Native-like Performance of Pragmatic Features: Speech Acts in Spanish

This study had as its goal to investigate how nonnative speakers (NNSs) of Spanish were able to perform pragmatics which in various ways resembled that of native speakers (NSs). The study focused on three advanced NNSs of Spanish who had contributed data six years earlier to a corpus of NS and NNS speech acts of complimenting, apologizing and refusing. The purpose was to do a contrastive analysis comparing the pragmatic performance of NNSs and NSs in order to capture both similarities and areas where highly competent NNSs displayed knowledge gaps, however subtle. The subjects responded to a language background questionnaire regarding their learning of Spanish and also completed a learning style preference survey. They were then asked to revisit their earlier performance in pragmatics from the corpus data and to describe the strategies that they used to produce their highly-rated performance in Spanish pragmatics at that time. The findings revealed ways in which the three subjects differentially imitated NS behavior, and provided insights as to how they arrived at native-like behavior in their facial expressions, use of clicks, physical contact practices, colloquial language, and cursing. The subjects’ reported learning style preferences appeared to be generally consistent with the strategies that they reported using for dealing with the pragmatic features of interest, such as the way that they dealt with cursing. Downloaded from Brill.com03/09/2022 01:12:10AM via free access 2 vacas Matos and Cohen 10.1163/26660393-bja10034 | Contrastive PragmaticS (2021) 1–30


Introduction
Any effort to describe the ability of NNSs to approach native-like performance requires identifying norms, which can be problematic. There are so-called ideal norms -where even the identification of the population to serve as the reference point for norms can be controversial -in contrast to judgments by NSs and advanced target-language (TL)1 users as to whether the observed pragmatic behavior fits their sense of what is appropriate in that situation within that context. One area where language educators often recommend that learners find examples of bona fide pragmatics in action is from TV talk shows and films. But the issue arises as to whether performance of pragmatics in the media actually mirrors unscripted, real-life behavior. The problem with examining TV interviews is that the interviewees may be on their best behavior. With regard to films, the actors' performance tends to be highly scripted and rehearsed. For instance, in order to make leading heroes in films more attractive to viewers, the script may have them behave less boorishly than they might actually behave in a given real-world situation. Furthermore, in real life both NSs and TL learners do not usually have the luxury of rehearsing their lines before delivering them, resulting in an unpolished delivery which in and of itself could result in pragmatic failure.
In addition, within the English as an International Language (EIL) and the English as a Lingua Franca (ELF) movements, there is an increasingly prominent view that it is inappropriate for teachers to use as their standard for performance that of NSs when it comes to TL pragmatics. Rather than laying the burden on students to conform to the norms of the TL speech community, there is a growing attitude that speech communities should simply be more 1 For the purpose of this article, target language (TL) will generally be used to refer to secondlanguage (L2) or foreign-language (FL) learning or use situations. In reality, especially given increasing access to the Internet worldwide, traditional definitions of L2 vs. FL learning and use may inaccurately describe the situation, which is more one of a continuum in terms of the nature of exposure. In other words, learners who are ostensibly in an FL situation may actually expose themselves more to native-like pragmatics than learners who are in an L2 situation but do not avail themselves of possible opportunities (such as through personal interactions on Zoom, streaming media, and the like).
accepting of performance in pragmatics that in the past perhaps would have been considered inappropriate -possibly even a cause for pragmatic failure. Without disputing the validity of these arguments, it may still be the case that many teachers and students worldwide do not share this view. Rather, they may hold the view that, regardless of how close learners of a language come to mastering the appropriate pragmatics for a given situation, there is value in being aware of what NSs tend to do and approximating these norms to whatever extent the given students wish to, given their learning style preferences and language learner strategy repertoire. The view taken in this study was to consider both positions as potentially valid and, in that spirit, to investigate further what being communicatively successful might actually entail.
This paper reports on a study based on a corpus of three speech acts performed both by NSs of Spanish and by intermediate and advanced English-speaking learners of Spanish from the US. The focus was on what it entailed for NNSs to be rated as highly successful from a communicative point of view on selected pragmatic features when performing three common speech acts -compliments, apologies, and refusals. The study took the stance of applying a close-up lens to pragmatic behavior. It considered what performance of pragmatics behavior might entail, realizing that both NSs and NNSs in a given speech community might prefer to view such behavior as interpretable through multiple norms, and thus rendering traditional contrasts in the pragmatics of a given speech community compared to another one as inappropriate.
A review of literature focusing on what we have and have not known with regard to native-like performance of TL speech acts will now be provided. Then, the methods for utilizing corpus data on L1 and TL performance will be presented. Next, the findings with regard to strategies for both the learning of TL pragmatics and for being rated as outstanding performers of speech acts will be identified and described. The discussion will consider both the advantages of this type of investigation as well as the limitations, and will offer suggestions for future research and possible pedagogical implications.

