14th October 1993





In the previous chapter, I outlined a particular theoretical approach

to understanding cognition: a "sociocultural" approach. I believe it

helps clarify psychological issues arising within educational

practice. I suggested that this framework could have a particular

value for addressing issues relating to the educational use of new

technology. However, it is not the only source for such inspiration:

other theoretical frameworks in Psychology also claim insights into

these same practical problems. In the present chapter, two such

psychological perspectives that have been attractive to

educationalists will be identified and discussed.

The aim of this Chapter is partly to bring the socio-cultural view

into sharper focus, by contrasting it with two traditions of

psychological theorising that are more established and may be more

familiar. Each of these presents a distinctive perspective on

understanding teaching and learning. Moreover, they each have

influenced the deployment of technology within education. Evidently,

the nature of that particular influence will be of special concern to

us here. The first of these theoretical frameworks corresponds to the

mainstream of cognitive psychology - I am choosing to label it

"experimental cognitive psychology" to capture a preference of

methodology that reflects a distinctive worldview. The second is the

Piagetian constructivist theory of development: one that has been

particularly influential within educational and developmental




The term "cognitive psychology" describes a broad church. Any

critical comparisons made here with cultural approaches do

not apply wholesale. The modern term "cognitive science" is hardly

more satisfactory as a target: it seems to embrace all sorts of theory

- including some very much in tune with the cultural ideas described

in the last chapter (eg., Hutchins, 1991). Nevertheless, I do feel

that within all this variety there is a general direction of

thought dominating current cognitive psychology - especially where it

is preoccupied with the experimental method. Perhaps another phrase

that captures well this tradition is "human information processing".

To some cognitive psychologists this phrase may sound old-fashioned.

Their research literature tends now to refer more to "symbol

manipulation". However, modern cognitive writers slip easily enough

into using the "information processing" vocabulary of the core

metaphor (eg. Vera and Simon, 1993). It still effectively captures

the private, computational nature of their approach - and that is the

key to any contrast with cultural perspectives.

The idea of defining and studying a mental "mechanism" dedicated to

information processing has enjoyed great appeal among psychologists.

With it, intelligence is understood in terms of managing a flow of

information from sensory systems, and from representational devices

implicated in the "coding" and "storage" of experience. Recent

advances in information technology itself, have encouraged computation

metaphors as a basis for modelling the cognitive "apparatus" (eg.

McShane, Dockrell and Wells, 1992). The work of Newell and Simon is

seminal in defining this agenda. The link with functions performed by

computers is easily identified in their writing:

The theory posits a set of processes or mechanisms that produce the

behavior of the thinking human....The processes posited by the theory

presumably exist in the central nervous system; they are internal to

the organism...Our theory posits internal mechanisms of great extent

and complexity, and endeavours to make contact between them and

visible evidence of problem solving.

(1972, p.9-10)

It is true that these concepts are under debate given new developments

in cognitive science (such as connectionism), but there remains a

widely-shared focus on a mental, symbol-manipulating apparatus. An

apparatus for encoding, symbolising and decoding - the computational

mechanism within our heads, if you will. In such a vision, culture

and socially-organized experience are important, but they are often

considerations to be grafted on afterwards. They become the

independent variables of research. Their possible relevance to

something central, basic and cognitive easily becomes an issue to be

explored later, in experiments designed to "control" such things.

This language of cognitive psychology has intuitive appeal. More so

perhaps, as we have become accustomed to ascribe "intelligence" to

computers themselves - an analogy encouraging a whole language of

explanation common to minds and machines. Such a language also offers

some sense of assurance to the seductive enterprise of *measuring*

intelligence: it seems credible that we might check the fine tuning of

this apparatus by administering suitable cognitive *tests*.

So, I am identifying characteristics of modern cognitive psychology

that are appealing, but they are in some tension with my preferred

cultural approach. The principle conceptual tension relates to

defining appropriate units for analysis. Within experimental

cognitive psychology this has generally been the individual - a

circumscribed mentality typically examined apart from any embedding in

socio-historical context. Only recently have some cognitive

psychologists considered that intelligent activity may need to be

analysed as something achieved through coordination with an external

environment of supportive intellectual resources. Hence, recently,

there has emerged an attempt to model such coordination in terms of

"distributed artificial intelligence" (eg., Gasser, 1991). By

contrast, within cultural psychology that mentality has always been

seen as inherently continuous with an historical, technological and

social context. Intelligence involves mediated action and any

conceptual vocabulary for studying it must accommodate mediational

means. There is also a methodological tension between these

approaches. Cognitive psychology has tended to practice a

laboratory-based style of science. Psychological processes are

abstracted for controlled study: "context" may be only conceived as a

catalogue of variables, any of which might be manipulated in such

settings. Cultural psychologists, on the other hand, are uneasy with

the controlled experiment (although, admittedly, they may still be

seeking a distinctive empirical strategy of their own).

The methods and analytic vocabulary of cognitive psychology have

furnished real understandings. Many psychologists influenced by

cultural alternatives are certainly happy to admit this (eg. Greeno

and Moore, 1993; Wertsch, 1991b). The approach has allowed progress

on at least some sorts of practical concerns. Surely this is so, or

it would not still be flourishing? Yet while there may be real

progress arising from this cognitive tradition, the reach of its

insights may be less than the investment might lead us to expect. In

my own view, the mainstream of cognitive psychology remains a

tradition of theorising that narrows our perspective on psychological

functions. It is particularly constraining when we turn to address

the needs felt by educational practitioners: the theory is weak in

addressing issues of learning or cognitive *change*. Arguably, this

has meant that cognitive psychology has had rather little impact on

educational practice (Saljo, 1987), or it has had localised and,

sometimes, rather distorting effects. Desforges' (1985) observation

is pertinent here: he notes that most influential psychological

theories of learning have been formulated with no empirical reference

to what happens in classrooms. This is strange, given that classrooms

are the places where learning is explicitly organized. In other

words, these theories refer largely to conceptualisations of the

abstracted human cognitive system. One theme that is neglected in the

laboratories within which such theories are generated is the situated,

interactional nature of teaching and learning. That social quality of

education is of special interest to the cultural approach and

underpins the discussion of computer use pursued in this book.

Earlier, I used the case of literacy to illustrate features of a

cultural psychological analysis. The same case serves to mark some of

the difference between a cultural and a cognitive approach. For

cognitive psychology the topic of literacy largely reduces to theories

of skilled reading. Conventional "psychological" accounts of reading

are often couched in terms of information decoding and processing

mechanisms. The processes referred to by these models tend to dwell

on a "bottom-up" analysis of reading: approaching it as an activity

founded on fundamental information handling skills and to be

cultivated through increased understanding of the nature of those


Such empirical work is elegant and persuasive and it does advance our

understanding of an extraordinary childhood achievement. Yet it is

incomplete. It fails to capture reading as an activity: as an

achievement that is located within a framework of

culturally-determined possibilities. Reading as we actually

experience it is carried out for purposes, and the written word enters

into and re-organizes our interactions with the world in a wide

variety of ways. Making it accessible to novices will depend upon

some sensitivity to an elaboration of its place within the novice's

world. But there is little suggestion within information processing

theories that, for example, reading difficulties are approachable in

terms of this cultural context (Brown and Campione, 1986); or little

characterization of reading as an activity involving mastery over

cultural resources (Heath, 1983); or little implication that we think

"through" the written word - that our modes of reasoning reflect

enforced contact with this medium (Scribner and Cole, 1981; Olson,

1986). This is not to say that this broader, cultural view of the

achievement does not inform a lot of what is done about developing

reading in classrooms. I think that it does. The point is merely

that any such insights will rarely have their origins in influences

from traditional cognitive psychology.

I believe the limitation of much cognitive psychology resides in an

agenda that is too focussed and constraining. The approach creates a

determination to conceptualise cognition in terms of general

symbolic processes (say, relating to attention, memory or reasoning),

and to investigate such processes in the environment of laboratories.

In these impoverished settings, contexts can be controlled such that

properties of cognitive processes can be documented in their

abstracted states. This is not to deny the reality of central

cognitive processes that might be shared by all members of our

species: characteristics arising under common circumstances of

biological and cultural heritage. Neither do I wish to deny that it

might sometimes be informative to study these in abstraction.

