xu, Case, & Wang, 2009 ), course or grade levels (e.g., Bardovi-Harlig, 2008, 2009, 2011;Félix-Brasdefer, 2007;Koike, 1996;Pinto, 2005), length of formal instruction (e.g., Bonganets, Kellerman, & Bentlage, 1987), and duration of residence in the target language community (e.g., Blum-Kulka & Olshtain, 1986;Decapua & Dunham, 2007).Among these proficiency determinants, standardized test scores and course or grade levels are considered to be more reliable indicators of proficiency because standardized test scores allow for comparison across test takers, and course/grade levels are usually determined by a systematic evaluation within a course or program, including placement tests, teacher assessment and oral interviews.In contrast, length of formal instruction and residence in the target language community are less reliable ways of determining proficiency due to the large variability in amount and quality of L2 exposure and instruction that cannot be accounted for by such measures.Therefore, the present study synthesized studies in which levels of L2 proficiency were indicated by standardized test scores and course/grade levels.

Pragmatic competence
Pragmatic competence consists of pragmalinguistics and sociopragmatics (Leech, 1983;Thomas, 1983).According to Thomas (1983), pragmatic failure can be broken down into two types: pragmalingusitc and sociopragmatic failure.Pragmalinguistic failure is fundamentally a linguistic problem, "caused by differences in the linguistic encoding of pragmatic force" (Thomas, 1983, p. 99), whereas sociopragmatic failure results from "different perceptions of what constitutes appropriate linguistic behavior" (Thomas, 1983, p. 99).The distinction between these two types of failure parallels the dichotomy between pragmalinguistics (functional aspect of pragmatic competence) and sociopragmatics (social aspect of pragmatic competence), which can be found in the definition of pragmatic knowledge of Bachman and Palmer's (2010) model of communicative competence.According to Bachman and Palmer (2010), pragmatic knowledge includes functional and sociolinguistic knowledge.Functional knowledge represents the knowledge of using linguistic forms to realize pragmatic functions, such as using would you to make a request, while sociolinguistic knowledge is the knowledge of using linguistic forms appropriately according to different situational variables (e.g., social status, familiarity, power relationship, and degree of imposition), such as choosing polite forms when speaking to people of higher social status (for a review, see Kasper & Rose, 2002;Taguchi, 2015).
The present study adopts Bachman and Palmer's (2010) definition of pragmatic knowledge, which includes functional (pragmalinguistics) and sociolinguistic knowledge (sociopragmatics) and is reflected in the knowledge of different target pragmatic features such as speech acts, routines, implicatures, address forms and discourse markers.

Method
The following steps were taken to locate relevant studies in the online academic databases.First, key words were chosen and sorted into two groups.Group 1 comprised words related to pragmatic competence, including pragmatic competence, sociocultural competence, interlanguage pragmatics, sociolinguistic competence, sociopragmatic competence, interactional competence, speech acts, routines, and implicatures.Group 2 consisted of words related to general L2 proficiency, such as L2 proficiency, L2 competence, and L2 grammatical competence.Second, all word combinations, created by mixing each word in Group 1 with each word in Group 2, were searched for in multiple databases, including Education Resources Information Center (ERIC), Linguistics and Language Behavior Abstracts (LLBA), Journal Storage (JSTOR), Modern Language Association Bibliography (MLA), and Sage Journals.This search process yielded 421 studies, of which irrelevant studies were eliminated based on the criteria listed below.At the end, 28 studies (marked with an * in references) were included.The following inclusion and exclusion criteria were used: 1. L2 proficiency was one of the independent variables investigated.2. The study was a data-driven empirical study of L2 pragmatic competence.3. The study had a cross-sectional design, and learners' groups in the study were determined by different L2 proficiency levels.4. Participants' L2 proficiency levels were determined by standardized test scores or course/grade levels.5. Studies using a sample of children were excluded.6.Studies with a longitudinal design were excluded.7. Studies of instructional effects on L2 pragmatic competence were excluded.1Following Norris and Ortega (2006), the 28 studies selected were coded for two types of study features: substantive and methodological features.Therefore, my coding scheme to profile each study (see Appendix A) included the following features: author and publishing date, research question(s), target language, target pragmatic feature(s), modality (comprehension, production or awareness of target pragmatic features), instrument/data source, sample size, groups (L2 proficiency levels), proficiency measure(s), and findings (see Appendix B for the table of the study profiles).I also coded each study for the findings in order to answer my synthesis question.

