The Use of Repetition as Self-Repair of an Efl Learner

Self-repair is one of the language learning strategies of second language learners that are distinctly different from each other. This study aims to investigate the use of repetition as self-repair in academic speaking. The researcher tried to find the types of lexical elements which frequently repeated in the academic speaking and dig up the underlying reason behind the repetition as self-repair. A case study was conducted using human instruments and interviews to collect the data. Once the data filled, the researcher analyzed the data by implementing a data archive, trying to expand the codes, evaluating the data sets, and wrapping up the data. The findings from this study revealed that the participants in this study might have used English in speech situations in which they would have used recycling the mistake word. The result of the study found that the participant did indeed self-repair Type D, which consisted of repetition and replacement of one lexical item, and nouns became the most repeated lexical item in the academic speaking. This was a delay strategy as she waited to restart the speech as she needed to think about the next word so that the listeners didn’t misinterpret her. Twisted tongue and slip of tongue as a result of the participant’s first language interference contributed to the repetition of certain lexical elements as self-repair.

of English. The study described the location of self-repair mostly happened in lexical-grammatical forms (pronouns) and syntactic constituent structure in English. Self-repair gives a chance for students to repair their errors themselves without the teacher's involvement. Self-repair is one of the language learning strategies of second language learners that are distinctly different from each other. Sharwood (1994) defines language learning strategies as the effort learners make to improve their already-existing competence of the target language.
Pioneering research about self-repair was conducted by Garfinkel, 1991, Schegloff, andJefferson (1974) who used the term "errorcorrection" instead of "self-repair" to describe those phenomena in a spoken expression. She then introduced the concepts of 'repair' and 'selfrepair' in her next research with Schegloff and Sacks. Schegloff (1987) were the first to differentiate between error correction and repair. Correction is described as one of many possible types of repairs. Meanwhile, the repair is not only limited to error replacement but also deals with any sort of 'trouble' in spontaneous speech. Rieger (2003, p. 48) Defines 'repair' as error correction, word search, and use of hesitation pauses, lexical, qua-INTRODUCTION Indonesia, where English is learned as a second language, there is indeed a tremendous difficulty to move from a teacher-centered classroom to a student-centered classroom. Second language learners are prone to make an error and/or mistake in speech production. (Murcia & Olshtain, 2000) pointed out that in some ways speaking can be considered the most difficult skill to acquire as it requires command of speech production sub-skills like vocabulary retrieval and choice of grammatical patterns, and sociocultural competence. Broughton, et. al. (2003). Teaching English as a Foreign Language (Education Books) believed that a learner will always make errors as an unavoidable and necessary part of the learning process. It is an inevitable phenomenon since second language learners are not completely mastering competence. Sometimes, they produce faulty utterances caused by a lack of attention, fatigue, or carelessness. The fault is accidental and can be self-repaired (Eckerth and Tavakoli (2012). In 2009, (Hellermann researched a well-documented practice of mundane conversation (self-initiated self-repair) as performed during the interactions of one adult learner sciousness among them of the notion of self-repair, thereby making them less reliant and thus increasing their linguistic competence.
Self-repair strategies in monolingual speech are highly organized, but they may vary from language to language (Fox et al., 2010). In this case, self-repair mechanisms from both languages are activated and speakers can access strategies from either one language or the other (Hlavac, 2011). Hlavac, (2011. established a classification that includes repairs: Unfilled pauses; Filled pauses (non-lexicalized); Filled pauses (lexicalized); False starts (pre-positioned); Backtracking, repairs (post-positioned); Explicit preempting (pre-positioned); Explicit justification (post-positioned); Paralinguistic markers: laughter, nervous coughing (pre-and post-positioned); Equivalents from Croatian accompanying English code-switches. (Hlavac, 2011, p. 3798) Concerning self-repair mechanisms, the level of language selection is extremely important. If highly skilled bilinguals do not choose languages at the lexical level, they do not make speech mistakes in bilingual contexts or, more specifically, in code-switching situations. It must be stated that not all self-repair mechanisms occur only at a lexical level. Fehringer and Fry (2007). said that error repairs and recycling do not necessarily require reinitiating of the processing process. The most interesting process to analyze the level of language selection is the self-repair mechanism of replacement, because it assumes the replacement of a lexical item from one language by an equivalent item of the other language. Fox and Jasperson, (1995) classify self-repair into seven different types. Type A consists of the repetition of a lexical item. Type B involves the replacement of a cut-off word. The word that is being replaced is the one that contains the repair initiation. Type C consists of the repetition of several lexical items. Here the repairing segment and repaired segment are lexically identical. Type D consists of repetition and the replacement of one lexical item. Part of the repaired segment is repeated in the repairing segment, and another part is replaced. Type E consists of the repetition of a clause or phrase and the addition of new elements before the repetition. The added lexical items modify the clause or phrase or add background information. Type F is a variation of type E; it also consists of repetition plus the addition of new elements. These new elements are said to be 'a matrix construction' that frames the repetition. Type G consists of abandoning the portion of the talk that is being cut off and a restart. The restart may or may not be semantically si-lexical, or non-lexical pause fillers, immediate lexical changes, false beginnings, and instantaneous repetitions.
