Optimizing Interaction in the Second Language Classroom. (Linguistics)
Pawlak, Miroslaw, Studia Anglica Posnaniensia: international review of English Studies
In recent years we have witnessed a significant shift of emphasis as far as second language acquisition research is concerned. This shift is reflected in the fact that many researchers have started to recognize the importance of the processes taking place in the classroom in second language development. Such a change of heart was effected by the failure of most input-output studies to produce conclusive results on the efficiency of various teaching methods and, consequently, by the recognition of the second language classroom as a cultural entity in which the whole learning process is predetermined by the interaction taking place between its participants as well as by the context in which this interaction occurs (cf. Allwright 1988; Allwright -- Bailey 1991).
As a result, a considerable amount of research has been carried out on interaction in second language classrooms as well as the factors that influence it. Researchers have focused their attention on participation which constitutes a direct manifestation of interactional processes. In an attempt at discovering the participation structure in an instructional context, much attention has been paid to turn-taking, initiative, teacher talk, learning strategies and repair. A considerable amount of research has also been directed towards pinpointing the differences in interaction between naturalistic settings and the second language classroom and finding out what sort of influence these differences can exert on second language development (cf. Allwright -- Bailey 1991; Ellis 1992, 1994; van Lier 1988).
The main goal of this study is to explore the ways in which the quality of interaction taking place in a typical polish high school second language classroom affects the teaching of speaking skills. An attempt will also be made to determine how different aspects of educational discourse can be shaped in order to make the acquisition of these skills as effective as possible. At the outset the relationship between the notions of input and interaction will be addressed, and a few comments will be made on the facets of classroom interaction to be investigated. Then an analysis of two sets of transcribed data will be carried out in accordance with the ethnographic tradition. The first set of transcripts, or the baseline data, will make it possible for us to determine to what extent the interactional patterns found in an instructional setting deviate from those encountered in general conversation, and in what ways they can be altered to promote real-life interaction and negotiation of meaning. The data collected in the second stage of the study come from English language classrooms in which some aspects of interaction were modified to make discourse more similar to communication in naturalistic settings. The patterns of interaction found in those classes will be analyzed and some comments will be made as to whether the modifications made the classes more conducive to the development of the speaking skills.
2. Input and interaction in second language acquisition
Research theory and practical experience all point to the fact that the samples of target language that learners are exposed to, often referred to as input, play a crucial role in second language acquisition. The recognition of the second language classroom as a social setting in which language learning takes place by means of "meaningful interaction", however, made it necessary for researches to explore the relationships between input and interaction. Although the indispensability of exposure to the target language was not called into question, some researchers argue that input in itself can no longer be regarded as the only factor determining the route and rate of second language acquisition.
The paramount importance of input, and to be more precise "comprehensible input" in language learning was emphasized by Krashen and later became the cornerstone of the so-called natural approach. Basing his theory on research into first language acquisition, Krashen (1981) draws a distinction between acquisition and learning. Acquiring a language means developing ability in a language by using it in natural, communicative situations. Learning, on the other hand, differs from acquisition in that it consists in attaining a conscious knowledge about grammar. In Krashen's view, conscious rules mastered by means of learning can only perform a limited function of an editor monitoring the utterances initiated by acquisition. In other words, it is the acquired system that is responsible for the generation of utterances and the sole function performed by the learned system is making changes to those utterances.
Krashen hypothesized that acquisition takes place only when people understand the messages in the target language, and such reasoning laid foundations for his famous Input Hypothesis. The hypothesis states that language is acquired (not learned) by understanding input that is a little beyond our current level of (acquired) competence, or to use a Selinker's (1972) term, our level of interlanguage development. Krashen (1985: 2) called this kind of input i+1 where i designates the learner's current level of competence and +1 designates the stage where input is challenging (immediately following i in some natural order), but still comprehensible with a certain degree of effort. In addition to that, Krashen claims that "listening comprehension ... is of primary importance in the language program and that the ability to speak ... fluently in a second language will come on its own with time" (Krashen -- Terrell 1983: 32). In other words, speaking ability emerges by itself after the acquirer has built up his compete nce through being exposed to comprehensible input. The above claim resulted in the promotion of the so-called silent period in language teaching and in the rejection of incomprehensible input; that is to say input by far exceeding the learner's competence. As Krashen and Terrell (1983: 19) put it: "incomprehensible input ... does not seem to help language acquisition".
