The Contribution of Negotiation of Meaning to Language Accuracy in an EFL Setting through a Focused Task

Despite the extensive research on the negotiation of meaning (NoM) using unfocused tasks, to our knowledge no study portrays to what extent NoM contributes to learners’ language accuracy, grammar, and vocabulary, in particular, using a focused task. Therefore, this research set out to obtain in-depth information about the contribution of NoM to language accuracy when learners were assigned to engage in a focused task discussion. The participants, treated as three dyads based on their English proficiency, were taken purposely from one class of the sixth-semester students in the Department of English Education in the Faculty of Education at Universitas Lampung, Indonesia. They were paired based on their English abilities (high-high; high-low; low-low) and their intimates among the 24 students in the class. All their utterances produced during a fifteen-minute focused task discussion were recorded and analyzed following the theory of NoM and language accuracy. It was discovered that the learners were engaged in NoM during the focused task discussion, and it contributes to the students’ language accuracy. Therefore, to optimize the contribution of NoM to language accuracy, the topic to be discussed should be based on the learners’ output (focused task) since the interlocutor can offer assistance as part of the negotiation of meaning process. The suggestions for future research are also discussed. * Corresponding author, email: nainggolan.flora@yahoo.com Citation in APA style: Flora, Mahpul, & Sukirlan, M. (2021). The contribution of negotiation of meaning to language accuracy in an EFL setting through a focused task. Studies in English Language and Education, 8(1), 294-312. Received August 8, 2020; Revised December 12, 2020; Accepted December 15, 2020; Published Online January 3, 2021 https://doi.org/10.24815/siele.v8i1.17667 Flora, Mahpul & M. Sukirlan, The contribution of negotiation of meaning to language accuracy in an EFL setting through a focused task | 295