However, by itself, this is a very limiting empirical agenda. The

richness of intelligent action is not comprehensively expressed in

terms of central representational processes. It surely arises as a

consequence of the cognitive subject's participation in episodes of

interaction, organized within diverse cultural settings. Human agents

are ceaselessly coordinating with their environments: so the nature of

the environment and a description of this dynamic exchange needs to be

embraced within any analysis of intelligent action. Experimental

psychology has fared poorly in capturing and understanding this

variety - say in ways that are of significance to educational


So, the legacy of contemporary psychological theorising seems to

include a strong faith in the possibility of content-free mental

processes. Consider a particular case: many psychologists study human

*memory* through observation of how their experimental subjects cope

with lists of words. The enterprise assumes a cognitive function of

memory can be usefully "purified" in this way - isolated for

exploration in parametric studies. Thereby the components of a memory

system are abstracted; the assumption being that they are normally

obscured from us by the "noise" inherent in everyday circumstances of

remembering. This approach exemplifies the view that refined

experimental materials can help segregate and expose certain core

mental functions. If, however, one adopts the cultural perspective of

recognizing that thinking is mediated through artefacts and social

practice, then the device of a word list seems very constraining. To

be sure, there are culturally-organized situations where such

remembering might be valued as a system of activity, but the minor

significance of such circumstances bears no relation to the popularity

of the word list device in memory research.

Not only is the specification of context-free cognitive functions an

inspiration for psychological research, such an idea also underpins a

strand of everyday educational thinking. There is a common sense view

about learning that seems to regard schools as sites for the

transmission of general purpose intellectual skills. Yet this

perspective may require critical consideration. Against it, for

example, Lave (1988) offers a vigorous critique of schooling, in which

she questions a determination to teach cognitive skills as if they

could be disembedded from the particular routines in which they are

exercised. The tradition of class tests reflects this - situations in

which the testee's thinking is deprived of conventional support from

material resources or from classmates. The academic examination has

evolved as a highly ritualized and condensed ordeal conceived in this


So, I am characterising an influential mainstream of cognitive

psychology as being closely concerned with isolating human information

processing capacities; concerned to characterise a private mental

space wherein the manipulation of symbols serves to guide action. How

would we expect such an analysis of cognition to interpret and

influence the place of new technology in education - and how would it

differ from a culturally-influenced approach? Some observers conclude

that the practical contribution of cognitive psychology in this area

remains modest (Beynon and Mackay, 1993; Laurillard, 1987). I believe

its influence may be visible in very general terms - in terms of a

broad theoretical orientation to computer-based practice - but less

visible at the level of specific software design.

Yet, a greater influence at just this level might be possible. I have

stressed that an information processing analysis has its strengths.

The issue is more one of its scope. So, for example, where a

significant part of a learner's task is to decode and manipulate some

symbolic system (say, in relation to the written word, mathematical

conventions, music etc.), then we might expect a cognitive

psychological analysis to be helpful in designing certain kinds of

challenging computer-based experiences. Similarly, the considerable

volume of research concerned with the acquisition of reading skills

might be expected to inspire a certain class of computer-based

learning resources. In fact, this does not seem to be happening on

any scale.

On the other hand, I believe there has been quite a strong *indirect*

influence from the cognitive tradition of psychology. It is indirect

because it tends to legitimise (without always being explicitly

invoked) certain common sense conceptions of cognition: in particular,

a faith in content-free thinking "skills". This, in turn, tends to

encourage a certain kind of educational purpose in relation to

computer-based learning. At the same time, the approach has

discouraged attention to the social fabric that is created during

learning. This is a context that it seems easy for most researchers

to overlook, yet which is of key importance to a cultural perspective.

I shall take up this theme in a later section of this chapter: where

each approach is considered in relation to its perspective on

computers for learning. First, however, the approach of cultural

psychology can be usefully contrasted with the second influential

movement within cognitive developmental theory: Piagetian

constructivism. I shall suggest similar limitations to this approach

as a comprehensive perspective on computer-based learning.






Educational practice has not only been influenced by theory within

*cognitive* psychology. Developmental psychology has also been

influential. In fact, it is widely accepted that psychological theory

has had its greatest impact on education through the developmental

work of Piaget. In this section I shall make a contrast between the

Piagetian "constructivist" approach and the cultural theory discussed

earlier. My main concern will be the priority the different theories

assign to autonomous discovery versus guided intervention as

underpinning the growth of understanding.

Piaget's work offers us a number of ideas regarding educational

practice that we now virtually take for granted. From him we derive a

strong sense of the distinction between development and learning; in

this scheme, what is termed development has an inwardly driven quality

and it must proceed in advance of formal teaching. Indeed, its

progress must underpin any strategy of organized teaching. The course

of such development depends upon things the child discovers

independently. Moreover, these discoveries emerge from opportunities

to act upon the material world and to reflect upon the consequences,

conditions and outcomes of such actions.

One feature of Piagetian theory echoes a theme identified above in

discussion of mainstream cognitive psychology: there is a commitment

to "central process" cognitive functions over more situated,

context-specific achievements. Piagetian researchers are not so fond

of laboratories as cognitive psychologists. However, their research

remains vulnerable to the criticism of neglecting context. Their

account of childhood reasoning fails to capture the senses in which it

is located in a social and cultural setting (Donaldson, 1978). I will

not review this issue again but, instead, focus on a further feature

of Piagetian thinking that is in tension with cultural perspectives -

one that has attracted more conspicuous unease in recent critical

commentary. This is the sparse reference to social factors in

Piaget's account of development.

It is frequently claimed that Piaget neglected social influences on

cognition. There is some injustice in such claims: at least, the more

vigorous of them. Piaget did not overlook interpersonal aspects of

cognitive development. Indeed, his early treatment of egocentric and

sociocentric thinking (Piaget, 1926) might be said to anticipate more

contemporary "social" perspectives on cognition. Nevertheless, it is

fair to claim that this concern played no significant role in the bulk

of his later theorising about cognitive development. Moreover,

although his concern for the social may be real, Piaget has been

largely interpreted *as if* his theory did not accommodate

interpersonal influences. Thus, for example, computer-based learning

materials inspired by Piaget may chose to neglect consideration of the

social context of their use - this is no less significant a trend

because it is based on interpretations of Piaget rather than an honest

appraisal of his ideas.

In contrast, what seems to be best known about Vygotsky and the early

socio-cultural theorists is that they promoted a view of development

strongly stressing the impact of social events. Moreover, this was

not simply a plea for more social psychology - at least, not if that

merely implied more attention to "social variables". Vygotsky was not

advocating a social psychology in the familiar sense of an

inter-personal psychology, one that reduces social experience to the

consideration of exchanges between individual social actors. His view

has us more thoroughly immersed in the social nature of life. Whereas

Piagetian and other psychological theory tends to start from the

individual and derive accounts of the social; Vygotsky proceeds the

other way round. For Vygotsky the social dimension of conscious

mental life is fundamental - 'individual consciousness is derivative

and secondary' (Vygotsky, 1979, p. 30).

At first encounter, Vygotsky's alternative might seem to have a

natural appeal. It prioritises the role of people (parents and

teachers, for example) in supporting the child's psychological

development. However, highlighting the role of other people may

necessitate querying another appealing perspective: one that puts

children themselves at the centre of their own development. This is

very much a view we associate with Piaget.

The nature of cognitive development in Piaget's scheme of things is

attractively captured in the term "constructivism": children are

actively "constructing" their understanding of the world. Surely we

will feel an affinity towards such a view. It offers us (personal)

responsibility for our psychological nature; it does not leave us

passive victims of social forces. In this spirit, Wood articulates a

widely-felt caution when warning that we should be wary of attributing

to 'the effects of instruction what is properly an achievement of the

child' (1989, p. 59). But as Wood's own work elegantly demonstrates,

any simple dichotomy of instruction and autonomy is misleading. In a

cultural view, the term "social" is used to help us see beyond those

(important) interpersonal activities that characterise direct tutorial

instruction. It identifies the fact that any material environment of

learning will manifest a structure reflecting some particular

socio-cultural history. Pupils learn interpretative practices that

have this social grounding. The appropriation of such practices may

not be exclusively within the contexts of face-to-face interaction.

