The Use and Functions of Discourse Markers in EFL Classroom Interaction

Los usos y las funciones de los marcadores del discurso en la interacción en el aula de inglés como lengua extranjera

Claudia Marcela Chapetón Castro*

Universidad Pedagógica Nacional, Colombia, Email: cchapeton@pedagogica.edu.co Address: C/. Colli Vehi 95 p. 302 (08026), BarcelonaSpain.

The aim of this paper is to investigate classroom interaction in the context of English as a foreign language being the teacher a nonnative speaker of the language. One specific aspect of classroom interaction and language use is the focus of attention, namely discourse markers (DMs). Using data from an EFL class, this study describes the occurrences and frequencies of DMs. It also provides an account for the main functions of DMs as they were used by a nonnative teacher of English and five adult students of EFL. A qualitative analysis reveals that discourse markers fulfill a number of textual and interpersonal functions which may contribute greatly to the coherent and pragmatic flow of the discourse generated in classroom interaction.

Key words: EFL classroom interaction, discourse analysis, discourse markers, nonnative teacher, adult EFL students

El artículo que aquí se presenta intenta investigar la interacción que ocurre en el aula de inglés como lengua extranjera cuando el profesor de inglés es nonativo. Un aspecto específico de la interacción en el aula y del uso del lenguaje es la presencia de los marcadores del discurso (MD). Con base en datos empíricos, este estudio pretende describir las ocurrencias, la frecuencia y las funciones principales de los MD. El análisis cualitativo de los datos revela que los MD cumplen funciones tanto textuales como interpersonales que pueden facilitar y contribuir al flujo coherente y pragmático del discurso generado en la interacción de aula.

Palabras clave: Interacción en el aula, análisis del discurso, marcadores del discurso, profesor de inglés no nativo, estudiantes adultos de inglés como lengua extranjera

Introduction

One common characteristic of English as a Foreign Language (EFL) classrooms is that the teachers may be nonnative speakers of the language they are teaching. From my experience as a nonnative teacher of English as a foreign language and, as a studentteacher educator, I consider that research on classroom interaction based on an analysis of the discourse can be very illuminating for two main reasons: First, it may contribute to gaining a better understanding of what happens inside the EFL classroom and second, it provides a valuable possibility to examine and describe the language used by nonnative teachers and students of EFL. Of course there has been research on this issue. A seminal publication on classroom interaction by Sinclair & Coulthard (1975) provides a comprehensive review, traced back to the late 1940s, of the considerable amount of research on the language used by teachers and pupils in classroom practices. An important contribution on discourse analysis for language teachers was made by McCarthy (1991) who provided not only a sound theoretical framework and descriptions based on research but also practical activities which sensitized teachers towards the language used inside their own classrooms. On the same line, CelceMurcia & Olshtain (2000) propose a discourse and context based perspective on language teaching and learning to redefine the roles for teachers, learners and materials. With the exception of the notable work by Llurda (2005) who explicitly addresses and puts together the research conducted in different EFL settings such as Catalonia, the Basque Country, Hungary and Brazil, the language used by nonnative Englishspeaking teachers and students remains largely unexplored.

The aim of this exploratory study is to investigate classroom interaction in the context of English as a foreign language being the teacher a nonnative speaker of the language. One specific aspect of classroom interaction and language use is the focus of my attention, namely discourse markers. Therefore, the occurrences of discourse markers will be explored and described both quantitatively and qualitatively with a grounded approach method in mind. Thus, I did not formulate, and seek to validate, hypotheses but rather took simple statistical analyses as a starting point for a qualitative analysis of the functions served by discourse markers in this particular classroom setting.

The research questions guiding this smallscale study are:

• How frequent are discourse markers (DMs) in the EFL classroom discourse sample under scrutiny here?

• Which DMs occur? How frequently do they occur?

• Which DMs are used by the teacher?

• Which DMs are used by the students?

• What are the prevailing functions of the DMs employed in classroom interaction by the teacher and by the students?

The next section will present a brief literature review on the main concepts which are central to this study followed by a description of the characteristics of the participants and setting, the instruments and procedures for data collection and the analysis of the data. The article finally closes with a discussion of the results and the conclusions.

According to van Dijk (1997) discourse is a form of language use which includes the functional aspects of a communicative event. It means that people use language in order to communicate ideas, beliefs or emotions in social events and situations such as an encounter with friends or a lesson in the classroom. This also suggests that in these communicative events, the participants do not limit themselves to using the language or communicating: they interact. As Douglas (2001) points out, discourse analysis is the examination of language used by the members of a speech community which involves looking at both language form and language function. In this study language is viewed as social interaction that takes place within a classroom community, among adult students and a nonnative teacher of EFL. As mentioned earlier, one specific aspect of classroom interaction and language use is the occurrence of discourse markers. This literature review deals with the two central concerns of this study: discourse markers (DMs) and studies on the discourse of nonnative EFL teachers.

