Language Learning & Technology ISSN 1094-3501 October 2020, Volume 24, Issue 3 pp. 120–135 ARTICLE Online learning negotiation: Native-speaker versus nonnative speaker teachers and Vietnamese EFL learners Pham Kim Chi, FPT University Nguyen Van Loi, Can Tho University Abstract Online English language teaching can now be facilitated by communication technology, which allows easy access to interaction with native speakers. Nevertheless, this industry subscribes to an assumption that native speaker English teachers (NESTs) are the gold standard of language whereas the non-native speaker English teachers (NNESTs) are inferior educators (Walkinshaw & Duong, 2014). Rare research has provided evidence of the negotiation produced by NESTs versus NNESTs with EFL learners online and its impact on the learners’ output. Thus, the current study narrows this empirical gap. Drawing upon a database of 30 five-minute interaction sessions between 30 teachers (15 NESTs and 15 NNESTs) and 30 basic level Vietnamese EFL adult learners, the study revealed similar negotiation of meaning functions as reported in previous research. However, the NESTs used more elaboration while the NNESTs used more confirmation checks, clarification requests, and reply clarification. Qualitative analysis further indicated that the NNESTs provided more productive support, encouraging the learners’ output, than the NESTs did. This implies that although online voice interaction creates an environment for EFL learners to practice, language educators and teachers, regardless of status, should heed how to handle it so that online learners can benefit from both comprehensible input and opportunities for pushed output. Keywords: Computer-mediated communication, output, meaning negotiation, EFL learners Language(s) Learned in This Study: English APA Citation: Chi, P. K., & Loi, N.V. (2020). Online Learning Negotiation: Native-Speaker Versus Nonnative Speaker Teachers & Vietnamese EFL learners. Language Learning & Technology, 24(3), 120–135. http://hdl.handle.net/10125/44743 Introduction Opportunities to practice speaking with English native speakers are crucial to improve English learners’ proficiency. In most situations, learners sharing the same L1 feel communication in the L2 is inauthentic (Zhao, 2005) and frequently switch to L1 to solve communication problems (Smith, 2005); they have few opportunities to use the target language to communicate with different L1 speakers or native speakers outside the classroom (Chiu et al., 2007). Yet, technology now can allow anyone with an Internet connection access to real conversations with English speakers worldwide. Online classes with English teachers both native and non-native are valuable language practice opportunities for EFL learners who lack daily exposure to English. Engaging in interaction with these English teachers, they are likely to receive comprehensible input, produce modified output through negotiation of meaning (NoM) and receive negative feedback that facilitates language acquisition (Long, 1996). According to Long (1996, p. 418), negotiation of meaning is the process in which learners and competent speakers provide and interpret signals of their own and their interlocutor’s perceived comprehension, thus provoking adjustments to linguistic form, conversational structure and message content, or all three until an acceptable level of understanding is achieved. However, the English language teaching industry in the East and Southeast Asia currently subscribes to an assumption that native English speaking teachers (NESTs) are the gold standard of spoken and written language, whereas non-native English speaking teachers (NNESTs) are inferior educators due to their lack of native-like linguistic competence (Walkinshaw & Duong, 2014). As Medgyes (1992) Pham Kim Chi & Nguyen Van Loi 121 noted, NESTs are commonly deemed preferable to NNESTs regardless of proficiency or experience. Presumably, few NNESTs have experienced the target culture or attained a proficiency level deemed adequate for providing English language learners with “good target language models” (Snow et al., 2006, p. 275). This belief is in line with learners’ ideology for successful communication. Conversation with native English speakers is what EFL learners generally cite as the most desired form of language study (Terhune, 2013). Despite recent debates on English nativeness, the distinction between NESTs and NNESTs continues to exist in “the minds of general public” (Pacek, 2005, p. 243). Kachru’s (1997) model indicates that English in the Inner Circle represents the traditional historical and sociolinguistic bases and is currently used as the native language. In this regard, a NEST is defined as an “inner circle” instructor for whom English is the home language; a NEST is a native of the United States, Australia, Ireland, the United Kingdom, Canada and South Africa (Seidlhofer, 2004). The Outer Circle expanded towards Asia and Africa, where English serves as a useful lingua franca between various ethnic and language groups who may speak English for practical purposes. This circle includes India, Nigeria, Bangladesh, Pakistan, Malaysia, Tanzania, Kenya, non-Anglophone South Africa, the Philippines and others. The Expanding Circle involves the rest of the world's population who use English mainly as a medium of international communication. In this study, the term NNEST covers educators from numerous nations in the outer or expanding circle locations; they are speakers of world Englishes (Seidlhofer, 2004, p. 210). Many studies have investigated types of NoM in text-based interactions (Patterson & Trabaldo, 2006, Akayoglu & Altun, 2009). In synchronous voice-based computer-mediated communication (CMC), NESTs were found to be beneficial in modifying EFL learners’ output (Bueno, 2012). However, research has yet to compare the quality of interactions between NESTs versus NNESTs with EFL learners in voice-based CMC. Thus, this current study narrows the empirical gap by examining NoM between Vietnamese EFL adult learners with NESTs versus NNESTs through online voice-based interactions. The study aims to investigate the two following research questions: 1. What functions of NoM are found in online voice interactions between NESTs and Vietnamese EFL learners, and between NNESTs and Vietnamese EFL learners? 