Conversation analysis and the ‘no gap, no overlap’ floor

SSJ 1974 provide some of the key formulations of conversation analysis. They start with the notion of an economy in social activities, in which a turn is something of value. This allows the development of the metaphor of possession, which is understood not as the right to speak but as actually speaking unopposed by others present.

From their data, SSJ propose that “It has become obvious that, overwhelmingly, one party talks at a time” (1974:669). A single speaker has the exclusive space to speak for a single turn, for whatever unit-type that turn comprises (as Judge Morland says to the supermodel Naomi Campbell in the opening quote). They acknowledge that more than one person may be speaking at any one moment: Such occurrences are “common, but brief” (1974:700). The rule set for constructing turns, among other things, provides for transfer of speakership “so as to minimise gap and overlap” (1974:704; emphasis added), not necessarily to eliminate it.

As a system for conversation, the model of turn-taking that SSJ provide deals with allocating who speaks, but locally, as the talk unfolds, at each Transition Relevance Place (TRP). Importantly for the analysis of the classroom data that we carry out, the turn-taking system can accommodate any number of parties, as well as parties joining and leaving an interaction. SSJ note that an increase in the number of potential speakers leads to turn size tending to be smaller. In addition, once there are at least four speakers, it is systematically possible for “schism” (Egbert 1997) to occur, such that the parties to one conversation divide and two conversations run concurrently. SSJ's account also allows for talk to lapse: It does not have to be continuous, a phenomenon which we refer to below as an “incipient” state of talk. It allows for pairs or groups of co-present parties to be selected as “next speaker.” Both these phenomena feature heavily in classroom talk. However, the focus on the turn as a singly occupied unit of value does not, in our view, allow us to account satisfactorily for what is going on in our data within the course of a lesson or class activity.

The collaborative floor

In her 1981 article “Who's got the floor?” Edelsky was possibly the first to turn the spotlight on the term “floor” and not use it in a taken-for-granted, colloquial way. Her notion of what she calls a “collaboratively developed floor” (abbreviated F2), which is characterized by more than one person speaking at a time, contrasts with the one-person-at-a-time floor (F1) proposed by SSJ. This distinction arose from Edelsky's not being able to work out who “had the floor” at any one moment in her data corpus of administrative meetings in a university department. However, it is not free from problems. Edelsky refers to the length of the stretches of talk she could not assign to one party as the floor holder, which last from 1.8 to 45.9 seconds (1981:391). This implies that there are some very fast shifts of floor. Two questions arise from this: Are such shifts real for parties to the talk? And what is consequential about such an apparent shift for the activity in hand? Furthermore, Edelsky left some of her data out of consideration, since the F1/F2 distinction did not capture all the talk that she had recorded. In trying to isolate F2s using objective criteria, “[w]hat remained were either singly produced floors (F1's) or a very small number of uncategorised episodes, which will be eliminated from further discussion” (p. 409). This is an issue we will return to in the analysis of our classroom data when we attempt to account for talk that might otherwise be so discarded.

Development of the collaborative floor

Coates's view of the “conversational floor” is summarized in a recent article on whether Deaf signers orient to a one-at-a-time turn-taking model or to a “more collaborative model” (Coates & Sutton-Spence 2001:507). Building on Edelsky's claim that there are two modes of organization for conversation, and that males and females seem to show different preferences – the former for one-at-a-time, the latter for “all-in-together” – Coates defines a single floor as one where “one speaker speaks at a time,” and a collaborative floor as one where “the floor is potentially open to all participants simultaneously” (p. 511; emphasis added). In an earlier discussion of women's talk, she invokes the musical analogy: “Speakers contribute simultaneously to the same theme, like several instruments playing contrapuntally” (1989:111). She makes a strong claim for this as a key feature of women's talk: Their voices “meld” to produce a collective voice. What she calls the “construction of talk” is shared by female friends “in the strong sense that they don't function as individual speakers” (1997:117; emphasis in original). This implies that the turn can be shared by more than one party to the talk, something we will discuss in relation to our classroom data.

