Teachers' decision making in relation to the use of talk in the classrooms (particularly collaborative talk) is interesting for a number of reasons. First, talk remains by far the most used medium of instruction and classroom action-Spoelders (1987) reports that two thirds of classroom time is spent on talk-and yet we seem to know little about why teachers use the talk they do. Second, classroom talk is typically dominated by triadic dialogue or an I-R-E (Initiate-Response-Evaluation) pattern of talk noted in the literature as consistent across grades and subjects (Spoelders 1987; Wood 1992). Such tight teacher control of talk remains when alternative forms of working such as collaborative and peer work are consistently recommended for middle years classrooms (Epstein, 1990). Third, and perhaps most interestingly, studies seem to indicate that teachers utilise traditional forms of talk, even though they know about, and have been trained in the use of, alternative forms. Two studies in particular are pertinent to this issue. Alvermann, O'Brien and Dillon (1990), found that the teachers in their study knew in abstract form what made for a 'good discussion' but that this apparently valuable form of talk was rarely utilised in the teachers' classrooms.

The second study reported by Wood (1992) found that, even when trained to utilise strategies derived from research to make talk more 'thoughtful and productive', teachers did not continue these practices after the conclusion of the program. These studies confirm that there are many influences on teachers' programming that cut across issues related to knowledge about talk. If, as the studies cited here suggest, teachers' planning takes account of a complex of factors and is somewhat constrained, we needed to better understand what teachers do take into account when programming and using talk in their classrooms.

During 1995-97 I worked with a group of eleven teachers (6 lower secondary and 5 upper primary) in nine South Australian schools serving largely disadvantaged communities on a research project which involved the exploration of the use of talk in middle years classrooms. The project involved me in supporting the teachers to conduct their own research into classroom talk, while developing a meta-analysis of the kind of curriculum and pedagogy they were practising.

While the teachers were free to choose the kind of talk they focussed on in their research, all used the opportunity provided by the project to experiment with forms of collaborative or peer talk arising out of various group tasks the teachers established in the subjects they taught. During the project the teachers were supported to plan, implement and reflect on units of work involving such talk in a recursive action-research cycle. The teachers video- and audio-taped classroom talk in these units of work and were subsequently involved in reviewing these materials, writing reflections on them and showing and discussing their data and reflections with the other teachers in the group. The teachers also read, discussed and wrote responses to the meta-analysis I conducted based on the their own writing and data selection.

As a group the teachers were clear about the difficulties involved in establishing and maintaining collaborative talk activities in their classrooms. Most noted that they usually did not teach in this way and confirmed that the I-R-E pattern of talk, noted above, was dominant in their own classrooms. In spite of these difficulties, however, during the course of the project the teachers all succeeded in using collaborative talk and found that it was a valuable strategy for engaging adolescents in the curriculum, often providing successful learning and assessment experiences for students who usually did not succeed in their classrooms. The nature of those tasks and learning outcomes, however, is not the focus here (see Cormack, 1998; Cormack, Wignell, & Nichols, 1998 for a discussion of these issues), rather, it is the data the teachers provided on the issues they take into account when they plan activities involving talk.

In order to consider the factors teachers took into account in their planning, I analysed the participating teachers' own written plans, reflections and reports on what happened in their classrooms as they implemented collaborative talk activities. The first run of analysis was a process of identifying lexical chains where words or groups of words around a particular idea were highlighted and given a tentative label. A total of 381 different instances was identified. For example, "a relatively high proportion of students have been tested and found to require a negotiated curriculum plan", was labelled a comment on student skill/ability; "The class shows a keenness to participate in laboratory practical work" was classified as a comment on student motivation/attitude; and "I am in only my second year at this school and am still coming to terms with the programming and behaviour management techniques required to teach in schools of this nature" was categorised as a comment on the teacher's own experience. These categories were refined and checked for reliability with a fellow researcher and counts were conducted on the categories to gather a sense of their relative emphasis in the teachers' reports. This analysis suggested two major outcomes. The first of these was that there were key differences between teachers and their contexts which impacted on their work. The second outcome was that, alongside these differences, there were some discourses running across the teachers' reports of their planning and work which indicated some common frames shaping what they decided to do.