One focal point for confronting this tension might be Piaget's

compelling account of the active, constructing infant. In "Origins of

Intelligence" (1953), Piaget directs our attention towards the energy

and curiosity of children in their first two years of life. We are

reminded of their relentless exploration of the material world, of the

energy invested in acting upon it. In contrast, Vygotsky makes

persuasive use of the same infant explorations to illustrate his own

contrasting view: a view of early cognition as socially-mediated. He

highlights within this early period of development the basic process

whereby children come to access mediational means. Vygotsky notes

that the infant's act of grasping is re-mediated into a cultural tool

(the sign of pointing) by virtue of how the social environment reacts

to this simple activity. This was a perspective to be widely adopted

later in accounts of the social context for language acquisition (eg.

Lock, 1978). Vygotsky's interpretation of such events renders a

re-reading of Piaget's careful natural history a curious experience.

One becomes sensitive to a strange neglect of the structuring presence

of the observer himself. The parent/researcher Piaget locates and

manipulates domestic bits and pieces around which the playful

observations are made - yet, fails to incorporate an account of that

mediation in the final theory.

Perhaps there is something of a projective test entailed in witnessing

infant behaviour. One observer might be most impressed by an

autonomous and creative exploration. Another by the manner in which

the infant's environment is structured to afford particular

possibilities, and by how the inevitable presence of adults serves to

create structure through joint activity. (This same projective

quality might apply to our experience of witnessing classroom life.)

In reflecting on theoretical perspectives that help us think about

cognitive development, Goodnow (1990) expresses a contemporary

judgement on Piaget's vision of the world: '....essentially

free-market and benign. All the information is available. You may

help yourself and act for yourself. The only limitations are imposed

by the nature of your own abilities - the extent to which your schemas

or logical structures allow you to take in the information' (p. 277).

This is a characteristically modern view on the Piagetian tradition.

It regrets a failure to recognize ways in which the child's

environment is culturally structured and, in particular, ways in which

socially-organized encounters serve to constrain and afford activity

in the Piagetian "free-market" environment.

Yet, there is also a sense in the contemporary literature of wanting

to reach some compromise with Piaget. In particular, the powerful

motive to act upon the world that is implied within constructivist

accounts may be too important to risk losing sight of. We sense this

in the writing of some current cultural theorists. So, Scott et al

(1992) refer to "cultural constructivism" and in the Introduction to

his book "Culture and cognitive development", Saxe (1991) is able to

cite both Vygotsky and Piaget as working in the constructivist

tradition. Finally, the concept of "co-construction" has found favour

recently, apparently to capture a potent mix of social engagement and

creative exploration.

I am persuaded that Piaget's constructivism does fail to come to terms

with the inherently social nature of cognition. However, I am also

wary that a pendulum of opinion may now swing too carelessly towards a

version of the "social" view that is excessively narrow. Moreover,

this narrow view can not thrive as an exclusive alternative to

unadorned constructivism. Earlier in this chapter I contrasted a

societal with a social interactional interpretation of cognition's

social nature. My concern is that only the social interactional

interpretation will be recognized and, for this reason, it will be

assumed that all learning has to be organized within contexts of

face-to-face social exchange. Expressed as a crude theoretical

confrontation: the social encounter within Vygotsky's zone of proximal

development is taken as the *exclusive* model of learning - in

opposition to the reflective abstractions engaged in by Piaget's

privately constructing child.

Gelman, Massey and McManus (1991) have recently toyed with such a

contrast. They argue that learning may be supported by a great

variety of resources. Moreover, they suggest that learners can

themselves sometimes be the best judges of what defines a good

supporting environment. They may sometimes be more expert in this

than the adults (teachers and parents) who will often make these

decisions on children's' behalf. The argument is illustrated with

studies of young children learning from exhibits in a hands-on science

museum. One challenging finding is that accompanying adults often did

not take opportunities to scaffold their charges' investigations of

the learning materials to hand. So, cultural theorists should not

assume that adults are always obliging in this respect (although they

might be within the structured settings that we observe as

researchers). Another finding concerned a display dealing with

principles of motion: a computer-based demonstration was more

effective in helping children abstract the central ideas than was the

spontaneous support of accompanying adults. The warning here is that

interaction with sympathetic adults is not the exclusive route to

understanding - indeed this museum example highlights computers as one

alternative media.

I do not feel that observations of this kind seriously question the

need to view cognition and learning as socially mediated. Take a more

classroom-based example than the one described above: consider the

pupil who independently explores turtle Logo and discovers something

about, say, geometry. The learning is not scaffolded in the same

sense as the example above perhaps. But the pupil's discovery is

located in a classroom on a device that has been configured by

programmers and now used according to a schedule that is determined by

teachers and administrators. Classrooms, programmers, and teacher-led

routines are typical features of these learning circumstances:

features that reveal their socially-constituted nature. They may be

taken for granted as the (mere) backcloth against which learning

occurs, but this familiarity should not cause us to miss the

profoundly social nature of the educational setting. Moreover, having

done so, we may accept that a pupil's learning need not finally

crystallise in the particular context of (tutorial) interaction with

another person. In short, how we organize a pupil's concrete

experience at the focus of this socially-organized world need not

always demand the more intimate engagement of face-to-face discourse.

This possibility is explored in an essay by Davidson (1992). He cites

the case of Pascal, who independently reinvented much of Euclid's

geometry - an interest expressly forbidden by his father. Surely,

this illustrates an intellectual achievement that had to be divorced

from the tutorial support of face-to-face interaction? But

face-to-face exchanges are just one realization of the "social" nature

of mental life. While, in this sense, aspects of Pascal's

achievements are solitary in character, they do remain embedded in the

social world. The symbol systems that the young Pascal appropriated

arose within the conventionalising discourse of mathematicians. Yet

Davidson's example of Pascal is useful. It illustrates the

possibility of considerable intellectual achievement sustained outside

of social interaction. We are reminded that such solitary analysis is

possible; although we may believe that what is involved is more

certainly more likely to be precipitated within social interaction:

'..the inferences needed to construct knowledge do not differ

according to whether they are prompted by independently observed or by

socially received reasons - although their rate might vary

considerably' (1992, p.28).

Thus, the cultural perspective should not insist that the final path

of all new learning is a scaffolded social encounter. Neither should

it deny the creative energy that characterises our typically

constructivist attitude as individual learners. However, this does

not imply that the cultural and the Piagetian constructivist

perspectives are easily harmonised. In the end, the cultural view

distinctively analyses learning in terms of the social structure of

activity - and not in terms of changes within individuals. This

suggests a different style and content to any programme of empirical

research: a cultural approach will be more about the participatory

dimension of problem solving and of the cultural resources that can be

mobilised in the process. These two views of 'learning' also imply

different points of focus for the design of computer-based support.

What exactly is implied by constructivism for the design of such

support? Papert and his colleagues have been the most vigorous group

applying this theoretical framework to development of computer

environments for learning. Papert's original (1980) clash with the

prevailing conception of computer-as-tutor arose from his faith in a

pupil's capacity for self-directed discovery - and his mistrust of

didactic, teacher-led alternatives. With the Logo microworld, Papert

offered a marriage between this new technology and Piaget's

discovery-oriented, child-centred perspective on cognitive

development. Logo was a tool for the making of discoveries. The

child was thereby allowed to control the computer, not the computer

control the child. In this conception, the technology presents a

responsive and versatile environment within which children may extend

their discoveries of nature. Recently, Papert and his colleagues have

coined the term "constructionism" (Harel and Papert, 1991) to capture

their particular perspective on teaching and learning: one that arises

from their experience of developing these ideas in practice.

However, as we have already noted, there has been a growing unease

associated with the outcome of applying the constructivist agenda in

the form of Logo. Of course, enthusiasts can claim that the overall

vision has not been thoroughly implemented and that Logo still remains

the only fully developed illustration of what is possible (Papert,

1987). Yet, the record of success associated with this prime example

remains disappointing. A response to this might be to cite the worry

I am expressing here: the neglect of interpersonal support for the

learning. A teacher's participation does not seem to have been

written into the detailed design of such activities. Yazdani voices a

solution; although it seems to involve only a tinkering with this

balance of autonomy and collaboration. He comments: 'This shortcoming

may be alleviated by designing course-ware within which such activity

is fitted. Here a human teacher holds the pupil's hand when necessary

and leaves him free to explore when educationally effective'

(Yazdani, 1987, p.112). Thus, the social dimension of learning is

respected - but, here, only in the modest sense of teachers being

loosely-coupled to exploratory microworlds.

This hand-holding conception of teaching participation is, so far,

still a rather insubstantial middle ground. It is useful to warn

against constructivist theorisers putting so much faith in autonomous

learning that the social context is neglected. And it is useful to

query cultural theorists who might suggest that all learning must be

scaffolded through social dialogue. However, if teacher intervention

within pupils' computer-based learning is to be one way forward, we

need a greater appreciation of the particular forms that it must take.

