Introduction

The field of second and foreign language acquisition is influenced by many interrelated factors that have a direct impact on students’ achievements. After realizing that language learning is not exclusively affected by the students’ cognitive abilities but also by the affective domain, a plethora of hypotheses have been introduced emphasizing the role that students’ emotions play in the process of learning. The affective filter hypothesis postulated that language learning is partially related with certain affective aspects like motivation, attitude, anxiety, and self-confidence (Krashen, 1985). Students who possess a low affective filter are supposed to have a greater potential to absorb intake while others who possess a high affective filter are said to have lesser chance of absorbing intake since the input can be blocked from reaching their language acquisition device (LAD) (Du, 2009). What type of classroom climate prevails depends primarily upon the teacher’s verbal and non-verbal behaviour. Therefore, teachers must solicit for maintaining an emotionally supportive climate in order to promote better learning since there is evidence that such a classroom atmosphere can increase students’ motivation, interest, and engagement (Brackett et al, 2009). Foreign language interaction analysis (FLint) is among the systems that can provide a better understanding about the climate of the classroom (Brown, 2001). Particularly, it can be used to measure different interactional aspects including proportions of teacher-student talk, the type of influence maintained in FL classrooms, teachers’ propensity toward dealing with students’ emotions, motivation, and ideas. It allows also the examination of the way in which students’ responses are supplied and the ratio of students’ initiation during interaction.

As far as studying EFL oral discourse in Algeria is concerned, many studies were carried out to examine what happens inside the local classrooms. The major part of the studies relied mainly on qualitative studies. Also, most of the scholars made use of discourse analysis as an analytical tool, as well as questionnaires and interviews, without resorting to analytical observation schemes for the examination of what happens throughout the delivery of lessons. As a matter of fact, there is a need to adopt new research instruments to the analysis of oral classroom discourse for the sake of covering new spheres of research that have not been sufficiently approached by researchers in relation to the local context of teaching EFL. To fulfil such a need, interaction analysis as a research method to the examination of classroom oral discourse can be used to build a clearer understanding about the nature of interaction that happens during EFL courses.

Furthermore, there is a lack of research, in the local context, about certain aspects of the teaching-learning process, such as : the climate of EFL classrooms, proportions of teacher and student talk, the nature of feedback provision, the ratio between choral responses and individual responses, students’ initiation, and the tendency of teachers towards reacting to students’ ideas and feelings. Through the use of Flint framework all these aspects can be investigated through a systematic analysis of data. This research tool can help both teachers and researchers in gaining further insights about the features of interaction, especially in regard to the nature of the teaching verbal behaviour applied whilst delivering lessons. It must be noted that to our best knowledge, this research is the first to implement the Flint system for the analysis of interaction in EFL classrooms within the local context of Algeria.

1. Literature review

Progress in learning a foreign language is underlined by numerous interwoven factors impacting learners’ achievements in the language they want to obtain a command over. Since a major part of learners’ exposure to FL is associated with the time they spent in classrooms, the verbal interaction that learners engage in with their teacher is believed to have a direct correlation with their outcomes in the target language (Alwright, 1984). Besides the essential elements that make up the construct of interaction (e.g. input, negotiation of meaning, output, feedback), the social-emotional climate of the classroom also adds up as an important factor affecting language acquisition and learning in general (Krashen, 1981 ; Reyes, Brackett, Rivers, White, & Salovey, 2012). This is where the role of teachers’ verbal behaviour raises to the surface since the way in which language educators conduct their courses is partially determinant of the affective dimension embedded in oral discourse (Alwright, 1984 ; Alonso-Tapia & Nieto, 2018).

Interaction analysis, as a research method, allows investigating the communication patterns that occur in classroom (McKay, 2006). It permits the researcher to observe teacher-student talk systematically by classifying the units of verbal behaviour according to a set of predetermined categories. In the same context, Moskowitz model of foreign language interaction analysis enables the exploration of the interactional categories constituting teacher-student talk and can offer a clearer understanding about certain features of the climate involved in the teaching-learning process (Brown, 2001), such as the level of students’ motivation and control exerted by teachers. Also, it permits the examination of teachers’ behaviour in terms of the kind of influence exerted during the delivery of courses through the quantification of verbal behaviour. Moskowitz (1971) points out that the system can be used to determine whether the teacher applies direct or indirect influence and also to measure the amount of talk sustained by the teacher and the students along with many other aspects of classroom discourse.