INTRODUCTION
Over the past few years, negotiation of meaning (NoM) has gained popularity in the field of a second language or foreign language education. Negotiation of meaning, which takes place while learners are involved in communication in the target language, is believed to be able to help them with their language development, particularly their language accuracy (Watanabe & Swain, 2008). Thus, there has been a steady increase in the number of analytical studies involving NoM. For instance, in face-to-face manners, when the native speaker (NS) and non-native speaker (NSS) interacted (Kitajima, 2009;Pica, 1994) when teachers interacted with students (Ko et al., 2003;Pica & Doughty, 1985) when NSSs interacted with fellow NSSs (Azkarai & Agirre, 2016;Baharun et al., 2018;Farangis, 2013;Nurazizah et al., 2018;Yufrizal, 2007), and via messenger (Arslanyilmaz & Pedersen, 2010;Bower & Kawaguchi, 2011;Castrillo et al., 2011;Luciana, 2005;Yazigi & Seedhouse, 2005;Zhu & Carless, 2018). The learners do not care too much about linguistic problems (Baharun et al., 2018;Nurazizah et al., 2018), even they sometimes use their native language if they have problems expressing their ideas in the target language (i.e., English). This kind of negotiation will not help learners with their language accuracy. The unfocused task, the task that is not bound to certain linguistic elements, is widely investigated by these previous studies. It might be one of the factors that makes a little contribution of NoM to learning outcomes, especially to linguistic problems, such as grammar and vocabulary.
Therefore, despite the extensive research, to our knowledge no study portrays to what extent negotiation of meaning contributes to learners' language accuracy, grammar, and vocabulary, in particular, using a focused task which was based on Swain's (1985) theory, Ellis' (1991) and Swain and Lapkin's (1995) statements about output hypothesis. This theory of output hypothesis implies that when a learner utters or produces expressions or sentences, she/he is just trying her/his hypothesis on the knowledge she/he has had in the target language. If the listener knows that there is an error made by the learner, who is the speaker, she/he will make a necessary correction on the utterance. In this way, the learner can also get necessary input from her/his utterances (output) and, therefore, there will be some development of her/his language accuracy. However, if there is no correction, the learner will possibly conclude that what she/he has uttered is errorless, even from the grammatical perspective, which makes room for fossilization (Swain, 1985). Since it was of great interest to the researchers to provide proof that would establish its contribution, it was of paramount importance for the researchers to carry out this research to specifically put to the test the contribution of NoM to language accuracy for learners. This is based on Bourke's (2008) statement that by noticing the errors, the learner can improve their language accuracy.
Besides, Hartono and Ihsan (2017) suggest that further researchers examine the contribution of NoM to language accuracy. Since far too little attention has been devoted to this matter, it is also reasonable to take this research as an effort to confirm whether or not it is a conundrum. That is why this research focuses on the contribution of NoM to language accuracy in the English Foreign Language (EFL) context, from the viewpoint of the language ability level: high-high dyads (both learners have good ability in English), high-low dyads (one has good ability and one has the low ability in English), and low-low dyads (both have the low ability in English). It is reasonable to extrapolate Kitajima's (2009) finding on the majority of Japanese, which is those with various English ability to engage in NoM steps while being interviewed by native speakers. Watanabe and Swain (2008) affirm that the more proficient student dominates the interaction of a pair, the contribution of the less proficient one is rather dim. To raise the language accuracy of learners, it is advisable to group learners based on their English ability (Bourke, 2008;Hartono, 2017;Luciana, 2005). Taking this into account, the researchers also looked deeply into the NoM done by learners with different levels of English ability.
To sum up, having looked into all the studies above, this study was conducted for the following reasons: 1) the generalizability of the published research on this issue is limited. None of the research clearly describes the contribution of NoM to language accuracy, particularly, grammar and vocabulary while the learners with different levels of English ability are engaging themselves in a discussion as a focused task, and 2) all the prior research had an unfocused task as their instruments failed to spotlight the contribution of NoM to language accuracy. Therefore, in this research, a focused task (linguistics error found during the unfocused task) is used as a topic for the discussion. This is based on the output hypothesis theory proposed by Swain (1985), and Swain and Lapkin (1995), and language awareness proposed by Bourke (2008). With all this in mind, the researchers put forward the following research questions: To what extent does NoM contribute to language accuracy when high-high dyads, high-low dyads, and low-low dyads are engaged in a focused task discussion? Specifically, the research questions are formulated as follows: 1. What types of negotiation of meaning do the high-high dyads, high-low dyads, and low-low dyads produce when they are engaged in a focused task discussion? 2. What is the contribution of negotiation of meaning to language accuracy when the high-high dyads, high-low dyads, and low-low dyads are engaged in a focused task discussion?

LITERATURE REVIEW
In this section, the theoretical concept of NoM and controversial issues on NoM in language development are described. Finally, a theory of designing a focused task for language accuracy is also elaborated. Pica and Doughty (1985) and Pica (1994) stated that during the discussion, the learners do some efforts so that the purpose of communication can be achieved well. To put it differently, they make NoM such as comprehension check, clarification request, and confirmation check (repetition, modification or correction, completion, or elaboration). Comprehension check takes place when the speaker intends to make sure whether he is understood by the interlocutor, for example: Do you understand? Do you know the meaning of…? etc. Clarification request is the effort of the listeners to Flora, Mahpul & M. Sukirlan, The contribution of negotiation of meaning to language accuracy in an EFL setting through a focused task | 297 understand the speaker's utterances (Sorry, pardon me, I don't understand) and confirmation checks request for confirmation through repetition, modification, or correction, completion, or elaboration. Request for confirmation through repetition is done by repeating the utterances, either his own or his interlocutor. The function is to make sure they have a mutual understanding. Modification or correction is done if the listener notifies something wrong with his interlocutor's utterances while completion or elaboration is the listener notifies the speaker has a problem or makes a pause in expressing his idea. Through the process of NoM, the listener and speaker have the sensibility to recognize the difficulty faced by their interlocutor in understanding what their interlocutor means. Varonis andGass (1985, as cited in Yufrizal, 2007) have developed a model to resolve such non-understanding sequences (known as the negotiation of the meaning).