2
Review of Literature

2.1
Approaching Native-like Performance in TL Pragmatics Studies showcasing remarkable native-like performance in adult TL acquisition are not that common in the research literature. While there is some evidence that adult learners can achieve impressive abilities in a TL, others would maintain that it is quite rare. A seminal study investigated how a British woman, Julie, whose onset of Arabic was not until age 21, in an untutored setting in Cairo for 2½ years, was able to pass as an NS (Ioup, Boustagui, El Tigi and Moselle, 1994). Measures used to assess her achievement included a speech production task, a grammaticality judgment task, a translation task, an anaphoric interpretation task and an accent recognition task. The findings showed that Julie reportedly achieved native-like control over Egyptian phonology totally by mimicking the sounds. She had no noticeable foreign accent, made few mistakes in morphology and syntax, and had good control of the lexicon, including conventionalized forms. Her success was attributed to talent at language learning, spearheaded by exceptional cognitive flexibility in processing L2 input and in organizing it into a system.
A more recent, large-scale study of Spanish/Swedish bilinguals (n = 195) involved subjects with differing ages of acquisition (<1-47 years), all of whom had identified themselves as potentially native-like in their L2 (Abrahamsson and Hyltenstan, 2009). NS judges determined that, while a majority of those who acquired Swedish before the age of 12 were perceived as NSs of Swedish, only a small minority of those who had started their L2 acquisition after age 12 were. Furthermore, when a subset (n = 41) of the early-onset group was scrutinized in linguistic detail by means of a battery of 10 highly complex, cognitively demanding tasks and detailed measurements of linguistic performance, representation and processing, the results revealed that only a few actually exhibited native-like behavior on all the measures of L2 proficiency that were administered. The findings of the study led the researchers to surmise that not only were adult learners unlikely to attain native-like ability in an L2, but that in addition, attaining native-like ability appeared less common among child learners than had previously been assumed.
Studies that focus on native-like performance in the pragmatics of a TL are especially scarce in the literature, perhaps in part because of the difficulties obtaining blanket agreement as to norms for native-like performance in pragmatics. Empirical research has often focused on control over one specific discursive feature within the domain of pragmalinguistic. For example, one study looked at the extent to which US college-student FL speakers of French attained control over the c'est-cleft, which called for an ability to make inferences as to how inclusive the reference was, which depended both on syntax and on knowledge of pragmatic aspects of discourse (Destruel and Donaldson, 2017). The results from 40 learners at three FL proficiency levels revealed a development towards native-like behavior, especially for those in the high-proficiency group. While such studies focusing on a specific syntactic form suggest that native-like ability in pragmalinguistics can be attained, there are numerous other aspects of performance in pragmatics with a broader scope and potential impact where NNSs have clearly not attained such native-like ability. This would include various nonverbal behaviors which may have sociopragmatic as well as practice pragmalinguistic import.
Rather than focusing on outstanding native-like pragmatic performance, a few studies have scrutinized models of so-called "appropriate" TL pragmatics as portrayed in the media, since such sources have been recommended to learners. For example, a study was conducted comparing English NS compliments in interviews and films with those in naturally occurring data (Tatsuki and Nishizawa, 2005). The findings revealed that whereas these media were relatively reliable models of pragmalinguistic behavior in that there was syntactic similarity across the venues, behavior of a more sociopragmatic nature was less likely to reflect reality. With regard to gender distribution, for instance, in an analysis of interview data from Larry King Live, males were found to give and to receive as many compliments as females, whereas in naturally-occurring data females gave and received more compliments. In films, males were actually found to give and to receive more compliments than females.
Another underrepresented area in the research literature is that of what is called for in terms of nonverbal behavior in order for TL learners to be rated as native-like. A volume has appeared which deals with the issue of how to interpret each other's communicative intentions when relying on nonverbal communication cues given by Americans in face-to-face interactions (Gregersen and MacIntyre, 2017). The focus is on teaching gestures, posture, facial expressions, eye behavior, space and touch and prosody (i.e. the patterns of stress and intonation accompanying utterances), including a series of both decoding and encoding strategies for increased efficacy in nonverbal communication. There are numerous video clips such as one in which two American college students demonstrate 16 hand gestures in the clip "Say it with your hands!" (Gregersen and MacIntyre, 2017: 291).
The results of comparisons between NS and NNS performance are likely not to be fully informative in that there are invariably areas of comparative pragmatic behavior that are finessed despite the best of intentions. For example, it is possible that subtle instances of NNS divergence go unnoticed, although they do not reflect what NSs would do in the same situation. Likewise, it is possible that NNSs overuse certain structures, again not producing outright errors but resulting in more frequent use of certain forms than would be found in NS performance.
In principle, corpus linguistics can play a significant role in research on pragmatics (Rühlemann and Aijmer, 2015b), depending on the foci of given corpora and the manner in which the data were obtained. Corpora have the potential of providing researchers with insights into key factors in pragmatics (politeness, processability and relevance), speech acts (both in a synchronic and a diachronic perspective), pragmatic markers (discourse markers, stance markers and interjections), reference (e.g., deixis), turn-taking and prosody (see Rühlemann and Aijmer, 2015a).
However, most large corpora, like COCA,2 contain language samples which were collected without any particular pragmatics foci in mind. Even though there are corpora which do narrow the focus somewhat, such as MICASE,3 which represents spoken language from academic contexts, it can be challenging to extract robust samples of speech act behavior beyond greetings and leave takings. It is for this reason that research comparing L1 with TL performance in pragmatics is not usually based on large corpora -it can be difficult to identify examples of given types of pragmatic behavior through a form-to-function approach in the wealth of available data. Even if the data are identified, usually it is not possible to retrieve crucial contextual information about the data under investigation (Cohen, 2018: 127).
Up to this point, this literature review has identified limited research on achieving native-like performance in a TL altogether, as well as research on the verbal and nonverbal pragmatic behaviors of L1 users of a TL. The review has found there to be little research on the strategies used by NNSs both in learning TL pragmatics and in being rated as outstanding in performance of it -a gap which the current study aims to fill.