According to Rieger (2003), repair consists of the repairable, repair initiation, and the repairing segment. The repairable is not necessarily audible to the addressee and researcher but can be inferred from the presence of repair initiation and the repairing segment. Repair initiation can consist of a cut-off, a filler, or a cut-off and a filler combination, but in the case of repetitions, it may be non-observable. The repairing segment repairs the trouble that the speaker has perceived, for example by finding or replacing a word. Rieger suggests that the most commonly employed self-repair strategies are devices for delaying the utterance of the next lexical item. Among these devices, repetitions of lexical items are very prominent. Ochs and Schieffelin, (1983) they said that repetition is probably one of the most misunderstood phenomena in psycholinguistics. Repeat is a vital tool for children and adult learners of a second or foreign language to become more confident in the language of conversation. Perrin et. al (2003) state that there are four functions of diaphonic repetition, they are: (1) a taking into account function, which indicates that the interlocutor's intervention has been correctly heard and interpreted, in which case the interlocutor is encouraged to carry on with his or her speech turn; (2) a confirmation request function, which illustrates a problem related to what the interlocutor has said and paves the way for a subordinate repair exchange; (3) a positive reply function, agreeing with the interlocutor's intervention; through this intervention, the speaker makes it his own. It thus supports wrapping up the exchange by signaling the success of the current negotiation, and (4) a negative reply function, expressing disagreement with the interlocutor's previous intervention, which delays the closing of the exchange either by prolonging it, cutting it short or by introducing an element of controversy. Such repetitions concern allo-repetition, i.e., ''two-party repetition'' (Perrin et al., 2003). Spoken monologues, such as one-sided conversations or interviews, offer a fascinating alternate test case that helps one to explore patterns of self-repetition.
In this study, the research tries to investigate the use of repetition as self-repair in academic speaking. The research questions of this study are as follows: 1) What types of syntactic/lexical elements are most repeated?; 2) What may be the underlying reasons for repetitions?
This research hopes to help to raise con-and pragmatically linked to the abandoned part. According to Levelt (1983), the dimensions of self-repairs should be dealt with under the basis that the speaker constantly parses his own inner or transparent voice. If a mistake or improperness has been found, the speaker can, in a way, 'pass' the semantic properties of the initial expression to the correction. As to Rieger (2003), repair consists of the repairable, repair initiation, and the repairing segment. The repairable is not necessarily audible to the addressee and researcher but can be inferred from the presence of repair initiation and the repairing segment. People repeat themselves frequently in monologues or dialogues, a positive interaction approach will expect that dialog will minimize self-repetition as people respond effectively to each other's contributions.
Repetition is a natural, social feature, simply part of our daily actions and conduct, and not just a sign of a "disfluent" or "sloppy" speaker (Schegloff, (1987). Allwright and Bailey (1991) suggest that self-repetition is the perfect aim for students to achieve, as they can restore their communicative deficiency, be more precise and articulate in communicating without any teacher instruction, and will internalize the correct type of speech. The concept of repetition has often escaped the attention of linguists, mainly because cases of repetition were sometimes seen as redundant, imitative, sloppy, or even suggestive of a speaker's lack of good speaking skills 1. (Shimanoff, S., & Brunak, J. (1977) Repetition can be described as the act of repeating or reproducing verbal or non-verbal actions created by oneself or another in communicative contexts. A specific meaning of repetition usually refers to the act of reproducing the linguistic elements of the previous phrase (words and grammar) in precisely the same manner. Patrick (2014) concluded that comparison of the patterns of self-repetition for monolog and dialog shows that there is more total lexical repetition in monolog but this trend only occurs consistently at greater distances of turn/sentence. Contrary to this syntactic repetition, the discrepancy is more pronounced and is significantly higher in a monologue at all sentence lengths. Another clear probability is that the gaps between turns in speaking triggered by turns of other people in dialog cause loss or misunderstanding that leads to decreased repetition when there is no pause between successive turns in the monologue. As Patrick (2018) said that monologues are more repetitive because without the stimulus of contributions from others we are more likely to slip into our habitual lin-guistic routines. Perrin (2003) state that there are four functions of diaphonic repetition, they are: (1) a taking into account function, which indicates that the interlocutor's intervention has been correctly heard and interpreted, in which case the interlocutor is encouraged to carry on with his or her speech turn; (2) a confirmation request function, which illustrates a problem related to what the interlocutor has said and paves the way for a subordinate repair exchange; (3) a positive reply function, agreeing with the interlocutor's intervention; through this intervention, the speaker makes it his own. It thus supports wrapping up the exchange by signaling the success of the current negotiation, and (4) a negative reply function, expressing disagreement with the interlocutor's previous intervention, which delays the closing of the exchange either by prolonging it, cutting it short or by introducing an element of controversy. Jean (2000) concluded that repetition as self-repair is a necessary feature of our linguistic structure and behavior, and so some tidiness and regularity must occur. Others are welcome to cast their nets, thus growing our stream of thought on these topics.