This analysis was written in rough form including identified quotations from the teachers' reports and provided to the teachers for comment. At a review meeting the teachers read, discussed and provided written feedback on the analysis. The teachers approved the general direction of the analysis and wrote further comments they wished added to the analysis. The following two sections discuss in turn the different factors teachers considered, and the discourses that appeared to be constituting teachers' decision making.

Teaching is a complex process and this was reflected in the data on the issues teachers commented on when describing their planning and programming of talk. Across the group, the categories of factors the teachers highlighted as impacting on classroom talk could be placed into three broad groupings, those which represented:

The term 'reading' is used to emphasise that how teachers describe their students, their work contexts and themselves is a product of the representations that are available to them in different times and places. Teachers use available discourses to construct their own and their students' subjectivities and deploy a variety of rationalities to explain why things are the way they are.

Differences between teachers showed up in two ways. First, there were aspects of their work situation or students that were clearly different from other teachers in other sites and this had an impact on what they decided they could or would do. Another way of saying this is that teachers worked in different contexts which led to differences in the ways they worked. For example there were clear differences in teaching and using talk reported in the teachers' classrooms between those who had all boys classes and those who had all girls classes.

The second way that differences between teachers were highlighted was that the teachers did not always respond in the same way to similar factors or saw them as having different relative importance. For example, none of the primary teachers even mentioned the issue of curriculum content coverage, while this was a key issue for two of the high school teachers. Over the course of the project, some of the teachers changed the ways they thought about particular factors adding another way that difference emerged.

Comments categorised as a "readings of the student" included statements about students' characteristics that teachers made in their descriptions of their work and their planning. All statements that highlighted a characteristic of students were categorised, whether or not the teacher went on to explain how this characteristic impacted on their planning or programming. Any statements which did refer to the impact of characteristics were highlighted as most pertinent for this analysis. In all six main ways of representing the students were categorised as:

The first three categories were the most common and are focussed on here. The most common characteristic mentioned was students' ability or skill. Teachers tended to give these comments without saying how the students' ability or skill impacted on what they did with them in relation to talk. Comment on ability/skill were both negative and positive. Negative comments tended to describe problems that implied a restriction on what it was possible to do with the students. Positive comments suggested skills that students had which implied possibilities for action.

This last comment provides some indication of the reasoning being employed by the teachers, in that students' ability was seen as impacting on how directly the teacher needed to be involved in the activity or talk to ensure that the students gained from it as learners. The more needy the students were perceived to be by the teacher, the more the talk needed to be directed by the teacher. In contrast, able and skilled students were perceived to need less direction and support.

Generally, the comments on students' attitudes and motivation were similar to those on ability with a similar set of implications for the forms of talk that could be employed.

Especially for the secondary teachers, issues of attitude and ability combined into a reading of the 'class' as a group which, depending on the mix, offered different challenges and possibilities for the teaching program. The focus on students' abilities and attitudes also arose partly because of the involvement of the teachers in the research project. All of the teachers reported 'knowing' their students better as a result of their involvement which in turn led them to have more to say about them as individuals.

A category of comment that was related to ability and attitudes were those on students' 'personality' and approach to relating with peers, referred to as 'social skills'. The primary teachers tended to provide most comments in this category usually in relation to comment on individuals or groups who stood out as particularly problematic or able as talkers. 'Personality' and social skills were implicated in talk in much the same way as ability and attitude because of the impact these attributes had on students' ability to work with the teacher and each other and therefore constrained the kinds of activities and groupings that could be established.

Three of the other categories grouped under the heading "readings of students" were social and family features; cultural and language features; and gender. These were not as frequently mentioned as the first three categories but still featured strongly.

Information on students' social and family lives as well as the cultural groups they identified with and the languages they spoke tended to be provided by the teachers when profiling their classes and schools at the beginning of their reports. Usually no direct comment was made on how these features impacted on their teaching and the use of talk in the classroom except in the most general terms. There was sometimes an implied link between such factors and what it was possible for teachers to do in school. Comments on students' gender were possibly prevalent because of the number of single sex classes the teachers in the group taught (two of the teachers were in single sex schools). The perceived qualities of boys as distinct from girls, were cited as strongly influencing the kinds of activity the teachers planned and the sorts of talk they expected in those activities.

