25th October 1993







The form of interaction I shall discuss in this Chapter is not the one

that first comes to mind as being "collaborative". I am not

intending to conjur up the image of pupil and teacher engaged together

in a sustained interaction around some computer task. Such intimate

interaction is a relatively rare luxury in most classrooms. So,

teachers and pupils in focussed and protracted collaboration at

computer-based problems is not my present interest. Instead, I will

be concerned here with another sense in which pupils and teachers

interact in the presence of computers. This is the more commonplace

sense in which pupils are engaged in some activity involving their

teachers but where those teachers' contributions are more indirect, or

mediated, or deferred. That is, they make only intermittent contact

with the task or refer to it on occasions when it is not actually in

progress. These are still encounters that are conceived to support

pupils' learning. The joint involvement with teachers arises because,

to varying degrees, teachers will define, interpret or intervene in

what is going on, or in respect of what has previously happened. This

is a loser sense of teacher-pupil-task interaction but it is typical

of children's normal classroom experience and is properly considered

one sense of joint "interaction" in relation to, or inspired by, some


My plan for the discussion of such circumstances is as follows. In

the first section below, I shall describe one experience of my own

involving the implementation of a primary school computer activity.

The point of this example is to identify certain problems that

surface when we focus upon teacher-pupil-task interaction of the kind

defined above. This example prompts consideration of just what it is

that might be done or said by teachers in support of their pupils'

computer-based experiences. I will suggest that what gets said in the

course of such supporting talk is something that matters. Moreover,

we should understand more about it. Yet, the analysis of

instructional talk that guided our consideration of interacting *with*

computers turns out (at first sight) to be less helpful to a

consideration of situations where the interaction is *in relation to*

them. The crux of the problem is that the processes invoked in

theorising about ZPD encounters are too intimate: they assume a degree

of person-to-person interaction that may not be so easily promoted in

the situations that I wish to consider next. This shortfall prompts a

return to theory in the final sections of the chapter: conceptions of

instructional discourse are extended to encompass categories of

classroom talk that reach beyond the contexts of intimate tutorial


Briefly, my argument will be that, through analysing zones of proximal

development, cultural psychology has furnished useful insights into

the character of instructional interactions. However, this analysis

may be too focussed upon interactions of the traditional tutorial

kind. In practice, much real instructional discourse is, instead,

embedded in a more open-ended and communal kind of interaction. Thus,

it is not as intimate as ZPD conceptions suggest. For example,

instructional talk is often concerned with linking the current

activity to previous events that the participants have jointly

experienced. So, it is not simply concerned with supervising

the actions that might effectively complete some current problem

solving task - although the jigsaws and puzzles of much ZPD research

might encourage this image.

In order to address these senses of instructional interaction among

pupils and teachers, I shall refer to a further concept that aims

to systematise classroom talk. This is the discourse principle

discussed by Edwards and Mercer (1987) and referred to by them as

"common knowledge". Their work will prove a useful basis for

integrating accounts of instructional exchanges that refer to a wide

range of classroom circumstances. What I shall suggest is that this

framework is one that leads us again to recognize the central

importance of intersubjectivity. That conclusion, again, urges upon

us the need to locate computer-based learning in a context of

interpersonal support.






Teachers could be forgiven some irritation with cultural psychology -

at least, insofar as it seems to foreground the zone of proximal

development as a framework for instruction. That characterisation of

instructional talk may not seem in tune with the reality of classroom

life. When this zone is realized for empirical purposes it will

usually be a rather peaceful place. It will not be densely populated,

the participants will tend to be mutually engaged, and the action will

be allowed to proceed relatively undisturbed. Most formal education,

however, does not permit any abundance of such relaxed encounters. As

Tharp and Gallimore (1988) comment in discussing ZPD interactions as

*classroom* phenomena: '..conditions in which the teacher can be

sufficiently aware of the child's actual, inflight performance, simply

are not available in classrooms organized, equipped, and staffed in

the typical pattern.' In short, the opportunity for dedicated and

focussed interaction around a task is something of a luxury.

Teacher-pupil ratios mitigate against such pleasures; pressure to

sustain order may do so also. Thus, the model of instructional

interaction promoted within cultural theorizing may seem too remote

from what routinely can be achieved within the realities of


There is certainly some discrepancy to be confronted here, but the

mismatch is not as dramatic as I have sketched it. There is no reason

why devices identified within studies of more intimate tutorial

encounters should not be reproduced in the busier context of class

instruction, albeit with less intensity. Newman et al (1989)

illustrate this in their development of the appropriation concept for

ZPD activity: their studies were successfully grounded in whole

classroom settings. Nevertheless, if these (ZPD) theoretical concepts

are claimed to play a significant role in systematising educational

practice, they may need some elaboration - in order quite clearly to

include the organization of discourse under the normal and busy

conditions of classroom life. As it happens, I believe the concepts

of intersubjectivity and socially-shared cognition prove valuable in

helping us bridge this gap - one between instruction as it gets

modelled for theory building and instruction as it often gets

practised in institutionalized settings.

To demonstrate the kind of social encounter that we must consider

under the present heading, I have chosen to dwell on an example of the

implementation of one (primary school) computer programme that I

have designed and observed myself. The programme is modest in its

aims. It is in the drill and skill tradition but can claim some of

the engaging properties illustrated in the estimation/harpooning

example praised by Scott et al (1992) and discussed in Chapter 1. It

is therefore likely to hold some appeal within primary school

settings. It was conceived in response to discussions with teachers

regarding the problem of moving children towards understanding

multiplication in relation to repeated addition.

Children have to guess the number of squares making up the snake

figure (using the screen pointer to select a number from an array at

the top of the screen). The programme generates number targets by

randomly selecting a value between 2 and X, the value of X increasing

after each correct choice. The task is difficult because not only do

the numbers get bigger, but the snake is constantly repositioning its

starting point and then uncurling again. Such a cycle involves the

snake uncurling into each of a sequence of N x M rectangular matrices,

where N and M are factors of the current target number, T. Sometimes

a number, T, will allow only one such matrix (a prime number: 1 x T).

