Effect of Context on Types of Hesitation Strategies Used by Iranian EFL Learners in L2 Oral Language Tests

Hesitation strategies appear in speech in the form of filled or unfilled pauses, paralinguistic markers like nervous laughter or coughing, or signals which are used to justify units in the coming utterances in which the speaker struggles to produce. The main functions of these forms of hesitation strategies have been associated with speech planning or accessing speech difficulties. The present paper reports on a study which investigated the effect of context on hesitation strategies employed by Iranian university students. Previous studies on hesitation strategies used by beginner or advanced L2 learners revealed that beginners mostly leave their hesitation pauses unfilled which cause their speech to sound disfluent, and advanced learners tend to use various fillers in order to sound like native speakers. The study examines this phenomenon across different contexts in an oral L2 test situation. The respondents were a group of students registered in the Tertiary English Language Program at a university in Kuala Lumpur, Malaysia. The frequency and use of silent pauses, silent pauses and fillers, fillers, and non-lexical words were examined in four different L2 oral tasks. The aim was to identify the effect of context on the types of hesitation strategies employed by the EFL learners in speaking tests.


Introduction
Very often, as we speak we change our minds about what we want to say and as a result hesitate.There is also no doubt that only few of us speak completely fluently without hesitating or making slips.Spontaneous speech therefore is often intermingled with pauses, hesitations, "err words, truncated words, repetitions, prolonged sounds, repairs, etc.In fact, "There is some sense in the idea that one of the very first things to learn in a foreign language is how to hesitate in it."(Schmid, Groningen, Fa¨gersten, & H¨ogskola, 2010) Bejarano, Levine, Olshtain, &Steiner (1997), andClennel (1995) said that using communication strategies can improve learners' skills for interpersonal communication (Nakatani, 2005) Since 1970s, a major focus of second language acquisition (SLA) research has been on communication strategies.Faerch and Kasper (1983) were pioneers in proposing that communication strategies are verbal plans within a speech production framework.They adapted a psycholinguistic approach in their study, and emphasized on two stages of production: "a planning phase and an execution phase".(cited in Wnnaruk, Prinyajaran, & Suranaree, 2008).They claimed that the planning phase involves developing a plan to achieve a communicative goal.During this stage, the language user selects those rules and items which can best contribute to the plan of making "proper" verbal behavior in line with the original communication goals.At the execution phase, learners resort to communication strategies in order to cope with executing problems in their planning phase.

Disfluencies
Disfluencies are strategic devices which signal the speaker's "under construction utterance".They have also been characterized as the automatic effects of cognitive burdens, particularly during speech production management (Nicholson et al, 2003).Spontaneous speech contains all kinds of disfluency phenomena such as silent pauses, hesitations, repetitions, fillers, grammatical errors, misselected lexical items, self-corrections, prolongations, false starts, slips of the tongue, etc., which occur because of disharmony between speech planning and execution stage.In fact, speech disfluencies are defined as phenomena interrupting the flow of speech without adding propositional content to an utterance.(Menyhárt, 2003) Therefore, if we consider that a native speaker's knowledge of language is imperfect, then L2 speakers will definitely have to cope with even bigger and more complicated communication problems.Research has indicated that less proficient language users encounter many communication problems.(Kaivanpanah, Yamouty, & Karami, 2012)