FLint system was inspired from the consecutive works of Flanders though which the latter developed an influential system called Flanders interaction analysis system (FIACS) dedicated for the examination of interaction within pedagogical settings (Brown, 2001 ; Chaudron, 1986a). Flint system is actually an extension of FIACS designed specifically to analyse the classroom events that occur in FL classrooms (Moskowitz 1971). It assigns twenty categories of interaction to the verbal and non-verbal events happening throughout the ongoing of lessons as it targets both teacher and student talk. However, teacher talk is given much attention in this system in account to the assumption that classroom discourse is dominated by teachers, considering their pedagogical role of maintaining guidance, course delivery, and the orchestration of interaction. In this respect, the common belief in literature about language studies is that teachers’ verbal behaviour dominates two thirds of the whole time allocated to language lessons (Sinclair & Coulthard, 1975).

FLint model identifies two type of teachers’ verbal behaviour, namely indirect influence and direct influence. Applying a more indirect influence during the delivery of lessons is believed to be more effective in encouraging learners to participate and to increase their involvement in classroom discussion (Flanders, 1974). There are six categories that fall within the frame of indirect influence. They are labelled by instances in which the teacher issues a move that involves dealing with feelings, praising or encouraging, joking, using students’ ideas, literal repetition of students’ responses, or asking questions. On the other hand, there are six categories that pertain to the direct influence which teachers apply during FL lessons. They occur in moments that involve giving information, correcting without rejection, giving directions, directing pattern drills, criticising student behaviour, and criticising student response.

Indirect influence is believed to guarantee a better quality of interaction because when teachers persist in fostering indirect teaching there would be a greater possibility of promoting a good classroom climate for the delivery of the course and increasing students’ freedom in class. Flanders (1974) states that all the actions falling into the indirect pattern of behaviour tend to increase and reward students’ participation, and to give the students the opportunity to become more influential inside the classroom. On the other hand, the direct influence that teachers exert during verbal interaction tends to be more power driven, managerial and authoritative in nature. Flanders (1965) points out that “direct influence consists of those verbal statements of teacher that restrict freedom of action, by focusing attention on a problem, interjecting teacher authority or both” (p. 9). That is to say, it decreases the freedom of students as it either bring about passive learning or calls for their compliance with the teachers’ instructions, directions, or criticism.

As it is mentioned earlier, students’ talk is also taken into consideration in this model and it allocates three interactional categories to describe the verbal behaviour of students. These involve specific student’s response, choral students’ response, and student’s open ended or initiated response. It also assigns separate neutral categories that represent moments in which silence, confusion, or laughter take over in the classroom. Additionally, the use of non-verbal behaviour and native languages is covered through the last utilities labelled “e” and “n”. The following table elaborates the categories of interaction accounted for in FLint model :

Table01 : the FLint system (adapted from Moskowitz, 1971, p. 213)