Discourse Markers: Definition, Characteristics and Functions

In her influential work on discourse markers Schiffrin operationally defines them as "sequentially dependant elements which bracket units of talk" (1987, p. 31). She suggests that DMs are used in discourse because they provide "contextual coordinates for utterances". That is, they contribute to building the local coherence which is jointly constructed by speaker and hearer in their discourse structure, context, meaning and action during interaction. They serve to show how what is being said is connected to what has already been said, either within a speaker's turn or across speakers' turns. In her research, she focuses on eleven discourse markers: oh, well (particles), and, but, or, so, because (conjunctions), now, then (time deictics), and you know, I mean (lexicalized clauses). In the relevant literature, there are studies which deal, whether generally or specifically, with a wide scope of DMs, however, difficulties arise as there is no agreement among scholars when they refer to their terminology, classification and functionality¹.

Brinton (1996) points out that DM has been the most common name suggested for "seemingly empty expressions found in oral discourse", however, she proposes the term pragmatic markers, as pragmatic "better captures the range of functions filled by these items"². Although Brinton acknowledges the fact that there has been little agreement on the items that can be called pragmatic markers, she compiles an inventory of thirty three markers³ that have received scholarly attention and proposes a broad number of characteristics typical of these words. Those characteristics were later taken up by Jucker & Ziv (1998) who reordered them to combine features that pertain to the same level of linguistic description: phonological and lexical, syntactic, semantic, functional and sociolinguistic features. Some characteristics of DMs, according to Brinton (1996) and Jucker & Ziv (1998) are:

a. DMs are predominantly a feature of oral rather than of written discourse.

b. They appear with high frequency in oral discourse.

c. They are short and phonologically reduced items.

d. They may occur sentence initially, sentence medially and finally as well.

e. They are considered to have little or no prepositional meaning, or at least to be difficult to specify lexically.

g. They seem to be optional rather than obligatory features of discourse. Their absence "does not render a sentence ungrammatical and/or unintelligible" but does "remove a powerful clue" (Fraser, 1988, p. 22 as cited by Brinton, 1996, p. 34).

h. They may be multifunctional, operating on the local and global levels simultaneously though it is difficult to differentiate a pragmatically motivated from a nonpragmatically motivated use of the form.

The different studies of DMs distinguish several domains where they may be functional, in which there are included textual, attitudinal, cognitive and interactional parameters. Accordingly, as stated by Jucker & Ziv (1998) DMs have been analyzed as textstructuring devices that serve to mark openings or closings of discourse units or transitions between them. Also, they serve as modality or attitudinal indicators, as markers of speakerhearer intentions and relationships, and as instructions on how given utterances are to be processed or interpreted. Thornbury & Slade (2006) argue that DMs and other interactional signals such as response elicitors (right?, Ok?) and attention signals (hey!) are crucial to the collaborative organization that takes place in conversation as streams of talk are segmented into "loose topically coherent" macrostructures:

Topics are broached, commented on, developed, extended, replaced, retrieved... and all this conversational flux is continuously shaped and negotiated by interactants. Crucial to this collaborative organizational "work" is the inserting of discourse markers and other interactional signals into the stream of talk. (Thornbury & Slade, 2006, p. 57)

As Brinton (1996) claims DMs are grammatically optional and semantically empty but they are not pragmatically optional or superfluous, instead, they serve a variety of pragmatic functions. She presents an inventory of ten functions which she groups into two main categories (based on the modes or functions of language identified by Halliday, 1973). First, the textual function which is related to the way the speaker structures meaning as text, creating cohesive passages of discourse, using language in a way that is relevant to the context. And second, the interpersonal function which refers to the nature of the social exchange, that is, the role of the speaker and the role assigned to the hearer. Table 1 presents my understanding of the inventory of functions devised by Brinton:

Central for the development of this study is Hellerman & Vergun's (2007) approach to DMs as they incorporate pragmatic functions in their definition. As these authors state, DMs are words or phrases that function within the linguistic system to establish relationships between topics or grammatical units in discourse, that is words such as so, well, and then. DMs also serve pragmatic functions, as a speaker uses them to comment on the state of understanding of the information about to be expressed using phrases such as you know, I mean. They may also be used to express a change of state, such as the particle oh; or for subtle commentary by the speaker suggesting that what seems to be the most relevant context is not appropriate e.g. well. Thus, the DMs are understood in this paper as lexical items that serve textual, pragmatic and interactional purposes. And, as Schiffrin (1987) and Brinton (1996) claim, their usage is optional, not obligatory as DMs could be taken out of an utterance without altering neither its structure nor its propositional content. Research on DMs has abounded since the 1980s⁴. Studies include analyses and descriptions of their use in different languages. DMs have also been examined in a variety of genres and interactive contexts, and in a number of different language contact situations as pointed out by Schiffrin (2001), who provides a rich discussion on the three different perspectives to approach DMs and summarizes recent studies that have contributed to understanding how DMs work.

Müller (2005) analysed the use of seven DMs in conversations of native and nonnative speakers of English in Germany and USA.

Regarding the study of DMs in classroom settings, Chaudron & Richards (1986) investigated the comprehension of university lectures by nonnative speakers of English living and studying in The United States, that is, in English as a Second Language (ESL) contexts. Chaudron & Richards (1986) made use of four different versions of the same text with different categories of discourse markers (baseline, micro, macro, or micromacro versions). Overall results showed that macromarkers produced better text recall than micromarkers. It was hypothesized that micromarkers do not provide enough information to help in making content more salient. Implications for the teaching of listening skills in ESL settings were discussed as well.

