I. INTRODUCTION
We do not learn language merely to convey information. Through language, we maintain relationships, show consideration for others, avoid conflicts, and protect one another’s face. In other words, language functions as a social tool through which people establish and maintain interpersonal relationships (Upton, 2006, Chapter 7). For example, the expressions “I can’t” and “I’d love to, but I already have plans” convey the same propositional meaning, yet their social effects are entirely different. Even when the intended meaning is identical, the choice of expression can substantially influence how the speaker is perceived and how interpersonal relationships are maintained. Therefore, grammatical accuracy alone cannot guarantee successful communication (Bardovi-Harlig & Hartford, 1993). Rather, learners need pragmatic competence, namely the ability to select appropriate expressions according to situations, contexts, and interpersonal relationships.
Pragmatic competence can be defined as “knowing how to say what to whom when” (Bardovi-Harlig, 2013, p. 68). In this regard, speech acts play a crucial role in communication. Speech acts refer not merely to linguistic utterances but to social actions performed through language (Austin, 1975). Requests, apologies, compliments, and refusals constitute fundamental units of actual communication and must be performed appropriately according to contextual and interpersonal factors (Kasper & Rose, 2002).
Among various speech acts, refusals are considered particularly challenging, because speakers must reject others’ requests or suggestions while simultaneously maintaining social relationships. As a result, indirect intentions and contextual meanings become more important than literal meanings. Speakers often avoid direct refusals and employ various mitigating strategies such as apologies, excuses, empathy, and gratitude to protect the hearer’s face (Beebe et al., 1990). Furthermore, the same refusal intention may require different linguistic strategies depending on contexts, interpersonal relationships, and social distance. Due to these characteristics, second language learners often experience difficulties in performing refusal speech acts appropriately. Therefore, refusal provides an appropriate domain for investigating the development of pragmatic competence.
The complexity and context-dependency of refusal speech acts suggest that learners cannot acquire pragmatic competence merely through learning linguistic forms. Thus, effective pragmatic learning requires opportunities to experience various realizations of speech acts in authentic contexts. Previous studies in L2 pragmatics have consistently emphasized the importance of authentic and contextualized input (Bardovi-Harlig et al., 2017; Yilmaz & Koban Koç, 2020) as well as the necessity of explicit instruction and appropriate guidance (Bardovi-Harlig & Vellenga, 2012; Taguchi, 2010).
From this perspective, corpora and multimodal materials can serve as valuable instructional resources for pragmatic instruction. Corpora provide large amounts of authentic and naturally occurring language data (Bardovi-Harlig & Mossman, 2022), while multimodal input, such as films and videos, offers rich contextual information regarding interpersonal relationships and communicative situations (Bardovi-Harlig & Vellenga, 2012; Furniss, 2016). In particular, a corpus is especially suitable for studying pragmatics because corpus compilers are inherently concerned with the contexts in which language is produced (Staples & Fernández, 2019).
Nevertheless, corpus-based language learning has predominantly focused on vocabulary and grammar instruction (Boulton & Cobb, 2017), while its application to L2 pragmatic instruction, particularly speech act learning, has received relatively limited attention. One challenge lies in the difficulty of identifying and retrieving speech acts from corpus data, as pragmatic meanings are not always realized through fixed linguistic forms (Staples & Fernández, 2019). Consequently, previous corpus-based pragmatic instruction has largely focused on pragmatic routines and conventional expressions (e.g., routines for agreement, disagreement, and clarification, such as “I agree,” “I don’t know,” and “That’s a good point”) whose relatively formulaic nature makes them more amenable to corpus-based learning (Bardovi-Harlig, 2023; Bardovi-Harlig, Mossman, & Su, 2017; Bardovi-Harlig, Mossman, & Vellenga, 2015). In contrast, context-dependent speech acts, such as indirect refusals, remain relatively underexplored. Unlike grammatical structures or lexical items, indirect refusals require learners to consider contextual factors, interpersonal relationships, and speakers’ intentions in interpreting their meanings.
To address these limitations, the present study integrates film-based contextual input with corpus exploration. By first presenting indirect refusals in authentic interactional contexts through film and subsequently guiding learners to examine their usage patterns in corpus data, this study aims to help learners develop a deeper understanding of the relationship between linguistic forms, pragmatic functions, and contextual meanings, which can lead to enhanced pragmatic competence. The following research questions guided the study:
1. How does corpus-informed instruction using multimodal input affect EFL learners’ pragmatic competence in indirect refusal speech acts?
2. What are learners’ perceptions of corpus-informed instruction for learning indirect refusal strategies?
II. LITERATURE REVIEW
1. Indirect Refusal Speech Acts and Pragmatic Competence
Refusals are considered to be among the most face-threatening acts because they contradict the listener’s expectations and threaten some aspects of the hearer’s positive face (Brown & Levinson, 1987; Eslami, 2010). When refusals are not appropriately expressed, they may cause unintended offense and disrupt communication (Taguchi, 2013). In addition, since using inappropriate refusals can have a negative influence on speakers’ interpersonal relationships, speakers usually employ several strategies to avoid offending their interlocutors (Beebe et al., 1990).
