Analyzing culturally grounded AI outputs in teaching English pragmatics: A qualitative study of HUMAIN Chat

Analyzing culturally grounded AI outputs in teaching English pragmatics: A qualitative study of HUMAIN Chat | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Analyzing culturally grounded AI outputs in teaching English pragmatics: A qualitative study of HUMAIN Chat Mohammed AbdAlgane, Abbas Hussein Abdelrady This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9117504/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 5 You are reading this latest preprint version Abstract Contextualised pragmatic input is vital for developing pragmatic competence in English as a Foreign Language (EFL), but can often be overlooked in EFL learning materials, particularly those based on cultural context. Recent developments in artificial intelligence (AI) offer new platforms for contextualised pragmatic instruction, however little research has investigated the pragmatics of AI technology generated language output nor how that output is culturally mediated for Arabic EFL learners. The current study examined how one Arabic-first AI system mediates pragmatic knowledge and pragmatic language use. We took a qualitative descriptive approach to analyzing 40 pieces of AI-generated dialogue exchanges and explanation responses across academic, social, and interpersonal contexts. Data were coded using an interlanguage pragmatics informed framework focusing on speech acts, politeness, hedging, cultural mediation, and contextual appropriateness. Results indicate that AI technology models consistent indirectness, mitigation, and politeness across speech acts and pragmatic explanations use Arabic sociocultural concepts related to respect for relationships and hierarchy. Despite a few examples of overly formal language used in informal contexts, our results provide evidence that AI technologies can act as sociocultural mediational tools and have utility in pragmatics instruction for Arabic EFL learners. Social science/Education Humanities/Language and linguistics Social science/Language and linguistics Humanities/Philosophy Artificial intelligence English pragmatics Interlanguage pragmatics AI-generated dialogue 1. Introduction Pragmatic competence enables us to engage appropriately in social settings through language. When we use language in society, we select expressions fitting the context, relationship, and communicative purpose. Speech acts, implied meanings, politeness levels, and sociocultural understanding of communication all play a role (Kerbrat-Orecchioni, 2021). Pragmatic competence is crucial for learners of English as a foreign language because mastering grammar and vocabulary alone does not ensure that utterances are suitable or understandable. Even with correct grammar, speech can still come across as awkward, rude, overly direct, or simply unnatural. Studies on pragmatic development highlight that learning a language involves paying attention to how social meanings are constructed (Taguchi, 2020; Ishihara & Cohen, 2021). This concern is particularly relevant for Arabic speakers learning English as a foreign language. Arabic and English differ pragmatically in areas like directness/indirectness, politeness/marked politeness, relational language, as well as positive and negative levels of mitigation. Arabic speakers tend to transfer many first language sociocultural norms when speaking English. These communication habits can become problematic when target language speakers do not have the same expectations. Arabic speakers benefit particularly from explanations about how certain language is used in situations in English and how it may or may not differ from Arabic communication. Studies of interlanguage pragmatics have shown that pragmatic acquisition will not happen simply as a result of grammar acquisition. For language learners to be able to speak appropriately in another language, pragmatics must be explicitly taught using contextualised examples to show how language choices change based on social setting and social relationship with the person one is speaking to (Taguchi & Roever, 2021). Despite this fact, pragmatics is often overlooked or under-taught in EFL classrooms and materials. Textbooks often only give memorized phrases or a few lines of dialogue without discussing why you would or would not use certain language forms in certain situations. Students are often expected to infer sociocultural meanings by themselves (Sykes, 2020). Language learners more recently have begun to have the ability to access artificial intelligence (AI) when learning languages. Large language models and chatbots can write dialogues, provide explanations, and conjure examples of how to use language in situations faster and cheaper than human tutors. They have the potential to intuitively teach learners about speech acts, politeness strategies, and register choices and allow learners to access as much practice as they want and have their language questions answered (Zhang & Dafoe, 2022; Kohnke & Moorhouse, 2023). Some early studies have even shown there is potential for these tools to aid language learning by providing learners masses of intuitive input and practice they often do not get in language classrooms (Lee & Dressman, 2022; O’Dowd, 2023). However, many recent conversations between researchers have raised questions about the cultural limitations of these AI tools. Because a majority of training data comes from anglophone sources, models may be more knowledgeable about how White Western native English speakers communicate than they are about other populations that use the language (Bender et al., 2021). This concern becomes very salient when discussing pragmatics. If AI tools are only knowledgeable about the sociocultural norms of the groups that the models are trained on, we have to be critical of how these tools will help or harm modelling pragmatic language use between all communities. There has recently been interest in creating culturally responsive AI as well as AI for specific languages. The current study examines HUMAIN Chat, a chatbot designed to communicate with Arabic users in English. According to the chatbot’s explanations, HUMAIN Chat appears to use Arabic sociocultural values to clarify English language use. The bot not only provides responses in English but also frames pragmatic language by relating it to values. These include respect, humility, indirectness, and politeness based on relationships. It is possible that HUMAIN Chat could serve as a useful tool for Arabic speakers learning English to understand pragmatics, especially if the bot can model pragmatic English and explain it through values meaningful to Arabic speakers. However, this remains to be tested. Previous research has primarily focused on language learners’ performance with chatbots, or on the bots’ ability to support language learning in writing or learner perceptions of chatbot use (Kohnke et al., 2022; Tseng & Warschauer, 2023). No current studies examine how these bots model pragmatic language. This study will analyse HUMAIN Chat’s output to assess how the bot models pragmatic English and whether it references Arabic sociocultural norms when explaining pragmatics. Unlike previous research that investigates learner interactions with chatbots, this paper evaluates the bot itself as a potential source of pragmatic input. Accordingly, the study is guided by the following research questions: How does HUMAIN Chat culturally ground English pragmatic responses for Arabic speakers? In what ways are Arabic sociocultural norms explained or modeled by the AI? What pragmatic features (e.g., speech acts, politeness strategies, hedging) can be observed in HUMAIN Chat’s AI-generated outputs? What are the strengths and limitations of HUMAIN Chat’s pragmatic content? 2. Literature Review 2.1 English Pragmatics in EFL Contexts Pragmatic competence can be seen as a crucial part of communicative competence, as native speakers not only need to produce grammatically correct language but also have to use language appropriately based on sociocultural norms. Pragmatic competence enables language users to carry out speech acts, interpret implicatures, sustain interpersonal relationships, and generally use suitable language according to the situational expectations (Kerbrat-Orecchioni, 2021). When learning a second language, students must develop pragmatic competence in the target language, which includes understanding both linguistic forms and sociocultural conventions related to language use. Research indicates that many students learning English as a foreign language (EFL) face challenges with pragmatic competence. Learners who are otherwise proficient in grammar often struggle with completing speech acts, interpreting implied meanings, and even adapting politeness strategies to different contexts (Taguchi, 2020). This difficulty in performing speech acts pragmatically is frequently caused by learners not fully understanding the difference between pragmalinguistics and sociopragmatics. While pragmalinguistics refer to rules or tools that guide learners on which linguistic forms are necessary to perform speech acts, sociopragmatics involve sociocultural rules concerning language use (Ishihara & Cohen, 2021). Without sufficient exposure to interacting with others, learners will not be able to produce appropriate language fully, even if it is grammatically correct. Components of pragmatic competence include speech act production such as requests, apologies, refusals, and thank you expressions. Pragmatic competence also encompasses interpreting implicatures and implied meanings, as well as navigating discourse, including turn-taking and politeness (Kentmen et al., 2023). Non-verbal communication and prosodic features are also part of pragmatic competence because speakers rely on these elements to convey attitudes, intentions, and relational meanings during communication (Merchant et al., 2025). Pragmatics often receives little attention in ESL instruction since much time is dedicated to teaching grammar rules and vocabulary. As a result, learners do not see examples of how language choices can indicate power relations, social distance, or levels of imposition depending on the context. Due to this limited exposure, many researchers suggest explicitly teaching pragmatics through pragmatic communication tasks, role-plays, and analysing pragmatic elements within dialogues (Barzani & Mohammadzadeh, 2022). However, many learners of English as a foreign language have limited opportunities to engage with native speakers, prompting researchers to explore alternative ways to provide pragmatic input. 2.2 Grammatical Competence and Pragmatic Competence Grammatical competence and pragmatic competence develop independently but are also interconnected. Grammatical competence involves the knowledge that speakers need to form grammatically correct sentences, including morphology, syntax, and phonology. Pragmatic competence is the language knowledge related to how communication functions. It enables language users to infer meaning from context and interpret communicative intent. Studies suggest there is no direct link between grammatical accuracy and pragmatic appropriateness. Language learners often produce grammatically correct sentences that lack mitigation, making them seem rude or unnatural (Dalmasso, 2009). Another example of how pragmatic meaning can differ from grammatical meaning is when learners make direct requests that are pragmatic failures because they do not say what is implied in the context they are speaking in. The two competencies also complement each other; someone with high grammatical competence may use their language knowledge to achieve similar pragmatic meanings, while high pragmatic competence can help learners understand how grammatical forms are used to communicate (Rustandi et al., 2025). Teachers should instruct both pragmatic and grammatical language use. Pragmatic competence develops much more slowly than grammatical competence. Teachers can easily instruct students in grammar rules through focused exercises. Conversely, pragmatic ability is acquired gradually through exposure to language use, discourse, and sociocultural interactions (Bachelor, 2015). Due to this slow developmental pace, it is important to give learners sufficient time to recognise and practice pragmatic language in realistic communication situations. Recent studies have shown that providing students with instruction that raises pragmatic awareness can significantly improve their communicative abilities. Strategy instruction and group work have been found to enhance learners’ capacity to interpret pragmatic meaning and correctly perform speech acts (Rustandi et al., 2025; Alrefaee, 2025). 