Turn-Taking in Colloquial Indonesian

Participants in a conversation follow certain basic rules in obtaining turns. Some questions addressed in this paper are, firstly, how turns are allocated in Indonesian conversation. Secondly, what are the signals given by a current speaker willing to yield to a next speaker (i.e. turn-yielding signals). Thirdly, what repair mechanism is employed by the participants when the turn-taking procedures are broken. The data used in this study were taken from an informal conversation of four Indonesians from Jakarta. The data were collected by using a non-participatory observation with a recording technique. The analysis shows that the turn-taking system in the Indonesian language confirms the model of turn-taking system developed by Sacks et al (1974), covering the occurrence and the recurrence of speaker change, the overwhelming tendency for one party to talk at a time,the common occurrences of more than one speaker at a time, and the common switches of speakers with no gapsand no overlaps.


Introduction
Turn-taking in a conversation is an important topic in discourse-analysis research as seen from the considerable number of studies and discussions on turn-taking, especially in English conversations (see, among others, Orestron (1983), Sacks et. al. (1974), and Nofsinger, (1991)).Much research has shown that participants in a conversation follow certain basic rules in obtaining turns.Richards and Schmidt (1983) said that participants in a conversation do not take turns randomly, but they are bound to a certain turn-taking system which governs a conversation.
Dealing with the turn-taking system, Sacks et al. (1974) state that the main feature of the turn-taking system is how turn-taking in a conversation is organized in terms of allocation of turns and construction of message units.In terms of allocation of turns, Sacks et al. propose two basic techniques, which are, firstly, current speaker may select next speaker, and secondly, next speaker may self-select.If, neither of these techniquesis applied, the current speaker may continue holding the floor.In terms of construction of message units, Sacks et.al. said that participants may construct the units in the forms of sentential, clausal, phrasal, or lexical construction.This implies that in taking a turn, a speaker may produce a single word.The turntaking system proposed by Sacks et al has become a basis for much research in turn taking strategies across languages.
So far, however, turn-taking systems discussed by scholars mostly refer to research done in English conversations.The data of turn-taking behaviors of speakers of languages other than English are very rare.Therefore, their turn-taking behaviors are not adequately understood.
Nothing is known about turn-taking behaviors in Indonesian.This paper aims to discuss turn-taking systems in the Indonesian language, especially in an informal conversation.Some points I am going to discuss are, firstly, how turns are allocated in Indonesian conversation.Secondly, what are the signals given by a current speaker willing to yield to a next speaker (i.e.turn-yielding signals).Thirdly, what repair mechanism is employed by the participants when the turn-taking procedures are broken.In addition to this, this analysis will also discuss gaps and overlaps occurring in Indonesian conversation.The analysis of the features above will be very important before discussing whether or not the turn-taking models suggested by Sacks et.al.are also applicable to Indonesian conversations.Before analyzing turntaking systems in Indonesian conversation, however, I am going to explain about what is meant by a turn.The understanding of the concept of turn-taking will be important before coming to the analysis and the discussion of turn-taking in this paper.Some scholars have proposed the definition of a turn.Goffman (1981) defines a turn as an opportunity to hold the floor.According to Sacks et al (1974), a change of turn is indicated by a change of speaker, and it takes place at a turn-relevance place, that is a place where a speaker change can legitimately occur.It implies that signals such as minimal responses like 'hm, yeah, hmh, hmm', that do not occur at transition relevant places, are not considered as turns ( Reid, 1995).This is very reasonable because minimal responses or back channeling sometimes may be meant to encourage the current speaker to continue holding the floor.In addition to this, they may also be used to show that the person producing minimal responses is still listening to or following the conversation.For the purpose of this study, I refer the definition of a turn as given by Goffman (1981).Back channeling or minimal responses which do not occur at a transitionrelevance place are not considered as an attempt to take a turn.