In their aim at determining if consultation of a corpus of classroom discourse can be of benefit in language teacher education, Amador, O'Riordan & Chambers (2006) examined the uses of discourse markers in French and Spanish. A quantitative analysis showed the low number of occurrences of DMs in both a French class and a Spanish class while a qualitative analysis described the main functions of DMs identified in classroom discourse. These functions were categorized into five groups considering mainly the role of the teacher in the classroom: To introduce a new topic or activity; to motivate or encourage the pupils; to call the pupils' attention; to recap or clarify what has been said; to rephrase what has been said.

In a recent research Hellerman & Vergun (2007) investigated the frequency of use and some functions of three particular discourse markers, well; you know; and like in classroom interaction and inhome interviews. 17 adult learners of English as a second language at the beginning level, provided the data of this 5year research project. Their results suggest that the students who use more discourse markers are those who are more acculturated to the US and use them outside their classroom. After this overview on discourse markers, a brief account on research regarding nonnative EFL teachers discourse will be presented.

NonNative EFL Teachers

To address this issue, it would be perhaps important to refer to what is meant by native speaker of English. In this study, a native speaker of English would be a person who speaks only English, or a person who learned another language later in life but still predominantly uses English as L1.

The teacher participating in this study is a nonnative English speaker as his L1 is Spanish (as it will be later dealt with in section 3.1). The language used by nonnative teachers in the EFL classroom has been addressed by relatively few scholars. By applying standard discourse analysis procedures, Cots & Diaz (2005) studied the nonnative teachers' classroom performance looking mainly at the construction of social relationships and the way linguistic knowledge is conveyed.

Their analysis suggested that teacher talk might be a continuum that locates teachers' discourse somewhere between a discourse of power and a discourse of solidarity and that gender variables may be more relevant than nativeness in order to understand interactional styles in the EFL classroom. Frodden, Restrepo, & Maturana (2004) conducted a research project on foreign language teachers' discourse and practices with respect to assessment in two Colombian universities. Their main aim was to contribute to the improvement of nonnative English teachers' assessment practices. Pineda (2004) examined how adult EFL students and nonnative teachers constructed meaning in the classroom when dealing with critical thinking related tasks, the metacognitive processes involved, the types of interactions built around the tasks and how they influenced language competence and critical thinking. Chang (2004) explored the relationships between five EFL nonnative teachers' identities and the impact on their teaching practices in Taiwan. The study proved that the five participants' knowledge of multiculturalism and language awareness, their Chinesecentered education, and their educational and personal experiences were evident in their teaching. As Müller (2005) asserts little is known about DMs usage by nonnative speakers and, as I see it, even less is known about their usage by nonnative EFL teachers.

Methodology

The Participants

The participants in this study are adult male and female students of English as a foreign language, and one male nonnative EFL teacher. The total number of students in this class is five. There are two male and three female students. Their ages range from 19 to 22. They live in Spain but they come from different places: three of them come from Catalonia, having Catalan and Spanish as their first languages. Another student is from Italy, his mother tongue is Italian. The other student comes from a LatinAmerican country and his first language is Spanish. They are in their fourth year English course and their current proficiency level, according to the classification parameters of the institution where they currently study, is upperintermediate. They attend EFL classes every Saturday morning from 10:00 to 13:15 during each academic semester.

The Setting

The EFL class analysed to develop this study was located at a language center functioning in the city of Barcelona, Spain. It is a language school with 15 years of experience in language teaching. They offer reduced groups with a maximum of eight students and a communicative approach to the language with the purpose of helping their students achieve a good command of both spoken and written English. Teachers monitor the students' progress by means of regular exams, attendance records and pedagogical advice. There are EFL classes scheduled during week days and also on Saturday mornings. Every session on Saturday morning lasts three hours.

Instruments and Procedures for Data Collection

The class recorded was the first session after Christmas holidays and the students talked about what they had done during their holidays. Participants talked about the traditions to celebrate Christmas in their countries: Spain, Italy and Colombia. After that, they talked about "worstcase scenarios and ways to prepare for disasters" which is a topic developed in their textbooks as part of the initial program of the course. This classroom activity combined reading with speaking practice; that is, with oral interaction.

Two different instruments were used to gather the data. First, I designed a questionnaire in order to collect background information of the course and to create a profile of the students. This form, used once with the group of students under scrutiny, was filled in by the teacher and consisted of two main sections: information regarding the nature of the course and students, and, a second section in which a brief description of the particular tasks developed in this class was required. This instrument was really important as it provided valuable information which contributed to a better understanding of the interaction that took place in the classroom.