While, agreeing with Yazdani (1987) that 'three elements are needed -

students, teachers and the computational system' (p.112), his

suggestion that teachers are furnishing the 'emotional context' does

not take us far enough. Such a "missing bit" view also is implied

within recent work by Lepper and colleagues (Lepper and Chabbay, 1988;

Lepper, Wolverton, Mumme and Gurtner, 1993). Their analysis of

limitations in computer-based tutoring suggests that what is

characteristically missing from such systems is a social capacity.

However, they believe that the difficulties of simulation have arisen

in respect of particular social skills: those concerned with managing

the affective and motivational dimensions of learning. These are

awkwardly fractured from a cognitive dimension inherent in the social

discourse of instruction (a dimension whose management is thought to

be more tractable).

I doubt that constructivist environments of computer-based learning

can be made more complete by simply appending human support - as if it

were just another resource to call upon. In particular, I question

whether the main gap that such interventions would fill is a gap in

the 'emotional context' of learning. In the next chapter, I will

consider how the nature of teacher-pupil collaborative interaction

typically goes further than this. However, here I have merely been

concerned to identify what is entailed by the theoretical orientation

of constructivism in general terms. There is much that is valuable in

this perspective. Perhaps it is only the advent of computer

technology that has associated it with the controversy of displacing

human tutoring. I will return to constructivism in the second part of

the present chapter, where I shall consider how each of the

perspectives I have outlined - cognitive, constructivist and cultural

- influence the way in which computers may be used in educational







In this chapter, I have been exploring psychological frameworks on

cognition and learning. The point has been to find a useful

conceptual vocabulary: one that might help systematise our thinking

about teaching and learning supported by new technology. I am anxious

to prepare for dealing with one characteristic of computer-based

learning - the ease with which it separates from the interpersonal

communication typical of learning from other resources. In Chapter 1,

we noted that this dislocation from the social exchange of classrooms

worried practitioners. The point of adopting some distinctive

theoretical framework is to help articulate the basis for such

concerns. Until recently, practitioners will have found little

support within psychology for any commitment they may feel towards

regarding learning as a socially-organized achievement. Now there is

a framework that helps develop this idea. I outlined the framework

in Chapter 2 - the socio-cultural approach.

In the present chapter, the cultural approach to cognition has been

contrasted with others from the mainstream of psychology. "Approach"

is a carefully chosen word in this context. The contrasts I have been

making identify fairly loosely-knit collections of ideas. Others

exploring such contrasts have suggested that these approaches might be

best characterised as "worldviews" (Agre, 1993) - rather than

"theories". Is there a focal point for that collection of ideas

comprising the cultural worldview on cognition? One good contender is

found in Suchman's characterisation of this framework: 'constructing

accounts of relations among people, and between people and the

historically and culturally constituted worlds that they inhabit

together' (Suchman, 1993, p.71). She contrasts this with a

perspective more typical of mainstream psychology 'dedicated to

explicating those processes of perception and reasoning understood

traditionally to go on inside the head' (op. cit., p.71).

This *social* perspective on cognition is the one which I am most

anxious to capture here. In what follows, I shall call upon this

perspective, as well as others that have arisen above in reviewing the

various worldviews. I shall consider how they each furnish a

distinctive approach to the realization of computer-supported

learning. In my own view, the worldview of cultural psychology offers

the most promising way forward. However, concrete examples of

practice guided by a cultural perspective may make that claim more

convincing. Such examples will be elaborated in later chapters. What

should be possible first is a general situating of the

culturally-inspired approach: locating it in relation to other

theoretical options. Characterising my preferred framework in this

way should create a useful springboard for the themes of collaboration

to be explored in the remainder of this book.

This review of theory will comprise four brief sections. The aim is

to reflect on a tension between artefact and theory: between

computers-for-learning and psychologies-of-learning. So, first, I

shall make some observations about microcomputers as physical objects

that get located in classrooms. It will be argued that their material

nature invites a certain pattern of use. Then, I shall highlight

certain central features of cognitive, constructivist and cultural

worldviews that interact with this material nature in important ways.

In fact, I hope this review will highlight a certain chemistry

involving theory and artefact: one that may be underpinning

distinctive (and controversial) varieties of educational practice.



Some relevant features of computers as artefacts


One distinctive feature of this technology is its physically

self-contained nature. This is significant for determining how gets

used in teaching and learning. It is true that computers can be

hooked up to other classroom equipment (say, in support of scientific

measurements) but, in general, they naturally function as

circumscribed screen-plus-input devices. A consequence of this design

is that they are readily positioned in their own classroom locations;

perhaps physically set apart from the main arena of educational


This need not be so, but other properties of computers surely

encourage this dislocated status. Many programmes written for them

may strike teachers as explicitly designed to function as stand-alone

activities. Suchman (1987) identifies three properties of the

technology that indirectly may promote this way of thinking. She

notes that computers have reactive, linguistic and opaque properties.

They are *reactive* because user actions are typically and immediately

met with (non-random) machine re-actions. Suchman argues that this

encourages the attribution of a sense of purpose - a capability,

perhaps, for sustaining self-contained interactions. Then, they are

*linguistic* because users employ common language to control them

(rather than buttons and levers). Suchman argues that this reinforces

talking about our encounters with computers in the same terms we talk

about human interaction. Finally, they are *opaque*. This argument

echoes Dennett's (1978) claim that the urge to ascribe intention to

human behaviour arises from our inability to see inside each others

heads - our mutual opacity. So, similarly, we may ascribe intention

and purpose to these opaque machines. The argument also refers to the

complex, yet structurally undifferentiated nature of the computer: the

fact that functioning is not readily describable in terms of

particular, graspable local events. This encourages a reification -

we speak of "its" activity.

In summary, then, this technology has both a self-contained and an

interactive quality. So, it may strike practitioners as offering a

focal point within a particular classroom niche: one where autonomous

and lively activity can be supported. The apparently purposeful,

challenging nature of the medium sustains pupils' interest and keeps

them busy. They may be seen as productively "getting on with it".

Conversely, the technology does not seem strongly to invite teacher

intervention. This, of course, may be very welcome in the demanding

environment of a full classroom.

How may learning in such an environment be interpreted? What happens

to pupils when they to go there to learn? Perhaps they retreat to the

computer area and occupy themselves until a task is complete.

(Educational software will often furnish a comfortable sense of

closure - the exercises are done, the pattern is constructed, the

treasure is found, and so on.) At the end, what is there to show for

this investment? Sometimes there will be visible, maybe tangible,

products; but, even then, the *process* whereby this creativity was

possible is invariably lost. What was actually done is too often

hidden from tutorial support (from teachers and others). It is

hidden simply because of the transitory nature of interactions

mediated by screen displays. The task may create a conclusion, and by

some criterion or other it may be satisfactory; but there will

probably be no recoverable record of the pupils' landmarking

achievements along the way to their final product. Moreover, as

Hoyles comments: '..correct answers may well hide incorrect processes'

(1985b, p.10) - a point noted and well illustrated in relation to

computer work by Moore (1993).

I have suggested that the structural properties of this technology

naturally promote a rather dislocated pattern of use. This is for

two reasons. The technology is materially self-contained, and it is

also "interactive". Together, these properties serve to sustain

circumscribed pupil engagements. This may or may not be seen

as a desirable style of working. So, what is its relation with

psychological theories about good circumstances for learning - are

they in harmony, or not? I shall argue below that some of the

theoretical themes discussed earlier seem actively to endorse this

pattern of use. Others can be made consistent with it. But others

caution that we should configure the place of computers-for-learning

differently: we should make them more central to the social fabric of

educational communities.




The artefact in conjunction with cognitive theories

Cognitive psychological perspectives were outlined earlier in this

chapter. At their core is a commitment to explaining intelligent

activity in terms of mental information processing. Human knowledge

is understood in terms of stored symbols; thinking or reasoning

is understood in terms of symbol manipulation. According to Vera and

Simon (1993), these symbols are patterns - of some kind; they are

organized into symbol structures through sets of relations. Their

physical representation in the brain is assumed, but not known; and,

for purposes of psychological analysis, not needed to be known. Such

theories, long-established within Psychology are now attracting some

criticism. Clancey (1993) has cited numerous commentators who argue

that this view of cognition has exercised an unwelcome influence:

serving to distort creativity within a variety of human enterprises.