2. Is there any difference between the two types of dyads in negotiation functions and how does NEST and NNEST language use impact the amount of learner output? Literature Review This section discusses the role of interaction and NoM in second language acquisition (SLA), CMC interactions and previous studies on CMC interactions. Interaction and NoM in SLA Interaction is commonly defined as communication between two interlocutors (Fernández-García & Martínez-Arbelaiz, 2002). From SLA perspectives, interaction has a more extended meaning. Long (1981) defined interaction as functions manifest in linguistic forms such as expansion and clarification. Interaction is also defined as social behaviour that occurs when one person communicates with another (Ellis, 2003). By this, interaction may take different functions depending on the context and the needs of communicators. In the early to mid 1980s, research on interaction focused on the role of input in SLA. Long (1981) examined NS-NNS conversation and NNS-NNS conversation, observing that interaction between NS- NNS dyads generated more comprehensible input than that between NNS-NNS dyads. Interaction has been proved to benefit language learners because it triggers NoM which in turn produces comprehensible input and feedback that can promote language acquisition (Long, 1996; Pica, 1994). Swain’s (1985) Output Hypothesis further reinforces the position of interaction. According to Swain (1985), when L2 learners attempt production, they test out their assumptions about the target language. Through receiving feedback or signals during interaction, learners are pushed to produce more comprehensible output, being aware of the gap between their existing knowledge and what they intend to convey, which in turn forces them to extend their inter-language syntactically. Swain (1985) stressed that language production facilitates acquisition when learners experience communication failure. 122 Language Learning & Technology Central to interaction is NoM, which is considered an essential element of language acquisition in many SLA studies (Pica, 1994; Ellis, 2003). For Pica (1994, p. 494), NoM refers to “the modification and restructuring of interaction between interlocutors when they experience comprehension difficulties.” Similarly, Ellis (2003) described it as “the process by which two or more interlocutors identify and then attempt to resolve a communication breakdown” (p. 346). Long’s (1983) framework of interactional modifications which involves clarification requests, confirmation checks and comprehension checks is useful for SLA studies on NoM. Interaction in CMC Computer-mediated communication (CMC) involves a variety of forms of synchronous, asynchronous or real-time interaction. With strong technical features, CMC has attracted many researchers in both synchronous and asynchronous modes, and research on computer-based interaction in SLA has also continued to proliferate. CMC has extended boundaries to surpass time and location for language learning and teaching. As Olaniran (1996) stated, CMC promotes group communication via electronic mail, voicemail, and videoconferences among organizations. Research (e.g., Warschauer, 2003) shows that CMC can be a good tool for second language (L2) learning. Studies comparing CMC interaction and face-to-face in L2 concluded that quiet learners are more expressive in CMC than in face-to-face interactions (Kern, 1995; Warschauer, 1996). Web-conferencing platforms (e.g. Adobe Connect, Blackboard Collaborate, and Big Blue Button) provide educators with synchronous, multimodal communication opportunities. Synchronous CMC resembles real-time interactions (Smith, 2003) and may elicit higher-quality spoken output (Warschauer, 1996). According to Fernández-García and Martínez-Arbelaiz (2002), not only does synchronous CMC in the L2 classroom give learners opportunities to negotiate meaning, but it also enables them to do so at their own pace. NoM Framework for CMC The current study involves synchronous CMC mediated by BigBlueButton, an open source web conferencing system for online learning. While it is mainly premised on Long’s (1983) Interaction Hypothesis for the values of interaction and NoM in language acquisition, it is not restricted to his move framework. Studies have suggested that negotiation routines and strategies in CMC are different from those in face-to-face interactions (Fernández-García and Martínez-Arbelaiz, 2002; Smith, 2003). Learners were found to participate more in CMC compared to face-to-face conversation (Warschauer, 1996). Research on CMC (e.g., Kötter, 2003; Jepson, 2005; Patterson & Trabaldo, 2006; Akayoglu & Altun 2009; Samani & Noordin, 2014) thus has expanded the framework to include more negotiation moves. Different frameworks have been employed to describe NoM functions in both asynchronous and synchronous interaction. The categories of Patterson (2001) were utilized for analysis of NoM moves in studies on e-mail chat and Instant Messaging (e.g. Patterson & Trabaldo, 2006); the negotiation model described in Varonis and Gass (1985) was used to identify negotiation routines in task-based email exchanges (e.g., Kitade, 2006) and in face-to-face interactions and CMC (e.g., Shim, 2007). Long’s (1983) framework was employed in research on text chat versus voice chat (e.g., Jepson, 2005); negotiation strategies defined by Long (1983), Pica and Doughty (1985) and Tarone (1980) in face-to- face exchanges were used in online discussions among learners (e.g., Lee, 2001); the NoM framework proposed by Patterson and Trabaldo (2006) was employed in research on text-based learners’ interaction (e.g., Samani & Noordin, 2014). This study relied on the taxonomy developed by Patterson and Trabaldo (2006) which was based on Patterson’s (2001) work on Computer-Assisted Class Discussions, and extended by Akayoglu and Altun (2009), as presented in Table 1. There are two main reasons for using this framework. First, the taxonomy was based on various resources in CMC environment as previously reviewed. More importantly, it was well suited to the research questions and study focus. Pham Kim Chi & Nguyen Van Loi 123 Table 1. Taxonomy of NoM functions by Patterson and Trabaldo (2006) Functions of NoM Explanation Confirmation check (ConC) Asking for confirmation of a previously made statement to be sure they have understood correctly. Clarification request (ClarR) Requesting clarification for an ambiguous statement, which may cause misunderstanding. Vocabulary request (VocR) Requesting a vocabulary word or phrase in the target language Comprehension check (ComC) Asking if the other person understood what was said or written, and generally expecting that he or she has understood. Reply clarification (RClar) Clarifying their previous statement as a result of request (clarification request). Reply confirmation (RConf) Confirming a statement when someone requests confirmation with expressions like “yes,” “OK,” “you are right.” Reply vocabulary (RVoc) Giving a meaning of the words or phrase as a result of request. Reply comprehension (RCom) Replying to comprehension check or indicating that statement was understood. Elaboration (Elab) Elaborating the meaning of a previous statement no matter whether the previous statement belongs to them. Correction – Self correction (Cor) Correcting an error made by another speaker or self-correction of one’s own error. Confirmation* (Conf) Confirming the previous statement although there is no confirmation request Elaboration request* (ER) Requesting elaboration if they do not have an idea about the speaker’s utterance, and requesting extra information. Reply elaboration* (RElab) Elaborating their own statement to make it clear as a result of request (elaboration request). *Functions of NoM added by Akayoglu and Altun (2009) Previous studies on interaction in CMC Until