Audiorecordings were also used. As the data were collected in an indoor setting, the type of recording equipment was selected accordingly. With the consent of the participants, a light, portable audiorecorder of professional quality was tested before the recording session and used to record the participants' oral interaction. Following Calsamiglia & Tusón's (1999) suggestions on how to deal with oral data for discourse analysis, the quality of the recording was verified at the end of the session in order to make sure that it was intelligible. Once the recording session had been completed, a digital copy was made and kept for backup. Then, an initial process of transliteration of the audiorecorded class began. Afterwards, a 25minute fragment of the session was taken as the main focus of attention in order to develop this paper. The fragment was chosen because it constituted the most representative and richest section in terms of oral interaction among the participants. This selected fragment was transcribed using specific transcription conventions which were very useful in providing the maximum transmission of contextual information and to ensure accuracy. The audio recording was transcribed directly into a computer file using the Sound Scriber program created by Breck (1998) at the University of Michigan, which aides in the transcription of digitized sound files and has several userconfigurable features. Occasional speech errors made by participants were not corrected; instead, they were transcribed as they had actually occurred. An instrument for the transcript was designed including information about the date, site, and key issues regarding the participants, context and the sample transcription.

Data Analysis

Bearing in mind the research questions posed to develop this smallscale study, I aimed at quantitatively and qualitatively relevant results. The quantitative side of the analysis was performed by the use of descriptive statistics. It consisted of simple statistical analyses such as lexical size and frequency counts in order to show the occurrences and distribution of discourse markers in the discourse. Taking Brinton's (1996) inventory of 33 items that can be considered DMs, I developed the quantitative analyses using the latest version of a computerresearch tool called AntConc, a freeware multipurpose corpus analysis toolkit designed by Laurence Anthony at Waseda University.

The qualitative analysis consisted of the identification and description of the pragmatic functions of discourse markers. To complete these tasks, I based my analysis mainly on the functions proposed by Müller (2005), Brinton (1996) and Schiffrin (1987).

Regarding the first research question posed to carry out this exploratory study, I first analyzed the general lexical size and frequency. As shown in Table 2a, the total number of words in the sample taken for the development of this paper (of transcribed oral data) is two thousand one hundred. The most frequent word of this sample is the definite article the, with 93 occurrences accounting for 4.43% of the data. It was followed by the nominative pronoun I with 90 occurrences (4.28%). The fourth most frequent word is the DM and with 74 occurrences (3.52%). This information may be unsurprising. Words such as the, I, and and are highly frequent in spoken communication. To give an example, McCarthy & Carter (1997), who used a far bigger sample (330,000 words), identified the, I, you and and as the four top words used in spoken English. However, a distinction between content and function words might be relevant. Thus, Table 2a shows the distribution of content words and function words in this sample of EFL classroom talk. Most of the highfrequency words are function words which consist of the 66% of the whole sample, while content words represent 34% and comprising words such as family, day and have, the first to appear with 19 occurrences each. McCarthy & Carter (1997) also found that over sixty percent of their data consisted of function words. A closer look at the data reveals that DMs occur 398 times. These occurrences correspond to 19% of the total corpus and to 30% of function words as shown in Table 2b.

Concerning the occurrence and frequency of DMs, Brinton's (1996) inventory of 33 items was considered as a basis. Using the concordance lines provided by the AntConc computer program, I analyzed each one of the instances in which DMs occur. Since some items from Brinton's inventory may also serve other functions different from their use as discourse markers, it was relevant to distinguish DMs from those cases. I made a distinction between nondiscourse marker and marker functions based on the list of features given in Table 1. The following extracts from my data illustrate that a) some items function as discourse markers and, therefore, were included as part of the analysis and b) some cases in which the items were serving as nondiscourse marker functions were excluded:

Excerpt (1) shows the use of well as a discourse marker: In line 107, the teacher asks S3 a question which is answered in line 112. "Well" has been previously used by the student to mark his/her response (in line 110). Here, well is used as a response marker by the student, thus, it was included in the analysis.

b) (2) 50 S2: So. ah: () I don't remember very well

In this example, well collocates with very and is an adverb. It is not fulfilling any discourse marker function. Therefore, it was excluded.

In this case, so is qualifying the adjective cheap. It was excluded because it was used as an adverb of degree or manner.

In this case, if was excluded because it was used as a conditional.

The above excerpts (1)(5) illustrate that the use of lexical items is dependent on the local context and sequence of talk in classroom interaction. Thus, these are two important factors to consider when making decisions on what to exclude or include as a discourse marker in the analysis. Table 3a shows the occurrences and frequencies of DMs in this study. The most frequent DM (and) occurs 74 times. Among other very frequent DMs we have uh huh / mhm (44 occurrences), ok and so (23 each), followed by but (19 occurrences). It is interesting to see that some DMs occurred only twice (now, and stuff/things like that, sort/kind of) or once (actually, just). In addition, some other markers from Brinton's inventory did not occur (after all, almost, anyway, basically, go "say", if, mind you, moreover, say, therefore, you see).

Based on the characteristics assigned to DMs by scholars such as Schiffrin (1987), Brinton (1996) and Jucker & Ziv (1998), I identified three more items that served as discourse markers in this sample taken from classroom interaction. Table 3b shows the occurrence and frequencies of these three DMs. The most frequent items are um / e with 50 occurrences. Yeah occurs 42 times and eh? only once.

As stated by Thornbury & Slade (2006) and by Schiffrin (2001), DMs often become combined. In my data, I found combinations such as and then (7 occurrences), ok and (3 occurrences), oh yeah, oh really, mhm and, well but, well um, and well, ok well, yeah mhm, well now, yes I know, ok so, ah ok, ah yeah, like yeah and so ah.