Krashen's (1985) input hypothesis has been criticized on a number of counts. Some researchers pointed out that positive input may not always be sufficient to provide the learner with enough information about a particular structure, and, in such circumstances, specific negative feedback is indispensable (White 1987: 107). It was also argued that depriving learners of the opportunity to speak in the early stage of second language development does not make sense since the learner's own output can serve as a means of trying out the hypotheses about the target language (Gregg 1984: 88). The most important criticism, however, was connected with the fact that by calling into question the importance of speaking, Krashen's theory downplayed the role of interaction in second language acquisition. Some researchers argue that it is dubious whether mere exposure to input, even if it is comprehensible, actually promotes language development, and that interaction in itself is productive because it reflects the effort on the part of the learner to comprehend input.
Long's (1983) model, for example, emphasizes the role of conversation in getting comprehensible input. In his view, it is a two-way exchange of information that makes the less competent speaker provide feedback on his comprehension, which results in a negotiated modification and, thus, supplies him with comprehensible input. This model, however, continues to acknowledge the decisive role of comprehensible input and is challenged by Allwright and Bailey (1991:122) who argue that "language acquisition can perhaps best be seen, not as the outcome of an encounter with comprehensible input per se, but as the direct outcome of the work involved in the negotiation process itself." To put it differently, it is the effort to negotiate meaning by means of interaction rather than the intended outcome of that work in the form of comprehensible input that promotes language acquisition. In Allwright and Bailey's view, the term "negotiated interaction" refers to the modifications that occur in conversations between competen t and less competent target language speakers and aim at resolving communication breakdowns.
The indispensability of meaningful interaction in second language acquisition is accepted by many researchers. Swain (1985, 1993), for instance, suggested that when producing utterances in the target language learners can oftentimes notice a linguistic problem as a result of feedback they receive from interlocutors and this realization can, in turn, cause them to modify their output. As a consequence, they are forced into a more syntactic processing mode and set about analyzing the structures, thus spurring the development of their interlanguage. The validity of this proposal, known as the "output hypothesis", was corroborated in a number of studies (e.g., Pica et al. 1989; Nobuyoshi -- Ellis 1993) which indicate that output modifications do occur in response to such conversational moves as clarification requests and confirmation checks.
The significance of interaction in second language development is also borne out by studies into the processes accountable for first language acquisition. On the basis of an analysis of caretaker-child interaction, some researchers emphasize the reciprocity of interaction over input. According to Ellis (1984: 88): "It is by using his linguistic resources in production that the child is able to clarify both the semantic and grammatical systems of the language" and to attain this end "it is important that he can get into conversation and play its part in sustaining it." (1) He adopts the same line of reasoning in the case of second language acquisition and claims that successful language learning occurs "not when the teacher provides an input with x features, but when reciprocal interaction occurs" (Ellis 1984: 89).
Another case for the significance of interaction is made by Rivers (1988) who also bases her claim on research into the psychological processes involved in the comprehension and production of the native language. She argues that some striking differences between listening and speaking prove that it is highly unlikely for intensive listening alone to lead to fluent and effective production of utterances. Her argument is that when listening, learners interpret what they are hearing mainly by means of the sequence of semantic elements that allows them to construct a plausible message from what they are hearing, and rely to a large extent on inference. To put it differently, mere exposure to input encourages learners to rely on semantic cues and hinders them from focusing on the syntax.
Speaking, on the other hand, enables the learner to control the language that he intends to use by selecting both the lexical items and syntactic structures to be employed. Speakers need grammar to express their meaning precisely and consequently to retain the listener's attention. As a result, it can be claimed that neither comprehension nor production can lead to the other in some incidental and subconscious way. As Rivers (1988: 7) puts it: "comprehension and expression of meaning are in constant interaction in real-life communication".
3. Facets of classroom interaction investigated in this study and their influence on the development of the speaking skills
The transcribed data were analyzed with regard to such facets of classroom interaction as turn-taking, initiative, topic and activity management. It was also attempted to pinpoint the common denominator of all those processes by describing different interaction types and ultimately the participation structure to be found in the second language classroom. Before moving on to the analysis itself, it seems warranted to take a closer look at those phenomena as well as the framework that was adopted in their analysis.