2.2
The Role of Learning Style Preference in the Learning and Performance of TL Pragmatics Another factor which could possibly impact comparisons between NS and NNS performance in pragmatics is that of individual differences, both in terms of intralingual and interlingual behavior. While not usually related to pragmatic performance, learning style preferences could potentially help to explain differences in the choice of strategies for performing TL pragmatics.
Whereas much of the learning style preference literature focuses just on sensory preferences (i.e., visual, auditory and kinesthetic; see Dörnyei and Chan, 2013), cognitive style preferences are likely to play a significant role in how learners deal with language tasks (Cohen, 2012). An early attempt was made by Ehrman and Leaver (2003) to conceptualize and categorize style preferences that were cognitive in nature, taking into consideration preferences that learners had with regard to cognition in learning language, featuring 10 subscales. These scales included, for example: -From random-intuitive to concrete-sequential: from preferring not to follow particular learning sequences to preferring to learn in a step-by-step manner; -From leveler to sharpener: from preferring to overlook linguistic distinctions in favor of focusing instead on similarities to preferring to notice differences and to seek distinctions. In order to achieve a more comprehensive measure of learning style preferences, the Learning Style Survey (Cohen, Oxford, & Chi, 2002) included scales not only for sensory/perceptual preferences, but also for cognitive style preferences and for personality-related (e.g., extroverted-introverted, impulsive-reflective) style preferences.4 Here is an example of each type of learning style preference: -Sensory learning style: visual -preference for charts, graphs, or a picture; -Cognitive learning style: concrete-sequential -preference for working stepby-step and following directions carefully; -Personality-related learning style: extroverted -active, interaction-oriented, and outgoing. Some of the learning style preferences interface with each other such that they jointly influence the completion of given tasks. For instance, more concrete-sequential performers of pragmatics might envision in their minds a continuum from least apologetic to most apologetic structures in the TL and select forms accordingly. Instead, more random-intuitive performers may perform the apology according to what just feels appropriate. Learners' learning style preferences are not necessarily fixed, but rather may vary along a continuum, depending on the situations in which they find themselves, with learners engaging in style-stretching out of their comfort zone (Cohen, 2012;Wang, 2018). For example, if learners tend to be impulsive in nature, they may nonetheless choose to be more reflective in order to achieve greater accuracy in their performance of TL pragmatics.
The current study aims to fill another gap in the research literature, namely, that of investigating the role that learning style preferences may play alongside language learner strategies such that TL learners are considered native-like in their pragmatic performance.

2.3
Research Questions The above literature review suggests that there is still much to learn about the strategies that learners use to native-like pragmatic performance, as well as how to mobilize corpus data in this effort. The goal of this study was to investigate the extent to which learners could explain the strategies that they used to achieve impressive results, as well as their level of awareness regarding NS behaviors absent from their performance.
Such research needs to take into account the issue of norms for NS behavior needs and the extent to which teachers and learners abide by some idealized norms. While teachers may wish to refrain from pronouncements as to what TL learners must do if they wish to be pragmatically appropriate, they still may feel obliged to provide learners with guidelines as to the types of pragmatics that would tend to characterize likely NS behavior in the given speech community. Needless to say, it is up to learners to determine the extent to which they will engage in what is considered appropriate pragmatic behavior, since learners have the right to exercise agency (Ishihara, 2010).
In an effort to apply a close-up lens to just what makes nonnative performance in pragmatics native-like, the current study addressed the following research questions: 1.
What strategies do outstanding NNS performers of speech acts report using to learn TL pragmatics? 2. What strategies do NNSs of a TL report using in order to perform the speech acts of complimenting, apologizing, and refusing in an outstanding manner? The pragmatic features of interest were intonation (for sarcasm or emotional support), nonverbal noises, facial expressions, gestures, body posture (dealing with personal space) and colloquial language (including cursing).

Prior Corpus Study as the Source for the Current Study
The Spanish Multimodal Corpus of Speech Acts5 was collected between 2014 and 2015, with the aim of investigating the extent to which Spanish language learners' level of pragmatic performance was consistent with their overall language proficiency level (Vacas Matos, 2017). What Vacas Matos found was that the results of her study were contrary to her expectations in that both high-and intermediate-level proficiency students had a low level of pragmatic ability. The few exceptional performers -in that their pragmatic performance was considerably above that of their peers -became candidates for the current study (see 3.2, below).
The COREMAH corpus was collected from students who were either at the B1 or C1 level of the European Framework of Reference for Languages, and from NSs. Vacas Matos conducted thorough analysis of the video clips which provided not only verbal but also valuable nonverbal behavior. Strategies of interest included the handling of gestures and facial expressions, cursing, proximity between the speakers, and physical contact.
The corpus compared Spanish NS performance on the speech acts of giving and responding to compliments and apologies, and refusing requests or offers (Vacas Matos, 2017). The data consisted of 180 recorded and transcribed video interactions produced by 48 English L1 advanced and intermediate speakers of Spanish and by 24 NSs of Spanish.
The videos showed that Spaniards engaged in physical contact when refusing an offer while the Americans refrained from it, and that the Spaniards also used more gestures and other means of nonverbal communication in order to display their emotions and intentions. In addition, Spaniards were found to be more likely than Americans to refuse a complement the first time it was offered because it would have been considered rude to accept it the first time around. It was also found that in giving compliments, the Americans smiled a lot, whereas Spaniards did not, since they felt that they had to look serious while insisting on the compliment that they were making. As was the case with refusals, the Spaniards used physical contact in order to intensify the compliment and to react to it.
The TL learners used more expressions of remorse in their apologies and responses to them, in contrast to the NSs, who resorted to excuses as well as cursing. Spaniards also used far more non-conciliatory strategies, especially displaying their reluctance to apologize through gestures signaling anger and reproach, even when they were apologizing. In contrast, the Americans made use of laughter and smiles, and refrained from cursing and from raising their voices.