Other relevant observations on the classroom floor

Susan Philips 1972 identified four “participant structures” as constituting a framework for teacher-controlled interaction. These structures correspond to the sense in which we want to define a floor as a set of possible rights to speak for a certain activity. For example, in one such participant structure, the teacher may address all the pupils, or one in the presence (and with the assumed attention) of the rest; for their part, a single pupil may address the teacher, and some or all pupils can respond in unison.

Hugh Mehan 1982 uses the term “floor” to mean, in effect, speaking. He talks of pupils “getting” and “holding” the floor, as if occupying some physical (as well as temporal) space in order to be able to talk while others remain silent. The pupil has to locate an appropriate juncture to “get the floor,” which Mehan argues is not after every turn, but after every “Initiation-Reply-Evaluation” sequence. So, to “hold the floor” involves having what you say attended to by others, and we are concerned to build the notion of listenership into our account.

Coates 1996 argues that the collaborative floor is “radically,” “qualitatively,” and “quantitatively” different from one-at-a-time turn-taking. In the classroom data we find many instances of the type of features that she describes as “collaborative,” yet it is clear that classroom talk is in many ways an asymmetrical speech exchange system, very different from the talk that occurs among the women's friendship group that Coates describes. Coates & Sutton-Spence 2001 contrast the asymmetry and formality of the classroom with norms of friendly conversation, where informality and symmetry are key. However, we believe that the classroom data we have cover a wide range of activities in terms of levels of formality and participation. This enables us to move the analytic focus away from the binary oppositions between asymmetry and symmetry, F1 and F2, collaborative and one-at-a-time floors.

A constrained floor: Doing the register

Taking the register and doing the lunch administration are very constrained activities in class, short and clearly bounded. Here we look at just three features: (i) the basic talk organization of the activity; (ii) how even within such a restricted activity, “other things” can happen, and how they fit in; and (iii) when the existence of some subsidiary floor is allowed or not. In this way, we hope to start to show that it might be necessary to distinguish some “official” floor that ties in with the activity, what that floor can bear or tolerate, and what will attract censure.

The basic organization of a tight floor

The activity, once under way, comprises a series of two-part sequences in which the teacher reads a pupil's name from an ordered list. If present, the named pupil replies yma (Welsh, ‘here’). Although any utterance would be evidence of their presence, this is the only one they do produce, and no other pupil replies for the nominated pupil. The basic two-part exchange, for a pupil who is present and attending to the activity, is thus simply:

In formal terms, there is potentially one such sequence for every name on the list of pupils. The sequence consists of two turns. In the slot for the first turn, the teacher is constrained to produce just one name from the list, in order. The production of this first part constrains the second part to a specific speaker, if present, producing one specific lexical object, the Welsh word yma, in the slot opened for him or her. The pupil has to produce this immediately. The floor organization is that of teacher and the first pupil on the list, for one adjacency pair, then teacher and the next pupil, and so on. The rest of the class are to attend and stay silent.

Changing floors when the activity is interrupted

In example (2), taking the register is interrupted by a pupil from another class. The interruption is bounded by the sequence of a knock on the door and dewch i mewn ‘come in’ (lines 9–10) at the start, and the sequence thank you and pupil leaving (lines 42–43) at the end. After a pause, the teacher resumes the activity from the point where she left off (line 44), without any repeat of a turn or sequence, and without naming the activity. Upon the teacher's saying Ian, Ian says yma, showing their mutual orientation to the resumption of taking the register. The activity in progress thus resumes to completion, once the insertion sequence is dealt with.