The predominance of comment about students' ability, attitudes and 'personalities' as well as about their homelives and cultural and linguistic characteristics demonstrated how, for teachers, the students were the immediate context and that perceived differences between students as individuals and as collective groups affected what it was possible to do in using talk.

As well as commenting on the way students' characteristics impacted on their planning, teachers also discussed the way their own experience, institutional position and other personal factors affected what they did. Such comments were categorised and grouped under the heading 'readings of themselves'. The four most often mentioned categories in this grouping (with the first two predominating) were:

Teachers' characterisations of their personal beliefs or approach to teaching were the most frequent kind of comment they made about themselves. This was often expressed in terms of 'my' approach to teaching, 'the way I do things', or 'beliefs' about teaching. When teachers spoke in these terms it was sometimes to say that their beliefs/approach allowed them to facilitate talk and provide a less restrictive program that involved students in the program.

However the most frequent comments were on the kinds of approach the teacher took to working with students which gave an indication of their approach to classroom talk.

An allied category of comments to those already described included those labelled as referring to teachers' experience or study as a factor in how talk was used in the program. All of the teachers in the project were experienced practitioners and their various experiences had an impact on what they did. For example, one teacher mentioned her experience in literacy work as providing a useful background for the research into talk.

Three teachers highlighted particular study or courses they were involved in or had completed. It seemed that the teachers were adapting strategies explored in particular curriculum areas such as literacy and science to their work on talk. Interestingly, for two of the teachers, experience was mentioned in a negative sense in that it had not prepared them for what they now faced in a new school settings.

Four teachers made comments which showed that they regarded their position in the school as an influence on planning. Position in school was mainly reported as being a factor that facilitated their being able to use talk in the classroom. One of the teachers explained how her role made it easier to focus on teaching rather than managing behaviour.

Another reported that his position provided a high level of freedom to make decisions about content and pedagogy.

However, position in the school could be read as being a negative factor in planning for talk. One teacher had been appointed acting principal and because of this:

A factor that two of the teachers took into account was the impact that a talk curriculum, or particular activities involving talk, would have on their work as well as on the students. The feeling here was that using talk in the classroom involved additional work for the teacher and this prospect had an impact on what the teacher might do.

The teachers' comments in this grouping of 'readings of themselves' reinforced a view that teacher characteristics do have a strong influence on what gets done in classrooms. These teachers showed that their own experience and feelings of confidence; the training and study to which they had access; their institutional role; and their views on the kind of work involved can all influence the shape and nature of the talk program. Because of the differences between teachers in these characteristics, variation between the ways talk would be used and valued in the different classrooms could be expected.

The third broad grouping of teachers' comments included statements about the school, faculty, or worksite in which the teachers worked. The four most often- mentioned categories were:

The largest category of comments referred to aspects of the school 'culture' or related characteristics which impacted on the teachers. Features mentioned included collegiality, staff turnover, supportiveness of parents, and school size. For the teachers the most important of these was the collegial context within which they worked. This was cited as a positive by the teachers who said that supportive colleagues and leaders made a real difference to morale and opened up spaces for experiment and positive action with students, or even provided much needed practical support through team teaching.

One teacher also mentioned the constraints caused by his reading of other staff who he perceived to be judging the quality of teaching according the noise level of the class. Other teachers reported this to be a factor in discussions but did not mention this in their written reports.

Resource issues were also frequently mentioned. These encompassed features such as facilities, learning materials, teacher-student ratio and access to additional personnel. Comments such as the following were typical, showing a direct link between resources available and what it was possible to do in talk.

Another category of comment which was frequently made referred to the impact of the nature of the subject being offered and related timetabling issues on the talk program. All but one of the comments on these issues were made by secondary teachers. The following comment from a science teacher shows how issues of curriculum affected what occurred in the classroom.

The nature of the subject was sometimes seen to be facilitative in providing a place for talk.

No matter what the subject, though, the secondary school timetable, could have an impact on opportunities for talk.

A related issue to that of curriculum and timetabling was that of assessment. The demands of assessment were perceived to be a limiting factor by some of the secondary teachers. A key issue in this respect was the assumed individualistic nature of assessment as a measure of performance in the context of talk, which is by definition an interactive activity involving two or more people. The teachers who commented on this issue indicated that the use of talk in assessment introduced challenges that traditional approaches could not meet.