In many cases, refusals are realized indirectly rather than explicitly. Instead of directly saying “No” or “I can’t,” speakers frequently use expressions such as apologies, excuses, gratitude, empathy, wishes, or alternative suggestions (Beebe et al., 1990). Such indirectness allows speakers to reduce the imposition of the refusal and protect the interlocutor’s face. Searle (1975) argues that indirect speech acts involve a discrepancy between the literal meaning of an utterance and the speaker’s intended meaning. Therefore, successful interpretation of indirect refusals requires listeners to infer implied meanings by considering contextual and interpersonal information.
Unlike grammatical knowledge, which primarily concerns linguistic forms, indirect refusals require learners to integrate contextual information, social relationships, and communicative intentions. Thomas (1983) points out that pragmatic failure occurs when speakers produce linguistically correct utterances that are socially inappropriate. Similarly, Kasper and Rose (2002) argue that L2 learners often experience difficulties in selecting appropriate speech act strategies because they lack sufficient pragmatic knowledge and sociocultural awareness. Previous studies have reported that L2 learners tend to overuse direct refusals, rely on a limited number of strategies, or transfer pragmatic norms from their first language (Beebe et al., 1990; Taguchi, 2015).
Because the appropriateness of refusals varies according to social distance, power relationships, and situational contexts, learners must determine not only what to say but also how, when, and to whom it should be said. According to Cohen (2008), native speakers’ intuitive pragmatic competence can be understood as knowing how and when to use particular speech acts appropriately, whereas non-native speakers often have difficulty acquiring such knowledge. In this respect, the ability to perform indirect refusals appropriately constitutes an important component of pragmatic competence.
Bardovi-Harlig (2013) defines pragmatic competence as knowing “how to say what to whom when” (p. 68), emphasizing the importance of contextual appropriateness in language use. And Deda (2013) similarly describes it as the ability to understand, construct, and communicate meanings that are accurate and appropriate to the social and cultural contexts in which communication takes place. According to LoCastro (2003), it is described as the ability to judge and interpret the meaning created by the interlocutors via their joint acts. The acts contain not only linguistic signals but also non-linguistic ones within socioculturally performed activities. Therefore, the development of indirect refusal competence can serve as an indicator of learners’ pragmatic development.
Moreover, because indirect refusals rely heavily on implied meanings and contextual information, learners need opportunities to observe how such expressions are used in authentic communicative situations. Exposure to realistic contexts and authentic language use may help learners understand not only the literal meanings of expressions but also their interpersonal and pragmatic functions. Hence, instructional approaches that provide contextualized input and opportunities for pragmatic interpretation may play an important role in the development of learners’ pragmatic competence.
The following sections address two crucial instructional resources, films and corpora, which are essential for the development of EFL learners’ pragmatic competence.
2. Films as Contextualized Input for L2 Pragmatics
Hymes (1972) conceptualizes language as a socially grounded and contextualized activity. Similarly, Auer (2009) argues that language learning is not merely a matter of memorizing fixed linguistic rules, but rather a process of learning how contexts are brought about through interaction. In other words, he emphasizes the development of contextualizing skills through which meanings are co-constructed in social interaction. Within this framework, the role of providing contexts for learners cannot be overlooked.
Learners are exposed to three major sources of input: teachers, instructional materials, and other learners (LoCastro, 2003). Among these, instructional materials encompass a wide range of resources, including traditional textbooks, computer-assisted learning, and multimodal sources such as videos, films, and television programs. In particular, films have been recognized as valuable resources for L2 pragmatic instruction because they provide authentic and contextualized language that is often absent from textbook dialogues (Abrams, 2014; Heidari et al., 2020). Unlike the often artificial speech acts presented in textbooks, language in films is embedded in meaningful social contexts and demonstrates a wide variety of linguistic formulae, allowing learners to observe authentic and interactional language use (Grant & Starks, 2001). Moreover, films enable learners to analyze language at the discourse level while interpreting speakers’ intentions through multimodal cues such as facial expressions, gestures, intonation, and body language, and to understand the situational appropriateness or inappropriateness of particular pragmatic features (Usó-Juan & Martínez-Flor, 2021). These features make films an ideal medium for teaching pragmatic strategies because they highlight pragmatic features through rich, contextualized conversations (Cohen, 2005). In addition, films are engaging and motivating, as they present real-life information and depict characters with whom learners can readily identify and empathize (Abrams, 2016; Washburn, 2001).
