649 Studies in Second Language Learning and Teaching Department of English Studies, Faculty of Pedagogy and Fine Arts, Adam Mickiewicz University, Kalisz SSLLT 7 (4). 2017. 649-671 doi: 10.14746/ssllt.2017.7.4.5 http://pressto.amu.edu.pl/index.php/ssllt Discourse markers in non-native EFL teacher talk Gloria Vickov University of Split, Croatia gvickov@ffst.hr Eva Jakupčević University of Split, Croatia ejakupcevic@ffst.hr Abstract The present study aims to investigate the use of discourse markers (DMs) in non- native (Croatian) EFL teachers´ talk with primary and secondary school students. The study concentrates on the occurrences and frequencies of DMs, but it also provides an account of the function distribution of the three most frequently used DMs (ok, so, and). The quantitative and qualitative analyses of the recorded tran- scriptions reveal that the teachers use a variety of DMs, almost exclusively the ones typical of classroom management and classroom discourse organization, with no significant differences in the patterns of DM use with the primary and secondary school students. The DMs fulfill a number of structural and interper- sonal functions mainly aimed at providing coherent and stimulating classroom discourse. The findings of this study might contribute to raising awareness of the diversified functions of DMs, which could facilitate non-native EFL teachers´ over- all lesson organization and structuring of particular teaching segments. Keywords: discourse markers; non-native (Croatian) EFL teachers; classroom discourse 1. Introduction Discourse markers (DMs hereafter) are multifunctional linguistic units with a principally non-propositional function and serve to connect segments of spoken Gloria Vickov, Eva Jakupčević 650 and written discourse (Fraser, 1999; Müller, 2005). They are syntactically inde- pendent and semantically optional, their main role being to maintain the cohe- sive relations in discourse. In addition, DMs are used to achieve a variety of in- teractive functions essential to the relationship between discourse participants, such as politeness, repairs, attention-getting and feedback (Iglesias Moreno, 2001). Functions of these units thus exist both on the textual and interpersonal levels, facilitating understanding for the discourse participants, guiding them through interpretation, and helping them to choose the right meaning out of all possible meanings (Aijmer & Simon-Vandenbergen, 2006; Brinton, 1996). Theoretical accounts of DMs are marked by an extraordinary diversity of terminology, definitions and taxonomies. Over the years, these units have been referred to as pragmatic particles (Ostman, 1981), discourse particles (Schourup, 1985), discourse connectives (Blakemore, 1992) and pragmatic markers (Fraser, 1996), to name just a few. Furthermore, a variety of theories and perspectives have been used to consider discourse markers, such as the theory of discourse coherence (Schiffrin, 1987), relevance theory (Blakemore, 1992) or the grammat- ical-pragmatic perspective (Fraser, 1996). However, in spite of this diversity, it is generally agreed that these units contribute to the pragmatic meaning of utter- ances and thus play an important role in the pragmatic competence of speakers (House, 2013; Lenk, 1998; Müller, 2005), enabling them to use language in a way that is culturally, socially and situationally appropriate (Fung & Carter, 2007). Con- sequently, insufficient and incorrect use of DMs may reduce communication ef- fectiveness and may cause failure in interpersonal and intercultural interaction (Martinez, 2004; Wierzbicka, 1991), leading the speaker to be perceived as impo- lite, and his social behavior or commitment to the conversation to be brought into question (Crystal, 1988; Iglesias Moreno, 2001; Lam, 2009). The importance of DM use in everyday native speech is an indicator of its relevance for foreign language learners who need to be aware of the textual and interpersonal effects achieved by the consistent use of DMs in order to com- municate adequately in specific contexts (Furko & Monos, 2013; Müller, 2005). But despite the fact that natural communication in a language largely depends on the appropriate use of DMs, these linguistic units appear not to receive enough recognition in foreign language teaching materials (Lam, 2009; Müller, 2005; Vickov, 2010). Existing EFL studies largely point to non-native students´ poor command of English DMs at different proficiency levels, in that they use DMs less or differently, or use a narrower range of these units than native speak- ers (Buysse, 2012; Liao, 2009). This seems to be caused by a variety of factors, one of them being artificial or inappropriate language input in the classroom (Fung & Carter, 2007; Müller, 2005). Even though language teachers are said to be responsible for about two thirds of classroom discourse (Ellis, 1994) and their Discourse markers in non-native EFL teacher talk 651 language use is very likely to be at least as influential as the textbooks are (Mül- ler, 2005), there is a serious lack of research focusing on non-native teachers’ use of DMs (Yang, 2011). In an attempt to shed more light on this issue, the present study investigates the use and functions of DMs in Croatian EFL teacher talk with students at two proficiency levels, in the final years of primary and secondary school. The findings of the study will allow for a better understanding of DM use in non-native EFL teacher talk. The study may also prove useful in defining the characteristics of EFL classroom discourse as well as in providing guidelines for teacher preparation courses. 2. Discourse markers in EFL classroom context Over the last decade, the interest in the process of acquiring and using discourse markers by non-native speakers of languages, especially English, has generated a considerable amount of research (e.g., Buysee, 2012; Fuller, 2003; Fung & Carter, 2007; Hellermann & Vergun, 2007; Lam, 2009; Liao, 2009; Müller, 2005). However, although English has been increasingly learned as a foreign language taught by non-native speakers of English in a formal pedagogical setting, there has been notably less research conducted in the classroom context (Chapeton Castro, 2009). The existing empirical research is mostly focused on EFL/ESL learners´ use of DMs in spoken language and is mainly conducted on small sam- ples of learners at the advanced level. The findings suggest that EFL learners use DMs less frequently than native speakers (Weinert, 1998), and often use them in inadequate ways, that is ways in which they would never be used by native speakers (Yang, 2005). Furthermore, some of the findings are indicative of an overuse of particular DMs such as, for example, but, and and so (Paatola, 2002; Vickov, 2010). Possible causes of such language behavior may lie in the unnatu- ral language input in the formal educational context and pragmatic fossilization (Trillo, 2002) or L1 interference (Jalilifar, 2008; Vickov, 2014). It could also be a consequence of an under-representation of DMs in English textbooks and ques- tionable competences of non-native EFL teachers in using DMs in spoken class- room discourse, as pointed out in one of the most influential studies in this area, conducted by Müller (2005). In providing valuable insights into the