91 OTHER-INITIATED SELF-REPAIRS IN STUDENT-STUDENT INTERACTION : THE FREQUENCY OF OCCURRENCE AND MECHANISM

The current study is intended to explore the typology of other-initiated self-repair in term of its frequency of occurrence and to investigate the mechanism of self-repair in an academic setting. To obtain the required data, a classroom conversational analysis was conducted in which the researchers observed a group of graduate students (19 active EFL speakers) in a state university in East Java, Indonesia. Based on the qualitative data analysis, the results of the study revealed that all typologies of other-initiated selfrepair were used depending on the aims the interlocutors try to gain since all typologies have diverse aims and ways. Of all typologies, confirming checks were dominantly used by the participants. Regarding the mechanisms, three types of mechanisms were found; full two-cycle repair, background check, and short-cut repair sequence.


INTRODUCTION
Troubles in speech production naturally appear in spontaneous speeches and conversation in student-student interaction.The use of English as second or foreign language may trigger the natural troubles in speeches as a part of learning process.When the troubles are detected, acts to repair may occur.The act of repairing can be initiated by the speakers or the interlocutors.The act of repairing which is initiated by the speaker is called self-initiated self-repair (e.g., Wisrance, 2017;Trisanti, 2017;Sato & Takatsuka, 2016) while the act of repairing initiated by the interlocutor is called other-initiated self-repair (Schegloff, Jefferson, & Sacks, 1977).
To the both topics, other-initiated self-repair is infrequent to discuss since the phenomena of self-initiated self-repair is more frequent to occur in daily communication.The self-initiated selfrepairs in EFL teaching are frequently revealed to know whether or not self-initiated self-repairs arise in the classroom discourse (e.g., Wisrance, 2017;Trisanti, 2017;Sato and Takatsuka, 2016).The study regularly involves learners from higher education.Lyster & Ranta (1997) stated that the occurrence and the success of repairing possibly will only happen when learners have acquired an adequate level of English proficiency.
The studies of other-initiated self-repairs are existent eventhough these are not as burst as the study of self-initiated selfrepairs.In 2008, Svennevig employed a study to investigate whether or not there is a frequent preference in the choice of repair initiation techniques.Svennenvig (2008) claimed a preference to address troubles in conversation as a hearing trouble is used mostly rather than to address them as troubles of understanding or acceptability.In requesting for hearing repair, mostly the interlocutor shows explicit request by saying ''what did you say?'' (Svennevig, 2008).Furthermore, a library-research of other-initiated self-repair was done by Dingemanse and Enfield (2014).In their research, they provided an overview of research methods and conceptual framework of otherinitiated self-repairs study.Based on the review, they recommended the qualitative analysis of individual cases and their environment in linguistics systems can be collaborated with a quantitative and comparative perspective to struggle on the organisational details of a possibly universal system for other-initiated repair.
However, the investigation of other-initiated self-repairs in depth and in details focusing mainly on frequency of the occurrence and the mechanisms of repairs has not been done.Thus, in addressing other-initiated self-repairs of graduate students' of EFL, the present research is intended to depict the frequency of occurrence and the mechanism of other-initiated self-repair in an Indonesian academic setting.

LITERATURE REVIEW Speech Production in Student-Student Interaction
In the classroom contexts, speech production possibly will occur during student-student interaction.Interaction among students occurs when students interact to other students.Student-student interaction arises during answering and asking questions, making comments, and conducting open discussions (Robinson, 1997).To add, discussion may offer any condition which highlights on sharing activities of the students.
In particular, EFL university student in Indonesia put into practise a lot to speak English fluently by doing many kinds of activities such as individual or group presentation, asking and questioning activity, and open discussion.Open discussion in cooperative group, for instance, can craft clarification of ideas and perspectives in a context free of teachers' initiation in the classroom (Gillies, 2006).During classroom discussion, students with confidence become more independent to interact with the other students.Students will not rely on the teachers who consistently initiate to build any communicative interaction (Nunan, 1992).By having discussion in the classroom, it is potential to build up student-student interaction.They are projected to be more active and initiative in the classroom by interacting with others: commenting, asking, answering, and sharing their best knowledge to other.In these activities, students may utter sentences with intension and other may notice mistakes of others' utterances.