METHODS
This study used a case study since secondlanguage learners might make a variety of mistakes that are different from each other. The focus of this study is to know how the use of repetitions as self-repair and the reason for usage. As Yin et. al., (2012) suggested, the writer applied descriptive design in the study included a single unit of analysis, i.e. academic speaking especially repetition used by the second language learner as a strategy to self-repair.
The participant is AA, a 23-year-old woman, she pursues her master's degree in one of the state universities in Indonesia. She had learned English since in the 4th grade of elementary school and spoke English frequently after she enrolled in English Education Department. Although she is an English Education Department student, she found that speaking is the most difficult skill because of her first language interference and the complexity of the second language (English). She said that it was very difficult to express what she thought and pronounce it correctly.
The data of this study were collected by implementing a triangulation method Johnson and Turner (2003) ; Tashakkori and Teddlie 2003), namely a method used to check and establish validity in this study by analyzing a rese-arch question from multiple perspectives (Simon (2009)). The primary data was collected through a record of her academic speaking. She was presenting certain material for a course. There are two videos collected from her presentation. The duration of each video is around 4-10 minutes. An in-depth interview was conducted in an informal setting to know the factors influencing the use of repetition as self-repair in her speech.
The recorded data (video of academic speaking) were transcribed. In this study, exposure to the verbal aspects of speech was the most essential thing, without neglecting other tools related to self-repair. The transcript, therefore, included all verbal components, as well as all pauses and speech fillers. For these analysis repetitions were given special attention as a significant factor. Therefore, they were specifically and carefully included in the document.
The coding process was conducted after data transription. It was a very complex and detailed procedure to classify all repetitions as selfrepair. Lampert, M.D., Ervin-Tripp, S.M., (1993) describe coding as the process of classification and labeling. On the surface, coding seems to be quite an easy task including "(a) identify the information that you wish to recover, (b) select mnemonic abbreviations or numbers as codes to represent that information, and (c) do it-match codes to actual cases in your database.'' (Lampert and Ervin-Tripp, 1993: 169) The process of coding was followed exactly as described to classify the self-repair elements. The repetitions coded were their linguistic structure, such as repetitions of prepositions, nouns, and verbs. Those repetitions that for obvious reasons were exceptions in coding and thus not examined because they cannot be considered self-repair. Consideration was given to the co-occurrence of interruptions, distracting background noise, and prosodic features (such as loud, soft, and emphasized speech) to determine whether the repetitions belong to the self-repair class or not.
The qualitative method was used to analyze the data. First, the collected data, self-repaired utterance, were categorized into seven types of self-repair proposed by Fox and Fox, B. & Jasperson, R. (1995) This study focused on A, C, D, E, and F since only those types of self-repair held repetition functions. Then the researchers classified the most frequent repetitions (lexical elements) including the repetitions of noun (RN), Adjective (RAJ), verb (RV), prepositions (RP), and conjunction (RC). The qualitative analysis is in the context of interactional sociolinguistics with its basic assumption that the meaning, structure, and use of language are related socially and culturally (Schiffrin (1994)). In analyzing the situated meaning of each repetition, repeated forms in the academic speaking contexts under investigation could be matched with recurring functions. Self-repair subcategories could thus be described in terms of form and function in academic speaking to get a general conclusion.