Hesitation
Hesitations are pauses of varying lengths, which are not usually left unfilled.They usually occur when a speaker finds himself in a position where he lacks the words to use or struggles with cognitive or verbal planning.Even native speakers fill hesitations when they speak and use fillers including non-lexical fillers like lengthening or stretching sounds, quasi-lexical fillers, repeating lexical items, and finally lexical fillers.(Rieger, 2003) Pauses in spontaneous speech have provided a rich source of data for several disciplines.They have been used to improve "automatic segmentation of speech", "classification of patients with acquired communication disorders", "the design of psycholinguistic models of speaking", and the "analysis of psychological disorders".Hesitations can mostly be found in conditions in which the speaker cannot find desired words to use and appears in either the cognitive or verbal planning stages.Native speakers have a bank of various fillers to use whenever they need to and these fillers might be "non-lexical fillers" -which appear in the form of lengthening or stretching some sounds-"quasi-lexical fillers", "repetitions" of one or several lexical items", and "lexical fillers".
Hesitation strategies have been categorized under the large class of 'disfluencies' or 'self-repairs'.Conversation analysts identify fillers as self-repair strategies, repair initiators, or indicators.
Hesitations carry little or even no meaning by themselves and can only be understood by clues in the context or situation.This implies that pragmatic markers mainly function to monitor discourse and conversation.(Erman, 2001) Brighton (1996) states that hesitations like "you know"," you see", and "sort of" have often been 'stylistically stigmatized and negatively considered specially in written or formal discourses'.As a result, they are not expected to be found in academic lectures.Similarly, Webber (2004) claims that "you know" and "I mean" are quite rarely used outside casual conversation.(Erman, 2001)

Mechanism of Disfluencies
Disfluencies usually occur in stuttering, blocks, prolongations, hesitations; (part of/complete) word repetitions, and self-corrections.These are all related to self-monitoring processes in which speakers check their speech quality.In other words, disfluencies are evident in individuals who stutter and suffer from planning problems within their internal speech, and often, as they attempt to correct their errors, they cannot prevent disfluencies from occuring.Monitoring accounts have attempted to explain continuity between those who stutter and those who do not: disfluencies in both these groups are caused by the same mechanisms, but emerge more frequently among those who stutter.Phonological deficiencies cause a lot of phonological speech errors internally which are identified and edited by the speaker's self-monitor.In fact, disfluencies are the product of an editing phase (interrupting and restarting).(Robert, Hartsuiker. Corley, Lickley, Russels, 2003) Looking from another perspective at how disfluencies occur, as a speaker selects a word, other phonologically related and similar words like the intended word are activated.At first, the activation of these words is similar, although when the activation is complete, the intended word is of a higher asymptotic value.Kolk & Postma(1997) discuss that in case a response is made during the phase in which activation is in process (rather than at full activation), the probability of competing rather than selecting the intended word increases which consequently leads to error occurrence.They suggest that a speaker recognizes these errors through the use of the perceptual system in case they are produced overtly, and then a monitor existing in the learner's linguistic system, responds by interrupting and starting correction.Thus, word repetition and hesitation (not actually errors in themselves) have been considered as indicators of underlying errors which are identified and interrupted before they manifest in the form of overt errors in speech output (cited in Howell, 2003).According to the editor theories of monitoring, there is a monitoring system in the speaker's cognitive domain which demands the existence of an editor to replace incorrect speech output through the production processes.It is likely that the editor itself contains a system of rules on its own which checks the output.(Komos, 1999)