functions of four discourse markers (so, well, you know and like) in native (American) and non-native (German) discourse on the basis of a substantial parallel corpus of spoken English at an advanced level, Müller has drawn readers´ attention to the link between pedagogical practices and language use suggesting that English textbooks and non-native EFL teachers seem not to provide a fully adequate ba- sis for illustrating native-like DM usage. However, even native teacher talk, as shown by Hellermann and Vergun (2007), seems to be deficient in the domain Gloria Vickov, Eva Jakupčević 652 of DM use, suggesting that the specific and unnatural classroom context might be to blame for the lacking language input. Nevertheless, some studies (e.g., De Fina, 1997; Fung & Carter, 2007; Walsh, 2006) have suggested the idea of DMs being an inevitable part of suc- cessful classroom management and a key part of pragmatic knowledge in gen- eral, making them indispensable not only to language learners but also, perhaps even more importantly, to teachers, who are among the most important sources of language input in classroom discourse. Fung and Carter (2007), for example, consider DMs to form the basic fabric of talk in pedagogical setting. In the con- text of teaching, DMs have been found to aid the comprehension of lectures (Flowerdew & Tauroza, 1995) and, conversely, speech lacking these units can contribute to the misunderstanding of academic monologues (Jung, 2006). Re- search also shows that the importance of DM use for language teachers is two- fold as they simultaneously perform both social and educational functions in the classroom. Firstly, as successful teaching stems from, among other things, suc- cessful interaction management in the classroom (Allwright, 1984), DMs have an especially important role in enabling teachers to structure their discourse, in that way making it possible for learners to interpret the communicative de- mands of the context and participate in activities more successfully (De Fina, 1997). As suggested by Walsh (2006), teachers can use DMs to signal the pro- gression of a lesson and indicate the beginnings and ends of particular teaching stages. This makes the methodology explicit and allows for more planning and thinking time, along with aiding the students´ understanding of both the content and the interaction patterns. Furthermore, DMs also help establish interpersonal relationships in the classroom, creating a better atmosphere for active participation (Chapeton Cas- tro, 2009; Othman, 2010). They facilitate the process of interpretation and social involvement in spoken interaction, and are essential to the maintenance of con- versational cooperation, ensuring that interactions go on smoothly (Lam, 2009). All in all, by reducing understanding difficulties and social distance between the teacher and the learners, these language units contribute to the effectiveness of learning in general and help create a shared space between the teacher and the students (Walsh, 2006). Simply put, DMs in teacher talk can help learners understand not only the teacher’s language but also the purpose of a specific activity and how it relates to the context. This was confirmed, for example, in the study conducted by Chapeton Castro (2009), in which she investigated the use of DMs in the classroom context by a non-native teacher who was shown to use DMs effectively to organize discourse and to fulfil interpersonal functions. In addition to the significant roles DMs play in the management and inter- personal relations in the classroom, the teachers’ use of these units is also vital Discourse markers in non-native EFL teacher talk 653 because teachers teaching in the EFL context are an important and prevailing source of input for their learners. According to Walsh (2006), modeling the tar- get language is one of the key activities that foreign language teachers perform, and the fact that they frequently represent the only source of foreign language input for the learners highlights the importance of studying the actual language they use. The finding that DMs seem to be rather rare or completely non-exist- ent in EFL textbooks (Celce-Murcia, 2001; Müller, 2005) additionally reinforces the importance of teacher talk for the acquisition of these units by EFL learners. However, despite the relevance of teacher talk for the learners´ acquisi- tion of DMs and the rising interest in the research regarding non-native foreign language teachers in general (Llurda, 2005), there is almost a total lack of re- search focusing on foreign language teachers´ use of DMs in the classroom en- vironment (Rongrong & Lixun, 2015; Shahbaz et al., 2013; Yang, 2011). At best, this area is merely touched upon in the context of foreign language learners’ use of DMs (Hellermann & Vergun, 2007; Mihaljević Djigunović & Vickov, 2010; Müller, 2005; Nejadansari & Mohammadi, 2015). Due to all this, the present study seeks to extend the current knowledge about non-native EFL teachers´ use of DMs. 3. The present study The aim of this study is to provide more insight into the use and functions of DMs in Croatian EFL teacher talk with students at two proficiency levels, more specifically, in the final years of primary and secondary school. In the final year of primary school Croatian students are 13-14 years old and have been learning English for at least eight years, while at the end of secondary school they are 17- 18 years old on average and have been learning English for 12 years. The two different education levels were chosen for two reasons. Firstly, they represent important thresholds in the Croatian education system, and secondly and more importantly, we wanted to see if Croatian EFL teachers adapt their language, DM use in particular, to their students’ age and proficiency levels. In other words, taking into consideration the secondary school students’ higher level of lan- guage proficiency, we wanted to see if the teachers ́use of English DMs with the secondary school students would be more intensive and richer than with the primary school students, with respect to the frequency and types of DMs used. In identifying DMs in our study, we relied on the framework provided by Fung and Carter (2007), who define DMs as “intra-sentential and supra-senten- tial linguistic units which fulfill a largely non-propositional and connective func- tion at the level of discourse” (p. 415). Their framework enables a functionally- based analysis and a systematic classification of the different roles DMs perform in the pedagogic register. As argued by Yang (2011), Fung and Carter´s multi- Gloria Vickov, Eva Jakupčević 654 functional framework of DMs is effective in that it “provides a descriptive model to analyze DMs on different levels” (p. 105). It provides a context-based model to analyze DMs from a functional perspective for classroom discourse, which can be thus applied to investigate the use of DMs not only in non-native EFL students´ spoken discourse but also in teacher talk. In this taxonomy, DMs are classified into four categories: interpersonal (e.g., indicating response, marking the attitude of the speaker), referential (i.e., marking textual relationships be- tween verbal activities surrounding the DM), structural (e.g., opening and clos- ing of topics, summarizing, holding