recently, researchers have investigated NoM in various CMC modes with different types of participants. Following Long (1983), Lee (2001) examined the negotiation strategies of learners of Spanish, using ParaChat program and found that the learners tended to use comprehension checks, clarification checks, requests and self-repairs. Clarification checking was the most salient move. The study further indicated that these strategies facilitated learners’ input and output. Patterson and Trabaldo (2006) found that learners of Spanish and learners of English were engaged more in Instant Messaging than email, and used NoM functions somewhat differently, specifically clarification requests, vocabulary requests, confirmation, and comprehension checking. Using a similar framework, Samani and Noordin (2014) indicated that ESL learners most frequently used confirmation, confirmation checks and elaboration and least frequently used reply confirmation, reply clarification and comprehension checks in text-based CMC. In Kötter’s (2003) study between German learners of English and American learners of German in online Tandems learning, clarification requests and elaboration were often employed, and notably the learners produced a similar number of confirmation checks and clarification requests. Bueno (2012) examined synchronous voice-based CMC from three dyads: Spanish learner of English-Spanish learner of English, Spanish learner of English-Turkish teacher of English, and Spanish learner of English-American learner of Spanish. The study indicated that dyads with different L1s produced NoM and modified output significantly higher than dyads of the same L1. A few studies have also focused on NoM patterns between native speakers of English and EFL learners. Akayoglu and Altun (2009) involved EFL learners in Turkey and native speakers of English in negotiating by text chat through Tapped In, the online workplace of an international community, where teachers, professional development staff, university faculty, students and researchers gathered to learn, collaborate, share and support one another using online features as conversation transcripts automatically emailed, text-based chat or private messaging. These researchers found that the native speakers frequently used clarification requests, confirmation checks and elaboration requests while the 124 Language Learning & Technology learners often utilized reply confirmation, confirmation and reply elaboration. Shim (2007) analyzed NoM of a NEST with EFL Korean, Chinese, Colombian and Brazilian learners in Daedalus Interchange chats and face-to-face modes. The study revealed that the teacher used confirmation more than other signals in both communication modes. The teacher often decoded learner messages in face-to-face exchanges while elaborating on her understanding of what the learners had said. The interest in CMC learning has grown recently in Vietnam as a number of Vietnamese learners found this mode novel and motivating, as compared with traditional teacher-fronted classes, and importantly, what they liked most was the freedom to share ideas (Nguyen, 2011). Pham et al. (2014) investigated the frequency and patterns of interaction of Vietnamese learners and instructors from Vietnam and other countries in asynchronous CMC forums. Using descriptive statistics and content analysis based on Salmon’s (2003) model, they showed that communication forums enhanced learners’ English communication with instructors and participation in online interactions. Besides that, instructors paid more attention to guiding and motivating the learners to interact. To sum up, research has explored NoM in various settings and participants; but a paucity of research has compared NoM between teachers of different L1 backgrounds and EFL learners. To gain further insights into the difference in interaction between them and to provide further empirical evidence for more rigorous interaction, the current study examined the NoM between NESTs versus NNESTs and Vietnamese EFL (VEFL) learners in online voice synchronous CMC. It seeks to further understand how teachers create facilitative conditions of SLA in an online environment. Methods Interaction analysis and database The current study was mainly based on the analysis of a database of 30 interaction sessions video- recorded, including 15 NEST-VEFL learner dyads and 15 NNEST-VEFL learner dyads. All the thirty VEFL learners were working adults with the same basic level of proficiency as measured by a proficiency test administered by the company. The 15 NESTs consisted of males (12) and females (3) from three English speaking countries: England (7), the US (5) and Australia (3). They all have university degrees and TESOL certificates. The 15 NNESTs consisted of males (5) and females (10) from the Philippines (8), Ukraina (1), Poland (1) and Serbia (5). They all also have obtained university degrees and TESOL certificates. The general method of this study is computer-mediated discourse analysis (CMDA), which was first proposed by Herring (1995; cited in Herring, 2001). As cited in Herring (2001), interactive exchanges can be the subject of CMDA, including analysis of logs of verbal interaction such as characters, words, utterances, exchanges, etc. The basic methodological orientation of CMDA is qualitative in that it makes observations of discourse phenomena and quantitative in that those phenomena are coded and counted and summaries of their relative frequencies are produced. Each recording was transcribed and analyzed quantitatively and qualitatively to answer the two research questions. The coding and counting approach was adopted using the taxonomy described above. To ensure the reliability, one colleague in the field of TESOL was asked to re-code the interactions based on the coding taxonomy. The inter-coder reliability was tabulated using SPSS statistical package; the reliability was established by using inter-coders of a sample of the data. Procedures To create the database of video recordings, the first author obtained permission from her online education company, where she worked as a teaching assistant. As an assistant teacher, the researcher was allowed to record each lesson session for writing evaluative reports to the company. The participants were made aware of the video recording activity as part of the company’s work for quality appraisal. In this way, the recordings were taken for interaction analysis, and to ensure anonymity, each teacher was given a code which was then used in reporting the data (e.g, NEST1, NNEST1). Each online lesson lasted 45 minutes and usually followed a format. The lesson started with a warm-up interaction between the teacher and learners, then peer interaction occurred among learners, followed Pham Kim Chi & Nguyen Van Loi 125 by pronunciation practice, and finally teacher-learner interaction which lasted around five minutes each. These teacher-learner interaction sessions formed the database for the study. Coding procedures NoM functions were coded, using the taxonomy described above. Examples of each type of NoM are provided below. Table 