Positive evaluation of student answers in classroom instruction

Within the context of teacher/whole-class instruction sequences, researchers have associated teacher evaluation of pupils’ answers to forms of traditional pedagogic discourse, also referred to as ‘triadic dialogue’, ‘monologic discourse’, ‘recitation’ and ‘Initiation-Response-Evaluation (IRE) sequences’. Teacher evaluation has also been associated with pupils’ low levels of participation. Explanations and solutions offered by prior research are mainly based on functional categories of actions, characterizing forms and functions of teacher questions and follow-up moves in IRE sequences. Using Conversation Analysis to investigate collections of positive evaluations in video-recorded lessons in two primary school classes, we propose an interactional explanation of the phenomenon and of its predominant use. We show that teachers systematically select the formats of their positive third-turn receipts not only to evaluate pupils’ answers for their abstract truth value, but also with respect to the role of each question–answer in the whole activity. We demonstrate that, in this way, teachers convey judgements about the question within the activity; thus, adding a constitutive property to the pedagogic practice and providing students with interpretive resources for a common understanding of pedagogic goals and procedures.


Introduction
Teacher evaluation of pupils' answers is one of the most complex, important and controversial issues in research on classroom interaction. On one hand, teachers' third-turn evaluations or 'follow-up moves' (Sinclair and Coulthard 1975) are widely recognized as being at the heart of 'instruction' in Western cultures: adults ask children questions to which they know the answers, then evaluate the correctness, accuracy or appropriateness of children's answers (Sinclair and Coulthard 1975;Mehan 1979;Drew 1981;Brice Heath 1983;Pine 1992;Grosse and Tomasello 2012). However, whilst teacher evaluation of students' answers seem central to instruction, teachers' evaluating work is also sometimes regarded as the means or mechanism for retaining control over matters of 'correctness' (Drew 1981, 260-261). Teacher evaluation has attracted criticism as the hallmark of conventional pedagogic discourse (also referred to as 'triadic dialogue' and 'recitation'), responsible for engendering pupils' low level of communicative engagement (Mercer 1995;Nassaji and Wells 2000;Nystrand et al. 2003;Alexander 2008;Mercer et al. 2009;Lyle 2008). Teacher evaluation is by far the most widely recognized feature of traditional classroom interaction known as the initiation-response-evaluation (IRE) model (Sinclair and Coulthard 1975;McHoul 1978;Mehan 1979), also associated with other properties of this type of instruction talk, such as highly codified test questions, the teacher's role of 'primary knower' (Berry 1981;Tsui 1989;Nassaji and Wells 2000), and pupils' short and unelaborated answers. These associations have encouraged scholars to search for ways to enhance pupils' level of engagement in pedagogic activities through more dialogic strategies, relying on greater use of 'genuine' questions combined with other forms of third-turn receipts such as comments, requests for further information (Nassaji and Wells 2000, 385) and uptakes (Nystrand et al. 2003, 145-147), 1 all of which might contribute to higher levels of pupils' participation (Nystrand et al. 2003, 147-148).
However, despite recent measures in education policies encouraging teachers to adopt more innovative and dialogic discourse styles, it appears that the IRE format and thirdturn evaluations are still predominantly used in pedagogic activities, especially when teachers are addressing and engaging the whole class (Nassaji and Wells 2000;Mercer et al. 2009;Lyle 2008). It is the very absence of the teacher evaluation of students' answers in this first example that highlights what an important part evaluations play; in extract 1, the teacher's non-response to the answer water (line 4) causes students to revise the answer (line 5), seeking the 'correct' answer that the teacher so evidently is looking for (see her repeat of the question in line 7).
(1) Water (PM:ET:2: natural sciences) [From a third-year class in an Italian primary school. T is asking each student in turn to name one component of soil] When the teacher fails to ratify the answer first offered, but instead re-asks the question in line 7 (now in the passive form), more pupils adopt the alternative answer 'clay' (e.g. line 9 is produced collectively); they thereby display that they have understood the absence of a direct ratification of 'water' (offered in line 4) as indicating that it was incorrect. Hence, the teacher's third-turn evaluation is not an optional expansion of the question-answer (Q-A) pair in instruction sequences; participants themselves orient to evaluation as an intrinsic part of instruction activity.
However, it is worth emphasizing that by 'evaluation' we mean both (1) the action of evaluating students' answers, conveying (even implicitly as in extract 1) a judgement about the correctness or otherwise of that answer and (2) the production of an evaluationdedicated third turn to register, receipt and evaluate the answer as positive or negative, again either directly or indirectly. Our focus here is specifically evaluating students' answers positively. Based on observations in an Italian primary school, we show that teachers use a restricted range of practices or techniques to indicate a positive evaluation of students' prior answers. Moreover, the selection from among this range of practices is not random; the evaluative technique or practice is closely connected to the specific type of pedagogic sequence of which it is a constitutive part or element. Therefore, our principal findings are: there is a range of five principal practices or techniques by which teachers evaluate an answer to their question as 'correct' and these practices are employed differentially, according to the specific kind of pedagogic sequence the teacher is building.