Previous Studies on Hesitation
The earliest formal study of disfluency commenced in 1950s in three separate disciplines.Wendell Johnson and his colleagues were pioneers who studied stuttering as a form of hesitation.Then linguists, like Frieda Goldman-Eisler (1958) were attracted to this idea too.Disfluency was also studied within psychotherapy and can be best remembered by the work of George F. Mahl (1956) and colleagues who carried out several studies in this field.
Early studies of hesitation strategies in L2 conversations focused on L2 beginner learners, and the findings proved that beginners are more interested to leave their hesitation pauses unfilled which consequently causes their speech to sound disfluent.On the contrary, native speakers use various kinds of fillers to fill their hesitation pauses, like "the lengthening of sounds", "quasi-lexical fillers (uh, uhm)", "lexical fillers (well, you know etc.)", and "repetitions".Bilinguals also use a variety of fillers in their native language as well as in their non-native language.In fact, they use idiosyncratic fillers in their L2.Idiosyncratic fillers are those words uttered by a speaker based on his/her own knowledge to fill in pauses in speech.It is usually a lexical filler which is used more often than all other lexical fillers.They are as flexible as, but at the same time seemingly "classier" than, "quasi-lexical fillers" which are used by bilinguals in their non-native language due to overgeneralizations and also productions of 'uhs' and 'uhms.'Fillers are also counted as a part of self-repair strategies which need to be further studied by conversation analysts who occasionally consider them as special cases which do not need to be analyzed.
The suggestion that the patterns of filled pauses might be language-specific has been proven in researches which were carried out in many different languages; Dutch (de Leeuw, 2004), French (Dewaele, 1996;Duez, 1982), German (de Leeuw, 2004;K¨unzel, 1997), Italian (Giannini, 2003), Japanese (Watanabe & Ishi, 2000), Korean (Trofimovich & Baker, 2006), Russian (Riazantseva, 2001;Stepanova, 2007), Spanish (Edmunds, 2006), and Swedish (Horne, Frid, Lastow, Bruce, & Svensson, 2003).(cited in Schmid et al., 2010) Hieke (1981) was one of the first few researchers who discovered that non-native speakers use more self-repairs compared to native speakers.Wiese (1984) found out that L1 and L2 productions entailed different processes in his study of self-repairs.Wiese also showed that L2 speakers employed a larger number of self-repairs than L1 speakers did.He stated that L2 speakers' error in speech is more than L1 speakers', and L2 speakers tend to correct their own errors more than L1 speakers do.He also proposed that L2 speakers required more time to plan their utterances due to their inadequate knowledge of their L2, and they showed less automatization in processing their second language when compared to what they did in their L1.O'Connor (1988) studied the speech of beginner and advanced L2 learners and found out that beginners use fewer self-repairs than advanced learners.They tend to employ various kinds of self-repair such as corrective repairs (rather than anticipatory repairs (covert repairs)), but advanced learners utilize more anticipatory self-repairs.Temple (1992) focused on self-repair in the speech of L1 and beginner L2 users.She analyzed speech and repair frequency in both groups and found out that native speakers seem to speak twice as fast when compared to non-native speakers because of the frequent and skillful application of fillers.In contrast, non-native speakers mostly leave their hesitation pauses unfilled, produce more false starts, and leave more errors uncorrected.(Rieger, 2003) This pattern was also indicated by two studies which compared the use of filled and unfilled pauses between speakers of American English and German (O'Connell & Kowal, 1972;O'Connell, Kowal, & H¨ormann, 1969).In these experiments, a group of speakers were asked to read some stories and retell them.It was found that some of the stories were phrased predictably but the rest contained contextually unexpected sentences.Interestingly, all speakers produced more and longer unfilled pauses in unexpected reading conditions and longer pauses in the unexpected retelling conditions, which supports the belief that the occurrence and length of unfilled pauses increase in cognitive tasks.The results also showed that pauses were only found in the retelling tasks (but the reading tasks did not show many filled pauses), and for this disfluency marker, group differences were identified: While the Germans made more filled pauses in the unexpected conditions than in the predictable ones, the American English speakers showed the reverse.There was no major increase of filled pauses in the more demanding tasks.
Even though hesitation markers like unfilled pauses, retractions (withdrawing), and repetitions have been said to have a non-language specific function, and that they merely signal or resolve the cognitive problem of a speaker's lexical or information retrieval, studies have shown that CDMs (cognitive disfluency markers) increase in those contexts where the speaker deals with a cognitively more challenging and complicating tasks.(Schmid et al, 2010)