For other numbers, there may be several matrices according to the set

of its factors. All the possibilities are illustrated for the case

of T=12 in the schematic diagram of Figure 3.1. During the activity

itself, they would be displayed successively and not simultaneously as

pictured here. The leading square has a face drawn upon it. This

cycle of repositioning and unfolding the various factor matrices for a

target number, T, continues until the guess is made.

-------- Insert Figure 3.1 about here --------

The speed at which these cycles occur can be set in advance but it

would be such as to make a simple 1-2-3... counting strategy soon

prove too limited. The child has to start finding estimating

strategies to keep the game going (a score can be incorporated to

motivate this: it is calculated as the sum of correctly guessed

targets so far and it is zeroed when the activity is restarted after

an incorrect guess). Six and seven year olds will quite quickly adopt

repeated addition strategies to cycles where a single row or column of

a matrix can be counted - 4+4+4 giving 12 for example. However, the

achievement of special interest is that whereby they move from

generating a solution by repeated addition to the more economical and

powerful one based upon multiplication of columns and rows - 3 x 4

giving 12.

This programme has been observed in three classes of 7 year-old

children (Crook, 1986), all at a stage where they are studying factors

and gaining multiplication table knowledge. I would like to extract

from these observations two (related) points concerning the use of

this programme. The first is expressed in Figure 3.2 which

illustrates changing performance on the activity with experience: it

shows changes in the average highest number that was correctly guessed

in successive sessions.

--------- Insert Figure 3.2 about here ---------

The curves show results for children working singly (S) and working as

pairs (P). At two points indicated, all children had a session of the

activity alone. The curve appears to have reached a ceiling. Because

some of these children were working together, it was possible for the

observer to gain more insight into their strategy - as it might

sometimes be articulated within their conversations. It was apparent

that, for these children, making the move from repeated addition

(which works quite well here up to about T=15) towards multiplication

did not occur naturally. This was surprising, because the format of

the task seemed well suited to affording that move: repeated trials

should serve to highlight the point where an existing strategy tended

to break down. Moreover, there was no time pressure forcing hasty and

unreflective decisions - the snake keeps moving repetitively until the

pupil is ready to make an estimate. There is also little doubt that

the children seemed highly motivated to improve on their last effort.

This limit on progress illustrates the first point I wish to make:

creating an *occasion* for extracting some new understanding may not

be enough. This is a point that has been most effectively elaborated

by Perkins (1985) in relation to classroom activities more generally

(but especially those based on computers). He cautions against too

easily assuming 'the opportunity does the teaching by itself' (p. 13).

The conception of the present programme exposes (in myself) a certain

naive faith in this principle. On some definitions the activity is a

drill. It confronts the pupil with a succession of discrete problems

converging on no particular creative endpoint. However, we suppose it

might offer more than the chance for sheer practice: the scheduling of

the problems and the visual representations used might encourage the

pupil to "stumble" into a firmer conception of multiplication.

Unfortunately, it seems they do not.

From some teachers' point of view this might be an attractive

programme: it is engaging and can be offered without the need of

supervision or support. Yet, this independence of operation is both

its strength and its weakness. Reluctantly, I concluded that in some

classrooms where it was used with enthusiasm it might as well not have

been used at all.

This brings us to a second observation about this programme. In two

of the classrooms, it was deployed intermittently in the computer

corner over a period of two school terms. It was judged to offer an

agreeable activity for the pupils, one that invited them to use number

skills in a playful manner. In the third class, its presence was more

conspicuous. On the walls were pictures of matrix-like patterns and

bendable snake-like models had been constructed from empty food

cartons. In discussing number topics with the class, it was not

unusual for the teacher to make use of blackboard illustrations that

echoed those in the computer activity. The difficult concept of a

prime number was referenced to the children's familiar (and

frustrating) problem of making a good guess for those long snakes that

only ever uncurled into one straight (1 x T) line. In short, for

these children, their experience with this simple computer activity

was drawn into the wider context of classroom life. There it was

mobilized to support the public discussion of number and to inform

various other creative activities.

The children in this class made greater progress on the computer

activity itself - although we can not make too much of this

observation as the difference in treatment was not a planned one in an

experimental sense. However, it is surely a persuasive idea that such

widespread classroom appropriation of the procedure and imagery of a

computer activity would make a difference to its impact. So, we

should be wary of our faith in the 'opportunity doing the teaching by

itself': with computers in particular, this faith may cause us to

neglect the effort of integrating the task into the public life of the


Lest in Chapter 1 I appeared too tolerant of drill and skill computer

activities, this present discussion identifies my own reservations

about computer-as-tutor. It is not that this software is

intrinsically suspect on the basis of some common design

characteristic. Nor that the principle of dense practice in some

problem domain must necessarily be avoided. The limitation of this

form of computer implementation is that it may cultivate a faith in

the self-contained effectiveness of such activity. Of course, this is

a species of criticism that might be applied to the support of any

classroom activity that is allowed to lose its context. However, it

does seem that this possibility might be especially real for the case

of activities supported by computers.

I am using this simple example to develop a point of view about

computer-supported learning: namely, that it is necessary to

incorporate it carefully into the collaborations that characterise

organized learning. Moreover, I believe that this is important to get

right, because many computer-based resources offer something quite

distinctive (and potentially powerful) as educational resources. So,

even the very simple activity described above exemplifies the

possibility of fashioning a rather novel experience. It allows a

rather distinctive kind of encounter with number: one that is hard to

reproduce in other media. The visual representation of number that is

achieved is actually no different to that which children may encounter

with standard classroom materials (Cuisinere blocks). However,

something distinctive is added by the incorporation of simple

animation and the opportunity for pupil interaction (number selections

and their evaluation). The number matrices move at a pace that

invites the discovery and exercise of new strategies for enumeration.

This, in turn, will periodically confront the pupil with the need to

make informed guesses (to estimate) - something that, traditionally,

is hard to cultivate in the early years of mathematical experience.

So, there is often something distinctive and powerful about the

experiences that computer-based activities can offer. The present

example illustrates this possibility within the most modest of

formats. However, it also illustrates another feature of computer use

in educational settings; namely, the danger of their dislocation from

a main stream of educational discourse (cf. Pelgrum et al 1990).