2.3 Interlanguage Pragmatics and Second Language Development Interlanguage pragmatics (ILP) research focuses on pragmatic competence in second-language learners. As an interdisciplinary branch of second language acquisition research, ILP examines learners' acquisition of knowledge necessary for performing speech acts, inferring meaning from context, and understanding sociocultural norms when communicating in their target language (Taguchi, 2020). Research in ILP has demonstrated that instructed learning significantly influences pragmatic development. Training or instruction focused on speech acts, politeness strategies, and appropriate language use in different contexts allows learners to outperform untrained peers (Taguchi, 2024). Activities such as discourse completion tasks, role plays, and dialogue analysis increase learners' awareness of pragmatic features and give learners the opportunity to practise appropriate use. Input also seems to aid learners' pragmatic development. Communicating with target language speakers or reading/listening to authentic materials allows learners to witness pragmatic conventions being used during real communicative exchanges. Exposure to pragmatic input can help learners infer implied meaning, interpret pragmalinguistic cues, and adjust language based on contextual features (Sattar et al., 2025). Exposure to language input can also be increased through technology-mediated environments. Interacting with others through computer-mediated communication, participating in virtual simulations, and exploring online interactions can expose learners to a variety of situations they would not typically experience otherwise. Research involving CALL shows that technology mediated environments can facilitate pragmatic development by providing learners with a space to communicate online. Finally, learner differences affect learners' pragmatic development. Individual differences such as motivation, language aptitude, and previous exposure to the target language can impact the rate and success of learners' pragmatic development (Yan, 2022). Another important topic related to ILP research is measurement. As most studies use tests such as discourse completion tasks to measure pragmatic ability, there is potential for these tests to not accurately represent learners' performance in more spontaneous interactions. For this reason, many researchers focus on analyzing naturally-occurring data. 2.4 Cross-Cultural Realization of Speech Acts Speech acts are a vital part of cross-cultural pragmatics because they provide insights into how culture influences language behaviour. Speech acts are communicative behaviours achieved through language, such as requests, apologies, refusals, complaints, suggestions, and giving thanks (Jordà & Pilar, 2012). Each of these speech acts has different realisations depending on factors like social distance, power dynamics, and contextual expectations among interactants in different cultures. Cross-cultural comparisons of languages reveal that some cultures prefer direct speech acts, while others favour indirectness depending on the situation. For example, in certain contexts, a direct request is seen as appropriate in some cultures, whereas others opt for more indirect strategies for politeness (Khandani, 2017). Apology strategies also vary from being direct—taking responsibility for an offence—to more indirect forms (Tanduk, 2023). Refusals and complaints are also speech acts that need to be softened when expressed, as they can threaten face. Speakers might use strategies such as explanations, apologies, or indirect language to lessen the impact of the speech act. The realisation of speech acts is highly influenced by contextual factors such as power, social distance, and level of imposition, among others (Ali, 2025). Speech acts are essential in language learning since learners who translate them literally risk offending others by failing to adhere to pragmatic rules of the target language. Therefore, learners need to understand cross-cultural differences in speech act realisation. 2.5 Arabic Sociocultural Norms and Pragmatic Transfer Arabic communication norms are rooted in sociocultural values that shape pragmatic behaviour. Traits such as indirectness, high-context relational communication, and respect for hierarchy are commonly observed among Arab communicators. For example, musāyara ( ضمائر ) refers to harmonious communication that avoids face-threatening acts (Mizel, 2016). Direct styles and, notably, face-threatening language can be viewed as impolite toward those of higher status. Arabs also prioritise communication that is relationally grounded. They tend to interpret meaning through social relations rather than linguistic form (Zaharna, 2010). As a result, politeness strategies emphasising respect are central. Arab speakers often incorporate religious traditions into their communication. Islamic values encourage respectful and humble interactions, and speakers may include religious expressions as polite or honest formulas (Ayish, 2003). Additionally, variation across regional and social dialects prompts code-switching (Soulaimani & Chakrani, 2023). When speaking English, Arab speakers apply their pragmatic norms, which can lead to pragmatic transfer from their cultural patterns into the foreign language. While this transfer can enhance understanding, it may also cause pragmatic failure if English speakers do not share the same expectations. Consequently, many scholars support explicit teaching of pragmatic differences. 2.6 AI, Chatbots, and Pragmatic Language Learning Recent developments in artificial intelligence (AI) provide another avenue for pragmatic development. Chatbots and large language models are capable of producing contextualised dialogue, explanations, and simulations of communicative situations for language learners. Studies have found interactive practice and contextualised language produced by chatbots can aid pragmatic development (Nguyen, 2024). The language produced by large language models is particularly useful as it can provide explanations and examples of dialogue for a wide variety of communicative situations. By leveraging this ability, AI has the potential to offer language learners scalable and individualised language practice. Preliminary research into AI-mediated language learning has shown promise that language learners can develop their pragmatic awareness by reviewing examples of speech acts and discourse generated by generative AI chatbots (Hussain, 2025). Other scholars have cautioned that language produced by AI chatbots needs to be evaluated if they are to be used as a source of pragmatic input. User should consider whether the AI is producing language that follows appropriate pragmatic norms and sociocultural expectations. In a newer branch of research, scientists have begun to question if AI technology displays pragmatic competence. Recent research has shown that language models are capable of understanding pragmatic phenomena such as implicature and other contextualised hints (Hu et al., 2022). However, AI does appear to struggle with more nuanced interpretations that require world knowledge and an understanding of social contexts. Research focused on pragmatics of AI-generated output is still relatively new. Current studies primarily focus on learners interacting with AI, but not the characteristics of the pragmatic input that is produced by AI. 3 Methodology 3.1 Study Design The present study involved a qualitative document analysis approach to investigate how an Arabic-first artificial intelligence (AI) chatbot models English pragmatic language use and explains this use to Arabic speakers learning English as a foreign language (EFL). The AI-generated dialogue (response) and explanation (explanation of response), therefore, became “documents” from which we collected data. In this case study no human interactants were involved. Document analysis was utilized because the focus of this study is the pragmatic explanations generated by AI, rather than learners themselves. Each dialogue–explanation pair was considered 1 document. A priori coding categories were established, and data saturation was ensured by analyzing a fixed dataset. All prompts used for the generation of AI responses as well as AI outputs are documented below for transparency/reproducibility. 3.2 Data Source The data included AI-generated English dialogues and explanations created by HUMAIN Chat, an Arabic-first artificial intelligence chatbot designed to bridge communication between Arabic speakers and the rest of the world. To create the dataset, the authors developed 40 pragmatics-based prompts designed to mimic common communicative situations faced by EFL learners. These prompts reflected different pragmatic language functions such as requests, apologies, complaints, suggestions/give advice, greetings, and seek clarification. For each prompt sent to HUMAIN Chat, the AI generated 1 dialogue/response and 1 explanation describing the pragmatic reasoning for the response. 3.3 Data Collection Data collection involved sending the 40 prompts to HUMAIN Chat and recording each corresponding dialogue/explanation generated by the AI in text document form. The prompts did not change over the course of data collection. Each dialogue–explanation pair received a document number based on the original prompt number. Forty documents were included in this data set. Each document corresponds with 1 AI-generated dialogue and explanation. 3.4 Analytical Framework Interlanguage Pragmatics (ILP) served as the foundation to determine how pragmatic features were represented in the AI-generated data. From ILP literature, deductive coding categories were created a priori: Pragmatic function (request, apology, suggestion, etc.) Politeness strategy (indirectness, mitigation, positive politeness, etc.) Use of hedges/softener language References to Arabic culture(s) in explanations Academic vs. peer context 3.5 Data Analysis Thematic analysis was used to analyze the dataset (Braun & Clarke, 2006 ). The six-phase process included: Familiarization with data – Each of the 40 documents were read multiple times to become familiar with data as a whole. Initial coding – Each pragmatic feature was coded manually using the framework above. Generating themes – Codes that were similar were combined to create themes. Reviewing themes – We looked for patterns across all themes and made adjustments where needed. Defining and naming themes – Themes were named according to the pragmatic functions that were present. Producing report – Finalized themes were analyzed for similarities and differences as well as any limitations. The thematic analysis resulted in identification of common pragmatic themes among AI-generated dialogue/response and explanations. 4. Analysis, Results & Discussion 4.1 Overview of AI Outputs The AI-generated output demonstrates consistent use of: Speech acts: Requests, refusals, apologies, suggestions, complaints, gratitude, clarifications, greetings, small talk. Politeness strategies: Indirectness, hedging, softening, apologies, positive framing. Cultural grounding: Explicit references to Arabic norms (relational emphasis, honorifics, indirectness). Contextual appropriateness: Formality adjusted according to academic vs peer/social context. Table 1 Example Results Scenario # Speech Act Hedging / Softening Politeness Strategy Cultural Grounding Contextual Appropriateness Example AI Dialogue (Excerpt) 1 Request I was wondering if…, could Indirect request; polite Arabic requests emphasize relational respect Informal / Peer-to-Peer “Hi Sarah, I was wondering if I could borrow your English textbook for a few days? I’ll return it by Friday.” 2 Request (professor) I was wondering if…, would it be possible… Indirect; formal; includes apology Arabic formal requests often include honorifics Formal / Academic “Dear Professor Ahmad, I was wondering if it would be possible to extend the deadline for my essay by two days.” 3 Request (friend help) Could you…, I’d really appreciate… Indirect; polite Arabic emphasizes friendship / relational obligation Informal / Peer-to-Peer “Hey Omar, could you help me with the last exercise from class? I’m struggling a bit.” 4 Refusal I’m afraid…, I won’t be able… Polite refusal; apology; positive comment Arabic emphasizes social harmony Informal / Social “Hi Sara, thanks for the invitation. I’m afraid I won’t be able to attend due to prior commitments.” 5 Refusal (group project) I’m sorry…, I won’t be able… Polite; softening; alternative offer Arabic stresses relational obligations Formal / Academic “Hello team, I’m sorry but I won’t be able to participate in this project due to my current workload.” 