Both in real-life and instructional contexts interaction is not random and the matter of who gets to speak is governed by certain regularities. The existence of those regularities is an outcome of the fact that interlocutors adhere to seldom-stated, but deeply ingrained, tacit norms which impose a rigid structure on any kind of discourse. The main purpose of describing turn-taking mechanisms, as those norms are often referred to, is to study the systematic nature of speaker change in different settings and to determine "how one utterance follows another in a rational, rule-governed manner" (van Lier 1988: 94).
Made up of many potential participants whose competence in the target language varies, the second language classroom is extremely vulnerable to turn-transition and distribution problems which can often lead to breakdowns in communication. In order to ward off such difficulties, classroom discourse is structured and has some distinctive features that make it different from interaction in naturalistic contexts. Among those features are the need for centralized attention (one speaker at a time, rare instances of simultaneous talk), the predominance of the teacher in deciding in advance what is going to happen in the classroom and the fact that the ideal performance or sequence during the entire lesson or its parts is often predetermined and learners are instructed about the permissible ways to participate.
If the rules of turn-distribution instituted in the second language classroom are extremely strict, transition ceases to be problematic and consequently ceases to be governed by negotiation, competition or personal initiative. As for participants, they become more concerned with trying to comply with the rules than resolving transition and distribution problems and that, in turn, leads to a significant diminution in their influence on classroom discourse. (2)
The classification of turns that that has been utilized in this study is based on the model proposed by Allwright and Bailey (1991: 128), but a few alterations have been made in order to tailor it to our needs. (3) As a result, it has been assumed that the turns encountered in the second language classroom fall into three categories:
1) self-initiated turns which can be further subdivided into:
-- bidding for turns which includes both verbal and non-verbal behavior
-- unbidden turns which include listening responses, inter-turn repair and repair initiation
2) teacher-initiated turns which also fall into two groups:
- by direct nomination using both verbal and non-verbal clues
- by general nomination also using verbal or non-verbal clues
Another aspect of classroom discourse that can have a considerable impact on the development of the speaking skills is the personal involvement of students, commonly referred to as initiative. According to van Lier (1988), both in general conversation (4) and in the second language classroom, initiative can be expressed by retrospective action (self-selection), prospective action (allocation), topic and activity management. What differs the second language classroom from naturalistic contexts is the degree to which the participants of the two types of discourse are allowed to express their initiative in all the four areas.
3) another-learner initiated turns
Some researchers argue that a second language lesson should be viewed as an extremely constrained speech exchange system as it is in many respects similar to a formal, pre-planned ceremony. According to McHoul (1978) and Sinclair and Brazil (1982), for instance, interaction in the second language classroom deviates from natural conversation in that it is only the teacher who has the power to direct speakership in any creative way, often restricts the learners' opportunities to take initiative and assumes a controlling role in conducting a fairly ritual conversation. As for the students, they are reduced to the status of pawns that have little impact on who the next speaker will be and that can participate in classroom proceedings only after they have been nominated by the all-mighty teacher.
Every interaction can also be characterized by its topic and the kind of activity it involves. In the analysis that follows topic and activity are not viewed as units in terms of which discourse is structured, and the focus is on more inclusive concepts of topic and activity orientation. In the second language classroom, it seems to be the extent to which the balance between the two orientations is skewed that makes it more or less conducive to the development of the speaking skills. The prevalence of topic-oriented discourse is generally considered to foster meaningful interaction and the predominance of activity orientation is bound to hamper active participation and negotiation for meaning.
The interplay between the two different types of orientation and such additional variables as the locality and non-locality of the elements responsible for the organization of interaction as well as academic task and social participation structure (5) enabled van Lier (1988: 171-173) to distinguish the following interaction types (ITs):
IT 1 (conversation), which is predominantly locally constructed since it is impossible to predict the participants' contributions and the main aim is to establish "informal talk"; the focus of interaction is neither on a specific topic or instructional content nor on specific activity rules; the usual rules for conversation are in operation; both teacher and learner can be primary speakers.
IT 2 (telling), where the organization of interaction is predominantly non-local since the teacher has decided in advance that some information should be imparted to the learners; the focus of interaction is on the topic; as for social rules and constraints, they apply in terms of giving appropriate listening responses and asking for clarification, participation is of a limited kind; the teacher is the primary speaker, students can self-select, but no change of topic or allocation of speaker other than the learner is open to them.