Organizational practices within modern economies are cited as one

broad example: Winograd and Flores (1986) make such a case in relation

to the deployment of new technology in business. We must include

among these examples the danger of distorting effects in respect of

computers and education. Cognitive theories, in which all intelligent

activity is understood in terms of computations in a private mental

world, tend to cultivate a particular stance on the psychology of

learning. Traditional cognitive theorising does not adequately

challenge the organization of computers into self-contained

arrangements - as characterised above. In fact, it may tend to

reinforce them. It may offer some endorsement to the dislocation of

computer work from a mainstream of socially-organized educational


I suggest that this arises through the promotion of two particular

ideas. The first concerns the possibility of empowering computers to

participate in something like human conversational dialogue. If this

is thought possible, then software authors will attempt machine-based

reproductions of the dialogues associated with instruction. The

second idea concerns conceptions of the outcomes of "learning": how

the symbolising cognitive apparatus gets altered by participation in

educational activities. Some characterisations of such outcomes are,

again, in harmony with more self-contained modes of arranging

computer-based learning. I will briefly elaborate each of these two

cognitive themes in order to clarify their bearing on educational


First, on the conversational potential of this technology: It was

noted above that the reactive, linguistic and opaque properties of

computers combine to create a sense of interactivity. Such

impressions have intrigued cognitive theorists. They have pursued the

goal of refining this interactivity to make it appear increasingly

intelligent; increasingly reflecting the feel of human communication.

Yet, Suchman (1987) questions whether orthodox cognitive approaches

are likely to succeed in this aim. She evaluates their agenda in

relation to the problem of designing a particular instructional

sequence - one that can be incorporated into a piece of office

equipment, and delivered so as to guide its operation by a novice

user. Even this apparently simple task is shown to be very difficult

to bring off. Her empirical observations suggest that simulating real

open-ended educational discourse will turn out to be a particularly

daunting task.

However, Suchman's view - which I find persuasive - is not simply

that, in practice, it proves hard to automate instruction

sequences. It may be an inherently flawed enterprise. She argues it

is grounded in a mistaken conception of how human communication is

managed. This conception is one derived from cognitive theory. It

reflects a preoccupation with explanations of behaviour built around

plans: resources arising from cognitive computations and supposedly

guiding our actions. In cognitive theory, the execution of plans is

supposed to explain individual actions. It is also invoked to explain

joint activity, particularly communication. In such a model,

effective communication is said to depend upon our capacity to read

the underlying plans of others. During communication, this mutual

interpretative ability allows us to recognize and exploit access to

bodies of shared understanding - data to drive our exchanges. Such a

computational model may seem to offer the prospect of simulating

discourse (say, instructional discourse) within a human-computer

interaction. A machine might be programmed to have the same symbolic

computing capacity as a human communicator. So it might generate, and

recognize in others, plans. It might, thereby, communicate. Yet this

is a controversial account of what must be simulated in order to

reproduce human discourse. I shall examine the controversy in the

following chapter, where the possible instructional capabilities of

computers are considered in a little more detail. For present

purposes, I only wish to identify this tradition of cognitive

theorising as offering authority to the idea of computers

"instructing". We need not dwell at once on whether this authority is

credible when closely analysed. The point is, it forms some part of a

climate of beliefs: beliefs that this technology can, at least

sometimes, be trusted to conduct the traditionally social task of

instruction. Moreover, to some extent, we can leave it to go about

doing this.

Faith in the possible simulation of instructional conversation is one

product of cognitive psychology that I believe reinforces the social

dislocation of classroom computers. The second concerns how we

conceptualise the very ingredients of cognition - and, thus, how we

conceptualise what gets learned. There is an orthodox cognitivist

perspective on this. It may originate in the attraction many

cognitive psychologists have towards research on human skilled action

(cf. Miller, Galanter and Pribram, 1960). What research on skills

tends to encourage is an expression of the computations of mental

activity in terms of "thinking skills". From there, the metaphor of

intelligence as underpinned by a cognitive "toolkit" is easily

appropriated. This form of analysis is nicely illustrated in a book

on computers and learning written from within the cognitive

psychological tradition by Underwood and Underwood (1990).

These authors define an approach for using computers in learning whose

aim is: provide children with the ability to think. This is achieved

by providing them with the basic skills - the cognitive toolkit - and

with experiences of the use of different combinations of the

components of their toolkit in problem-solving exercises.


I find this acquisition/application distinction difficult to sustain;

a different psychological model may be more helpful. The distinction

suggests that we offer learners activities that equip them with the

basic tools and, then, activities that allow them to reach for these

tools and apply them (in creative and novel combinations). But where

does the break occur: how clearly can learning tasks be identified as

being about acquiring new tools versus using them?

It is implied teachers need to involve computers in 'two broad classes

of activity' (Underwood and Underwood, 1990, p. 29) but they turn out

to be hard to differentiate. In fact, these authors resist taking the

differentiating step that might tempt many practitioners: offer

children one class of focussed activities (drills) that give dense

practice in the underlying skills while, at other times, offer more

open-ended problems in which the skills are integrated and applied.

This is the very route that such theorising might encourage.

However, Underwood and Underwood's own preference is an exclusive

promotion of activities in the open-ended category. This is welcome.

But hard to derive from a toolkit model. Their rejection of

programmes offering intensive skill-centred practice is made on

the grounds of them being less motivating (op. cit., p.34). Thus,

presumably, they are less efficient. However, even the early

developers of computer drills identified the very opposite problem as

common: students "trapped" into a sustained use of such activities

(cf. Noble, 1991, chapter 7). In any case, when under- or over-

motivation does not seem a conspicuous problem, why not provide skills

practice in a drilled way - rather than tolerate the low density

practice inherent in open-ended programs?

The notion of discrete, general purpose thinking skills may encourage

just this kind of reasoning. The analogies driving this theoretical

framework are seductive. We do, intuitively, recognize that certain

ways of thinking and reasoning seem to have generality: they seem to

get mobilized within a variety of particular situations. So this

invites reification. "Tool" is one metaphor that follows. Although

there is a mixed bag of metaphors possible: for example, "skills",

"tools" and "cognitive health" (op. cit., p.38). What they share,

however, is a focus on cognition as involving core mental processes

that are deployed freely across problem solving domains. Yet do such

processes provide the best explanation of the continuity that we

sense within our thinking and reasoning? I believe such theorising

trades too much on becoming "equipped" with private, self-contained

mental resources. However, what is relevant to the theme of the

present chapter is the classroom practice potentially encouraged by

such an analysis. The analysis may tempt the cultivation of core

tools through efficient but decontextualised activities. It may,

then, tempt a belief that when a learner is witnessed to be exercising

a particular cognitive tool, so the work of acquisition for that core

skill has been achieved. It is available and we merely have to

*motivate* its application elsewhere. Computers may seem to furnish

circumscribed settings in which the use of a given tool can be

precipitated; or suitably honed for application in other settings - a

kind of workshop environment. In other words, the availability of the

tool is not contextually studied in this analysis. So, it becomes a

perspective that may distract attention from overarching social

processes in instruction: the very processes I would suggest are

implicated in creating such cross-situational continuities of thinking

and reasoning that we intuitively recognize do occur.

In summary, cognitive psychology fails to take seriously the social

context of computer-based learning. Indirectly, then, it may

legitimise a dislocation of classroom computers: a social decoupling

already suggested by the material nature of this artefact. I have

argued that, firstly, a cognitivist orientation encourages attempts to

develop machine-based instructional dialogue. Evidently, such

computer-as-tutor applications separate the technology from a social

context: the whole point is to reproduce and replace (human) teacher

intervention. Secondly, cognitive psychology neglects the

contextualised nature of learning in favour of the view that

what is acquired are general-purpose ways of acting. So, it

encourages a "bottom-up" analysis of mental life: an orientation

suggesting a toolshed of discrete, cognitive resources. Within such a

framework, computers might seem attractive and self-contained

laboratory environments for a forging of the tools.




The artefact in conjunction with constructivist theory


Cognitivist and constructivist perspectives have plenty of common

ground. For one thing, constructivist theories (as epitomised by

Piaget's work) share an interest in defining general purpose cognitive

structures. For another, they do not favour contextualist analyses of

cognition. Thus, some of the worries rehearsed in the previous

section apply again in this one. However, there are two aspects of

constructivist thinking that I wish to highlight as distinctive to the

tradition. They are each very relevant to this problem of integrating

computers into a social context of learning. The first concerns a

pupil-centred view of learning; the second, again, concerns the

application of a certain tool-for-thinking species of metaphor.