A growing body of empirical research has demonstrated the effectiveness of multimodal resources, including films, television series, and sitcoms, in L2 pragmatic instruction. For example, Abrams (2014, 2016) found that film-based instruction helped learners develop greater interactional competence by enabling them to interpret pragmatic meanings in relation to interpersonal relationships and situational contexts. Similarly, studies using films, sitcoms, and video-based metapragmatic instruction have consistently reported improvements in learners’ performance of various speech acts as well as their pragmalinguistic and sociopragmatic competence (Alerwi & Alzahrani, 2020; Alsmari, 2020; Hashemian et al., 2016; Omar & Razı, 2022). More recently, Dayani et al. (2024) found that film-based instruction significantly enhanced nursing students’ motivation to perform speech acts such as requests and refusals by providing more authentic and engaging communicative contexts than traditional textbook-based instruction. Likewise, Babayev (2026) demonstrated that films served as an effective and engaging instructional resource, promoting not only language acquisition but also learners’ cultural understanding and awareness more effectively than conventional textbook-oriented approaches.
Nevertheless, films as instructional materials are not without limitations. L2 learners may find it difficult to determine whether expressions encountered in particular scenes are specific to those contexts or are widely used across different communicative situations. Therefore, learners first need to understand the authentic contexts in which particular expressions are used through films and then inductively verify their broader usage patterns through corpora. Such an integrative approach can compensate for the limitations of each instructional resource while maximizing their complementary strengths.
3. Utilizing Corpora in L2 Pragmatic Instruction
Sinclair (1991) defined a corpus as “a collection of naturally occurring language text, chosen to characterize a state or variety of a language” (p. 171). Similarly, Kübler and Zinsmeister (2015) defined a corpus as an electronically accessible collection of texts or spoken transcripts sampled to represent a particular language or linguistic domain. Corpora include various forms of language data, such as everyday conversations, institutional discourse, academic interactions, interviews, service encounters, and even television programs and films through subtitles or transcripts. Because corpora consist of naturally occurring language produced in authentic communicative situations, they provide valuable resources for language learning and instruction.
Corpus linguistics, as a research methodology, investigates language use through the systematic analysis of authentic language data (Biber et al., 1998). By examining recurrent patterns and contextualized examples, corpus-based approaches allow learners to observe how language is actually used rather than relying solely on invented textbook examples. In particular, corpora have considerable potential for pragmatic instruction because they preserve the contexts in which language is produced and provide multiple realizations of language use (Bardovi-Harlig & Mossman, 2022). As Staples and Fernández (2019) argue, corpora are particularly suitable for studying pragmatics because pragmatic meaning is inherently dependent on contextual information.
Despite the growing importance of corpus linguistics in second language acquisition research, relatively little attention has been paid to L2 pragmatics (Rabadán Gómez, 2021; Staples & Fernández, 2019). This limitation partly stems from the relatively small number of spoken corpora and the methodological challenges involved in identifying pragmatic phenomena such as speech acts within corpus data (Rühlemann & Aijmer, 2015). Consequently, corpus-based language learning has predominantly focused on vocabulary and grammar instruction (Boulton & Cobb, 2017), and, particularly in higher education, has been widely applied to vocabulary and academic writing instruction (Li et al., 2025), while its application to pragmatic instruction remains comparatively limited. However, the use of corpora for pragmatics is meaningful in that it compensates for learners’ limited exposure to the target language through authentic and contextualized discourse (Romero-Trillo, 2018).
Nevertheless, several studies have demonstrated the pedagogical potential of corpora in L2 pragmatics, which is particularly important in EFL contexts where learners have limited opportunities to encounter and produce authentic target language use (Yilmaz & Koban Koç, 2020). For instance, Caprario et al. (2022) showed that corpus-based analysis helped learners understand the pragmatic functions of the discourse marker I mean beyond its literal meaning. Likewise, Bardovi-Harlig et al. (2015, 2017) and Furniss (2016) reported that corpus-informed instruction effectively enhanced learners’ comprehension and production of pragmatic routines while promoting pragmatic awareness and learner autonomy. In addition, Alsuhaibani (2022) found that both corpus-based learning and consciousness-raising instruction significantly improved learners’ pragmatic competence in compliment responses, although no significant difference was found between the two instructional approaches. Similarly, Bouzekria et al. (2023) demonstrated that corpus-based instruction effectively facilitated Algerian EFL learners’ acquisition of pragmatic routines, suggesting that such routines can be successfully developed in classroom settings without requiring study-abroad experiences.
These findings suggest that corpora can serve as valuable instructional resources for developing pragmatic competence. However, indirect refusals are highly context-dependent speech acts, making it difficult for learners to identify their pragmatic functions through corpus searches alone. In contrast, films provide rich contextual cues, such as interpersonal relationships, situational contexts, and nonverbal behaviors, but offer only a limited range of language examples. Corpora, on the other hand, expose learners to diverse authentic language use but may not always provide sufficient contextual information for pragmatic interpretation. Therefore, combining film-based contextual input with corpus exploration may compensate for the limitations of each approach by helping learners connect linguistic forms with their pragmatic functions across contexts. Building on this rationale, the present study integrates films and corpus exploration to facilitate learners’ understanding of indirect refusal strategies and pragmatic competence.