the floor), and cognitive (e.g., denoting the thinking process, reformulating, elaborating). Fung and Carter (2007) also pro- vide criteria concerning whether a linguistic item or expression can be defined as a DM. These criteria include position, prosody, multigrammaticality, indexi- cality and optionality and can be summarized as follows: · DMs are difficult to place within a traditional word class; they are drawn from different grammatical and lexical inventories; · they are syntactically optional: DMs can be omitted without the truth value of the utterance being affected; · they are often sentence-initial, although they may occur at the begin- ning, in the middle, or at the end of a discourse unit or form a discourse unit of their own; · they have little or no semantic meaning in themselves; · they are multifunctional; · they are prosodically independent and are largely separate from the ut- terances they introduce; there are prosodic clues that occur with DMs and these include pauses, phonological reductions, and separate tone units; · they function as indexical expressions to signal the relation of an utterance to the preceding context and to assign the discourse units a coherent link. Fung and Carter´s (2007) defining criteria are in line with some of the most highly cited researchers in DM studies such as Schiffrin (1987), Brinton (1996) and Müller (2005). 3.1. Research questions The study addressed the following research questions: 1. What is the frequency distribution of DMs in the classroom talk of Cro- atian EFL teachers? 2. What is the frequency of use of DMs in primary and secondary class- rooms? 3. What are the functions of the most frequently used DMs? Discourse markers in non-native EFL teacher talk 655 3.2. Method 3.2.1. Participants The participants in our study were six EFL teachers (Ts 1-6), none of whom were native speakers of English and none of whom had spent more than a year in an English-speaking country. Three of the teachers were recorded while teaching primary school students and three while teaching secondary school students. All of them were experienced teachers, with an average of eight years of work experience at school. They all taught in private foreign language schools and worked with learners representing different age groups. 3.2.2. Procedure Two 45-minute classes were recorded for each teacher, although the duration was somewhat varied due to external factors (teachers being late or finishing early), as can be seen in Table 1. In total, 501 minutes (8 hours 21 minutes) of teacher talk in the classroom were recorded, yielding a corpus of 32,681 words. Table 1 The duration of recordings and number of words per teacher T1 T2 T3 T4 T5 T6 Total Minutes recorded 87 73 80 90 85 86 501 No. of words 6974 5098 5266 5145 7285 2913 32681 No. of DMs 1082 474 477 738 590 283 3644 DMs/1000 words 155.15 92.98 90.58 143.45 80.99 97.16 111.50 No. of different DMs 40 28 39 32 34 18 Three of the teachers were recorded while teaching classes in the final year of primary school (8th year of learning), and three in the final year of sec- ondary school (12th year of learning). The researchers were not present during the classes so as to cause as little interference as possible. Instead, the teachers were provided with digital voice recorders and instructed to turn them on, place them on their desks and teach a regular class. The classes included a variety of language activities ranging from grammar and conversation exercises to the ac- tivities focused on developing reading and listening skills. 3.2.3. Analysis For our analysis, quantitative and qualitative methods were combined, ranging from a macro-investigation through the careful reading of the transcripts and using the concordance analysis software AntConc (Anthony, 2014) to a more detailed Gloria Vickov, Eva Jakupčević 656 micro-discourse analytic examination. The recordings were transcribed, with the transcription presented in standard orthographic form. The transcripts were first analyzed in detail by reading through them and marking the occurrences of DMs, based on the definition and framework provided by Fung and Carter (2007). The status of a DM needs to be contextually-referenced, which is why the initial analysis of the transcripts was performed manually. Through this process, a list of 51 identified DMs was compiled. During the identification process par- ticular attention was paid to those linguistic units which only act as DMs in cer- tain contexts, such as or and and, which are often used as conjunctions. Another example, described in more detail in Section 3.3.3.2, is the word so with a vari- ety of non-DM uses which had to be carefully identified and excluded. These units were one of the main reasons why the initial analysis of the transcripts was performed manually; machine-aided analysis cannot differentiate between their uses. However, AntConc (Anthony, 2014) was used after the initial analysis to perform searches for specific DMs and double-check their numbers, and to check the contexts in which problematic DMs (e.g., or, and and so) appeared. The final, qualitative analysis of the most frequent markers was per- formed by reading through the transcripts once more and extracting examples in order to determine their main functions and common patterns of use. As De Fina (1997) noted, markers may assume activity-specific functions which may be different from the ones described by existing research, and it is important to emphasize that the interpretation of their meaning cannot be performed out of context, which is why further careful reading of the transcripts was necessary. We used a combination of top-down and bottom-up approaches in our qualita- tive investigation, looking at the available research and finding examples for the most frequent DMs, as well as supplementing them with the additional func- tions detected in our corpus. 3.3. Results and discussion 3.3.1. The frequency distribution of DMs in the classroom talk of Croatian EFL teachers There were 3,614 occurrences of DMs in our corpus, which amounted to 111.50 DM tokens per 1,000 words. In total, 51 different DMs were identified, although 20 of those appeared in fewer than 10 instances, which is why they were ex- cluded from Table 2. Out of the remaining 31 DMs, 8 of them had more than 100 occurrences in the corpus. Discourse markers in non-native EFL teacher talk 657 Table 2 The most frequent DMs DM No. DMs/ 1000 words DM No. DMs/ 1000 words Ok 755 23.11 Also 35 1.08 So 548 16.77 Oh 29 0.89 And 368 11.27 Excellent 27 0.83 Mhm/aha 265 8.11 I don't know 25 0.77 Yeah/yes 262 8.02 Ah 24 0.74 Right 234 7.17 And then 22 0.68 But 224 6.86 Let's see 21 0.65 All right/alright 140 4.29 Well 21 0.65 Or 96 2.94 Actually 18 0.56 Because/cause 78 2.39 Of course 18 0.56 Now 66 2.02 First 17 0.53 Then 61 1.87 Let's say 16 0.49 Like 55 1.69 Fine 14 0.43 What about 46 1.41 For example 11 0.34 Good 37 1.14 You know 11 0.34 Great 37 1.14 As can be seen from Table 2, four DMs alone, ok, so, and and mhm/aha, account for more than half of the DM occurrences in the corpus. This is in line with previous research where these markers have been mentioned as typical of the classroom environment (Chapeton Castro, 2010; Fung & Carter, 2007; Oth- man, 2010). Moreover, a shared attribute of the