In short, we report the systematic associations between the various (positive) evaluation formats and the broader sequences of which they are a part, and particularly between the evaluation formats and the organizational features that distinguish different pedagogic activities, realized through teachers' prior questions.
Without in any way underestimating the extensive relevant (Lemke 1990;Nystrand and Gamoran 1991;Mercer 1995;Nassaji and Wells 2000;Nystrand et al. 2003;Alexander 2008;Mercer et al. 2009), it is clear that the explanations offered so far of the IRE model and of the teacher evaluative turns are based largely on functional categories, leaving unexplored how evaluations are generated through the organization of interaction itself. Although previous studies have provided generalized descriptions of the pedagogic functions of teacher evaluation, there is not yet a comprehensive account of how positive evaluations are constructed, and indeed what makes them positive and recognizably positive (i.e. to students). In line with some recent conversation analytic work on teacher-led classroom discussions (Lee 2007(Lee , 2008Hellermann 2003Hellermann , 2005, our aim is to take this a stage further, giving close attention to the local contingencies of interaction in instruction sequences. Using Conversation Analysis as our methodological perspective, first we identify and explore the practices most frequently used by teachers to evaluate student answers positively. We then consider further the differential distribution of these positive evaluation formats as they occur in the larger instruction sequences, observing some systematic relationships. The analysis here shows that positive evaluation is a constitutive property of instruction activities, providing pupils with interpretive resources of the ongoing activity and, thus, enabling them to participate effectively in the interaction.

Data and methods
The data are video recordings made in an Italian primary school, of naturally occurring scheduled classes in a variety of subjects, in which teachers adopt a conventional 'chalk and talk' method, especially during the instruction phases of the lessons. They normally face the children, who are seated in parallel rows, and address the whole class or single students publicly. The corpus consists of approximately 80 hours and about 50 class periods in two third-year classes (ages 7-8). Two cameras were used in each classroom where the two classes took most of the lessons, one camera filming the teacher and the other the students. Participants included 50 pupils and their four teachers. Subject consent was obtained from all participants according to the Italian law n. 675/1996 'Tutela delle persone e di altri soggetti rispetto al trattamento dei dati personali', which provided safeguards with regard to the treatment of personal data at the time when recordings were made (1999)(2000). Participants' names and any references to participants' private life have been anonymized. The recordings were transcribed according to the conventions widely adopted in Conversation Analysis (a glossary of these conventions is provided in the appendix), and translated into English.
Our analytical approach is Conversation Analysis (Schegloff 2007;Heritage and Clayman 2010;Sidnell and Stivers 2012), an empirical and inductive method relying on the systematic analysis of audio tapes and videotapes of naturally occurring interactions to identify and describe the practices which participants methodically use and to which they orient in their own social behaviour, in producing and understanding their own and others' conduct in everyday social interactions. In this study, we focus on excerpts during the early stages (first 30 minutes) of 10 lessons, in which extended teacher-led discussion takes placeoften going over key aspects of previous lessons or reminding students of what has previously been covered. We identified 145 instances of questions asked by teachers in this phase of the class, the answers to which were positively evaluated by teachers.

Positive evaluation practices
In a sample of 200 teacher evaluations produced by four teachers in seven classes, 72.5% of students' answers were evaluated positively, and only 10% were evaluated negatively; in 17.5% of the cases identified, teachers treated the answer as not quite correct (inaccurate, inappropriate or incomplete). These percentages held across teachers and topics, suggesting that positive evaluations play a uniformly central role in formal primary instruction. In the 145 instances in our corpus in which teachers positively evaluate students' prior answers, we identified five principal construction formats, 3 namely: explicit positive assessments, verbatim full repetitions of students' answers, embedded repetitions, or repetitions accompanied by other turn components, formulaic receipt particles, direct transition to the next Q-A sequence.
In our analysis we look at the action that has been crystallized as referred to with the blanket term of (positive) 'evaluation' (Lee 2007); in which we found significant variations both in the construction and in the delivery of positive evaluations. We go on to demonstrate that teachers select from among these formats according to the larger design and pedagogic aims of the activity. Each of these five formats, and hence practices for evaluating students' answers positively, is outlined in the following sections.

Explicit positive evaluations
The clearest and most direct way in which teachers positively evaluate students' prior answers is to respond with an explicit evaluation terms such as 'bene' (good) and 'bravo' (good boy).
(2) [PM:LT:5a:geometry] [T is demonstrating how angles work, by showing that, when changing direction, a body rotates and draws an angle on a plane surface. In the fragment, the teacher is eliciting a specific term that designates a type of line in geometry.] The teacher's question in lines 1-2 is formed through a non-interrogative practice, known as 'designedly incomplete utterances' (Koshik 2002;Margutti 2010). It may be noted that in responding 'bravo' in line 6, the teacher evaluates the student's answer with what is literally a complimentary assessment of the student. Most explicit positive evaluations follow this pattern; in only two instances the teacher evaluated the answer itself, as illustrated in example 3 (in which the initial question and the student's reply are not shown).