If we examine this sequence in terms of the floor, there are several shifts. The activity is under way in its routine form (lines 6–7), as described above. At the knock on the door, the teacher and the pupil from another class (“Boy”) establish a new floor between themselves (line 10) to deal with the interruption, while the class remain silent. The taking of the register is suspended while this interruption is dealt with. Out of this one-to-one floor, in order to complete the inserted sequence (which is in effect, another activity) the teacher turns to address the whole class (line 20), a floor which allows any individual pupil (“Pupil”) to reply to her question. When this insertion sequence is complete, the teacher (in line 42) resumes the one-to-one floor with the boy from the other class again, to complete the “interruptive” activity. The register-activity floor pattern then resumes as before. The danger of an analysis that focuses solely on the floor shift is that it does not recognize that a different activity has come into play. So we want to claim that it is precisely this change of activity that requires a different floor organization, and that this organization is accomplished locally and flexibly without any party's drawing attention to it.

When you can whisper and when you cannot

In (3) and (4), a stretch of talk from a lunch administration activity, we get insight into the question of whether other floors can coexist with an activity's main floor. The lunch administration process follows the same sort of pattern as doing the register, with a succession of one-to-one floors between the teacher and each pupil. Those pupils who have yet to be called have to pay some attention to ensure that they respond promptly; those toward the end of the list can relax their attention somewhat in the early part of the activity, and those who have had their names called out and have responded no longer need to attend until the whole activity is over.

Here, the teacher has started the activity (line 15) but has to call for quiet before continuing, in response to one pupil's overlapping Celine's utterance (but not obliterating it; the teacher does not ask for a repeat of the utterance, and Celine does not offer one). Thus, to get the activity under way, the other pupils have to be silent, or rather, to maintain a certain level of near-silence that allows the activity to proceed unhindered. Despite the teacher's call for quiet, after a brief pause whispering by some pupils is audible (from line 20 on), and this continues for several iterations of the “teacher – pupil on list” adjacency pair.

The whispering clearly does not prevent the progress of the lunch administration activity until the point at which it has become something else. The teacher calls for a stop to the chatting, an illegitimate category of action by the rest of the class, that is seemingly incompatible with the focus of the main activity in hand. The teacher follows this demand with an account for her request for silence (line 37):

Her call for silence occurs just as Tariq is about to deliver his second-pair part, which he then does, but in overlap with the end of her turn. Without prompting, he repeats cinio once the teacher has got the silence she asked for, and she acknowledges his orientation to the resumption of the activity.

Shortly afterward, there is an example (5) in which a boy does not respond immediately. What is interesting is that Reuben's failure to attend to the routine and respond promptly (given that the order of names read out each day is the same) is accounted for by the teacher on the basis that he was chatting, invoking the same illegitimate category as above. On the recording, though, there is no audible evidence that Reuben, or any other pupil, was even whispering:

These extracts show how a subsidiary floor may develop, even in a very constrained activity. To summarize: For the activity of lunch administration, pupils should not talk unless nominated to do so, and they should absolutely not “chat,” but they may “whisper”; chatting impedes the progress of the activity, whispering does not. Thus, there are:

(i) a legitimate floor for the activity that gets it done, the succession of teacher to single pupil adjacency pairs;

(ii) an illegitimate floor that attracts censure, in that, whatever the volume and level of involvement by pupils, if it impinges on the main floor, it will be characterized as “chatting”; and

(iii) an allowable, though not officially ratified, subsidiary floor (or floors), where pupils may whisper, so that the main floor can progress without problems of audibility or attention.

A “loose” floor: Out in a South London street

If we are to consider taking the register as an example of how tight and constrained the organization of talk can be, then this next example might well represent the other end of a continuum. One recording from the South London corpus seems to us to contain the loosest possible constraints on talk within a single activity that we have come across in our data. In this particular recording, eight pupils from the class, supervised by the researcher (JT) rather than a teacher, are walking along some of the streets outside the school. The activity they are engaged in involves assessing aspects of the environment and grading them, such as the state of the pavement, grass verges, presence of litter, and so on. There is thus a definite goal to the activity to be achieved, but a fluid grouping of parties moving along the street. The setting changes as they move. We examine four aspects of this activity that bear on the organization of talk.

There is what we call an “incipient” state of talk throughout, in that any party can start a turn or sequence when talk lapses. There is a fluid mix of groupings of speakers, and multiple floors are evident throughout, but there rarely is complete silence.

