Negotiation of meaning in Chinese non-native speaker–Spanish native speaker communication in assessment context

The article describes the characteristics of the negotiation of meaning (NOM) in oral interaction between native Spanish speakers and Chinese immigrant speakers of Spanish by assessing the context through 15 recordings, in which the candidates orally interacted with their interviewers. The goals of the research were to observe whether, in this context, NOM occurs; identify and classify the most frequent negotiation strategies of meaning in terms of their function; determine whether reactive negotiation strategies or preventive strategies predominate; establish which of the two interlocutors leads the negotiation; and determine whether this negotiation is effective. It has been observed that NOM is present in 11.21% of the total number of conversational turns (n = 4,379). The most frequently used strategies were confirmation check, comprehension check response, and confirmation. Of the strategies used (n = 491) by participants and interviewers to negotiate the meaning in this context, 64.95% was of a preventive nature. The NOM was effective in 90.02% of the cases.


Negotiation of meaning in oral interaction
Many sociocultural approaches consider language development as something inter-mental, indissolubly linked to social interaction. The authors supporting these approaches maintain that learning cannot be acquired in isolation (Foster & Snyder Ohta 2005). Indeed, interactionists of the second-language acquisition (SLA) not only emphasise the importance of the input comprehension (something that would coincide with the ideas of Krashen 1982Krashen , 1985 but also assure that it can be modified to be comprehensible if initially it is not (Baralo 1999) and, therefore, achieved through adjustments in the interaction (Long 1996).
Based on previous authors who had studied the processes of learning and first-language acquisition, such as Vygotsky (1978), Wood et al. (1976), and Bruner (1978), Long (1981Long ( , 1996 developed his Theory of Interaction, in which the negotiation of meaning (NOM) plays an important role.
(2) Which kind of task is more efficient for achieving a greater degree of interaction and negotiation (Blake 2000, Hardy & Moore 2004, Jeong 2011? (3) Which kind of group of informants is the most appropriate in this type of evidence (Oliver 2002, Blake & Zyzic, 2003, Lee 2004, Sotillo 2005, Jeong 2011, Kawaguchi & Ma 2012? (4) Which are the most frequent and/or effective types of strategies for NOM in SLA (Fernández-García & Martínez-Arbelaiz 2002, Kötter, 2003, Foster & Snyder Ohta 2005, Smith 2005, Bower & Kawaguchi 2011)? (5) The specific features of certain groups of students, depending on the factors such as their age (Oliver 2002, Morris 2005 or their nationality (Abdullah 2011, Yi & Sun 2013. Framed within this theoretical context, this study is intended to contribute to the teaching of Spanish as a foreign language, clarifying the way by which the NOM between the NSs and the learners of Spanish as a language of migration takes place.
Those who learn Spanish to use it in the controlled context of the classroom often need to face communicative situations that require a greater knowledge of the language than the one they possess. That is why strategies for NOM will be very useful to compensate gaps in language knowledge and to repair communication problems, especially when the level of linguistic competence of the NNS is low.
The importance of the NOM has been highlighted by the Common European Framework of Reference for Languages (CEFR; Council of Europe 2001)on which the most recent Spanish teaching curricula are basedwhich includes this strategy within the strategic competence that is part of the communicative competence. The CEFR approach is action oriented for students to learn coping with different communicative tasks. Therefore, it emphasises the importance of sociolinguistics and pragmatics as well as their associated competences in addition to the linguistic content (Council of Europe 2001).
In addition, when oral interaction occurs in a context of assessment of language use, Amador (2010), Doquin & Martín Leralta (2014), Baralo et al. (2016), among others considered that the training of teachers and examiners is essential for the correct development of the exam, since the role of the interviewer during the OEI test can affect the results. For this reason, the results of this study are also intended to contribute to facilitating the work of teachers and examiners by providing them with data on the production of Chinese-speaking learners of Spanish and the way in which they interact with Spanish NS. To do so, the study examines NOM in the OEI test of the examination for the Lengua Española para Trabajadores Inmigrantes (LETRA) diploma (Spanish Language for Immigrant Workers).