most frequently used markers is their high degree of multi-functionality, as elaborated on in Section 3.3.3, which makes their frequency less surprising. When it comes to the teachers’ in- dividual trends of DM use (Table 3), five out of six teachers used ok most fre- quently, while one used so most often. All of the teachers had the following markers among their top ten: ok, so, and and yeah/yes. Table 3 Five most frequent DMs per teacher These results indicate that there was not much variety in the teachers’ use of these top DMs, which is not surprising due to the limitations of the classroom T1 No. DMs/ 1000 words T2 No. DMs/ 1000 words T3 No. DMs/ 1000 words So 249 35.71 Ok 105 20.6 Ok 116 22.03 Right 113 16.21 So 87 17.07 So 68 12.92 And 106 15.2 And 50 9.81 Yeah/yes 48 9.12 All right/alright 105 15.06 But 46 9.02 But 35 6.65 Ok 103 14.77 Yeah/yes 43 8.43 All right/alright 33 6.27 T4 No. DMs/ 1000 words T5 No. DMs/ 1000 words T6 No. DMs/ 1000 words Ok 166 32.27 Ok 201 27.6 Ok 64 21.98 Mhm/aha 149 28.97 And 70 9.61 And 46 15.8 So 67 13.03 But 54 7.42 So 36 12.36 And 66 12.83 So 41 5.63 Mhm/aha 35 12.02 Right 49 9.53 Right 40 5.5 Great 28 9.62 Gloria Vickov, Eva Jakupčević 658 environment and the tasks the teachers are expected to perform within it, such as organizing the flow of discourse, allocating turns, responding to students and evaluating their responses. However, as qualitative analysis will also show (see Section 3.3.3), individual differences were visible in the patterns of use and pref- erences for specific DMs. 3.3.2. The frequency of use of DMs in primary and secondary classrooms To answer the second research question we addressed in our study, we com- pared the teachers´ use of DMs with students at two proficiency levels: primary and secondary school levels. As can be seen from Table 4, the values for the secondary school EFL teachers are higher throughout. They used 2033 DMs in total, as opposed to 1611 DMs used by primary school teachers, but when this is viewed taking into account the number of DM tokens per 1000 words, the difference is smaller: While secondary school teachers used 117.26 DMs per 1000 words, their primary school counterparts used 105. When it comes to the average number of different DMs used, this is also greater for the secondary school teachers (36) as opposed to the primary school ones (28). Table 4 Teachers’ use of DMs: secondary versus primary school Secondary school Primary school T1 T2 T3 T4 T5 T6 Total words: 17338 Total DMs: 2033 DMs/1000 words: 117.26 Average number of different DMs: 36 Total words: 15343 Total DMs: 1611 DMs/1000 words: 105 Average number of different DMs: 28 Although the secondary school teachers used a greater variety of DMs and used them more frequently, there was less of a difference between the two groups than could be expected taking into account the fact that the secondary school students had been learning English four years longer than their primary school counterparts had; it is logical to assume that the teachers’ language should be adapted to the higher proficiency level of the students with a wider range of DMs. The reason for this might be that all of the classes were recorded in private schools, where we could expect a higher level of students´ language proficiency, meaning that they are already quite fluent by the end of primary school, and this can allow teachers to talk more “freely,” thus using a similar number of DMs with the secondary school group. Another reason, however, may be that the teachers´ use of DMs seems to be largely conditioned and con- strained by the specific features of the classroom environment. Classroom talk occurs within specific time boundaries, is guided by social and institutional Discourse markers in non-native EFL teacher talk 659 norms and has precise goals, which imposes certain restrictions upon teachers. They are the ones organizing the discourse, and they plan the procedures within the classroom and are responsible for the management of turns and the evalu- ation of responses. It might be because of these constraints that, in spite of the secondary school students’ higher language proficiency, their EFL teachers failed to use a greater variety of DMs than while teaching at the primary school level. Finally, a possible cause of not having employed a larger variety of DMs in the talk with the secondary school students may simply lie in the teachers’ commu- nication style or even in their own poor competence in this area of language production. Of course, this is only an assumption as the teachers from our sam- ple were not tested with respect to their competence in the area of DM use. 3.3.3. The functions of the most frequently used DMs In addressing the third research question related to the functions of the most frequently used DMs, our study included the qualitative analysis of the markers ok, so, and and. The three DMs were chosen both on practical and theoretical grounds. Similar to Müller (2005), Fuller (2003) and Liao (2009), the rationale for choosing them is related to the fact that they occurred in significant numbers in our corpus for meaningful qualitative analysis. In theoretical terms, we were interested in DMs with a range of different functions which might be relevant for EFL teachers. These markers are frequently discussed in literature as units useful for classroom management (House, 2013; Othman, 2010). Furthermore, they are also among the DMs most commonly used by native speakers in the pedagogic sub-corpus of CANCODE (Cambridge and Nottingham Corpus of Dis- course in English), as determined by Fung and Carter (2007). Finally, the findings of other studies seem to confirm these markers are the prevailing ones in the non-native EFL teacher talk (e.g., Chapeton Castro, 2009; Rongrong & Lixun, 2015; Shahbaz et al., 2013). 3.3.3.1. OK The DM ok is the most frequently and diversely used marker in our corpus, as well as in the talk of five out of the six teachers, which is in line with previous research describing its many functions both outside and inside the classroom (House, 2013; Othman, 2010; Schleef, 2004). According to House (2013), ok is one of the most common, broadly acceptable and adaptable DMs, especially useful for ESL speakers owing to its ability to realize many interactional functions in different positions with minimal linguistic and cognitive effort of the speaker. Gloria Vickov, Eva Jakupčević 660 A detailed illustration of the functions of ok in the lecture genre is found in Othman (2010), who has described its uses both with rising and falling intona- tions. The former marks a progression or confirmation check, functioning as a re- sponse elicitor, a seek of assurance and a device for partitioning different points of information. Ok with a falling tone functions as an attention-getter, especially in transitions between activities. In this way, it can serve as a signpost for opening or closing topics, a tool for shifting lecture mode. Fung and Carter (2007) have listed this function in the structural category, while ok also functions in the inter- personal category, indicating responses such as agreement, confirmation and acknowledgement. Additionally, Chapeton Castro (2009) has listed the use of ok as a turn taker and turn giver, which was very commonly found in our corpus. As