Neither aspects of the approach are inherently incompatible with the

computational orientation of cognitive psychology. Rather, they

represent considerations that are simply more central to a

constructivist view.

Cognitive psychologists often turn to constructivism when they

confront questions of how conditions for learning should be organized.

Unfortunately, their own perspective on the nature of cognition

has inspired few rich theories concerned with managing the processes

of learning. Theorising tends to focus on the varieties of

structuring that must be imposed on to-be-learned material -

structuring that will best harmonise with human information processing

strategies. For some, this may seem to place the pupil in an

unattractively passive posture: a hapless processor of well-packaged

knowledge. Cognitive psychologists sensitive to the atmosphere of

actual classrooms (cf. Underwood and Underwood, 1990) may turn to

constructivist ideas. They may argue for a learning environment that

respects the exploratory, creative basis of cognitive change. In such

a Piagetian constructivist environment, the learner acts upon the

world, and then abstracts understandings from private reflection on

the consequences of that action. There is a metaphor within this of

the learner as (experimental) scientist. In general terms,

psychologists concerned with information processing may relate

favourably to this image. Although, it is not clear whether learning

that is pupil-centred in this sense can be systematised comfortably

within their preferred vocabulary of cognitive computations.

The significant point is that this theorising tends to direct

attention towards the crafting of environments suitable for such

learning-by-discovery: how to define settings where pupils may explore

and, thereby, construct new understandings for themselves.

Inevitably, at the same time, pupil-centred theorising distracts our

attention from the interpersonal dimension of learning: how it is

supported by the interventions of other people. In such a

climate of theorising, computers may be appropriated to provide just

this kind of self-contained, facilitating setting. In a book

concerning constructivist computer environments, the following is


But education - real education - is not something performed on

someone, nor is it something one *gets*; it is something one *does*

for oneself (Falbel, 1991, p.30).

The idea that it might be a *collaborative* achievement is not an

option in this list. Of course, in principle, actual constructivist

computer environments might well incorporate opportunities for support

at the level of tutorial collaboration. But, in reality, the

interactivity of computer-based microworlds can render this support

(apparently) less necessary: pupils who get stuck can often experiment

their way forward. Indeed, such autonomy may seem very desirable.

There is a second constructivist theme that I believe can seduce us

into neglecting the social context of education. It concerns how to

define the nature of structural changes associated with learning

something. Again, we encounter the tool metaphor, but in a rather

different guise. Constructivism dwells upon pupil activity. It

encourages the design of environments in which that activity may be

creative. So, under this tradition, computers are unlikely to be

deployed in their more didactic tutorial mode. Instead, an ideal

constructivist computer environment might be one that furnishes the

pupil with some generic tool (text processor, spreadsheet etc.). At

other times, the computer might offer rather more specific

"microworlds" for pupils to explore or control: again, we may say they

are enjoying a "*tool* to think with". If a pupil succeeds in

effectively controlling such a tool, the designers of these

environments obviously seek to characterise what cognitive benefits

have resulted.

Sometimes the framework of cognitive psychology, as presented above,

might be invoked. It might be said that the computer environment is

allowing the exercise or integration of core cognitive skills. But

the ambitious designer might wish to identify more far-reaching

impacts than refinement-through-meaningful-practice. In such cases,

effects on the learner are naturally expressed in terms of

cognitive-structural changes. Most straightforwardly, it may be

claimed that acquired ways of acting with the computer tool get

internalized. This thereby creates cognitive structures to be

understood as new private tools of thought. In Salomon's (1988)

apposite phrase, we witness IA, or 'artificial intelligence in

reverse'. "Reversed", because the intelligence acquired originates in

the artefact: the human computer-user internalises computational

resources encountered in the course of controlling the technology.

For Salomon and others there is, thus, a 'cognitive residue' arising

from mastering certain computer tools or controlling certain


"Internalization" is an attractive concept to express such an outcome;

although not the only one. One alternative is that proposed for the

'cognitive residue' associated with learning computer programming.

It is claimed that programming in Logo is a cognitively "powerful"

experience because it makes concrete for the novice programmer certain

strategic processes fundamental to problem solving (Lawler, 1987). An

example might be the mastery of writing "procedures" - self-contained

units of code, the building blocks of a program. Here, it is argued,

what mediates the creation of a useful cognitive residue is not

"internalization". It is some (more constructivist) process of

reflective abstraction - reflection centred about these procedures as

products of thought. Through successful programming, learners are

empowered to confront and examine a concrete representation of their

problem solving processes.

Such interpretation converges on a view of learning similar to the

cognitive psychological theorising outlined above. The key similarity

(for present purposes) is that the outcomes of effective

computer-based learning experiences are characterised in tool-like,

domain-general terms. (On the other hand, a difference might be that

the cognitivist orientation tends to focus on activities that afford

"practice" from the learner; while the constructivist orientation

tends to focus on activities that afford "reflective abstraction".)

Regarding the similarities of position, I should stress that neither

cognitive nor constructivist theorists will necessarily highlight

these claims about tools. They will not necessarily foreground the

general-purpose character of a given cognitive acquisition - as if it

were always a strong empirical claim. It would be fairer to say that

such claims emerge from the literature as implications that invariably

are easily drawn. This is simply because cognitive and constructivist

empirical work on computer-based learning largely ignores the issue of

contextual constraints on acquisitions. It proceeds *as if* what is

learned has generality - or takes the form of new cognitive tools.

Both themes discussed in this section probably enjoy a natural appeal

and, thereby, may readily underpin classroom practice. Firstly, the

notion of pupil-centred learning probably resonates will with

prevailing ideology in cultures where individual agency is highly

prized. Secondly, the notion of acquiring general tools of thought is

an attractively straightforward metaphor for intellectual change

during development. It fits with our fashioning the classroom to be a

general purpose environment - rather than another particular context

where rather particular things get done (Lave, 1988; Guberman and

Greenfield, 1991). Again, my conclusion is as in the previous

section. An influential tradition of psychological theorising

conspires with the material nature of the computer to encourage a

certain pattern of use. That pattern of use makes it too easy for

computer-based learning to fracture from a mainstream of

socially-organized educational practice.




The artefact in conjunction with cultural theory


I see the framework of socio-cultural theorising as a check on this

fracturing tendency. However, first, note the continuity between

cultural and constructivist thinking - the shared commitment to an

active learner. Cole and Griffin's remark: '...thinking is always and

everywhere the internalization of the means, modes and contents of the

communications activities that exist in the culture into which one is

born' (1980, p.356). This suggests an image of learners actively

abstracting order from the world - as a consequence of their

participatory encounters within it. However, compared to traditional

constructivist thinking, the cultural version pays more attention to

an "order" in the world that specifically arises from social history

(artefacts, institutions, technologies etc.). It also sees processes

of social interaction as more important in supporting cognitive change

- where traditional constructivism may stress processes of

(independent) reflective abstraction. Finally, socio-cultural

theorising is less ready to conceptualise general (tool-like)

cognitive changes as the outcome of learning encounters.

These points of contrast identify two themes that I wish to highlight.

They are each theoretical commitments relevant to guiding the

educational use of computers. The discussion of them will involve

some consolidation of points that I have already made. The two themes

concern: (i) the cultural perspective on conceputalising the

cognitive change arising from "learning" and, (ii) how this

perspective characterises environments that are supportive of learners

- environments of education.

The theoretical concepts central to these two themes are,

respectively, mediational means and social interaction. In relation

to them, the following typically is claimed. Appropriation of

mediational means is central to characterising the achievements of

learning. Social interaction is crucial to supporting the effort

towards such achievements. However, these propositions may sometimes

seem tangled. For social interaction can also constitute the actual

mediational means to be acquired - as well as being implicated in the

process of educational support (on occasions where other, more

materially-based forms of mediation are being appropriated). Social

interaction can constitute interpretative practices that we acquire as

new mediational means: such 'ways with words' may mediate new forms of

interaction within our environments. We may often analyse

participation within social interaction in terms of these outcomes.