Pupils' enthusiasm during work on computers may easily mislead us into

thinking learning is proceeding in pace with engagement. So, in this

case, there was a notable discrepancy between some teacher's

judgements of what the children were doing (namely that it was a

helpful activity with which they were making progress) and the

researcher's finding which was that, often, it was of no apparent help

at all. The researcher has detailed records of what pupils were

actually doing - the trial by trial choices and estimates were

captured and stored by the software as part of a research exercise.

These records were not so easily extracted by the teacher and the time

needed to study them might not be available. Indeed, such feedback is

not normally incorporated into the design of commercial software and

teachers have, therefore, to rely on more informal glimpses of what is

happening with an activity.

This example should illustrate the broader sense of teacher-pupil-task

interaction that I am considering in this section. This is not the

intimate interaction that might occur in a sustained encounter between

a teacher and pupil(s) - as the latter worked at a particular

computer-based task. Much instructional communication is not so

intense as this. It comprises instead the reflection, review and

integration that teachers impose upon children's activities; this may

be done intermittently as those activities progress, or

retrospectively when closure on them has been reached (Edwards and

Mercer, 1987).

Such communication is more extended in that it can embrace larger

numbers of participants than is typically implied by

culturally-influenced research on ZPDs. It is more extended in that

it may come and go across a longer period of time: the social

interaction that organizes some particular learning activity in class

need not be restricted to the tidy closed session-like encounters

that often characterise research modelling of instructional discourse.

In our modest concrete example above, the teachers could interact

*in relation to* the computer in the more extended sense of them

explicitly drawing shared knowledge of a program into the community

life of the classroom. The computer activity then becomes (for some

classes) a resource that organizes discussion intermittently in time

and, perhaps, requiring talk that involves large groups or the whole

class. Surely, much social interactional work that is done in

classrooms is necessarily done at this level - albeit complemented by

the more intimate tutorial contacts that psychological commentators

are more comfortable with from their research traditions.

My concern in this section has been to illustrate what might be

necessary to create an effective teacher-pupil-technology

interaction: a social interaction *in relation to* technology. The

example suggests the potential of computer-based experiences when they

are fully assimilated into this social dynamic of classroom

interaction. It also cautions a limitation in what might be achieved

when the pupil-technology component of the exchange is isolated from

this dynamic. Much more empirical work is necessary to determine the

force of this caution. In the absence of such research, I shall take

a different approach. This will entail reviewing more general

arguments for the potency of classroom discourse: the proper

management of computer-based experiences might be inferred from

observations of talk organized around *other* kinds of learning

activity as it has been documented in research.





I shall discuss two perspectives on the mediating role of

teacher-organized talk. In each case a form of *continuity* is being

created for the participants. Firstly, there is a kind of lateral

continuity. This is required in respect of pupil activities that

might otherwise be left isolated as practical experiences - where, in

reality, they are conceptually related in significant ways. What is

addressed here is the problem of achieving transfer of learning:

allowing pupil understandings to generalize in important ways to new

situations. Secondly, there is a kind of longitudinal continuity.

This might be described as the creation of a kind of narrative state:

furnishing a recognized platform for the next set of explorations. It

arises in talk which is used to knit together the sequences of

disparate actions and observations that constitute some learning

exercise. Sometimes, such experiences may have been organized over

quite extended periods of classroom time and the integration is a

substantial responsibility. This form of continuity may be implicated

in transfer also, but it is more implicated in the empowerment of

fresh instructional talk.

In the two sections that follow, I will expand upon both of those

senses of the social basis of learning: the creation of both lateral

and longitudinal continuities. In doing so, I shall make references

to how computers might enter into such communications and, thereby,

become effective resources for the support of teacher-pupil

interaction. In the first of these sections, I shall consider more

closely the grounds for viewing transfer of learning as a

socially-organized achievement. This identifies the lateral

continuity that must be created. In the second, I shall consider

longitudinal continuity: the creation of an integrating "common

knowledge" within a learning community.

(1) Creating lateral continuities: transfer of learning


In discussing cultural psychology in the last chapter, we noted that

it was a theory emphasising the situated nature of human

understanding. Learning becomes, in Vygotsky's words, 'the

acquisition of many specialized abilities for thinking' (1978, p.83).

We also noted a price to be paid for doubting that generalized

thinking skills should have a key role in theorising: some other basis

is required for explaining how learners manage to transfer their

knowledge from one situation to another - for undoubtably they do.

The first thing to be said about transfer is that we may intuitively

exaggerate how easily it is achieved: reviews of research into

spontaneous transfer suggest it does not readily occur (D'Andrade,

1981; Detterman, 1993; Lave, 1988; Pea, 1988; Perkins and Salamon,

1987; Resnick, 1987). So, we must discover more about what has to be

done to make transfer happen. Current accounts of transfer are

strongly cognitive, their focus being upon the mediating influence of

private, mental structures. A recent example is one proposed by

Hatano and Inagaki (1992). They suggest that contextualised knowledge

is "desituated" (transfers) when the learner synthesises a certain

kind of abstract representational device: a mental model of the

relevant domain. This synthesis is made possible when learners have

enjoyed a particular range of encounters with the domain. But what

precipitates such mental modelling? Although a mental model may prove

a helpful way for us to conceptualise part of the process, the idea

must be complemented by some account of the origins of such models.

That is, some account of how children's concrete experiences in a

domain are best organized to facilitate the proposed representational

synthesis - the supposed basis of the transfer.

Some of what matters may reside in the structure of tasks that are

offered to encourage learning (for example, the pace or

predictability of action, or the nature of feedback). However, it may

be hard to judge whether a highly differentiated and flexible task

environment is better described as contributing to a single rich model

in the learner - or to the cultivation of a greater variety of

situated responses. One more straightforward kind of significant

supportive experience comes from the social environment of learning.

Understandings may be enriched in the sense proposed by cognitive

psychologists (and, thus, transfer more readily) if learners enjoy a

certain pattern of tutorial interaction with other people. One

pattern might involve pressure to articulate knowledge to others (see

Chi, Bassock, Lewis, Reimann and Glaser (1989), and see also further

discussion of peer processes in the following chapter). However,

another pattern of potentially useful social relations is closer to

the concerns of the present discussion. It arises from the supportive

intrusions into our activity that are made by those who are more

expert than us: occasions where teachers and other experts act to

impose a certain interpretative framework on our actions. This is a

proposal more actively pursued by some working in the tradition of

cultural theorising.