Review of studies of proficiency effect on pragmatic competence
Findings of selected studies can be categorized into three groups: (a) positive effects; (b) almost no effect; and (c) mixed effects that vary depending on the nature of pragmatic features considered.
First, the positive role of L2 proficiency was documented in three studies of pragmatic transfer (Koike, 1996;Maeshilba, Yoshinaga, Kasper, & Ross, 1996;Wannaruk, 2008).For example, Maeshilba et al. (1996) investigated transfer of apology strategies from L1 Japanese to L2 English.Participants were 30 Japanese EFL learners with intermediate English proficiency (TOEFL scores from 400 to 500), 30 with advanced English proficiency (TOEFL scores from 510 to 627), 30 English NSs, and 30 Japanese NSs.The instrument was a dialogue construction questionnaire with items representing different social variables such as gender, social distance and relative social status, and degrees of severity of the committed offense.The following is an example (p.182): At a friend's home Ann and Bill are both 35 years old and are good friends.Ann borrowed a computer magazine from Bill.Unfortunately, Ann spilled coffee on the magazine and damaged it.She is now returning it to Bill.Bill: What happened to my magazine?Ann: Bill: All participants completed the dialogues and rated situations on a 5-point scale for five contextual factors: severity of offense, offender's obligation to apologize, likelihood for the apology to be accepted, offenders' face loss, offended party's face loss, and social distance and dominance.Findings showed that in situations where Japanese and American apology patterns were the same, learners had positive transfer from Japanese apologies, but there was no significant difference between advanced and intermediate learners.However, in the situations where Japanese and American apology patterns differed, advanced learners transferred their apology behavior from Japanese to English in only two instances, whereas intermediate learners did the same in six instances.In essence, more advanced learners had a better ability to minimize negative L1 transfer.Wannaruk (2008) also found a positive proficiency effect on negative L2 pragmatic transfer.She used a discourse completion task (DCT) to examine 40 Thai EFL learners' pragmatic transfer in refusals to invitations, suggestions, offers and requests.Participants were assigned to three groups: lower intermediate, intermediate and upper intermediate groups based on their scores on the university's Graduate English Test (no details about the test).In addition, 40 American and 40 Thai NSs participated.Data from each group were compared in terms of frequency of refusal strategies.Consistent with the findings of Maeshilba et al.'s (1996) study, lower proficiency learners had more negative L1 transfer in L2 refusals than their higher proficiency counterparts.These findings suggest that increased proficiency can improve L2 learners' ability to control negative L1 transfer.
Similarly to transfer of pragmatic performance, the positive proficiency effect on comprehension and identification of speech acts was reported in two studies (Cook & Liddicoat, 2002;Garcia, 2004).Cook and Liddicoat (2002) investigated L2 English learners' comprehension of three types of requests: direct, conventional indirect and nonconventional indirect.A direct request delivers the intention of the request in grammatical, lexical or semantic forms specialized for this pragmatic function such as imperatives (e.g., please) and want statements.A conventional indirect request expresses the intention of the request by using fixed linguistic conventions such as would/could you + verb phrase.A nonconventional indirect request (e.g., hinting), on the other hand, is not realized in fixed linguistic forms or conventional expressions, so the interpretation of this type of request requires the hearer to comprehend intention behind the utterance by using contextual information.A multiple-choice questionnaire consisting of five scenarios per request type was administered to 100 ESL learners in Austria and 50 Austrian English NSs.The ESL learners were divided into low and high proficiency groups according to their IELTS (International English Language Test System) or TOEFL scores.Findings revealed that NSs comprehended significantly more requests than high-proficiency group.High proficiency learners outperformed low proficiency learners on all three types of requests, but the difference was smaller for direct and conventional indirect requests than that for unconventional indirect requests.In other words, there was a positive proficiency effect on comprehension of requests, but the effect was mediated by degrees of directness and conventionality of speech acts.Garcia (2004) found a positive proficiency effect on identification of speech acts.She designed a speech act identification task that consisted of four types of nonconventional indirect speech acts: requests, suggestions, corrections and offers.In the task, participants chose the right speech act according to the dialogue.56 ESL learners formed high and low proficiency groups (proficiency determined by TOEFL scores).19 English NSs also participated.Consistent with Cook and Liddicoat's (2002) study, NSs and high proficiency learners had significantly higher identification scores of target speech acts than low proficiency learners.However, there was no significant difference between NSs and high proficiency learners, suggesting that high proficiency learners can become native-like in identification of speech acts.