III. METHOD
1. Participants
The participants in this study were four undergraduate students who did not major in English, with a mean age of 24.75 years. They were selected because they represented EFL learners with limited exposure to authentic English use and pragmatic instruction. None had overseas experience and reported limited engagement with English learning, as well as relatively low perceived needs and interest in learning English. They also had no prior experience with corpus use or pragmatic instruction. In addition, all participants except for P1 reported low-intermediate English proficiency, which was generally consistent with their standardized English proficiency test scores (see Table 1). Given their limited exposure to authentic English and lack of prior instruction in pragmatics, particularly indirect speech acts, they were considered an appropriate group for examining how corpus-informed instruction could support the development of pragmatic competence among low-proficiency EFL learners.
2. Material
1) Written Discourse Completion Test (WDCT)
A WDCT was used to examine learners’ production of refusal speech acts. Although DCT responses may not fully reflect actual language use, they provide insights into what learners consider appropriate in given situations (Golato, 2003). Although this method has been criticized due to the lack of interactionality, it can control several contextual variables (Cyluk, 2013) and also represent prototypical responses that indicate participants’ pragmalinguistic knowledge (Kwon, 2004). Since DCTs use straightforward scenarios and usually require a single response, they can also be advantageous for lower-level learners to decrease anxiety and burden (Youn & Chaipuapae, 2022). For these reasons, the WDCT was employed to analyze changes in the frequency and appropriateness of indirect refusal strategies before and after the instruction.
2) Corpus of Contemporary American English (COCA)
COCA was used as the primary corpus resource for the instructional sessions. As one of the largest freely available corpora of American English, it provides authentic language examples and enables learners to explore patterns of language use through concordance and collocation searches (Hu, 2015). Participants used a paid COCA account to avoid limitations on the number of searches and access to available functions. Since searching with the corpus search terms was required to obtain accurate concordance lines, examine contexts clearly, and avoid search restrictions, participants were instructed on this procedure prior to the corpus activities. Specifically, they were asked to distinguish between the expressions they selected from the film script and their corresponding corpus search terms before conducting corpus searches.
To prevent excessive cognitive overload among participants, the instructor pre-screened the indirect refusal expressions appearing in The Intern (Meyers, 2015) by checking their frequencies and ensuring that each corpus search returned no more than 50 concordance lines. Four samples of the indirect refusal expressions among twelve used in this instruction and their frequency information are presented in Table 2.
3) The Film The Intern
The film The Intern was selected because it portrays workplace interactions in which indirect refusals frequently occur. The film depicts the relationship between Ben, a 70-year-old senior intern at a fashion startup company, and Jules, a young CEO in her 30s, portraying how the two characters from different generations and backgrounds learn from each other and grow through their workplace interactions. Moreover, diverse interpersonal relationships, such as those between supervisors and employees, colleagues, clients, and family members, appeared, allowing learners to observe how refusal strategies vary according to power relations, social distance, and communicative purposes. Its authentic and context-rich dialogues provide learners with opportunities to examine how indirect refusals are realized in socially appropriate ways across different situations. Conversation 1 is an example of a film dialogue from The Intern containing an indirect refusal.
Jules: So, you don’t wanna transfer?
Ben: Not really, sir, I’m terribly sorry. Excuse me.
Jules: Okay. Well, then looks like you’re stuck with me.
Ben: Great, I’m excited.
Jules: I will email you when I have something for you to do.
Ben: Or I could just stop by a few times day, check in...
Jules: I’ll email you. Also don’t feel like you have to dress up. I mean, we’re super cas here.
Ben: Well, I’m comfortable in a suit, if it is okay.
(Conversation 1, The Intern)
4) Corpus Journal
The corpus journal worksheet consisted of target expressions, corpus search results, and pragmatic analysis. Participants conducted corpus searches using the expressions learned during each session, completed the journal, and submitted it to the instructor after every session. The journal activity aimed to capture learners’ reflections on corpus exploration, their interpretations of target expressions, and the development of their pragmatic understanding of indirect refusal strategies (see Figure 1).
3. Procedure
The study was conducted over five sessions. In Session 1, participants completed a pre-test WDCT designed to assess their use of indirect refusal speech acts. Following the pre-test, explicit instruction was provided on the concept of indirect refusals, including their pragmatic functions and common realization strategies.
Sessions 2 to 4 focused on corpus-informed instruction. Each instructional session lasted approximately 60-90 minutes and centered on a single scene from The Intern, except for the fourth session, which covered two scenes. Participants first engaged in script-based discussions using selected scenes in which indirect refusal expressions were embedded in authentic workplace interactions. After analyzing each scene, participants discussed which expressions could be interpreted as indirect refusals and explained their reasoning. During this process, the instructor provided guidance and scaffolding to help learners identify the target expressions and understand their contextual meanings and pragmatic functions. In this study, scaffolding involved providing learners with hints that encouraged them to interpret the target expressions appropriately by considering the conversational context and the relationships between the characters. The instructor also fostered an environment in which participants could actively share their opinions and engage in discussions. At the end of each session, the instructor summarized the target indirect refusal expressions based on Beebe et al.’s (1990) taxonomy, after which participants selected two expressions for subsequent corpus exploration.