( In line 3, the teacher formulates an explicit positive assessment of Luca's answer, conveyed by the adjective 'interesting'.

Full -verbatim and freestanding -repetitions of students' answers
Teachers commonly endorse students' prior answers through some form of full or verbatim repetition of those answers, as illustrated in the following examples.  (0.6)((turn to look at the list of previously elicited answers on the blackboard )) 16 ! T con"chi:glie, ((she writes it on the blackboard)) shells (a few students after) By reproducing the exact linguistic content of students' responses, teachers ratify their answers as 'expected' and, as such, 'correct'. Repeating another speaker's words is in some respects intrinsically affiliative (Pomerantz 1984, 66-67;Schegloff 1996, 78-81). However, there are characteristic features of the prosody of such affiliative repeats: very often the repetition is delivered with prosodic features that add emphasis to the student's production by means of sound stretching (as in extract 4, line 7; extract 5, line 5 and extract 6, line 12), raising the amplitude (extract 4, line 7 and extract 6, line 12) and upand-down pitch shifts (extract 4, line 7 and extract 6, lines 8 and 16).
Hellermann (2003) noted a similar prosodic pattern in a sample of 177 instances of teacher lexical repeats of student answers (of which, 111 were repetitions in IRE sequences) during 25 hours of discourse in a grade 12 physics class and in a grade 11 American history class. Hellermann (2003, 88-92) found that repetitions that are oriented to as positive assessments display a specific affiliative prosodic pattern. This affiliative quality of teacher repetitions rests on four features, comparing the repetitions to the student responses: (1) the tempo of the repetition -its placement in relation to the student answers, synchrony and rhythm, (2) same pitch level as students' answers, (3) falling pitch contour and (4) duration of the repeat lengthened (in comparison with students' versions) by sound stretching.
Our analysis of the prosodic packaging of teacher repetitions, based on Hellermann's four indicators, shows that affiliation and positive assessment in our data are managed through slightly different means, though closely related to the same indicators. 4 First, the placement of these repeats is contiguous, and often in overlap with the answer, as for preferred types of actions. Second, the pitch level of the teacher's repeats is often higher than the surrounding talk, mainly because teachers raise the volume of voice, as in extract 4, line 6 and extract 6, line 12. Third, their intonation contour is always more dynamic than that of the answer they evaluate, with up-down pitch contour, as in extract 4, line 7 and in extract 6, line 16. However, the repeats in our data do not have a distinct falling intonation; rather, we find a slightly rising intonation, as in extract 6, line 8, displaying the teacher's orientation to the position of each individual IRE sequence as one in a series. As we will show, these cases belong to a specific type of instruction sequence in which subsequent questions elicit responses in a list of same-set items (see also extract 7 and endnote 7). Thus, the slightly rising intonation indicates this sense of an unfinished project. However, it is worth highlighting that there is a prosodic pattern shared by the majority of stand-alone verbatim repeats in our corpus through which the accented items of the repeated word are articulated and thereby emphasized. In extract 6, the student answers have a tenser articulation in the second syllable of the word, whereas in the teacher's repeats, the emphasis withdraws in the first syllable. Combined with a distinctive higher pitch in the word initial syllables, this produces an intonation movement similar to the up-down contour reported by Gardner (1997) and Selting (1994) as associated with heightened involvement in assessments. 5 Returning to the intrinsically positive (affiliative) character of verbatim repeats of a prior speaker's word(s), Schegloff (1996) notes that a verbatim repeat not only registers that prior response, but also confirms its correctness. Schegloff demonstrates that specifically confirming the prior speaker's correctness (rather than simply agreeing with prior speaker) makes salient that the speaker (i.e. the one repeating the other's words) already knew the answer. The base component of the action that does 'confirming', therefore, is that the speaker agrees with circumstances of which he has primary access. It is significant that this same distribution of knowledge among speakers, whereby the speaker who repeats has primary access to the information that gets repeated, is present also in our data. The questions whose answers are repeated in third position are, indeed, questions for which teachers know the answer in advance (also referred to as 'test' or 'knownanswer' questions) and for which they have primary epistemic status. These aspects, together with a high density of repetitions and of repetition-involved practices in third position, contribute to the students' understanding of this practice as doing positive evaluation.
Finally, it is worth noting that as well as repeating an answer as a means to confirm its correctness in the kind of freestanding repeats illustrated above, such repeats/confirmations may be reinforced by other accompanying conduct such as writing the answer on the blackboard, nodding and other gestural activities, and moving immediately on to a next question. For example, in extract 7, the teacher follows a repetition by declaring that the answer will be written on the blackboard (line 6).