Trouble Sources in Student-Student Interaction
In student-student interaction, trouble may occur and it potentially interrupts the flows of interaction in the classroom.In terms of troubles faced by the interlocutor, troubles of hearing, understanding, and acceptability, a study has been conducted by Svennevig (2008).The findings showed that in repairing troubles of hearing and understanding there was a strong tendency towards providing nominee solutions to the trouble (checking hearing and understanding) over merely representing the presence of a trouble or specifying the nature of it.Referring to the universal frequency of those trouble sources and the ways to repair the troubles, speakers are subjective towards expressing solutions offered to troubles of hearing and understanding and towards merely demonstrating the nature of the trouble regarding the troubles of acceptability (Svennevig, 2008).

Repairing Strategy: Self-Repair
Self-repairs are formed in response to a linguistic problem in speech, such as the use of inappropriate lexis, pronunciation, or syntax (Pillai, 2006).Self-repairs arise when speakers discover and modify errors in the speech production (e.g., Levelt, 1989Levelt, , 1999)).Selfrepair has two unalike types which are self-initiated and otherinitiated self-repairs (Schegloff et al.,1977).Self-initiated self-repair is employed when speakers grasp that they need to repair their utterances without encouragement from another participant, for example, 'They are going to New -um, Zealand'.Differently, otherinitiated self-repairs arise where other participant prompt a communicative need and signs a particular utterance for the need in communication, such as asking about part(s) of the preceding utterance.
The Overlooked Self-Repair: Other-Initiated Self-Repair Self-repair can be initiated by an interlocutor when an interlocutor detects any inappropriate or even does not catch what the speaker is saying.Schegloff et al. (1977) stated that self-repairs can be from other initiation -known as other-initiated self-repairs.In his study, Björkman (2014) reported that there are six typologies of otherinitiated self-repairs: paraphrasing, repetition, overt questions, clarification request, co-creating the messages, and word replacement.
Fundamentally, paraphrasing is done to state the same contents expressed with different words (William et al., 1997: 312).Repetition is a frequent approach which is important to achieve as collaboration (Mauranen, 2010).Just as Rieger (2003) assumed that repetition is the most communal type of repairs of a particular segment.Overt question strategy is basically employed when the interlocutor raises questions about the preceding utterances due to unclear statements, and the interlocutor confirms to check (Jamshidnejad, 2011).Clarification request is used when interlocutors ask for additional or extra explanation or more details of what the speaker has uttered (Dornyei and Scott (1997).Co-creating the messages of anticipation introduced by Björkman (2014) occurs when speakers fill in gaps in each other's utterances in an effort to produce a comprehensive utterance (Kirkpatrick, 2007).Word Replacement in this content is as same as immediate lexical changes notion promoted by Schegloff et. al. (1977).Word Replacement depicts clearly language use or the content matter.This strategy relates to the changes of inappropriate words into the appropriate words to elude misleading mainly to the interlocutors' interpretation.

Other-Initiated Self-Repair Mechanisms
The mechanism of other-initiated self-repair portrays the patterns of self-repairs by speakers due to other's initiation.It habitually involves the use of interruption marker, such as 'sorry', 'excuse me', or the reuse of element of the preceding speech (Schegloff, 1977).The other-initiated self-repairs sequences are arranged by adjacency pairs, and that they are sequences per se (Schegloff, 2007).Hutcby and Wooffitt (1998) recommend three positions of repairing; representing trouble source, depicting next-turn repair insertion (NTRI), and representing the process of repairing.In more detailed, the mechanism of other-initiated self-repair is elaborated as follows (adopted from Svennevig, 2008).

Short-cut repair sequence:
Trouble source turn B: The girl is sad.