Rogoff and Gardner (1984) develop this point to help characterise the

very earliest learning experiences. They show how adult guiding

participation within infant and preschool problem solving can serve to

identify for a child links between contexts of novel and familiar

problems. The same point is made within the Laboratory of Comparative

Human Cognition's discussion of more formal learning settings (LCHC,

1983). They propose that the key to transfer will often lie in how

other people (who are more knowledgeable) do and say things that

identify the links between contexts (cf. Pea, 1989). Organized

environments for learning (say, classrooms) will expose us to an

interpretative layer of discourse that is imposed upon our activities

- teacher talk. Educational practice involves the provision of

distinctive tasks around which this is arranged. Within these tasks,

action can be organized and, then, tutorial interventions serve to

indicate for learners the overlap among them. Pea (1988) has

elaborated this view and juxtaposed it with traditional psychological

theories that suppose transfer is mediated by "common elements" within

the material contexts implicated. Pea argues that such "sameness" is

not intrinsic to things, and detected by us as such; sameness is a

sociocultural concept. It lies within category types the thinker has

appropriated in the course of socially-organized activity. Expert

participants in this activity contribute interventions that serve to

achieve this.

This conception of transfer invites more research in which such

proposed social processes can be properly exposed and understood.

Some movement in this direction is exemplified in the recent studies

of Newman, Griffin and Cole (1989): here an attempt is made to analyse

the sense in which the "same task" can be encountered in new

situations. Their conclusion is that socially-mediated processes are

central to how we discover this continuity: in particular, they

highlight the significance of teacher-mediated appropriations of pupil


A perspective of this kind may be appealing to practitioners, for it

identifies a crucial ingredient of educational experience as being

within their hands. It also has a special significance in relation to

the use of computers. I have argued above that this is a technology

with properties that allow it to be easily dislocated from classroom

life. If this happens, then the valuable concrete experiences that

computers provide may not be referenced within the discourse that

helps mediate transfer. I argued in the last chapter that the nature

of computer-based tasks readily *encourages* this marginalization.

How often is this actually the fate of computer activity in


Probably too often - given that the opportunity for children to use

the technology independently is seen as a favourable property. Some

commentators will remark positively on classroom computers because of

the opportunity they can provide to release a teacher's time and

attention (Clements, 1987; Fraser, Burkhardt, Coupland, Philips, Pimm

and Ridgeway, 1988; Lepper and Gurtner, 1989). These observations

relate to work in schools; the same approving view of computers is

expressed in relation to undergraduate teaching (Hague, 1991). From

extensive classroom observation, Eraut and Hoyles (1989) comment:

'..assigning pupils to work on computers allows the teacher to attend

to the rest of the class in peace, and to give more individual

attention to those pupils who are not working on computers'. Yet

Eraut has reflected elsewhere (1991, p. 203) that this is an

unfortunate trend: computer work, he argues, needs a great deal of

planning in relation to other activities. This is very much my own

point. Limited planning of work - and the loss of talk that exploits

the continuities and connections thereby created - may undermine the

breadth of learning that is supported by computer-based activities.

This possibility gains credibility when we think again of the

particular case of Logo projects. The vision of transferable skills

arising from Logo work appealed to many practitioners but its

realization has been elusive. There have been some reports of

effective transfer (eg., Clements and Gullo, 1984) but the consensus

is that such achievements are not easy (Simon, 1987). Pea and Kurland

(1987) have interpreted these difficulties in terms that refer to the

social context in which Logo is experienced. They propose that there

has been a serious neglect of the role that teachers must play in

organizing and interpreting the children's activity. They argue that

the impact of the Logo experience is correlated with the extent to

which there is an external participation of this kind. Others who

have reviewed research in this area have endorsed the idea that adult

intervention is important (Keller, 1990; Krendl and Lieberman,

1988; Noss and Hoyles, 1992) The history of Logo is one that alerts us

to the difficulty with which learning experiences generalize - without

social resources to bridge contexts. Logo pioneers may not have

denied the relevance of a culture of use but until recently

(Papert, 1987; Harel and Papert, 1991) this was never a dominant theme

in the promotion of the activity. As Hoyles (1992) notes, this

impressive resource typically risks being compartmentalised in at

least two senses. Bureaucratically, it risks being bolted onto the

curriculum; Psychologically, it can become an experience for pupils

that is unnaturally separated from a mainstream of classroom-based


Regarding the collaborations of teachers and pupils *in relation to*

computers, what has been said so far may be summarised as follows.

The technology has characteristics that allow its use to be easily

separated from the mainstream of class activity. To some extent,

this might be viewed as liberating: a self-contained quality that

allows teachers to focus their energies more intently elsewhere.

However, we should be wary of this seductive strategy for deploying

computers. In the present section, I have identified the role of

social interaction in promoting the "lateral continuity" of schooled

achievements: helping the transfer between different situations of

practice. The Logo experience reveals that even the most engaging and

ingenious computer environment can fail to penetrate pupils learning.

At least, this is what seems to happen for young children when more

experienced collaborators (i.e., teachers) are not quite closely

involved with the activity.

Moreover, Ryan's (1991) evaluation of a full range of computer-based

learning interventions identifies "teacher pretraining" as the most

significant predictor of the outcomes. This must also hint at the

central importance of teachers being able to cross-reference

computer work with other experiences: such options will not be

so readily available to teachers who are not themselves

confident with the focal activity.

There is much useful research to be done that clarifies the role of

classroom talk in the sense that I have framed it in this section.

We must surely accept that that social processes have *some*

importance in creating continuity between disparate activities: I

would submit that they are of central importance to this continuity.