Participants then explored their self-selected expressions through corpus searches in COCA. They then examined authentic language examples and recurring patterns of use. As a follow-up assignment, participants completed corpus journals in which they recorded corpus examples of the target expressions and reflected on their meanings, usage patterns, and pragmatic functions based on their own interpretations (see Figure 1). During corpus exploration, participants also encountered examples that did not function as indirect refusals. Rather than excluding such cases, the corpus journal assignment encouraged learners to compare and analyze concordance lines that were similar to or contrasted with the target expressions identified in the film. The purpose of the assignment was to promote learners’ pragmatic awareness by helping them interpret language use across different contexts. After submission, the journals were used solely for data analysis, and individualized feedback was not provided in subsequent sessions.
In Session 5, participants completed a post-test WDCT to measure changes in their production of indirect refusal speech acts. Finally, semi-structured interviews were conducted to explore participants’ perceptions of the instructional activities and their learning experiences throughout the study. The information on the scenes used throughout the sessions and the flowchart of the overall instruction are presented in Figure 2 and Table 3.
4. Data Analysis
The WDCT consisted of six items in both the pre-test and post-test (see Appendix A. The WDCT responses were analyzed using two measures: appropriateness and frequency. The appropriateness of the participants’ responses was evaluated by two raters using a six-point Likert scale adapted from Taguchi’s (2013) rating scale of appropriateness. To establish inter-rater reliability, an intraclass correlation coefficient (ICC) analysis was conducted. The ICC values were .941 for the pre-test and .675 for the post-test, indicating excellent and moderate levels of agreement, respectively, based on the criteria proposed by Koo and Li (2016). When the two raters assigned identical scores, the rating was retained as the final appropriateness score. For responses with discrepant ratings, the final appropriateness score was calculated by averaging the ratings independently assigned by the two raters. In order to analyze changes in the use of indirect refusal strategies in greater detail, the frequency of linguistic expressions was also examined based on Taguchi’s (2013) coding framework, which applies Beebe et al.’s (1990) taxonomy.
To further examine individual differences in learners’ pragmatic development and learning patterns, the researcher, who also served as the instructor, conducted an in-depth analysis of each participant’s WDCT performance and corpus journal data. The two data sources were investigated together to develop individual learner profiles, focusing on how each participant interpreted and applied indirect refusal strategies. In addition, the corpus journal data were analyzed from a holistic perspective to identify broader patterns in learners’ pragmatic understanding across the instructional sessions. Because each participant selected two expressions for exploration in each session, the expressions most frequently selected across the six sets of corpus journal entries were examined to identify recurring patterns in each learner’s interpretation and understanding of pragmatic expressions.
Lastly, the interview consisted of four general questions concerning participants’ overall perceptions of the instruction, the strengths and weaknesses of corpus exploration, their understanding of indirect refusal strategies, and suggestions for improving future instruction (see Appendix B). The interview responses were examined on a question-by-question basis, with commonalities and differences across participants identified within each area. Where relevant, recurring themes that emerged from participants’ responses were also considered to capture their shared and individual perspectives. Given the small number of participants, individual responses were examined in detail to provide a more in-depth account of their experiences and perceptions.
IV. RESULT
1. Quantitative Overview of Learners’ Pragmatic Development
The comparison of appropriateness rating scores between two tests and overall descriptive data are shown in Table 4.
First, there was a noticeable decrease in the use of direct refusal strategies. In the pre-test, all participants frequently employed direct refusal expressions such as “I can’t”, “It’s impossible”, and “This is not fair”, resulting in a total of nine occurrences. In the post-test, however, the frequency decreased to four occurrences, indicating a reduction of more than half.
Second, there was a balanced increase in the use of various mitigating strategies, including wish, empathy, and gratitude expressions. The most noticeable change was observed in the use of the wish strategy (e.g., “I wish I could”, “I’d love to but”), which increased from one to eight occurrences. Furthermore, expressions of empathy and gratitude also appeared more frequently in the post-test. This finding suggests that the participants became less dependent on a single refusal pattern centered on statements of regret (e.g., “I’m sorry”) and began to employ a wider range of face-saving strategies.
Third, the quality of excuse and reason strategies was strengthened. All participants provided excuses or reasons in every item of both tests, resulting in the maximum frequency of 24 occurrences. Compared with the relatively simple reasons found in the pre-test, participants produced more contextually appropriate and elaborated explanations, such as wedding anniversary and lack of information. These findings suggest that learners became more capable of constructing plausible and situation-specific reasons to support their refusals. Comparison of refusal strategy frequencies between two tests is illustrated in Table 5.
