Interaction has been central to theories of second language acquisition and pedagogy since the 1980s. Rivers explained the interactive perspective in language education: “Students achieve facility in using a language when their attention is focused on conveying and receiving authentic messages (that is, messages that contain information of interest to both speaker and listener in a situation of importance to both). This is interaction” (Rivers, 4). One of the first researchers to consider the importance of interaction was Hatch. She “made a pivotal and indelible mark on the field of SLA through her publication of two seminal papers on language learning and interaction” (Pica, 494). Hatch proposed that researchers should look towards interaction for insights into language learning development rather than the syntax of the language. In other words, she hypothesized that learners made progress as a result of real-life interaction rather than communicative competence arising out of the continuous practice of structures (Macaro, 172).

In SLA classroom, the interactional input of the teacher is also part of the teacher talk which is considered as the main source of language input. Krashen proposed the Input Hypothesis to explain how learners’ interlanguage develops and grammatical features are acquired when learners are exposed to input that contains grammatical features a little beyond their current level of competence (Krashen and Terrell, 32). The Input Hypothesis is explained in detail through the “i+1″structure. “i” stands for the current level of language competence of learners while “1” stands for the item that they are going to acquire. Therefore, the “i+1” structure indicates that learners are able to learn the language by being exposed to the input containing knowledge a little beyond their current level of competence.

Extending Krashen’s Input Hypothesis, Long put forward the Interaction Hypothesis, which he held could make learners’ SLA development possible (Long, 420). The Interaction Hypothesis differs from the Input Hypothesis in that it puts more emphasis on how to make input comprehensible. Krashen believes that “clues based on the situation and the context, extralinguistic information, and knowledge of the world make comprehension possible” (Richards and Rogers, 182). He proposed the term “premodification” which means to make input comprehensible by simplifying or modifying the input before exposing it to learners by using common or familiar words, phrases and sentences. While admitting the role of premodification, Long lays more emphasis on the interactional modification.

Through observing learners interacting with native speakers, Long concluded that what the former were doing was trying to improve the quality of the input they were receiving from the native speakers by giving them some verbal feedback or others to demonstrate that they had not understood. In this way, following Krashen’s earlier model, the native speakers’ input should become more fine-tuned to the immediate needs of learners thus allowing the latter to understand more easily. Therefore, the input will be made comprehensible as a result of interaction. Long proposed that interactionally modified input comes about as a result of the use of confirmation checks, clarification requests and comprehension checks by the two parties in a conversation (Macaro, 172).

While the first two hypotheses concentrate on the teachers’ input, Swain’s Output Hypothesis advocates that learners should be provided with more opportunities of producing output. In order to acquire a new language, it is not sufficient to notice it and keep silent. In order for the acquisition to occur, learners must also use the language in verbal production. First, it increases the intensity of the noticing on the new item as input. Second, it forces learners to attend to the construction of the new language before and during output. Third, it encourages teacher confirmation that the output is correct and provides evidence for learners’ hypotheses about the target language. This has been known as the Output Hypothesis. To sum up, the Output Hypothesis claims that output can promote language acquisition under certain conditions by allowing learners to produce output. Moreover, Swain believed that in order for students to achieve native-like language competence, they need to be pushed more in their output by providing them with more opportunities to use the target language in the classroom (Swain, 429).

In sum, the above three important theories of second language acquisition combine to reveal the effects of different types of classroom interaction on language learning from different angles. Therefore, in the field of second language acquisition, it is reasonable to state that classroom interaction plays a crucial role in promoting language learning.