2. NoM coding scheme with examples Functions Examples Clarification request H2: I want to know how much is the fee? NNEST2: for? for what? Confirmation * H3: in February fourteen twenty eighteen NNEST3: so it is in February, February fourteen Confirmation check H1: my skin is dull and I want to make it brighter NNEST1: you want your skin to make it brighter? H1: yes Correction or self- correction L12: features, features NEST12: that’s OK and features Elaboration NEST8: you can learn on the Internet L8: yes, internet goes to any corner of the world, every field, economy, education, art Elaboration request * H6: I think Ha Long Bay is the most beautiful place in my country NNEST6: Ha Long Bay? and where is that? can you explain? H6: I know Ha Long Bay is the famous place in the world Reply clarification NNEST14: maybe some bubble wrap H14: bubble wrap what? NNEST14: it is a packing material and it prevents damages, packing company uses it for their devices Reply confirmation NNEST6: Eiffel Tower? H6: yes, I think it is attractive because structure, architect a long time Reply elaboration * NNEST6: and do you visit some other parts of French or only Paris? H6: *Functions of NoM added by Akayoglu and Altun (2009) In order to ensure the reliability of coding, 20% of the entire data set was coded by two independent coders. The coders agreed on 90% of their coding, suggesting that the data were coded with strong consistency. Data Analysis In order to answer the first question, the interactions of 30 pairs of participants were analyzed and coded for NoM functions. The Conversation Analysis transcription was based on Atkinson & Heritage (1984). The codes were then quantified by counting out of the total words of the teachers, the total number of NoMs per 100 words, and the total number of each type of NoM per 100 words. The data was entered into SPSS for statistical analysis. The differences between the two groups were statistically analyzed by using Mann-Whitney U test because the assumption of a normal distribution was not met in this small sample. The comparison is made in the number of words, NoM functions per 100 words and specific NoM functions. The differences in interaction quality was further examined and illustrated by qualitative interaction analysis. Results First, the distributions of NoMs and specific NoM functions are analyzed for NESTs, NNESTs, and VEFL learners. Then quantitative comparisons between NESTs and NNESTs are presented. Finally, a qualitative analysis of learner output is conducted to analyze how NEST and NNEST NoM input affects 126 Language Learning & Technology learner output. NoM functions used in NEST and NNEST groups The NESTs produced far more words than the VEFL learners and also produced NoMs at a higher rate. Out of 4,083 total words produced across the 15 sessions, the NESTs produced 2501 words, while the learners produced 1582 words. Among 210 NoM instances across 15 sessions, the teachers used 3.38 NoMs per hundred words, while the learners used 1.76. As Table 3 reveals, four dominant NoM functions were observed among the NESTs. These functions included elaboration request (1.03 per 100 words), confirmation (0.86 per 100 words), elaboration (0.76 per 100 words), and correction (0.39 per 100 words). In contrast, functions dominant among the learners were reply elaboration (0.86 per 100 words), reply confirmation (0.27 per 100 words) and elaboration (0.24 per 100 words). Five negotiation functions in the framework were not observed, namely comprehension check, reply clarification, reply comprehension, reply vocabulary and vocabulary request. Table 3. Frequency of NoM and Ratio of NoM per 100 words in both dyads Function NESTs-VEFL NNESTs-VEFL Words = 4083; NoM = 210 Words = 4997; NoM = 239 NEST VEFL NNEST VEFL Freq Ratio Freq Ratio Freq Ratio Freq Ratio Elaboration request 42 1.03 3 0.10 64 1.28 3 0.06 Confirmation 35 0.86 8 0.20 35 0.70 8 0.16 Elaboration 31 0.76 10 0.24 7 0.14 2 0.04 Reply elaboration 4 0.10 35 0.86 1 0.02 41 0.82 Correction or self-correction 16 0.39 0 0.00 5 0.10 0 0.00 Reply confirmation 4 0.10 11 0.27 9 0.18 14 0.28 Confirmation check 5 0.12 4 0.10 17 0.34 5 0.10 Clarification request 1 0.02 0 0.00 10 0.20 6 0.12 Comprehension check 0 0.00 0 0.00 0 0.00 0 0.00 Reply clarification 0 0.00 0 0.00 6 0.12 6 0.12 Reply comprehension 0 0.00 0 0.00 0 0.00 0 0.00 Reply vocabulary 0 0.00 0 0.00 0 0.00 0 0.00 Vocabulary request 0 0.00 0 0.00 0 0.00 0 0.00 In other words, on average, the NESTs used more NoMs than the VEFL learners. They often utilized elaboration requests, confirmation, elaboration, and correction. In return, the learners used reply elaboration, reply confirmation and elaboration more frequently, and elaboration requests and confirmation checking less often. In contrast, the NNESTs and VEFL learners produced a similar number of words across the 15 sessions. Out of 4997 words, the teachers produced 2636 words and the learners did 2361 words. There were 239 NoM moves out of 15 sessions among which the learners were engaged in negotiation at a ratio of 1.70 NoMs per hundered words, whereas the NNESTs engaged at a rate of 3.08 NoMs per hundered words. Table 3 further reveals nine specific NoM functions the NNESTs were engaged in. The ones frequently used were elaboration requests (1.28 per hundred words), confirmation (0.70 per hundred words), confirmation checking (0.34 per hundred words), and clarification requests (0.20 per hundred words). The less frequent ones involved elaboration (0.14 per hundred words), reply elaboration (0.12 per hundred words), and correction (0.10 per hundred words). Like the NESTs, the NNESTs used elaboration requests most frequently. This means that reply elaboration, which is the corresponding response function, was most frequent in the learners (0.82 per hundred words); followed by reply confirmation (0.28 per hundred words) in return to confirmation checking by the NNESTs. The learners rarely used confirmation checking, clarification requests and reply clarification. They also seldom asked Pham Kim Chi & Nguyen Van Loi 127 for elaboration (0.06) and elaborate on their ideas (0.04). Comprehension checks, reply comprehension, reply vocabulary and vocabulary requests were not used. In a word, the data indicated that like the NNESTs, the NNESTs made use of more negotiation functions than the VEFL learners in their interactions. They preferred elaboration requests, confirmation, and confirmation check respectively, while the VEFL learners frequently used reply elaboration and reply confirmation; and this difference will be discussed in a later section. Quantitative differences between NESTs and NNESTs Table 4. Frequency of NoM functions per 100 words compared between NESTs and NNESTs Functions NESTs NoM Rate NNESTs NoM Rate Elaboration request 1.03 1.28 Confirmation 0.86 0.70 Elaboration 0.76 0.14 Reply elaboration 0.10 0.02 Correction or self-correction 0.39 0.10 Reply confirmation 0.10 0.18 Confirmation check 0.12 0.34 Clarification request 0.02 0.20 Reply clarification 0.00 0.12 Total 3.38 3.08 A Mann-Whitney U test was used to compare the