IT 3 (elicitation), which is predominantly non-local since the content matter has been selected and sequenced in advance and it has been decided that it must be systematized by a process of asking and answering questions; the focus is both on topic and activity since some information must be collected and presented in a recitation format; social rules and constraints do not apply since it is the teacher who controls participation; he can either ask undirected questions (followed by self-selection or bidding) or specifically allocated ones.
IT 4a (ritual), where interaction is predominantly non-local since such activities as repetition, minimal pair practice or substitution drills are the prevalent ones; the focus is on activity; social rules have been superseded by ritual rules and constraints; participation structure is identical with that in IT 3, but chorus responses are possible.
IT 4b (group work), which can be turned into any of the types enumerated above depending on the kind of task set by the teacher; it differs from all the previous types in that the learner is the primary speaker.
The above typology makes it possible to discern a close relationship between interaction types and participation structures, and on its basis the following patterns of participation in the second language classroom can be pinpointed (van Lier 1988: 174): (6)
PS 1: T/L -- LIT: a conversation to which both the teacher and the learner can contribute by introducing/changing the topic, self-selecting or allocating the next speaker;
PS 2:T -- (L): the teacher has to impart some information while the students to self-select in order to seek clarification, make additions or indicate the level of comprehension;
PS 3: T -- LL/L: the teacher asks undirected questions and the students self-select either by speaking or bidding for turns; this may result in chorus responses or speech overlaps;
PS 4: T -- L: the teacher allocates specific students either to respond to some question or to perform some action (as in a repetition or substitution drill);
PS 5: L -- L/(T) -- learners work in groups trying to solve a task set by the teacher; the teacher may interfere in this kind of work thus changing the character of participation.
According to van Lier (1988: 163), it is IT 4 that constitutes the backbone of the lesson and the remaining types of interaction can only be viewed as some transition stages which are sooner or later bound to be turned into IT 4. As a result, interaction in the second language classroom is predominantly different from general conversation and strips the learners of the chance to negotiate meaning in the target language. It should be noted, however, that teacher-dominated IT 4 can also have some instructional value and, most importantly, it can be turned into group work in which students are capable to interact with each other and thus engage in some meaningful discourse.
(4.) Some comments on the source and nature of the transcribed data
The data analyzed for the purposes of this study were collected over a period of three months in a typical high school environment where the curriculum requires students to attend three English classes a week during the first three years of instruction and four classes a week in the last year. During two of the three (four) classes, the students are divided into groups of more or less sixteen and once a week the teachers are expected to teach the whole class of over 30 students. Due to various, sometimes totally unpredictable, circumstances the number of contact hours is often dramatically reduced, which entails a lack of continuity and is undoubtedly highly detrimental to the acquisition of both speaking and other skills necessary to master a second language.
As for the subjects of the study, the experiment focused on a group of third year students, some of whom had been exposed to more or less extensive instruction in English before they came to high school. In consequence, the class could be described as a mixed ability group in which the linguistic competence of some students is utterly stunning and the incompetence of others makes it virtually impossible for them to actively participate in normal classroom proceedings. This disparity makes it extremely difficult for the teacher to decide on techniques that both of the extremes could benefit from, and sometimes he has to choose the lesser of two evils trying to gear the content of teaching as well as the activities to either the less or more competent learners. (7)
The choice of lessons transcribed for the purposes of this study is deliberate and stems from the desire on the part of the author to capture the most typical aspects of interaction occurring in the course of his classes. This objective is facilitated by the fact that being both the researcher and the subject of study, the author was in a position to sift through the data and ultimately decide what constitutes the most salient features of the classes he conducts and what is merely a deviation from the ordinary. The lessons selected for analysis were primarily devoted to working on different kinds of texts and illustrate the most conspicuous techniques and activities employed by the teacher. It should also be noted that the emphasis was placed predominantly on the lessons during which some meaningful interaction and negotiation of meaning are likely to occur. Therefore, classes whose subject matter did not go beyond grammatical explanation and the application of grammatical rules in totally predictable context s were excluded from the analysis. (8)
5. Interaction in a typical high school class -- how does it differ from general conversation
5.1. Turn-taking and initiative
Predictably enough, the data analyzed for the first stage of this study seem to indicate that it is the teacher who plays a dominant role in classroom proceedings. It is him who knows in advance what the focus of instruction is going to be and who makes sure that the plan he came up with will be implemented in the course of a given lesson. In order to accomplish this goal it is necessary for him to control the turn-taking mechanisms by imposing certain rules on classroom discourse, thus altering the patterns of turn-transition and distribution that normally govern general conversation.