The positive proficiency effect was also documented in five studies of pragmatic production, with three of them investigating production of speech acts (Dalmau & Gotor, 2007;Pinto, 2005;Trosborg, 1995), and the other two examining pragmatic production at discourse level (Al-Gahtani & Roever, 2012;Geyer, 2007).
For example, Pinto (2005) used a 4-item DCT to examined L2 Spanish learners' production of requests.He first identified similarities and differences in request strategies between English and Spanish NSs and then compared these strategies with those used by L2 learners.Participants included 20 American learners of Spanish whose general L2 proficiency levels were determined by their course levels from Group 1 to Group 4.He found that the learners moved towards target Spanish norms from lower-level (Group 1 & 2) to higher-level proficiency (Group 3 & 4) regarding the token frequency of request strategies.Lowerlevel learners were more pragmatically ambiguous than higher-level learners.For example, they were more likely to transfer the pseudo permission strategy in English (e.g., can I + verb phase) to L2 Spanish request.This negative transfer caused ambiguity because Spanish NSs did not know whether the function of this sentence pattern was to ask for permission or a favor.This ambiguity may be due to lower-level learners' limited knowledge of L2 pragmalinguistic forms.
The positive proficiency effect on speech act production was also evident in Dalmau and Gotor's (2007) study of L2 English apologies.They used an 8-item DCT to examine the production of apologies of 78 Catalan learners of English across three different proficiency levels (intermediate, advanced and proficient).Participants' proficiency levels were determined by their scores on the university's placement test that included oral and written parts (p.312; no details provided).In addition, 26 English NSs were recruited to make a comparison with English learners.They found that learners with higher proficiency had a greater range of apology strategies and were less likely to use nontarget-like apology expressions.Moreover, more proficient learners used more lexical intensifiers (e.g., very, really, so), but their overall token frequency of intensifiers was still significantly lower than that of NSs, and they still had problems producing accurate pragmalinguistic forms (e.g., I'm sorried; p. 307).
Aside from speech act production at the sentential level, two studies showed that proficiency affected L2 learners' pragmatic production at discourse level (Al-Gahtani & Roever, 2012;Geyer, 2007).Geyer (2007) used the Japanese Oral Proficiency Interview (OPI) to investigate L2 learners' production of selfqualification segments in Japanese.Self-qualification segments were defined as "parenthetical statements within a discursive unit" (p.339), which were used to mitigate previous statements.Japanese self-qualifications were frequently introduced with contrastive markers, such as demo, kedo, and ga, each of which is equal to but, though or although in English.Participants were divided into four levels: superior, advanced, intermediate and novice.Findings showed that lower proficiency learners were able to produce self-qualifications, but they often did not use contrastive markers.Intermediate and advanced learners employed contrastive expressions in self-qualifications, but not always appropriately.Compared to lower proficiency learners, higher proficiency learners had a larger repertoire of lexical and grammatical markers to qualify their utterances.Moreover, higher proficiency learners were able to use contrastive expressions in conjunction with other discourse features such as foregrounding main points of the utterance and putting the qualification segments in the background.In other words, learners' increased L2 discourse knowledge led to more native-like production of self-qualification segments.