Our study
This study describes the characteristics of the NOM in oral interaction between NSs of Spanish and Chinese-speaking learners of initial Spanish level in the context of the certification assessment, focusing on the OEI test of the LETRA diploma promoted by Antonio de Nebrija Foundation and the Directorate-General of Immigration of the Community of Madrid (Baralo & Estaire 2011).
We opted for this student profile due to the lack of studies on Chinese speakers who study Spanish as a language of migration in a context of linguistic immersion, particularly from the point of view of NOM and assessment. In addition, there is a widespread perception that Chinese candidates have communication problems with the interviewer and that they negotiate meaning poorly, so we strongly believe that empirical data are needed to support or disprove this assessment.
2 Methodology 2.1 Research questions 1. Is there NOM between the NS and the non-native Chinese speaker in the context of the OEI test for the LETRA diploma? 2. If yes, which type of communication strategies is the most common in this context? 3. If there is NOM in this context, are the strategies related to either reactive or preventive acts? 4. If there is NOM, who takes the lead in this context, the interviewer or the candidate? 5. If there is NOM between the interviewer and the Chinese-speaking candidate in this context, is it effective?
The first research question analyses whether there is an NOM in the selected samples. The second research question explores the frequency of variable used of each NOM strategy included in the study: clarification request, clarification, confirmation check, confirmation, comprehension check, comprehension check response, elaboration request, elaboration, and overt indication of lack of linguistic knowledge (OILK). The definition of these strategies for NOM can be found in Annex 1.
In order to answer the third research question, we need to clarify that reactive strategies are those used after a mistake or lack of understanding in communication; on the other hand, preventive strategies are those used without any previous mistake or lack of understanding in communication.
To answer the third research question, type 2 strategy used as a response to a type 1 strategy will not be taken into account since it does not appear as a direct consequence of a communication problem, but it is related to another strategy that has been previously used to solve a problem. The number of times that the strategy of OILK has been reactive or preventive will be quantified.
Then, to address the fourth research question, the interlocutor who initiates more frequently the NOM will be considered the leader, discarding from the repertoire of identified strategies those that do not initiate negotiation but continue it (as in the case of the third research question). As for the fifth and final research question, the strategy for NOM is considered the most effective, which allows for the maintenance of communication without making changes in subject matter after attempting negotiation, thereby avoiding unintended subject changes due to lack of understanding, situations in which lengthy silences appear after attempting the NOM, situations in which the NNS fails to understand the interviewer despite the latter having attempted to negotiate the meaning, or situations in which the NNS answers something that is not being asked.

Research design
This study aims to contribute to the research projects IN.MIGRA2_CM (H2015/HUM-3404) and EMILIA (FFI2017-83166-C2-2-R) by compiling data and analysing the oral linguistic corpus of learners of Spanish as a language of migration Nebrija-INMIGRA. To do so, we carried out a descriptive study based on a corpus of oral productions of 15 candidates of Chinese origin for LETRA diploma, 10 women and 5 men, in interaction with their respective native or very competent Spanish interviewers. These two groups will be represented with the labels PAR (participant, candidate) and INT (interviewer) in the transcripts and in the analysis of the results.
A mixed methodology is used in order to initially quantify and compare the frequency of use of each type of strategies for the NOM, promoting a focused descriptive qualitative study. This way, the procedure is carried out as follows: 1. Drawing up a list of the ten types of strategies for the NOM that will be analysed in the samples. They are identified by the following abbreviations: Overt indication of lack of linguistic knowledge OILK 2. Compilation of the strategies for NOM present in the oral interactions included in the transcripts and further data processing in Excel. 3. Identification, classification, and count of the situations of NOM used by interviewers and participants in the transcripts according to the type of strategy. 4. Classification of the strategies for NOM used by interviewers and participants according to their reactive or preventive nature. 5. Identification and classification of the strategies for NOM initiated by the interviewer or the participant. 6. Identification and classification of the strategies for NOM used by interviewers and participants that have been effective and ineffective. 7. Analysis of results to detect the differences between informants and/or types of strategies.