As illustrated in Tables 4 and 5, a descriptive analysis of the WDCT responses reveals positive qualitative and quantitative shifts in the participants’ pragmatic production. Although the sample size limits generalizable statistical inferences, individual performance indices present distinct trajectories of development—most notably, a sharp decline in direct rejections alongside a diversified deployment of face-saving mitigators such as Wish and Statement of Alternative strategies (see Appendix C).
2. Individual Profile Analysis
1) Participant 1 (P1)
P1 showed relatively impersonal and firm refusal strategies in the pre-test, but added expressions demonstrating consideration for the hearer’s feelings in the post-test. Moreover, whereas he directly used expressions such as “impossible to accept” in the pre-test, he began to present positive opinions first, such as “I really want to participate,” to mitigate the force of refusal. He also employed more indirect strategies, including an indefinite reply (“I have no idea how I can tell you…”), which implicitly conveyed the difficulty of refusing rather than expressing rejection directly.
According to the corpus journal data, P1 primarily focused on grammatical and structural aspects, noticing how subtle grammatical changes, such as tense shifts, contribute to politeness and interpersonal distancing. He recognized that small grammatical modifications can make refusals sound softer and less direct. For instance, when analyzing the expression “I was really looking forward to it,” he noted that the corpus examples did not explicitly convey an indirect refusal. Instead, he inferred that the pragmatic effect resulted from the use of the past tense, which created psychological distance from the present situation. Likewise, when examining “If it is okay,” he focused on how its meaning varied according to sentence structure, distinguishing its use in as if constructions from its function as a simple conditional expression.
2) Participant 2 (P2)
P2 showed the most dramatic change, moving away from relatively aggressive and direct refusal strategies toward a wider range of indirect refusal expressions that considered the hearer’s face. In the pre-test, he frequently used blaming reasons and direct refusals such as “unfair, go back” and “I can’t believe your attitude.” In contrast, his posttest responses demonstrated a much softer attitude through the use of positive opinions, mitigated expressions, and wish strategies. For example, in the recommendation-letter item, he shifted from the direct response “I can’t believe your attitude, so I can’t write a letter” to a more sophisticated combination of strategies: “Oh good idea. I wish I could but I have busy schedule can we talk later?” He also frequently added promises of future acceptance, such as “I’ll do it next time,” in an effort to minimize potential relational damage.
According to his corpus journal data, P2 focused particularly on tone control and conversational management, interpreting expressions in terms of how they regulate interactions and maintain interpersonal relationships. For instance, he considered “I hope you don’t mind” useful for politely closing conversations, while interpreting “Can we talk later?” as a strategy for managing conversation flow rather than simply postponing it. He also viewed “I’m comfortable in a suit” as an expression that enables speakers to disagree while respecting the interlocutor’s opinion, thereby softening the overall tone of interaction. These findings suggest that P2 primarily interpreted pragmatic expressions through their interpersonal and conversational functions.
3) Participant 3 (P3)
P3 produced more sophisticated refusal discourse in the post-test by incorporating promises of future acceptance and selecting more formal and contextually appropriate expressions, building on her existing tendency to express empathy and provide alternatives. Compared with the pre-test, her alternatives became more practical and situation-specific. For example, she replaced the relatively simple suggestion, “Buy coffee at a 24-hour self-service café,” with the more appropriate response, “How about asking another professor?” when refusing to accept a student as an academic advisee. She also began to employ promises of future acceptance, such as “I’ll make it for you next time,” and expressed refusal in a clearer and more formal manner by replacing lengthy explanations with concise justifications, such as “for the sake of fairness.”
According to her corpus journal data, P3 demonstrated an empathy-oriented analytical pattern, focusing on the emotional impact of expressions and their function in comforting or reassuring the hearer. For example, she recognized that “I am terribly sorry” can be used not only to express regret in refusal situations but also to comfort people experiencing difficult circumstances. Likewise, she interpreted “It is all good” as an expression that reassures the hearer and facilitates a smooth closing of conversations. These findings suggest that P3 primarily interpreted pragmatic expressions through their emotional and interpersonal functions.
4) Participant 4 (P4)
P4 demonstrated more sociolinguistically mature and considerate refusal discourse in the post-test by moving beyond simple apologies to acknowledging the interlocutor’s concern through gratitude, empathy, and mitigating strategies. Although he already employed empathic expressions such as “I’m sorry to hear that” in the pre-test, his post-test responses became more elaborate by combining them with wish strategies and other supportive moves. For example, in Item 4, he mitigated the refusal by first expressing a wish, saying, “I wish I could participate...” Likewise, in Item 5, he shifted from the direct response “But we can’t negotiate” to a more considerate refusal by first acknowledging the interlocutor’s goodwill, saying, “I’m glad you’re concerning...” before declining the suggestion. These findings indicate that P4 increasingly attempted to protect the interlocutor’s positive face while performing refusals.