2.1.2 General Classroom Interaction and Learning

When students learn in a classroom setting, a primary source of knowledge comes from teacher talk and teacher-student interactions, as the processes and transactions involved in the construction of meanings are mediated through the use of verbal communication. Central to Vygotsky’s sociocultural theory of human learning is the idea that social interaction plays an essential role in the development of cognition. First, cultural development appears between people on an inter-psychological plane and then inside the learner on an intra-psychological plane. This applies equally to voluntary attention, logical memory and the formation of concepts. All the higher functions originate as actual relationships between individuals (Vygotsky, 57). In addition, through the role of mediation, students are able to transform skills that lie in the zone of proximal development (ZPD, a term formulated by Vygotsky to refer to the area of students’ potential development). Williams and Burden state that “mediation is a term used by psychologists of the social interactionist school to refer to the part played by other significant people (the one usually with more knowledge, e.g. a teacher) in the learners’ lives, who enhance their learning by selecting and shaping the learning experiences presented to them” (Williams and Burden, 67). That is to say, the notion of the teacher assisting student performance through the “zone of proximal development” suggests that teachers can adjust the discourse on the inter-psychological to support the students’ evolving understanding of knowledge or development of complex skills. Therefore, Vygotsky’s theory implicates that learners should be provided with socially rich environments so as to explore knowledge domains with their fellow students and teachers.

2.1.2.1 Interaction Pattern of Initiation-Response-Feedback (IRF)

The three-part exchange structure known as “triadic dialogue” has been found to be common in classroom discourse. This discourse format typically consists of three moves, i.e. initiation (usually via a teacher question), student response and teacher evaluation. It is more commonly referred to as “IRE” (Mehan, 64). Accordingly, the teacher poses a closed question that is basically information-seeking, that requires a predetermined short answer, and that is aimed at the recall or lower-order cognitive level. He or she then praises or confirms correct answers and corrects those that are wrong. Sometimes, the three-part exchange structure is also known as “IRF”–initiation, response and feedback or follow-up as the third move may not necessarily be an explicit evaluative remark (Sinclair and Coulthard, 54). For instance, Wells has discussed various ways in which the teacher can provide feedback by encouraging students to express ideas, generate hypotheses and test them (Wells, 1986: 50).

The triadic dialogue, which is typical of traditional classroom teaching, is usually considered to restrict students’ thinking as students’ responses remain somewhat short and standardized, thus minimizing their role in the co-construction of meaning. Although such conventional teacher-questioning practices have thus been criticized, some researchers have pointed out that it is consistent with educational goals. For example, Newman, Griffin, and Cole argued that the three-part exchange has “a built-in repair structure in the teacher’s last turn so that incorrect information can be replaced with the right answers” (Newman et al, 127). Such an opinion is considered appropriate in that “the responsibility of teachers is viewed as ensuring that students acquire the knowledge that is normative within a particular culture” (Chin, 1316). In addition, Wells has argued that, when used effectively, “it is in this third step in the co-construction of meaning that the next cycle of the learning-and-teaching spiral has its point of departure” (Wells, 1993: 35). Therefore, the triadic dialogue can prove to be beneficial to classroom teaching if the teacher is willing to interact with students further. “An instance of this would be when the teacher asks a question that stimulates further productive thought, based on their evaluation of students’ previous responses. In such a case, the teacher would be guiding the development of students’ ideas by successively building on their contributions in a reciprocal manner” (Chin, 1315).

As for patterns of discourse, Mortimer and Scott expanded the IRE or IRF structure by identifying the IRFRF chain where the elaborative feedback from the teacher is followed by a further response from a student (Mortimer and Scott, 41). This form contributes to a dialogic interaction. As part of the feedback, the teacher could encourage a student by repeating his or her comment or asking for explanation. By establishing this pattern of discourse, the teacher is able to further explore students’ ideas.

As Wells suggested that the third step of the IRF sequence might have potential for productive discourse (Wells, 1993: 10), Chin established an analytical framework based on questioning to explore the specific forms of feedback. “Four aspects of classroom discourse (namely, content, type of utterance, thinking elicited, and interaction pattern) constitute the elements of the “questioning-based discourse” analytical framework” (Chin, 1322). She studied several science lessons from Year 7 which were observed across a variety of lesson structures. In her study, Chin identified four different types of feedback. The follow-up or feedback given by the teacher in the IRF sequence usually takes the form of a comment or statement followed by either another question, or further statements that expound more subject-related knowledge. Therefore, the feedback of the triadic dialogue could consist of a “comment-question” (C-Q) or “statement-question” (S-Q) couplet in which the question in the couplets may be considered as overlapping with the initiation of the next IRF sequence. However, Chin points out that, if there are no more questions asked, it takes the form of a “comment-statement” (C-S) couplet. Sometimes, feedback comprises only comments or statements. Chin’s findings suggest that “by changing the third move of an IRF questioning sequence from an explicit evaluation to one that includes “responsive questioning,” teachers can make their classroom discourse more thought-provoking and stimulate more elaborate and productive student responses” (Chin, 1340).