The classes usually start with the teacher announcing what the subject of the given contact hour is going to be and what exactly the students are expected to be doing during the forty-five minute class that follows. Here are two examples of such a state of affairs:
"OK, so today we are going to first of all take a look at the final episode of THE VISITOR as we didn't manage to do it yesterday ... plus we are going to listen to some of the dialogues that you prepared for today ... so first of all, let's take a look at ..." "All right, so those are the different ways of keeping fit ... and now we are going to take a look at the text ... and we are going to do it like this ... I will read this text to you in just a minute and as I am reading, please try to find answers to the following questions ..." (9)
When the lessons are in progress, it is again the teacher who announces that the time has come to move on from one topic or activity to another, tells the students what they are supposed to read or listen to and how much time they are expected to devote to one particular activity. In the majority of cases, it is also him who decides what words or structures require some additional practice, whether or not the answers provided by the students are sufficient and whether any further elucidation of a given pronunciation, semantic or syntactic problem is required. Even if a given class veers off course and the subject is temporarily changed, in most cases it is again the teacher rather than the students who is responsible for such a development as is the case in the following example:
(when one of the students has apparent difficulty using a structure she should already be familiar with) "Wow, there was a test today, right? ... this will be the most important grade that you are going to get this semester, it will be on this test and the test that will be on Monday ... on vocabulary ... the two most important grades that are going to skew the balance ... meaning will be very decisive about your grade ... what is 'unsinkable'? ..."
The initiative on the part of the students is very seldom expressed explicitly and is usually confined to cases in which the student asks for clarification or some additional information regarding the pronunciation or meaning of a given word or phrase. In a number of such cases, students initially use their mother tongue and only after they have been admonished by the teacher do they switch to the target language. Another characteristic of the lessons is that students' contributions which have the potential of bringing about even a momentary deviation from the subject are few and far between and only a handful has been identified in the transcribed data. Here is one of the few examples:
(4) (when discussing different ways of saying 'over the weekend') "T. You do, great; what a good memory you have ... S. I remember a mistake. T. Mistake ... what kind of mistake? S. There was a word "important" ... and I thought it was 'imported' (pronounced like "impotent") T. Impotent?! (laughter) S. Imported. T. Oh, ... imported .. I don't remember because ..."
Despite their apparent unwillingness to take the floor, however, the students do seem to exert a more or less explicit influence on classroom proceedings. First of all, it is the learners' verbal responses signaling to what extent the information that the teacher is trying to impart is being taken in that affect the amount of time and practice that he allots to a given problem. If, for example, there is hardly any response to a question the teacher poses, he tries to follow up on it by asking some additional ones in the hope of getting at least some of the students to come up with the correct answer or, in some cases, any answer at all. The same holds true for some of the structures and vocabulary items that seem to pose serious problems for the learners. It can also be argued that some of the mistakes committed by the students make it possible for the teacher to focus on problems that could have gone unnoticed otherwise and often provide students with both interactionally modified input and extra practice op portunities. The following exchange illustrates this point:
(5) "S. It was sailing across the Atlantic Sea. T. Ocean. (a few minutes later) T. ... by the way, what is the difference between the sea and the ocean? Magda, what do you think? What is the difference between the sea and the ocean? S. ... ah ... what is the ocean? T. Ocean ... like the Atlantic ocean and the Baltic Sea. S. Aha ... a no tak. T. All right ... S. The sea is ... ah ...smaller and ... T. How do you say 'plytsze'? S. Deeper ... ah ... no ... T. More shallow, OK? S. More shallow".
Apart from being in control of the subject matter and the different activities that characterize a given lesson, the teacher also exerts a considerable influence on turn-taking mechanisms. Most of the turns are teacher-initiated and there seems to be an even distribution between direct and general nomination. It should be noted, however, that direct nomination is predominantly used as the last resort after the teacher has received no response to his general nomination and has attempted to paraphrase his questions in a variety of ways. Seemingly, if it were not for personal solicits, some students would never get the chance to participate because they will never volunteer unless called upon. Self-initiated turns also occur, but their incidence is rather low and they mostly perform the function of responses to questions posed by the teacher. As for another-learner initiated turns, they are virtually non-existent during normal classroom proceedings and abound only during role-plays.