Research Method
The data used in this study are informal or colloquial conversations by Indonesian speakers from Jakarta.In this case, the data of informal conversations were taken from four Indonesians who are close friends.They are about 19 years old, and they use Indonesian in their home.They were recorded when they were having a conversation in their dormitory in Canberra.
The data that were taken by using an observation method with the recording technique, were then transcribed before they are available for further analysis.As pointed out by Tao (1996), an initial problem for any discourse-oriented linguist is dividing the flow of speech into useful units for analysis.He further said that in many studies dealing with information flow in conversational discourse, an intonation unit has been taken as the natural unit of discourse.An intonation unit (IU) is defined by Du Bois et al.(1993:47) as 'a stretch of speech uttered under a single coherent intonation contour'; the criteria for identification of IUs are discussed in Du Bois et al. (1993).
Following Tao (1996) and Du Bois et al. (1993), I take the IU as a natural unit of spoken discourse and proceed from this prosodic unit to analyze turn taking system in Indonesian.In this study, therefore, the conversational data are segmented into IUs.To identify the speaker of a given turn in the conversation, a capital letter is inserted at the beginning of the turn followed immediately by a colon (:).The capital letters 'A', 'B', 'C', and 'D' are used here to identify different speakers.To get more natural data, the spoken data used for the analysis is the one taken after about five minutes of the recording.

Result and Discussion
The analysis of turn-taking will be divided in to four parts, which are, allocation of turns, indicators of ends of turns or turn-yielding signals, repair mechanism.In addition to this, I am going to discuss gaps and overlaps as we may need to know how often they occur in the conversation.Besides, we may want to know how gaps and overlaps occur in the conversation.

Allocation of Turns
The data of 230 turn exchanges show that two techniques of turn-allocation are employed by the Indonesian speakers, which are, firstly, the current speaker selects a next speaker, and secondly, the next speaker self selects.Most turn-takings are predominated by the addressed question technique.In this case, it is very common that the current speaker selects the next speaker.The following segment of the conversation shows that speaker D (line 54) selfselected a turn, then he allocated the turn to speaker A by addressing his name 'Yung' followed by a question.Speaker A (line 55) took the turn but he might not be clear with what speaker D was asking about, therefore, he was asking for a clarification by saying 'apaan' (' what do you mean ?' ).Speaker D took the next turn (line 56) and clarified his question so that speaker A understood that speaker D was, in fact, asking about his job training.(what for ?/I don't think I need to go around ) 224.C :malas, (.) nggak ada duit, bajet (.) tipi:s, (.) gitu aja (don't feel like it (.) no money (.) budget (.) short of (.) just that ) The data also show that the current speaker may continue holding the floor if the two techniques, which are, the current speaker selects the next speaker and the next speaker selfselects himself are not applied by the speakers.In the following example, Speaker C (line 65) selected himself to take a turn to response to A's talk (line 63).Speaker C might hope that speaker A or B would then take the turn but in fact speaker A produced a minimal response 'ahh' which was not actually intended to take a turn.Speaker C (line 67) then continued his turn to repeat his point by paraphrasing his former utterance but still speaker A and B did not take the turn at the transition relevance place.After 2.0 seconds of pause, speaker C then continued his talk by changing the topic.An interesting point from this data is that why speaker A and C did not take a turn to give response to C's talk.The response produced by speaker C (line 65 and 67) might be interpreted by A and B as a criticism for A so that it might make speaker A feel uncomfortable.Therefore, A and B did not intend to take the turn to give response to C's utterance. 62