(7) [PM:LL:1a:history] [T asks each student in turn to name something needed by prehistoric peoples. In line 1, T names student, who is thereby selected to answer next] Following her verbatim repeat in line 5, the teacher next asks Cristina about something else, about when Giovanni had changed direction (in order to establish how far in each direction Giovanni went before changing direction) -thereby consolidating the sense that 'twice' was the correct answer to how many times? Thus, displaying the answer as 'correct' is achieved partly through the repeated answer being treated as having been part of a progression towards something more, a step in a sequence. 8 In cases such as this in extract 8, where the verbatim repeat is followed by asking a next question, the prosody of the repeat is different from the stand-alone repeats illustrated in extracts 4-6, which are more dynamic and have a characteristic 'hyperarticulation' (Ogden 2006(Ogden , 1762. By contrast, the quality of the repeat due volte in line 5 as assessing is conveyed through a distinct falling intonation and by a tenser articulation of the initial syllable of the word 'due'. Furthermore, by continuing on to the next question right away in a rush-through (Schegloff 1982, 76), the teacher conveys that the answer is 'correct' not only per se, for its truth value, but also because it is exactly the right stepping stone to move ahead in the larger sequence.

Embedded and accompanied repetitions
Teachers' repetitions of students' answers may frequently be embedded in some expanded version of the answer. The student's answer in line 3 is a noun phrase, which the teacher repeats in the context of a fuller syntactic form (clause). Whereas in verbatim repetitions teachers accomplish positive judgements mainly through prosodic variations, here they achieve this by means of syntax. By including the repetition of the answer as part of a larger unit, the teacher emphasizes his/her primary access to the information elicited with the question.
In other cases, the repetition may be followed by or combined with formulaic receipt tokens, as in extract 10.  Through such expanded repetitions, teachers do more than just acknowledge the correctness of the answer. For instance, the explicit positive assessment 'va bene' (see Section 3.1) in extract 10 underlines its additional sequence-closing function (Schegloff 2007, 118). 9

Particles
The third construction format teachers may use to evaluate prior answers positively consists of particles such as mm or oh without any accompaniment, as in these two examples. Whilst the particles in these third-turn positions have no precise lexical 'meaning', nonetheless they convey the teacher's endorsement of the prior answers. The evidence for this is partly that they are treated as accepting the answer as correct; in no case does the teacher follow such particles by initiating repair on the prior answer. Moreover, there is evidence from research in a variety of other interactional settings that they work to indicate that the speaker has been informed by the other (see Heritage 1984 on oh as receipting question-elicited informings; see also Schegloff 2007 on this type of oh as proposing sequence closure). Notice that in line 6 of extract 12, the stretching of the sound and its final intonation contour register the answer as long-awaited information, endorsing it as correct and adding a confirmation nuance of meaning to the receipt.
Likewise, mm, in the position of third-turn receipts of answers in IRE sequences, as in line 5 of fragment 11, works as an evaluation token. Its up-down intonation contour ("mm:::), resulting from the initial voice higher pitch and stressed articulation of the sound, followed by a more relaxed and stretched vocalization, is similar to assessment-like mms: an Mm with rise-falling contour was found to be an expression of heightened involvement, showing affinities with assessment tokens. It can be seen as a weak version of items such as Wow, Great, or Amazing. This up-down intonation contour is characteristic of many assessments. (Gardner 1997, 147) Owing to these prosodic features, the receipt-token mm in extract 11 assesses the answer positively, but the flat final intonation gives it the characterization as a weak positive assessment. By contrast, the stretching of the sound and the final intonation contour of oh in extract 12 (line 6) endorses the answer as totally correct, the positive evaluation being highlighted by the particle's role as closing-relevant.

Direct transition to the next Q-A sequence
The final practice used to evaluate answers positively consists of simply moving on to ask a next question, without having made any evaluation, explicit or implicit -indeed, without having acknowledged the prior answer, nor repeat the students' answers, nor even utter a particle in acknowledgement of the prior answer. This is the most elliptical form of evaluation, the affirmative work relying on the very absence of any overt evaluation. The arrowed lines in the following examples show teachers asking next questions.  In neither case does the teacher make any overt assessment of the prior answer (given in lines 1-3 in each example). Notice that whereas the teacher in example 1 also did not provide an overt assessment, that was understood by the students to indicate that the initial answer (water, line 4) was incorrect; moreover, the teacher subsequently re-asked the same question (line 7). Hence, the absence of a third-turn receipt is not sufficient to convey a positive evaluation; it is the movement to a next (different) question that indicates to the students that the prior answer was correct, by being sufficient in order to progress to the next question.