Discourse markers were used differently by the participants in this study. In relation to the third and fourth research questions posed to develop this study, Table 4 shows two categories in which DMs were classified according to whether they were used by the nonnative teacher (TT) or the adult EFL students (SS). The total number of DMs used by the teacher was 244 (61%) while students used them 154 times (39%).

The fact that students used 39% of the total DMs may confirm De Fina's (1997, p. 337) concern on the "dominant role of the teacher in the classroom". However, these results contradict those obtained by Amador, O'Riordan & Chambers (2006, pp. 9091), who found that pupils "use hardly any discourse marker" (3%) being the teachers the ones who used 97% of the DMs identified in classroom interaction. Regarding the use of DMs by the teacher, this study shows that this nonnative teacher uses a great deal of DMs once, and some DMs are repeatedly used, as shown in Table 4. In contrast, Amador, O'Riordan & Chambers (2006) found that "the four native speaker teachers use a relatively limited number of DMs (9, 4, 10, 8)". The total number of DMs used by the teachers in Amador, O'Riordan & Chambers' study came to 253, accounting for 97% of the total (ibid.). Though this raw number (253) is very close to the occurrences identified in the discourse of the nonnative teacher participating in this smallscale research (244), it instead accounts for 61% of the total. This may suggest that the nonnative teacher's role might not be as "dominant", in De Fina's words, and thus may allow a slightly more space for students to participate in classroom interaction.

However, differences in the quantity of DMs used by native and nonnative teachers and students in classroom interaction may be related to a variety of factors and methodological issues. In Amador, O'Riordan & Chambers' study, the classes recorded were "intended simply as examples of classroom interaction" (2006, p. 86), but no clear details were given on the kind of tasks or activities developed while recording. In contrast, as explained in section 4.3, the particular sample analysed to develop this paper consisted of 25 minutes in which students were asked to talk about a recent experience. Although this activity was proposed and guided by the teacher, it was mainly studentcentered and pupils were free to participate, intervene and express themselves using the target language. This issue may explain the high number of times in which students use DMs like and, um/e, yeah, mhm, no and well as shown in Table 4.

After having looked at the occurrences, frequencies and distribution of DMs, I decided the following section of this paper would address the last question related to the general functions of DMs in classroom interaction. In order to identify and describe their main functions, I analyzed each discourse marker in its context of use; that is, I considered both the local context and the sequence of talk in which they occurred during classroom interaction. The initial twenty two lines of the whole transcript are included in Table 5 in order to illustrate the qualitative analysis that was performed on the entire dataset. As is shown in Table 5, a variety of DMs are present to aid the speakers in the construction of their discourse and meaningmaking during classroom interaction. The functions I identified are both textual and interpersonal.

The textual functions of markers are more related to the construction of discourse coherence. For instance, so, in line one is used by the teacher in order to initiate his discourse. So is also used by the teacher in cases 4 and 8 as a result marker and with the purpose of emphasizing and structuring his discourse coherently. He also uses a couple of fillers such as um to fill a momentary hesitation probably occasioned by "the demands of realtime processing pressure" (Thornbury & Slade, 2006, p. 56). In line seven, the teacher uses and then to signal continuity and to mark the temporal connection and sequential dependence on the discourse. Student 1, in line 10, takes the turn and volunteers to interact by using the DM yeah. The teacher assigns the turn using the DM ok. S1 uses the filler um, in lines 12 and 20, as a delaying tactic to fill a momentary hesitation, to sustain discourse and to hold the floor. Most of the uses of the DM and in this extract are related to its textual function of showing continuity and adding new information (cases 17, 21 and 27). However, and, in case 25, is used by the student not only to mark continuity and thematic connection but also as a turn keeper showing that even though she has been interrupted, she still holds the floor. The use of because in line 16, as a marker of cause, not only has the textual function of introducing new information (exams at the university) but also provides an explanation or reason connected to the previous information ("I tried to study") which, as I see it, contributes to the coherence of the discourse as it expresses the relation of relevance between the preceding utterance and the context. Case 30 in line 22 shows the way the student indicates the end of her turn. However, she uses the lexical phrase "that's all" which is not considered a DM by any of the scholars previously referred to. Another example that illustrates this issue is observable in the following excerpt:

(6) 127 TT: Yyou (bis) went to VISIT your family.

128 S3: si.(.) yes. to visit ↓yes. ↓and no more.

130 S4: e: well I: I sleep a lot.

Student 3 closes her turn using the expression "and no more" as shown in line 128 of the transcription. The student's indication of the end of her turn makes the teacher assign a new one (line 129) to student 4 who uses the discourse marker well preceded by a filler as a turn taking signal. The analysis of the data showed that relinquishing the floor is sometimes unmarked; that is, sometimes students do not use any DMs to indicate a close but instead, it is the teacher who closes their turn by using DMs such as ok or well.