In view of all the aspects of turn-taking discussed above, it can be argued that the interaction that occurred during the classes analyzed for this stage of the study can be placed somewhere in the middle of the continuum ranging from relatively unconstrained and totally constrained speech-exchange systems. (10) It is the teacher who, in most cases at least, exercises control over classroom proceedings by selecting the subject matter to be taught, the techniques and activities to be adopted as well as by making the overwhelming majority of decisions concerning turn transition and distribution. On the other hand, however, the rules governing turn-taking mechanisms are not as strict as could be expected, and the teacher tries to encourage rather than stifle initiative on the part of the learners. This is accomplished by employing general nomination whenever possible and rarely rejecting a piece of information or a change of topic volunteered by the students. (11)
5.2. Topic and activity
The classes selected for the purposes of this study are all more topic than activity-oriented for the reasons outlined in the previous section. All of them focus either on a given text and the processing of information it contains or on discussing a given topic. Obviously, activity orientation is indispensable and it can be easily discerned as the students are often required to practice pronunciation by repeating words and phrases, read specific excerpts from texts in a stated order or answer some questions predetermined by the teacher. None of those activities, however, is carried out for its own sake and all of them serve the primary purpose of discussing the subject or text at hand. It should also be noted that very rarely does activity orientation take on the form of tedious repetition and substitution drills, the only exception being practicing pronunciation. As a result, the students have the opportunity to focus on the content of the messages that they are trying to produce rather than on the necessity to adhere to some rules and rituals.
Topic-orientation per Se, however, does not guarantee the existence of a sufficient number of practice opportunities for the students and does not automatically make second language classroom discourse akin to general conversation. Consequently, it is the teacher's responsibility to come up with a combination of topics and activities that would be the most favorable to promoting negotiated interaction. This can be attained through passing some of the responsibility for the subject matter of some of the classes on to the students as this is bound to pique their interest in classroom proceedings and consequently enhance their participation. It might also make sense to get the learners to work on information-gap activities in small groups in the hope that the ensuing interaction will enable them to negotiate for meaning and practice turn-taking mechanisms.
5.3. Interaction types and participation structures
The transcripts of lessons analyzed for the purposes of this paper seem to repudiate van Lier's (1988: 163) supposition that second language classroom discourse tends to be dominated by IT 4a (ritual) and that the remaining types of interaction can only be viewed as some deviations from the ordinary which sooner or later are bound to be turned into IT 4. The transcribed data show that although interaction is predominantly non-local, it is IT 3 and IT 2 that play the most significant role and constitute the backbone of the lessons. As for IT 1, which is characteristic of general conversation, it does not occur very often and some of the slots in which it could occur are taken over by the students' mother tongue. A similar comment could be offered on the incidence of IT 4b (group work) which is relatively rare and can be discerned only during role-plays.
As far as IT 4a (ritual) is concerned, it appears to perform a highly marginal function and contrary to van Lier's claim, it takes up only an insignificant fraction of a given contact hour. In fact, its role is confined mostly to helping students solve their pronunciation problems and therefore the only form it takes on are repetition drills. Although it is usually the teacher who makes the decision as to which items call for some extra practice, in a number of cases those drills are triggered off by the students' blatant inability to pronounce a given word or phrase or their explicit request for assistance. An example of such a case follows:
(6) "T ... the assassination of John Kennedy S. ... was ... T. ... is ... S. Reg ... (seemingly not able to pronounce the word properly) T. Regarded. S. Regard ... T. ... ded ... repeat, 'regarded' S. Regarded ... as one of the most tragical ..."
The function that van Lier ascribes to IT 4 is taken over by IT 2 and IT 3, and it is their constant interplay that constitutes the main body of the lessons. In most cases, before moving on to presenting or discussing a given text, the teacher first focuses the students' attention on the subject matter of the lesson and tries to elicit some information about that subject matter or relate it to students' personal experience. As soon as the text has been introduced, the teacher elicits some information that it contains by either having the students answer some pre-questions, asking some additional questions about the contents of the text or some related aspects, or getting the students to explain the meaning of some words and phrases in English. All these techniques are indicative of IT 3, where the focus is on both topic and activity and where turn-taking mechanisms are controlled by the teacher who asks directed or undirected questions.