Indicators of Ends of Turns
The data show that turn-yielding signals or indicators of ends of turns are indicated by the falling and rising intonations, and the use of tag particles such as 'kok', 'la ji', ' kan', 'dih', 'gitu' at the ends of turn-constructions units.
In the following example, A (line 183) self-selected his turn, and he ended it with a rising intonation, as shown with capital letters.Even though his sentence is not in a question form, it functioned as a question.Speaker C (line 184) responded to A's query using the same type of utterance.In turn, A (line 185) replied but with a slightly lower intonation.The rising intonation is very common for asking a question.In this case, the current speaker may select a particular person to take a turn if the question is directed the particular person, but the current speaker may also allocate the turn to any persons in the conversation.If the second alternative happens, overlaps sometimes occur in the conversation.
The following example shows that speaker B (line 2) selected himself to take a turn.He asked a question to other persons in the conversation.Since he did not select a particular person, speaker C (line 3 ) and B (line 4 ) talked or responded B's question at the same time so that overlap could not be avoided.The response produced by speaker C is not heard or is not clear but the one produced by speaker A is clear.However, speaker A might not understand what B was asking.As we see from the data (line 2) speaker A did not finish his question.He did not produce a complete question so that he did not also use a rising intonation.Therefore, A was asking for a clarification by asking a question 'apa ?' (what ?).However, B's response which was intended to give a clarification (line 5 ) might not also be clear for speaker A; therefore, A did not response / did not take the turn directly.Speaker A (line 7) gave his response / answer (i.e. which is the emphasis of C's answer) until after C (line 6) gave a response to B's question.
The use of particle 'la ji' produced by speaker A is meant to give an emphasis.
(it's no problem , isn't it ?) Tag particles are also used by the speakers of 'Bahasa Indonesian' to end their turns.
Each of these particles is literally meaningless but when it is attached to a sentence, it alters the degree of politeness of the sentence.In addition to this, certain particles, such as 'kan' and 'ya' can function as the formatives of a tag question when they are attached at the end of a sentence.
Therefore, they indicate the completion of a syntactical construction.The data show that almost all sentences which end with a particle are followed by the change of a speaker.Look at the use of the particles (ie.la ji (line 7), kan (line 57), kok (line 58) ) in the following examples.
(why not eating at the one downstairs (at the canteen ) ) 57. D : udah (.) udah masa trial lu sudah selesai kan ?(your training period is finished, isn't it ?) 58.A : gua berhenti kok ( I quit ) 59. D : keNAPA ? (why) In the example above, speaker A (line 7 and 58) ended his turn by using the particle 'la ji' and 'kok'.These particles are used by the speaker to show that he ended his turns and he might want other persons to take a turn.This is different from the use of the particle 'kan' (line 57) which is used by the speaker to ask for response or answer from the next speaker.Speaker D (line 57) who ended his turn with the particle 'kan'might want the next speaker to give an answer to his question.D's question was responded by A who gave the explanation concerning to his training /job.

Repair Mechanism
Turn-taking system in 'Bahasa Indonesia' is predominated by one speaker speaks at a time.When there are interruptions, the parties involved in the conversation employs several ways of repairs.One technique found in the data was repeating a part of theformer utterance as shown in the example below.
102.D :tadi gua (.) itu pas ke kamar lu nyari (.) lha Yung, Yung mana Yung?= terus gua tanya si: itu Vivi [(.) di bawah itu kan lagi main sam aitu: (I went to your room to find Yung but I couldn't so I asked Vivi who was downstairs playing with .. ) 103.A : [hm 104.A :Vien 105.D :sama yang Afrika itu: (with the African ) In the example above, A (line 104) self-selected to make a correction, that is, by completing D's utterance (line 102).After A terminated his turn, D (line 105) self-selected a turn and repeated part of his previous utterance.Self repair mechanism by repeating part of an utterance can also be seen in the following example.Speaker D (line 210) started talking when C had not finished his turn so that it results in the occurrence of an overlap.In his utterance, D self repaired by cutting off his utterance, that is, 'ud-udah'.In addition to this, he also self repaired by repeating some parts of his utterance ( forward repair ), that is 'udah kayak'.