Resolution T I R RR
Notes: 1. Trigger (T) which invokes or stimulates incomplete understanding of the hearer (i.e., the utterance which causes misunderstanding), 2. Indicator (I), which is the listener's signal of incomplete understanding (i.e., of misunderstanding), 3. Response (R) is the original speaker's attempt to clear up the unaccepted-input, and 4. Reaction to Response (RR), which is an element that signals either the listener's acceptance or continued difficulty with the speaker's repair.
In this study, the stages above are summarized Figure 1. In short, this figure is used to analyze the NoM produced by each dyad during the discussion.

Figure 1. Negotiation of meaning
In conclusion, during communicative exchanges, the learners could restructure the conversation to make communication run well.

Controversial Issues on Negotiation of Meaning
There have been several longitudinal studies involving NoM in learning English as a second or foreign language. However, the previous researchers have a different opinion on it, especially their contributions to learners' language development. NoM occurring during learners' interaction can help them with their language acquisition development (Gass, 1997;Long, 1996). In line with this, Swain (1985) and Swain and Lapkin (1995) affirm that learners can also acquire comprehensible input from the output on the condition that there are corrections from the interlocutors. Besides,  Language and Education, 8(1), 294-312, 2021 Farangis (2013) and Samani et al. (2015) with their findings, confirm that NoM can be of considerable help in the improvement of second language acquisition. In short, NoM, which takes place while learners are being involved in communication in the target languages, is believed to be able to help the learners with their language development. However, this study did not give empirical data indicating the contribution of NoM to language accuracy.

| Studies in English
Conversely, Foster et al. (2005) deduce that even though NoM could help learners modify input for the better for their own sake, it is not a good way of expanding their second language acquisition (SLA). This statement was supported by Nurazizah et al. (2018) who state that the Malaysian learners use their native language (Malay) and gesture if they do not know how to express their ideas in English. In short, it is still questionable whether NoM can contribute to the quality of the learners' language, especially on language accuracy. Before this statement, Shortreed (1993), and Van der Branden (1997, as cited in Yufrizal, 2007 infer that the primary function of NoM relates to how learners make attempts to keep communication running and make sure the people engaged in it understand each other. In agreement with it, Castrillo et al. (2011), Azkarai and Agirre (2016), Hartono and Ihsan (2017), and Nurazizah et al. (2018) posit that NoM in second language acquisition is an attempt made by learners to cope with. Despite their linguistic errors, learners keep on being involved in communication as long as they understand what others say. By having different types of tasks, the learners in Malaysia did the same type of NoM when they interact with each other using English (Baharun et al., 2018). Similar to the previous studies, this study did not give empirical data indicating the contribution of NoM to language accuracy. It is safe to infer that researchers have not treated NoM in adequate detail as there has not yet been a finding proving that it can be wellinternalized by learners.

Designing a Focused Task for Language Accuracy
Referring to the previous findings and the instruments used by the researchers, most of whom use unfocused tasks, that is the task that is not bound to certain linguistic elements (see, for example, Abukhattala, 2013;Aksu & Gönül, 2014;Westera, 2011). These studies lead to the view that by giving communicative tasks, the students have time to practice the target language and they feel enjoy during the learning process. Farangis (2013) drew the inference that, in task-based activities, learners are very active and have more chances to use the target language, and it will make positive impacts on their understanding of the L2 acquisition development. Side by side to it, Mahpul and Oliver (2018) advocate task complexity by stating that it affects learners' utterances, in this context, positively. Those who suggest that communicative task such as task-based language teaching (TBLT) is needed in learning English. Mahpul and Oliver (2018) and Kanoksilapatham and Suranakkharin (2019) state that task-based learning activity was able to improve Thai learners' English ability.
In line with this, in the present study, the focused task was formulated based on the linguistic problems (grammar and vocabulary) coming to the surface while they were being engaged in the unfocused task (free topic). In other words, the task of this research is called a focused task. It was done based on the theory of Swain's (1985), Ellis' (1991), and Swain and Lapkin (1995) statements about the output hypothesis. This theory of output hypothesis implies that when a learner utters or produces expressions or sentences, she/he is just trying his hypothesis on the knowledge she/he has had in the target language. If the listener knows that there is an error made by the learner, who is the speaker, she/he will make a necessary correction on the utterance. In this way, the learner can also get necessary input from her/his utterances (output) and, therefore, there will be some development of her/his language accuracy. Pollard (2008) defined accuracy to refer to the correct use of language. In line with this, Ahangari and Barghi (2012) stated that accuracy is the ability to use the language correctly, and grammar instruction in any language teaching/learning program mainly aims at uplifting accuracy in learners for better communication. However, if there is no correction, the learner will possibly conclude that what he has just said is errorless even from the perspective of grammar, which makes some room for fossilization (Swain, 1985). This statement was supported by Richard and Rodgers (2001), who suggest that during the task, the learners are expected to foster processes of negotiation, in form of modification input and it is beneficial for second language learning and the learning involving communicative activities or task is better for learners (Ahmad & Rao, 2013).
To make the communicative task, there must be an information gap activity, one person has certain information that must be shared with others to solve a problem, gather information, or make decisions (Richards & Rodgers, 2001). In line with this, information gap can force students to negotiate the meaning because, during the discussion, the students have to make others understand (Arslanyilmaz & Pedersen, 2010;Azkarai & Agirre, 2016;Baharun et al., 2018;Bower & Kawaguchi, 2011;Castrillo et al., 2011;Farangis, 2013;Luciana, 2005;Nurazizah et al., 2018;Yazigi & Seedhouse, 2005;Zhu & Carless, 2018). Thus, in this study, the information gap task has been considered to find out the quality of students' utterances while performing the negotiation of meaning during the focused task.