Profile of study participants
The study participants consisted of 15 candidates and their respective LETRA exam interviewers. In the LETRA exam, candidates must perform three tasks related to oral interaction: an interview on personal information, training, and job; a description of an image related to a profession/job and a subsequent conversation on some aspects related to that profession/job; and finally a conversation about an image related to the administrative-labour field and with administrative documents (a mock test is accessible at www.diplomaletra.com). Candidates are expected to reach an A2-n level, slightly lower than an A2 level of the CEFR (Council of Europe 2001), with "n" standing for grammatical competence (González & Rosario Guerra 2010).
As has been reported in other studies on the linguistic acquisition by migrants, it can be difficult to obtain their personal information, as they are often reluctant to provide their data. In our study, we obtained data about the age of 12, the sex of 14, the time spent in Spain of 12, the mother tongue and languages learned of 15, the time they spent learning Spanish of 12, and the academic training and current occupation of 11 participants.
The profile of the 15 candidates is homogeneous in terms of their mother tongue and preferred language for communication, but heterogeneous with respect to other individual variables that might affect their level of oral communicative competence in speaking Spanish. Their period of residence in Spain ranges between 1 and 18 years, with an average of 6.6 years spent in the country (SD = 5.6). Great differences have also been observed in the length of time they have been learning Spanish, which varies between 1 and 11 years, with an average of 4 years of language learning (SD = 3.7).
Of the informants, 6 were men and 8 were women, who were aged between 16 and 49, with an average of 31.3 years (SD = 12.4). Six report having university-level studies, four claim to have other higher education studies, three have completed their compulsory education, and four did not provide any data at all on their academic training. All the participants passed the oral interaction test of the LETRA exam, except for two participants: PAR10, the youngest participant, who has spent the least amount of time in Spain and in learning the language, although she has passed the exam, thanks to her performance in the rest of the tests (reading comprehension, written expression, and audiovisual comprehension); and PAR15, a candidate who lives in Spain and has been learning the language for 2 years. The average score obtained in the oral interaction test is 6.44 (SD = 1.63). The interviewers consisted of 8 men and 7 women who were aged between 30 and 62 (mean 39.43, SD = 8.39), all with university degrees and professional experience in teaching Spanish to immigrants (five of them:  INT1, INT3, INT4, INT5, and INT8), social care for immigrants (seven of them: INT6, INT7, INT9, INT11, INT12,  INT13, and INT14), or intercultural mediation in migrant support organisations (three of them: INT2, INT10, and INT15). All have passed a compulsory training course of 100 h in order to be able to participate as evaluators of the LETRA diploma, which includes a section on training in communication strategies for the oral test aimed at reducing the variability between them when it comes to the oral test (Doquin & Martín Leralta 2014).
The duration and speaking turns of each oral interaction are given in Table 1.

Corpus transcription and compilation
The productions were transcribed in the CHAT format (Codes for the Human Analysis of Transcripts) using the CLAN program

Results
Through the available samples, we found examples for each NOM strategy included in the study (see examples in Annex 2), and all the research questions formulated for the analysis could be answered by examining these data.

Answer to research question 1
Interviewers (INT) used a total of 282 negotiation strategies, while participants (PAR) used a total frequency of 209, which constitutes a total of 491 negotiation strategies between interviewers and participants, for a total of 163.37 min of recording time.
Observing only the strategies produced by the interviewers, in most cases (73%) between 10 and 25 strategies are produced, a result that differs from that obtained by observing only the frequency with which participants use negotiation strategies, wherein in most cases (64%) between 5 and 15 strategies are found.
The length of the recordings used as a sample for this research ranges from 7.12 to 15.17 min, 10.89 being the total average (SD = 2.35).
In addition to the length of the recordings, the number of times that the interlocutors have used strategies for NOM based on the number of speaking turns has been taken into consideration.
Taking into account the total number of speaking turns that take place in the whole sample (4,379 turns) and the average number of speaking turns that occur per conversation (291.93 turns, SD = 65.82), we infer that there is an average of 32.73 strategies per 291.93 turns (SD = 14.20), which means that approximately 11.21% of the speaking turns is actually strategies for NOM.