Gaps and Overlaps
The data show that of the 230 turn exchanges there are 17 overlaps, 5 of which occur at a transition-relevance place and 12 overlaps occur very close to a transition-relevance place.The overlaps at a transition-relevance place happen when the current speaker did not select a particular person to take a turn.This could happen when the current speaker asked a question to other persons in the conversation.Consequently, two speakers might take a turn at the same time when they gave response to the current speaker's question.In the following example, speaker C (line 9) and A (line 10) took a turn at the same time when they gave a response to B's question (line 8).
8. B : =koknggak yang (.) dibawah ?, (why not having the one downstairs ) 9. C : [mahal ( it's expensive) 10.A : [(ma ) In the following example, speaker D (156) continued to hold the floor (i.e. after 1.8 seconds pause) because he might feel that no other speakers would take a turn.However, speaker A (157) at the same time took a turn so that an overlap could not be avoided.The data also show that overlaps could occur very close to a transition-relevance place.
It happens because the next speaker takes a turn when the current speakers' utterance is very close to a transition-relevance place.In this case, the overlapped part is usually brief.In the following example, the overlap occurs between D and A's utterance.Speaker A started a talk before D finished his turn because A might think that he knows how D's utterance is going to be completed.Therefore, A could gave a response to D's utterance before the utterance is accomplished.In this case, A could project how D's utterance is going to be accomplished.The data also demonstrate that most turn exchanges occur without gaps.Of the 230 turn exchanges, there are only 11 gaps, occuring between one to two seconds long, and 13 gaps of less than one second long.This shows that the speakers, who are really well-acquainted, tend to minimize gaps during their speech transition.

The Relevance of Turn-Taking Model by Sacks et.al for the Indonesian Data
In this part I am going to discuss whether the turn-taking system in Indonesian conversation as we have analyzed above relevant to the turn-taking model as proposed by Sacks et al (1974).First of all, let us see the Sacks et.al. turn-taking model. Sacks et al (1974: 700-701) propose a model of turn-taking in any conversation which they think must include the following facts : (1)The occurrence and the recurrence of speaker change, (2) the overwhelming tendency for one party to talk at a time, (3) the occurrences of more than one speaker at a time are common, but brief, (4) a speaker switches with no gap and no overlaps are common, (5) the techniques of turn-allocation are obviously used, or a current speaker may select a next speaker (as when he addresses a question to another party ) or other persons may self-select in starting to talk, (6) talk can be continuous or discontinuous, (7) repair mechanisms exist for dealing with turn-taking errors, and therefore,if two persons speak at the same time, one of them will stop prematurely, thus repairing the trouble.
If we compare the turn-taking model with the findings in the analysis above, we will see that the turn-taking behaviors of the Indonesian speakers confirm the turn-taking system proposed by Sacks et al.The occurrence of one party talks at a time is very common in Indonesian.The data analysis above shows that most turn exchanges occur without overlaps.The occurrence of overlaps, however, is also common but they are usually very brief.The data indicate that the form of the overlapped part is a word or a part of a word, and so it will not affect the running of the conversation.This fact also implies that the speakers follow some rules or norms in their turn taking by letting others know what they are doing, and this is very common in other cultures (see Nofsinger, 1991) The Indonesian data also show that the speakers of the Indonesian language employ the techniques of turn-allocation as proposed by Sacks et al, which are a current speaker may select a next speaker (i.e. when he addresses a question to another party) or parties may self-select in starting to talk.Finally, the data also demonstrate that repair mechanisms exists dealing with turn-taking errors, that is, when overlaps occur between two persons as they talk at the same time.

Conclusion
We have discussed turn-taking behavior of the speakers of Indonesian, especially in an informal conversation.A number of features which characterize the turn-taking system have been analyzed and discussed.The analysis shows that the turn-taking system in Indonesian confirms the model of turn-taking system developed by Sacks et al (1974).The analysis also shows that the speakers of Indonesian employ turn-yielding signals as cues they use when they 209.C : di sininya berat badannya kayaknya nam[bah ( here I feel like to gain my weight ) 210.D : [ud-udah kayak ha ha udah kayak kingkong lu (you are (fat ) like Kingkong ) This also happens to speaker D (line 115 and 116) who terminated his turn by lowering his intonation at the end of each turn-construction unit.
In the following example, speaker A (line 112, 114, and 116) ended his utterance by a falling intonation.