The interpersonal functions of DMs are precisely more related to the reactions, responses and relations built by the participants during interaction, that is, to the role of the speaker and hearer during the social and communicative exchange. Interpersonal functions of DMs are revealed in the following examples as shown in the excerpt in Table 5: In line 2, the teacher uses right, and also ok (in lines 3 and 9), both with rising intonation, in order to check understanding and seek the students' agreement on his proposed activity. Student 1 responds in line 8 using ok to express understanding and agreement. It is interesting to see that the teacher uses mhm (cases 15, 16, 18 and 22) as a backchannel signal, thus, providing permanent feedback to student 1 "signaling that the message has been understood and confirming that communication is on course" (Thornbury & Slade, 2006, p. 58) while S1 continues to hold the floor. Cases 23 and 24, yeah and ah, are examples of DMs used by the interactants as response markers. As I see it, the teacher uses ah also to confirm his previous assumption which had been expressed as a question in line 17 (at the university?). The combination of two DMs as in case 28, oh yeah, is used by the teacher as a reaction marker which also has the interpersonal function of conveying agreement. He agrees with the student about the common act of eating a lot during Christmas.

The following excerpts (7), (8) and (9) taken from the data further illustrate the textual and interpersonal functions of DMs in the interaction of this EFL class:

(7) 38 TT: Try to: remember, you said you forgot but=

39 S2: .........................=it wwas u:m

molto good <laugh>

40 TT: [<laugh>]

41 S2: because I: traveled to my island, u:m the twenty two of December, and then

43 TT: ...............................[You went to] to your what? e:

44 S2: [eh?]

45 TT: to your ISLAND?

46 S2: yeah. I'm from (.) e: Sardegna

47 TT: ↑Oh really? [Oh] I thought you were from (bis) the main land, from Italy=

48 S2: .............................[yeah]=No, no (bis). I stayed in my island, it is in the mediterraneo

49 TT: yeah, I know

Excerpt (7) shows that participants use DMs such as um (lines 39, 41) and e (lines 42, 43, 46) as pause fillers to indicate they keep holding the floor. In lines 41 and 42, the student uses three DMs that aid in the construction of his discourse: because indicates the inclusion of new information; and then marks temporal connection and so is used as a sequential marker. In line 44 the DM eh? fulfills an interpersonal function: it is used by the student to express a reaction to the preceding question of the teacher, signaling his lack of understanding and his need to listen to the question again. The teacher also uses reaction markers in line 47: Oh really, with upward intonation, is both expressing a response (of surprise) and requesting confirmation from the student. In lines 46 and 48 the student uses yeah as a response and confirmation marker of the ongoing discourse. The DM oh used by the teacher in line 47 as a reaction to the confirmed information overlaps with the students' response marker yeah.

Excerpt (8) is preceded by a communicative event in which student four is mainly narrating what she did during Christmas and on her birthday at the beginning of January. S4 is interrupted by S2 who says that his birthday was also at the beginning of January. In line 148, student 2 tells the participants that his mom's birthday was on the same day:

149 S4: ↑u:h nice. <laugh> a:nd

150 TT: and MY birthday was the 13th (.) =of January=

151 S1: ................when?=...........=↑u:h

152 S2: congratulations!

153 SS: <laughing>

154 TT: <laughs> ↓OK

155 S4: a:nd I invited my friends to: to lunch (.) no (.) to dinner.

156 TT: To have dinner, mhm.

In line 149 student four responds with the DM uh and, in her attempt to regain her turn, she uses the DM and to signal her willingness to continue with her narration. As shown by the transcription conventions, S4 is interrupted by the teacher who takes the floor also using the DM and. Student 1 shows his response to the ongoing discourse about birthdays by using the reaction marker uh in line 151. After some natural laughing, the teacher uses the DM OK, in line 154, as an explicit turn giver which aids student four in acquiring the floor. The DM and in line 155 signals that S4 still holds the floor even if she has been interrupted (turn taker and turn keeper) and it also shows continuity, thematic connection and the addition of new information. Finally, in line 156, the teacher uses the DM mhm after providing some corrective feedback to the student.

(9) 175 TT: What about you (.) Carlos?

176 S5: Well, the same of mmy partners here [ I (bis) ate a lot,

177 S1: .............[mhm <laughing>]

178 S5: and I worked [also On the kings's day [I ↓worked until six or seven[

179 TT: ................[mhm].............. [mhm]........[mhm]

In line 177, student one interacts with student 5 by using mhm as an agreement marker while the teacher uses mhm to provide permanent feedback and as a confirmation marker that the communication is on course.

As the analyses reveal, discourse markers fulfill a number of textual and interpersonal functions which contribute greatly to the coherent and pragmatic flow of the discourse generated in classroom interaction. The above described functions of markers such as so, because, and, ok and yeah are examples of "their apparent multifunctionality" (Schiffrin, 1987, p. 64). As previously shown, DMs may be used simultaneously in several different ways. Research has revealed, as Müller (2005) argues, that generally the discourse markers studied by scholars fulfill more than one function or at least have subfunctions as is the case here. I do agree with Schiffrin on her assertion that DMs are contextdependant so that they "can gain their function through discourse" (2001, p. 60).

These and other examples from the data illustrate how DMs function. Table 6 summarizes the functions of DMs used by participants in this specific class sample. Again, it is clear that they can be multifunctional and that they serve both textual and interpersonal functions.