Answer to research question 2
The most common strategies for NOM in the analysed corpus are confirmation check (CON1), comprehension check (COM2), and confirmation (CON2) (see Table 2) The most frequent NOM strategy used by all participants and interviewers is the confirmation check (CON1), with a frequency of 129 of the 491, which constitutes 26.27% of the total negotiation strategies in the sample. This strategy is followed by the comprehension check response (COM1) and confirmation response (CON2), with a frequency of 106 and 100, respectively, which means 21.58% and 20.36% of the total NOM strategies found in the analysed recordings.
The least used strategy is the elaboration request (ELA1), with a frequency of 14 out of 491, i.e. 2.85% of the total strategies for NOM found in 163.27 min of the recordings.
When studying the results obtained by participants and interviewers separately, it is important to highlight that the strategy most frequently used by interviewers is the confirmation check (CON1), with a frequency of 102 of 282, which is the total amount of NOM strategies used by interviewers (36.17% of these), while the strategy most frequently used by participants is the confirmation (CON2), with a frequency of 81 of 209, which is the total amount of NOM strategies used by participants (38.75% of the total number of NOM strategies produced by participants).
When comparing these results to those given in Table 2, it is noticeable that the most used NOM strategies by both interviewers and participants do not coincide with the most used strategies when they are analysed separately, as can be seen in Figure 1.
Regarding the frequency of use of each strategy, the following graphs show the information corresponding to the candidates and the interviewers, respectively.  The relationship between the use of different strategies by each interlocutor can be better understood from Figures 2 and 3. Thus, for example, the greatest occurrence of the CON1 strategy by interviewers corresponds to that of CON2 by candidates. The cases in which a significantly higher use of a strategy is observed in a specific interaction are discussed in Section 4.

Answer to research question 3
Preventive strategies have a total frequency of 215 of the total amount of strategies (331), which means that 64.95% of them have a preventive character and, therefore, most of the strategies included in this study have a preventive nature.
The data also provided information on the most frequent strategy of preventive nature in the case of interviewersthe comprehension check response (39.88%)and in the case of participantsthe OILK (61.90%). They also informed that the most frequent strategy of a reactive nature: the confirmation check in both cases (48.75% and 43.58%, respectively).

Answer to research question 4
The total amount of NOM strategies initiated by interviewers is noticeably higher than those initiated by participants, with only 23.56% of NOM situations initiated by participants, while 76.43% was initiated by interviewers.
A more detailed study of the data shows that the confirmation check strategy (CON1) is the preferred option used by interviewers to start negotiation (101 of 253), which means that the interviewer uses it 39.92% of the times to initiate the NOM. In the case of the participants, the most used strategy to initiate the NOM is OILK, with a frequency of 29 of the 78 negotiations initiated by this interlocutor, which means that participants used it to initiate the negotiation in 37.17% of the cases, followed by the strategy of comprehension check, with a frequency of 24 of 78, i.e. 30.76% of the cases.

Answer to research question 5
According to the obtained data, the fifth research question also received a positive response; 90.02% of the strategies that were used by participants and interviewers was effective, with a frequency of 442 of 491. These were effective 94.23% of the times when they were used by interviewers and 84.21% of the cases when the participants used them. A more detailed analysis of the data shows that the most effective strategy is the confirmation check (CON1), with a frequency of 120 of 491 (24.43% of all strategies), whereas the most ineffective is the OILK, with a frequency of 15 of 491, representing a 3.05% of the total ( Table 3).
By comparing the number of times the strategy has been used with the frequency in which it has been effective (by both participants and interviewers), it can be established that the most effective and most used strategy is the comprehension check response (COM2), since it is effective in 99.05% of the cases, with a frequency of 21 of 26.  At the other end is the least effective strategy in terms of the number of times it has been used: the OILK was used only by participants and effective in only 48.27% of cases (14 of 29), followed by clarification (CLA2), which was effective in 61.90% of cases (13 of 21).