Positive evaluation practices and pedagogic activities 10
In the previous section, we have reported the formats for teacher positive evaluations broadly speaking in terms of a continuum from those formats in which students' answers are evaluated more explicitly to those that are most implicit, indirect or elliptical. Their differential frequency of occurrence is summarized in Table 1, from which it is clear that repetitions of one form or another (i.e. whether simple verbatim repetitions or embedded repetitions) are the most commonly used method of positively evaluating students' prior answers. The next most frequent practice is the one that precisely does not (overtly) evaluate the prior answer, but instead simply moves on to ask a next question. Explicitly positive evaluation formats are rather infrequently used, whilst the two examples of the use of particles shown above are the only two occurring in our sample of 145 positive evaluations (there are further instances in our larger data corpus. 11 ) According to the IRE model, it is sufficient for a teacher to select from among the available positive evaluation formats -any one from among them -in order to fulfil the requirements of the IRE sequence structure. No further account is provided than that teachers evaluate the correctness or otherwise of prior answers. Thus, little is known from the existing literature about any possible linkage between each (positive) evaluation format and the specific pedagogic activity types in which it occurs. 12 Therefore, for instance, although there is a literature on how teachers (or caretakers) differentiate their verbal behaviour to convey negative or positive evaluation (Tarplee 1996;Hellerman 2003), or how they manage to progress the interaction when students fail to answer (Drew 1981;McHoul 1990;Macbeth 2004;Lee 2007), no previous research has shown that the format of positive evaluations is associated with the specific type of activity thus enacted.
Our investigation of the differential distribution of these positive evaluation formats in different pedagogic activities reveals that the format of the third evaluative turn is a constitutive property of the design and construction of the broader pedagogic activity, not only because it accomplishes the objective of ratifying the answer as correct, but particularly because it provides students with interpretive resources for understanding more broadly the type of activity being enacted and teachers' pedagogic goals. We go on to show that freestanding teachers' verbatim (full) repetitions of students' prior answers are systematically associated with recursive sequences following the same questioning procedure, in which the pedagogic activity is to test whether students are able to follow the procedure accurately enough to arrive at 'right' answers -in other words, accumulating similar (kinds of) Table 1. Distribution of third-turn positive receipts of students' answers (total 145).

Third-turn format types (instances) Percentage
Full repetitions (51) 35.20% Embedded or accompanied repetitions (54) 37.24% Direct transition to next question (24) 16.55% Assessment terms (8) 5.51% Particles (2) 1.37% Miscellaneous (6) 4.13% answers to similar (kinds of) questions. By contrast, the other formats that we have identified may be used, individually or in combination, in what Levinson describes as a more Socratic method of instruction (Levinson 1992, 90), in which teachers develop a line of reasoning or argumentation through a series of questions-and-answers, each of which stands as a building block supporting the next stage in the progression of the instruction sequence. In this second pedagogic activity, each question (and the information thus elicited) has a different weight and role in the progression of the instruction sequence.
4.1. Pedagogic activity I: questions designed to be answered by following the same procedure as was used previously to answer correctly In teacher evaluation formatted as freestanding verbatim repetitions, the exact lexical replication of the answer is the only compositional unit of the turn before the next question is asked. 13 Example 15 is a case in point. The teacher is engaging the class in observations of pictures representing views of different human settlements in various surroundings, some of which are more natural than others. She asks the students to answer in succession to what is essentially the same question -to name either an artificial or a natural element in one of the pictures selected. The aim of the activity is to build/distinguish the notions of artificiality vs. naturalness, and accordingly to verify whether students can classify items correctly. In line 7, Marco's answer is acknowledged with a lexical repetition, delivered (as we have seen above) with dynamic intonation contour (Ogden 2006), displaying the affiliative/positive quality of the repetition. However, in addition to its dynamic intonation movement, what makes the repetition an acceptance of the answer is also the manner in which the talk progresses after the repetition. Hence, three main features are associated with the positive character of this evaluation: (1) the repetition replicates the answer verbatim but is phonetically more dynamic; (2) after the repetition, the teacher asks a next question without adding any other component; and (3) the question which follows (lines 9 and 15) has the same format as the preceding one (line 5) (the teacher interrupts her first attempt in line 5 when she discovers a problem that another student is having).
Thus, if we look beyond the single IRE sequence and broaden our analytical focus to include the serial organization of questions in this sequence, each question requires students to follow the same procedure as they did to arrive at a previous 'right' answer. Therefore, extract 15 consists of a series of same-format questions that further the pedagogic activity of testing whether students have understood the procedure by which a 'right' answer can be arrived at. In this context, the acceptance of the answer is heard as the result of both the repetition format and the production of the next same-format question in a series. It is, therefore, only when the teacher moves to the next stage in the instruction sequence that the previous answer is understood as having fulfilled the prior question's requirements, so that it is now possible to progress.
In such sequences, teachers design each next question (lines 15 and 27) according to the same format or formula as they used for the first (line 5); by producing a series of same-format questioning turns, the teacher creates a self-replicating and recurrent instruction context. Starting from the second questioning turn, it soon becomes clear to students that they are to re-apply the same procedure each time. Evaluating an answer as appropriate in this activity context is not only a matter of assessing its intrinsic or objective 'truth value', but also a matter of assessing its appropriateness for the requirements of the instruction procedure thus enacted. As Lee has argued, 'questioning sequences are designed not only to elicit answers from students but also to build resources for them' (Lee 2008, 258). The teacher's selection of the format with which to evaluate a prior answer, and hence that answers to sameformat questions are ratified by same-format evaluations, contributes to the instruction potential of instruction sequences. As we will see in the next section, this observation is supported by the evidence that teachers adopt different constructional choices in evaluating student answers in the context of activities where responses to differently formatted questions are ratified by quite different evaluation formats.