The teachers' comments on their institutional context discussed here confirm that factors external to teachers and students are a key influence on how talk gets used in classroom programs. However, it is also clear that these influences need to be considered alongside and in combination with teachers' readings of themselves and of their students. The number and scope of comments canvassed in this section provides an illustration of the multitude of factors that affect teachers' use of talk in classroom programs.

The discourses working across teachers planning

As well as showing the differences between teachers and their work contexts, the analysis discussed above indicated that there were some commonalities between teachers and across the range of factors they discussed. In the main, these commonalties emerged from the language teachers used to describe students and how they planned to work with them in the classroom.

This is unsurprising. There are powerful educational discourses on learning, childhood/adolescence, gender and schooling which serve to construct ways of interpreting experience and which help to constitute teachers. The teachers themselves pointed this out during project discussions. In one meeting they discussed animatedly, the way that staffroom stories about students and their families could work to present a negative view of the community that impacted on how they school responded to students (this was significant because the majority of the schools in the project served identified disadvantaged communities).

The teachers found they had few ways of talking about their students at their disposal that did not produce their students as problems. The teachers felt that they were working against the grain in trying to represent their students as capable rather than lacking because they were surrounded by representations of their students as deficit. This experience is captured by the notion of discourses 'speaking' through the teachers or, as Cherryholmes (1988 p. 14) puts it, "we are captives of our discourses-practices .... They control us, not the other way around." A science teacher in the group hinted at this in a response to an early draft of my analysis when she commented on the ways external standard setting represented student ability in a particularly powerful way.

In order to follow up this insight, the teachers' descriptions of what they took into account in their planning for teaching were analysed to consider the ways of talking and writing about students that operated across the group. The guiding question was: "What are the dominant discourses operating in the ways teachers constitute their students"? However, being mindful of the discussions held with the teachers about how they were both influenced by, and attempted to resist, deficit perspectives, the analysis also noted instances where teachers were resisting those discourses, or finding gaps, spaces and contradictions between dominant discourses to talk about their students and their work in different ways. The definition of discourse used was:

Three key constructions of students and their work could be seen to operate across the teachers' descriptions. These were constructions of students according to

A range of intersecting discourses informed these constructions of the students which are discussed below.

When teachers described their students and their impact on the teachers' planning for talk, concepts of 'ability' were continually employed. This occurred either through direct attributions of ability to a student or groups of students (eg 'mixed ability', 'able students', 'least capable student') or a reading-off of ability from tests or other techniques for judging students' performance (eg 'below the state average'). Sometimes students were referred to by institutional labels attached to children who were provided extra learning support or who were designated as having a set of (dis)abilities (eg 'adaptive ed students'; 'students of high intellectual potential'). Broadly speaking these labels were used to indicate the 'special' nature of the students as individuals or groups that had to be accounted for when planning for talk.

Versions of ability operated strongly in four of the teachers' reports and were used both to indicate a factor taken into account in designing work and as a lens for examining and explaining student performance. Three overlapping versions of ability were identified:

The use of ability had varying effects on teachers' planning. Sometimes ability provided an explanation of students' observed behaviour.

For this same teacher ability was a foundational concept, providing a perspective on students that could not be questioned, even by apparent good performance in tasks involving talk.

These labels tended to be used as explanatory in terms of the kind of teaching approach and/or work that the teachers could employ with the class. For example, ability had an impact on grouping in terms of who should work together.

Ability had an impact on what the teacher expected from students (even if this was not always borne out), or predicted for their learning trajectories. Ability also had an impact on organisation of the class and approach to teaching.

Not all of the teachers used the construction of ability and one actively resisted it. In her writing there was a consistent refusal to use the labels inferring fixed ability - her writing maintained a strict reporting of observable behaviours with an emphasis on what the student could do as opposed to what they could not do, in spite of the fact that her class consisted of students with severe behavioural and learning difficulties.

There was also evidence that conceptions of ability were being questioned by the teachers, especially as a result of the close contact and observation arising from their involvement in the project. For two of the teachers, as a result of the close involvement and observation of student talk in the project, undifferentiated conceptions 'ability' came to be broken down, if not abandoned.