number of words between NESTs and NNESTs, VEFL learners (with NESTs) and VEFL learners (with NNESTs), NESTs and their VEFL learners, NNESTs and their VEFL learners. The result showed that the number of words uttered by NESTs and NNESTs was not significantly different (p > .05). However, the number of words produced by the learners with NESTs and the learners with NNESTs showed a significant difference (p = .021). The learners interacting with NNESTs produced more words than those interacting with NESTs across the fifteen sessions. The NESTs produced far more words than their VEFL learners (p = .022), while the NNESTs and their learners showed no difference in this respect (p = .191). Table 5. Differences in NoM functions used by NESTs and NNESTs ER Conf Elab Relab Cor RConf ConC ClarR RClar Mann - Whitney U test 108.5 110.5 64.5 90 88 84 55 52 82.5 Asymp. Sig. (2- tailed) 0.867 .932 .031 .148 .209 .153 .009 .002 .035 As further shown in Table 4 and Table 5, among the nine NoM functions produced by NESTs and NNESTs, significant differences were observed in the use frequency of elaboration (p = .031), confirmation checking (p = .009), clarification requests (p = .002), and reply clarification (p = .035). The NNESTs used more clarification requests (ClarR), reply clarification (Rclar), and confirmation check (ConC), whereas the NESTs used elaboration (Elab) more often, and this difference will be discussed in a later section. Quantitative differences between VEFL learners with NESTs and NNESTs The Mann-Whitney U test of the NoM differences between VEFL learners when interacting with NESTs and NNESTS (Table 6) further revealed significant results in elaboration (p =.046) and reply clarification (p = .007). The learners interacting with NNESTs used more elaboration while the learners interacting with NESTs used more reply clarification. Table 6. Differences in NoM functions used by VEFL learners in the two groups of dyads ER Conf Elab Relab RConf ConC ClarR Rclar Mann - Whitney U Test 92 95.5 74 104.5 97 105.5 90 67.5 128 Language Learning & Technology Asymp. Sig. (2-tailed) .189 .408 .046 .733 .485 .694 .073 .007 Qualitative evidence Elaboration, confirmation check, clarification request, and reply clarification were significantly different between NESTs and NNESTs. Elaboration and reply clarification were also significantly different between VEFL learners in dyads with NESTs and with NNEST. Therefore, these functions are taken further for the qualitative analysis. Elaboration As indicated, the NESTs used elaboration more frequently, and this happened when the learners failed to provide sufficiently comprehensible information. Although elaboration with explanations and examples is useful to generate comprehensible input for learners, it is likely to limit the time for learner production. The following excerpt from Session 8 in a NEST-VEFL learner dyad showed that the NEST could have blocked opportunities for the learners to respond and move the talk forward. NEST8: ok, do you know any disadvantages of learning online? L8: it is advantages of learning online - it is ((pause)) NEST8: [ok L8: [I join - the forum there and I study the group ((pause))= NEST8: =ok, very good, you also talk about advantages. I will give you some advantages. As I said before I want to tight a tie on Youtube and having videos, if I do not remember something, I can go back and watch the video again. I can also see the video if I want to, and I can pause or start to follow steps, step by step and do slowly to make sure that everything I do correctly. And one disadvantage or maybe students are shy speaking English, learning on the Internet will help build up confidence, but they don’t have experience talking to someone, one on one or face to face, learn online, participate online when they talk to actual person, they are very shy because there are a lot of difference when I am talking to you here and talking to you in person across to me. A big difference and if you are shy, face to face interaction, you will be very shy, ok, very good job Despite providing rich comprehensible input, NEST8, an American teacher appeared to block the opportunity for L8 to practice output by taking the floor, instead of asking prompt questions to elicit the learner to speak further. Also, after providing the input, the teacher neglected to check whether the learner understood it or not. The teacher’s elaboration strategy, in this case, unintentionally prevented the learner’s elaboration, which seemed to be limited to “it is advantages of learning online - it is ((pause))” or “I join - the forum there and I study the group ((pause))”. However, elaboration was more frequently used by VEFL due to elaboration prompting from NNESTs, leading to more production. As the excerpt below from Session 3 illustrates, a Filipino teacher maintained the conversation by encouraging H3 to elaborate. NNEST3: =where do you celebrate Tet holiday?= H3: =yeah family celebrate the Tet holiday because it is the most important holiday in my country, yes family so fun and happy when Tet holiday is coming= NNEST3: =what about traditional food? Do you cook traditional food during the holiday? Does your family cook traditional food?= H3: =my mother every year, my mother cooks traditional cake such as Chung cake or Tet cake and nowadays, people can eat in the restaurant so many traditional food and after eating yes ((pause)) NNEST3: go on? after eating?= H3: =after eating, we can visit relatives, sometimes you visit pagodas, we have happy new year The learner, H3, elaborated more on Tet celebration, traditional food, and activities following the Filipino teacher’s elaboration requests “where do you celebrate Tet holiday?”, what about traditional food? Do you cook traditional food during the holiday? Does your family cook traditional food?, go on? Pham Kim Chi & Nguyen Van Loi 129 after eating?” In this case, these questions gave the learner the opportunity to elaborate and produce more output. Confirmation check Confirmation check helps to ensure conversational exchanges are comprehensible, and as used by the teacher together with other functions like elaboration requests, also pushes learners to elaborate on what they have said. The following excerpt from Session 9 in a NNEST-VEFL learner dyad illustrates that the NNEST9, a Filipino teacher, combined confirmation checking and elaboration requests to encourage the learner to speak more. NNEST9: so when you use your camera, what do you like to photograph? You have any subject?= H9: =subject, no, I like ah ((pause)) I do not have subject, I just like taking photos and video= NNEST9: =ah ok yes, you just take photos of anything, right?= H9: =yes= NNEST9: =uh huh so do you have a lot of photos and video in your phone?= H9: =sorry?= NNEST9: =I think you have a lot of photos in your phones now?