Teacher talk	Indirect influence	1-Deals with feelings : In a non-threatening way, accepting, discussing, referring to, or communicating understanding of past, present, or future feelings of students. 2- Praises or encourages : Praising, complimenting, telling students why what they have said or done is valued. Encouraging student’s to continue, trying to give them confidence. Confirming answers are correct. 2a-Jokes : Intentional joking, kidding, making puns, attempting to be humorous, providing the joking is not at anyone’s expense. Unintentional humor is not included in this category. 3- Uses ideas of students Clarifying, using, interpreting, summarizing the ideas of students. The ideas must be rephrased by the teacher but still recognized as being student contributions. 3a- Repeats student response verbatim : Repeating the exact words of students after they participate. 4-Asks questions : Asking questions to which an answer is anticipated. Rhetorical questions are not included in this category.
	Direct influence	5-Gives information : Giving information, facts, own opinion or ideas, lecturing, or asking rhetorical questions. 5a- Corrects without rejection : Telling students who have made a mistake the correct response without using words or intonations which communicate criticism. 6- Gives directions : Giving directions, requests, or commands which students are expected to follow. 6a-Directs pattern drills : Giving statements which students are expected to repeat exactly, to make substitutions in (i. e., substitution drills), or to change from one form to another (i. e., transformation drills). 7-Criticizes student behavior : Rejecting the behavior of students ; trying to change the non-acceptable behavior ; communicating anger, displeasure, annoyance, dissatisfaction with what students are doing. 7a-criticizes student response : Telling the student his response is not correct or acceptable and communicating by words or intonation criticism, displeasure, annoyance, rejection.
Student talk		8- Student response, specific : Responding to the teacher within a specific and limited range of available or previously shaped answers. Reading aloud. 8a- Student response, choral : Choral response by the total class or part of the class. 9- Student response, open-ended or student initiated : Responding to the teacher with students’ own ideas, opinions, reactions, feelings. Giving one from among many possible answers which have been previously shaped but from which students must now make a selection. Initiating the participation.
Silence, confusion, or laughter		10-Silence : Pauses in the interaction. Periods of quiet during which there is no verbal interaction. 10a-Silence-AV : Silence in the interaction during which a piece of audio-visual equipment, e.g., a tape recorder, filmstrip projector, record player, etc., is being used to communicate. 11-Confusion, work-oriented : More than one person at a time talking, so the interaction cannot be recorded. Students calling out excitedly, eager to participate or respond, concerned with task at hand. 11a- Confusion, non-work oriented : More than one person at a time talking, SO the interaction cannot be recorded. Students out-of-order, not behaving as the teacher wishes, not concerned with the task at hand. 12-Laughter : Laughing, giggling by the class, individuals, and/or the teacher. e- Uses the native language : Use of the native language by the teacher or the students. This category is always combined with one of the 15 categories from 1 to 9. n- Nonverbal : Nonverbal gestures or facial expressions by the teacher or the student which communicate without the use of words. This category is always combined with one of the categories of teacher or pupil behavior.

As illustrated above, each category of interaction has a corresponding numerical number that represent it. This permits the researcher to observe what happens inside the classroom by coding each category according to its identifying number in an observational sheet. In this model, assigning codes to classroom verbal events recurs consistently every three seconds throughout the period of observation (Moskowitz, 1971). Even if the observed behaviour does not change, the same category will be recoded again. Eventually the observer will get a chain of numerical data (tallies) representing all the verbal and non-verbal interchanges that had occurred during classroom communication. The series of category numbers that the researcher gets after the coding procedure are subsequently arranged into pairs in order to be tabulated into a matrix for the sake of facilitating the calculation procedure. That is to say, the encoded data will be next decoded based on the formulated tallies. Eventually, through the use of the calculation formula that consist of data derived from the matrices, the researcher will be able to calculate different ratios and percentages, which reflect diverse aspects of teacher-student talk and provide insights about the nature of interaction that took place in EFL courses.

2. Methodology

This study involves a descriptive research that was carried out based on a real-time systematic observation of EFL classrooms. The gathered data was examined though classroom interaction analysis as it employed FLint system for encoding and decoding data. The same system was used for the identification of the interactional patterns that took place during EFL courses. For this purpose, the data were analysed statistically to determine the ratios reflecting the nature of interaction and the aspects of teacher-student talk.

2.1. Research setting and participants

This research was conducted in four EFL classrooms selected randomly from two Algerian universities, namely the University of Badji Mokhtar at Annaba and the University of Chadli Bendjedid at El Tarf. It involved four different teachers and 78 third- year students of EFL studying in the departments of English language at the two mentioned academic institutions. In order to conduct the study, a signed document was approved by the administrations of the two universities and all participant gave their oral consent to take part in the study.

2.2. Instruments and Procedure

This study implements FLint as a means to analyse classroom interaction. In order to collect data, the researcher coded the patterns of interaction that occurred during EFL lessons in real-time at an interval of three seconds. An audio recorder was used to back up the data retrieved through observation. For the sake of maintaining an accurate analysis with regard to the features of interaction in EFL classroom, a number of ratios were calculated based on the encoding procedure and the matrices derived from systematic observation.