This last quotation is illustrative of the power of the discourses working in the representations of the classrooms - here ability in reading was split from ability in writing which enabled an explanation of potentially contradictory evidence that did not challenge the basic construct.

It should not be assumed that through encouraging talk in classrooms and closer teacher involvement with student talk for learning, that teachers will automatically question basic assumptions about ability. This next quote, while about 'personality traits' rather than about 'ability' (although both belong to a psychological discourse), is illustrative of the way that 'knowing' the students better may lead teachers to feel they can better speak the 'truth' about students.

Whether or not this has positive effects for students, particularly those from disadvantaged and diverse communities, may well depend on the labels and constructions teachers employ in speaking that 'truth'.

All but three of the teachers' reports referred to family or homelife or social group in profiling their students. Teachers in the most disadvantaged government schools, tended to highlight descriptions of family circumstances and/or the extent of poverty as indicated by receipt of government support or the complexity of the school as indicated by the range of cultural groupings of students and numbers of special education students.

When describing the aspects of their classroom and school contexts they felt were important, most of the teachers tended to provide a list of factors using labels with a sociological flavour.

As this last quote shows, alongside the listing of sociological labels was a concern with the family circumstances of students. Teachers provided commentary on the nature of families, their perceived circumstances and in some cases, their perceived 'supportiveness'.

Clearly the students' social circumstances, especially their family lives as inferred by levels poverty and other indicators, were significant for the teachers in thinking about how they should work with those students. For example, absenteeism was cited by three of the teachers as a problem and connected directly to homelife and parenting.

Two kinds of responses were evident in the ways teachers moved from acknowledging the difference, diversity and constraints of their students to making decisions about how to design a program for them. The first of these responses was that this difference and difficulty defined a 'lack' which must be overcome.

Homelife as limited or poor was clearly seen by teachers working in schools with significant numbers of poor students as an explanation for school problems - what was being asked for in the school was presented as very different from, or even opposed to, what occurred in the home.

This represented a measuring of students' lives against what is valued or required in and for school. One of the teachers listed the sorts of questions that she believed teachers in her school asked in relation to students' backgrounds providing an example of how family and community life was defined in relation to an assumed school norm.

The second way that teachers responded to the issue of students background problems and differences was to find ways of resisting seeing their students as problems and to question how students' lives were used by schools in explaining success and failure.

Some of the teachers attempted to reconceptualise a response to the students that didn't blame them for lacking what the school was charged with teaching. These responses came both from the teachers' reports and from discussions in the meetings where such issues were made an explicit focus of discussion. Responses included rejecting the implication that families were responsible for delivering what the school required and asking what the school was doing.

Another response was considering diversity and difference as a potential resource while other responses included recasting complexity as a rich diversity and focusing on perceived 'strengths' of the students and their families even if this was not officially recognised by the school.

This last approach proved very powerful for some of the teachers in the group as a means of viewing their students as (cap)able. One of the teachers who did this most successfully tended to rely on a child-centred discourse connected with wholistic approaches to the teaching of literacy which emphasised describing what students can do, rather than what they can't.

For this group of teachers, the issue of 'these students' as a significant context for their work was ever present and discourses which spoke of them in terms of the features of their social group and family lives were a constant presence, whether in constituting the students as problematic or lacking in terms of what the school wanted, or in terms of something that had to be worked around, resisted or ignored.

Thus teachers were caught between a view of students as lacking and needing compensatory, even oppositional training, in school on the one hand, and another view which attempted to see these students in positive ways. The problem as the teachers identified it was that a language was not available that helped them both to acknowledge the vast differences between the cultures of home and school and to work positively and constructively with those differences. The pervasiveness of the discourse of deficit allowed them little room for manoeuvre in their planning.

Gender

A third key construction operating in the teachers' descriptions was gender. This is perhaps unsurprising given the fact that two of the teachers worked in single sex schools (one boys, one girls Catholic school) and that three of the teachers in the government schools were involved in teaching single sex classes at some time during their work on the project. However gender was also raised by other teachers in the group as an issue in their planning and teaching. Discussion of gender often related to discussion of behaviour management or 'discipline' issues. It was boys, in particular, who were discussed in this way:

This same teacher compared her 1996 year 8 co-ed maths class with her 1995 all boys maths class. She reported that her 1996 students were willing to work independently and collaboratively, something the boys could not do. In the primary setting where two of the teachers in 1996 split their two year 6/7 classes into an all boys and all girls class there was a similar reading of the boys as incapable of working collaboratively.