Discussion
In Samani et al. (2015), the percentage of each NNS-NNS conversation aimed to negotiate the meaning was set at an average of 2.10 of the conversation (526 words of a total of 24,980 were used to negotiate meaning); while in Bower & Kawaguchi (2011), the number of speaking turns aimed to negotiate meaning consists of 20% in the case of conversations of the NS-NNS type (Americans who study Japanese) and 7.2% in the case of conversations of the NS-NNS type (Japanese English students). Therefore, it can be confirmed that the results obtained by Bower & Kawaguchi (2011) in the case of Japanese students are similar to those obtained in this study in which 11.21% of the speaking turns are NOM strategies. However, they do not match with those of Samani et al. (2015) and Bower & Kawaguchi (2011) in the case of American students, with whom the NOM occurs more frequently.
The aforementioned studies use samples taken from synchronous conversations mediated through written messages on the Internet, which could explain the differences, as these authors point out.
As for the most frequent negotiation strategies in communication between NS and NNS, Samani et al.'s (2015) results differed from those of this study. They analysed the NOM strategies used by 14 Malay, Chinese, and Indian students while speaking to each other in English (NS-NNS) through synchronous chat communication. While for Samani et al. (2015) the most common strategies were confirmation (26.61% of the total NOM strategies used), elaboration (21.48%), and confirmation check (15.78%), closely followed by elaboration request (15.59%); the strategies most used by the participants in our research are confirmation (CON2), with 38.75% of the strategies used by this group followed by OILK, 13.87% and confirmation check (CON1), 12.91%. Therefore, confirmation (CON2) and confirmation check (CON1) are included in both studies as one of the three most used strategies by learners, but elaboration (ELA2) occupies the fifth position in our study. The OILK was not included among the variables in Samani et al.'s study (2015), although it could be related to the strategies of vocabulary request and vocabulary check. On the other hand, it is noticeable that the comprehension check response has never been used in the work by Samani et al. (2015), while in our study this strategy occupies the fourth position and are used by 11.96% students. This might be due to the fact that sometimes the comprehension check response is a repetition of what the interviewer has previously said; therefore, since in this case it is a chat-mediated interaction, the use of this type of strategy is somehow limited.
Abdullah (2011) also studied the most frequent negotiation strategies, with Chinese and Indonesian students learning English. These students had face-to-face conversations in English; and after analysing the results, it was concluded that the most frequent strategy in conversations was other repetitions, with 27.39% of the total amount of strategies used; confirmation check, 26.09% and self-repetition, 15.22%. The results obtained by Abdullah (2011) differ from ours, as the only coincidence is that confirmation check is one of the most frequent strategies used by the participants. However, it should be taken into account that the variables used in our study and that of Abdullah (2011) are different, as some strategies are not included in one of the two studies or are grouped differently, so the comparisons are inaccurate. Nevertheless, the author's reflection on the use of repetition stands out because, based on Chen (1990), he states that repetition (a strategy that has not been considered independently in the study) is characteristic of Chinese speakers because they perceive it as polite, although in our study it is infrequent.
On their part, Foster and Snyder Ohta (2005) analysed the oral production in class of English and Japanese learners from different backgrounds and found that some participants did not use any NOM strategy at the time when their oral production was examined. More specifically, 5 Japanese students of a total of 19 and 6 English students of a total of 20 did not use any NOM strategy. In our study, all participants use strategies for NOM but even though the results do match with those of Foster & Snyder Ohta (2005), not all participants use the same strategies and they never apply all of them.
Concerning the reactive or preventive nature of strategies for NOM in communication between NS and NNS, Nassaji (2010), who studies attention spontaneously produced (unplanned, so NOM is included), obtained similar results; preventive acts (69%) are more frequent than reactive ones (31%) and preventive ones are initiated by the teacher in most cases. Nassaji (2010) noted that this may be due to an attempt on the teacher's side to anticipate problems that students might encounter in communication, a consideration that could also apply to the context of our study.
Regarding leadership of NOM in communication between NS and NNS, certain similarities with Nassaji's study (2010) can be traced: 60.5% of NOM is initiated by the teacher, while only 39.5% is initiated by students. Nassaji (2010) attributed these percentages to the authority role played by the teacher in the classroom, which could also explain why the same thing happens in the context of assessment, as observed in this research. Lee (2004) also mentioned that it is the NS who leads the conversation in his research, which studies the interaction between Spanish NS and American NNS who communicate through the Internet. In this study, students' opinions about this type of interaction are collected; and some mention that they feel more intimidated when interacting with NS than when interacting with NNS. Bearing these statements in mind, one might think that perhaps in the LETRA exam participants may feel self-conscious or afraid of not knowing how to negotiate meaning on the grounds that they feel pressured speaking to a native and also being evaluated.
It is noteworthy that the OILK is used frequently by participants, as it is the most widely used despite the assessment context (which differs from what Abdullah, 2010, found). This would contradict the general conception of Chinese learners not overtly expressing incomprehension.
Regarding the effectiveness of strategies for NOM in the communication between NS and NNS, Bower & Kawaguchi (2011) studied the conversation in English and Japanese between Japanese and Australians in eTandem sessions through chat, and they established that in the English exchanges 87.90% of the negotiations is effective, with 86.60% effective NOM negotiations initiated by NS and 88.90% initiated by NNS. In the case of Japanese exchanges, the percentage of effective NOM negotiations is even higher, i.e. 91.50 (95.20% of them initiated by NNS and 98.50% initiated by NNS). Both results coincide with those obtained in our study, in which 90.02% of the strategies used were effective -94.23% in the case of NS (interviewers) and 84.21% in the case of NNS (participants).
Based on profile information of our study participants, the only individual variable that seems to be related to an increased use of some strategies for NOM is the linguistic proficiency of non-native candidates. Thus, in the interviews of the only two participants who did not pass the oral interaction test, a considerably higher number of CON1 strategies were used by the interviewers (22 by INT10 and 24 by INT15 compared to an average of 4.3 by the rest of the interviewers) and CON2 by the candidates (18 in PAR10 and 14 in PAR15 compared to an average of 3.7 by the rest of the candidates). Likewise, the only candidate (PAR13) with the highest score in the test (score 10) is the one who employed the NOM strategies the least.
The variable length of stay in Spain in the case of the candidates, the age and sex of candidates and interviewers, and the occupation of the interviewers do not seem to be related to the number and type of NOM strategies that were employed. Regarding the variable length of stay in Spain, it is noteworthy that some of the longest established candidates in Spain are those who have most recently begun learning the language (PAR5, 14 years and PAR6, 11 years, with only 1 year since they began attending Spanish classes). This could explain why the strategies for NOM they use to talk to natives are similar between both long-and short-standing migrants in Spain, as many of the former have not maintained any regular communicative contact with the natives for most of their time in the arrival country.