The ratio showing whether teachers provided feedback with or without rejection is calculated by comparing the frequency of categories 5a and 7a (5a/7a). The ratio that demonstrates whether students responded individually or chorally to teachers, when being directly nominated or prompted to respond, was measured by comparing the frequency of occurrence of categories 8 and 8a (8/8a). Moreover, the following ratios were calculated : teacher talk (TL), student talk (ST), silence, confusion, or laughter (SCL), direct teacher talk (DTT), indirect teacher talk (ITT), indirect to direct ratio (I/D), teacher response ratio (TRR), and student initiation ratio (SIR). For this purpose, the following formula were used, drawing on the calculation procedures suggested by Moskowitz (1971) and Flanders (1970) :

TT = (categories 1+2 + 2a +3 + 3a +4+5 + 5a +6 + 6a +7 + 7a)/NX100
TT refers to the proportion of the teacher’s verbal behaviour sustained during interaction.

ST = (categories 8 + 8a +9) /NX100
ST indicates the ratio that represents the frequency of students’ verbal behaviour.

SCL = (category 10 + 10a +11 + 11a +12)/NX100
SCL refers to the proportion of time in which silence, confusion, or laughter took over the classroom.

DTT = (categories 5 + 5a +6 + 6a +7 + 7a) /NX100
DTT represents the ratio of the teacher’s action that restricted students’ participation.

ITT = (categories 1+2 + 2a +3 + 3a +4) /NX 100
ITT denotes the tendency of teacher’s actions that encouraged students’ participation.

I/D = (categories 1+2 + 2a +3 + 3a +4)/ (categories 5 + 5a +6 + 6a +7 + 7a)
I/D identifies whether the teacher was direct or indirect in terms of the influence exerted during interaction.

TRR = (categories 1+2 + 2a +3 + 3a)/ (categories 1+2 + 2a +3 + 3a +6 + 6a +7 + 7a) X100
TRR It represents the ratio for the teacher’s propensity toward reacting to the ideas and feelings of students.

SIR = (Category 9) / (Categories 8 + 8a +9) X100
SIR indicates the proportion of student talk identified by the observer as an act of an initiation.

In addition, the overall mean values are calculated, by adding up the total frequencies of the pertinent categories that correspond to each ratio in classes A, B, C, D, following the previous calculation formula.

3. Results and Discussion

3.1. Results

The results showed that teachers talk occupied a proportion of time that varied from 67.92 % to 72.98 % in the observed sessions. The average of teacher talk in all sessions accounted for 70.80 % of classroom interaction. The time allocated for students talk was significantly lesser since only 24.63 % of the total time of the observed courses pertained to students discourse as the proportions of student talk ranged from 22.13 % to 26.58 %. Silence or confusion as a pattern of interaction took up merely an average 4.56 % in the four observed classrooms. The higher proportion of silence or confusion involved a value of 5.48 % while the lower value was equal to 3.82 %. The following table illustrates the results corresponding to each lesson along with the mean percentage of the four observed sessions :

Table 02 : The distribution of TT, ST, SCL and corresponding mean percentages

Variables	Class A	Class B	Class C	Class D	Mean	Minimum	Maximum
TT ( %)	71.69	72.98	67.92	71.46	70.80	67.92	72.98
ST ( %)	24.19	22.13	26.58	24.71	24.63	22.13	26.58
SCL ( %)	4.11	4.87	5.48	3.82	4.56	3.82	5.48

The verbal behaviour of teachers which correspond to indirect teacher talk varied in a range of value that extended from 28.28 % to 37.36 % of the total time pertaining to teacher-student talk. The mean percentage of indirect teacher talk was 32.56 %. On the other hand, direct teacher talk accounted from 34.32 % to 40.32 % of the observed verbal behaviour that pertain to EFL instructors. The calculated mean percent of direct teacher talk of the four sessions was 38.15 %. The results showed that three teachers exerted direct influence while only one teacher maintained indirect influence in teaching EFL since one teacher marked a value of 1.08 which greater than 1. The lowest ratio pertained to the teacher in class C whose verbal behaviour marked a ratio of 0.71. The calculated mean ratio of indirect to direct teaching was 0.85 (less than value “1”) which correspond to the average value in the four examined classes. The table below demonstrates the results :