Types of repetition
The data of this study were collected from three videos of participant's academic speaking projects with a total duration of 33 minutes. All the speech and self-repairs were transcribed and coded as shown in Table 1. The transcribed audio then classified into seven types of self-repetition proposed by Fox, B. & Jasperson, R. (1995); they are Type A, Type B, Type C, Type D, Type E, Type F and Type G. From 3086 words, there are 94 items identified  Table 1, it shows that the participant mostly (40.3%) did self-repair Type D where she repeats and replaces only a lexical element. Part of the repaired segment was repeated in the repairing segment, and another part is replaced. In the example "...may need a bridge to reach their profe-their personal and professional batteries."; the participant repeats "their'' in the repairing segment, and she repairs "profe-" in the repaired segment into "personal". The repetition employed by the participant because she twisted her tongue to pronounce the word and wanted to repair it directly. Then the least type of selfrepair done by the participant was Type F (1.6%) where she repeated a clause "It is based on.." but she added a new element or 'matrix construction' "Oh no, I mean" intended to give the signal of repairmen before the repairing segment.
To know the most repeated lexical element by the participant, the researcher analyzed the frequency of each repetition employed by her. It was observed that participants mostly repeated five main lexical elements when applying the selfrepair process in her academic speaking. These were repetitions of noun (RN), Adjective (RAJ), verb (RV), prepositions (RP), and conjunction (RC). The frequency of lexical element repetition in academic speaking will be shown in Table 3.
The interpretation of data revealed that along with these five types, the participants mostly did the repetition of Noun as self-repair. She made 25 noun repetitions in her academic speaking. The participant employed the repetition of those lexical items because she wanted to postpone the subsequent lexical item production. It allows the speaker to keep the floor at a potential transition-relevance place (TRP) and/ or it gives her time to engage in linguistic and/ or cognitive preparation, i.e. she may check for a particular word or structure, or think about the substance of her expression (Rieger, 2003). In the example (1), the participant used the repetition of "knowledge" to correct her pronunciation since she argues that sometimes she twisted her tongue in pronouncing English. There is a short delay between the repetition of "teacher "to adjust the correct intonation of the speech.
(1) AA = "Through an active mental process development and then create and build the meaning and knowled-knowledge so teacher-teacher can set learnercentered" The repetition of adjectives was quite frequent too in the participant's academic speaking. Example (2) shows the case of double repetition of the adjective "instant". She was insecure about her pronunciation, then she tried to repair it immediately but she employed the mispronunciation "instante" and repeated it once again into correct pronunciation. The participant's insecurity came from the abundant miss-pronunciation produced in her academic speaking while she tried to make the audience understand her speech. She sometimes remembered her first language sounds, so produced English sounds like her first language.
(2) AA= "it takes it takes time to learn, so learning is not instant instante instant" The least of all repetition employed by the participant was the repetition of conjunction. She repeated the conjunction such as "and", "for", "but" and "so". This helps to pause the production of the next lexical item or to postpone any TRP. Conjunctions were often repeated without any other components, but occasionally pauses, sound lengthening, quasi-lexical and lexical fillers are pronounced between the first and second use of the conjunction. In the example (3), the participant used a quasi-filler between "and then" conjunction.
(3) AA = "and then... and then the next is about social constructivism" From those analyses, most of the repetition was used to delay the next lexical item production so the participant had time to think about what appropriate words she would say. Twisted tongue and slip of tongue are also the factors of doing repetition as self-repair. Repetitions some- times utilize a pause between the first and second utterance of the lexical item in the academic speaking.
The underlying reasons for repetitions as selfrepair This research confirms the presence of repetitions of the essential elements of L2 among IL speakers. The analysis results indicate that this academic speaking may consist of any grammatical category element. Participants have used this strategy to reinforce the assumption that this strategy is universal instead of specific to language learners. Repetition emerged as one of the key communication strategies to facilitate communication between the speaker and the audiences. In the Non-Native Speaker, we found preposisiton, verbs, adjective, conjunction, noun, and conjunction may all be subjected to repetition.