3.2
Participants An analysis of data from this COREMAH corpus resulted in the selection by Vacas Matos of three students who at the time of the data collection were in the Spanish MA program at Middlebury College, VT, and had a C1 proficiency level on the CEFR6 scale. All three were selected from the larger corpus because Vacas Matos considered their performance in pragmatics to be truly outstanding. In fact, it was notably better than that of the other NNSs in the speech act corpus, and, in fact, more similar to that of the speech act performance of those NSs who performed the same role in their NS-NS pairs as did the NNSs in their pairs.
The three participants in the study will be referred to as Santi, Genevieve and Maya. Maya had studied four summers in Vermont, while Santi and Genevieve had studied for two summers in Vermont and had gone to Madrid for the year abroad program.7 The following descriptions of the three subjects' language backgrounds are based on their responses to a Language Background Questionnaire which elicited each subject's personal story of language learning (30-60 minutes). The questionnaire focused on Spanish: years studying the language, the number of languages that they spoke, and the nature of the contact that they had had with other cultures.
Santi had had four years of Spanish language study, starting in high school and resuming in college, after a hiatus. Then, he spent 23 months altogether studying abroad in Argentina, Spain and Perú. His Spanish teachers were from Cuba, El Salvador, Chile, Spain, Uruguay, Argentina, Mexico and Puerto Rico. While he reported having been a teacher in a charter school for several years, at the time of the study he had an administrative position with a nonprofit linguistics organization.
Genevieve had studied Spanish for 18 years starting in high school, and had spent 26 months of that time studying in Spain. Her teachers were from Spain, Cuba and Mexico, and she also had had contact with Salvadorians. She had been a Spanish teacher for different middle and high schools for 5 years prior to the collection of the corpus data and for six years subsequently.
Maya was the subject who had spent the most time abroad (6 years), living in Spain. She had begun her study of Spanish in high school. While most of her teachers had been from Spain, she reported having some contact with Mexican and Cuban Spanish speakers as well. She had been a Spanish teacher at different schools for 14 years prior to the collection of the corpus data and for six years subsequently.
While it was noted above that all three of the participants were deemed "outstanding" in their performance as NNSs of Spanish pragmatics, it was also noted that their performance was not homogeneous. For instance, Maya did better than the other two subjects in the "refusal of help" situation. She was offered help numerous times and her refusals included more of the strategies used by Spaniards, both in verbal excuses and in the use of body language.
As another example, Genevieve was the best of the three in terms of her responses to compliments, by first expressing doubt as to the truth value of the compliment, even denying it or playing it down, until she finally accepted it. Her performance suggested that she was experienced with how Spaniards dealt with compliments, as witnessed by the number of turns it took her to reject or at least doubt the compliment several times whether verbally or in gestures, before she finally accepted it.
Santi, on the other hand, was not so good with the speech act of complimenting in that he accepted the compliment the first time around, although he did so a bit jokingly. While his body language was like that of Spaniards, he reported in his interview that, given he was a modest person, he exercised his agency to give a more typical American response rather than to engage in multi-turn denial of the compliment before accepting it. In the speech act of apology, on the other hand, Santi reacted more like a Spaniard, asking for reasons for the infraction while openly displaying his annoyance, eventually joking about it at the end of the clip. His nonverbal body language showed, as with the other subjects, how experienced he was in these kinds of situations, and how adept he was at behaving as if he were highly offended.

3.3
Instruments 3.3.1 Learning Style Survey For this study, the Learning Style Survey developed by Cohen et al. (2002) was used in order to get a sense of how learning style preferences might have contributed to the learning and performing of TL pragmatic link. This survey was selected in order to obtain a comprehensive measure of learning style preferences, including not only the traditionally popular sensory/perceptual style preferences, but also the cognitive and personality-related style preferences.

3.3.2
Interview Protocol about Strategies for Pragmatics An interview protocol was created in order to explore what makes someone pragmatically native-like (30-60 minutes). The probing questions for the interview were based on Vacas Matos's analysis of the speech act performance on compliments, apologies, and refusals by the American learners of Spanish and by the NSs who contributed to the COREMAH corpus. Special attention was given to areas where NNSs used pragmatic behaviors typical of the NSs' responses in the corpus -for example, knowing when to increase their volume for effect, when it was appropriate to be rude or to make excuses.
The Interview Protocol was designed to prompt subjects to analyze the pragmatic features of interest (noted above, 2.6) in their performance in the video clips that they had made for that corpus and to compare their performance to that of NSs (see Appendix B for the Interview Protocol). There were both questions meant for all three subjects, as well as personalized questions for each of the three subjects, based on a careful analysis of their respective performances.

3.4
Procedures for Data Collection The three subjects were first asked to complete the self-scored Learning Style Survey8 and to familiarize themselves with a Language Background Questionnaire. The initial meeting involving the three subjects and the two investigators was conducted over Zoom in September of 2020. That meeting and all subsequent meetings were video-recorded and transcribed. The purpose of the first meeting was to familiarize the subjects with the design of the study, to make sure that all of them were comfortable with the plan, to talk with them collectively about their learning styles and to begin collecting their language background data. They had known each other from having completed the MA in Spanish through Spanish immersion at Middlebury College. Hence, participation in this study served as a form of reunion as well, adding a convivial atmosphere to the meeting.
Subsequently, the subjects met twice individually with the Vacas Matos over Zoom to discuss their strategies for both formal learning of pragmatics and for more informal acquisition of it.9 The researchers provided the participants with some basic background on pragmatics, which included the meanings for terms used commonly by linguists to describe pragmatic behavior. For instance, in the first interview, the distinction was made between teaching about the cultural behavior of Spaniards in contrast to the teaching of pragmatics, which involves both language and culture. This distinction enabled the three participants to report whether their Spanish teachers actually taught pragmatics.
In the second meeting, the subjects discussed individually with the Vacas Matos their linguistic background through the Language Background Questionnaire. During the third individual meeting, they were asked to share their responses to the Interview Protocol about Strategies for Pragmatics which they had been given in advance so as to focus their attention on the pragmatic features of interest noted above. Then, with these features in mind, they watched the videos of themselves engaged in compliments, apologies and refusals from the corpus five years earlier, as well as the video clips of several NSs who were also part of the corpus.
After the completion of the three individual interview sessions, the three subjects and the researchers once again met together in a joint Zoom session to discuss the study. This final group meeting provided an opportunity for the subjects to make any further comments if they wished to do so, and also for the investigators to possibly pose any further queries with regard to performance in pragmatics.

3.5
Procedures for Data Analysis After the video-taped interviews were transcribed, descriptive statistics and content analysis were performed on the responses. First, attention was given to the Language Background Questionnaire and to the Learning Style Survey in order to make comparisons across the three subjects. In addition, the subjects' performance on the three speech acts from the earlier study involving the COREMAH corpus was re-analyzed and compared to that of native speakers. The analysis involved annotation of utterances and accompanying nonverbal behavior. For example, counts were made as to the number of times they apologized, and if they did, the number of times that they touched the other speaker when they apologized.
Analysis of the video clips focused on both similarities and differences in performance among the participants, due both to features of the speech acts themselves and to individual differences. The subjects' retrospective explanations for their video-recorded performances elicited by means of the Interview Protocol about Strategies for Pragmatics provided further data for the analysis. Vacas Matos extracted and transcribed interview data that included descriptions of the strategies that subjects reported using. Then, through Zoom videoconferencing, the two researchers exchanged observations and comments, discussing and reaching an agreement on the interpretations of speakers' strategies.