Pedagogic activity II: questions developing a line of reasoning
In Section 4.1, we have seen that there is a systematic relation between the construction format of evaluations (freestanding verbatim repeat) and the kind of pedagogic activity that is managed through the teacher's questions (following a procedure to arrive at the right answer). Here, we consider a contrasting pedagogic activity pursued through sequences in which each Q-A pair -together with the teacher's evaluation -puts in place a building block in the progression towards an end point in which the teacher will conclude through having demonstrated some point, argument or concept. The pedagogic activity is thus different from that described in the previous section; this difference is reflcted in the different design of each question in the series, and particularly in the formats with which teachers evaluate students' answers to these 'building block' questions -they are not constructed as freestanding verbatim repeats, but instead take other construction formats.
Extract 16 is from a geometry lesson where a simple demonstration of angles has just taken place, consisting in Giovanni pacing the room according to the teacher's instructions. The teacher has instructed him to keep his right arm raised parallel to the floor, and to turn right or left a couple of times. Starting from the door, Giovanni takes three steps ahead to reach the teacher's desk; then he turns left and moves a few steps in the direction of the bottom of the room. Finally, he turns a second time, to the right, and takes eight more steps, ending by the window. The aim of this is to show that, each time Giovanni changes direction, his body and arm rotate, drawing an angle in the air. Notice that the form and character of each of the questions in lines 1, 17, 26 and 29 (single-arrowed lines indicate teacher questions) are different. In line 1, the teacher asks 'how many times did he change. . .?'; in line 17, she asks 'when (did he change)?' and in line 26, she asks 'what did he do..?' Parenthetically, we already begin to see the contrast in terms of the pedagogic activity illustrated in extract 15, in which the teacher uses the same format when she asks 'Can you name for me (a natural/artificial element)?', but to different students in a round -whereas here she is asking different-format questions of the same student. Furthermore, it is worth recalling that in the preceding activity, the same-format series of questions is designed to elicit information organized as a list of items (artificial or natural elements) belonging to a specific set (urban human settlements) -this being designed to direct them in grasping the meaning of the terms 'natural' and 'artificial' by comparing the items elicited in each of the two sets. By contrast, the activity illustrated in extract 16 involves developing a line of reasoning, building towards a final point, the concept of 'angle'. In this case, the instruction sequence consists of a series of different-format questions, each having a different role and weight in the larger sequence. Like 'stepping stones', these questions are designed to lead students to a concluding message or point. The different function played by each question in this 'argumentative' type of pedagogic activity is evidenced by the different format of the questions and, consequently, by the different practices used to confirm the answer as correct.
As it happens, after the first 'correct' answer (line 16), Cristina fails to answer the question in line 17, so the teacher asks the question several times (lines omitted). Having not received a satisfactory answer, the teacher takes a different route. Rather than asking Cristina to say 'when' did Giovanni changed direction, as she did in line 17, the teacher asks instead about each turning point. When the teacher asks the question in line 26, she also physically repeats the demonstration. The third question in the fragment (line 29) differs from both the two previous ones: it elicits a response through a non-interrogative format, and is deployed in third position, thereby evaluating the students' answers (lines 27 and 28) with a direct transition to the next Q-A sequence (Section 3.5). 14 This practice marks the answer (to which the question responds) as providing information elicited in the service of a superordinate purpose. Finally, when the students fail to answer -the 0.8 second pause in line 30 -the teacher answers it herself. So here in extract 16, each question has a different format; the teacher's aim is to build towards a particular conclusion through concatenated steps requiring a different procedure to be followed to answer correctly, each question also being designed to handle specific local contingencies (such as, for instance, those in which a student fails to answer).
However, the key point to highlight is that the teacher constructs her evaluations through different formats (double-arrowed lines indicate the position of evaluations). In line 17, she evaluates the correctness of Cristina's answer by first repeating that answer ('twice') and then adds her next question ('when Cristina?'). Therefore, instead of standalone verbatim repeats (the form used to evaluate students' answers in the previous pedagogic activity), here the teacher constructs a two-unit turn, where, after the repeat, the teacher moves immediately to a next and different question, progressing the sequence further (Section 3.2). Evidence that the turn is specifically designed in that way, to do that pedagogic work, and that stand-alone lexical repetition is not the appropriate evaluation practice here is the teacher's rush-through (indicated with the symbol <, line 17), pulling together the lexical repetition of the answer and the subsequent question as two elements or construction units of the same turn to build one utterance 'unit'.
So now we see that, in sequences having this kind of 'progressive' instructional approach in which teachers embark on a step-by-step progression towards a conclusion, each question is designed differently -and most importantly, the evaluations are constructed in ways other than verbatim stand-alone repeats. Moreover, each evaluation format in this excerpt is different from the other; so, different evaluation formats are used according to the interactional contingencies of the particular Q-A sequence involved.