Participants
The participants were treated as three dyads based on their English proficiency. They were taken purposely from one class of the sixth-semester students in the Department of English in the Faculty of Education at Universitas Lampung, Indonesia. These three dyads were paired based on their English abilities (the high-high learners, the high-low learners, and the low-low learners) and their intimates among the 24 students in the class. The high English ability is the learner who can express her/his ideas without too great of an effort. She/he searches for words occasionally but only one or two unnatural pauses, while the low ability has to search for the desired meaning. The listener can understand a lot of what is said, but she/he must constantly seek clarification, and cannot understand many of the speaker's more complex or longer sentences (Heaton, 1991). This was done based on the suggestion proposed by Luciana (2005), Bourke (2008), and Hartono and Ihsan (2017) who state that by noticing the errors, the learners can improve their language accuracy and the theory of output hypothesis (Swain, 1985;Swain & Lapkin, 1995) that the learners can get input from their output. The intimate of the students was gathered from the questionnaire (by asking them to write five persons they are close to) and the students' English ability was gathered from the lecturer of the Speaking class.

Research Instruments
The instrument of this research is a focused task. It was planned based on the students' linguistic problems (grammar and vocabulary) when they did a discussion on unfocused task (the students were given 24 personality traits and were asked to choose five most important ones which a teacher should have, and they were also asked to provide the reasons why they chose the personalities). In short, the focused task came from an unfocused task.

Data Collection Procedure and Data Analysis
The data were collected by the following procedures: 1. The linguistic problems (grammar and vocabulary) of the dyads at the unfocused task were written and informed to the whole dyads so that they also knew what the other dyads' linguistics problems were. Then, the dyads were asked to discuss the solutions to each other's problems. This is the topic for the focused task. 2. The whole utterances made by the students during a fifty-minute focused task discussion were recorded and transcribed. Based on the transcript of each dyad, the following steps were taken for analysis: 1. Respectively counting the C-units (an independent clause with its modifiers) 2. Coding the types of NoM (comprehension check, clarification request, and confirmation check: repetition, modification or correction, and completion or elaboration) all the dyads had done when they had the focused task discussion. 3. Counting the types of NoM (comprehension check, clarification request, and confirmation check: repetition, modification or correction, and completion or elaboration) all the dyads had done when they had the focused task discussion. 4. Analyzing which type of NoM (comprehension check, clarification request, and confirmation check (repetition), modification or correction, and completion or elaboration) had made the contribution to the language accuracy (grammar and vocabulary) for each dyad. This was done based on the recording transcript of their utterances of each dyad, contextually. 5. To have more valid data about the NoM's contribution to language accuracy, an interview was conducted. The process was recorded and the results were transcribed for further analysis.