Findings
In Section 4.1, findings with regard to the subjects' learning of pragmatics will be presented, and then in Section 4.2, the focus will shift to the strategies that the subjects used in order to perform pragmatics in a native-like fashion.

Strategies That Outstanding NNS Performers of Speech Acts Used to Learn Spanish Pragmatics
The Subjects' Learning Style Preferences The following are results for the self-scoring Learning Style Survey: Santi reported being auditory, random-intuitive, open-oriented, global, a synthesizer, inductive, extraverted and impulsive.10 Genevieve reported being visual, concrete-sequential, closure-oriented, detail-oriented, analytic, deductive, introverted and reflective. Maya reported being visual, random-intuitive, closure-oriented, detail-oriented, a synthesizer, inductive, extraverted and impulsive.

4.1.2
The Subjects' Reported Language Learning Strategies First the strategies that all three subjects shared are presented, then those shared by two of them, and then those unique to one person.

4.1.2.1
Strategies Shared by All Subjects The following are strategies that Santi, Genevieve and Maya reported in common for learning pragmatics: -Watching media in an active way -taking notes, transcribing, stopping and going back in the time sequence to review what was said and how, analyzing the scenes and even making role-plays from them for a class; -Taking note of cultural differences among NSs and NNSs and how these differences influenced the performance of pragmatics; -Observing intonation and playing with it when speaking Spanish and even teaching it; -Paying attention to curses, asking NSs how to use them in and then using them often enough themselves so that their use would become more or less automatic in appropriate situations; -Accepting physical contact as practiced by Spaniards -to the point of embracing it as part of their local modus vivendi; -Engaging in regular interactions with NSs in order to gain control over pragmatic elements of the language and culture.

Strategies Common to Two of the Subjects
The following are strategies that both Santi and Genevieve reported using. 10 Three of the Learning Style Survey items are not reported here, since they were not relevant to this study.
-Attempting to detach themselves from their American selves when using Spanish in class -so that whatever they said, they would not lose face over it; -Scrutinizing metalinguistically the role of connectors and pause fillers in discourse; -Giving as much metalinguistic attention to colloquialisms; -Noticing when and how NSs used sarcasm for pragmatic effect; -Observing the facial expressions and gestures used by NSs and then making a conscious effort to imitate them; -When they lacked correct words and phrases, making an effort to employ nonverbal means of communication that they had seen NSs use -for example, clicks and gestures. Both Santi and Genevieve shared an attitude that the acquisition of pragmatics was a process which they had to be conscious about from the very beginning. They noted a need to play around with the language, which included imitating NSs. They felt that ultimately it might become natural, but initially it would involve a high level of general awareness and attention to specific types of pragmatic behaviors.

4.1.2.3
Strategies Unique to Just One of the Subjects All three of the subjects reported at least one strategy that they alone used. Santi noted that he had strategies that he reserved for the learning and use of different languages, like journaling or writing down situations and the specific words used for it, reflecting metalinguistically about a situation and the pragmatics related to it, or using a virtual forum11 to understand it better. He reported being an autonomous learner, who was used to devising and deploying a number of strategies to enhance his language learning. As he noted in his interview: "This cannot be taught to students … you gotta want it … I think you can teach people the tools, to say like, here is the model of someone that has done this, and this is what they did to get there." Additionally, he reported having his own theories about how his NS culture influenced the process of learning and acquiring pragmatics. For instance, he thought that being raised as a Catholic helped to explain his behavior in numerous interactions with Spaniards. The following were some of the strategies that he alone reported using: -Frequently asking people how to react (e.g., what the right words would be) in a given situation, such as at a surprise birthday party; 11 Wordreference.com.
-Using tools such as dictionaries and online forums to find out which expressions worked and how to react verbally in different situations; -Keeping a journal in Spanish and, while writing, reflecting on the specific linguistic forms that he was using; -Using flashcards expressly targeting pragmatics, including nonverbal communication such as gestures; -Writing down idioms that he had heard in conversations in order to look for their meaning afterwards; -Quizzing himself and preparing for future conversations. Santi's attitude toward strategizing can be summarized as follows. He expressed the view that learners needed to be proactive if they wanted to gain excellence in their pragmatic development. He went on to point out that, whereas it is possible to teach learners strategies such as the ones that he used, it would be wrong to assume that these learners would not only learn them and but also make use of them in their own performance. He felt that learning and using pragmatics was a highly individual matter. It was not something that could be rubber-stamped.
Genevieve reported only one strategy unique to her -a strategy for identifying what it was she should be learning. She reported asking at least two NSs about the same pragmatic behavior. For example, she would ask them how to react to something or if something was correct/normal in Spain.
Finally, Maya's approach to strategizing about pragmatics was different from that of Santi and Genevieve. She did not take such a proactive and programmed approach to the deployment of strategies in the TL community as they did. She expressed the view that much of her pragmatics was acquired without effort -rather, by simply being in the environment and communicating with the NSs. She felt that her being an actress helped enormously with the process of learning and acquiring pragmatics.