Both of these sets of comments reveal the reasoning that occurred to impact on decision making about how to work with students on talk. Put simply, for some teachers boys could not be trusted to work and talk independently of the teacher on tasks that required collaboration. This meant that such talk was either avoided or set up in such a way that the boys could be 'monitored' for 'silly off-task behaviour'.

The teacher at the all boys school who joined both that school and the project in 1996 commented on the difference she perceived in working for the first time at an all boys school.

In telling contrast to the teachers quoted here, the teacher in the all-girls school did not raise any issues of behaviour management, described his school as having "no major behaviour management problems". It seems no coincidence that researching talk in his own class in he described some questions which focussed on students who were 'quiet'.

While the discourse of gender was most obvious in the language of the teachers working in single sex classes or schools, it was not confined to these reports. Usually in relation to behaviour, boys were mentioned as problem, for lack of cooperation, disruption, and 'off-task' behaviour.

This gendered discourse has an enormous potential to influence how teachers plan for talk in classrooms. For these teachers, the talk that was valued in collaborative activities often equated with cooperative talk conducted at a distance from the teacher's direct supervision. Given the gendered discourse illustrated here, girls were more likely to be constituted as good at talk and boys as problematic.

Students as talkers and workers were constructed in these classrooms through the interconnections between a variety of discourses. While these three key constructions of ability, home life/social group and gender have been discussed separately here, they also worked in combination to produce views of students and lines of reasoning on what it was possible to do in the classrooms with talk. For example, it was possible to see that constructions of ability combine with those of student home life and gender to produce some students as 'other,' and as incapable or untrustworthy products of an impoverished background.

This analysis leads to a potential explanation for the fact that talk practices in classrooms have changed little over time in spite of the interventions of research and professional development programs. Constructions of students and the discourses that inform them that have been identified here have a strong purchase on the ways that schooling and teaching are thought about and enacted and they are not restricted to considerations of talk in classroom programs. The longevity and power of these discourses, constituting as they do the institutions of schooling and their credentialling of students, will not be addressed simply through a training program on better talk in the classroom. School change must address the relationships between teachers' practices and broader issues of how schools are organised socially (Elliott 1993). This analysis suggests that that school change must also address the relations of teachers' practices with how schools are situated in their communities. The teachers in this project illustrated the difficulty of planning talk programs for students if they are constituted as 'other' to the school and its intentions.

However, this was not the whole story. There was evidence of the teachers finding gaps and spaces between these constructions and the discourses that informed them and using other ways of constituting their students that resisted this 'othering'. The teachers, through the process of research and reflection, did find alternative ways of viewing their students (eg as capable; as different rather than deficit) and found examples of work and action in their classrooms that served to contradict the ways that dominant discourses constituted their students (eg boys who stayed on task, poor students who learned complex mathematical concepts, less 'able' students who succeeded).

By introducing the concept of the constitution of students through discourse into discussions at project meetings with the teacher researchers, it became possible for the teachers to think otherwise about their students, and to imagine them as successful learners and talkers. This enabled them to establish ways of talking and working in their classrooms for which there was usually no space.

The first half of the analysis which discussed the range of factors teachers take into account in their planning served to illustrate the complexity involved in working with diverse student populations in classrooms. It was not only differences among students that indicated variation between classrooms, there were differences between the teachers and their experience and views about teaching, and between the institutional contexts within which they worked, that all made for particular combinations of factors affecting what it was possible to do in each site.

This analysis also revealed the complexity involved in changing classroom practices in relation to talk (something the research shows is notoriously difficult to do). Programs for change that do not address each of the contextual fields discussed above - teachers' views on students, on themselves and on their institutional context - and acknowledge that they will intersect differently from school to school, and classroom to classroom seem unlikely to succeed. Add to this, issues raised above in relation to the discussion of discourses including the need to reevaluate the relation of the school to its community and the ways educational discourses constitute students and teaching and it is possible to measure the scope of work required to effect significant and lasting change to school talk practices in the middle years.