2.1.2.2 Teacher Questioning and Student Participation

As a prominent part of classroom discourse, teacher questioning plays an important role in classroom teaching and has been the focus of linguistic and pedagogical studies (Nunan, 192). There are several reasons why they are so commonly used in teaching.

They stimulate and maintain students’ interest.

They encourage students to think and focus on the content of the lesson.

They enable a teacher to clarify what a student has said.

They enable a teacher to elicit particular structures or vocabulary items.

They enable teachers to check students’ understanding.

They encourage student participation in a lesson. (Richards and Lockhart, 185).

Second language researchers have proposed that teacher questions play a crucial role in language acquisition. “They can be used to allow the learner to keep participating in the discourse and even modify it so that the language used becomes more comprehensible and personally relevant” (Banbrook and Skehan, 142).

In addition, many previous classroom-based studies have focused on the taxonomy of teacher questions. Barnes identified the closed-ended and open-ended questions (Barnes, 12). Moreover, a display question is a question to which the questioner already knows the answer while a referential question is a question where the teacher does know the answer and is genuinely interested in hearing the answers from students (Long and Sato, 268). And there are questions that either assist or assess (Tharp and Gallimore, 52). According to Brock, referential questions can increase students’ language output in class and thus promote language acquisition. “An increased use by teachers of referential questions, which create a flow of information from students to teachers, may generate discourse which more nearly resembles the normal conversation learners experience outside of the classroom” (Brock, 49).

Although at a theoretical level, referential questions are likely to trigger more immediate output than display questions, the distinction is too simplistic to offer an in-depth understanding of teacher questioning as it focuses on the types of questions rather than the abilities elicited. The nature of questioning in constructivist-based teaching environment is different. In such an environment, the teacher’s intent is to elicit what students think, to encourage them to elaborate on their previous answers and ideas, and to help them construct conceptual knowledge. Therefore, questioning can diagnose and extend students’ ideas and the teacher can engage students in higher-level thinking including analysis, application, synthesis and evaluation as those questions are open-ended and requiring one-sentence or two-sentence answers (Brookhart, 5).

Teaching questioning takes another form of “a reflective toss” in the feedback move of the IRF sequence (van Zee and Minstrell, 1997a: 216). A reflective question is posed to a student by engaging his or her previous response to a teacher question, thus extending the teacher-student interaction and further exploring students’ ideas. A reflective toss usually consists of three parts, i.e. a student statement, the reflective question and additional student statements (Chin, 1319). Furthermore, the teacher’s use of a reflective toss serves a series of subgoals. They include using questions to help students clarify their meanings, consider a variety of views and monitor the discussion and their own thinking (van Zee and Minstrell, 1997b: 266).

While most studies centered on the role of teachers in classroom, student participation is directly related with the quality of teacher-student interaction. Student participation has become a hot issue in the field of SLA research. Tsui conducted a survey among 38 teachers on the elements of reticence in middle school classrooms in Hong Kong and discovered five influential factors: students’ low English proficiency, students’ fear of making mistakes and getting laughed at by others, lack of wait time for students to think due to teachers’ intolerance of reticence, uneven allocation of turns to students and teachers’ incomprehensible input (Tsui, 148-155). Karp and Yoels carried out a one-month observation program in the 10 classes at an American private university and identified the “consolidation of responsibility” (Karp and Yoels, 429). On one hand, teachers would only call on some specific students to answer their questions. On the other hand, some other students would remain silent in class as they were rarely called upon. Through the observation of fifteen classroom sessions and out-of-class interviews with two female and two male students, Morgenstern discovered that “there were many opportunities for student speech, but a core of five to six students seemed to monopolize these opportunities. Student actions and attitudes, recorded by observation and interview, revealed four tacit rules for class participation: (1) do not ask stupid questions; (2) do not waste the teacher’s time; (3) do not waste class time; and (4) try to find the answer before asking the teacher. Some students function under the assumption that only those with the most knowledge should speak, thus assuming a hierarchy of knowledge” (Action and Inaction: Student and Teacher Roles in Classroom Participation).