4.1.3
The Relationship between Learning Style Preferences and Strategy Choices While there were clearly strategies shared by all three of the subjects, there were differences in their learning style preferences. The result was that they achieved high ratings in their use of Spanish through complementary paths.
Santi could be considered the best example of someone who depended largely on conscious use of strategies in his learning of pragmatics. In terms of his learning style preferences, rather than being concrete-sequential in his learning, he was random-intuitive, open-oriented, global and a synthesizer.
Genevieve's reported approach involved being observant and reflectivetaking a good look before she leapt. She attributed her success to her ability to analyze language and cultural elements and subsequently to perform them. With regard to style preferences, she indicated that she was an introvert who liked to do things in sequence, an analyzer, and a person who liked to know the rules and then use them. Her reported strategy use appeared to be consistent with these learning style preferences.
Maya reported that what worked for her was to submerse herself in the TL culture, and that her continuous relationships with local people was what contributed greatly to her informal acquisition and formal learning of pragmatics. She reported that this approach enabled her to get high ratings for pragmatics. Maya was more of an extrovert, impulsive, and random-intuitive on the one hand, and on the other, she was closure-oriented, someone who focused on details, and a synthesizer. Her reported strategy use seemed to be consistent with these learning style preferences.
Santi and Maya's learning style preferences of being globally (big-picture)oriented, extroverted, and impulsive reportedly had some influence on how they dealt with cursing in Spanish, just as Genevieve's more analytic, concretesequential, introverted and reflective learning style preferences influenced to some extent how she dealt with it. While the cursing of the first two was more intuitive and free-flow, Genevieve was more reflective in her approach to cursing, preferring to take a more reflective rather than impulsive approach.
Overall, the three subjects reported sharing many learning style preferences. However, Santi reported being more auditory in contrast to Genevieve and Maya, who said that they were more visual. Santi and Maya were reportedly more random-intuitive and global, in contrast to Genevieve's reportedly being more concrete-sequential and detail-oriented. Santi and Maya described themselves as more extroverted and impulsive, while Genevieve described herself as more introverted and reflective. Of course, while these labels provide some measure of descriptive power, one should consider that these learners might also have stretched outside their comfort zone of learning style preferences.

Strategies Used by NNSs in Order to Perform Speech Acts in a Manner
Rated as Outstanding During the second individual interview, the subjects were asked about their awareness of acting like an NS during the sessions recorded for the COREMAH corpus six years earlier. For every speech act, they first watched their own performance, offering a detailed description of what they did like an NS, and afterwards, they watched a pair of NSs performing the same speech act. In some way, the three of them were surprised at their native-like performances. Some of them acknowledged that they were doing it unconsciously, reflecting the degree to which they had internalized these behaviors.
For instance, Santi reported thinking that he may have adopted some of the gestural behavior from soap operas he had watched in Argentina when he was living there. He expressed his surprise at finding that he had acquired this nonverbal behavior so well since he was focusing just on his verbal language. At times, the subjects were fully conscious that they were using TL gestures, and at other times using gestures from the L1, and aware that their interlocutors did not use or understand them.
Whereas at times discussion of pragmatic performance in individual sessions involved talk about pragmatic differences across cultures, at other times the participants were not able to articulate how they arrived at using appropriate TL pragmatics. Maya, for example, attributed her prowess in the local pragmatics as a function of being immersed in the language and culture.
The following are the findings for the three subjects, including mention of the level of awareness that the speakers exhibited with regard to how they used the TL so well.
-Gestures: All three subjects indicated that they were keenly aware of gestures. Not only did Santi indicate that he was highly aware of the use of gestures, but actually reported using a book called Che Boludo 'Hey, Man' that explained them. He said that he used them as a tool to compensate for his limited ability to communicate in the TL, Spanish. Genevieve was also very aware of gestures, and referred to her former roommate's gestures as an example. She went on to describe perceived differences between Americans and Spaniards -that Americans were more expressive with their facial expressions and Spaniards with their gestures.
Likewise, Maya described herself as very aware of gestures, stating that "gestures are double communication, to establish a feeling." She added that she often used gestures in English as well. She noticed that Spaniards used more gestures than Americans. She expressed the view that the use of gestures was important if someone wanted to be considered an advanced speaker of Spanish.
-Facility with facial expressions: Santi was acutely aware of the importance of using appropriate facial expressions in TL interactions and gave some examples of the expressions that he would use. Genevieve was not so conscious of them, and Maya was not conscious of them at all. -Physical contact with their conversation partner: Despite the fact that he came from a very affectionate family, Santi still found that at first it was difficult for him to greet Spanish men by kissing them. So, except for the kissing part (which, by the way, he had also experienced in Argentina), when he got to Spain, Santi reported finding it easy to incorporate physical contact into his routine. Although Genevieve reported being a "toucher" in her normal English-language interactions, she still had to adjust to the difference in the physical space between speakers in the US as compared to Spain, with people keeping less distance from one another in the latter case. She said that, while she grew to incorporate the pragmatic norms for physical contact within Spanish culture, nonetheless she would never initiate coming nearer to an interlocutor than was the custom in the US. Maya reported that she fit right in when it came to physical contact within Spanish culture because that was how she perceived herself as behaving in the US culture as well. She noted that, although she was from a very affectionate family, she was nonetheless shocked at first by how close Spaniards would be to each other when they interacted. -Clicks: Santi realized that clicks were meaningful in Spanish, and he reported always trying to imitate them. Genevieve indicated that she was also very aware of clicks -and in other languages as well. She referred to a similar dental click /ts/ used in both Jordan and Spain (which meant 'no' in both cultures), and how the use of this click sound among Americans in the US could signal disrespect or even disgust. She said that she tried to imitate these clicks -both how they were pronounced and where they occurred in an utterance -and that she especially used them consciously when she was living in Jordan. -Raising one's voice: Santi considered speaking more loudly to be culturally appropriate, since it was his view that Spaniards enjoyed adding drama to their interactions in a passionate way ("like, over the top"). As he put it, on occasion he would notice "some sort of release of pressure instead of keeping it under wraps, and pushing down your feelings. You could really release the feelings that you're having at that moment." Maya also perceived Spaniards as being a bit more histrionic than the somewhat more reserved Americans. Genevieve shared an observation that Americans seemed to talk more loudly when they were outside the United States. -Intonation: While all three subjects indicated that they were acutely aware of intonation, they reported dealing with it in different ways. Santi remarked that he paid attention to intonation as part of his active listening in order to understand more fully what was being said. As he put it, "intonation comes first, and then, words." For him, tone was the first clue to the intention of the speaker and the one that dictated how he would react. Genevieve reported that sarcastic intonation caught her attention -that for her, sarcasm was the thing that stood out even more than physical contact and gestures. Nonetheless, she noted that she did not teach intonation to her students because she felt that their grasp of Spanish was too low for them to be able to get it. Maya commented that she was aware of intonation in Spanish, as well as in English. She thought that intonation was a very important way to transmit feelings, and so she felt that it was important to teach her students Spanish intonational patterns. -Colloquial language, fillers and markers: Santi reported really enjoying the learning of these features since "managing them is what differentiates a person who is learning Spanish from a Spanish speaker." The above-mentioned book that he had purchased to help him with gestures, Che Boludo, dealt with colloquial language as well. He reported taking the book with him wherever he went. He also used internet tools like the WordReference Forum to learn more about these various features. Genevieve also loved the topic. Not only did she report being very aware of colloquial language, fillers and markers, but in addition she indicated that she was teaching them to her students. She noted that she herself used them constantly, but admitted being surprised at how frequently she was, in fact, using them when she observed her performance on the videotaped clips in the corpus. In addition, she pointed out that some of these features were easier for her because they had direct equivalents in English. Maya stated that she had learned these pragmatic features in the streets of Spain. She noted that she paid particular attention to colloquial expressions when used in movies and TV series that she watched online, and that she would explain their meaning in English to her partner. -Cursing: Santi said that he was very attentive when it came to curses -that he loved cursing, while at the same time clarifying that he used curses for emphasis, not out of anger. He reported that he asked NSs about the meaning of curses, tried to imitate the way that NSs perform them and that he was continually learning new ones. Genevieve reported having a teacher who taught them curses, and that whenever she had an opportunity, she would ask NSs about them. She went on to say that she swore much more in Spanish than in English, and said that, while she understood heavy-duty cursing, she preferred to teach her students how to use mild curses. Maya indicated that she paid attention to curses and asked NSs about them, although she did not remember making any efforts on her part to imitate them. Rather, she felt that they just came naturally. Maya recounted that her ex-boyfriend's family used them all the time in Spain, and that she herself cursed a lot in Spanish (e.g. by using the Spanish equivalent of the f-word, joder).