Yet, at other times (particularly in classrooms) we may wish to

analyse social processes in terms of their contribution to the mastery

of some other mediational means: in terms of a system of learning


Thus, the contemporary cultural literature happens to have much more

to say about social interaction than any other theme. Put the other

way round, it has so far been distracted from saying very much about

other forms of mediation: other technologies and artefacts entering

into cognitive activity. For example, in a recent review, Wertsch and

Kanner (1992) can fill several pages identifying studies of social

interaction within learning. But when they come to summarise

empirical work on mediation, their discussion is much shorter.

Moreover, their discussion is exclusively about "ways of talking" as

mediational means. This is quite legitimate: language practices are

potent examples for any discussion of this topic. However, a focus on

language tends to blur the separate consideration of "social

interaction" and "mediation" - a distinction being made here (and one

made by others such as Wertsch and Kanner). This blurring may

contribute to the popular impression that cultural theory is only

concerned with interpersonal interaction. While, strictly, this may

be a false impression, it remains true that there is scant empirical

research in the cultural tradition that is about other mediational

means than spoken or written language. So, there are few examples of

a socio-cultural analysis applied to learning where material artefacts

are central to the mediational process (although see Hutchins (1986)

for one relevant analysis). This is particularly disappointing for my

present purposes here, as I am anxious to locate such a potential case

for analysis - computers - within this cultural framework.

Yet, such an analysis remains viable. It can be applied most

naturally to styles of computer-based learning described in Chapter 1

in terms of computer-as-pupil/tool. In what follows, I shall consider

the general directions a culturally-influenced account of

computer-supported learning might take.

First it may be necessary to accept a different unit of analysis from

that typical in psychological theorising. In characterising human

cognition, Psychologists have focussed on the individual (and the

individual's covert cognitive structures). Cultural approaches

suggest attention to "functional systems": systems of cognitive

activity in which individuals are participants - interacting with

various mediational means in the interests of achieving some goal.

There is a parallel here with physiological systems (a link developed

by certain early cultural psychologists (eg., Luria, 1973), through

their practical interest in neuropsychology). For example, when

"respiration" is termed a physiological function - this does not mean

it is something with a simple location in some tissue or other (cf.

treatments of "memory" located thus as a psychological function).

Respiration is properly termed a "functional system" because it

comprises a variable mechanism organized towards an invariant goal.

That is, the constituent activities of a system such as respiration

are versatile: they may, for example, permit substitutions or

reorganizations in the face of disruption applied to that system. So,

we may say the same for a psychological function: disruption of an

existing reliable mechanism may allow reorganisation through alternate

mediational means.

The case for orienting towards functional systems - rather than

individual cognitions - has been more fully argued by Newman et al

(1989, chapter 4). They suggest that efforts to characterise

*internal* cognitive representations should be greatly aided by being

able to observe their 'external analogues' in functional systems of

activity (op. cit., p.73). In other words, the intrapsychological

plane of mental life is comprehensively related to an

interpsychological plane of practical activity, and theorising should

explore that dialectic (cf. Hutchins, 1986). An advantage for our

present interest is that this perspective can take much more seriously

the status of external cognitive supports: such as the various

material technologies that enter into problem solving efforts. We

are, thereby, lead to ask how computers - as a particular example of

interest - might enter into such functional systems as mediational


Consider cases from the computer-as-pupil/tool range of possibilities.

A pupil's production of geometrical shapes might be analysed as a

functional system of activity - typical of those cultivated within

classroom life. The resources of Logo-based turtle graphics could

then be viewed as new mediational means that are inserted into this

system. Logo provides children with a new device for manipulating

certain familiar graphic products: generating visual patterns through

controlling the execution of various computer commands and procedures.

In the course of all this, we say that the manner in which the

activity of *doing* geometry is organized gets altered. Our

analysis of this achievement turns away from expressing it in terms

of changes in underlying cognitive storehouses of knowledge. It

turns instead to characterise mediational changes in cognitive

*practice*. To take another example, the effects of computer-based

word processing on the functional system of children's writing

also might be analysed in just this way. Word processing technology

re-mediates the activity of producing text. This is reflected in

new ways of presenting, composing and communicating ideas in

writing. In these examples, what is changed is how we go about being

"mathematicians" or being "story writers" (or, on other occasions, it

might be changes within our efforts to be "biologists", "designers"

and so forth).

Of course such changes may not always be welcomed. In fact,

some criticism of educational technology is grounded in worries about

its mediational effects. For example, in Chapter 1, I noted Cuffaro's

(1984) concern that access to highly manipulable computer painting and

drawing tools will narrow a pupil's experience of graphical

representation. The expressive possibilities in becoming a visual

artist, it is argued, can be effectively constrained by this new

mediational means. Another example would be Johnson's (1991) concern

that database programmes might re-organize the activity of being an

historian, leading it to become dominated by arid data manipulation.

Roszak (1986) voices similar concerns over a much wider range of

intellectual activities.

More often, the effects of introducing computers into a functional

system are greeted favourably. For instance, this is frequently the

case for the examples of Logo and word processing. However, both

enthusiasm and criticism arises from the fact that re-mediation

involves more than creating some expected intellectual product more

quickly, or more profusely or more fluently. This narrower sense of

"amplifying" what is achieved may be claimed (cf. Cole and Griffin,

1980). What is more interesting is how the "prospects" for the

learner may have altered. The creation or elaboration of functional

systems promises extended resources for new patterns of intellectual

involvement - in mathematics, writing or whatever. New mediational

means may sometimes narrow such future creative possibilities (as

critics can claim) but they may also enrich them. For example, word

processing may well make the writer more productive (more words, more

quickly etc.). But it may also have other effects. It may alter the

way the system is carried out. So, computer-mediated writing may

afford new ways of editing text. It may alter the way in which this

system relates to, and enters into, other systems of activity. Thus,

computer-mediated writing may permit the products to contribute to new

patterns of communication with others (eg., desk top publishing or

electronic mail). It may also displace or undermine earlier skills and

patterns of exchange.

Raising these possibilities brings us to what I think is a

particularly important contribution of cultural theorising to the

interpretation of computer-based learning. It concerns how we

understand any extended "reach" that new forms of mediation might

create for a given functional system: how the learner is

empowered to think more effectively in new situations. Traditional

approaches propose that learning involves changes in underlying

cognitive structures: these are then brought into play to enrich new

activities. The cultural approach sees the extension of a functional

system of activity as arising more from organizing and directing

pressures arising from the socio-cultural environment. More exactly:

first, the approach tends to regard new learning as strongly situated

or circumscribed within the particular contexts of activity that are

arranged for it. Second, it assumes generalisation and elaboration

occur because of various forms of socially-organized intervention.

Particularly in school, steps are actively taken by others to

appropriate newly organized functional systems into other, associated

systems of activity. I shall say more about this in the next chapter.

For the moment, this leads us to confront and define the social

organization that does usually characterise educational settings.

For, the business of encountering new mediational means is claimed to

be strongly supported by such social engagements. Recall that

mediation was the first of two cultural psychological themes I am

highlighting here - the second being a focus on the socially-grounded

nature of cognition. I will conclude this section by bringing into

focus the nature of the social context that pupils currently encounter

during schooled learning. The word "social" applied to this context

is intended to identify more than face-to-face exchanges between

people. However, I shall suggest that a richer sense of "social" that

is intended still requires such interpersonal exchanges to be

prominent - if the societal context of activity is to work effectively

as an environment for learning. This key role for social

*interaction* is a theme I shall then develop in the next chapter.

As I have stressed in the last chapter, cultural psychology promotes

two senses in which cognition and cognitive change are "socially"

grounded. Firstly, learning has a social quality by virtue of its

relation to practices of interpersonal exchange - participating in

relevant discourse and joint activity. Secondly, learning is socially

grounded in the "societal" sense: it takes place within an

historically-defined and socially-organized framework of cultural

resources - artefacts, rituals, symbol systems and so forth.

The first of these two senses of social context has been more easily

grasped, and more widely implicated in educational theorising. It is

evidently relevant to my concern that computer-based learning may not

be assimilated into the social fabric of learning. It might lead to

the suggestion that teachers intervene more in (solitary)

pupil-machine learning. This would create more tutorial input - in

the social interactional sense. Such solutions encourage more

assertive or more carefully contrived episodes of instructional

intervention: teachers playing more part in their pupils'

explorations. Lepper and Chabay (1988) and Yazdani (1987) toy with

possibilities of this general kind. Perhaps they may be appropriate

where the form of activity in question is exploratory

(computer-as-pupil); although it is less clear how more assertive

involvement is created for more closed types of exercise

(computer-as-tutor). However, I believe dwelling only on this form of

solution will miss the significance of a cultural analysis as implied

by the second sense in which learning is viewed as socially grounded -

the societal sense.