The interview indicated that the participant was a student of a university in Surakarta, Indonesia, she took a graduate program in English Education Department. She claimed that she is Indonesian so the first language is Bahasa or Indonesian language then the second language is English. She had learned English for around 10 years since she was in elementary school. AA: "Okay my first language is Bahasa or Indonesian language". "Then my second language is English" AA: "I learn English for more 10 years, it starts when I am in elementary school until now" The interview, the researchers may be concluded that she actively practices her English academic speaking in her daily life, formal, and nonformal, this the excerpt AA: "Yes I must practice my English academic speaking daily because consciously and I am a student of English education so at least 3 days in a week I do English academic speaking because I can practice it in college in a way interact with my friends and lecturer " Birnbaum et. al. (1977) characterizes the speech of adult second language performer and ESL student who shows signs of overuse, During a conversation that lasted just longer than 15 minutes, that there was not a single lengthy speech which is not riddled with gaps, false endings, repetitions, and other voice repairs___there are far more than 69 examples of reparation (not counting gaps). The data from the entire procedures conducted to analyze the participant's repetition might suggest the following interim conclusion.
First, the participant did repetition. She assumed that English is difficult to learn considering the participant has learned English since elementary school, especially in pronunciations, because in her case, she found that a major obstacle which might possess of the appropriate vocabulary and diction when the informal situation we should use a structured speaking rather than in informal. According to Aba (2013), He clarified that learning English as a foreign language was not an easy feat. There are certain aspects of language which need to be mastered, such as language skills, which consist of listening, speaking, writing, and even language aspects such as grammar, syntax, and vocabulary. AA: "Yes there are some obstacle or difficulties when I did English Academy speaking it is about pronunciation for vocabulary and also structure because I know that in academic speaking we should do use the appropriate vocabulary or dictions and use the right structure it is different with non-academic speaking we can use slang then we can forget about the structure" Second, the participant realized that she made a mistake, then the participant might find a strategy to overcome her mistake such as, repeat her word and change to the right word, she explained the purpose of doing repetition was making her statement more clear and understandable for the audience to reach her point. Ellis and Barkhuizen. (2005) said that learners should turn to a contact technique, a systemic ability to which the speaker has recourse while having trouble in communicating himself to keep the conversation running. They are ignored because of their oral fluency, but still want this kind of support Borland and Pearce (2002) AA: "The purpose of doing repetition is making my statement more clear and understandable for the audience because I think that when I speak incorrectly whether about pronunciation or structure the audience did not get my points so I repeat the word to make it clear" Third, another related reason is perhaps the participant's first language interference, which is Indonesian and/or the Javanese as her mother-tongue, was some factors that affected her repetition such as slip of tongue. L2 complexity, English pronunciation, contributed to her twisted tongue. As stated by Masangya & Lozada (2009)  For the last conclusion in the interview, it was dealing with satisfaction after she made a repetition, she felt confident after doing repetition. The data might suggest that the discourse that is produced spontaneously tends to be able to demonstrate a certain nature of a learner's ability that may not be missed out. Prassetyo, (2016) explored the convenient error made by the participant that makes the participant feel more inspired. In the field of Second Language Acquisition, it can be implied that error analysis might provide convincing data portrayal about interlanguage errors made by the participant. However, it should be noted that, first of all, some of the errors might not be observable. AA: "Of course, I feel satisfied after doing repetition. It can make me more confidence in speaking because I feel that what I'm talking but can't deliver well to audiences and I feel that the audience more understand because I used to ride for pronunciation and structure" Since the repetitions, namely, to allow the speaker to gain time without losing the floor while searching for a word or construction, the participant in this study might have used English in speech situations in which they would have used recycling the mistake words. The results have indicated that the English often possesses difficulties for many people, even students of the English education department. Mastery of the basic rules in speaking takes time. Despite the resources made available for exposure of the language intended, errors or mistakes are still bound to happen. However, such instances are not a clear indication of failure to language acquisition but learning should be constant and must seek other tools for improvement.

CONCLUSION
Making a mistake for second language learners is an inevitable phenomenon. They are not completely mastering competence. Sometimes, they produce faulty utterances caused by a lack of attention, fatigue, or carelessness. To correct the mistakes or errors, learners have their strategy to self-repair, one of the strategies is doing a repetition of lexical items. This study analyzed the repetition as self-repair produced by a second language learner in academic speaking. The participant of this study mostly made selfrepair type D where she repeated and replaced the lexical item. Repetition of nouns frequently employed by the participant to delay the time in producing the next items in her speech. She did the repetition also due to twisted and slip of tongue as a result of her first language interference. Her first language sounds affect her pronunciation in pronouncing English. English often possesses difficulties for many people, even students of the English education department because of its complexity.
This study hopes to give a view about repetition as self-repair as a learning strategy. However, this study is still lacking data collection since this study was conducted during COVID-19. The next study might employ a longer duration of speech, so there will be various cases of selfrepair that can be analyzed. The next study might also compare repetition as self-repair between Bahasa Indonesia and English.