Summary of Findings
This study focused on three advanced NNSs of Spanish who had contributed data six years earlier to a corpus of NS and NNS performance on the speech acts of complimenting, apologizing and refusing. The subjects responded to a language background questionnaire which asked about their learning of Spanish, with a focus on pragmatics. In addition, they also completed a survey to determine their learning style preferences. They were then asked to revisit their earlier pragmatic performance in the corpus data and to explain how they had achieved such excellence in their performance of Spanish speech acts at that time.
The subjects were found to have similarities and differences in the strategies that they used in their performance of pragmatics with regard to features of interest such as gestures, facial expressions, clicks, colloquial language, fillers, markers, and curses. It was found that the subjects' descriptions of strategy use were by and large consistent with their reported learning style preferences. For example, how subjects positioned themselves along the continuum from introverted to extroverted was generally consistent with the strategies that they reported using both for learning and for performing the pragmatic features of interest in the study (such as cursing).

5.2
Limitations One limitation of this study was that selection of the three subjects as being outstanding performers on the three speech acts in the corpus was conducted only by Vacas Matos, in consultation with Cohen. The selection process did not include input from other NSs regarding their impressions as to how nativelike the subjects' performance was in comparison to that of the other NNSs in the corpus.
In addition, in order for this study to capitalize on the existence of a rich corpus of data on L1 and TL Spanish pragmatics, the three subjects were called upon to recollect how they had learned their pragmatics prior to the corpus study six years earlier. Furthermore, based on their viewing of the video clips, they were called upon to reconstruct the strategies that they had used in order to produce what was deemed outstanding performance. Hence, it is possible that attrition had impacted the subjects' recall both of their strategies for learning pragmatics and of the strategies that they had selected for use in their video-recorded speech act performances.
There was no provision for subjects to offer verbal report (VR) data after engaging in speech act performance. While collecting VR data during performance is problematic, it might have been possible to collect such data between the speech act vignettes included in the collection of the corpus. That way, assuming the subjects had been oriented to focus on the pragmatic features of interest, it would have been possible to collect strategy information while it was still fresh in their minds.
Another limitation was that, although Vacas Matos conducted most of the interviews in Spanish, which gave her an impressionistic sense of the subjects' abilities with pragmatics at the time of the interviews, no formal assessment was conducted, such as by having them perform the same three speech acts again. Such assessment could also have been used as a means for determining the impact of attrition on their performance of speech acts.

5.3
Interpretation By having subjects comment on the video clips of their speech act production for the COREMAH corpus, it was possible to identify and describe in this study a number of strategies in performance areas that until recently had tended to be ignored in empirical studies on TL pragmatics. The descriptions covered speech acts such as cursing, matters of prosody such as intonation as a means for signaling sarcasm and nonverbal behaviors such as facial expressions, gestures and ways of dealing with physical contact. Direct access to a corpus with both L1 and TL pragmatics made it possible to conduct a comprehensive contrastive analysis of what contributed to the NNSs' excellence in the performance in pragmatics, which is not usually the case with studies involving the more typical discourse-completion or role-play data.
The benefit of having subjects complete a learning style preference survey was that it helped to demonstrate how excellence in a language performance area such as pragmatics could be achieved by individuals having different learning style preferences. A major finding from the analysis of the learning style preference data was that although the three individuals reported somewhat differing styles and strategies for dealing with the learning and use of the TL, they all were rated as outstanding performers in pragmatics. In addition, no one strategy repertoire was found to be more beneficial than another. Likewise, it was found that sometimes the same strategies were reportedly used in different ways. The data revealed that, while much of the subjects' performance was native-like, there were gaps -either due to the influence of the L1 (whether inadvertently or by choice) or due to a lack of knowledge.
Whereas the viewing of the video clips provided a means for jogging the subjects' memories as to their strategies for producing stellar performances, undoubtedly some attrition had set in, which could have influenced the reliability and consequently the validity of their explanations.
While both Genevieve and Maya were at the time of this study Spanish teachers, and consequently still actively engaged in teaching about and performing pragmatics, Santi was not. The fact that his contact with Spanish was significantly less than what it had been may have had a deleterious effect on his ability to respond to the investigator's queries.