(10) 70 …special dates, the holidays like the: 24th (.) 25th

(11) 242 …that is a very (.) like (.) the most important DAY of the holidays in Colombia,

(12) 253 and then I spent, a month, less than a month, like yeah twenty some days um in

As regards the distinct functional uses of discourse markers, it was observable that both the students and the teacher made use of these items to fulfill textual and interpersonal functions in the EFL classroom. Generally, students mainly used DM to serve textual functions. Specially, they made great use of pause fillers and turn keepers (e.g. um, and, e) and of the DM and to signal new information and continuity. In relation to interpersonal functions, cooperation and agreement markers were the most commonly used by the students (e.g. yeah). Textual functions of DMs were highly used by the teacher as well. In the construction and organization of classroom discourse, the teacher used the DM and to indicate sequences, continuity and new information. OK was often used as an opening and closing frame marker and it was very useful in the organization and assignment of turns during interaction. The teacher also used a variety of DMs that fulfill interpersonal functions such as backchannel signals, checking understanding markers, response and reaction markers and confirmation markers. On the whole, the prevailing uses of the discourse markers identified and analyzed in this smallscale study fulfill textual functions that aid the participants in structuring the classroom discourse coherently.

Finally, there was another element present in classroom interaction worth mentioning: laughter. Even though it is not considered a DM, it has attracted my attention; first, because it is very frequent; it appears 32 times and also, because it is used both by the nonnative teacher and the five adult EFL students. In agreement with Coates (1997), I consider that laughter was used by the participants to signal their constant presence, a way to say "we are here, we are participating". Laughter also occurred to signal amusement and surprise, but as I see it, one of the most important functions of laughter in classroom interaction may be to release tension and to create a relaxed, comfortable atmosphere in which everyone is welcome to participate. That is, the joint creation of a relaxed setting where the main goal is not only the exchange of information but the construction and maintenance of good social relations.

Conclusion

DMs have been widely studied by researchers even if discussions on terminology and definable issues are still unresolved. However, there seems to be general agreement on the fact that the production of coherent discourse is an interactive process that requires speakers to draw upon communicative knowledge and pragmatic resources. The fact that most of the studies on DMs have focused their attention on native (or bilingual) speakers of English who acquire this pragmatic competence in their childhood might be an indicator of the need to further explore and systematically investigate the language used by nonnative English teachers.

One of my goals with this exploratory study was to describe the occurrences and frequencies of DMs in EFL classroom interaction with the teacher being a nonnative speaker of the language. Results showed that DMs occurred 398 times, which corresponds to the 19% of the total sample of recorded and analyzed classroom data. It was also found that most DMs were used by the nonnative teacher (61%) while students' use of DMs accounted for 39%. It was also observed that and was the DM most frequently used by both the teacher and the students and that some DMs such as say, therefore, you see or anyway were never used.

I also aimed at providing an account for the main functions of DMs in classroom interaction. In general, DMs were used by the nonnative teacher and the five adult students of English as a foreign language to serve structural, pragmatic and interactional purposes. As I see it, and in agreement with Müller (2005), DMs contribute to the pragmatic meaning of utterances and thus play an important role in the pragmatic competence of the speaker. As Schiffrin (2001) explains, DMs tell us not only about the linguistic properties (semantic and pragmatic meanings and functions) and the organization of social interactions, but also about the cognitive, expressive, social and textual competence of those who use them.

Even though the adult EFL students from this smallscale study used less that 40% of the total DMs, they in fact used them with several textual and interpersonal purposes as previously discussed in the analysis. However, this might be an indication of the need to conduct further research in order to make informed decisions about the implicit or explicit teaching of DMs in the EFL classroom. Studies along this line might be an important contribution to the development of the pragmatic competence of the learners.

Though this exploratory study may not allow for generalizations on the discourse particularities of the nonnative speaker community, it might serve as an awareness raiser for the need to consider further research along the line of nonnative speakers of the language and mainly on EFL classroom interaction. It is true, as Llurda (2004) points out, that the transformation of English as an international language has brought with it many changes to the teaching profession which should not be overlooked. Further research on the differences and similarities between native and nonnative teachers' discourse might help us identify and characterize those changes Llurda refers to. More specifically, research on DMs and classrooom interaction may be illuminating, first, because the functions and contexts of DMs are so broad and are part of the basic tools through which discourse can be understood and, second, because this kind of research agenda may throw light on the multifaceted reality in which the English language is used both by nonnative teachers and learners.

¹ For a comprehensive review on a whole range of terms, definitions, features and functions assigned to discourse markers by different scholars see Brinton, 1996; Jucker & Ziv, 1998; González, 2004; Müller, 2005.

² Brinton (1996, pp. 3031) presents a detailed examination of the various definitions given to DMs in relation to the different functions identified as central and therefore assigned to DMs by different scholars.

³ The complete list will be shown later on (Table 3a) as it served as the basis for the quantitative data analysis of the present study.

⁴ For a summary of the most significant research see Schiffrin (2001, pp. 5467) who addresses the most remarkable authors and their focuses on research regarding DMs.