RESULT
This study aimed to obtain NoM's contribution to language accuracy while the learners were discussing a focused task. In this research, the divisions of NoM were based on Pica and Doughty (1985) and Pica (1994), namely: comprehension check, clarification request, and confirmation check (repetition, modification or correction, and completion or elaboration). First, the number of NoMs are elaborated, then followed by each type of NoM and its contribution to language accuracy. To make it clear, it can be seen in Figure 2. Even though the length of time given was 15 minutes, they did not use this maximum time since they were able to finish the task in time. For example, H1-H2 dyads had 13.25 minutes, H-L dyads 10.25 minutes, and L1-L2 dyads 14.3 minutes. Based on Table 1, it can also be seen that C-units produced by H1-H2 dyads was 122, H-L dyads was 131, and L1-L2 dyads were 150 C-units. Comprehension checks were made at the focused task, but only by the H-L dyad. There were 11 Cunits (40%). While clarification requests were made by all the dyads, and the one making the most clarification requests were the L1-L2 dyad, with 27 C-units (18%), followed by the H-L dyad, with seven C-units (26%), and the H-H dyad, with one Cunit (5%). As for confirmation checks (repetitions) and modification or correction, the H1-H2 dyad made nine (33%), and th e L1-L2 dyad made five (19%), none was made by the H1-H2 dyad. The last one is the NoM in the form of completion or elaboration. Only the H1-H2 dyad did it. There were 18 C-units (15%). To have clear information about its contribution to learners' language accuracy, each type of NoM is elaborated in the next sub-sections.

Contribution of Negotiation of Meaning (NoM) to Language Accuracy
This section shows the result of each type of NoM that contributes to language accuracy. It is elaborated based on the types of negotiation: comprehension checks, 14.3 0 302 | Studies in English Language and Education, 8(1), 294-312, 2021 clarification requests, and confirmation checks (repetition, completion or elaboration, and modification or correction).

Contribution of comprehension checks to language accuracy
The result can be seen in Table 2 and Figure 3.   Figure 3.  Pica and Doughty (1985) and Pica (1994) stated that comprehension checks are the utterances made by a speaker to make sure whether what he has just said is understood by other interlocutors. At this focused task, the NoM produced by H-H dyads is 19, but none belongs to comprehension checks. The H-L dyads produced 13, and six (17%) belong to comprehension checks but it only gives a contribution of five (83%) to language accuracy. The comprehension checks produced by H-learner were mostly in the form of raising intonations, such as in the following extracts.

Extract (1) H : (Speaking Indonesian) Karena kan "kind" itu adjective, sedangkan kan rumus Simple Present
Tense adalah subject, verb, object. Jadi supaya bikin adjective-nya jadi verb, kita buatnya pake to be, to be nya...is, kan? [Because "kind" is an adjective and the simple present tense formula is "Subject + verb + object", to make the adjective a verb, we use a verb to be.  H-learner used the Indonesian language because she wanted her pair to understand. Besides, she felt comfortable to explain the grammar using her native language. This is in line with Nurazizah et al. (2018) who state that the Malaysian learners use their native language (Malay) and gesture if they do not know how to express their ideas in English.
During the NoM, 17% of comprehension checks made by H-learner does not give a contribution to the L-learner in this dyad, as in the following extract.
Extract (3) H : Because it is…actually…about the conditional sentence. Do you remember the rules of a conditional sentence? (Comprehension checks) L : Hmmm... H : Conditional type one, two, and three. Type 1, "if I am rich ... I will buy car". Type 2, "if I were rich, I would buy a car". Type 3, "If I had been rich. I would have bought a car". You get the point? (Comprehension checks) L : Hmmm. Okay…later...later I will learn about...aaa...about conditional sentence. Okay, the next is number 5.