In which case, computer-based learning is vulnerable to remaining

highly context-bound - unless more is done to integrate them into the

mainstream of classroom talk. One programme of empirical research

that lends further credibility to these conclusions (while not

explicitly addressing the same issues) is summarised in the next

section. There, I consider the social-discursive creation of more

longitudinal continuity. This research described is the work of

Edwards and Mercer (1987) who have studied and documented talk in

primary school classrooms. They demonstrate the extent to which such

settings are indeed saturated with interpretative, sense-imposing

teacher talk. However, they frame its significance more in terms of

what I am calling longitudinal continuity - rather than the lateral

continuity I have discussed here as constituting the achievements of



(ii) Creating longitudinal continuities: common knowledge


Common knowledge, as it will be defined here, is closely tied to the

pivotal concept of intersubjectivity. This is the concept that helped

us in the last chapter to identify a recurrent theme within cultural

accounts of instructional processes: namely, the human capacity for

projecting and interpreting mental states in others. In that

discussion, I argued that the potency of instructional talk

depends upon its exploitation of intersubjective understandings. In

what follows, I discuss more of what lies behind this achievement. In

particular, I consider how effective instructional communication is

facilitated by the prior construction of common ground. For

researchers, it is a challenging empirical task to determine how and

why we construct such mutual understandings; to determine what makes

us curious about uncovering common ground, and motivated to act in

ways that exploit its potential. So, just how this is achieved in the

special case of instruction is my special concern in the present


Outside of educational research, the study of common grounding is often

approached through attention to the management of conversations at the

moment-to-moment level (cf. Clark and Brennan, 1991; Clark and

Schaefer, 1989; Schegloff, 1991). For these researchers, uncovering

the nature of a shared context will involve studying something that is

done with language as it is used on the fly - talkers creating

continuity by reacting to whatever circumstances happen to arise as

their interaction unfolds. Extensive participation in routine

conversation equips us with resources that make these achievements

seem natural and spontaneous (see Forrester (1992) for a perspective

on how children master this during development). However, in the

circumstances of teaching, this is only part of what must go on. It

is the part that was discussed in the last chapter. The creation of

common ground for instructional purposes must be much more of a

contrivance: something constructed across sustained and orchestrated

patterns of talk. It is something that depends upon a conscious

investment of discursive effort; this effort being exercised over

extended periods of shared time and space. The consequent

achievements only become visible if we research beyond the

moment-to-moment level of conversation; if we concentrate on more

protracted structures of social exchange.

In the case of instruction, understanding the creation of common

ground may require particular attention to the more material context

of communication: the special environments within which problem

solving and learning gets organized. In other words, structured

interaction with material resources may provide participants with

important mutual reference points for their common grounding efforts.

This may be especially the case where joint problem solving is located

within media with distinctive structural properties - such as

computers, perhaps. In any case, to discover more of how mutual

understandings are typically created in classrooms, we must look at

the patterns of conversation that have been documented to occur within

them. I will turn to this next.

The empirical study of classroom talk is a relatively recent research

interest (see Cazden (1986) for a systematising review). If there has

been some neglect of the topic, this might reflect the

learning-through-*activity* emphasis of much contemporary educational

theory. In commenting on this situation, Edwards (1990) identifies a

common attitude that he sees exemplified in the dictums of the

(British) Nuffield Maths project: 'I hear and I forget...I see and I

remember...I do and I understand'. These are principles offered in

a spirit of defining good classroom practice. Edwards regrets the

prioritising of the experiential, activity-centred ideal. However,

this may not be a assessment of what the Nuffield scholars are

promoting. Are they implying that the words of teachers are good only

for hearing - and then forgetting? That gloss on the Nuffield

philosophy may be an exaggeration of orthodox views on talking versus

acting. Nevertheless, the orthodoxy may still need examining. There

may well be grounds for arguing that educational theory has downplayed

(not denied) the significance of things that get "heard" - teacher

talk - in the interests of prioritising the impact of pupils' "doing".

Perhaps some prejudices regarding instructional talk arise because of

a stark contrast conjured up between "telling" and "doing". In this

contrast, talking is framed as something indirect; a substitute for

the real thing. The real thing is acting to discover for oneself.

Certainly, wherever teaching is being organized, there may be a

tension of this kind to be identified and addressed. However, very

little routine instructional talk seems to be easily forced into this

simple contrast. In reality, such talk gets organized in *relation*

to pupil activity - not in contrast with it. Which is not to imply

that the reality of classrooms is one of always talking within

intimate tutorial dialogue (perhaps in the formats reviewed in Chapter

4). Yet, typically, actual talk may share the contextualisation of

those occasions and yet may not need to be analysed in exactly

the same terms. Some talk that is not part of an intimate verbal

dialogue may still be organized to enter into potent (dialogic)

relations with things that pupils are *doing*. Or it may be organized

to operate back upon such prior activity. In this manifestation, talk

that seems merely "heard" may still be playing an active part in

pupil's constructions of knowledge.

This idea is central to Edwards and Mercer's (1987) discourse analytic

work in primary classrooms. Theirs is particularly important research

to mention here. For, it reinforces our concerns regarding the need

for teachers to be *in contact with* pupils and computers. But it

also serves further to bridge the gap between the intimate tutorial

setting of typical ZPD research and the more diffuse circumstances of

classroom teaching.

Edwards and Mercer's approach is distinguished by its concern to

characterise classroom talk in terms of its role in defining

"context". Their approach to this is a timely one. Students of

language-in-use have only recently come to terms with the

incorporation of context into their analyses (Cazden, 1986). However,

in "contextualizing" an utterance, researchers may often appeal only

to features of the physical situation in which the talk is organized,

or to details of what has very recently been said. Context in this

sense is concrete and accessible. Indeed, this is convenient for

those researchers who are armed with audio and video recording devices

and who may be preoccupied with capturing the here-and-now.

Unfortunately, such professional conveniences may distract researchers

from considering whether what they capture is what they need; whether

what they capture allows them properly to situate the dialogue. The

understanding of what gets achieved with talk may require researchers

to access contextual information that simply is not included within

here-and-now records of this kind.

Specifically, what Edwards and Mercer propose is that we should

recognize the context of talk as incorporating *intermental*

achievements. Context is not just stuff-out-there, directly

accessible to all the participants and available for action replay in

our research recordings. It incorporates mutual understandings:

intermental by virtue of arising from whatever history of joint

activity is common to the conversants. If we now seek to understand

what guides the momentary formulation of particular utterances, we

must do so through attention to this backlog of "common knowledge".