What does this second sense of cognition's social character imply for

the effective incorporation of computers? It requires us to attend to

the kind of consideration expressed in a quotation from Scribner

(1990) used earlier in this chapter. She encourages us to give more

consideration to 'how cultural communities this world over organize

activity settings for the "social transfer of cognition"' (p.93).

This invites us to attend more to the social fabric that surrounds and

supports the learner - the established traditions and technologies

that constitute 'activity settings'.

To observe a school and to declare that the learning therein has a

"socially-organized basis" need not suggest that pupils' activities

are constantly guided by intimate tutorial contact. They may be a lot

of the time, but - particularly where computers have penetrated - what

pupils are often doing may have a solitary character. That is, it may

not incorporate much face-to-face social interaction. However,

schools and other traditional sites for the organization of learning

are socially-defined in a more profound sense than this. Thus,

schools and their classrooms are structures comprising conventions,

technologies, rituals, architecture and other distinctive features

that have evolved through our cultural history. This sociocultural

identity therefore influences learning through its' structuring of

motives, goals, values, priorities, assessment and so forth. It

specifies a pattern of relations among domains of knowledge that

define a "curriculum"; it describes social roles and rules for the

management of attention; it exercises various links with other

community structures, including families; it furnishes various kinds

of technologies and resources to support study; it imposes a

particular style of cognitive work through the organization and

management of physical space. Of all the human activity we may

suggest to be "socially-grounded", learning within the setting of

institutionalized schooling must be one of the most vivid examples.

How can the incorporation of computers into education bear upon its

"social" nature in this sense? At one level, the classroom computer

is merely another technological intrusion into an environment that is

already preoccupied with technologies and their deployment. This

invites a cultural line of analysis that dwells upon the social

history of their development: revealing how the activity of designers

and innovators has imparted to this technology characteristics that

afford or constrain how it might now be used (cf. Habermas, 1987;

Noble, 1991). However, my purpose in raising this second sense of

learning as socially-grounded was not to open up the possibility of

that particular discussion. My purpose was to make a more

straightforward point: namely, that this social fabric of education

does evoke a participant structure. There are ways in which people

must act in order to coordinate it and, thereby, make it work for

pupils. Such a sophisticated sociocultural system has to be kept

knitted together; the system will incorporate roles, responsibilities

and expectations that serve to support such integration. Through it,

various kinds of continuity and connection may be created that define

the overall experience of "school" or "learning". Particular things

get done and said to integrate pupils' discrete learning experiences

into a richer texture of shared communication and understanding.

One way to put this, is to claim that the classroom community will

resist the creation of "vacuums" for learning to occur in: resist the

possibility of learning becoming divorced from other past and present

understandings achieved in the overall setting. However, it might be

hard for the participants to appreciate that this quality of

continuity does characterise their enterprise - hard to identify it

from the familiarity of being inside it, as it were. The significant

point is that this consideration of a schooling context returns us to

an interest in the organization of communicative activities. But now

these activities arise, not in the form of intimate tutorial guidance

(the most accessible sense in which "learning" is recognized as being

about social interaction). Instead, they arise through the demands of

sustaining a certain socio-cultural fabric.

Naturally, I am suggesting that this fabric is important. What pupils

achieve depends upon schools grounding learning in a socio-cultural

context understood in this sense. But I am also suggesting that the

way we deploy new technology in classrooms is very relevant to the

quality of such a context. So, the tension between computers and the

social character of education arises not simply because we have some

rigid belief that learning must always be mediated by a human

instructor acting in a tutorial relationship to a pupil. There may be

a need to preserve comprehensive opportunities for such focussed

guidance. However, there is also this other, more overarching sense

of social interaction that needs consideration. Computer-based

activities may encourage a dispersal of learning into discrete and too

self-contained experiences. This technology has a capacity for

motivating and directing problem solving activity to be a fairly

autonomous experience. There is something attractive to this. But

it does serve to moderate some of the preparatory or synthesising

interventions that teachers naturally offer.






In this chapter, I identified two well-established traditions of

psychological theorising: cognitive psychology and constructivism. I

sketched their features in a way that I hope allows the character of a

third possible perspective (cultural psychology) to be appreciated

more clearly. I then considered the relationship of these theories to

computer-based learning. My argument has been that there is a certain

chemistry involving the material nature of the technology and ideas at

the core of orthodox psychological theorising about learning. The

product of this chemistry is that computers are all too readily

separated from the mainstream of classroom life. Their material

properties tend to invite this, and prevailing theories about learning

and cognition do not adequately challenge such a trend.

Is there evidence that dislocation from the social context of learning

can undermine what pupils achieve? The little direct evidence

available tends to point this way. Thus, to some commentators,

evaluation studies of learning with certain computer microworlds (such

as Logo) are strongly suggestive: they imply that the impacts of

learning in these settings are not readily visible outside of the

microworlds themselves (eg., Simon, 1987; Yazdani, 1987). The

shortfall of these microworlds in practice is well expressed in

reflections from one commentator working inside the cognitive

tradition of artificial intelligence (AI):

AI has not yet been able to translate its successful attempts with

microworlds to the real world. There seems to be such a level of

increase in complexity of the domain when moving away from the

microworlds to the real ones, that most lessons need to be relearned.

(Yazdani, 1987, p.112)

The evidence suggests that the problems are located in poor contact

between the pupils' microworld activity and traditional forms of

teacher intervention (Pea and Kurland, 1987). These conclusions are

fairly compelling - although there is debate as to whether

constructivist agendas for educational technology really did intend

that computer-based learning should separate from socially-organized

instruction (Papert, 1987). Whatever, the intention, it seems that

this dislocation does easily happen. It is also clear that, in their

current work, the early innovators now pay much more attention to

specifying a supportive social framework (Ennals, 1993; Resnick, 1991;

Turkle and Papert, 1991).

My account here of a cultural perspective on learning and cognition

should indicate that this approach will have a particular interest in

computers as they are applied to education. This is partly because a

central concept in the cultural approach is "mediation". Cognition is

viewed as human activity mediated by the technologies, artefacts and

rituals that have emerged within the history of a particular culture.

Cognitive change and development then involves the appropriation of

such mediational means as those represented in new technology. I

suggested above that the proper unit of analysis for understanding

this cognition was a functional system: coordinated cognitive

*activity*, often incorporating forms of external support (as well

illustrated by the example of computers). Mediation, as it occurs

within functional systems of activity, helps us understand the nature

of cognition. What of learning, or cognitive *change*? Here again

the cultural perspective should have a special interest in computers.

This is because their common pattern of use can conflict with a second

key concept in this tradition of theorising: the social basis of


A concern with mediation was certainly one concept central to the

cultural approach of Vygotsky and his contemporaries. But, alone, it

would not set the socio-cultural theorists apart as a distinctive

movement within psychology. What does set them apart is the

combination of their concern for mediation and their commitment to a

related claim: the claim that cognition is *socially* constituted.

Cultural psychology makes sense of "learning" by reference to the

social structure of activity - rather than by reference to the mental

structure of individuals. We are encouraged to analyse learning by

paying careful attention to two features of its typical format.

Firstly, learners bring established functional systems of cognition to

any new situation for learning: we should try to understand their

nature. Secondly, other people (particularly teachers) act in ways

that allow these systems to be re-mediated or elaborated. Social

interaction, it is claimed, makes possible the extension and

transformation of existing cognitive systems. This is a theme to be

taken up further in the next chapter.

If cultural theory alerts us to the social texture inherent in

educational activity, what is recommended when we sense its disruption

- as we might in the face of new technology? Evidently, any answer to

this will arise from better understanding of the interpersonal

*processes* that normally are active within classrooms. The repair

will have to be organized at these points. I shall consider what we

understand about these processes in the next chapter, using it to

start deriving preferred ways of deploying new technology. A key

concept will be that of socially-shared cognition or - to employ a

term in contemporary favour - "intersubjectivity". I shall argue that

the opportunities to create and exploit this state underpin much of

what is achieved in formal educational settings. We discover it not

just in the face-to-face social interactions that constitute

"tutoring" in these settings but also in more extensively mediated

forms of social exchange. The significant challenge will be to

determine how computers can "enter into" these practices - rather than

sidestep or undermine them.