Table 03 : Percentages of ITT, DTT, and I/D ratios along with the average values

Variables	Class A	Class B	Class C	Class D	Mean	Minimum	Maximum
ITT ( %)	37.36	33.47	28.28	30.75	32.56	28.28	37.36
DTT ( %)	34.32	39.51	39.64	40.32	38.15	34.32	40.32
I/D (ratio)	1.08	0.84	0.71	0.76	0.85	0.71	1.08

By comparing the categories 5a and 7a, this study determines whether teachers provided corrective feedback through or without rejection. The results revealed that the amount of feedback supplied without rejection (cat 5a) is always greater than the amount of turns in which the teacher rejects students’ wrong responses. The calculated average ratio indicates that category 5a was 4.23 times greater than the instances of occurrence of category 7a. The comparison of category 8 with category 8a showed that specific student’s responses are always more frequent than choral responses. The average frequency of category 8 is 6.18 times greater than category 8a. The following table demonstrates the results :

Table 04 : the ratio of categories 5a/7a and 8/8a along with the calculated average ratio

Variables	Class A	Class B	Class C	Class D	Mean	Minimum	Maximum
Cat5a/Cat7a (ratio)	4.12	2.75	4.5	5.4	4.23	2.75	5.4
Cat8/Cat8a (ratio)	3.93	9.37	7.43	8.07	6.18	3.93	9.37

Teacher responses to students’ feelings acts as an index for the emotional climate in the classroom. The findings disclosed that teachers’ response ratios to students’ feelings, attitudes, and ideas varied from 88.28 % to 97.41 %. The mean percentage of teacher response ratio was 93.21 % indicating a high tendency toward fostering the emotional climate in EFL classrooms. The proportion of student talk that corresponded to the acts of initiation (SIR) ranged from 60.28 % to 66.83 %. The average of SIR across EFL classrooms was 62.60 %. The results are illustrated in the table below :

Table 05 : the percentages of TRR and SIR along with the calculated mean values

Variables	Class A	Class B	Class C	Class D	Mean	Minimum	Maximum
TRR (%)	88.28	93.30	96.22	97.41	93.21	88.28	97.41
SIR ( %)	61.08	60.28	66.83	60.68	62.60	60.28	66.83

3.2. Discussion

The distribution of classroom time is reflected in the proportions of interaction behaviour embodied either in teacher talk (TT), student talk (ST), and silence, confusion or laughter (SCL). The results have shown that the major part of EFL lessons was dominated by TT since the overall mean corresponding to TT constitutes 70.80 % of interaction in the observed classrooms. Such a high proportion of TT contradicts the assumption that teachers’ inclination toward maintaining speech decreases as the age and the level of learners upgrade. On the other hand, the mean percentage of ST was equal to a value of 24.63 % in the same observed sessions. This can imply that EFL lessons were teacher-centred rather than student-centred since the proportions of TT occupied more than two thirds of classroom time in all the examined lessons. Also, the findings suggest that teachers must work to increase ST since there is a positive correlation with the amount of ST and language learning achievement (Sedova et al., 2019). Furthermore, the moments in which SCL took up over classroom interaction constituted only 4.56 % of EFL lessons which indicates that teacher-student talk was characterized with a continuity of discourse as the flow of communication was rarely disrupted or paused during class discussion.

The findings that correspond to the type of influence that teachers exert during interaction showed that teachers tend to sustain more direct talk while delivering EFL, because DTT covered 38.15 % of instructors verbal behaviour compared to a lesser value of 32.56 % pertaining to ITT. Drawing on the previous results, it can be said that teachers’ behaviour was not favourable for promoting students’ freedom of action and participation, considering that only one teacher was able to maintain an adequate indirect influence during interaction. The mean I/D ratio was equal to 0.85 (less than 1) indicating that direct influence prevailed inside EFL classrooms and consolidates the previous claim. In this regard, Flanders (1965) states that direct influence of teachers restricts the students’ freedom as it makes them momentarily more dependent on their instructors. This implicates that teachers should put more emphasis on adopting a more indirect approach in order to maximize students’ freedom of action and encourage participation (Mangal & Mangal, 2009). Nevertheless, it is important to note that the average I/D ratio (0.85) is close to the satisfactory value recommended to be maintained in the teaching-learning process.