Al-Gahtani and Roever (2012), on the other hand, used a role-play task to examine sequential organization of requests.Participants were 26 male Saudi learners of Australian English who were divided into beginning, lower-and upper-intermediate, and advanced groups (proficiency determined by course levels).Moreover, a cloze test was administered to upper-intermediate and advanced participants, and their self-reported IELTS scores were also collected to ensure that these two groups represented different proficiency levels.All participants completed three request-making role-play tasks, each of which required a long conversation that varied in the power relationship between interlocutors (equal, less and more powerful).For example, Situation 1 (p.49; equal interlocutor relationship) required participants to ask his housemate to buy some bread from a supermarket when the housemate was watching TV and wanted to stay home.Three role-play tasks were conducted individually and audio-taped.Findings showed that there was a positive proficiency effect on learners' sequential organization of requests.Compared to lower proficiency learners, higher proficiency learners produced more preexpansions (e.g., greetings, summons prior to request) and insert expansions (e.g., negation about timing and other details of the request).However, only four out of the eight higher proficiency learners included justifications or reasons in the request turn in the situation where they were in a higher status position, but lower proficiency learners had the same performance regardless of the power relationship.In other words, higher proficiency learners had better pragmatic performance at discourse level, but their sensitivity to the power relationship in target situations was not significantly better than that of lower proficiency learners.

Almost no proficiency effect on pragmatic competence
In contrast to the studies summarized above, four studies reported almost no proficiency effect2 on pragmatic competence (Allami & Naeimi, 2011;Niezgoda & Roever, 2001;Shardakova, 2005;Takahashi, 1996).For example, Niezgoda and Roever (2001) investigated awareness of grammatical and pragmatic errors in 48 ESL and 124 EFL learners.Participants were divided into high and low proficiency groups based on their scores on the placement and achievement tests.The instrument was a 20-item video-and-questionnaire task originally designed by Bardovi-Harlig and Dörnyei (1998).In the task, participants watched each scenario twice and read the written description of the scenario, then indicated the appropriateness/correctness of the last utterance (target speech act utterance) in each scenario by checking yes or no.The following is an example (p.71): The teacher asks Peter to help with the plans for the class trip.T: OK, so we'll go by bus.Who lives near the bus station?Peter, could you check the bus times for us on the way home tonight?P: #No, I can't tonight.Sorry.Was the last part appropriate/correct? Yes No If there was a problem, how bad do you think it was?Not bad at all___:___:___:___:___:___Very bad And, how would you revise it?
As displayed above, the target speech act utterance for each scenario was in bold, and there were two questions next to it: The first question asked about the appropriateness and correctness of the utterance, and the second asked about the perceived severity of the error, if any.If the answer to the first question was yes, the participants proceeded to the next scenario.Findings showed that low proficiency learners recognized significantly more pragmatic than grammatical errors, whereas high proficiency learners identified slightly more grammatical than pragmatic errors.In other words, high proficiency learners' awareness of pragmatic errors was lower than low proficiency learners'.Allami and Naeimi (2011), on the other hand, used a DCT to investigate Iranian EFL learners' production of refusals.In this study, 30 Iranian EFL learners were divided into three groups: lower-intermediate (TOEFL scores lower than 400), intermediate (TOEFL scores from 400 to 530) and upper-intermediate proficiency (TOEFL scores higher than 530).In addition, 31 Iranian NSs also participated.All participants completed a 12-item DCT representing four different refusals: refusals to requests, invitations, offers and suggestions.Within each type of refusals, three situations differed in social status (higher, equal or lower).Comparison between Iranian and American NSs showed that American patterns of refusals were less subject to the impact of social status.In other words, American NSs did not change the use of refusal formulas according to social status.The reverse was true for Iranian NSs, showing that Iranians were more sensitive to social status.This perception of social status was transferred to L2 English by Iranian learners.Compared with lower intermediate and intermediate learners, upper intermediate learners transferred more L1 social norms to L2 request, leading to more pragmatic failures.