Conclusions
In the light of the collected data, we can affirm that there is NOM between the interviewer and the Chinese speaker candidate in the context of assessment of Spanish as a language of migration, and that it is present in 11.21% of the total number of speaking turns with an average of 32.73 negotiation strategies in each test. This NOM is carried out by both interlocutors although it is the interviewer who leads it, initiating the negotiation in 76.43% of the cases found in the sample. The NOM produced is described by analysing the most frequently used strategies, including confirmation check, comprehension check response, and confirmation, which constitute 26.27%, 21.58%, and 20.36%, respectively, of all the NOM strategies used by the participants and interviewers alike; 64.95% of the strategies used by all of them to negotiate meaning in this context was of preventive nature. As for the effectiveness of the NOM carried out, it has been observed that it is effective in 90.02% of the casesin 94.23% of the occasions the interviewers used negotiation strategies, while the participants used them in 84.21% of the cases.
Although the results of our study do not appear to have been influenced by age, length of residence in Spain, or educational level, it would be interesting, in the light of the conclusions of Mavrou & Martín Leralta (2018), to increase the number of samples to further study how age, time of residence in Spain, and educational level influence the way in which Chinese students who take the OEI test for the LETRA exam negotiate meaning with their interviewer.
This research provides a classification of strategies for the study of NOM that may be useful for the scientific community, given that there is no unanimity concerning the strategies that should be included and, in some cases, they are not clearly justified and/or defined.
According to previous studies, the NOM between two or more NNS is more abundant than that between NS and NNS; open interview and image description tasks do not promote NOM, and the assessment context can be quite an influential factor as it affects attention, level of anxiety, and willingness to communicate. However, bearing in mind that the purpose of this research is to describe how the NOM between Spanish NSs and Chinese students of initial level occurs in such a specific context as the certification assessment of the level of Spanish, it has to be understood that the conditions for the NOM cannot be modified; otherwise, it would be impossible to achieve the intended objective.
In view of the obtained results, we suggest replicating this study to compare the results with other instances in which the students' origin is different or in which the type of task and context change.
To conclude, we would like to underline the importance of NOM as a communicative strategy and a key component of strategic competence. We support that NOM should be included within the curriculum of language teaching and of the qualification criteria of any test that intends to evaluate the competence in OEI in second languages, since it enables the student to understand and be understood by others. The study of the characteristics of negotiation according to the different student profiles will allow teachers to meet the demands of each profile, since they will be able to offer student support in the areas they need. Taking as reference Long (1980: 82-3 in Foster & Snyder Ohta 2005, these strategies are expressed by one of the interlocutors to ask for clarification on something the other has said. It may include expressions that explicitly state that the message has not been understood (such as "I don't understand it"), yes or no questions, or pre-arranged questions that cause the interlocutor to provide new information, add information that has already been provided and/or express himself/herself in other words. Samani et al. (2015:21) define the "reply clarification" strategy as one that serves to "clarifying his or her previous statement as a result of request (clarification request)." In the case of this research, clarifications will be considered to be that which Samani et al. (2015) have defined as a "reply clarification" and also those clarifications that are given without having been preceded by a clarification request, in which case one could speak of preventive negotiation strategies, which are used not when the conversation has broken down but when one of the interlocutors believes that what has been said previously may not be clear to the other and decides to anticipate the problem.