References

Amador, C., O'Riordan, S., & Chambers, A. (2006). Integrating a corpus of classroom discourse in language teacher education: The case of discourse markers. ReCALL, 18(1), 83104.        [ Links ]

Breck, E. (1998). SoundScriber transcription program for Windows. Michigan: The University of Michigan Regents. Retrieved January 27, 2008, from the Micase Homepage. Web site: http://www.lsa.umich.edu/eli/micase/soundscriber.html        [ Links ]

Brinton, L. (1996). Pragmatic markers in English. Grammaticalization and discourse functions. Berlin/New York: Mouton de Gruyter.        [ Links ]

Calsamiglia, H., & Tusón, A. (1999). Las cosas del decir. Manual de análisis del discurso. Barcelona: Ariel.        [ Links ]

CelceMurcia, M., & Olshtain, E. (2000). Discourse and context in language teaching: A guide for language teachers. Cambridge: Cambridge University Press.        [ Links ]

Chaudron, C., & Richards, J. (1986). The effect of discourse markers on the comprehension of lectures. Applied Linguistics, 7(2), 113127.        [ Links ]

Chang, S. (2004). A case study of efl Teachers in Taiwan: Identities, instructional practices and intercultural awareness. Dissertation Abstracts International: The Humanities and Social Sciences, 65(4), 1218A1219A.        [ Links ]

Cots, J., & Diaz, J. (2005). Constructing social relationships and linguistic knowledge through nonnative speaking teacher talk. In E. Llurda (Ed.), Nonnative Language Teachers: Perceptions, Challenges and Contributions to the Profession (pp. 85106). New York: Springer.        [ Links ]

Coates, J. (1997). The construction of a collaborative floor in women's friendly talk. In T. Givón (Ed.), Conversation: Cognitive, communicative and social perspectives (pp. 5590). Amsterdam: John Benjamins.        [ Links ]

De Fina, A. (1997). An analysis of Spanish bien as a marker of classroom management in teacherstudent interaction. Journal of Pragmatics, 28(3), 337354.        [ Links ]

Douglas, A. (2001). Discourse analysis for language teachers. Retrieved June 13, 2008, from the Center for Applied Linguistics: Digests. Web site: http://www.cal.org/resources/digest/0107demo.html        [ Links ]

Frodden, M., Restrepo, M., & Maturana, L. (2004). Analysis of assessment instruments used in foreign language teaching. Íkala, Revista de Lenguaje y Cultura, 9(15), 171201.        [ Links ]

González, M. (2004). Pragmatic markers in oral narrative: The case of English and Catalan. Amsterdam: John Benjamins.        [ Links ]

Halliday, M. A. K. (1973). Explorations in the functions of language. London: Edward Arnold.        [ Links ]

Hellerman, J., & Vergun, A. (2007). Language which is not taught: The discourse marker use of beginning adult learners of English. Journal of Pragmatics, 39(1), 157179.        [ Links ]

Jucker, A., & Ziv, Y. (1998). Discourse markers: Descriptions and theory. Amsterdam/ Philadelphia: John Benjamins.        [ Links ]

Llurda, E. (2004). Nonnativespeaker teachers and English as an International Language. International Journal of Applied Linguistics, 14(3), 314323.        [ Links ]

Llurda, E. (2005). Nonnative language teachers: Perceptions, challenges and contributions to the profession. New York: Springer.        [ Links ]

McCarthy, M., & Carter, R. (1997). Written and spoken vocabulary. In N. Schmitt, & M. McCarthy. (Eds.), Vocabulary: Description, Acquisition and Pedagogy (pp. 2039). Cambridge: Cambridge University Press.        [ Links ]

McCarthy, M. (1991). Discourse analysis for language teachers. Cambridge: Cambridge University Press.        [ Links ]

Müller, S. (2005). Discourse markers in native and non native English discourse. Amsterdam: John Benjamins.        [ Links ]

Pineda, C. (2004). Critical thinking in the efl classroom: The search for a pedagogical alternative to improve English learning. Íkala, Revista de Lenguaje y Cultura, 9(15), 4580.        [ Links ]

Schiffrin, D. (1987). Discourse markers. Cambridge: Cambridge University Press.        [ Links ]

Schiffrin, D. (2001). Discourse markers: Language, meaning and context. In D. Schiffrin, D. Tannen,& H. Hamilton (Eds.), The handbook of discourse analysis (pp. 5474). Massachusetts: Blackwell Publishers.        [ Links ]

Sinclair, J., & Coulthard, M. (1975). Towards an analysis of discourse: The English used by teachers and pupils. London: Oxford University Press.        [ Links ]

Thornbury, S., & Slade, D. (2006). Conversation: From description to pedagogy. Cambridge: Cambridge University Press.        [ Links ]

Van Dijk, T. (1997). Discourse as structure and process. London: Sage Publishers.        [ Links ] ]]> 2006 18 1 1 83104 1998 1996 1999 2000 1986 7 2 2 113127 2004 65 4 4 1218A1219A 2005 85106 1997 5590 1997 28 3 3 337354 2001 2004 9 15 15 171201 2004 1973 2007 39 1 1 157179 1998 2004 14 3 3 314323 2005 1997 2039 1991 2005 2004 9 15 15 4580 1987 2001 5474 1975 2006 1997