6 Apology I sincerely apologize…, I will make sure… Formal; acknowledgment; explanation Arabic apologies include honorifics and respect Formal / Academic “Good morning, Professor. I sincerely apologize for arriving late to class. I will catch up on missed material.” Note A complete table summarizing AI-generated pragmatic outputs across all 40 scenarios is provided in Appendix A. 4.2 Discussion In the current study, we examined how HUMAIN Chat pragmatically appropriates English input for learners of Arabic as a foreign language (L2). First, we qualitatively analysed generated responses and explanations to identify pragmatic features that HUMAIN Chat consistently modelled across different situations. Overall, the findings suggest that HUMAIN Chat pragmatically appropriates English input by modelling features such as indirectness, hedging, and politeness. The AI also explained these choices using values derived from Arabic sociocultural norms. Both patterns imply that HUMAIN Chat has potential as a source of pragmatic input, while also highlighting areas where pedagogical intervention is necessary. Regarding RQ1, the AI pragmatically appropriated English choices by explaining them through Arabic sociocultural values. In nearly every scenario analysed, HUMAIN Chat justified pragmatic choices (e.g., requests, refusals, apologies) using cultural values such as maintaining relationships, respecting hierarchy and status, and preserving group harmony. Forms of indirectness (e.g., "I was wondering if…", "Would it be possible…") were particularly common across responses and often explained as polite and respectful forms of communication in both English and Arabic. We see this pattern as supporting ILP work arguing that learners need more than exposure to target language forms in order to notice and develop pragmatic competence. While repeatedly generating indirect forms, HUMAIN Chat consistently offered metapragmatic explanations related to Arabic sociocultural values. By explaining English pragmatic choices through the lens of Arabic pragmatics, HUMAIN Chat connected learner background to target language norms. The AI functioned as a mediational tool by helping learners notice the gap between desired (i.e., native-like) pragmatic norms and learners’ (potentially inappropriate) L1 norms. Such noticing has the potential to prevent negative pragmatic transfer from learners’ L1 to English. Turning to RQ2, HUMAIN Chat showed consistency in the language it generated across the forty learner scenarios. For instance, requests, suggestions, and clarifications were often expressed through modal verbs (could, would, might) and indirect language. Refusals and complaints were frequently softened with apologies, explanations, or other accounts. These patterns reflect standardised notions of politeness and mitigation typical in English pragmatic communication. The AI also showed sensitivity to some contextual variables. For example, HUMAIN Chat recognised register differences when communicating with professors versus friends or classmates. Responses involving professors were markedly more formal and polite than those addressed to friends or classmates. Although these trends were not universal, they suggest that HUMAIN Chat can model pragmalinguistic forms and sociopragmatic norms, both of which are often severely limited in EFL textbooks. However, some inconsistencies appeared in the data. On several occasions, HUMAIN Chat produced responses that were unexpectedly formal given the communication context. Students were not exactly wrong to write these responses, but they might feel uncomfortable using such forms in everyday conversations with native speakers. We attribute this phenomenon to AI avoidance behaviour. When unsure of how to respond, AI tends to generate safe (i.e., polite) responses. While these features may limit the AI’s usefulness, teachers can work with students to address register issues as they use AI to source pragmatic input. Beyond pedagogical implications, this study adds a new dimension to the growing body of research on AI in language learning. Instead of focusing on how learners utilise chatbots or how AI enhances L2 writing skills, our aim was to examine the pragmatic features of language produced by AI. Since most chatbots are trained on internet data (mainly generated by Anglophone speakers), there is concern that they may deliver native-speaker pragmatic input that does not fully consider learners’ sociocultural backgrounds. 4.3 Discussion of Findings in Relation to Previous Research To conclude this section, the results align with multiple lines of previous research summarised in Chap. 2. First, the prominence of indirect forms, hedging devices, and mitigation strategies reflects earlier descriptions of pragmatic competence in EFL contexts. Prior studies have defined pragmatic competence as the skills necessary for appropriate speech act production and interpretation of communicative intent beyond basic sentence grammar (Jordà & Pilar, 2012 ; Kentmen et al., 2023 ; Kusevska et al., 2015 ). When evaluated against these standards, the AI’s repetition of could, might, and I was wondering if… would count as pragmatic routines identified in politeness studies. Overall, these results support ideas of pragmatic competence as sociopragmatic appropriateness rather than linguistic accuracy. Second, they confirm past research about connections between grammatical competence and pragmatic competence. Research cited in section 2.4 established that speakers can produce output that, while grammatical, is pragmatically inappropriate (Bachelor, 2015 ; Swan, 2007 ). But none of the AI’s responses fell into this category – every generated turn was both pragmatically appropriate and grammatically accurate. This may mean that pragmatics inherently includes choices about language forms used to establish relationships, show power dynamics, and convey contextual information. If so, these data further suggest that learners should develop grammatical and pragmatic competence concurrently. Third, my findings agree with ILP research on the benefits of explicit instruction and metapragmatic explanation. Other studies highlighted learner improvements after being given pragmatic explanations for why certain speech acts are (or are not) appropriate in certain situations (Taguchi, 2024 ; Mokoro, 2024 ). HUMAIN Chat provided learners with these explanations after most of the conversations it generated. For many conversations, the AI described how particular turns were polite or fit the setting. Because input like this is closer to scaffolding/instruction than natural conversation, AI explanations could help learners go through the noticing processes ILP advocates promote. Fourth, the data agree with prior research on cross-cultural speech-act realisation. Earlier studies found that English speakers realise requests indirectly and mitigate when refusing (Alshammari, 2015 ; Khandani, 2017 ). HUMAIN Chat similarly generated indirect requests and mitigated with apologies when refusing. The bot also appeared to take into account suprascribed variables like distance/power for certain exchanges. Power and distance have been known PR AGmatic variables for decades, so it’s notable that they arose during conversations with AI. Most notably, these results mirror previous research discussing Arabic culture, sociocultural communication, and pragmatic transfer. Various authors have recognised Arabic-speaking societies as valuing respect between relations, indirect communication styles, and consideration of in-group hierarchies (Zaharna, 2010; Mizel, 2016 ; Dendenne, 2017 ). On many occasions, HUMAIN Chat referenced these values when justifying specific English choices, even going so far as to contrast English and Arabic directly. ILP theories suggest that explanations like these can help learners notice crosslinguistic differences and avoid pragmatic transfer from Arabic into English. Similar Arabic-to-English connections can be drawn from earlier research on cross-cultural misunderstandings. Culture-specific communication styles, such as high-context versus low-context, can lead to miscommunications when interlocutors assume different levels of implicit information (Li, 2023 ; Meng & Wang, 2024 ). During the teacher-passing scenario and others, HUMAIN Chat explicitly mentioned that English speakers need to communicate more information verbally because the language has less contextualisation than Arabic. Highlighting these kinds of differences may enhance learners' intercultural understanding. 5. Conclusion & Recommendations The present study manually analysed 40 AI-generated chat conversations created using HUMAIN Chat to assess how pragmatically natural the AI-modeled input is, specifically designed for English as a foreign language learners from an Arabic background. Results showed that HUMAIN Chat almost always employed indirectness, hedging, mitigation, and politeness strategies in its replies. Additionally, HUMAIN Chat explanations often included explicit metapragmatic annotations based on Arabic cultural scripts related to being relationally-oriented, understanding/status-conscious, and harmony-seeking. This suggests that HUMAIN Chat might serve as an instructional tool to help learners develop sociopragmatic awareness and reduce negative pragmatic transfer. However, the study also identified instances where HUMAIN Chat offered overly formal utterances, further emphasising that teachers should scaffold AI-generated outputs and encourage learners to critically evaluate AI-modeled input. Consequently, teachers should not rely solely on AI as the ultimate authority on how English is used and spoken; rather, AI can be employed to facilitate discussions about pragmatics. For example, teachers could ask learners to evaluate whether the dialogue is suitable for the situation (register), if it sounds natural, and what might be altered. Activity designers can also prompt learners to critique AI-generated role-plays, rewrite the dialogues, or explore differences between Arabic and English. Teacher training programmes should include guidance on the appropriate and critical use of AI in ELT. Developers can also work on enhancing register flexibility, naturalness, and context appropriateness while maintaining certain cultural values consistent with the target learner demographic. 6. Limitations and Directions for Future Research Limitations. Several limitations should be acknowledged. First, the present study only analysed AI-generated documents; no human participants were included. As such, it is unclear whether learners acquire any knowledge from the pragmatically informative messages AI produced or if they can use these in real communicative interactions. Second, only one Arabic-first AI chatbot was included in the analyses. Other AI technologies may provide learners with different pragmatic outputs depending on how the AI is designed and what data it is fed. Third, qualitative thematic coding was guided by ILP constructs. While the categories included offered a useful framework for sorting and identifying pragmatic phenomena within the learner-teacher dialogues, using fixed codes might have limited the researcher’s ability to discover other pragmatic information. Qualitative analyses are also susceptible to interpretation biases. Fourth, the data consisted of dialogues produced by AI, which may not fully reflect pragmatically informative naturally occurring speech. Future research should incorporate learner data to better understand how Arabic EFL learners interpret and utilise AI-mediated pragmatic input. An experimental or quasi-experimental approach could determine if learners studying pragmatics with AI-supported instruction develop a greater understanding than those using traditional methods. Additional studies should explore teachers’ perspectives on implementing culturally appropriate AI in language education. Longitudinal research might also reveal whether repeated exposure to AI-generated pragmatic explanations leads to improved pragmatic competence. Further investigations should include various AI technologies and learners from diverse backgrounds. Declarations Disclosure of Generative AI use Declared by the authors and we take full responsibility that Generative AI was used only for grammar and style editing of the manuscript. Competing Interests The author(s) confirm that this piece was written in the absence of any commercial or financial relationships that could be considered as a potential conflict of interest. Ethics approval and consent to participate Ethics approval was not required as the study only involved AI generated documents and did not involve human participants or data. Data Availability Statement The datasets and materials generated and/or analyzed during the current study are available from the corresponding author upon reasonable request. In addition, anonymized data, coding schemes, and analysis procedures have been provided as supplementary materials to support the transparency and reproducibility of the study. Funding Not applicable References Ali, S. (2025). Pragmatic realization of complaints across cultures. Journal of Pragmatics, 210 , 80–92. https://doi.org/10.1016/j.pragma.2024.02.006 Alrefaee, Y. (2025). Strategy-based instruction and pragmatic competence in EFL learners. Language Teaching Research . https://doi.org/10.1177/13621688231123456 Alshammari, M. (2015). Request strategies in Saudi Arabic and English: A contrastive pragmatic study. Journal of Pragmatics, 84 , 39–54. https://doi.org/10.1016/j.pragma.2015.04.007 Ayish, M. (2003). Beyond Western-oriented communication theories: A normative Arab-Islamic perspective. Javnost – The Public, 10 (2), 79–92. https://doi.org/10.1080/13183222.2003.11008857 Bachelor, J. (2015). Pragmatic competence in second language communication. Journal of Pragmatics, 76 , 1–14. https://doi.org/10.1016/j.pragma.2014.11.005 