At the outset of the recording, as the activity is about to get under way but they have not yet started the specific tasks, we hear a wide variety of talk: casual (lines 7, 20–22), organizational interspersed with comments (lines 9–15), playful (line 24), and to do with the tape recorder itself (lines 25–32). Note that none of this is problematic, as the parties are making their way to the point in the setting where they can start the tasks that comprise the official activity itself.

This is a task-based activity in which pupils are rating items, entering numbers on sheets, and totaling them. Most of the talk on the recording reveals the pupils to be getting on with the task, but occasionally a subgroup veers off task or undertakes some other activity. In (7), some boys start singing (line 641), and when called to order by the adult present (line 651), one boy provides the justification that they have done it. The implication is that, having finished their part of the official activity, they are now free to do something else; being outdoors allows them to do something that they would not get away with in class. The adult refocuses their attention to the official activity: Whatever it is they have done needs to be checked before their part in the activity is officially over. This is done flexibly at that moment (lines 651–654), not by prior decree, and it resolves the way the boys' activity has shifted off task:

In the next example, the pupils are still engaged in this institutional activity, derived from the school curriculum, but the “setting” is out in public, which offers opportunities for a wider interaction with social space than is found in the classroom. Here a girl notices an advertisement in a shop window, and the talk among these pupils then takes an off-task trajectory, as discussed above:

The activity itself involves moving around in the streets outside the school, observing certain features of the scene around them. The potential for this off-task interaction is thus inherent in what they are doing. However, the pupils in general maintain a focus on the task throughout the recording: Excerpt (8) comes in a transitional phase as they are heading back to the school to collate their figures. There is this brief period when the activity is suspended while a change of setting is effected.

The role of moderator in the talk, routinely taken by the teacher in whole-class activities, differs here. The adult in authority is not their teacher, and the pupils refer to the researcher variously as Miss or Joanna, a use of first name not usually allowed with teachers.

However, at various points the researcher does make calls to order, to stop the multiplicity of floors when they involve off-task talk, and to ensure that the activity is getting done:

Order is thus maintained by the researcher through occasional checks on progress and temporary restorations of a single floor for the whole group. However, pupils also feel able to place limits on the behavior of other pupils as they go about the task. In (12), a girl twice makes comments to boys, reprimands that are then given stronger force by the researcher's taking them up. Her comments relate first to general behavior, and second to talking at all. The open nature of the setting and the freer nature of the activity, working in pairs and individually, means that the moderator of the activity cannot be present in every floor, even when she can hear talk in other floors.

Thus, any of the multiple floors through which this activity gets done is self-regulating.

Incipient floor

Many activities in class involve a general silence for individual work, or for reading. It turns out, though, that quite a lot of talk does in fact occur in these sessions. One feature is that there is a predominant activity for the class, while the teacher may be carrying out some other activity, such as marking work or general administration. There is the possibility of interruptions from outside the class, and from events inside the class requiring action. What we call here an “incipient floor” involves a floor with extended silences (where that is a base norm for the activities in hand), but potentially open throughout for certain actions that are not predictable at the outset.

In a few minutes from a single classroom session, we can observe some examples of how talk is occasioned. Silent reading is in progress for most of the class. Ian realizes that his nose is bleeding and is told by the teacher to go and get a tissue: an unpredictable event arising in the class, and not related to the activity. Sheila is typing up some material on the computer for the teacher in preparation for the class assembly presentation. She has a problem with the computer font size and refers it to the teacher, who gets Guillermo to stop his reading and help her. The teacher then asks Masoon about a letter of absence, as she works on some administration; a moment later, she asks Anisha to go to the school secretary with a message. Anisha thus has to break off from her work. We can observe here the mixture of events that provoke the initiation of some talk, and that they can be initiated by pupils as well as by the teacher.