The intimate nature of these context-building achievements might seem

to make them inaccessible to interpretative research. However, this

context of talk need not be private in the sense of being impenetrable

to understanding from outside. If it is possible to record

conversational strategies for creating and referring to shared

understanding, and if those strategies allow some systematization,

then we may make useful progress. An appeal of Edwards and Mercer's

formulation is that it focuses on the actual discursive moves

involved in creating common knowledge - particularly as such moves

might be exercised in more ritualized settings of communication. So,

in the setting of a classroom, we might identify the dialogue devices

that typically are invoked for this purpose; or we may be able to

reveal the ways in which such talk is most effectively interlaced with

pupil activity. Those aims are very much the concern of Edwards and

Mercer's own empirical work.

In situations such as classrooms, groups of individuals regularly

interact over extended periods of time. Their achievements may depend

on how effectively a sense of overlapping, compatible understanding is

constructed: Edwards and Mercer refer to this in terms of the creation

of "continuity". Continuity is the intermental context that is

developing through time. Their own empirical contribution is then

twofold. Firstly, they demonstrate the extent to which teachers *are*

actively engaged in creating such continuity: talk is relentlessly

deployed for that purpose. Secondly, they illustrate some of the

commonly used discourse devices that serve this purpose.

Regarding their first achievement, it is valuable to have documented

how much of teacher talk is invested towards the active construction

of joint understanding. Edwards and Mercer point out one

straightforward reflection of this agenda: their teachers were

generally reluctant to admit into class talk too much personal

material from pupils' own outside lives. Wertsch (1991a) has also

noted how instructional settings generally restrict discourse to

experiences that are bounded within them. The image that dominates

this research is of teachers actively commenting upon, elaborating and

posing questions around what their pupils have experienced as a

community. There is certainly plenty of pupil *activity* in these

classrooms and that is the feature that the teachers themselves were

most anxious to highlight and cultivate. What is easier to miss from

within the participant's perspective is how far this activity is

filtered through an interpretative and teacher-regulated discourse.

This takes us to Edwards and Mercer's second achievement: cataloguing

and evaluating the discourse devices that are put to work in this way.

These include techniques of organized recapping that allow the

creation of a shared *memory* of what had happened when pupils had

done things (cf. Edwards and Middleton, 1986). Also included are

techniques of cued elicitation that served to solicit an agreed

account of what was currently happening. As an observer, there may be

a natural inclination to judge this scene as one in which pupils are

merely having their dormant knowledge innocently extracted from them.

However, a more coherent account of events stresses that what matters

about this talk is the active creation of an agreed, shared

understanding of what the community experienced. The teacher is

summarising, challenging, questioning and so on, in ways that both

check current positions and update the evolving shared context - the

backlog of "common knowledge". This common knowledge is generative:

it becomes the platform for new understandings and new connections to

be made. A teacher's contribution towards the progress that can be

realized in a classroom depends in an important sense upon being able

to exploit this intermental creation.

Edwards and Mercer do not use this formulation directly to

address transfer of learning, in the way that the issue was discussed

above in relation to computer work. However, a distinction they make

between ritual and principled knowledge makes contact with the problem

of forcing understandings to go beyond the immediate settings of their

acquisition. *Principled* understandings are those that do apply

widely across settings and lead to greater reflective self awareness.

Edwards and Mercer's analysis highlights discourse practices that may

be more likely to cultivate such understanding: this analysis is

consistent with the hope that we may make progress on the transfer

problem through considering how classroom talk is related to the

organization of pupil's classroom experience.

In short, this is an analysis that both demonstrates the natural

penetration of teacher talk into pupil activity and also furnishes a

conceptual vocabulary to understand its strategic management.

Nevertheless, I feel there are two limitations to this account and

they each matter for our interest in educational technology.

Firstly, Edwards and Mercer's inspiration is to conceptualise the

context of classroom activity in terms that incorporate the

intermental achievements of conversation. This challenges our

inclination to prioritise the immediate and material setting of action

as being what mainly comprises its "context". Yet, their shift of

emphasis may have led them, in their own research, to swing the other

way: effectively neglecting how talk is contextualised within

material circumstances. Thus, they do demonstrate organizing

principles of teaching talk in relation to the conduct of specific

classroom tasks. However, there is little consideration given to how

the format of those tasks serve to constrain or facilitate the

particulars of the talk that emerges - how the talk is materially

situated. So what is documented in the research is not itself

contextualised in this spirit. It is one insight to note that the

institutionalization of social relations in the form of schools

encourages certain styles of sense-making talk. But that insight

needs to be elaborated: capturing how particular formats of activity

afford particular realizations of this talk. Such a focal orientation

to settings themselves is rare within the literature. (One example is

evident in the work of Cook-Gumperz and Corsaro (1977). They have

made a clear case for attending to the physical (play) environment as

creating or denying discourse possibilities involving children and

adults in preschool.)

I have tried to make this point elsewhere in relation to computers as

artefacts within children's social, rather than cognitive, development

(Crook, 1992a): suggesting that we could begin more of our analysis

from "things themselves" and consider what kind of social interactions

they afford. Here, I am suggesting that the same strategy could apply

in our thinking about interactions organized to support learning. The

key question might be: What interactions can arise between teachers

and learners if they chose to interact around the particular contexts

of computer-based activities? I suggest this is a point of entry for

empirical work: research that might characterise a social context for

computer use incorporating pupils, teachers and computers themselves.

This entails defining the possible collaborative interactions of these

participants *in relation to* the technology. There is very little

observation organized in this tradition; although an example might be

one study of secondary mathematics teaching that shows how the

presence of computers can create a less teacher-led, more discursive

style of class interaction (Fraser, Benzie, Burkhardt, Coupland,

Field, Fraser, and Phillips, 1984). This finding is rather welcome -

in view of my second concern regarding Edwards and Mercer's


This second concern is captured in a critique of the work by Cazden

(1990). Edwards and Mercer's account dwells on the teacher-dominated

character of classroom talk. In highlighting the frequently directive

nature of this talk, it fails to present a sufficiently constructive

alternative agenda. Their analysis dwells on identifying limitations

in the way discourse is deployed: they are particularly concerned with

discursive moves used by teachers to *control* the development of

continuity and context; or with teachers' efforts to maintain the

"fiction" of extracting what children already know; or with their

management of the dilemma of respecting self-discovery while ensuring

the curriculum is "properly" learned. Edwards (1990) comments about

these observations: 'Typically this dialogue turns out to be no simple

negotiation between equals but a process that is dominated by

teachers' concerns and aims and prior knowledge'. In short, the

emerging picture is a somewhat gloomy one. It is thin on examples of

what might be considered effective or creative practice.