used by our participants, ok performs both structural and interpersonal functions, appearing to serve as a useful classroom management tool for all of our teachers. Structurally, it is most commonly used in our corpus in its sentence- initial position, as a topic opener or a means of shifting the lecture mode, holding the students’ attention and making transitions between activities more salient: (1a) T3: Ok, in your notebooks, what’s the last thing that we had written, possibility, probability? (1b) T4: Ok, today our plan is the following . . . In our corpus, ok is very frequently used in combination with other DMs. Within the structural function, it is sometimes followed by now, to emphasize the topic shift/opening function: (2) T3: Ok, now, nominal relative clauses is something you have been using naturally without thinking twice about it . . . The interpersonal uses of this marker are much more common in our sample. Among the most frequent instances is the sentence-final position of ok with ris- ing intonation as a progression check or a seek of assurance, used by the teacher not expecting an answer, but making sure they have the students’ attention: (3) T2: I think you will have the same type of exercise, just to paraphrase, ok? Similarly, this use of ok appears in the sentence-medial position, allowing the teachers to partition the information and in this way make it easier for students to follow the discourse flow: (4) T5: But, um the third unit starts with page 24, ok? And finishes with the with page 31 so actually it's like seven seven pages right? Discourse markers in non-native EFL teacher talk 661 It is interesting to note that the sentence-final function of ok is especially prom- inent with one teacher, T5. For this teacher, the sentence-final position is used almost like a fixed formulaic utterance, repeated after nearly every turn. This might be a result of the teacher’s personal linguistic habits, or it can be caused by extra-linguistic factors such as the content of the class itself (e.g., revising for a test, prompting the teacher to feel the need to emphasize key information), or the fact that the students are teenagers whose attention needs to be maintained: (5) T5: Now, exercise eight um go through these five questions and in like two or three minutes I’ll ask you to answer these questions, ok? Furthermore, ok is used both sentence-initially and finally as a response marker, allowing the teacher to acknowledge what the student has said. However, this is frequently not its only use, as a single DM does not necessarily perform only one function at a time. In her analysis Chapeton Castro (2009) has given exam- ples of ok functioning as a response marker, and at the same time as a turn giver, which was also common in our corpus, especially when the students are answer- ing a series of questions and the teacher acknowledges the answer and passes the turn to another student. Similarly, ok is often used to both respond to the answer and introduce the follow-up question: (6a) T5: Ok, all types of food or just some? (6b) T5: Ok, in McDonald’s, ok, what food can cause addiction? A common response technique in our sample is for the teachers to restate the students’ words after hearing an answer, which is commonly followed by a re- sponse-marking ok: (7a) T2: He was scary, ok. (7b) T6: Mhm the air is polluted, ok. As was mentioned before, ok is commonly used in combination with other DMs. In our corpus, it is most frequently paired up with but, and and so, used to acknowledge what the speaker has said and indicate that the message needs to be modified in some way. With so, what follows is usually another question (as in Example 8a) or a summary of what the student has said (as in Example 8b): (8a) T1: Ok, so what do you mean by real things? (8b) T4: Mhm, ok, so she had to go through the book to find the answers, ok¸ and what about Paula? Gloria Vickov, Eva Jakupčević 662 This combination is used very frequently by T1, who uses it to start a great number of turns with the function of opening or shifting a topic and prefacing questions: (9) T1: Ok, so tell me a little before we continue with the book, how was your prom, so Tom knows what to expect tonight? Similarly, the combination ok, and is used to acknowledge the answer and pref- ace a follow-up question (as in Example 10a) or provide additional information (as in Example 10b): (10a) T5: Ok, and what was um um the experience, was it a positive one, was it a negative one? (10b) T2: Ok, and food can be spoilt as well, right. Finally, the combination of ok and but, which is mostly a marker of contrast, is frequently used by our teachers to correct or modify what has been said by the students (see Examples 11a and 11b). In this way it is used to soften the impact of the correction. (11a) T3: Ok, but we are still talking about a past modal here. (11b) T5: Ok, they were misled by their food, ok, but a group of teenagers sued McDon- ald’s and what were charges. To sum up, ok is the most frequent marker in our sample, accounting for 755 out of 3644 DM tokens (over 20% of all the DM tokens), as well as 23.11 out of 1000 (2.3%) of all the words in the corpus. Ok performs a variety of functions in our corpus, mostly interpersonal, but also structural ones. The examples show that the participants are able to use the potential of this multifunctional marker and augment it by using it in combination with other common markers. Finally, in line with the quantitative and qualitative data on the use of ok in our sample, we can conclude that the use of ok largely depends, among other factors, on the teaching style of individual teachers, as well as on the type of classroom activity, once again emphasizing the importance of taking extra-linguistic factors into account. 3.3.3.2. So Closely following ok, there are 548 occurrences of the DM so out of the total number of 3644 (over 15%) DMs in our sample. This is in line with the pedagog- ical sub-corpus of CANCODE used by Fung and Carter (2007) as the native corpus in their study, where so is also the second most frequent DM. It is important to note that the word so can have a variety of non-discourse marker functions (Müller, 2005), which we took care to exclude. Some of them are listed below Discourse markers in non-native EFL teacher talk 663 with examples from our corpus, shedding light on the common issue of identifi- cation of DMs in general, which can often be problematic due to the multi-func- tionality of many words that operate as DMs. For example, when not a DM, so appears in our corpus as an adverb of degree (as in Example 12a), to express purpose (usually in combination with that, as in Example 12b), in some fixed expressions (as in Example 12c), and as a substitute (as in Example 12d): (12a) T5: Daria and Marija, that's so nice, you have a question? (12b) T1: So like I said always check what is said before and always make sure what is written after so that it fits in completely . . . (12c) T3: I’m just exceptionally smart and intelligent and so on. (12d) T4: No, I don’t think so, it’s B1 level from last year, ok? Similarly to ok, so is versatile as a DM, performing, along with its referential function, a variety of structural functions. In her analysis of EFL speaker data, House (2013) has found that so is used as a useful