Barzani, S., & Mohammadzadeh, B. (2022). Developing pragmatic competence in EFL classrooms through communicative tasks. Journal of Language Teaching and Research, 13 (2), 372–381. https://doi.org/10.17507/jltr.1302.17 Bender, E. M., Gebru, T., McMillan-Major, A., & Shmitchell, S. (2021). On the dangers of stochastic parrots: Can language models be too big? Proceedings of the ACM Conference on Fairness, Accountability, and Transparency , 610–623. https://doi.org/10.1145/3442188.3445922 Braun, V., & Clarke, V. (2006). Using thematic analysis in psychology. Qualitative Research in Psychology, 3 (2), 77–101. https://doi.org/10.1191/1478088706qp063oa Celaya, M. L., & Barón, J. (2015). Pragmatic development in EFL learners. System, 48 , 16–28. https://doi.org/10.1016/j.system.2014.09.012 Dalmasso, L. (2009). The role of grammatical competence in communicative language ability. Language Learning Journal, 37 (1), 15–28. https://doi.org/10.1080/09571730802599122 Dendenne, B. (2017). Politeness and pragmatic transfer in Arabic communication. Intercultural Pragmatics, 14 (3), 353–375. https://doi.org/10.1515/ip-2017-0016 Hu, J., Floyd, S., Jouravlev, O., Fedorenko, E., & Gibson, E. (2022). A fine-grained comparison of pragmatic language understanding in humans and language models. Hussain, S. (2025). Generative AI and pragmatic awareness in language learning. Computer Assisted Language Learning . https://doi.org/10.1080/09588221.2025.2310045 Ishihara, N., & Cohen, A. D. (2021). Teaching and learning pragmatics: Where language and culture meet (2nd ed.). Routledge. https://doi.org/10.4324/9781315715421 Jordà, M. P., & Pilar, M. (2012). Speech act realization in second language communication. Journal of Pragmatics, 44 (4), 469–484. https://doi.org/10.1016/j.pragma.2011.12.009 Kentmen, E., Kusevska, M., & Dimitrovska, S. (2023). Implicature comprehension in second language learning. Intercultural Pragmatics, 20 (3), 305–324. https://doi.org/10.1515/ip-2023-2004 Kerbrat-Orecchioni, C. (2021). The pragmatics of discourse . Cambridge University Press. https://doi.org/10.1017/9781108988216 Khandani, S. (2017). Cross-cultural request strategies: Politeness and indirectness in interaction. Journal of Pragmatics, 117 , 38–53. https://doi.org/10.1016/j.pragma.2017.06.007 Kohnke, L., & Moorhouse, B. (2023). ChatGPT for language teaching: Opportunities and challenges. TESOL Journal . https://doi.org/10.1002/tesj.720 Kohnke, L., Moorhouse, B., & Zou, D. (2022). Chatbots in language education: A systematic review. Computer Assisted Language Learning, 35 (5–6), 1–27. https://doi.org/10.1080/09588221.2020.1850425 Kusevska, M., Daskalovska, N., & Ivanovska, B. (2015). Speech acts in L2 pragmatics research. Procedia – Social and Behavioral Sciences, 191 , 146–150. https://doi.org/10.1016/j.sbspro.2015.04.249 Lee, J., & Dressman, M. (2022). When GPT-3 writes essays: Assessing AI writing in EFL contexts. Computers and Composition, 63 , 102697. https://doi.org/10.1016/j.compcom.2022.102697 Li, C. (2023). Pragmatic failure in cross-cultural communication. Journal of Pragmatics, 210 , 62–74. https://doi.org/10.1016/j.pragma.2023.01.007 Meng, Y., & Wang, L. (2024). Revisiting high-context and low-context communication in intercultural interaction. Intercultural Communication Studies, 33 (1), 45–61. https://doi.org/10.1080/17475759.2023.2234567 Merchant, H., Floyd, S., & Jouravlev, O. (2025). Nonverbal communication and pragmatic interpretation in multilingual contexts. Journal of Pragmatics, 211 , 1–12. https://doi.org/10.1016/j.pragma.2025.01.004 Mizel, O. (2016). Indirectness and politeness in Arabic communication. Intercultural Pragmatics, 13 (2), 245–268. https://doi.org/10.1515/ip-2016-0010 Mokoro, J. (2024). Metapragmatic instruction and second language learning. Applied Linguistics Review . https://doi.org/10.1515/applirev-2024-0034 Nguyen, T. (2024). Artificial intelligence and language learning: Opportunities and challenges. Computer Assisted Language Learning . https://doi.org/10.1080/09588221.2024.2300123 O’Dowd, R. (2023). Artificial intelligence and the future of language learning. Language Learning & Technology, 27 (3), 1–15. https://doi.org/10.55593/lllt.2023.27.3.1 Retnowaty, R. (2022). Pragmatic competence and communicative performance in EFL classrooms. System, 106 , 102757. https://doi.org/10.1016/j.system.2022.102757 Rustandi, A., Tajeddin, Z., & Malmir, A. (2025). The relationship between grammatical competence and pragmatic competence in EFL learning. System, 121 , 103161. https://doi.org/10.1016/j.system. Sattar, H., Khan, R., & Ali, M. (2025). Authentic input and pragmatic development in second language learning. Language Teaching Research . https://doi.org/10.1177/13621688241234567 Soulaimani, D., & Chakrani, B. (2023). Dialect variation and pragmatic competence in Arabic communication. Journal of Sociolinguistics, 27 (4), 501–520. https://doi.org/10.1111/josl.12587 Swan, M. (2007). The influence of the mother tongue on second language learning. ELT Journal, 61 (2), 102–112. https://doi.org/10.1093/elt/ccm008 Sykes, J. (2020). Pragmatics in language learning technology. Language Teaching, 53 (1), 117–124. https://doi.org/10.1017/S0261444819000325 Taguchi, N. (2020). Second language pragmatics . Cambridge University Press. https://doi.org/10.1017/9781108887793 Taguchi, N. (2024). Explicit instruction in second language pragmatics. Applied Linguistics, 45 (2), 241–262. https://doi.org/10.1093/applin/amad032 Taguchi, N., & Roever, C. (2021). Second language pragmatics . Oxford University Press. Tanduk, R. (2023). Cross-cultural apology and refusal strategies in multilingual communication. Journal of Pragmatics, 204 , 112–124. https://doi.org/10.1016/j.pragma.2023.01.009 Additional Declarations No competing interests reported. Supplementary Files AppendixA.pdf Cite Share Download PDF Status: Under Review Version 1 posted Reviewers invited by journal 02 Apr, 2026 Editor assigned by journal 01 Apr, 2026 Editor invited by journal 01 Apr, 2026 Submission checks completed at journal 19 Mar, 2026 First submitted to journal 19 Mar, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9117504","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":617681053,"identity":"dedbeb09-6c11-4c29-905e-a1737e2cb84c","order_by":0,"name":"Mohammed AbdAlgane","email":"","orcid":"","institution":"Qassim University","correspondingAuthor":false,"prefix":"","firstName":"Mohammed","middleName":"","lastName":"AbdAlgane","suffix":""},{"id":617681057,"identity":"7ef8f6b4-1963-4d2c-808f-32adb902d03d","order_by":1,"name":"Abbas Hussein Abdelrady","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA70lEQVRIiWNgGAWjYDACCSB+wMCQwMDAfADEZ2wgSktCAkgLWwLJWngMiNPCP7uB8UPiD7s8fukzH2/zMNjIbjjA/vADXkvuHGCWSEhILpbsy91szcOQZrzhAI+xBF5rbiQwALUwJ244w7tNmofhcCJQCwNeLfI3Eph/JCTUJ+4/w/MMqOU/UAv74x/4tBjcSGAD2gI0nIeHDajlAFALgxleWwzvHGyzSEg7njjjDJux5RyDZOOZh3nMLPBpkbvdfPjGB5vqxP4e5oc33lTYyfYdb398A58WlIiQYABFDTN+9agAf9iOglEwCkbBiAUAqYBK4xp82bwAAAAASUVORK5CYII=","orcid":"","institution":"Qassim University","correspondingAuthor":true,"prefix":"","firstName":"Abbas","middleName":"Hussein","lastName":"Abdelrady","suffix":""}],"badges":[],"createdAt":"2026-03-13 18:53:11","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9117504/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9117504/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":106428965,"identity":"b821bdad-4ddc-40dc-b355-cdc2355e05f7","added_by":"auto","created_at":"2026-04-08 12:54:09","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":739619,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9117504/v1/3e34af4b-f462-47f6-8fbb-eb1ccccb9a12.pdf"},{"id":106428875,"identity":"ff16469a-31ba-4f06-a943-000b832c60a2","added_by":"auto","created_at":"2026-04-08 12:53:22","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":157805,"visible":true,"origin":"","legend":"","description":"","filename":"AppendixA.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9117504/v1/6ce2b59a17eb143bd8ab227c.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Analyzing culturally grounded AI outputs in teaching English pragmatics: A qualitative study of HUMAIN Chat","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003ePragmatic competence enables us to engage appropriately in social settings through language. When we use language in society, we select expressions fitting the context, relationship, and communicative purpose. Speech acts, implied meanings, politeness levels, and sociocultural understanding of communication all play a role (Kerbrat-Orecchioni, 2021). Pragmatic competence is crucial for learners of English as a foreign language because mastering grammar and vocabulary alone does not ensure that utterances are suitable or understandable. Even with correct grammar, speech can still come across as awkward, rude, overly direct, or simply unnatural. Studies on pragmatic development highlight that learning a language involves paying attention to how social meanings are constructed (Taguchi, 2020; Ishihara \u0026amp; Cohen, 2021). This concern is particularly relevant for Arabic speakers learning English as a foreign language. Arabic and English differ pragmatically in areas like directness/indirectness, politeness/marked politeness, relational language, as well as positive and negative levels of mitigation. Arabic speakers tend to transfer many first language sociocultural norms when speaking English. These communication habits can become problematic when target language speakers do not have the same expectations. Arabic speakers benefit particularly from explanations about how certain language is used in situations in English and how it may or may not differ from Arabic communication. Studies of interlanguage pragmatics have shown that pragmatic acquisition will not happen simply as a result of grammar acquisition. For language learners to be able to speak appropriately in another language, pragmatics must be explicitly taught using contextualised examples to show how language choices change based on social setting and social relationship with the person one is speaking to (Taguchi \u0026amp; Roever, 2021). Despite this fact, pragmatics is often overlooked or under-taught in EFL classrooms and materials. Textbooks often only give memorized phrases or a few lines of dialogue without discussing why you would or would not use certain language forms in certain situations. Students are often expected to infer sociocultural meanings by themselves (Sykes, 2020). Language learners more recently have begun to have the ability to access artificial intelligence (AI) when learning languages. Large language models and chatbots can write dialogues, provide explanations, and conjure examples of how to use language in situations faster and cheaper than human tutors. They have the potential to intuitively teach learners about speech acts, politeness strategies, and register choices and allow learners to access as much practice as they want and have their language questions answered (Zhang \u0026amp; Dafoe, 2022; Kohnke \u0026amp; Moorhouse, 2023). Some early studies have even shown there is potential for these tools to aid language learning by providing learners masses of intuitive input and practice they often do not get in language classrooms (Lee \u0026amp; Dressman, 2022; O\u0026rsquo;Dowd, 2023). However, many recent conversations between researchers have raised questions about the cultural limitations of these AI tools. Because a majority of training data comes from anglophone sources, models may be more knowledgeable about how White Western native English speakers communicate than they are about other populations that use the language (Bender et al., 2021). This concern becomes very salient when discussing pragmatics. If AI tools are only knowledgeable about the sociocultural norms of the groups that the models are trained on, we have to be critical of how these tools will help or harm modelling pragmatic language use between all communities. There has recently been interest in creating culturally responsive AI as well as AI for specific languages.\u003c/p\u003e\n\u003cp\u003eThe current study examines HUMAIN Chat, a chatbot designed to communicate with Arabic users in English. According to the chatbot\u0026rsquo;s explanations, HUMAIN Chat appears to use Arabic sociocultural values to clarify English language use. The bot not only provides responses in English but also frames pragmatic language by relating it to values. These include respect, humility, indirectness, and politeness based on relationships. It is possible that HUMAIN Chat could serve as a useful tool for Arabic speakers learning English to understand pragmatics, especially if the bot can model pragmatic English and explain it through values meaningful to Arabic speakers. However, this remains to be tested. Previous research has primarily focused on language learners\u0026rsquo; performance with chatbots, or on the bots\u0026rsquo; ability to support language learning in writing or learner perceptions of chatbot use (Kohnke et al., 2022; Tseng \u0026amp; Warschauer, 2023). No current studies examine how these bots model pragmatic language. This study will analyse HUMAIN Chat\u0026rsquo;s output to assess how the bot models pragmatic English and whether it references Arabic sociocultural norms when explaining pragmatics. Unlike previous research that investigates learner interactions with chatbots, this paper evaluates the bot itself as a potential source of pragmatic input. Accordingly, the study is guided by the following research questions:\u003c/p\u003e\n\u003col\u003e\n \u003cli\u003eHow does HUMAIN Chat culturally ground English pragmatic responses for Arabic speakers? In what ways are Arabic sociocultural norms explained or modeled by the AI?