According to his corpus journal data, P4 exhibited a context-oriented analytical pattern by examining how identical expressions convey different pragmatic meanings depending on the situation and the relationship between speakers. For example, he explained that “Can we talk later?” may function either as a conventional expression of refusal or as a genuine promise to continue the conversation, depending on the contextual cues. Likewise, he interpreted “I have to work late” as an indirect refusal in one context but as simple information sharing with family members in another. Overall, P4 consistently demonstrated a strong awareness that pragmatic meanings are shaped by context rather than by linguistic forms alone.
3. Overall Patterns of Corpus Journal Analysis
After reviewing each participant’s corpus journal entries, expressions that were frequently selected across participants were identified. The researcher then synthesized pragmatic insights by comparing and analyzing participants’ interpretations of the same expressions. Accordingly, Table 6 summarizes the recurring patterns identified in the corpus journal data, rather than presenting the verbatim responses of individual participants.
4. Learners’ Perception of Corpus-Informed Pragmatic Instruction
1) Overall Perceptions of the Instruction
Regarding the effectiveness of film-based instruction, all participants perceived that the use of film helped them understand indirect refusal expressions by providing contextualized and multimodal input. Unlike textbook-based learning, film allowed learners to interpret expressions together with situational information, including characters’ emotions and relationships. In particular, P2 stated “By watching the video itself, I could understand the situation through various sources of information (e.g., visual and auditory cues), which helped me better perceive the differences among refusal expressions.” Similarly, P1 mentioned that learning expressions along with the contexts and emotions presented in the film helped him retain expressions and recall them more easily in daily life. P3 and P4 also emphasized that learning expressions in realistic situations was more meaningful than memorizing isolated forms.
2) Learners’ Perceptions of Corpus Exploration: Benefits and Challenges
All participants generally perceived that exploring authentic language examples and their uses in various contexts through the corpus was an interesting experience. In particular, they reported that comparing examples used in similar and different contexts and understanding the broader contexts of certain expressions were informative. P1 mentioned the novelty of discovering that similar expressions could be used differently depending on the context. Additionally, P3 and P4 evaluated that they could understand diverse uses and contexts through corpus searches.
On the contrary, they also experienced initial difficulties and technical limitations. P2 and P3 mentioned that it was challenging to determine whether the retrieved examples matched the intended meanings and contexts within the massive data, as well as to become familiar with the corpus system at the beginning. Moreover, P1 and P4 pointed out limitations of the corpus platform itself, including the cost, inconvenient interface, and recurring system errors. These responses indicate that while participants recognized the usefulness of corpus tools for expanding their understanding of authentic language use, appropriate support and a more user-friendly environment are needed for more effective utilization.
3) Learners’ Understanding and Awareness of Indirect Refusals
Regarding their newly gained understanding of indirect refusal speech acts and conversational strategies, all participants reported that they had become aware that refusals are not simply expressed through specific linguistic forms. Instead, they viewed refusals as communicative strategies that require consideration of the interlocutor’s feelings and situational contexts. Focusing on two participants’ cases in this aspect, P3 highlighted the differences between Korean and English refusal strategies, noting that English speakers do not necessarily rely heavily on “I’m sorry” but use various expressions, such as “I have to-”, to convey refusal depending on the context. P4 emphasized that the expression “Let’s move on” helped him realize the importance of contextual and nonverbal cues in interpreting meaning beyond dictionary definitions. These responses suggest that participants came to perceive indirect refusals as context-dependent communicative strategies rather than merely expressions to memorize.
4) Learners’ Suggestions for Future Corpus-Informed Instruction
Regarding suggestions for future corpus-informed instruction, participants generally proposed that sufficient instructional time, richer learning materials, and appropriate support for corpus use would be necessary for more effective learning. P1 mentioned that more instructional time would allow learners to further examine the sentences retrieved from the corpus. He explained that “it was somewhat difficult to fully understand the atmosphere and situation in which the expressions were used only based on the paragraphs provided in the corpus.” Similarly, P4 expressed the need for more time to explore a wider range of refusal expressions that learners could use in various situations. P2 emphasized the importance of continuously updated corpus data and learning materials containing diverse expressions. These responses suggest that participants actively reflected on their learning experiences and proposed ways to improve the instructional design and learning environment.
V. DISCUSSION
The present study aimed to investigate whether corpus-informed instruction with multimodal input contributed not only to the improvement of learners’ pragmatic competence but also to their perceptions regarding indirect refusal strategies. It is notable that the overall appropriateness scores improved in the WDCT findings. Unlike the other participants, P1 exhibited little change in the appropriateness score. This finding may be explained by P1’s relatively higher English proficiency, which may have enabled the participant to demonstrate a relatively high level of pragmatic appropriateness in the pre-test. Consequently, the limited gain observed in the post-test may reflect a ceiling effect rather than a lack of instructional impact.