Where further elicitation turns out to be unfeasible due to the deficient linguistic competence of the students and IT 3 ceases to be sufficient, it is transformed into IT 2. That happens when the students are incapable of providing a satisfactory answer to a question posed by the teacher; they find it impossible to explain a given word of phrase because they have never encountered it before or because the explanation would require them to draw upon structures or lexical items that are over their heads; or when some comments on the meaning and usage of a new syntactic structure are required. In such cases, the teacher takes the floor and imparts some information to the students who are allowed to give listening responses and seek further clarification, but usually cannot change the topic of discourse. (12) An example follows:
(the teacher is trying to explain the word expense account that has apparently been misunderstood by some students) "... when you work for a company and you are the manager, OK, you have to meet some important people very often and of course you don't meet them at your office, but you meet them at a restaurant, at the hotel, ... and then somebody has to pay for it; so you have a special account out of which you can pay for those things ... to jest fundusz reprezentacyjny ... I would say ... expense account ... so he must have been a pretty important person ..."
As for IT 1, its incidence is rather low. The transcribed data suggest that classroom discourse turns into IT 1 either when the teacher chooses to stray from the subject by slipping in a comment or a mention that encourages the students to get engaged in conversation; or, which is extremely rare, when one of the students volunteers a change of topic. In a situation like that discourse takes on a certain measure of non-locality and the students are in a position to take the initiative by changing the course of interaction and competing for turns. It must be admitted, however, that instances of IT 1 are seldom fully spontaneous and are usually confined to just a few turns.
Another comment that should be made on the occurrence of IT 1 is that some contexts in which it could be expected are dominated by the students' native language. Although such a situation is understandable in cases where the message learners want to convey is too complex and exceeds the students' linguistic competence, they often draw upon their mother tongue to ask such questions as: "Whose handout is that?", "Is it for me?" or "What does it mean?". Beyond a doubt, most of them are capable of expressing those messages in the target language and their unwillingness to do so can be attributed either to mere laziness, lack of motivation or to their perception of the classroom as an artificial setting in which no real-life communication can ever occur.
The analysis also showed a very low frequency of group work (IT 4b) which is particularly useful as it can be turned into any interaction type depending on the format of the task set by the teacher. In fact, the only examples of group work are encountered when students are asked by the teacher to practice some role plays that they were asked to come up with in advance as part of an oral homework assignment. Although such activities are invaluable as they provide students with practice opportunities, students often simply memorize their parts and the whole conversation has nothing to do with negotiated interaction and is, therefore, dissimilar to interaction occurring in naturalistic contexts. Consequently, the tasks teachers devise for pair or group work should guarantee the existence of a genuine exchange of information, as this will supply the students with opportunities to practice turn-taking mechanisms and thus contribute to the development of their communicative competence.
Predictably enough, the prevalent interaction types determine the patterns of participation that are the most conspicuous in the course of the lessons selected for the purposes of this study. Since classroom discourse is dominated by IT 2 and IT 3, the most frequent patterns of participation are those in which the teacher is the primary speaker and the learners either self-select or are called upon to provide answers to teacher's questions. As interaction types change, so do the roles of the participants so that learners are allowed to considerably influence the nature and direction of classroom discourse in IT 1, and they become primary speakers in the course of pair or group-work.