The implied shortfall concerns how far understandings are "negotiated"

within talk. If Edwards and Mercer's summary picture is persuasive,

then one cause for concern will be its demonstration that teacher talk

is often reinforcing a model of knowledge that implies a ready-made

"grown up" version: this is the one to be learned. Teacher discourse

is frequently oriented towards creating a common knowledge that

reflects an "official" story. Faced with a perceived obligation of

this kind, the challenge teachers may feel is one of working out how

the official versions may be discretely precipitated from joint

activity. Unfortunately, the interventions that might be necessary to

achieve such innocent extractions can sometimes form a poor

representation of how knowledge is actually developed and negotiated

in investigative contexts outside of the school. So, even work in

natural science (which might seem the most straightforward area in

which to promote official versions) should enjoy careful discursive

management. Steps might be taken to ensure it is experienced as

knowledge derived in an atmosphere of conjecture, debate and argument.

This version of the *activity* of science was rarely witnessed within

the discourse of Edwards and Mercer's classrooms.

We, therefore, have an account of classroom talk that offers us the

useful framework of common knowledge and its creation, but which

presents a cautionary view on current practice. Edwards (1990)

suggests one way of advancing from this situation. He reiterates

Piaget's observations regarding the cognitive developmental

possibilities of talk among peers. Because of the equality of status

between pupils themselves, pupils working with their peers (rather

than their teachers) can create certain richer possibilities. They

can create arenas for the precious experiences of motivated argument

and reflection. Such situations of collaborative learning might

represent better conditions for acquiring the rhetorical skills of

knowledge-building than the conditions normally experienced within

teacher-dominated talk.

The notion of common knowledge will be an influential one in what

follows. I have expressed two limitations to the account arising from

Edwards and Mercer's (1987) work. Concern about both of them will

inspire empirical work to be described in subsequent chapters. There

I shall give more consideration to how the structure of computer tasks

constrains or extends the discourse organised in relation to them. I

shall also consider the particular issue of peer interaction

as it is mediated by computers - hoping that within such arenas we

might reveal interactions that allow learning through a more

negotiated form of discovery. This respects Edwards' own suggestion

for dealing with the more constraining nature of teaching discourse -

the tendency for official versions of knowledge to be imposed through

such talk.


Common knowledge and intersubjectivity


In the last Chapter, I reviewed some ways in which Psychologists

have characterised instructional talk - considering whether those

perspectives implied it might be simulated by technology. I suggested

that the various devices prominent in such talk could be integrated by

reference to the concept of intersubjectivity. So, success within

an intimate, tutoring style of discourse seems to depend upon the

participants cultivating an intersubjective attitude. In the present

chapter, the discussion of instructional talk has shifted towards

considering a more diffuse version of this activity. Here we have

considered exchanges that are not necessarily so closely coordinated

with activity-in-progress. They are likely to refer to prior events

and experiences. They are often communal in nature: directed towards

groups of pupils rather than individuals. I suggest that such talk

may also be understood by reference to the concept of

intersubjectivity. The point of such a move is to consolidate an

argument for firmly associating learning with the deployment of

intersubjective attitudes. If this association is granted, then it

must limit how far computers can be expected to reproduce the

interpersonal dimension of educational practice.

The key to the intersubjective nature of classroom talk (in those

formats characterised above) is its natural achievement of common

objects for attention: shared resources that the participants

understand to be shared, and which can serve as platforms for their

further communication. Edwards and Mercer's characterisation of this

achievement for classrooms is a novel and useful contribution to our

understanding of the discourse of learning. However, their

formulation of the common knowledge idea is not entirely novel. The

technical terminology is borrowed from established work on meaning and

human communication (eg., Lewis, 1969). I shall identify briefly

those parallel traditions of theorising that are relevant to my

concerns here. This will allow us to acknowledge that the "level" at

which Edwards and Mercer's research is directed remains distinctive:

it is a level particularly appropriate to the understanding classroom

practice and, thus, provides a good model for future research.

The conversational analysts Clark and Brennan (1991) capture the

spirit of what has been said above about classrooms in their general

observation: 'All collective actions are built upon common ground and

its accumulation' (p.127). However, common grounding activity -

within instruction or within any form of discourse - is not a unitary

phenomenon. It can be said to be managed at more than one "level", or

by the use of more than one set of communicative resources. This

notion is well captured within distinctions made by Krauss and Fussell

(1991). Firstly, they identify resources that may be available in

the form of knowledge about social category memberships. This

comprises existing beliefs and expectations about others that are

derived from general understandings of relevant social status or

social practices (eg. that which might follow from understanding that

someone is "a New Yorker"). Secondly, there is a further level of

resource; one that arises from the flow of direct, ongoing interaction

with a conversant. This is information gathered and exchanged "on the

fly". Conversational analysts have studied this second level of

communicative resource as it is created and discovered within everyday

talk (Clark and Schaffer, 1989; Sperber and Wilson, 1986). My own

discussion of the management of instructional discourse (Chapter 4)

addressed this same phenomena in the particular context of teaching

and learning.

The level at which Edwards and Mercer's research is pitched seems

somewhere in between those identified above - and one peculiarly

significant to organizational life such as that which arises within

classrooms. Their research defines communication that depends on

direct experience of social interaction (rather than abstracted social

category knowledge). Yet what is being deployed in this communication

does extend into a history of the conversants interaction; one that

goes beyond the momentary events of the present conversation. In

other words, the common knowledge analysis concentrates on a

longer-term continuity: a more protracted (but interpersonal)

construction of shared understandings. The analysis also helps us to

see this processes as an achievement motivated and guided by a

particular purpose - in this case, the management of learning. Thus,

within that important context of education, we may start to identify

the particular conversational resources that are available to make

communication effective.