For example, Takenoya (2003) investigated L2 learners' production of address forms and awareness of social factors underlying their choices of address forms.Participants were 52 American learners of Japanese and 85 Japanese NSs (53 males and 43 females).The learners formed three proficiency groups: beginning, intermediate, and advanced (proficiency determined by placement test scores).The instrument was a questionnaire with five tasks: (a) a DCT, (b) a conversation completion task, (c) an inference task, (d) a ranking task, and (e) a rating task.In the DCT, participants chose appropriate address forms from a list of 12 address forms.In the conversation completion task, the learners completed a dialogue with appropriate address forms.The inference task required the learners to read short dialogues between two Japanese speakers and to infer their relationship according to address terms used by the interlocutors.Findings showed that the learners in all three proficiency groups were able to produce native-like address forms.Higher proficiency learners performed better than lower proficiency learners, but there was no significant difference across proficiency levels.In terms of awareness of social variables, higher proficiency learners had better awareness of the gender factor in task situations than lower proficiency learners, but they did not have better awareness of closeness between interlocutors or addressee's social status than lower proficiency learners.Put differently, proficiency had different effects on L2 learners' awareness of different social variables involved in the use of address forms (e.g., addressee's gender, social status, and closeness between interlocutors).
Similarly, Taguchi (2006) revealed different proficiency effects according to social factors involved in task request-making situations.She used a role-play task to examine Japanese learners' production of requests in two types of situations.In one situation type, the power relationship between interlocutors was equal, the social distance between interlocutors was small, and the degree of imposition was low (PDR-low); in the other situation type, the listener had greater power, the social distance between interlocutors was large, and the degree of imposition was high (PDR-high).20 English NSs and 59 Japanese learners of English took part in the study.L2 learners were divided into high and low proficiency groups based on their TOEFL scores and teacher ratings of their oral proficiency.All participants completed the 4-scenario role-play task, and their overall appropriateness was evaluated with a 6-point scale by six experienced ESL instructors.Findings revealed that high proficiency learners had higher overall appropriateness scores than low proficiency learners, but the difference for PDR-low situations was smaller than that for PDR-high situations.In terms of request expressions, both groups underused mitigated-preparatory expressions (e.g., I'm wondering if + verb phrase), and overused hinting expressions in PDRhigh situations, but this deviation from native speaker pattern was greater for lower proficiency group.In PDR-low situations, on the other hand, both groups used more direct expressions, and there was no significant difference between the two proficiency groups.In other words, proficiency effects on L2 learners' production of requests varied according to social variables involved in task situations.Félix-Brasdefer (2007), on the other hand, examined L2 learners' production of three types of requests in Spanish: direct, conventional indirect and unconventional indirect in formal (learner-native speaker interactions) and informal (learner-learner interactions).Indirect requests were divided into two types: conventional indirect requests (query preparatory) and unconventional indirect requests (hinting).Participants were three groups of L2 Spanish learners in the USA: beginning, intermediate, advanced (15 students in each group).L2 proficiency was decided by course levels: Learners in the beginning group were taking second-semester Spanish courses; those in intermediate groups were taking sixth-semester Spanish courses; and advanced participants were taking eighth-semester Spanish courses.Participants provided their oral responses in seven request-making role-play situations.Findings showed that beginning learners used the highest percentage of direct requests in all situations, followed by the intermediate and advanced groups.In contrast, advanced learners used the highest percentage of conventional indirect requests, followed by the intermediate and beginning learners.However, there was no significant difference in production of unconventional indirect requests across proficiency levels (beginning with 6%, intermediate with 5%, advanced with 3%).These findings suggest that proficiency effects depend on the types of pragmalinguistic forms (degree of directness and conventionality) that learners are required to produce.