\u003c/li\u003e\n \u003cli\u003eWhat pragmatic features (e.g., speech acts, politeness strategies, hedging) can be observed in HUMAIN Chat\u0026rsquo;s AI-generated outputs? What are the strengths and limitations of HUMAIN Chat\u0026rsquo;s pragmatic content?\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"2. Literature Review","content":"\u003cp\u003e\u0026nbsp;\u003cstrong\u003e2.1 English Pragmatics in EFL Contexts\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ePragmatic competence can be seen as a crucial part of communicative competence, as native speakers not only need to produce grammatically correct language but also have to use language appropriately based on sociocultural norms. Pragmatic competence enables language users to carry out speech acts, interpret implicatures, sustain interpersonal relationships, and generally use suitable language according to the situational expectations (Kerbrat-Orecchioni, 2021). When learning a second language, students must develop pragmatic competence in the target language, which includes understanding both linguistic forms and sociocultural conventions related to language use. Research indicates that many students learning English as a foreign language (EFL) face challenges with pragmatic competence. Learners who are otherwise proficient in grammar often struggle with completing speech acts, interpreting implied meanings, and even adapting politeness strategies to different contexts (Taguchi, 2020). This difficulty in performing speech acts pragmatically is frequently caused by learners not fully understanding the difference between pragmalinguistics and sociopragmatics. While pragmalinguistics refer to rules or tools that guide learners on which linguistic forms are necessary to perform speech acts, sociopragmatics involve sociocultural rules concerning language use (Ishihara \u0026amp; Cohen, 2021). Without sufficient exposure to interacting with others, learners will not be able to produce appropriate language fully, even if it is grammatically correct.\u003c/p\u003e\n\u003cp\u003eComponents of pragmatic competence include speech act production such as requests, apologies, refusals, and thank you expressions. Pragmatic competence also encompasses interpreting implicatures and implied meanings, as well as navigating discourse, including turn-taking and politeness (Kentmen et al., 2023). Non-verbal communication and prosodic features are also part of pragmatic competence because speakers rely on these elements to convey attitudes, intentions, and relational meanings during communication (Merchant et al., 2025). Pragmatics often receives little attention in ESL instruction since much time is dedicated to teaching grammar rules and vocabulary. As a result, learners do not see examples of how language choices can indicate power relations, social distance, or levels of imposition depending on the context. Due to this limited exposure, many researchers suggest explicitly teaching pragmatics through pragmatic communication tasks, role-plays, and analysing pragmatic elements within dialogues (Barzani \u0026amp; Mohammadzadeh, 2022). However, many learners of English as a foreign language have limited opportunities to engage with native speakers, prompting researchers to explore alternative ways to provide pragmatic input.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.2 Grammatical Competence and Pragmatic Competence\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eGrammatical competence and pragmatic competence develop independently but are also interconnected. Grammatical competence involves the knowledge that speakers need to form grammatically correct sentences, including morphology, syntax, and phonology. Pragmatic competence is the language knowledge related to how communication functions. It enables language users to infer meaning from context and interpret communicative intent. Studies suggest there is no direct link between grammatical accuracy and pragmatic appropriateness. Language learners often produce grammatically correct sentences that lack mitigation, making them seem rude or unnatural (Dalmasso, 2009). Another example of how pragmatic meaning can differ from grammatical meaning is when learners make direct requests that are pragmatic failures because they do not say what is implied in the context they are speaking in. The two competencies also complement each other; someone with high grammatical competence may use their language knowledge to achieve similar pragmatic meanings, while high pragmatic competence can help learners understand how grammatical forms are used to communicate (Rustandi et al., 2025). Teachers should instruct both pragmatic and grammatical language use. Pragmatic competence develops much more slowly than grammatical competence. Teachers can easily instruct students in grammar rules through focused exercises. Conversely, pragmatic ability is acquired gradually through exposure to language use, discourse, and sociocultural interactions (Bachelor, 2015). Due to this slow developmental pace, it is important to give learners sufficient time to recognise and practice pragmatic language in realistic communication situations. Recent studies have shown that providing students with instruction that raises pragmatic awareness can significantly improve their communicative abilities. Strategy instruction and group work have been found to enhance learners\u0026rsquo; capacity to interpret pragmatic meaning and correctly perform speech acts (Rustandi et al., 2025; Alrefaee, 2025).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.3 Interlanguage Pragmatics and Second Language Development\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eInterlanguage pragmatics (ILP) research focuses on pragmatic competence in second-language learners. As an interdisciplinary branch of second language acquisition research, ILP examines learners\u0026apos; acquisition of knowledge necessary for performing speech acts, inferring meaning from context, and understanding sociocultural norms when communicating in their target language (Taguchi, 2020). Research in ILP has demonstrated that instructed learning significantly influences pragmatic development. Training or instruction focused on speech acts, politeness strategies, and appropriate language use in different contexts allows learners to outperform untrained peers (Taguchi, 2024). Activities such as discourse completion tasks, role plays, and dialogue analysis increase learners\u0026apos; awareness of pragmatic features and give learners the opportunity to practise appropriate use. Input also seems to aid learners\u0026apos; pragmatic development. Communicating with target language speakers or reading/listening to authentic materials allows learners to witness pragmatic conventions being used during real communicative exchanges. Exposure to pragmatic input can help learners infer implied meaning, interpret pragmalinguistic cues, and adjust language based on contextual features (Sattar et al., 2025). Exposure to language input can also be increased through technology-mediated environments. Interacting with others through computer-mediated communication, participating in virtual simulations, and exploring online interactions can expose learners to a variety of situations they would not typically experience otherwise. Research involving CALL shows that technology mediated environments can facilitate pragmatic development by providing learners with a space to communicate online. Finally, learner differences affect learners\u0026apos; pragmatic development. Individual differences such as motivation, language aptitude, and previous exposure to the target language can impact the rate and success of learners\u0026apos; pragmatic development (Yan, 2022). Another important topic related to ILP research is measurement. As most studies use tests such as discourse completion tasks to measure pragmatic ability, there is potential for these tests to not accurately represent learners\u0026apos; performance in more spontaneous interactions. For this reason, many researchers focus on analyzing naturally-occurring data.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.4 Cross-Cultural Realization of Speech Acts\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSpeech acts are a vital part of cross-cultural pragmatics because they provide insights into how culture influences language behaviour. Speech acts are communicative behaviours achieved through language, such as requests, apologies, refusals, complaints, suggestions, and giving thanks (Jord\u0026agrave; \u0026amp; Pilar, 2012). Each of these speech acts has different realisations depending on factors like social distance, power dynamics, and contextual expectations among interactants in different cultures. Cross-cultural comparisons of languages reveal that some cultures prefer direct speech acts, while others favour indirectness depending on the situation. For example, in certain contexts, a direct request is seen as appropriate in some cultures, whereas others opt for more indirect strategies for politeness (Khandani, 2017). Apology strategies also vary from being direct\u0026mdash;taking responsibility for an offence\u0026mdash;to more indirect forms (Tanduk, 2023). Refusals and complaints are also speech acts that need to be softened when expressed, as they can threaten face. Speakers might use strategies such as explanations, apologies, or indirect language to lessen the impact of the speech act. The realisation of speech acts is highly influenced by contextual factors such as power, social distance, and level of imposition, among others (Ali, 2025). Speech acts are essential in language learning since learners who translate them literally risk offending others by failing to adhere to pragmatic rules of the target language. Therefore, learners need to understand cross-cultural differences in speech act realisation.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.5 Arabic Sociocultural Norms and Pragmatic Transfer\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eArabic communication norms are rooted in sociocultural values that shape pragmatic behaviour. Traits such as indirectness, high-context relational communication, and respect for hierarchy are commonly observed among Arab communicators. For example, musāyara \u003cem\u003e(\u003cspan dir=\"RTL\"\u003eضمائر\u003c/span\u003e)\u003c/em\u003e refers to harmonious communication that avoids face-threatening acts (Mizel, 2016). Direct styles and, notably, face-threatening language can be viewed as impolite toward those of higher status. Arabs also prioritise communication that is relationally grounded. They tend to interpret meaning through social relations rather than linguistic form (Zaharna, 2010). As a result, politeness strategies emphasising respect are central. Arab speakers often incorporate religious traditions into their communication. Islamic values encourage respectful and humble interactions, and speakers may include religious expressions as polite or honest formulas (Ayish, 2003). Additionally, variation across regional and social dialects prompts code-switching (Soulaimani \u0026amp; Chakrani, 2023). When speaking English, Arab speakers apply their pragmatic norms, which can lead to pragmatic transfer from their cultural patterns into the foreign language. While this transfer can enhance understanding, it may also cause pragmatic failure if English speakers do not share the same expectations. Consequently, many scholars support explicit teaching of pragmatic differences.