If we look ahead to the point at which the teacher is now getting to the concept that she embarked on and was progressing towards in extract 16 -which is the concept of an angle in geometry -we can see that here too she uses a variety of question forms. Again, significantly she also constructs her evaluations through different formats. Her first evaluation in line 3 might seem to be an instance of the format used in the other pedagogic activity, a verbatim repeat of the students' answer; but it is not. This is only a partial repeat of the student's answer, 'rotazio:ne', evaluating this response positively but indirectly through completing her own turn -adding ''zio::ne' to her incomplete 'ro:ta:?'. Furthermore, this partial repetition does not display the features that characterize the hyperarticulation found in verbatim stand-alone repetitions, such as the up-down intonation contour. The stretching of the vowel sound and the falling intonation contour are rather linked to the teacher completion of her own turn.
Then, the teacher asks the question which brings the students to the key and concluding point of this phase of the lesson, asking them 'what do I form?' (lines 17 and 22) -to which they answer in unison ('angles', lines 24 and 25). The question has a rather elaborated delivery (arrowed lines). Having reached its first completion in line 9, the teacher does not give permission to answer to the students who bid for it (lines 10 and 11), being the question addressed to Cristina. In line 12, the teacher produces an increment to the question, which is followed by other bids to answer (lines 13 and 14); but Cristina still does not answer. In line 15, the teacher re-asks the question with the interjection 'eh', again without getting any response. At this point, she re-reformulates the question ('"co:sa faccio io ¼ "co:sa fo::rmo'), eliciting more bids to answer (lines 18 and 20). Only in line 22, does the teacher give students permission to answer, which they do in unison. As the delivery of the question is the result of complex processing, the evaluation is equally elaborated. The teacher's positive evaluation of this answer starts in line 26, in the form of a complex turn construction consisting of [repeat]  an angle what is it -it is formed -it is due to the change of]. By holding back other students who are competing to answer the question, and providing subsequent reformulations of the questions, the teacher highlights how the information thus elicited has a special status in the activity. Thus, once the answer is provided, the very complex format of the positive evaluation is designed to parallel this 'drawing answers together' to a conclusion. Finally, in this excerpt, this multiunit evaluation turn ends in line 28, with a designedly self-interrupted utterance, whose answer is constructed as an [embedded repeat] þ [next question] (line 32). So again, an answer is evaluated with a format constructed to fit a different question form.
The key point here is that the teacher does not use the stand-alone verbatim repeat used when engaged in the 'procedural' pedagogic activity outlined in the previous section. By choosing different formats to evaluate answers positively, the teacher builds a different pedagogic activity sequence in which students are led along a line of reasoning towards a concluding point or message.

Conclusions
We have focused on positive evaluations in classroom instruction organized as series of IRE sequences, in which teachers publicly address the whole class or single students publicly. In our data, other types of teachers' assessments -negative or other non-positive assessments -are very much less frequent, as documented above.
We identified five main formats through which teachers assess student's prior answers positively, of which repetitions of one form or another are the most common format. The analysis shows how each format is variously designed in terms of lexis, syntax, phonetics and other aspects of turn construction and sequential position -all working affiliatively to confirm the correctness of the answer. Other features such as the ways in which talk progresses to the next question also play a role in evaluating answers positively. We have demonstrated that participants themselves orient to these practices as accomplishing positive evaluation.
Our other principal finding is that teachers do not employ these formats/practices randomly; rather, they are systematically selected according to teachers' engagement in one of two specific pedagogic activities. In both of these pedagogic activities, questions are asked in a series; but each activity has a different internal organization, achieving a different instruction project. In one case, with concatenated same-format questions, teachers design an activity that aims to instruct students how to follow a procedure to arrive at the 'right' answer. In the other, characterized by a series of differently formatted questions, they develop a line of reasoning that leads to a final point. We have shown that there is a systematic association between these two instructing methods and the positive evaluation practices selected within each. In the former type, teachers use only freestanding verbatim repetition; in the latter, they construct their positive evaluation through different formats. By differentiating their evaluation techniques to reflect the overarching instruction method, teachers provide an in-built mechanism for students to recognize the activity in hand.