Aside from different types of pragmalinguistic forms, L2 proficiency also had different effects according to modalities of pragmatic performance (Bardovi-Harlig, 2008, 2009;Bardovi-Harlig & Bastos, 2011).For example, Bardovi-Harlig (2008) investigated the relationship between recognition and production of formulaic expressions in L2 pragmatics.Participants were 61 intermediate to advanced ESL learners (proficiency determined by course levels).Four tasks were administered to the participants: a self-reported recognition task, a context identification task, a DCT and a modified Vocabulary Knowledge Scale (developed by Wesche & Paribakht, 1996).In the self-reported recognition task, learners circled all the formulaic expressions that they knew from the given list of 63 items.In the context identification task, learners chose the situation in which they were likely to hear the target expressions.The DCT consisted of six scenarios that elicited the target formulaic expressions used in the recognition task.In the DCT task, participants read each situation and wrote their answers according to the situation.In the modified Vocabulary Knowledge Scale task, learners gave their recognition of six expressions.The following is an example (p.211): Results showed that all learners reported high recognition scores, but much lower production scores.Proficiency had a positive effect on the self-reported recognition task: Higher proficiency learners had higher recognition scores than lower proficiency learners.In contrast, there was not a significant increase in production scores across proficiency levels.These findings revealed that proficiency had different effects on recognition and production of formulaic expressions.
Finally, five studies revealed that the proficiency effect on pragmatic petence was mediated by length of stay in the target language community (Bardovi-Harlig & Bastos, 2011;Shardakova, 2005;Taguchi, 2011;2013;Xu, Case, & Wang, 2009).For example, Shardakova (2005) investigated effects of proficiency and length of stay in the target language community on L2 learners' production of apologies.Participants formed five groups: (a) 23 low proficiency American learners of Russian without in-country experience, (b) eight high proficiency learners of Russian without in-country experience, (c) 24 low proficiency learners of Russian with in-country experience, (d) 35 high proficiency learners of Russian with in-country experience, and (e) 41 NSs of Russian.Low proficiency was defined as OPI levels ranging from novice-high to intermediate-mid, while high proficiency was defined as OPI levels ranging from intermediate-high to advanced-mid.All participants completed a 21-item written DCT and an assessment questionnaire that required learners to rate interlocutors' power relationship and social distance in task situations as well as severity of the offense and obligation to apologize on a 3-point scale.It was found that more proficient learners employed a slightly greater number of apologies.However, increase in proficiency without study-abroad experience often resulted in overuse of polite strategies of apology.Moreover, increased proficiency had no effect on learners' perception of target social factors, but exposure to the target culture had a positive effect on learners' perception of target social factors involved in task situations.These findings suggest that compared with length of stay in the target language community, proficiency had a smaller effect on L2 learners' production of apologies and awareness of social factors involved in task situations.
In contrast, Xu, Case and Wang (2009) found that compared to length of stay in the target language community, proficiency had a more significant effect on L2 learners' awareness of pragmatic and grammatical errors.In their study, 126 ESL learners in the USA were first divided into high proficiency group (TOEFL scores above 550, N = 62) and low proficiency group (TOEFL scores below or equal to 550, N = 64).Then, each group was further divided into two subgroups according to length of residence (LOR): long LOR (more than 1 year) and short LOR (fewer or equal to 1 year).Therefore, there were four groups of participants: high proficiency with long LOR, high proficiency with short LOR, low proficiency with long LOR and low proficiency with short LOR.The instrument was a 20-senario questionnaire adopted from Bardovi-Harlig and Dörnyei (1998, see p. 13 for a sample item).Findings revealed that proficiency and length of stay in the target language community both played an important role in learners' awareness of pragmatic and grammatical errors.Compared to length of stay in the target language community, proficiency had a larger effect on L2 learners' awareness of pragmatic and grammatical errors, because the higher proficiency group may have more pragmalinguistic knowledge and be more sensitive to linguistic forms and their pragmatic functions.

Summary of findings and discussion
The synthesis question asks whether L2 proficiency affects adult learners' pragmatic competence.Findings from previous studies support an overall positive proficiency effect on L2 pragmatic performance.In particular, the existing literature has found that compared to lower proficiency learners, higher proficiency learners are better at minimizing negative L1 transfer in situations where L1 and L2 patterns differ (e.g., Maeshilba, Yoshinaga, Kasper, & Ross, 1996;Wannaruk, 2008).Moreover, more proficient learners are better at comprehending requests (Cook & Liddicoat, 2002) and identifying different types of speech acts (i.e., request, suggestion, correction and offer) (Garcia, 2004).With regard to pragmatic production, more competent learners have a larger repertoire of strategies to realize speech acts (e.g., Dalmau & Gotor, 2007;Trosborg, 1995), and are able to use more target-like pragmalinguistic forms such as intensifiers (e.g., Dalmau & Gotor, 2007), downgraders (e.g., Pinto, 2005) and contrastive markers (e.g., Geyer, 2007).
However, high L2 proficiency does not guarantee a native-like pragmatic performance.The positive role of proficiency in L2 pragmatics is mediated by pragmalinguistics and sociopragmatics aspects of pragmatic competence.
Pragmalinguistics involves knowledge of linguistic forms and their social functions.Therefore, it can be improved by increased proficiency which includes increased lexical, grammatical and discourse knowledge.In other words, proficiency has positive effects on L2 pragmalinguistics.This claim was supported by previous studies.For example, Dalmau and Gotor's (2007) study revealed that unlike lower proficiency learners, higher proficiency learners produced a greater range of apology strategies, more lexical intensifiers, and fewer nontarget-like apology expressions.These findings suggest that increased proficiency grants learners a wider range of pragmalinguistic forms to realize pragmatic functions.Also, Geyer (2007) found that more proficient learners employed a greater number of lexical and grammatical markers to qualify their utterances when they had to adjust their previous opinions, and they were capable of using contrastive expressions in conjunction with other discourse features, such as foregrounding main points of the utterance and putting the qualification segments in the background.These findings support the notion that higher proficiency learners have more lexical, grammatical and discourse knowledge, which allows them to outperform lower proficiency learners on pragmatic production at discourse level.
However, proficiency effects vary depending on different types of target pragmalinguistic forms, such as speech acts of different degrees of directness and conventionality.For example, Félix-Brasdefer (2007) showed a decline in the production of direct requests but an increase in the production of conventional indirect requests, and no significant difference in the production of unconventional indirect requests from lower to higher proficiency learners.One possible reason is that higher proficiency learners have more knowledge of fixed pragmalinguistic forms to perform conventional indirect requests, whereas lower proficiency learners have limited knowledge of pragmalinguistic forms; as a result, they overuse direct requests which only require simple pragmalinguistic forms.On the contrary, increased proficiency cannot improve production of unconventional indirect requests, which are not associated with fixed pragmalinguistic forms.
Aside from types of pragmalinguistic forms, different modalities of pragmatic performance can also mediate proficiency effect.For example, Bardovi-Harlig (2008) showed that proficiency only had a positive effect on recognition but not on production of formulaic expressions.Formulaic expressions are fixed or semifixed lexical strings that can serve pragmatic functions in situations.Therefore, the acquisition of formulaic expressions is similar to that of vocabulary knowledge, showing a gap between comprehension and production.These findings suggest that modalities of pragmatic performance should be taken into account when we discuss proficiency effect on L2 pragmatics.
The other aspect of pragmatic competence, namely, sociopragmatics, on the other hand, refers to the social aspect of pragmatic competence, which involves evaluation of social factors embedded in task situations such as interlocutors' social status, social distance and power relationship.It seems that compared to pragmalinguistics, sociopragmatics is less susceptible to the influence of proficiency because it does not directly focus on particular linguistic forms.For example, Takenoya (2003) showed that there was no significant difference in the production of address forms across proficiency levels, because higher proficiency learners did not have more native-like awareness of social variables reflected in task situations (e.g., addresses' social status and social distance between interlocutors).Similarly, Niezgoda and Roever (2001) found that high proficiency L2 English learners' awareness of pragmatic errors was lower than low proficiency learners'.The almost no effect of proficiency was also showed in Allami and Naeimi's (2011) study where Iranian learners with higher L2 proficiency transferred more L1 perception of social status to L2 English requests, leading to more pragmatic failures.These findings suggest that increased proficiency does not necessarily improve L2 learners' sociopragmatics.One possible reason is that L2 learners may need more time to realize differences between L1 and L2 sociopragmatic norms and make their decisions to conform to or resist the target norms.L2 learners may consciously choose to diverge from the target norms because they may want to maintain their foreign identity (e.g., Davis, 2007;Kim, 2014) or because the target norms do not match their perceived self-image as a second language speaker (e.g., LoCastro, 2001LoCastro, , 2012)).