Confirmation check CON1
Considering Long's contributions (1980: 81-2 in Foster & Snyder Ohta 2005, these strategies are expressed by one of the speakers immediately after the message issued by his interlocutor to ensure that the message has been correctly heard and understood. Confirmation checks are usually rising intonation questions that include the repetition of part or the entire message delivered by the NNS. They can be answered with simple answers such as "yes" or "mhmm" and do not require the NNS to provide new information.

CON2
They are used by any of the interlocutors (sender) to let the other (receiver) know that he has understood correctly the message that the first wanted to transmit. Samani et al. (2015: 21) defined "reply confirmation" as a function "confirming a statement when someone requests confirmation like 'yes,' 'ok,' 'you're right'." Comprehension check COM1 Based on Long (1980: 82 in Foster & Snyder Ohta 2005, it can be said that this type of strategy is any expression emitted to know the message has been understood by the interlocutor. Among the comprehension checks, there are common prefixed questions such as "OK" or "do you understand?", repetitions of the message emitted by the same NS or part of it with different intonation.
Comprehension check response

COM2
Although commonly associated directly with the confirmation check (Abdullah 2011, Samani et al. 2015, this study has considered that it does not necessarily have to be preceded by this strategy but can occur independently when one wants to let his or her interlocutor know that his or her previous utterance has been understood correctly. Among the expressions that have been considered comprehension check responses are "other repetitions" (partial, total, or modified) that serve the purpose described above even though the repetitions were understood by other authors (Oliver 2002, Kötter 2003, Abdullah 2011) as independent strategies. The "mhmm" interjection should not always be seen as a confirmation check response as it may serve to support the speaking turn.
Elaboration request ELA1 As defined by Samani et al. (2015: 21), this type of strategy refers to the request for a more complete response if one does not have a clear idea of what the interlocutor has previously said and therefore additional information is needed.

ELA2
Following Samani et al. (2015), this type of strategy is understood as the production of a statement that has been preceded by one that has not provided all the sufficient information or that has not made the message clear. The elaboration provides additional information about a previous utterance and can be preceded by an elaboration request; by another strategy stating that the earlier message has been perceived as incomplete; or even as a preventive strategy when the interlocutor detects that his previous message might not be understood and decides to anticipate himself or herself in order to avoid interrupting the conversation.
Overt indication of lack of linguistic knowledge OILK This strategy has been included in the study as a result of adapting and merging two different strategies present in other studies. These are the overt indication of understanding (Kötter 2003) and the non-verbal indication of understanding (Abdullah 2011). Through the overt indication of lack of linguistic knowledge, one of the interlocutors (usually the NNS) lets the other know that he or she has not understood the previous utterance or the previous reference because of language issues and not for other reasons (such as knowledge of the world). It is also used by the NNS when he or she wishes to answer in a certain way but fails due to a lack of linguistic knowledge. Some common expressions are "I don't know how to say it," "How do you say…?" or "I don't know what that is." Oral interaction among immigrants and native speakers  91 (continued on next page)