element supporting the speaker in the planning of upcoming moves, as well as a means of looking back on previous ones. A very detailed analysis of so in non-native speaker speech can be found in Müller (2005), who has listed and exemplified a variety of dif- ferent functions of the DM, some of which are also found in our sample. Firstly, as a referential marker, so is used to mark a result or consequence (as in Example 13a). Although this might be one of the more salient functions of so, it is not as frequent in our sample as the structural functions of this DM. Similarly, according to Müller (2005), the result can be merely implied, and not stated, such as in Example 13b, where the DM is sentence-final. (13a) T4: . . . you won’t be able to use your mind maps, so try to memorize when these things are used . . . (13b) T3: All of them all of them have one, so. As a structural marker, so is used in many ways to help the organization of dis- course and specific moves, for example to open a new topic or to indicate a topic shift (as in Examples 14a and 14b). It can likewise be used to indicate a return to the main thread of discourse after an interruption or digression from a topic (as in Example 14c), where the speaker returns to the topic after interrupting it with the DM right?, at the same time indicating to the listeners that the information which follows is important. (14a) T1: So, let’s start off with we’ll come back and forth, we’ll combine. (14b) T2: So, we’ll start with the second exercise. Gloria Vickov, Eva Jakupčević 664 (14c) T1: So, these are the main parts that you need to focus on before you put your um answer um the missing paragraph, right? So, the main ideas that they want to pre- sent and are usually put here towards the end of the paragraph. Another frequent function of so found in our corpus is to preface a summary (as in Example 15a), introduce a rewording (as in Example 15b) or give an example (as in Example 15c). This use is, not surprisingly, particularly common in parts of the lessons in which grammar or vocabulary are being explained. (15a) T1: to restore control by reintroducing a little disconnectedness, right? So, to not be constantly connected. (15b) T4: . . . and think of your own examples, so don't use examples from the book. (15c) T4: So, deception is like, let's say a lie, but um maybe in a more physical sense . . . However, in parts of lessons which focus more on conversations with the students, so is most commonly used to mark a speech act: preface a question (as in Example 16a) or an instruction (as in Example 16b): (16a) T1: So, what about you, your generation is not that much older than this generation, so um did you also have similar things to keep you distracted from other things? (16b) T3: So paraphrase it give me the same meaning and do the same. As mentioned before, so is sometimes combined with other DMs in our corpus, most frequently with ok, but also with alright. This combination is only used by T1 and T3, although it is T1 who uses it very frequently, almost as a fixed phrase, prefacing questions and summaries: (17a) T1: Alright, so, how could you explain that? (17b) T1: Alright, alright, so, it takes you emotionally, right? Moreover, it is interesting to note that T1 uses so most frequently of all teachers. In fact, with 35.72 occurrences per 1000 words, it is the most used DM by a single teacher. This, together with the variety of uses of this DM shows once again that the teaching style and the individual preferences of the teachers in- fluence their DM use greatly. 3.3.3.3.3. And The third most frequent DM in our sample, and, is not as represented in ESL literature as ok and so. The reason for this may lie in its low salience as a DM, and the possible difficulties this may present in the course of its identification and teaching. Moreover, it does not perform a variety of functions comparable Discourse markers in non-native EFL teacher talk 665 to the previous two, although it is obvious from the quantitative analysis of our data that our teachers find it useful. It is interesting to note that in the pedagog- ical sub-corpus of CANCODE, used by Fung and Carter (2007) as the native cor- pus, and is the most frequently used DM with 11.736 tokens per 10000 words, which is comparable to our frequency of 11.27 tokens. According to Fung and Carter (2007), and can function in three categories: referential, structural and cognitive. The first of those covers the uses of and as a coordinating device (as in Example 18a) or to express addition (as in Example 18b): (18a) T3: But I have my Kindle and I have a dictionary so after class I’ll look it up for you. (18b) T3: OK so this is we are talking about obligation and we said that we have must and have to what’s the difference? Structurally, and is used to denote continuity, as in Example 19a. This use is very com- mon in our corpus as it allows the teachers to keep the turn when necessary. Similarly, in Example 19b, the speaker is continuing a thought interrupted by the comprehen- sion check right? This case is similar to Example 14c, where the DM so is used in the same way. The combinations of right? as a check of progress or understanding and so or and thus aid the listeners in following the discourse flow, making it extremely useful in ESL classroom contexts where the success of the teaching is dependent on the lis- teners’ understanding, hindered by the use of a language foreign to them. (19a) T2: And, if you remember, there was also something else, there were three columns. (19b) T5: All of us watch a lot of movies and TV series, right? And that’s a good practice for you. In a similar sense, the participants in our study commonly use and to preface questions, using it to denote that additional information on a topic is needed, as exemplified above in the combinations of ok and and. In Examples 20a and 20b, we see and used on its own to seek elaboration on a topic from students, which makes it very useful in conversational parts of lessons. (20a) T6: And, where would you live, in town or countryside or village? (20b) T5: And, which bookstore is that? Finally, as a cognitive DM, and can be used by the speaker to denote their think- ing process, such as in Example 21a, where it is used by the speaker to indicate that they will digress from the topic to provide unrelated but important infor- mation, or in Example 21b, where it is used to preface a self-directed question. (21a) T1: And, yes, by the way, since before I forget these papers that you just took now so always have them with you . . . (21b) T4: And what else did I want to mention? Gloria Vickov, Eva Jakupčević 666 To conclude, although lacking the variety of functions presented by ok and so, and is used in diverse ways by the teachers, which shows it is a very useful class- room management tool. Table 5 Functions of the three most frequent DMs DM Category Function Ok Structural Opening and closing topics Indicating topic shifts Holding attention Marking transitions Introducing a follow-up question Interpersonal Progress check Partitioning different points of information Response elicitor Seek of assurance Indicating responses: agreement, confirmation and acknowledgement So Referential Expressing result or consequence Structural Opening topics Indicating topic shifts Returning