6. How were the patterns of interaction changed to approximate naturalistic settings
The lessons analyzed in the second stage of the study can no longer be described as typical as they were conducted with an eye to determining the feasibility of making classroom discourse more conducive to the development of the speaking skills. In pursuit of that objective, it was necessary for the researcher to change various aspects of second language classroom interaction so that the whole instructional setting could bear more resemblance to general conversation, which is generally considered to foster negotiated interaction and, thus, facilitate the mastering of the communicative skills in the target language. The areas in which the classes examined in this stage of the study differ from those analyzed above can be summarized as follows:
1. They are characterized by a considerable degree of unpredictability, which was accomplished by changing both the subject matter and the character of activities employed in the course of the lessons. The students were no longer confined to working on a given text and were asked to participate in discussions, engage in information gap activities and games. They also got involved in much more pair- and group-work than they do during regular classes. Although all of those activities had been, to a certain extent at least, pre-planned by the teacher, it was the students who came up with their own arguments and ideas and who were allowed to make any contribution to classroom proceedings as long as they were willing to use the target language. In order to make classroom interaction even more similar to general conversation, the teacher renounced the lion's share of the control he usually exerts over turn-taking, thus making it possible for the learners to self-select and compete for turns. At times, however, it was necessary for him to step in to ward off an inevitable breakdown in communication, to prevent some of the students from using their native language or to make them aware of the fact that the extent to which a cla ss had veered off its intended course was no longer acceptable. (13)
2. The number of instances in which the learners mother tongue could be regarded as an acceptable means of conveying messages was drastically curbed. It was made clear to the students prior to each activity that no utterances in their native language would be accepted unless they were compelled to fall back on them in the face of an unavoidable breakdown in communication or when they were indispensable to elucidating some linguistic problem. Although a restriction like that is extremely unlikely to occur in general conversation and can be viewed as yet one more factor making the whole instructional setting artificial, its imposition was necessitated by the students' reluctance to use the target language unless there was no other option. Such a step also seems justified in view of the fact that some of the slots that could naturally be occupied by IT 1 were typically taken over by the students' mother tongue. It seems reasonable to assume that if all of the students obeyed the rule, this constraint would parad oxically work to their advantage and stimulate more negotiated interaction.
3. It was decided to introduce activities which are generally considered to promote negotiated interaction as well as to make students partially responsible for the subject matter of some classes. As has been suggested above, those activities included information gap tasks in which the students were required to elicit some missing facts from their partners, numerous instances of group-work in which they had to express their views on a variety of topics, survival games where the learners were expected to come up with a plausible solution to a knotty problem and debates in which everybody was free to voice arguments for or against a given cause. In contrast to regularly conducted classes, it was the students who suggested topics they would like to discuss and their preferences were taken into account when designing the subject matter of the lessons conducted for this stage of the study.
4. It was attempted to curb the negative influence that the varying linguistic competence of the learners exerts on classroom proceedings. As has already been explained, the unwillingness to participate on the part of the students can be attributed to the fact that they lack the incentive to pitch in knowing that their more competent classmates will gladly dominate classroom interaction. In order to deal with that phenomenon, a number of different solutions were tried out. During debates the teacher made sure that both teams included an even number of the most competent students and set rules in such a way that at least several of the less proficient learners were required to take the floor. As for group work and information gap activities, sometimes the more advanced students were separated from the less advanced ones when conversation groups were set up; at other times mixed ability groups were allowed. All of these steps were aimed at finding an arrangement which would counter the pernicious influence of t he varying linguistic competence of the learners.
7. How did the various facets of interaction change as a result of modifications?
The transcribed data suggest that the changes outlined above did indeed make the classes more similar to general conversation. On the whole, the students exert much more control over classroom proceedings than they normally do and the activities devised by the teacher make it possible for them to compete for turns and expand on virtually any change of topic they volunteer. Another reason why classroom discourse bears resemblance to interaction occurring in naturalistic contexts is the marked prevalence of topic orientation and the fact that such explicit instances of activity orientation as repetition and substitution drills are few and far between. It is also evident that classroom proceedings are, at least when a given activity is in progress, dominated by IT 1, which is characteristic of a relatively unconstrained speech-exchange system where other interaction types perform but an auxiliary function. (14)
It would be highly hazardous to assume, however, that copying the interaction patterns prevailing in general conversation and superimposing them on classroom discourse has only advantages and is always bound to assist students in their efforts to master the target language. In fact, the data analyzed in this stage of the study indicate that excessive reliance on IT 1 and mere efforts to approximate the patterns if interaction in the second language classroom to those governing general conversation may produce an adverse effect and be at times even less effective than more constrained and thus seemingly less advantageous instructional settings in which IT 1 plays a peripheral role.
The analysis of the transcribed data shows that it is almost exclusively the more proficient students who dominate interaction in debates and in the course of group work, whereas the less competent students have noticeably fewer opportunities to participate than during the classes in which IT 2 and IT 3 prevailed. Such a state of affairs can be attributed to the fact that the lacking linguistic competence of the less proficient students precludes them from effectively competing for turns and the teacher is no longer there to employ direct nomination in order to make sure that even the least competent of the learners will get a chance to speak once in a while. The ramifications of such a situation are appalling as the intrinsic motivation of the less proficient students drops below acceptable levels and they simply