Contribution of clarification request to language accuracy in each dyad
The following Table 3 and Figure 4 describe the contribution of clarification requests to language accuracy in each dyad. To have a clear illustration, Table 3 is illustrated in Figure 4. NoM in the form of clarification request, produced by H-H dyads is only one (5%), and it gives a contribution to language accuracy since it can be internalized or it provides comprehensible input for the learner, for example:

1) High-high dyad
Extract (4) H2 : I change the sentence as "because the teacher knows the class condition and students better". H1 : Can you repeat it? (Clarification request) H2 : …because the teacher knows the class condition and students better.
The function of the negotiation of meaning in the form of the clarification request was just to make sure that the conveyed message was already understood by the other interlocutor, not for the sake of language accuracy.

2) High-low dyad
As for the H-L dyad, the clarification request made was seven (26%) and six (86%), and give a contribution to L-learner. The L-learner asked for explanations of the wrong grammar since she did not understand why they were considered grammatical errors. The following extract is one of them.
Extract (5) H : How about the number 5? (Clarification request) L : Hmm...Because I can't the...eee I... I do not understand. (Clarification request) H : So, my answer is "if she or he is not patient". Actually, this is the same...you know...same form with the second eee…number two because this is also a simple present tense, yes. And, in this sentence, the adjective is patient. So, we need 'to be'. L : Hmmm...adjective...to be "is". H : Yes.
Only one (14%) of clarification request does not give a contribution to L-learner, as in the following extract.

3) Low-low dyad
In the last dyad (H-L), the clarification request was made six times (5%). The Llearner asked for explanations of the wrong grammar since she did not understand why they were considered grammatical errors. The following is one of them.

Contribution of confirmation checks (repetition, completion or elaboration, and modification or correction) to language accuracy in each dyad
This type of NoM also happened during the interaction. Its contribution to language accuracy can be seen in Table 4 and Figure 5.  Table 4 is illustrated in Figure 5. During the discussion, only H-H dyad made confirmation checks in the form of completion or elaboration, that is 18 (95%). All the completions or elaborations in this case did not relate to language accuracy but only to make the communication run well.
Extract (8) H1 : How about number one? "I think the teacher has to patient". H2 : Number one. I wrote here the right sentence "I think the teacher has to be patient". H1 : Umm...my sentence is "I think the teacher should be patient". H2 : Yeah, actually "has to" means the same as "should". (Elaboration) H1 : Yeah, "should "and "has to". We just need to add "be". (Elaboration) H2 : Iya, we need to add "be" because "patient" is adjective. In the H-L dyad, only H-learner made nine (33%) confirmation checks in the form of modification or correction. It gives seven (89%) contributions to L-learner's language accuracy (grammar). An example can be seen in Extract (9) In this dyad (H-L) one (11%) did not give a contribution to language accuracy, as the following extract.
Extract (10) H : If the teacher had been objective, the students would not get disappointed. This is the third type of conditional sentence. The last is the L-L dyad. In the H-L dyad, L-learner produced five (18%) confirmation checks in form of modification or correction. It gives only one (20%) contribution to L1-learner's language accuracy (grammar). One example can be seen in Extract (11).
In this dyad, there are four (80%) confirmation checks in the form of modification or correction that do not give a contribution to L1-learner's language accuracy (grammar). One example can be seen in Extract (12).
Extract (12) L2 : And then how about number 10 "The teacher can be evaluated the...weakness".