However, a more substantial and meaningful achievement lies in the changes observed in participants’ frequency of using refusal strategies. The use of direct refusal strategies noticeably decreased, while excuse and reason strategies became more sophisticated. In addition, a wider range of indirect refusal expressions appropriate for specific contexts emerged more frequently in the post-test. These findings suggest a possible contribution not only to improving learners’ pragmatic performance but also to raising their pragmatic awareness of indirect refusal speech acts. Moreover, the learners appeared to develop pragmatic appropriateness by adjusting the degree of refusal according to situations and interlocutors and by managing conversations more strategically. This finding is consistent with previous studies and meta analyses reporting the positive effects of pragmatic instruction on learners’ speech act performance and pragmatic competence (Jeon & Kaya, 2006; Lingli & Wannaruk, 2010; Plonsky & Zhuang, 2019; Taguchi, 2015; Takahashi, 2001). Given the learner autonomy and inductive nature of corpus exploration, the present finding supports Wu’s (2025) study, which demonstrated that task-choice autonomy positively influenced Chinese EFL learners’ performance in refusal speech acts. However, the present study further suggests that corpus-informed instruction combined with multimodal input can promote learners’ contextual awareness and performance of indirect refusals.
The most important finding from the interview data is the change in participants’ perceptions of language. In other words, they recognized the importance of considering social factors, including contexts, interlocutors, and interpersonal relationships, and of employing mitigating expressions beyond their literal meanings. They also learned how expressions whose dictionary meanings differ substantially from their pragmatic meanings can convey specific intentions in refusal contexts. These findings suggest that the participants developed greater flexibility and context-sensitivity in their pragmatic understanding of language. Furthermore, although the participants acknowledged the cognitive burden associated with exploring large amounts of unorganized corpus data, they agreed that authentic and contextualized materials, such as COCA and film, facilitated their pragmatic learning. This finding is consistent with previous studies that have emphasized the value of authentic input and corpus-based instruction in developing pragmatic competence (Bardovi-Harlig, Mossman, & Su, 2017; Bardovi-Harlig, Mossman, & Vellenga, 2015; Bardovi-Harlig & Vellenga, 2012; Furniss, 2016; Yilmaz & Koban Koç, 2020). Therefore, careful instructional design and appropriate scaffolding are necessary to help learners effectively utilize large amounts of raw corpus data.
The corpus journals provided insights not only into what learners learned but also into how they developed pragmatic knowledge. Although all participants were exposed to the same instructional materials, they attended to different aspects of pragmatic knowledge during corpus exploration. P1 focused on grammatical forms, P2 on conversational management, P3 on emotional mitigation, and P4 on contextual distinctions. These findings reveal that the open and exploratory nature of corpus-informed instruction allows learners to construct pragmatic knowledge according to their own interpretive orientations and preferences. In other words, pragmatic competence did not develop uniformly across participants, but rather emerged through different areas of attention and interpretation. Hence, the participants’ corpus journals demonstrate that learner autonomy, together with appropriate instructional scaffolding, can facilitate individualized pathways of pragmatic development. As Taguchi (2026) argues in her recent paper, this finding supports the view that pragmatic competence is not a fixed and predetermined construct, but reflects learners’ agency and active meaning-making in interaction (LoCastro, 2003).
VI. CONCLUSION
The present study revealed that corpus-informed instruction could not only contribute to low-intermediate EFL learners’ pragmatic competence but also influence their perceptions of indirect refusal speech acts.
However, this study has several limitations. First, the small sample size limits the generalizability of the findings. Second, the relatively short instructional period may have constrained the extent of learners’ development. Longitudinal instructional designs are needed to investigate more sustained and sophisticated learning effects. Third, the present study focused exclusively on indirect refusal speech acts. Future studies should examine the effects of corpus-informed instruction on other speech acts and pragmatic features.
Despite several limitations, this study offers several pedagogical implications. First, corpora can serve as effective instructional materials by providing learners with large amounts of authentic and naturally occurring language data. Through inductive exploration of corpus data, learners are exposed to a variety of linguistic exemplars and patterns, which may promote learner autonomy as well as pragmatic understanding. Second, combining corpus exploration with multimodal input, such as films and videos, may facilitate L2 pragmatic learning by providing learners with realistic and contextualized situations. Therefore, teachers and educators should consider the complementary effects of these materials and utilize them according to learners’ proficiency levels and learning preferences. Finally, the findings suggest that appropriate teacher scaffolding plays an essential role in helping low-level proficiency learners effectively interpret and utilize authentic corpus data.
In conclusion, future research could employ larger samples and various research designs to investigate individual differences in pragmatic development. Since the corpus journals revealed distinct learning patterns among participants, further studies may explore how learner characteristics, preferences, and autonomy influence the development of pragmatic competence in corpus-based instruction. In addition, future studies may investigate the effects of a more learner-directed Data-Driven Learning approach on the development of pragmatic competence. Such research would provide further insights into the role of learner autonomy in corpus-based pragmatic instruction.


PDF Links
PubReader
ePub Link
Full text via DOI
Download Citation
Print