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.6 AI, Chatbots, and Pragmatic Language Learning\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eRecent developments in artificial intelligence (AI) provide another avenue for pragmatic development. Chatbots and large language models are capable of producing contextualised dialogue, explanations, and simulations of communicative situations for language learners. Studies have found interactive practice and contextualised language produced by chatbots can aid pragmatic development (Nguyen, 2024). The language produced by large language models is particularly useful as it can provide explanations and examples of dialogue for a wide variety of communicative situations. By leveraging this ability, AI has the potential to offer language learners scalable and individualised language practice. Preliminary research into AI-mediated language learning has shown promise that language learners can develop their pragmatic awareness by reviewing examples of speech acts and discourse generated by generative AI chatbots (Hussain, 2025). Other scholars have cautioned that language produced by AI chatbots needs to be evaluated if they are to be used as a source of pragmatic input. User should consider whether the AI is producing language that follows appropriate pragmatic norms and sociocultural expectations. In a newer branch of research, scientists have begun to question if AI technology displays pragmatic competence. Recent research has shown that language models are capable of understanding pragmatic phenomena such as implicature and other contextualised hints (Hu et al., 2022). However, AI does appear to struggle with more nuanced interpretations that require world knowledge and an understanding of social contexts. Research focused on pragmatics of AI-generated output is still relatively new. Current studies primarily focus on learners interacting with AI, but not the characteristics of the pragmatic input that is produced by AI.\u003c/p\u003e"},{"header":"3 Methodology","content":"\u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003e3.1 Study Design\u003c/h2\u003e \u003cp\u003eThe present study involved a qualitative document analysis approach to investigate how an Arabic-first artificial intelligence (AI) chatbot models English pragmatic language use and explains this use to Arabic speakers learning English as a foreign language (EFL). The AI-generated dialogue (response) and explanation (explanation of response), therefore, became \u0026ldquo;documents\u0026rdquo; from which we collected data. In this case study no human interactants were involved. Document analysis was utilized because the focus of this study is the pragmatic explanations generated by AI, rather than learners themselves. Each dialogue\u0026ndash;explanation pair was considered 1 document. A priori coding categories were established, and data saturation was ensured by analyzing a fixed dataset. All prompts used for the generation of AI responses as well as AI outputs are documented below for transparency/reproducibility.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003e3.2 Data Source\u003c/h2\u003e \u003cp\u003eThe data included AI-generated English dialogues and explanations created by HUMAIN Chat, an Arabic-first artificial intelligence chatbot designed to bridge communication between Arabic speakers and the rest of the world. To create the dataset, the authors developed 40 pragmatics-based prompts designed to mimic common communicative situations faced by EFL learners. These prompts reflected different pragmatic language functions such as requests, apologies, complaints, suggestions/give advice, greetings, and seek clarification. For each prompt sent to HUMAIN Chat, the AI generated 1 dialogue/response and 1 explanation describing the pragmatic reasoning for the response.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003e3.3 Data Collection\u003c/h2\u003e \u003cp\u003eData collection involved sending the 40 prompts to HUMAIN Chat and recording each corresponding dialogue/explanation generated by the AI in text document form. The prompts did not change over the course of data collection. Each dialogue\u0026ndash;explanation pair received a document number based on the original prompt number. Forty documents were included in this data set. Each document corresponds with 1 AI-generated dialogue and explanation.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003e3.4 Analytical Framework\u003c/h2\u003e \u003cp\u003eInterlanguage Pragmatics (ILP) served as the foundation to determine how pragmatic features were represented in the AI-generated data. From ILP literature, deductive coding categories were created a priori:\u003c/p\u003e \u003cp\u003e \u003col\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003ePragmatic function (request, apology, suggestion, etc.)\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003ePoliteness strategy (indirectness, mitigation, positive politeness, etc.)\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eUse of hedges/softener language\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eReferences to Arabic culture(s) in explanations\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eAcademic vs. peer context\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003c/ol\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003e3.5 Data Analysis\u003c/h2\u003e \u003cp\u003eThematic analysis was used to analyze the dataset (Braun \u0026amp; Clarke, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2006\u003c/span\u003e). The six-phase process included:\u003c/p\u003e \u003cp\u003e \u003col\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eFamiliarization with data \u0026ndash; Each of the 40 documents were read multiple times to become familiar with data as a whole.\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eInitial coding \u0026ndash; Each pragmatic feature was coded manually using the framework above.\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eGenerating themes \u0026ndash; Codes that were similar were combined to create themes.\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eReviewing themes \u0026ndash; We looked for patterns across all themes and made adjustments where needed.\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eDefining and naming themes \u0026ndash; Themes were named according to the pragmatic functions that were present.\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eProducing report \u0026ndash; Finalized themes were analyzed for similarities and differences as well as any limitations.\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eThe thematic analysis resulted in identification of common pragmatic themes among AI-generated dialogue/response and explanations.\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003c/ol\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003e4. Analysis, Results \u0026amp; Discussion\u003c/b\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003e4.1 Overview of AI Outputs\u003c/h2\u003e \u003cp\u003eThe AI-generated output demonstrates consistent use of:\u003c/p\u003e \u003cp\u003e \u003col\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eSpeech acts: Requests, refusals, apologies, suggestions, complaints, gratitude, clarifications, greetings, small talk.\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003ePoliteness strategies: Indirectness, hedging, softening, apologies, positive framing.\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eCultural grounding: Explicit references to Arabic norms (relational emphasis, honorifics, indirectness).\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eContextual appropriateness: Formality adjusted according to academic vs peer/social context.\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003c/ol\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eExample Results\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"7\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eScenario #\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSpeech Act\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eHedging / Softening\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003ePoliteness Strategy\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eCultural Grounding\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eContextual Appropriateness\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eExample AI Dialogue (Excerpt)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRequest\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eI was wondering if\u0026hellip;, could\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eIndirect request; polite\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eArabic requests emphasize relational respect\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eInformal / Peer-to-Peer\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026ldquo;Hi Sarah, I was wondering if I could borrow your English textbook for a few days? I\u0026rsquo;ll return it by Friday.\u0026rdquo;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRequest (professor)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eI was wondering if\u0026hellip;, would it be possible\u0026hellip;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eIndirect; formal; includes apology\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eArabic formal requests often include honorifics\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eFormal / Academic\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026ldquo;Dear Professor Ahmad, I was wondering if it would be possible to extend the deadline for my essay by two days.\u0026rdquo;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRequest (friend help)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eCould you\u0026hellip;, I\u0026rsquo;d really appreciate\u0026hellip;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eIndirect; polite\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eArabic emphasizes friendship / relational obligation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eInformal / Peer-to-Peer\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026ldquo;Hey Omar, could you help me with the last exercise from class? I\u0026rsquo;m struggling a bit.\u0026rdquo;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRefusal\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eI\u0026rsquo;m afraid\u0026hellip;, I won\u0026rsquo;t be able\u0026hellip;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003ePolite refusal; apology; positive comment\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eArabic emphasizes social harmony\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eInformal / Social\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026ldquo;Hi Sara, thanks for the invitation. I\u0026rsquo;m afraid I won\u0026rsquo;t be able to attend due to prior commitments.\u0026rdquo;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRefusal (group project)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eI\u0026rsquo;m sorry\u0026hellip;, I won\u0026rsquo;t be able\u0026hellip;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003ePolite; softening; alternative offer\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eArabic stresses relational obligations\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eFormal / Academic\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026ldquo;Hello team, I\u0026rsquo;m sorry but I won\u0026rsquo;t be able to participate in this project due to my current workload.\u0026rdquo;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eApology\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eI sincerely apologize\u0026hellip;, I will make sure\u0026hellip;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eFormal; acknowledgment; explanation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eArabic apologies include honorifics and respect\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eFormal / Academic\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026ldquo;Good morning, Professor. I sincerely apologize for arriving late to class. I will catch up on missed material.\u0026rdquo;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eNote\u003c/strong\u003e \u003cp\u003eA complete table summarizing AI-generated pragmatic outputs across all 40 scenarios is provided in Appendix A.\u003c/p\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003e4.2 Discussion\u003c/h2\u003e \u003cp\u003eIn the current study, we examined how HUMAIN Chat pragmatically appropriates English input for learners of Arabic as a foreign language (L2). First, we qualitatively analysed generated responses and explanations to identify pragmatic features that HUMAIN Chat consistently modelled across different situations. Overall, the findings suggest that HUMAIN Chat pragmatically appropriates English input by modelling features such as indirectness, hedging, and politeness. The AI also explained these choices using values derived from Arabic sociocultural norms. Both patterns imply that HUMAIN Chat has potential as a source of pragmatic input, while also highlighting areas where pedagogical intervention is necessary. Regarding RQ1, the AI pragmatically appropriated English choices by explaining them through Arabic sociocultural values. In nearly every scenario analysed, HUMAIN Chat justified pragmatic choices (e.g., requests, refusals, apologies) using cultural values such as maintaining relationships, respecting hierarchy and status, and preserving group harmony. Forms of indirectness (e.g., \"I was wondering if\u0026hellip;\", \"Would it be possible\u0026hellip;\") were particularly common across responses and often explained as polite and respectful forms of communication in both English and Arabic. We see this pattern as supporting ILP work arguing that learners need more than exposure to target language forms in order to notice and develop pragmatic competence. While repeatedly generating indirect forms, HUMAIN Chat consistently offered metapragmatic explanations related to Arabic sociocultural values. By explaining English pragmatic choices through the lens of Arabic pragmatics, HUMAIN Chat connected learner background to target language norms. The AI functioned as a mediational tool by helping learners notice the gap between desired (i.e., native-like) pragmatic norms and learners\u0026rsquo; (potentially inappropriate) L1 norms. Such noticing has the potential to prevent negative pragmatic transfer from learners\u0026rsquo; L1 to English.