DISCUSSION
The students were very active and felt comfortable while they were engaged in the focused task. It can be seen from the recording and transcript. They helped each other and sometimes they laughed. The empirical data of the focused task differ from the statement of Watanabe and Swain (2008) which implies that H-learners dominate a discussion and disregard L-learners' opinions.
Comprehension checks, as one type of negotiation of meaning (Pica, 1994;Pica & Doughty, 1985), took place mostly at the H-L dyad because the H-learner would like to ascertain that her explanation had been understood well by the L-learner. From extracts (1) and (2), it is known that the H-learner made comprehension checks like a teacher, but the L-learner kept comfortable and relaxed. It is indicated in the dialogue transcript. The L-learner always responded to the comprehension checks made by the other interlocutor. Explicit explanations in the source language (Indonesian) and consciousness-raising factors, as Ellis (2002) states, took place through the negotiation in the form of comprehension checks and one of the acquisition processes described by Richards (2002), i.e., noticing. In this research, with the comprehension checks in the form of raising intonations, the student was able to understand that such an adjective needs a verb "to be" before it in the target language, whilst the source language, i.e., Indonesian, does not have this rule. Nor does it have a verb of to be.
Considering all this, it can be said with confidence that NoM in the form of comprehension checks at the focused task can contribute to learning outcomes, especially grammar knowledge for the H-L dyad. Besides, H-learner in this dyad has multiple roles, as a teacher and friend. This empirical data support Rahim and Chun (2017) who investigated the teacher's belief in teaching English (the teacher should be a friend and facilitator). In these dyads, the H-learner also gets benefits because she has time to practice her ability in explaining the grammar. It is good for her as a teacher in the future.
In this research, all the dyads made clarification checks. The dyad making the fewest clarification checks was the H-H one, with no more than one C-unit (5%). The function of the negotiation of meaning in the form of the clarification request was just to make sure that the conveyed message was already understood by the other interlocutor, not for the sake of language accuracy (see Extract 2).
As for the H-L dyad, the clarification request was made by H-learner. The Llearner asked for explanations of the wrong grammar since she did not understand why they were considered grammatical errors (see Extract 5). The sentence they discussed at the focused task, i.e. "If she or he not kind", was a wrong utterance by the interlocutor (L) while they were engaged in the unfocused task discussion. The Llearner perceived the sentence as correct since she had translated it word by word from the source language, i.e., Indonesian: if (jika), she or he (dia), not (tidak), kind (baik). So, the literal meaning is "jika dia tidak baik". This sentence is already right in Indonesian. Therefore, the L-learner thought "if she or he not kind" was already right. This sentence would not be considered ungrammatical if the corresponding formula of the source language were identical to that of the target language (Newmark, 1988).
Based on the transcription of the students' utterances and a brief interview, the comprehensible input (i+1), which is described by Krashen (1985Krashen ( , 1994Krashen ( , 2003, could come into existence. The empirical data of this research, i.e., the transcript of the students' dialogue, and the brief interview indicate that the student understood her pair's explanation and was able to generate the sentence "she is pretty" and explain it. This input could be internalized by the student because: 1) the wrong sentence was one she had generated (output hypothesis by Swain, 1985), 2) there was an explicit grammatical explanation (Ellis, 2002), and 3) there was enough consciousness. Ellis (2002) avers that one of the ways of raising learners' consciousness is to provide them with an experience of using the target language. In line with this, Richards (2002) concluded that a few factors are determining the change of input into the intake (complexity, saliency, frequency, and need) and the three language acquisition processes (noticing, comparing, and integrating) highlighted by Ellis (2002) occurred when the H-L dyad had the focused task discussion, and the empirical data confirms that negotiation of meaning contributes to learning outcomes, especially grammar knowledge. This finding is in line with Zhu and Carless (2018) who stated that written comments given by peer, allows the writer to negotiate the meaning. Of the eight errors, four are related to Indonesian-English translation, and the other four are related to the opposite. These eight expressions were produced by the L-learner during the unfocused task discussion. This learner requested some help from her pair either directly or indirectly, and the pair provided her with the right aid. This finding is in line with Flora et al. (2020) that if one did not know how to say something in English, he asked for help from his friends during the activity using the roundtable technique.
In contrast to this result, Kötter (2003), taking his finding into account, believes Germans rarely request or make a translation while interacting with their friends in the target language. They did code-switching more. It is inferred from the transcription of their utterances during the discussion that the expressions were understandable to the learner. It seems possible that it happened because the expressions had initially been her output (Swain 1985) and the lecturer informed her that the help provided by her pair would be asked about. Roughly speaking, the NoM in the form of the clarification requests at the focused task, which derived from the errors of the learner had, made a considerable contribution to the language accuracy from the viewpoints of both vocabulary and grammar. The results of this study are in line with Ellis (2002), Bourke (2008), and Gilakjani and Ahmadi (2011). They elaborated that during the learning process, the learners get involved in such cognitive processes as noticing, hypothesis testing, problem-solving, and restructuring. This present study shows that during the NoM, the learners identified the errors they made and tried to solve the problems. It is not baseless to say that the NoM was worthless from the perspective of the contribution to L-L dyads learning outcomes. Both students put effort into expressing their ideas in English, yet, owing to their English limitations, the obstructions they had to get rid of were too big for them. When one of them stumbled upon a problem, the other could not provide enough help. Each discussion on a sentence or expression was ended with uncertainty.
There was even a sentence from one of them that had been right (when she translated the Indonesian sentence "dia tidak memiliki pengetahuan" into "she doesn't Flora, Mahpul & M. Sukirlan, The contribution of negotiation of meaning to language accuracy in an EFL setting through a focused task | 309 have knowledge"), but since she followed her pair's suggestion, the sentence turned ungrammatical. The possible reason why L1 followed L2's suggestion was because she did not have adequate knowledge about the grammatical rule of the sentence. Therefore, she translated it word by word. Probably, L1 just translated "have" into her first language (mempunyai) and then added "not" to negate it, so she finally agreed to her pair's (L2) suggestion and put "not" after "have" so that "doesn't have" turned into "have not". That being the case, the theories of the input (Krashen, 1985) and interactional hypotheses (Long, 1996), and the output hypothesis (Ellis, 1991;Swain, 1985;Swain & Lapkin, 1995) were not proven right in the case of the L-L dyad's learning outcomes because of limited knowledge. Negotiation of meaning on H-H dyads informs that they completed each other's sentences because they had known the rules of a sentence. They were also able to explain why the first sentence was not quite right. Though it seems there was no contribution of the NoM in the form of completion or elaboration, the learning process made other types of contributions. For instance, through the focused task discussion, the learners, who were prospective English teachers, had considerable time to explain or recall the grammatical points they had mastered. Lastly, the negotiation of meaning in form of confirmation checks (correction) gave contribution for L-learner on H-L dyad. L-learner in this dyad took benefits for her language accuracies. The input can be internalized by L-learner since her interlocutor gave an explicit explanation (Richards, 2002), and because the input comes from her output (Swain, 1985;Swain & Lapkin, 1995). The input could not be internalized because at the same time she got too much to understand, and automatically, NoM in form of correction did not give sufficient contribution to her language accuracy. This is in line with what Krashen (1985Krashen ( , 1994Krashen ( , 2003 suggests that the input should be i+1 (comprehensible input). However, NoM in form of corrections for L-L learners did not give an optimal contribution, since sometimes, none of them was able to give the right correction. Based on the interview, both these low learners want to have a pair who has decent English ability while the H-learners both in H-H and H-learners declared that this activity is good because they have time to practice and explain the grammar they have acquired before. This study has limitations due to the interview since not all types of NoM produced by them were asked deeper. In other words, the types of NoM that contribute and do not were mainly counted based on the transcript of their utterances, contextually.

CONCLUSION AND SUGGESTION
The empirical data highlight the description of NoM when the students did NoM during the discussion on focused tasks. Regarding the focused task, it is constructed by the linguistic problems encountered by learners at the unfocused task before it. It makes a considerable contribution to language accuracy, especially to the grammar knowledge of an H-L dyad. It is most probable that it happens because of the linguistic problems derived from the utterances (output) of the learners and the explicit explanations in the source language such as Indonesian. NoM at a focused task is of substantial benefit to H-learners, particularly in the form of chances to put their grammar knowledge into use. It is of uttermost importance for a prospective English teacher. As for L-learners, NoM contributes a lot to their language accuracy, grammar knowledge in particular.
The data obtained from the dialogue transcript and interview with the L-learner as the sample units indicate that they asked the other interlocutors questions and requested explanations from them comfortably, so the input could turn into the intake (can be internalized), consequently it contributes to the learner's language accuracy. Therefore, to optimize the contribution of NoM to language accuracy (grammar and vocabulary), the topic to be discussed should be based on the learners' output (focused task) since the interlocutor can give help. However, to have in-depth information, detailed questions related to the contribution of NoM to language accuracy need to be conducted. Besides, for future researches, it is needed to conduct similar research involving four to five students in a group with different English abilities so that they can get more input for their language accuracy.