Function of Disfluencies and Hesitation
Although disfluencies have been considered as a negative phenomenon which interrupts communication and wastes listeners' time, studies have shown that disfluencies might contribute in a positive way to a more efficient communication by giving extra time to the speakers to plan, and inform the listener about the mental attitudes of the speaker and planning difficulties faced by the speaker.When listeners expect the speaker to carry on talk instead of taking his/her turn, they might also understand reason for the hesitation and may even be able to predict the coming utterance.This will finally lead to two possibilities: whether to prepare for it or to offer help to the speaker to resolve the difficulty (Clark, 2002;Shriberg, 2005;Stenstroem, 1994).Filled pauses have been found to be frequent in dispreferred responses or embarrassing remarks (Finegan, 1994;Rose, 1998;Sadanobu and Takubo, 1995).(Watanabe, 2007), and silent pauses have been known to facilitate breathing, and enable the speaker to harmonize his/her speech processes, and at the same time allow the listener to better comprehend and digest what they have heard.Other types of disfluency phenomena are known as "errors", which almost always are distracting for the listener.Recently, a study on hesitations and disfluencies in speech, showed there is an instance of disfluency for every six words in spontaneous speech, although, in longer monologues the frequency of error is for every 3.6 words.( this does not include silent pauses) (Menyhárt, 2003).

Participants
In order to investigate the pragmatic markers Iranian EFL learners use during hesitation, the researcher selected a population of TEP (Tertiary English Program) students in a public university in Kuala Lumpur, Malaysia, who had already taken the IELTS exam and received a score of 5.5.Thus, the language competence of the population was almost the same, but their language background was checked for a higher congruency scale.The instrumentation which was used at this stage was a LBQ questionnaire (language background questionnaire), which helped the researcher to identify the most congruent participants in terms of their language background.The respondents were six males and six females whose first language was Persian, and these respondents had learned their L2 (In this case, the English language) initially at school, and then continued in language institutes or private classes in their home country.
Each participant had to take part in an oral L2 test consisting of 4 parts: Introduction, Conversation (General) Questions, Retelling a passage, and Picture description.Each of these tasks took about 10-12 minutes.
The collected data yielded about 140 minutes of English interaction between the student as the subject and the researcher played the part of an interviewer.

Research Instrument
The research instruments which were used in this study included a Language Background Questionnaire to find the most congruent subjects regarding their language background, twelve unseen passages (never seen before by the respondents) for a retell after a short time reading, twelve unseen pictures (never seen before by the respondents) for a description based on the participants' imagination, and finally 3 sets of general questions extracted from the assessment database of a language institute (with the prior permission granted from the institute).

Data Collection
For the data collection phase, the researcher recorded the sessions and then transcribed the recordings of the interview sessions.These recordings captured all pauses and even incoherent sounds the respondents produced.The recordings were checked several times in order not to skip even a short silent pause.The researcher then identified the hesitation strategies used and coded them as drawling, pauses, repeating words, using hesitation filler words and producing incoherent vocals.Not all pragmatic markers like "I think" indicated a hesitation strategy, so the researcher had to ensure that for the data analyzed the data actually functioned as hesitations in the utterances.Employing frequency counts on the transcribed

Method
The sessions with the respondents were digitally recorded.The findings of the pilot study contributed to the reliability of the results by showing that the participants needed to read the passages in a timely manner that is around around 1-2 minutes to get a gist of what they had to read.Also, the subjects of the passages were varied to prevent the passing and sharing of the general idea of the texts among the respondents.The topics centered on social, historical, and environmental issues.
Each session began with some explanations about the format of the test, and proceeded with conversation questions.This later continued with retelling a passage as the third part, and finally describing a picture was the fourth part of the test.