Moreover, the ratio between the instances in which teachers did not reject students wrong answers were 4.23 times greater and those in which rejection was observed. This pedagogical predisposition toward providing feedback without rejection was present in all the observed sessions. Such a propensity indicates that feedback provision was constructive in nature rather than being obtrusive or detrimental for the learning process since students’ would not be intimidated as a result of fearing harsh criticism (Ghalley & Rai, 2010). Supplying feedback without rejection can encourage learners to take part in discourse as it reduces the level of inhibition in terms of students’ participation (Mustapha, Nik, & Yunus, 2019). Furthermore, it suggests that teachers’ did not take students’ views and attempts for granted as their corrective interventions were mostly appropriate and constructive rather than being critical (Harmer, 2007). The comparison between categories 8 and 8a provides an index for whether students tend to respond specifically or chorally when prompted by teachers. The results showed that specific student responses were supplied with a rate which is 6.18 times higher than the case of choral responses. This may either imply that teachers’ solicited elicitations and nominations were mostly targeted toward individual students or that the nature of the elicitations were not conducive for collective responses.

Teacher response ratio (TRR) provides an index for the emotional climate of the classroom (Buch, 1975) ; it represents the tendency of teachers’ for the endorsement of learners’ emotions and ideas. The high ratio of TRR (93.21 %) reveals that teachers were not disposed toward displaying an authoritarian attitude or self-reference whilst conducting EFL courses (Tiwari & Pandey, 2013). Also, it attests to teachers’ predilection toward appreciating students’ ideas and feelings as it is considered to be an encouraging behaviour that increases students’ achievements and willingness to engage in class discussion (Amidon, 1966 ; Chowdary, 2004). Students’ initiation ratio (SIR) represents the frequency with which students took the initiative and participated voluntarily. The average ratio of 62.60 % obtained in this study is significantly high indicating that students were motivated, since they took the initiative to participate willingly in most of the cases without being called upon or invited by teachers (Harmer, 2007). It can further imply that students’ responses were not restricted to the scope of teachers’ elicitations as their contributions were not limited to a previously pre-empted narrow range of possible answers or responses. Also, such a high frequency of initiation can be correlated with the fact that students are adults (post-18), and as consequence, they were ready to take responsibility for their own learning without necessarily being granted with a permission to share their ideas.

Conclusion

The main findings of this study revealed that teacher talk occupied more than two-thirds of the classroom time and that most of teachers were disposed toward exerting a direct influence while conducting EFL lessons. The mean I/D ratio showed that teachers’ behaviour was prone to the recommended value of indirectness that should exist during interaction. The ratio that pertain to the amount in which SCL prevailed was very low suggesting that interaction was characterized with a continuity of discourse with few moments of silence or disruption. Feedback of teachers was generally provided without rejection of students’ wrong answers implying that teachers appreciated students’ contributions and promoted a supportive atmosphere. The comparison between students’ specific and choral responses (categories 8 and 8a) showed that students reacted to most of the teachers’ nominated elicitations by responding individually not collectively. However, most of students’ contributions were supplied either through initiations or open-ended responses indicating a high degree of motivation and willingness to engage in class discussion. In addition, teachers proved to have a very high tendency toward fostering students’ feelings and ideas in all the observed sessions. The former results suggest that the socio-emotional climate of the classroom was generally supportive and conducive for students’ participation. Nonetheless, teachers should endeavour to increase the amount of students’ talk and to exert a more indirect influence in order to promote a better quality of interaction during the conduction of EFL courses. Further studies need to be carried out in the local context of Algeria to acquire a better understanding about interaction in EFL classrooms and to establish FLint normative ratios that correspond the local context. Fulfilling such a need would allow researchers to maintain accurate interpretations about the results derived from the system and to gain a deeper understanding about both the nature of interaction and the climate of EFL classrooms.