Let us look in a little more detail at some examples of “incipiency,” because if there is to be a binary contrast here, it would be with some notion of a “continuous” floor. First, a pupil makes a request during a session of individual silent reading. The teacher agrees to it: The request is on task, reasonable, and properly delivered. Her agreement then comes with an instruction. The floor is a straightforward dyadic one, for the two parties alone, and the only overlap comes where the pupil is concluding the sequence Request-Acquiesce-Thank, when the teacher is about to extend the Acquiesce turn with a limitation on the permission being given. Given what is going on, first in the session as a whole and second in this particular sequence, it makes little sense to talk in terms of this being F1 rather than F2. Instead, it is a matter of how such a sequence is unproblematic within broadly “silent” concurrent activities.

In the next example, we see how the teacher's actions during such silent individual work initiate some talk. Pupils are liable to be interrupted in their work both individually (as here, where the teacher wants to sort out something about the assembly, or where there is an errand to run), and as a class, with time reminders. Repeating instructions for a task is another thing the teacher will do at occasional points during an activity when silent work is going on.

After a long silence, the teacher's single-person activity (planning the assembly) requires input from a pupil, whose own activity has to be interrupted for the two of them to form a dyadic conversational floor. Their talk is, however, available to be overheard by other pupils, although initially no other pupil is a ratified listener. Note, though, that Molly joins in (line 53) without in any way being nominated to do so, but presumably because her contribution is timely and relevant, since she is able to give the teacher the information she wants when Celine has sought a clarification. The two girls are then able to produce a final answer together, although Molly allows Celine, who was the primary addressee, to start to answer the teacher (line 58). Molly joins in after the hesitation, by which time Celine has recovered. Silence then resumes. The sequence amply deserves the characterization of “collaborative” between the three parties to the activity, and they each make contributions that “get the job done” together. Again, to invoke an F1/F2 distinction, and to see lines 58–59 as a shift of floor, does not seem to us to add to the analysis. Rather, Molly's contribution to the talk here can be better explained by the way it gets that current activity done.

A third way that talk starts up in an incipient floor is through interruptions. These examples are taken from a “Numeracy Hour” during which pupils are working individually on problems. Such interruptions (understood as not connected to the activity in hand) can originate within the class as well as from outside. The class has been quiet, when Ian notices something on the floor that it is legitimate for him to raise with the teacher – it is not his pen that is on the floor.

The second interruption is from outside the class, of the sort that can happen at any point during any activity. Various kinds of talk are going on here simultaneously. There has been a state of incipient talk revolving around the teacher (of the sort discussed above, with queries, helping individual pupils, general reminders); there is also whispering among pairs of pupils (allowed for those doing estimates), and open talking by a group at one table with a teaching assistant. When another teaching assistant knocks on the door with a request, the teacher has to establish a dyadic floor with her, suspending what she was doing before, but the rest of the class continues with its activities.

Multiple floors

As we have just seen, there can be multiple activities going on in a single setting, each with its own floor organization, and with some fluidity between conversational groups. For instance, in the “Literacy Hour” example (17), pupils have to complete a worksheet and then do silent reading for the rest of the session, but each does so at his or her own pace. The teacher carries on other activities while the class works, attending to administrative matters and marking work completed earlier, as well as monitoring progress on the current task to ensure that everyone does at least finish the worksheet. The teacher here checks (in Welsh) whether Anisha is reading – that she has finished the primary task, the worksheet – before sending her on an errand:

In the next segment, also from a Literacy Hour, pupils are working individually, helped by the teacher and a teaching assistant. There are questions from individual pupils to the teachers as they do their corrections. The teacher is taking a query from Ian at the same time as the classroom assistant is dealing with a similar query from another boy:

An analytic problem here can be discerning from the tape alone which utterances are in which floor, but we can start to sketch the complexity of the issue. Twenty-five or so pupils on this occasion are working individually, and there are two teachers helping, all in the single setting of the classroom. In one sense there is just one activity, filling out worksheets. The assistant and all the pupils stay in their places, while the teacher moves around the setting. There is thus the systematic possibility of at least two concurrent floors, relating to the same activity, and – as in (18) – sometimes the same topic within that activity. While the floor around the teaching assistant can be considered “stable” because it will most likely include only some combination of those pupils at her table, there will be several shifts of floor as the teacher moves from one group of pupils to another.