Results
The study aimed to investigate the effect of contexts on hesitation strategies employed by Iranian EFL learners during an oral L2 test.It was proposed in previous studies that the context or to put it in a more clear way the type of question, affected the frequency and type of hesitation strategies used.Respondents for the study were interviewed in four different contexts: Introduction, Conversation Questions (also called "general questions"), retelling a passage, and picture description.Interestingly, the total frequency of hesitation strategies differed one task in the oral test to another task in the test.This is illustrated in the following tables and graphs.  1 shows that the context of general questions show the highest rate of use of hesitations (228 times), and the smallest number of hesitation strategies were for the introduction section (59 times).To illustrate this further, the graph below is referred.(Figure 2) .This could mean that respondents of the study produced the highest number of hesitations when replying to "General Question" as they were more challenged for this task.Speaking about an unseen passage has been shown to be more hesitation rising than speaking about an unseen picture, and also self-introduction did not yield a big frequency of hesitation discourse markers perhaps because of the participants' background knowledge about it.To analyze the effect of context on hesitation strategies, all of the four contexts were analyzed in detail and the following results were obtained.

Introduction Section
The summarized data in Table 2 refers to the findings about the frequency of use of hesitations in the "Introduction" section.
Table 2. Total frequency of hesitations in the introduction section As seen in Table 2, the EFL learners produced quite a lot of hesitation filler vocals when they were introducing themselves, and the related count is considerably higher when compared to the use of the other hesitations.Then, "drawling", is shown to be the second most frequent hesitation with the frequency of nine.The respondents also used pauses (4 times) when introducing themselves, and interestingly, "Repeated Words" and Hesitation Filler Words" were not used for this task.

Conversation Questions
The frequency of hesitation strategies used by the participants in response to the "Conversation Questions" is summarized in Table 3.The participants of this study produced hesitation filler vocals of a much higher count compared to the other hesitations.The findings show that drawling is the second most frequent form of hesitation used by the Iranian EFL learners, and the lowest number of hesitation is for the use of Pauses for this section.

Speaking about a Passage
Table .4summarizesthe results of analysis on hesitation strategies employed by the participants while retelling a passage.According to the data seen in table.4,HFVs were recorded as the most frequently used hesitation strategy by the respondents while speaking about a passage.The second mostly frequently occurring hesitation is drawling (25 times).

Speaking about a Picture
This section presents the findings about "Picture Description" when the participants were asked to develop their ideas about an unseen picture.The results are summarized in Table 5.Table 5 shows that the Iranian EFL learners produced a considerable number of incoherent vocals while hesitating in this section.This could imply that they could use this native-like strategy in this section more frequently compared to the other contexts of the oral task.

Conclusion
The overall aim of this paper was to investigate whether context affects the frequency and type of hesitation strategies employed by EFL learners in an oral L2 test.The findings of the study showed, the type of questions and background knowledge are factors which can affect the production of hesitations.Based on the results, the highest frequency of hesitation discourse markers was found across the utterances in reply to general questions, but the lowest rate was recorded in the introduction section of the test.Thus, it can be concluded that background knowledge can affect hesitations.The more background knowledge one has, the less he/she hesitates.For instance, the participants already knew what to talk about themselves in the introduction part, so they did not hesitate much in that part, but as they were asked some unpredictable questions in the second part of the test, they got involved in the planning process of an utterance which resulted in an increased use of hesitation discourse markers in their speech in the L2.A comparative analysis between the frequency of hesitation markers across the four sections, proved that after general questions, retelling a passage seems more challenging than picture description, probably because when they read a passage and are expected to speak about it, they are bound to use the ideas in the paragraphs which mostly depends on their memory recall, but when they have to describe a picture, they have the freedom of talk, so they can rely on a bigger range of vocabulary and background knowledge, which might lead to a lower rate of hesitations.It does seem that understanding second language learners' use of hesitations is important as it can provide salient information on the cognitive aspects and challenges posed by different contexts of what happens when one has to speak in a language that is not one's native language.
phrases for native and non-native listeners.science direct.

Figure 1 .
Figure 1.Procedures employed in the study

Figure 2 .
Figure 2. Frequency of hesitation strategies in the four contexts

Table 1 .
Total frequency of hesitation strategies in the 4 contexts investigated

Table 3 .
Frequency of Hesitation Strategies in Conversation Questions

Table 4 .
Total Frequency of Hesitation Strategies in Speaking about a Passage