= H9: =yes, yes I have about three thousands pictures= NNEST9: =oh that’s a lot The teacher confirmed the information “yes, you take photos of anything, right?” and upon hearing the learner’s response, maintained interaction with an elaboration request “do you have a lot of photos and video in your phone?” In this case, the teacher pushed the learner to produce more output. Unlike the Filipino teacher in the preceding excerpt, the conversation in Session 9 below demonstrates the different way a NEST used confirmation checking followed by elaboration to negotiate with a learner. L9: yes, ((pause)) when my children have been vaccinated, they always have got a fever, I am so worry= NEST9: =ah, ok, so ah he still has a fever?= L9: =yes= NEST9: =but he has been vaccinated?= L9: =yes NEST9: =ok, perhaps, you should take him to the doctor and check out what the problem is= L9: =yes, I take my children to the doctor, the doctor has got it very uh it very- easy, he will uh - they will –uh- [they will= NEST9: =ok, you can take him to the doctor and maybe give him some medication. If he had vaccine, it should be ok The teacher confirmed the information “but he has been vaccinated?” then upon hearing the learner’s confirmation, elaborated “ok, perhaps, you should take him to the doctor and check out what the problem is” or “ok, you can take him to the doctor and maybe give him some medication. If he had vaccine, it should be ok,” rather than using other functions to encourage the learner to speak. Therefore, if the learner had no ideas for discussion, the interaction would stop. Clarification requests Clarification requests prevent misunderstanding and create smooth exchanges in authentic interactions. Below is an example from Session 9, a dialogue between a Ukrainian teacher and a VEFL learner. H2: I want to know how much is the fee?= 130 Language Learning & Technology NNEST2: =for? for what?= H2: =the fee for luggage, I want to check my baggage on the train so I want to know how the fee is. How much the fee is? NNEST2: =what is the fee? you can say how much does it cost? to check the baggage. It is 10 dollars for one suitcase, but you have 2 suitcases, it will cost you 15 dollars for everything In the conversation, the learner was allowed to provide better explanations, making her message clearer by saying “the fee for luggage, I want to check my baggage on the train so I want to know how the fee is.” Followed by “for? for what?” a request from NNEST2. On the other hand, the interaction between NEST15, an American teacher and L15 in the following excerpt from Session 15 seemed less successful in getting the message across. NEST15: =of course, you can pay by visa, so would you like some cookies or anything with you? or you just want to pay now = L15: = uh I want to pay by visa= NEST15: =ok ok sure, so I will bring the bill, here is the bill? does it look correct? the bill? L15: ((pause)) no NEST15: no? = L15: =yes, I want nine cookies bring home= NEST15: =ok, so I give you the bill can you pay 50,000 Vietnamese dong?= L15: = uh, fifty thousand? yes= NEST15: =yeah, fifty thousand= L15: =yes, it is correct exactly When the NEST15 asked “does it look correct?” the learner answered “no” and the NEST15 requested for clarification by asking “no?” with a raising intonation. However, the learner seemed not to understand the request and continued with “I want nine cookies bring home”, understood as an answer to the previous question “would you like some cookies or anything with you?” Although the teacher used a clarification request, the interaction was not as productive as with the case of NNEST-VEFL learner illustrated above. Reply clarification Reply clarification is to elucidate a previous statement. When learners fail to make sense of something, they may use this function, which may elicit the use of reply clarification from the teacher. In the following case from Session 14, in NNEST-VEFL learner dyad, the negotiation led the learner to ask for clarification to which the teacher replied with an explanation. Such negotiation provides further comprehensible input to the learner. NNEST14: ok, this is a good idea, and put them around the device, ((pause)) right?= H14: [yes NNEST14: =[maybe some bubble wrap H14: [bubble wrap what?= NNEST14: =it is a packing material that prevents damages, so I think bubble wrap some packing company uses it for their devices, good thank you very much= H14: =thank you teacher Evidently, H14 failed to know what the bubble wrap is for, which led to her to getting to know the meaning. The NNEST14, a Serbian teacher, explained it with a clear context which made the input comprehensible to the learner who then confirmed her comprehension with “thank you teacher”. The data of the study showed that throughout 15 sessions, reply clarification was not present in NEST-VEFL Pham Kim Chi & Nguyen Van Loi 131 learner dyads. This may be because the learners in these dyads did not have much chance to negotiate further. The frequent request for clarification by NNESTs also made a significant difference in their learners’ reply clarification compared with the VEFL learners in the NEST dyads, as illustrated in the excerpt from Session 2 in a NNEST-VEFL learner dyad. H2: I want to know how much is the fee?= NNEST2: =for? for what?= H2: =the fee for luggage, I want to check my baggage on the train so I want to know how the fee is. How much the fee is? By replying to clarification requests, the learner had another chance to negotiate for meaning, which could prompt more thoughts and longer sentences. H2 seemingly produced more extended output upon the request, giving the context and explaining why she wanted to ask the question: “the fee for luggage, I want to check my baggage on the train so I want to know how the fee is. How much the fee is?” In this case, the NNEST, an Ukrainian teacher failed to understand the learner or intentionally used a clarification request to encourage the learner to elaborate. To sum up, both quantitative and qualitative data indicate that there were marked differences in the patterns of NoM functions and the quality of interactions. The NESTs tended to provide more extended input than the NNESTs instead of further prompting learners to speak, and thereby unintentionally restricted the VEFL learners’ opportunities to talk. This is confirmed by the outweighing number of words produced by NESTs compared to that of their learners (p =.022) and the frequent use of elaboration. In contrast, the NNESTs tended to create more opportunities for their learners to produce more output by using confirmation checking, clarification requests apart from other functions. Discussion and Implications In this study, VEFL learners participated in online voice CMC interactions with either a NEST or an NNEST. The use of NoMs by the learners, NESTs, and NNESTs was analyzed both quantitatively and qualitatively. Significant differences were observed in the use of NoM functions by NESTs and NNESTs which impacted learner output. The results revealed that elaboration request was most frequently and equally used by both NESTs and NNESTs, and this function seems to be effective in encouraging learners’ production. This finding is consistent with previous studies (Akayoglu & Altun, 2009; Samadi & Noordin, 2014). In terms of online learning environment, voice interactions seem to be more effective