\u003c/p\u003e \u003cp\u003eTurning to RQ2, HUMAIN Chat showed consistency in the language it generated across the forty learner scenarios. For instance, requests, suggestions, and clarifications were often expressed through modal verbs (could, would, might) and indirect language. Refusals and complaints were frequently softened with apologies, explanations, or other accounts. These patterns reflect standardised notions of politeness and mitigation typical in English pragmatic communication. The AI also showed sensitivity to some contextual variables. For example, HUMAIN Chat recognised register differences when communicating with professors versus friends or classmates. Responses involving professors were markedly more formal and polite than those addressed to friends or classmates. Although these trends were not universal, they suggest that HUMAIN Chat can model pragmalinguistic forms and sociopragmatic norms, both of which are often severely limited in EFL textbooks. However, some inconsistencies appeared in the data. On several occasions, HUMAIN Chat produced responses that were unexpectedly formal given the communication context. Students were not exactly wrong to write these responses, but they might feel uncomfortable using such forms in everyday conversations with native speakers. We attribute this phenomenon to AI avoidance behaviour. When unsure of how to respond, AI tends to generate safe (i.e., polite) responses. While these features may limit the AI\u0026rsquo;s usefulness, teachers can work with students to address register issues as they use AI to source pragmatic input. Beyond pedagogical implications, this study adds a new dimension to the growing body of research on AI in language learning. Instead of focusing on how learners utilise chatbots or how AI enhances L2 writing skills, our aim was to examine the pragmatic features of language produced by AI. Since most chatbots are trained on internet data (mainly generated by Anglophone speakers), there is concern that they may deliver native-speaker pragmatic input that does not fully consider learners\u0026rsquo; sociocultural backgrounds.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003e4.3 Discussion of Findings in Relation to Previous Research\u003c/h2\u003e \u003cp\u003eTo conclude this section, the results align with multiple lines of previous research summarised in Chap.\u0026nbsp;2. First, the prominence of indirect forms, hedging devices, and mitigation strategies reflects earlier descriptions of pragmatic competence in EFL contexts. Prior studies have defined pragmatic competence as the skills necessary for appropriate speech act production and interpretation of communicative intent beyond basic sentence grammar (Jord\u0026agrave; \u0026amp; Pilar, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2012\u003c/span\u003e; Kentmen et al., \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Kusevska et al., \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). When evaluated against these standards, the AI\u0026rsquo;s repetition of could, might, and I was wondering if\u0026hellip; would count as pragmatic routines identified in politeness studies. Overall, these results support ideas of pragmatic competence as sociopragmatic appropriateness rather than linguistic accuracy.\u003c/p\u003e \u003cp\u003eSecond, they confirm past research about connections between grammatical competence and pragmatic competence. Research cited in section \u003cspan refid=\"Sec5\" class=\"InternalRef\"\u003e2.4\u003c/span\u003e established that speakers can produce output that, while grammatical, is pragmatically inappropriate (Bachelor, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2015\u003c/span\u003e; Swan, \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2007\u003c/span\u003e). But none of the AI\u0026rsquo;s responses fell into this category \u0026ndash; every generated turn was both pragmatically appropriate and grammatically accurate. This may mean that pragmatics inherently includes choices about language forms used to establish relationships, show power dynamics, and convey contextual information. If so, these data further suggest that learners should develop grammatical and pragmatic competence concurrently.\u003c/p\u003e \u003cp\u003eThird, my findings agree with ILP research on the benefits of explicit instruction and metapragmatic explanation. Other studies highlighted learner improvements after being given pragmatic explanations for why certain speech acts are (or are not) appropriate in certain situations (Taguchi, \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Mokoro, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). HUMAIN Chat provided learners with these explanations after most of the conversations it generated. For many conversations, the AI described how particular turns were polite or fit the setting. Because input like this is closer to scaffolding/instruction than natural conversation, AI explanations could help learners go through the noticing processes ILP advocates promote.\u003c/p\u003e \u003cp\u003eFourth, the data agree with prior research on cross-cultural speech-act realisation. Earlier studies found that English speakers realise requests indirectly and mitigate when refusing (Alshammari, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2015\u003c/span\u003e; Khandani, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). HUMAIN Chat similarly generated indirect requests and mitigated with apologies when refusing. The bot also appeared to take into account suprascribed variables like distance/power for certain exchanges. Power and distance have been known PR AGmatic variables for decades, so it\u0026rsquo;s notable that they arose during conversations with AI.\u003c/p\u003e \u003cp\u003eMost notably, these results mirror previous research discussing Arabic culture, sociocultural communication, and pragmatic transfer. Various authors have recognised Arabic-speaking societies as valuing respect between relations, indirect communication styles, and consideration of in-group hierarchies (Zaharna, 2010; Mizel, \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; Dendenne, \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). On many occasions, HUMAIN Chat referenced these values when justifying specific English choices, even going so far as to contrast English and Arabic directly. ILP theories suggest that explanations like these can help learners notice crosslinguistic differences and avoid pragmatic transfer from Arabic into English.\u003c/p\u003e \u003cp\u003eSimilar Arabic-to-English connections can be drawn from earlier research on cross-cultural misunderstandings. Culture-specific communication styles, such as high-context versus low-context, can lead to miscommunications when interlocutors assume different levels of implicit information (Li, \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Meng \u0026amp; Wang, \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). During the teacher-passing scenario and others, HUMAIN Chat explicitly mentioned that English speakers need to communicate more information verbally because the language has less contextualisation than Arabic. Highlighting these kinds of differences may enhance learners' intercultural understanding.\u003c/p\u003e \u003c/div\u003e"},{"header":"5. Conclusion \u0026 Recommendations","content":"\u003cp\u003eThe present study manually analysed 40 AI-generated chat conversations created using HUMAIN Chat to assess how pragmatically natural the AI-modeled input is, specifically designed for English as a foreign language learners from an Arabic background. Results showed that HUMAIN Chat almost always employed indirectness, hedging, mitigation, and politeness strategies in its replies. Additionally, HUMAIN Chat explanations often included explicit metapragmatic annotations based on Arabic cultural scripts related to being relationally-oriented, understanding/status-conscious, and harmony-seeking. This suggests that HUMAIN Chat might serve as an instructional tool to help learners develop sociopragmatic awareness and reduce negative pragmatic transfer. However, the study also identified instances where HUMAIN Chat offered overly formal utterances, further emphasising that teachers should scaffold AI-generated outputs and encourage learners to critically evaluate AI-modeled input. Consequently, teachers should not rely solely on AI as the ultimate authority on how English is used and spoken; rather, AI can be employed to facilitate discussions about pragmatics. For example, teachers could ask learners to evaluate whether the dialogue is suitable for the situation (register), if it sounds natural, and what might be altered. Activity designers can also prompt learners to critique AI-generated role-plays, rewrite the dialogues, or explore differences between Arabic and English. Teacher training programmes should include guidance on the appropriate and critical use of AI in ELT. Developers can also work on enhancing register flexibility, naturalness, and context appropriateness while maintaining certain cultural values consistent with the target learner demographic.\u003c/p\u003e"},{"header":"6. Limitations and Directions for Future Research","content":"\u003cp\u003eLimitations. Several limitations should be acknowledged. First, the present study only analysed AI-generated documents; no human participants were included. As such, it is unclear whether learners acquire any knowledge from the pragmatically informative messages AI produced or if they can use these in real communicative interactions. Second, only one Arabic-first AI chatbot was included in the analyses. Other AI technologies may provide learners with different pragmatic outputs depending on how the AI is designed and what data it is fed. Third, qualitative thematic coding was guided by ILP constructs. While the categories included offered a useful framework for sorting and identifying pragmatic phenomena within the learner-teacher dialogues, using fixed codes might have limited the researcher’s ability to discover other pragmatic information. Qualitative analyses are also susceptible to interpretation biases. Fourth, the data consisted of dialogues produced by AI, which may not fully reflect pragmatically informative naturally occurring speech. Future research should incorporate learner data to better understand how Arabic EFL learners interpret and utilise AI-mediated pragmatic input. An experimental or quasi-experimental approach could determine if learners studying pragmatics with AI-supported instruction develop a greater understanding than those using traditional methods. Additional studies should explore teachers’ perspectives on implementing culturally appropriate AI in language education. Longitudinal research might also reveal whether repeated exposure to AI-generated pragmatic explanations leads to improved pragmatic competence. Further investigations should include various AI technologies and learners from diverse backgrounds.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eDisclosure of Generative AI use\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDeclared by the authors and we take full responsibility that Generative AI was used only for grammar and style editing of the manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting Interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe author(s) confirm that this piece was written in the absence of any commercial or financial relationships that could be considered as a potential conflict of interest.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eEthics approval was not required as the study only involved AI generated documents and did not involve human participants or data.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData Availability Statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets and materials generated and/or analyzed during the current study are available from the corresponding author upon reasonable request. In addition, anonymized data, coding schemes, and analysis procedures have been provided as supplementary materials to support the transparency and reproducibility of the study.