to main thread Prefacing a summary Introducing a rewording Giving an example Marking speech acts: prefacing questions or instructions And Referential Coordinating device/expressing addition Structural Denoting continuity Prefacing questions: seeking or providing additional information Cognitive Denoting the thinking process 3.3.3.4. Summary of the functions of the most frequent DMs The qualitative analysis of the three most common DMs in our corpus (ok, so, and), as summarized in Table 5, shows that the participants are aware of the diverse functions of these units and use them according to their personal pref- erences and teaching styles. Structurally, all the three markers are used to de- note the discourse flow and help students follow its course by making important points in the teachers’ talk, such as transitions between activities, questions, instructions and explanations, more salient. They are also used in numerous combinations adding to their meaning. Their importance for the organization of classroom activities and discourse cannot be overstated, so it is not surprising that this use is prevalent in our corpus. However, the teachers also use the DM ok to fulfill key interpersonal functions, to check the progress of the discourse and partition information, again aiding students in understanding and following discourse. Also, they frequently use it to acknowledge the students’ answers Discourse markers in non-native EFL teacher talk 667 and contributions, at the same time asking for additional information or offering clarification, thus creating a stronger interpersonal bond with the students. On the whole, the three most frequent DMs in our study are used by the teachers to fulfill functions that help them structure the classroom discourse and estab- lish interpersonal relations with their students. 4. Concluding remarks In the present study an attempt has been made to shed additional light on the use of DMs by non-native (Croatian) EFL teachers. In particular, the study con- centrated on both frequency and function distribution of English DMs in the teacher talk with the primary and secondary school students, in order to gain a full picture of the teachers’ use of DMs as cohesive devices in the discourse gen- erated in classroom interaction. The results have shown that Croatian non-na- tive EFL teachers use a variety of DMs, with 51 different ones identified in our corpus. However, only four markers account for more than half of the total num- ber of occurrences in our corpus. These are the markers ok, so, and, and mhm/aha, which are, along with the equally frequent marker alright, typical of the classroom settings, more precisely, for transitions between teaching activi- ties, giving instructions, asking questions, and so on. The qualitative analysis re- veals that the three most frequently used DMs, ok, so and and, fulfill a number of structural and interpersonal functions which may contribute greatly to the pragmatic flow of classroom interaction. Teachers’ appropriate use of these markers is highly important as they are facilitative of students’ listening compre- hension (Flowerdew & Tauroza, 1995). As used by our participants, these mark- ers help construct coherent classroom discourse and interactional relationships with students, aiding them in following the course of classroom interaction as well as inviting them to actively participate. In this respect, ok has been con- firmed as a useful classroom management tool as it was used as a topic opener and a means of shifting the lecture mode, holding the students’ attention and making transitions between activities more salient. It is also used to partition the information, which makes it easier for students to follow the discourse flow. Furthermore, ok is employed as a response maker, allowing the teacher to acknowledge what the student said, and at the same time as a turn giver. The combination of ok and other discourse markers (and, but and so) enables the teachers to correct and modify their students’ utterances as well as to provide additional information. Similarly to ok, so is used in many ways to help the or- ganization of classroom discourse. Its referential and structural functions range from opening a new topic, indicating a topic shift, indicating a return to the main thread of discourse after an interruption or digression in a topic to prefacing a Gloria Vickov, Eva Jakupčević 668 summary and giving examples. And has proved to be very useful in conversa- tional parts of lessons as it is used to seek elaboration on a topic from students or to denote the speaker’s thinking process (e.g., teachers use and to indicate that they would digress from the topic to provide unrelated but important infor- mation). Structurally, and is mostly used to denote continuity, which allows the teachers to keep the turn when necessary. All of the described uses contribute to the students’ understanding and easier and more successful participation in classroom discourse, which makes DMs invaluable parts of the language of in- struction. The findings of our study seem to be corroborated by a number of similar studies, such as Liu (2006), Walsh (2006), Eslami and Eslami Rasekh (2007), Yu (2008) and Chapeton Castro (2009). The findings of this study might contribute to raising awareness of the di- versified functions of DMs which could facilitate non-native EFL teachers´ overall lesson organization and structuring of particular teaching segments. Further- more, as pointed out by Chapeton Castro (2009), the knowledge of the pragmatic uses of DMs might help teachers to establish more interpersonal relationships in the classroom and to create a more inviting atmosphere for active participation. Further findings of our research show that there is not much difference in the patterns of teachers’ DM use with primary and secondary school students. Although this is probably due to the constraints and specific nature of the class- room discourse, it might also indicate that the teachers lack the necessary com- petence in this area of language production. As a matter of fact, since correct usage of various DMs in different communication situations requires a native- like communicative competence, the question arises as to what degree non-na- tive EFL teachers are capable of illustrating authentic use of English DMs (Müller, 2005). In an attempt to answer the question, we suggest that further research be undertaken in the area of the teachers´ competence in using English DMs in both oral and written language production. We believe that this is important not only for establishing the level of the teachers´ discourse and pragmatic compe- tence, but also for defining the needs and guidelines for teacher professional training and preparation courses. As the present research is a study confined to six Croatian EFL teachers, no claims can be made as to its generalizability. Comparative, large-scale studies of DM use among EFL teachers of different L1 backgrounds, with students at different proficiency levels and within different educational contexts could pro- vide valuable insights into the general features and patterns of DM use by non- native EFL teachers. A comparative study with a native teacher corpus in the EFL classroom setting should be equally enlightening. Discourse markers in non-native EFL teacher talk 669 References Aijmer, K., & Simon-Vandenbergen, A.