5.4
Suggestions for Future Research Given the pioneering nature of this study, there are no comparable previous studies with which to compare it. Thus, this gap affords researchers a niche within which to conduct future research. One suggestion for future study would be to elicit speech act production from TL learners who have differing degrees of contact with the TL. The first group could be those in a relatively traditional FL situation -namely, a situation in which they have no contact with the language other than through language lessons and accompanying materials. The second group could be those in an L2 environment where the TL is spoken around them and in which they have regular contact with that language. The advantage of this research path would be to explore the extent to which, because of increasing global connections, language learners gain rich exposure to the pragmatic features of interest, regardless of their language learning environment.
Another direction for future research would be that of investigating the performance in pragmatics by heritage speakers of the TL. Whereas they may have difficulty in certain instructed areas such as grammar, heritage speakers may feel more comfortable when it comes to L1 pragmatics as manifested in certain domains, such as that of the home and the local community. It would thus be possible to investigate the extent to which heritage speakers of the TL transfer their dominant-language pragmatics to their TL pragmatics performance.
Especially since many teachers consider pragmatics something that should be addressed with more advanced students, another area for investigation would be that of the differential impact of instruction regarding the fine-tuning of pragmatics especially appropriate for students at different proficiency levels. Greetings, for example, are needed in the very first lessons of a TL course. Such would also be the case with expressions of gratitude, requests and other speech acts. Whereas there are certainly features in pragmatics that could be postponed until more advanced proficiency levels, a question worth exploring would be which of the more basic niceties of physical contact, prosody and the like, could be of benefit to beginning learners of a TL, given their language backgrounds and learning style preferences.

5.5
Pedagogical Implications Issues of facial expressions and gestures, along with matters of prosody, are not usually topics for instruction, both because such information is characteristically not found in textbooks and because even more basic information about, say, speech acts tends not to be taught widely. And if it is taught, it does not tend to be taught explicitly, although the research would suggest that explicit teaching is effective (see, for example, Nguyen, Pham, T. H. and Pham, M. T., 2015). So, it remains for classroom teachers to determine whether they wish to focus on matters highlighted in the findings of this study. Teachers could survey their students to determine whether they desire formal instruction, and if so, to select areas on which to focus.
One pedagogical implication of this study would be that some of the behaviors which are associated with NSs of the TL are usually beyond the areas that teachers are likely to include in their classroom instruction, for numerous reasons such as the assumption that learners are unlikely to remember the information. Ironically, pragmatic niceties may be just what students do remember.
Another pedagogical implication is based on the subjects' reporting that being in the midst of NSs enhanced their performance of pragmatics, namely that FL learners are at a disadvantage in this respect. Consequently, the development of pragmatic ability especially among FL learners would likely benefit from their becoming more explicitly aware of both similarities and differences in pragmatic behavior when comparing the L1 and the TL speech communities. Classroom activities could include those in which students engage in analyzing and critiquing their perceptions of the pragmatics in given interactions, where the interlocutors' age, gender, socioeconomic level and other factors come into play (see examples of such activities provided in Cohen, 2018).
A final implication could be that in areas such as degree of physical contact with an interlocutor, teachers may wish to honor their students' right to agency with regard to whether they opt to perform the pragmatics considered appropriate by the given TL speech community for that context. In other words, the teacher can have a conversation with students as to the degree to which they would make an effort to adhere to TL norms for pragmatic behavior. Some students may feel that embracing certain TL behaviors would jeopardize their own cultural identity. Hence, it could be the teachers' role to support learners in experimenting with TL pragmatics so "they will enjoy a greater level of acceptance or insight into the target culture" (Bardovi-Harlig and Mahan Taylor, 2003:6).

5.6
Conclusions Employing a corpus like the COREMAH, which was productively used in this study with both native and non-native speakers, provided an opportunity not only to see differences in speech act performance between NSs and NNSs, but more importantly to see ways in which NNSs performed like NSs, and to explore reasons for this. While acknowledging that performing like NSs is not necessarily essential for NNSs to be successful in their interactions, this study would suggest that approximating NS behaviors can serve at times to enhance such interactions. One caveat to bear in mind is that the perception of what constitutes ideal norms for performance needs to be somewhat fluid rather than being fixed.
By means of discussing both the learning and the performance of pragmatics with the three subjects in this study at length, it was possible to identify certain patterns that facilitated both the learning of Spanish pragmatics and its performance in given speech act situations. As indicated above, having the subjects complete a learning style preference survey assisted in the interpretation of data in that this information revealed that excellent, even native-like behavior was arrived at by learners who had, at least to some extent, their own repertoire of strategies for both learning and performing TL pragmatics. This finding underscores the reality that learners differ in their pathways to excellence, and that teachers need to be mindful of this.

Bio Notes
Marta Vacas Matos (Ph.D., Autonomous University of Madrid) created an open-access, Multimodal Corpus of Speech Acts in Spanish as part of her dissertation, and has been instructor at Accent Global Learning Madrid and IES Abroad Madrid, where she developed the new curriculum for heritage speakers in study abroad in Madrid. Additionally, she has been linguistics advisor for children's language learning apps, multimodal materials designer at the Language Media Center, University of Texas at Austin, and instructor in Spanish teacher development, University of Pittsburgh.
Have you tried to imitate any of this behavior? How do you feel about it?
11. How do you handle cursing? a. What is your reaction when someone curses in Spanish? b.
How much do you yourself curse in Spanish? If you do, can you give an example? c.
Have you ever asked a native speaker about a curse -perhaps one that they used -in terms of the significance of that curse in the specific instance? d.
Have you tried to imitate any of this behavior? How do you feel about it?
12. What about physical contact? a.
From your experience what are ways that speakers make physical contact with others? b.
Do you feel uncomfortable when someone touches you while speaking with you? c.
To what extent have you incorporated frequent hugs and kisses into your interactions with Spaniards?