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u0026nbsp;\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n \u003cli\u003eAli, S. (2025). Pragmatic realization of complaints across cultures. \u003cem\u003eJournal of Pragmatics, 210\u003c/em\u003e, 80\u0026ndash;92. https://doi.org/10.1016/j.pragma.2024.02.006\u003c/li\u003e\n \u003cli\u003eAlrefaee, Y. (2025). Strategy-based instruction and pragmatic competence in EFL learners. \u003cem\u003eLanguage Teaching Research\u003c/em\u003e. https://doi.org/10.1177/13621688231123456\u003c/li\u003e\n \u003cli\u003eAlshammari, M. (2015). Request strategies in Saudi Arabic and English: A contrastive pragmatic study. \u003cem\u003eJournal of Pragmatics, 84\u003c/em\u003e, 39\u0026ndash;54. https://doi.org/10.1016/j.pragma.2015.04.007\u003c/li\u003e\n \u003cli\u003eAyish, M. (2003). Beyond Western-oriented communication theories: A normative Arab-Islamic perspective. \u003cem\u003eJavnost \u0026ndash; The Public, 10\u003c/em\u003e(2), 79\u0026ndash;92. https://doi.org/10.1080/13183222.2003.11008857\u003c/li\u003e\n \u003cli\u003eBachelor, J. (2015). Pragmatic competence in second language communication. \u003cem\u003eJournal of Pragmatics, 76\u003c/em\u003e, 1\u0026ndash;14. https://doi.org/10.1016/j.pragma.2014.11.005\u003c/li\u003e\n \u003cli\u003eBarzani, S., \u0026amp; Mohammadzadeh, B. (2022). Developing pragmatic competence in EFL classrooms through communicative tasks. \u003cem\u003eJournal of Language Teaching and Research, 13\u003c/em\u003e(2), 372\u0026ndash;381. https://doi.org/10.17507/jltr.1302.17\u003c/li\u003e\n \u003cli\u003eBender, E. M., Gebru, T., McMillan-Major, A., \u0026amp; Shmitchell, S. (2021). On the dangers of stochastic parrots: Can language models be too big? \u003cem\u003eProceedings of the ACM Conference on Fairness, Accountability, and Transparency\u003c/em\u003e, 610\u0026ndash;623. https://doi.org/10.1145/3442188.3445922\u003c/li\u003e\n \u003cli\u003eBraun, V., \u0026amp; Clarke, V. (2006). Using thematic analysis in psychology. \u003cem\u003eQualitative Research in Psychology, 3\u003c/em\u003e(2), 77\u0026ndash;101. https://doi.org/10.1191/1478088706qp063oa\u003c/li\u003e\n \u003cli\u003eCelaya, M. L., \u0026amp; Bar\u0026oacute;n, J. (2015). Pragmatic development in EFL learners. \u003cem\u003eSystem, 48\u003c/em\u003e, 16\u0026ndash;28. https://doi.org/10.1016/j.system.2014.09.012\u003c/li\u003e\n \u003cli\u003eDalmasso, L. (2009). The role of grammatical competence in communicative language ability. \u003cem\u003eLanguage Learning Journal, 37\u003c/em\u003e(1), 15\u0026ndash;28. https://doi.org/10.1080/09571730802599122\u003c/li\u003e\n \u003cli\u003eDendenne, B. (2017). Politeness and pragmatic transfer in Arabic communication. \u003cem\u003eIntercultural Pragmatics, 14\u003c/em\u003e(3), 353\u0026ndash;375. https://doi.org/10.1515/ip-2017-0016\u003c/li\u003e\n \u003cli\u003eHu, J., Floyd, S., Jouravlev, O., Fedorenko, E., \u0026amp; Gibson, E. (2022). A fine-grained comparison of pragmatic language understanding in humans and language models.\u003c/li\u003e\n \u003cli\u003eHussain, S. (2025). Generative AI and pragmatic awareness in language learning. \u003cem\u003eComputer Assisted Language Learning\u003c/em\u003e. https://doi.org/10.1080/09588221.2025.2310045\u003c/li\u003e\n \u003cli\u003eIshihara, N., \u0026amp; Cohen, A. D. (2021). \u003cem\u003eTeaching and learning pragmatics: Where language and culture meet\u003c/em\u003e (2nd ed.). Routledge. https://doi.org/10.4324/9781315715421\u003c/li\u003e\n \u003cli\u003eJord\u0026agrave;, M. P., \u0026amp; Pilar, M. (2012). Speech act realization in second language communication. \u003cem\u003eJournal of Pragmatics, 44\u003c/em\u003e(4), 469\u0026ndash;484. https://doi.org/10.1016/j.pragma.2011.12.009\u003c/li\u003e\n \u003cli\u003eKentmen, E., Kusevska, M., \u0026amp; Dimitrovska, S. (2023). Implicature comprehension in second language learning. \u003cem\u003eIntercultural Pragmatics, 20\u003c/em\u003e(3), 305\u0026ndash;324. https://doi.org/10.1515/ip-2023-2004\u003c/li\u003e\n \u003cli\u003eKerbrat-Orecchioni, C. (2021). \u003cem\u003eThe pragmatics of discourse\u003c/em\u003e. Cambridge University Press. https://doi.org/10.1017/9781108988216\u003c/li\u003e\n \u003cli\u003eKhandani, S. (2017). Cross-cultural request strategies: Politeness and indirectness in interaction. \u003cem\u003eJournal of Pragmatics, 117\u003c/em\u003e, 38\u0026ndash;53. https://doi.org/10.1016/j.pragma.2017.06.007\u003c/li\u003e\n \u003cli\u003eKohnke, L., \u0026amp; Moorhouse, B. (2023). ChatGPT for language teaching: Opportunities and challenges. \u003cem\u003eTESOL Journal\u003c/em\u003e. https://doi.org/10.1002/tesj.720\u003c/li\u003e\n \u003cli\u003eKohnke, L., Moorhouse, B., \u0026amp; Zou, D. (2022). Chatbots in language education: A systematic review. \u003cem\u003eComputer Assisted Language Learning, 35\u003c/em\u003e(5\u0026ndash;6), 1\u0026ndash;27. https://doi.org/10.1080/09588221.2020.1850425\u003c/li\u003e\n \u003cli\u003eKusevska, M., Daskalovska, N., \u0026amp; Ivanovska, B. (2015). Speech acts in L2 pragmatics research. \u003cem\u003eProcedia \u0026ndash; Social and Behavioral Sciences, 191\u003c/em\u003e, 146\u0026ndash;150. https://doi.org/10.1016/j.sbspro.2015.04.249\u003c/li\u003e\n \u003cli\u003eLee, J., \u0026amp; Dressman, M. (2022). When GPT-3 writes essays: Assessing AI writing in EFL contexts. \u003cem\u003eComputers and Composition, 63\u003c/em\u003e, 102697. https://doi.org/10.1016/j.compcom.2022.102697\u003c/li\u003e\n \u003cli\u003eLi, C. (2023). Pragmatic failure in cross-cultural communication. \u003cem\u003eJournal of Pragmatics, 210\u003c/em\u003e, 62\u0026ndash;74. https://doi.org/10.1016/j.pragma.2023.01.007\u003c/li\u003e\n \u003cli\u003eMeng, Y., \u0026amp; Wang, L. (2024). Revisiting high-context and low-context communication in intercultural interaction. \u003cem\u003eIntercultural Communication Studies, 33\u003c/em\u003e(1), 45\u0026ndash;61. https://doi.org/10.1080/17475759.2023.2234567\u003c/li\u003e\n \u003cli\u003eMerchant, H., Floyd, S., \u0026amp; Jouravlev, O. (2025). Nonverbal communication and pragmatic interpretation in multilingual contexts. \u003cem\u003eJournal of Pragmatics, 211\u003c/em\u003e, 1\u0026ndash;12. https://doi.org/10.1016/j.pragma.2025.01.004\u003c/li\u003e\n \u003cli\u003eMizel, O. (2016). Indirectness and politeness in Arabic communication. \u003cem\u003eIntercultural Pragmatics, 13\u003c/em\u003e(2), 245\u0026ndash;268. https://doi.org/10.1515/ip-2016-0010\u003c/li\u003e\n \u003cli\u003eMokoro, J. (2024). Metapragmatic instruction and second language learning. \u003cem\u003eApplied Linguistics Review\u003c/em\u003e. https://doi.org/10.1515/applirev-2024-0034\u003c/li\u003e\n \u003cli\u003eNguyen, T. (2024). Artificial intelligence and language learning: Opportunities and challenges. \u003cem\u003eComputer Assisted Language Learning\u003c/em\u003e. https://doi.org/10.1080/09588221.2024.2300123\u003c/li\u003e\n \u003cli\u003eO\u0026rsquo;Dowd, R. (2023). Artificial intelligence and the future of language learning. \u003cem\u003eLanguage Learning \u0026amp; Technology, 27\u003c/em\u003e(3), 1\u0026ndash;15. https://doi.org/10.55593/lllt.2023.27.3.1\u003c/li\u003e\n \u003cli\u003eRetnowaty, R. (2022). Pragmatic competence and communicative performance in EFL classrooms. \u003cem\u003eSystem, 106\u003c/em\u003e, 102757. https://doi.org/10.1016/j.system.2022.102757\u003c/li\u003e\n \u003cli\u003eRustandi, A., Tajeddin, Z., \u0026amp; Malmir, A. (2025). The relationship between grammatical competence and pragmatic competence in EFL learning. \u003cem\u003eSystem, 121\u003c/em\u003e, 103161. https://doi.org/10.1016/j.system.\u003c/li\u003e\n \u003cli\u003eSattar, H., Khan, R., \u0026amp; Ali, M. (2025). Authentic input and pragmatic development in second language learning. \u003cem\u003eLanguage Teaching Research\u003c/em\u003e. https://doi.org/10.1177/13621688241234567\u003c/li\u003e\n \u003cli\u003eSoulaimani, D., \u0026amp; Chakrani, B. (2023). Dialect variation and pragmatic competence in Arabic communication. \u003cem\u003eJournal of Sociolinguistics, 27\u003c/em\u003e(4), 501\u0026ndash;520. https://doi.org/10.1111/josl.12587\u003c/li\u003e\n \u003cli\u003eSwan, M. (2007). The influence of the mother tongue on second language learning. \u003cem\u003eELT Journal, 61\u003c/em\u003e(2), 102\u0026ndash;112. https://doi.org/10.1093/elt/ccm008\u003c/li\u003e\n \u003cli\u003eSykes, J. (2020). Pragmatics in language learning technology. \u003cem\u003eLanguage Teaching, 53\u003c/em\u003e(1), 117\u0026ndash;124. https://doi.org/10.1017/S0261444819000325\u003c/li\u003e\n \u003cli\u003eTaguchi, N. (2020). \u003cem\u003eSecond language pragmatics\u003c/em\u003e. Cambridge University Press. https://doi.org/10.1017/9781108887793\u003c/li\u003e\n \u003cli\u003eTaguchi, N. (2024). Explicit instruction in second language pragmatics. \u003cem\u003eApplied Linguistics, 45\u003c/em\u003e(2), 241\u0026ndash;262. https://doi.org/10.1093/applin/amad032\u003c/li\u003e\n \u003cli\u003eTaguchi, N., \u0026amp; Roever, C. (2021). \u003cem\u003eSecond language pragmatics\u003c/em\u003e. Oxford University Press.\u003c/li\u003e\n \u003cli\u003eTanduk, R. (2023). Cross-cultural apology and refusal strategies in multilingual communication. \u003cem\u003eJournal of Pragmatics, 204\u003c/em\u003e, 112\u0026ndash;124. https://doi.org/10.1016/j.pragma.2023.01.009\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"humanities-and-social-sciences-communications","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"palcomms","sideBox":"Learn more about [Humanities \u0026 Social Sciences Communications](http://www.nature.com/palcomms/)","snPcode":"41599","submissionUrl":"https://submission.springernature.com/new-submission/41599/3","title":"Humanities and Social Sciences Communications","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Nature AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Artificial intelligence, English pragmatics, Interlanguage pragmatics, AI-generated dialogue","lastPublishedDoi":"10.21203/rs.3.rs-9117504/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9117504/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eContextualised pragmatic input is vital for developing pragmatic competence in English as a Foreign Language (EFL), but can often be overlooked in EFL learning materials, particularly those based on cultural context. Recent developments in artificial intelligence (AI) offer new platforms for contextualised pragmatic instruction, however little research has investigated the pragmatics of AI technology generated language output nor how that output is culturally mediated for Arabic EFL learners. The current study examined how one Arabic-first AI system mediates pragmatic knowledge and pragmatic language use. We took a qualitative descriptive approach to analyzing 40 pieces of AI-generated dialogue exchanges and explanation responses across academic, social, and interpersonal contexts. Data were coded using an interlanguage pragmatics informed framework focusing on speech acts, politeness, hedging, cultural mediation, and contextual appropriateness. Results indicate that AI technology models consistent indirectness, mitigation, and politeness across speech acts and pragmatic explanations use Arabic sociocultural concepts related to respect for relationships and hierarchy. Despite a few examples of overly formal language used in informal contexts, our results provide evidence that AI technologies can act as sociocultural mediational tools and have utility in pragmatics instruction for Arabic EFL learners.\u003c/p\u003e","manuscriptTitle":"Analyzing culturally grounded AI outputs in teaching English pragmatics: A qualitative study of HUMAIN Chat","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-08 12:53:09","doi":"10.21203/rs.3.rs-9117504/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"reviewersInvited","content":"","date":"2026-04-02T12:26:23+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-04-01T07:11:54+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2026-04-01T07:04:00+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-03-19T17:24:29+00:00","index":"","fulltext":""},{"type":"submitted","content":"Humanities and Social Sciences Communications","date":"2026-03-19T16:33:38+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"humanities-and-social-sciences-communications","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"palcomms","sideBox":"Learn more about [Humanities \u0026 Social Sciences Communications](http://www.nature.com/palcomms/)","snPcode":"41599","submissionUrl":"https://submission.springernature.com/new-submission/41599/3","title":"Humanities and Social Sciences Communications","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Nature AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"fc0935f9-2982-43c7-b40c-8dd5785dd1db","owner":[],"postedDate":"April 8th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":65736139,"name":"Social science/Education"},{"id":65736141,"name":"Humanities/Language and linguistics"},{"id":65736143,"name":"Social science/Language and linguistics"},{"id":65736144,"name":"Humanities/Philosophy"}],"tags":[],"updatedAt":"2026-04-08T12:53:09+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-08 12:53:09","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9117504","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9117504","identity":"rs-9117504","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}