Whatever the level research is pitched at, the idea of capturing the

control of communication in terms of mobilising intersubjectivity

seems appropriate. However, attending to an orchestrated setting like

a classroom may alert us to the case for framing communication in

terms of a dynamic, interactive achievement. I believe this is more

helpful than metaphors that tend to reify what is involved (such as

"common *grounding*" (Clark and Brennan, 1991) or "an

interpsychological cognitive *object*" (Newman et al, 1989), or even

"common knowledge". Such metaphors tend to reify the achievement.

They may even cultivate speculation about programming data structures

that might represent these cognitive objects. Whereas the real

problem for simulation is not one of reproducing some static cognitive

entity. It is a problem of capturing the sensitivity, empathy,

projection and improvisation that constitutes communicative

*interaction*. And that, as has been argued for instruction at the

more momentary level of conversation (Chapter 4), is an unlikely


This discussion of common knowledge indicates how human

intersubjectivity may be regarded from two perspectives when we

consider educational settings. On the one hand, it resides within the

prompting, monitoring, intruding talk that makes up

instruction-in-progress. On the other hand it resides in talk and

action that serve to create that which becomes held in common - and

known by the participants to be held in common. How might the impact

of access to such resources be optimised? Our answers are likely to

be framed in both motivational and cognitive language. Learners must

be *motivated* to adopt intersubjective attitudes. This may be a

question of cultivating links to their spontaneous goals and

priorities as they may be formed elsewhere; it may also involve

creating more equable opportunities for them to contribute to (or

negotiate) common knowledge. The more *cognitive* dimension to

optimising these resources requires us to study carefully the way

discourse and activity is coordinated within authentic instruction -

as exemplified in some of the socio-cultural empirical work discussed

here. In either case, there is a challenge to integrate new

technology with these practices, rather than allowing it to be

subverted by it. This is a challenge requiring us to attend to the

particular manner in which computers can become an activity setting in

relation to which common knowledge is effectively negotiated.








In this chapter and the previous one, I have considered two senses in

which computers might enter into the social fabric of educational

activity. In Chapter 4 I evaluated the idea that interaction with a

computer might be programmed to reproduce the social character of a

face-to-face tutorial dialogue: social interaction *with* computers.

This plan seemed too ambitious. We feel this as soon as we reflect on

the nature of the human conversation deployed during instruction.

Thus, I dwelt upon characterising such instructional talk; arguing

that it is grounded in the distinctively human capacity for

intersubjectivity - and that teaching involves the organized

management of that intersubjectivity.

Such management is partly a question of timely intrusions into

learning-in-progress. I invoked Suchman's analysis of "situated"

action to characterise the necessarily improvisatory and versatile

nature of this achievement; arguing that it is unlikely to be

simulated by computers. However, the management of intersubjectivity

entails a further set of discourse resources; these being concerned

with the more protracted building of socially-shared cognition. Such

a proposal arose from arguments developed within the present Chapter,

where social interactions *in relation to* computers have been

discussed. This involved acknowledging that the meaning of some

teaching utterance is rarely to be located in, or made manifest

through, its simple surface features - as if such meaning were

something to be generated by a rule-bound system of the sort that

computer programmers would seek to construct. Effective instructional

talk will be contextualised. Indeed, its utterances will derive their

impact from the skill with which speakers build upon mutually (perhaps

laboriously) constructed shared understandings. This richer,

intermental sense of "context" as defined for instructional

communication is not something to be captured in computer programs.

(Perhaps the intellectual work that tutorial language must be made to

do in order to create and exploit this context is at the root of what

can often make "teaching" such a peculiarly tiring activity.)

Unfortunately, there are grounds for fearing that computers remain

vulnerable to exclusion from this enterprise. I have drawn attention

to features of the activities they support that encourage this

dislocation from the classroom community. In particular, computers

have design features that readily encourage a pattern of use whereby

the activity is dissociated from the core of classroom life - and

where teachers may less readily engage with what pupils are doing.

This is a real problem, given important ethnographic claims that

discursive interventions within (other) pupil work are a persistent

feature of what teachers normally achieve (Edwards and Mercer, 1987).

Such interventions serve the important function of creating continuity

of experience. This continuity constitutes a "common knowledge" that

forms the platform for yet more new discourse and new activity.

Teacher-pupil collaborations in this sense may be less easily fostered

in contexts of computer work.

Because relevant research is scarce, there is still little to be said

on the question of how best to integrate teachers, pupils and the

particular settings created by computers. To pursue this theme here

would demand rather piecemeal considerations of particular

computer-based activities and their potential for incorporation in

class talk. This would make our discussion veer too much in the

direction of curriculum issues of a rather specialist nature.

However, I do intend to return to the themes raised here regarding the

pupil-teacher creation of socially-shared cognition. The discussion,

in a later chapter, will consider how computer-based activities can

offer some *generic* support to such efforts: how the technology can

offer an infrastructure within the classroom that underpins the

creation of various useful continuities - rather than undermines it.

This will involve us in considerations of how communication can be

mediated *through* computers rather than simply *with* them or *in

relation to* them. I shall outline a configuration of computers that

goes some way towards avoiding the breakdown of community-based mutual

knowledge - as it might otherwise occur in relation to computer work.

First, however, I wish to consider the other important sense in which

social exchange can be organized to involve this resource. This

discussion will centre on classroom "collaboration" as more

traditionally understood. Pupils may be invited to interact with each

other around computers: this technology may support peer-based

collaborative work. In fact, we shall see that the arrangement of

peer interaction *at* computers has attracted considerable

practitioner and research interest. Moreover, our discussion of

teacher-pupil talk in the framework of common\n knowledge did converge

on a claim by Edwards (1990) that classroom peer interaction should be

seen as a promising solution to some of the more problematic aspects

of teacher-led discourse. It might prove to be the forum in which the

*processes* of intellectual discovery and investigation are most

naturally modelled and made accessible to learners.

In the next chapter, I shall first introduce the literature that does

exist on this topic. I shall then review existing theoretical

frameworks for making sense of this literature. My own preference is

to emphasise the conceptual vocabulary of intersubjectivity that I

have introduced above. In a subsequent Chaper, I shall present some

class observations of my own to help develop that theoretical