In another Literacy Hour, there are a large number of conversational floors in class as the pupils carry out a mixture of individual and group tasks. Provided the parties whisper, this is “ratified” legitimate talk. In this activity, the floors are not fixed and necessarily continuous, but are “incipient” and allow for others to join and leave. Here, Christian and Ben have an exchange within their floor that the teacher overhears and joins to establish the way the task should be carried out:

Shifts of address

Many classroom activities can be said to have a “whole-class floor.” This is characterized by the teacher's addressing the class as one collective listener. She acts as moderator of the talk, in which pupils have to be selected to speak, and then when they do, they must speak to and through the teacher (cf. Philips 1972). Sometimes the teacher addresses just one pupil, but as an individual, not as a representative of the class, as in the case of a reprimand.

These shifts of address are generally brief and do not disrupt the whole-class floor in that they have no consequences for the activity; indeed, the reprimand is tied to the activity in hand because it is directly concerned with what the pupil needs to be doing to engage in the activity. The other pupils remain silent, recognizing that the constraints of the whole-class floor are not suspended for such shifts. In (20), the teacher first addresses Sean alone, but publicly, in that the class can overhear; she then addresses Anisha alone, but privately, as an aside when Anisha has returned from an errand, before nominating Jonathon to speak. Jonathon has raised his hand, but this action is also available to any other pupil. These shifts of address are accomplished without any confusion as to what is going on or any change in activity.

Keeping a whole-class floor going

Tightening up a floor

If example (20) represents a smooth and orderly series of shifts of address to accomplish different things, then (21) relates to the situation in which the teacher reestablishes not so much a single focus to the talk, but a tighter arrangement of talk when a loosening of constraints has occurred. However, this looser talk has not developed out of nothing: The teacher is guiding the pupils through the writing of a story summary and has turned to write on the board; the pupils are contributing to finding the right form of words, with overlapping voices. Although this has been done in a whole-class floor, that has not been of the “tight” form observed in the previous example, but in a much looser form which arises not from some indiscipline on the part of the pupils but from the exigencies of the activity at these moments. The flow of suggestions represents a way that the group can reach a consensus. What happens could be characterized as involving two different floors, and a forced move from a collaborative “overlapping floor” to one-at-a-time. However, close listening reveals that not all pupils are calling out, and that the overlaps are quickly resolved: Self-selection in a multi-party floor leads to just this situation. A better way to capture the feel of what is happening here is to see that, at this point in the activity, such pupil self-selecting and “looser” talk have been entirely appropriate and have moved the activity forward; but there is a flexible, teacher-decided moment where it is too loose to continue the activity.

The teacher calls out a forceful halt to the “open contributions” feel of the session. In lines 65–66 and 69, pupils respond with one thread of an answer, while in line 68 Ian picks up on the need to provide a complete sentence that takes up the teacher's mention of waking up. The problem for pupils is to know when something is appropriate and when it is not: As the teacher allows “calling out” to take place, they continue in that vein, taking their cue from the successive actions that are allowed. As soon as she recommences the previous pattern of writing on the board and seeking suggestions, the pupils do exactly what she has just told them not to. The problem for the parties is not calling out as such, or overlapping talk of itself, but the point when “looseness” becomes too loose.