Hearing repair initiation (hearing check)
A: sad? Understanding repair B: She's in her mourning

METHOD
This study is qualitative in nature; which for all objectives and purposes is emphasized on human beings in social circumstances (Robson, 2011).It uses the qualitative paradigm despite the fact that it espouses the methodological approach of Conversational Analysis (henceforth CA).Furthermore, this study scrutinized the organizational features of talk among students in the classroom interaction and the data was analysed using CA.The subjects of this research were a group of graduate students, 19 students, who learn English actively.They were involved because the researcher has personal interest to the group.The group is well-known as very active EFL-speakers in the classroom primarily at the moment when they have unlimited time to do questioning-and-answering troubles.This study was done in a class of graduate students of EFL in a state university in Surabaya.
In this study, observation was the main instrument used.The observation was done for 100 minutes in a linguistic class when there are three students presenting published scientific articles.The rest of students were listening to the presentation.Afterward, the group came to sections to discuss the papers; asking and answering activities.While doing the observation, the researchers recorded the students' interaction in the classroom by audio recording.It was then transcribed to complete the data taken during observations.To obtain the data, the researchers attended the classroom and note down all activities in the classroom particularly the occurrence and mechanism of other-initiated self-repair in student-student interaction.While observing, the researchers recorded all the utterances during the observation to notice all students' speech during the classroom activities through audio recording.
To answer the reserach questions, the data were analysed through several phases.The first procedure was to ascertain any kinds of phenomena occurred in the classroom interaction among the participants in this study.In this stage, the researchers re-read the notes they got from the observation to detect the occurrences of frequency and the mechanism of other-initiated self-repairs in the student-student interaction.Secondly, the researchers condensed the data that had been noted and transcribed.The researchers selected, eliminated, focused, simplified, and transformed the data that are linked with the phenomena of other-initiated self-repairs.The researchers then categorized the data based on the typologies of the repairs so that the researchers could exhibit the data clearly for the research question 1 and 2. The next step was to form regularities and patterns with respect to the occurrences of the phenomena and to display that these regularities were methodically produced and oriented to by the participants as normative organizations of action (Heritage, 1984).Finally, the data were displayed and reported in the form of sentences.The data were exhibited by using next-turn proof procedure since the data for those research questions involved conversational analysis.

FINDINGS AND DISCUSSION The frequency of other-initiated self-repair occurrence in studentstudent interaction
To answer the first research question, the researchers scrutinized the data in depth to exhibit the empirical proofs of the use of repairing strategies in speech errors through observing and recording transcription.Table 1 and 2 are presented to show the researchers' findings related to the frequency of other-initiated repair strategies based on the typology and trouble sources used by graduate students mastering English as FL in the classroom.The table envisions the frequency of other-initiated self-repair occurrence based on typology.Based on the observation and the audio transcription, the researchers found that clarification request, an interlocutor ask for explanation or more details of what the speaker has told, is dominantly used as other-initiated self-repair.It is used 8 times in 100 minutes.Similar to the study of Björkman (2014), clarification request was used dominantly as strategy to initiate others to repair, popularly known as other-initiated self-repair.Repetition then becomes the second typology of other-initiated self-repairs that is mostly used by students in the classroom context.In this study, students in 100 minutes used repetition 8 times to initiate others to repair the trouble sources in speaking.Meanwhile, overt questions were used 7 times and paraphrasing was used 5 times during the class discussion.Yet, co-creation of the message or anticipation occurred 6 times and word replacement was used only 2 times in 100 minutes.Therefore, to list the typology based on the frequency of occurrence in the classroom setting, the researchers concluded that the sequence of the typology are clarification request, repetition, paraphrasing, co-creation of message of anticipation, and word replacement.The result of this study is certainly similar to Björkman's seminar work written in 2014.
The findings on other-initiated self-repair occurrence based on trouble sources were presented in table 2. Based on the trouble sources, the frequency of occurrence of other initiated self-repair has no significant difference.Table 2 shows that based on trouble sources, troubles of understanding were significantly used.In total, other-initiated selfrepair used due to trouble of understanding occurs 11 times.Meanwhile, other-initiated self-repair used due to trouble of hearing occurs 9 times and other-initiated self-repair used due to trouble of acceptability occurs 8 times.The difference of the frequency of occurrence of other-initiated self-repair may be influenced by the setting and the moment when the research conducted the observation.As mentioned previously, the researchers administered the study when graduate students were doing presentation and having discussion on the published journals presented in the classroom.The discussion activity, such as asking and answering activities, may provide the tendency of having troubles of understanding since in asking-answering activity what students mostly did were asking what they do not understand or things that were not clear so that they need extra information about certain issues.The result of this study reported similar result of Svennevig's (2008) study revealing that hearing trouble occurred more dominant rather than other troubles.Svennevig (2008) reported based on his finding that trouble of understanding occurred 99 times, troubles of hearing occurred 71 times and trouble of acceptability occurred 50 times.To conclude, the result of this study in terms of the frequency of occurrence of other-initiated self-repair based on trouble sources is in line with the result of Svennevig's seminar work.
In addition, in the case of revealing the frequency of otherinitiated self-repair occurrence based on trouble sources in microlinguistic elements, the result of the study is exhibited as follows.
Table 3 The frequency of other-initiated self-repair occurrence based on trouble sources in microlinguistic elements The researchers found that the grammatical incorrectness and the lexis could guide the interlocutor to initiate others to repair.After analysing the data, the researchers found that grammar and pronunciation mistakes triggered the interlocutor to initiate others to repair the errors.The grammar error occurred twice and word choice error occurred and triggered the interlocutor to initiate others to repair the troubles 5 times and the inappropriate pronunciation occurred once.

The mechanism of other-initiated self-repair in student-student interaction
To respond to the second research question, the researcher used a similar instrument to the research question 1.The mechanisms of other-initiated self-repairs occurred in three different forms; full two-cycle repair sequence, background check, and short-cut repair sequence.To portray the details, the findings of the study were presented as follows.

Full Two-Cycle Repair Sequence
The full two-cycle repair sequences were found when students communicated with intention to check whether both the speaker and the interlocutor have similar comprehension about the topic they were discussing.Excerpt (1) displays how the mechanisms in the form of full two-cycle repair sequence occurred.Excerpt (1) shows that during interaction, a student as an interlocutor detected there was an inappropriate word that was uttered by the speaker while the interlocutor believed that what the speaker intended to say was different from what the speaker was actually saying.The trouble of acceptability then occurred so that the interlocutor initiated another party (the speaker) to repair by saying 'you mean the questionnaire?'.The way the interlocutor initiated the speaker to repair the speaking is word replacement.As stated previously, word replacement displays changes of inappropriate words into the appropriate words to elude misleading mainly to the interlocutors' interpretation.The interlocutor initiated other to repair by changing the word 'question' became 'questionnaire', the more appropriate word to use based on the topic they were discussing.The case is the interlocutor initiated to keep the content matter which is 'questionnaire' to be clearly understood.Björkman (2014) mentioned that word replacement in his seminar work is frequent related to language use even though in particular case, she found lexicogrammatical word replacement, and content matter also occurred.Thus, the instances provided in this study clearly depict how word replacement used in content matter.Excerpt (1) also depicts the response of the other to repair the trouble sources.In the excerpt (1), it can be seen that the speaker provided the confirmation by saying 'Yes, the questionnaire I mean' as the acceptance that both the speaker and the interlocutor agreed and understood the topic they were discussing.

Excerpt (1)
Excerpt (2) also depicts how other-initiated self-repair occurred in the form of full two-cycle repair sequence.The sequence was started when the speaker symbolized S1 uttered 'Oh, yes.(8 sec).oh.Yes in, in page 78, in..'.The trouble was detected by the S2 as the interlocutor when she failed to catch what the speaker mentioned which is number 78.The trouble detected in excerpt (2) was clearly about trouble of hearing.When the speaker said the number 78, the interlocutor may have difficulties in gaining the idea so that the interlocutor initiated the speaker to repair or restate what has been mentioned by uttering 'Sorry, what did you say?'.However, the trouble of hearing that appeared may not fully occurred since the interlocutor actually succeeded in identifying the number 78 with doubt in mentioning it.The sequence of turn-taking after the interlocutor initiated the speaker to repair and clarify what the interlocutor said, the speaker in excerpt (2) showed that he succeeds in receiving signal from the interlocutor by confirming 'yes'.The turntaking was continued by the interlocutor by signalling that he understood the repair initiation by uttering 'yes, it is on page 78?' and finally the sequence of turn was ended by S1's utterance.

Excerpt (2)
Similar to those previous two excerpts, excerpt (3) also depicts how full two-cycle repair sequence occurred.Other-initiated self-repair may occur when an interlocutor initiate to fill what the speaker failed to say.The strategy is popularly known as co-creation of the message or anticipation.As mentioned previously, Kirkpatrick (2007) stated that co-creation of the message or anticipation deals with how another participant of communication fill in gaps in each other's utterances in an effort to produce a comprehensive utterance.Excerpt (3) factually depicts how cocreation of the message or anticipation of other-initiated self-repair occurred.The trouble occurred when S1 failed to deliver a lexis 'clarification '. Excerpt (3) shows that the speaker may get a trouble to utter the word intended to mention so that the interlocutor initiated to complete or fill what the speaker was actually intend to say.The interlocutor initiated the speaker by uttering a word 'classification'.Then, the speaker confirmed it by saying 'Yeah, classification…' and elaborated his idea.In stating the further explanation, the speaker had a trouble to select a proper word which made the interlocutor had a question in her mind.However, the interlocutor seemed to understand what the speaker intended to say.Then, she ensured what the speaker's intention is and the interlocutor understood were in line so that the interlocutor decided to initiate by using clarification with the utterance 'you mean the classification of the learning strategy?'.When both the speaker and the interlocutor got similar understanding of ideas or topics they were discussed, the speaker then ended the sequence by saying 'yeah, the learning strategy.

Background Check
Background check as the sequence of other-initiated self-repair also happened in the classroom interaction.The researcher of this study found the mechanisms of other-initiated self-repair using background check pattern.The excerpts are presented as follows.
Excerpt (4) Background check as the mechanism of other-initiated selfrepair occurred not so long as the full two-cycle sequence occurred.The background check mechanism occurred also due to a trouble.When an interlocutor detected a trouble such as he or she did not know the meaning of particular word or phrase, he or she may ask directly about the meaning of the word or the phrase by repeating what the speaker has already stated.Excerpt (4) shows how an interlocutor initiated the speaker to repair or give an additional clue of the meaning of M mentioned by the speaker.In initiating the speaker, the interlocutor afforded to request a clarification from the speaker.When the speaker succeeded to catch the initiation of the interlocutor, the speaker then afforded to explain the letter M which stands for Mean.In excerpt (4), it seemed that the interlocutor needed an extra explanation of what the speaker explained by checking the background by saying 'Mean.Did you find on the page 78?'.
Excerpt (5) also presents how background check mechanism of other-initiated self-repair.In excerpt 5, S2 firstly completed the S1's utterance in the second line.To add, the communicative behaviour is well-known as 'anticipation' in this monograph (adapted from Kirkpatrick, 2007).Later on, S1 repeated S2's recommended phrase.Following that, S1 checked his understanding by continuing his ideas.To avoid misleading of understanding between two people speaking, the S2's utterence in the last line shows how filler such as 'Uhm…' seemed to be a method to avoid potential trouble in understanding.Then, the S2 anticipated it by flooring ideas by uttering a yes-no question that can be identified in line 4.

Short-cut Repair Sequence
Excerpt (6) comes from the same setting of collecting data of the prior excerpts.It presents how short-cut repair sequence occurred in other-initiated self-repair.since the context and the topic of their discussion is about how a researcher of a seminar work collected data in a questionnaire.In line 3, the speaker succeeded to catch the signal of the interlocutor's initiation so S1 directly repaired by repeating the word 'respondents.' The data of excerpt (8) below includes the instance of cocreating the message of anticipation.In excerpt (8), in line 2, it can be seen that S2 filled in the blanks of S1's utterance in an effort to assist the S1 to produce the complete utterance.Kirkpatrick's term of this type of anticipation is 'lexical anticipation' (Kirkpatrick, 2007).In line 2, the S2 completed the phrase the S1's utterance.The line 3 then shows that the S1 succeeded to understand the repair's initiation of S2 so that S1 accepted the initiation by saying 'multilingual students'.

CONCLUSION
In summary, it is found that of all typologies of other-initiated self-repair strategy, confirming checks, which cover paraphrasing, repetition, and overt questions were dominantly used among other typologies.Of all the three types of confirming checks, repetition was the most frequently used, followed by overt question, and paraphrasing.Clarification request as the other type of other-initiated self-repair strategy was also used as frequent as repetition.Meanwhile, co-creating of messages or anticipation was the third typology frequently used and word replacement was the less frequently used typology.In relation to other-initiated self-repair mechanism, three sequences: full two-cycle sequence, background check, and short-cut repair sequence, are administered.

Table 1
The frequency of other-initiated self-repair occurrence based on typology

Table 2
The frequency of other-initiated self-repair occurrence based on trouble sources