than text-based or email interactions and was preferred by learners (Patterson and Trabaldo, 2006). This raises a crucial question regarding the conditions for effective interactions EFL learners can have when choosing their interlocutors and learning environment. In addition, the result analysis indicated that NESTs and NNESTs used elaboration differently. The fact that the NESTs used more elaboration than the NNESTs is consistent with what Kitade (2006) found. Nevertheless, the current study found that too much teacher elaboration apparently reduces learners’ output practice. Instead of asking the learners to elaborate more on their responses or clarify their ideas, the NESTs tended to negotiate meaning by giving detailed explanations or examples. In contrast, the NNESTs used more confirmation checking, clarification requests, or elaboration requests to prompt further interaction and output. While provision of comprehensible input is essential to improve learners’ listening, output practice also contributes to language acquisition (Swain, 1985). It is crucial that teachers be aware of balancing these opportunities to create the most facilitative environment especially for adult learners who seek opportunities for speaking practice online. The current study demonstrates the crucial role of teachers in online voice CMC in supporting EFL learners’ learning through creating quality interactions. Learner involvement in NoM is only successful as long as learner output is considered as a key factor in the online environment, especially in cases where learners have rare opportunities to practice speaking with English-speaking teachers. The online learners in this study produced output in negotiating with teachers in both types of dyads, but the learners 132 Language Learning & Technology produced more output in interaction with the NNESTs than with NESTs. Such evidence is consistent with Lee (2001) who found that requesting clarification is the most salient negotiation function and with Jepson (2005) who found that it encourages the speaker to elaborate ideas on and sustain interactions. This result provides further evidence to reconsider the stereotype that NNESTs are inferior to NESTs, and that in terms of negotiation functions in CMC, an awareness of the role and purpose of interaction could be vital in supporting EFL learners to develop proficiency. It would be more effective if teachers of English, regardless of native status, are made aware of deploying interactions not only to provide comprehensible input and corrective feedback, but also to encourage and push learners to produce output. Despite the small sample of the participants, the findings of the present study have several implications for the research on the relationship between the online education industry, teachers, and learners in terms of second language acquisition. First and foremost, the results of this study suggest that online voice CMC interactions could be an ideal environment for EFL learners to practice the target language. The increasing availability and affordability of online chat technology will continue to create conditions that facilitate EFL learning. Evidence from this study suggests that voice chat synchronous CMC interactions offer a promising environment where learners can participate in meaning negotiation. This online environment seems to be supportive and sufficient to EFL learners in participating in distance learning without spatial barriers; the learners worked independently and learned actively with online teachers who did not see their faces during the interaction. This could create a secure learning environment compared to face-to-face interaction which poses face threatening risks for low proficiency learners. However, it might have contributed to misunderstanding which possibly accounts for the differences in patterns of NoM functions. Second, the EFL learners benefit from both NESTs and NNESTs in different ways. Evidence from this study showed that they received more comprehensible input from NESTs and produced more output in negotiating with NNESTs. In this respect, the status of nativeness is no longer crucial; rather it is vital to be aware of how to negotiate with learners as well as the purpose of online interaction sessions. However, this tentative conclusion needs corroborating evidence from future research with larger sample sizes. In particular, clarification requests, confirmation checks, or any functions which could move the dialogue forward, should be employed to produce learner output. Online learning undeniably brings foreign language learners in closer contact with expert users of English which increases the likelihood of exposure to rich input, but spending a lot of time to provide elaboration limits opportunities for learners who need to practice extended discourse in order to improve their speaking proficiency. From significant differences in learners’ use of NoM functions, this research suggests that VEFL learners could benefit more from input when interacting with NESTs whereas they would have more comprehensible output with NNESTS. Conclusion This study attempted to examine the use of negotiation of meaning functions in dyads with NESTs and NNESTs and compare the interactions between them, as well on their impact on learner production. The data confirmed that the teachers are generally effective in negotiating with the learners and the VEFL learners benefited from voice chats with both NESTs or NNESTs. However, the VEFL learners had restricted opportunities to talk when they interacted with NESTs due to NEST overuse of elaboration. On the other hand, NNESTs used more negotiation functions that encouraged learners to talk more. The differences in the number of words VEFL learners produced and the negotiation highlight important implications for the online teaching industry. The study further emphasizes the effectiveness of synchronous online voice CMC interactions in promoting VEFL learners involvement in conversations. It also helps to strengthen the important position of NNESTs in supporting EFL learners in CMC interactions. The study has two major limitations. The majority of NNESTs were Filipino, so the qualitative data was not maximally representative of NNESTs. Also, future research based on a larger database and different learner levels of proficiency would probably produce more insights into how online interactions are handled and how they could benefit second language learners. Pham Kim Chi & Nguyen Van Loi 133 Acknowledgements We thank the participants of this study. We especially thank the reviewers for insightful feedback, and the editors of Language Learning and Technology for their constructive comments and suggestions on editing. References Akayoglu, S., & Altun, A. (2009). The functions of negotiation of meaning in text-based CMC. In R. de Cassia Veiga Marriott & P. Lupion Torres (Eds.), Handbook of research on E-learning methodologies for language acquisition (pp. 291–306). Hershey, PA: Information Science Reference. Atkinson, J. M., & Heritage, J. (Eds.) (1984). Structures of Social Action: Studies in Conversation Analysis. Cambridge: Cambridge University Press. Bueno, M. C. (2012). Focus on form and negotiation of meaning in synchronous voice-based computer mediated communication: Effect of dyad. Procedia - Social and Behavioral Sciences, 34, 39–44. Chiu, T., Liou, H., & Yeh, Y. (2007). A study of web-based oral activities enhanced by Automatic Speech Recognition for EFL college learning. Computer Assisted Language Learning, 20, 209– 233. Ellis, R. (2003). Task-based language learning and teaching. Oxford: Oxford University Press. Fernández-García, M., & Martínez-Arbelaiz, A. (2002). Negotiation of meaning in non-native Speaker-non-native speaker synchronous discussions. CALICO Journal, 19(2), 279–224. Herring, S. C. (2001). Computer-mediated discourse. In D. Tannen, D. Schiffrin, & H. Hamilton (Eds.), Handbook of Discourse Analysis (pp. 612–634). Oxford: Blackwell. Jepson, K., 2005. Conversations and negotiated interaction in text and voice chat rooms. Language Learning & Technology, 9(3), 79–98. Kachru, B. B. (1997). World Englishes and English-using communities. Annual Review of Applied Linguistics, 17, 66–87. Kern, R. (1995). Restructuring classroom interaction with networked computers: Effects on quantity and characteristics of language production. Modern Language Journal, 79(4), 457–476. Kitade, K. (2006). The negotiation model in asynchronous computer-mediated communication (CMC): Negotiation in task-based email exchanges. CALICO Journal, 23(2), 319–348. Kötter, M. (2003). Negotiation of meaning and codeswitching in online tandems. Language Learning and Technology, 7(2), 145-172 Lee, L. (2003). Negotiation of meaning and codeswitching in online tandems. Language Learning & Technology, 7(2), 145–172. Lee, L. (2001). Online interaction: Negotiation of meaning and strategies used among learners of Spanish. ReCALL, 13(2), 232–244. Long, M. H. (1981). Input, interaction, second-language acquisition. Annals of the New York Academy of Sciences 379(1), 259–278. https://doi.org/10.1111/j.1749-6632.1981.tb42014.x Long, M. H. (1983). Native speaker/non-native speaker conversation and the negotiation of comprehensible input. Applied Linguistics, 4(2), 126–141. Long, M. H. (1996). The role of the linguistic environment in second language acquisition. In W. C. Ritchie and T. K. Bathia (Eds.), Handbook of research on second language (pp.413–468). San Diego, CA: Academic Press. Medgyes, P. (1992). Native or non-native: Who’s worth more? ELT Journal, 46(4), 340–349. 134 Language Learning & Technology Nguyen, L. V. (2011). Learners’ reflections on and perceptions of computer mediated communication in a language classroom: A Vietnamese perspective. Australasian Journal of Educational Technology, 27(8), 1413–1436. Olaniran, B. A. (1996). A model of group satisfaction in computer-mediated communication and face- to-face meetings. Behavior and Information Technology, 15(1), 24–36. Pacek, D. (2005). “Personality not nationality”: Foreign students’ perceptions of a non-native speaker lecturer of English at a British university. In E. Llurda (Ed.), Non-native language teacher: Perceptions, challenges and contributions to the profession (pp. 243–262). New York, NY: Springer. Patterson, P. J. (2001). Computer-assisted language learning: An analysis of discourse produced in computer-assisted and oral class discussions by Spanish learners. Ph.D. thesis, The University of Texas at Austin. Patterson, P. & Trabaldo, S. (2006). Negotiating for meaning across borders with CMC. Teaching English with Technology, 6(2), 1–15. Retrieved from https://www.tewtjournal.org/issues/past- issue-2006/past-issue-2006-issue-2/ Pham, T., Thalathoti, V. & Dakich, E. (2014). Frequency and Pattern of Learner-Instructor Interaction in an Online English Language Learning Environment in Vietnam. Australasian Journal of Educational Technology, 30(6), 686–698. Pica, T. (1994). Research on negotiation: What does it reveal about second language learning conditions, processes, and outcomes? Language Learning, 44(3), 493–527. Pica, T., & Doughty, C. (1985). Input and interaction in the communicative language classroom: A comparison of teacher-fronted and group activities. In S. M. Gass, & C. G. Madden (Eds.), Input in second language acquisition (pp. 115-132). Rowley, MA: Newbury House. Samani, E. & Noordin, N. (2014). Discourse analysis of students’ interactions in terms of negotiation of meaning in the Malaysian context. Middle-East Journal of Scientific Research, 19, 112–118. Retrieved from http://www.idosi.org/mejsr/mejsr19%28icmrp%2914.htm Seidlhofer, B. (2004). Research perspectives on teaching English as a lingua franca. Annual Review of Applied Linguistics, 24, 209–239. Shim, Young-Sook. (2007). Negotiation of meaning between an L2 teacher and student in face-to-face interactions and CMC. English Teaching, 62 (3), 265–288. Smith, B. (2003). Computer-mediated negotiated interaction: An expanded model. The Modern Language Journal, 87(1), 38–58. Smith, B. (2005). The relationship between negotiated interaction, learner uptake, and lexical acquisition in task-based computer-mediated communication. TESOL Quarterly, 39(1), 33–58. Snow, M., Kamhi-Stein, L., & Brinton, D. (2006). Teacher training for English as a lingua franca. Annual Review of Applied Linguistics, 26, 261–281. Swain, M. (1985). Communicative competence: Some roles of comprehensible input and comprehensible output in its development. In S. Gass and C. Madden (Eds.), Input in second language acquisition (pp. 235–253). Rowley, MA: Newbury House. Tarone, E. (1980). Communication strategies, foreigner talks, and repair in interlanguage. Language Learning, 30, 417–431. Terhune, N.M. (2013). Learning to learn digitally: Getting students on the road to autonomy. International Journal of Computer-Assisted Language Learning and Teaching (IJCALLT), 3(4), 9–24. Varonis, E. M., & Gass, S. (1985). Non-native/non-native conversations: A model for negotiation of Pham Kim Chi & Nguyen Van Loi 135 meaning. Applied Linguistics, 6(1), 71–90. Walkinshaw, I., & Duong, O. T. H. (2014). Native and Non-Native English Language Teachers: Student Perceptions in Vietnam and Japan. SAGE Open, 4(2), 1–9. Warschauer, M. (1996). Comparing face-to-face and electronic discussion in the second language classroom. CALICO Journal, 13(2), 7–26. Warschauer, M. (2003). Technology and social inclusion: Rethinking the digital divide. Cambridge, MA: MIT Press. Zhao, Y. (2005). Research in technology and second language education: Developments and directions. Greenwich, CT: Information Age Publishing About the Authors Pham Kim Chi is currently an English lecturer in the Department of English at FPT University, Vietnam. Her primary research interest falls in English language teaching and learning, interactions and technology. Email: chipk@fe.edu.vn Nguyen Van Loi is currently a senior lecturer and teacher educator in English Language Teaching at Can Tho University. His areas of research involve methodology in teaching and learning English, language teacher education and cognition, and second language acquisition. Email: loinguyen@ctu.edu.vn