Finally, our study provides an explanation for the resilience of evaluative third evaluative turns in the everyday teaching practice, suggesting that this activity can be considered -and certainly is considered by both teachers and students -as manifesting in a very real way the success or otherwise of the ongoing teaching and learning process. It is appropriate, therefore, that we should better understand the 'work' managed through evaluating answers, how teachers accomplish that 'work' and how the complexities of evaluating are related to the pedagogic aims of teachers in the classroom. Notes 1. 'We define uptake as occurring when one conversant, for example, a teacher, asks someone else, for example, a student, about something the other person said previously (Collins 1982, quoted in Nystrand et al. 2003. 2. With 'eh' is rendered in Italian an interjection that accomplishes many functions (Laurentia Dascalu and Vanelli 1996). 3. There was a small group of miscellaneous formats, including reformulations, which were so infrequent as not to warrant attention in this report. 4. Reasons for the different prosodic packaging of teacher repetitions to convey positive assessment might be associated to language-specific uses of prosody or, also, to the organization of the whole domain of practices involved in conveying positive/negative evaluation of students' answer. With regard to the latter, in our data, lexical repetition or repetition-involved practices are used to provide a routine positive evaluation of the answer, whereas other non-routine treatments of the answers that convey a milder affiliation or judgements of the answer as wrong, incomplete, inappropriate or vague employ other resources than repetition. Hellermann's study (2003) is different because it compares the prosodic and phonetic realization of repetitions used to do positive evaluations with those conveying judgements other than positive. It would be also wrong to compare our findings to those in Hellermann's paper because the two studies have different purposes: the import of prosody in the IRE exchange in Hellermann's vs. the organization of the domain of positive evaluation in IRE-shaped instruction sequences in our work. Another variable that might be responsible for the differences of our findings in comparison to Hellermann's is the different ages of the students who were recorded. 5. It is worth adding that these observations on the phonetics of teacher positive repeats are also in line with findings from a study by Ogden (2006) on the phonetic resources used to convey agreement/disagreement in ordinary conversation, where the author provides evidence for the existence of a distinct phonetic pattern for agreement in second assessments. 6. In this example the teacher repeats the answer with slightly rising intonation (as indicated by the symbol ¿). In Italian, not differently from English and other languages, repeating the answer with rising intonation might be heard as prompting self-correction. However, attention has to be drawn here to two contextual properties of the interaction, which suggest to exclude this function. First, the fact that the teacher is eliciting responses in a round (see background information provided); the rising intonation indicates suspension and characterizes the answer as one in a series, with other answers to come. Second, and in support of that, it is worth noticing that the same intonation contour is consistently used also to deliver the second unit of the turn, in which the teacher further ratifies the correctness of the answer by announcing that it will be written on the blackboard. 7. The interrogative adverb 'when' refers back to prior question in line 1; therefore, here, 'when' is elliptical for 'when he did he change direction?'. 8. See also McHoul who observes that whilst there is no explicit evaluation, '(the) answer is nevertheless shown to be acceptable by virtue of the teacher doing a thematic continuation of it', in McHoul (1990, 357). 9. To our knowledge, there are no studies available on the use of 'va bene' in Italian interaction, on which to ground a characterization of this use. However, from a very cursory inspection of a collection of ordinary interaction in Italian telephone calls, we register a very high density of 'va bene' in the closing of the calls, especially when speakers make final arrangements. 10. We use here the terms 'evaluation format' and 'evaluation practice' as referring to the same type of analytical object, to include also 'zero-evaluation' cases, in which the teacher does not produce any explicit (lexical) or implicit (phonetic) evaluation-dedicated turn or component, but moves to next question. Thus, although otherwise specified, here we consider these two terms interchangeable. 11. It is worth noting that a study by Koole (2012) based on a corpus of Dutch dyadic teacher-students interaction (ages 12-13) shows very different proportions. In that study, overt evaluations outnumber the other formats by covering the 57% of the 170 teacher evaluations of the sample. We do not have enough evidence to suggest that these different findings might indicate a cultural variation in the distribution of overt evaluation of student answers; especially if we consider that the Italian and Dutch contexts presents also variations concerning variables other than cultural background, such as the setting (dyadic vs. teacher/whole-class) and the students' age (8/7 vs. 12/13). In our corpus, we did not register any individual variation among the four teachers on this respect. Therefore, we can affirm that this is a commonly used practice. 12. Two studies in particular - Nassaji and Wells (2000) and Nystrand et al. (2003) -have addressed the relationship between the function of teacher third turn in IRE sequences and type of classroom discourse. As argued by Lee (2007), however, in that study, the function of the teacher third turn is defined by means of a predetermined and abstract coding system, whereby a fixed and limited repertoire of possible actions are identified (for instance, evaluation, justification, clarification, etc.) and related to a specific teacher role (primary knower vs. not-knowing interlocutor), types of questions (known-answer vs. genuine or exploratory questions) and classroom discourse modes (monologic vs. dialogic). 13. Note that examples 7 and 8 are variations of the full, verbatim and freestanding format through which teachers may deliver positive evaluations. In those cases, non-vocal conduct accompanied the (verbal) repetition -which, therefore, is otherwise freestanding (i.e. verbally). 14. Here, it is important to understand that 'girato' (Cristina's answer in line 28) is a verb (past participle of the verb 'to turn'), whilst the teacher's question in line 29 concerns the action 'giro' (noun, in English 'a turn') that has therefore been performed. Rising intonation (not necessarily a question) .hhh In-breath hhh.
Out-breath wohrd heh Audible aspirations within words, including in laughter word Talk quieter than surrounding talk WORD Talk much louder than surrounding talk Word Vocal emphasis "word# Marked variations in pitch in the following word/syllable (word) The transcriber's 'best hearing' of what is said, when unclear or uncertain < Indicate that the following talk sounds like it starts with a rush <word> Talk much slower than surrounding talk >word< Talk much faster than surrounding talk ((. . .. . .)) Text in italic in double parenthesis represent an effort to describe other forms conduct !
Indicates lines of particular interest