-M. (2006). Introduction. In K. Aijmer & A.- M. Simon-Vandenbergen (Eds.), Pragmatic markers in contrast (pp. 1 -10). Amsterdam: Elsevier. Allwright, R. L. (1984). The importance of interaction in classroom language and learning. Applied Linguistics, 5, 156-171. Anthony, L. (2014). AntConc (Version 3.4.3) [Computer Software]. Tokyo: Waseda University. Retrieved from http://www.laurenceanthony.net/ Blakemore, D. (1992). Understanding utterances: An Introduction to pragmatics. Oxford: Blackwell. Brinton, L. (1996). Pragmatic markers in English: Grammaticalization and dis- course functions. Berlin: Mouton de Gruyter. Buysse, L. (2012). So as a multifunctional discourse marker in native and learner speech. Journal of Pragmatics, 44, 1764-1782. Celce-Murcia, M. (2001). Teaching English as a second or foreign language. Bos- ton: Heinle and Heinle. Chapeton Castro, C. M. (2009). The use and function of discourse markers in EFL classroom interaction. Profile, 11, 57-77. Crystal, D. (1988). The English language. Harmondsworth: Penguin De Fina, A. (1997). An analysis of Spanish bien as a marker of classroom manage- ment in teacher-student interaction. Journal of Pragmatics, 28, 337-354. Ellis, R. (1994). The study of second language acquisition. Oxford: Oxford Uni- versity Press. Eslami, Z. R., & Eslami Rasekh, A. (2007). The effect of metadiscourse use on Iranian EFL learners’ lecture comprehension. Procedia – Social and Behav- ioral Sciences, 98, 1026-1035. doi: 10.1016/j.sbspro.2014.03.513 Flowerdew, J., & Tauroza, S. (1995). The effect of discourse markers on second language lecture comprehension. Studies in Second Language Acquisition, 17, 435-458. doi: 10.1017/S0272263100014406 Fraser, B. (1996). Pragmatic markers. Pragmatics, 6, 167-190. Fraser, B. (1999). What are discourse markers? Journal of Pragmatics, 31, 931-952. Fuller, J. M. (2003). Discourse marker use across speech contexts: A comparison of native and non-native speaker performance. Multilingua, 22, 185-208. Fung, L., & Carter, R. (2007). Discourse markers and spoken English: Native and non-native use in pedagogic settings. Applied Linguistics, 28, 410-439. doi: 10.1093/applin/amm030 Furko, B. P., & Monos, K. (2013). The teachability of communicative competence and the acquisition of pragmatic markers – a case study of some widely- used Business English coursebooks. Argumentum, 9, 132-148. Gloria Vickov, Eva Jakupčević 670 Hellermann, J., & Vergun, A. (2007). Language which is not taught: The discourse marker use of beginning adult learners of English. Journal of Pragmatics, 39, 157-179. House, J. (2013). Developing pragmatic competence in English as a lingua franca: Using discourse markers to express (inter)subjectivity and connectivity. Journal of Pragmatics, 59, 57-67. Iglesias Moreno, A. E. (2001). Native speaker – Non-native speaker interaction: The use of DMs. ELIA: Estudios de ingüística inglesa aplicada, 2, 129-142. Jalilifar, A. (2008). Discourse markers in composition writings: The case of Iranian learn- ers of English as a foreign language. English Language Teaching, 1(2), 114-122. Jung, E. H. S. (2006). Misunderstanding of academic monologues by nonnative speakers of English. Journal of Pragmatics, 38, 1928-1942. Lam, P. W. Y. (2009). Discourse particles in corpus data and textbooks: The case of well. Applied Linguistics, 31, 260-281. doi: 10.1093/applin/amp026 Lenk, U. (1998). Discourse markers and global coherence in conversation. Jour- nal of Pragmatics, 30, 245-257. Liao, S. (2009). Variation in the use of discourse markers by Chinese teaching assistants in the US. Journal of Pragmatics, 41, 1313-1328. Liu, S. X. (2006). Teachers’ discourse markers and classroom context. Journal of Hunan Institute of Humanities (Science and Technology), 4, 148-150. Llurda, E. (2005). Looking at the perceptions, challenges, and contributions... or the importance of being a non-native teacher. In E. Llurda (Ed.), Non-na- tive language teachers: Perceptions, challenges, and contributions to the profession (pp. 1-9). New York: Springer Science+Business Media. Martinez, A. C. L. (2004). Discourse markers in the expository writing of Spanish university students, IBERICA, 8, 63-80. Mihaljević Djigunović, J., & Vickov, G. (2010). Acquisition of discourse markers – Evi- dence from EFL writing. Studia Romanica et Anglica Zagrabiensia, LV, 255-278. Müller, S. (2005). Discourse markers in native and non-native English discourse. Amsterdam: John Benjamins. Nejadansari, D., & Mohammadi, A. M. (2015). The frequencies and functions of discourse markers in the Iranian university EFL classroom discourse. Inter- national Journal of Research Studies in Language Learning, 4(2), 3-20. doi: 10.5861/ijrsll.2014.840 Ostman, J.-O. (1981). You know: A discourse functional approach. Amsterdam: Benjamins. Othman, Z. (2010). The use of okay, right and yeah in academic lectures. Dis- course Studies, 12, 665-681. doi: 10.1177/1461445610376365 Discourse markers in non-native EFL teacher talk 671 Paatola, U. (2002). Interlanguage and interdiscourse system communication ap- proaches to Nordic English: Turn-initial but (Unpublished MA disserta- tion). University of Helsinki, Finland. Rongrong, D., & Lixun, W. (2015). Discourse markers in local and native English teachers’ talk in Hong Kong EFL classroom interaction: A corpus-based study. International Journal of Language and Linguistics, 2(5), 65-75. Schiffrin, D. (1987). Discourse markers. Cambridge: Cambridge University Press. Schleef, E. (2004). Gender, power, discipline, and context: On the sociolinguistic variation of okay, right, like, and you know in English academic discourse. Texas Linguistic Forum, 48, 177-186. Schourup, L. (1985). Common discourse particles in English Conversation. New York: Garland. Shahbaz, M., Sheikh, O. I., & Shahbaz Ali, M. (2013). Use of discourse markers by Chinese EFL professors: A corpus-based study of academic lectures by natives and non-natives. Journal of Education and Practice, 4(5), 80-89. Vickov, G. (2010). Usvajanje diskursnih oznaka u učenju engleskoga kao stranoga jezika (Unpublished doctoral dissertation). University of Zagreb, Croatia. Vickov, G. (2014). Investigating L1 influence on the acquisition of L2 discourse markers. In A. Peti-Stantić & M. M. Stanojević (Eds.), Language as infor- mation proceedings from the CALS conference 2012 (pp. 71-88). Frankfurt am Main: Peter Lang. Walsh, S. (2006). Investigating classroom discourse. New York: Routledge. Weinert, R. (1998). Discourse organization in the spoken language of L2 learners of German. Linguistiche Berichte, 176, 459-488. Wierzbicka, A. (1991). Cross-cultural pragmatics. The semantics of human inter- action. Berlin: Mouton de Gruyter. Yang, R. (2005, July). Conjunctions in international teaching assistants’ oral dis- course. Paper presented at The 14th World Congress of Applied Linguistics AILA, Madison, USA. Abstract retrieved from http://www.pur- due.edu/OEPP/abstracts Yang, S. (2011). Investigating discourse markers in pedagogical settings: A liter- ature review. ARECLS, 8, 95-108. Yu, L. Q. (2008). Interpersonal meaning of discourse markers by foreign language teachers in classroom. Journal of the Graduates of Sun Yat-Sen University (Social Sciences), 29(3), 129-135.