In summary, previous studies have found that more proficient learners have more L2 pragmalinguistc forms to perform pragmatic functions, but proficiency effects vary depending on the nature of target pragmatic features.With regards to sociopragmatics, higher proficiency learners do not necessarily have better knowledge of target social norms, which is essential to a native-like pragmatic performance.In other words, L2 pragmalinguistcs and sociopragmatics are not equally influenced by proficiency levels.Increased proficiency can expand L2 learners' linguistic repertoire of pragmatic performance, but linguistic readiness does not directly lead to native-like pragmatic performance.It is learners' awareness of target sociopragmatic norms and their willingness to apply these norms to their language use that decide the degree of appropriateness of their L2 pragmatic performance.

Conclusion and future research directions
This synthesis study has found that there is an overall positive effect of proficiency on L2 pragmatic performance.However, the proficiency effects vary depending on the two aspects of pragmatic competence.A native-like pragmatic performance requires sufficient knowledge of both L2 pragmalinguistics and sociopragmatics, and these two aspects of pragmatic competence are reflected differently in various target pragmatic features.Therefore, future studies should compare proficiency effects on different target pragmatic features, which would give insight into how different aspects of pragmatic competence are affected by L2 general proficiency.Data from these studies may also shed light on the threshold proficiency required to achieve targetlike pragmatic performance, because the threshold proficiency should include fundamental knowledge of linguistic forms to perform basic pragmatic functions.
Future studies should also investigate the interplay between proficiency and different social factors involved in task situations as this would elucidate what social factors can mediate the effect of proficiency on L2 pragmatic performance.Research in this domain will also enrich our understanding of which variables are more influential in performing a particular pragmatic feature.
Another direction for future studies is to investigate the interplay between proficiency and learner-related variables such as the learning context (e.g., Shardakova, 2005;Xue, Case, & Wang, 2009), length of formal instruction, and attitude toward target culture.Such research would help explain the extent to which proficiency effects on L2 pragmatic competence are mediated by learnerrelated factors (i.e., individual differences).Learners' individual traits such as motivation, willingness to conform to the target norms and identity, and their personal learning experiences may affect the relationship between proficiency and different aspects of pragmatic competence (i.e., pragmalinguistics and sociopragmatics).For example, future research may want to investigate why some L2 learners with high proficiency choose to diverge from the native sociopragmatic norms.
Finally, most of the studies investigated examined pragmatic performance at the monologic level such as production of speech acts (e.g., Wannaruk, 2008), which could not represent pragmatic competence at the dialogic level such as sequential organization (e.g., Al-Gahtani & Roever, 2012).Future studies should examine the magnitude of proficiency effect on L2 learners' ability to act adaptively in interaction (i.e., pragmatic adaptability), because sufficient knowledge of pragmalinguistic forms and the functions performed by those forms in different situations could provide L2 learners with the foundation of pragmatic knowledge, while pragmatic adaptability determines the actual quality of their pragmatic performance in the sequential context of conversations.
Instructions: Choose the ONE answer that best describes your knowledge.Circle the letter.Example: A piece of cake (a) I don't remember having heard this expression before.(b) I have heard this expression before, but I don't know what it means.(c) I have heard this expression before, and I think it means (d) I know this expression.It means something is easy (e) I can use this expression in a conversation.Give an example: I can do that.It's a piece of cake.