Moving from multiple floors to a single floor

In the next activity in the Literacy Hour (22), pupils are working individually or in pairs at their tables. The teacher is at what is known as the “top table,” engaged in some one-to-one talk with Emily. A comment prompts the teacher to issue a reminder about paragraphs, and she opens up the whole-class floor by calling Year Six. This organization of the floor is established immediately: The pupils recognize the call and listen. The teacher can now carry out one of the standard features of such a floor organization in a pedagogic activity – she can issue a “known answer” question to the whole class, then nominate a single pupil to answer, producing, as here, the classic sequence Initiate (Question-Bid to Reply-Nominate)-Reply-Evaluate (lines 108–112):

Restricted floor in setting

We have already seen one instance of this kind of floor in the section on incipient floor. The phenomenon here is of a conversational floor obtaining between some parties in a particular physical setting, where other parties who are not specifically ratified listeners for that floor have overheard what is being talked about and join in. Here, pupils are working individually or in pairs and should be writing their summaries silently. One girl, Molly, comes up to the teacher at the front of the class to ask a question. The floor that exists there is now a dyadic, teacher-pupil floor. In line 23, Sheila, from the back of the class, claims to know and then offers an answer to what is being discussed in this floor. Note that she was not only not a part of this teacher/pupil dyad, but was also physically distant from it:

Sheila shows that she is paying attention across the class (begging the question of whether she should have been), but the teacher accepts this intervention without calling for an account of any sort from Sheila. It contributes to the talk relevantly and is on-task. The most important factor is not that this was a floor restricted to the teacher, Molly, and no other, but that Sheila moved the activity along. It is thus the real social actions of parties that are key, not the formal mechanics of the arrangement of talk: The latter are the ways in which the former get done.

Chorusing

Chorusing is talk in which parties produce the same utterance at the same time, and deliberately so (cf. Lerner 1993). It is an extreme form of simultaneous talk in which identical lexical items are produced. In the following example, two pupils join in with the teacher's utterance in line 153 to produce “low-grade” chorusing:

There is a projectability to discursive actions, which for certain types of actions enables parties to produce the chorus effect. In (24), the pupils seem to be co-constructing the projected focus of the teacher's reading of a diary entry intended to explore the meaning of the phrase to get my goat.

The next two examples are of full-fledged chorusing. One pupil starts to join in, then several others come in with the teacher, getting the intonation and number of blahs exactly right. This is as highly collaborative and as simultaneous as talk can get, but analytically it is unclear to us what would be gained by trying to see it as a shift to an F2.

Through such chorusing sequences, the pupils are producing simultaneous talk that displays close alignment with what the teacher is doing; by joining in, they can demonstrate their active participation rather than passive listenership.

Utterance completions

Last, we turn to utterance completion, a phenomenon that forms a key part of Coates's argument for a collaborative floor. Generally, some hesitation by the speaker is taken by some other as a sign of a search for a word or phrase, and that other party offers a candidate completion. It is then up to the first speaker to accept or amend that offer. What is most noticeable in the first two examples (27–28, representative of many others in our corpora) is that the other's action in offering such a candidate completion is not taken as intrusive or even commented upon, but is, as it were, absorbed by the first speaker. It is, however, still for the first speaker to decide on its adequacy and thus determine the final outcome of the utterance.

In the first example, the teacher accepts both the form and the content of a completion offered by a pupil. In the second, when the teacher hesitates or pauses, a pupil offers a candidate completion in the form of a single word, but then the teacher's own completion differs; it is, however, in the same semantic area. Note how the boy then goes along with her formulation because she has used a tag question.

In (29), the teacher has given a complete utterance with an embedded tag question that is picked up on by a couple of boys, who give a minimal response. As soon as she seeks to extend the utterance with a conjunction, and then pauses, Nathan gives a perfectly plausible completion that, although she does not take it up, shows he has tracked her utterance:

In all these completions, we can see exactly what Sacks (e.g. 1995:93) means by the “mind reading” that goes on in interaction.

These utterance completions contrast with what are recognizably invited completions, in which the teacher is looking for pupils to come up with either a “known” answer, or in the case of a mathematics lesson, one that can be calculated:

There is another way in which utterance completions can be used as a resource. In (30), a pupil's suggestion after a teacher's pause allows for a cooperative construction of the scope of the task. The pupil has come up with a reasonable number that is available to the teacher to adopt. Thus, the pupil has used the opportunity for an utterance completion to initiate a possible negotiation of just how many questions they are going to have to write: