Agents and Artefacts.

John Pickering, Psychology Department, Warwick University.

In Social Analysis, March 1997, Issue 41 (1), pages 45 - 62.

email: j.pickering@warwick.ac.uk

Introduction.

Computer technology has given rise to agent like artefacts that are becoming intimately involved with social action. Present trends in this involvement show it to be deepening as agents become more autonomous and more capable of natural seeming interaction with human beings. As the trends unfold, circumstances will arise where it will be difficult or unimportant to distinguish between technologised human agents and humanised technological artefacts, especially for the young. Machines are now more than ever involved in the evolution of technology and skilled human practice as participants rather than mere tools.

The autonomous machine that concerned Samuel Butler in Erewhon is very much with us. The blending of people with machines, the posthuman condition, has been a commonplace of cyberpunk writing for almost two decades and is beginning to be discussed as a technological project (Pepperrell, 1996). Fiction aside, the fact is that technology has played an increasingly significant role in human cultural evolution for well over 2 million years. From the earliest stone tools to nanotechnology we see a progressive increase in the importance of artefacts to human practices.

This paper considers how computer technology fits in with this progression. It will consider the role of artefacts that simulate elementary psychological powers in the integration of technology with human practices. One aspect of this role will be to create skilled practices in which humans and machines interact socially. Rogoff's distinction between three levels of human sociocultural learning will used to suggest more specifically how these practices may arise (Rogoff, 1995). Also relevant here are recent developments in evolutionary theory and in psychology, as well as Ingold's critique of the conventionalised distinction between biology and culture (Ingold, 1996). Taken together, these elements provide a framework within which to address the assimilation of technology, especially during human development.

Social interaction with machine intelligence is becoming more realistic. It is not important whether machine intelligence is considered to be "the same as" human intelligence. Nor is it important to decide whether systems that learn and evolve are in fact autonomous agents. Whether the agency of artefacts comes from human design or whether it emerges as the artefact develops, what is important is that artefacts can become agent like. Simulation is enough. The possibility of interacting with artefacts as if they were agents will mean that they are treated as such. This will give them a role in creating the habitus of contemporary technocratic societies. Habitus here being used in the sense used by Bordieu to mean not only the material conditions of a culture but its sensitivities and values as well (Bourdieu, 1984). New values and sensitivities are being created as people and agents co-operate. As artefacts develop basic social abilities, they will participate in the development and transmission of skilled practices from one generation to another.

This participation is growing rapidly as computer technology changes the habitus of education, work, the arts and domestic life. The practices and values thus created are assimilated by individuals as they develop within technocratic cultures. As McLuhan pointed out, assimilating technology changes consciousness. Systematically observing this assimilation, especially by young, sets an important ethnographic agenda.

Growing up with technology.

Rogoff's analysis of sociocultural learning blends the pragmatism of Dewey with the action theory of Vygotsky and Leontiev. Knowledge is not abstract or formal in essence, although it may be made so. Rather, it arises through and is for action. As well as formalised education, knowledge and the values that go with it, is transmitted through cultural activities, technological practices and everyday social interaction, particularly that between children and the world of adult practices into which they are learning to fit. From these elements, Rogoff developed a three level model of qualitatively distinct but overlapping processes through which individuals assimilate sociocultural practices.

At the most explicit level is apprenticeship. This is a broadening of the conventional usage beyond the simple acquisition of craft or artistic skills. It is a collective term for activities such as education, training and instruction of all sorts. Apprenticeship is what happens when learners who know they are learning participate with teachers who know they are teaching to develop specific skills and knowledge.

At the next level there is guided participation. Here, explicit instruction is not involved. Nonetheless individuals learn though taking part with others in collective activities that leave a residue of skills and knowledge. These collective activities need not be mediated through direct social interaction. Rather they are mediated through any activities which transmit skills and knowledge from those who have them to those who do not. Instruction may not be the primary purpose of these activities, indeed, it usually is not, but it is an important and direct result.

Participatory appropriation is perhaps the least explicit and consensual mode of sociocultural learning. It is closely related to guided participation with the important distinction that it is personal, tends to be creative and to individualise knowledge. While the former two categories deal in knowledge and practices that are to some extent pregiven, this latter category also deals with ways of participation that originate with the individual. The former two categories are more likely to involve the individual in internalising the goals, knowledge and values of their sociocultural setting. The latter category, however, is more akin to the individual making for themselves a style and a unique set of practices which are the means to achieve goals they have set themselves.

In making this distinction, Rogoff asks basic questions about social learning such as: What is learned? Where and how is it stored? What are the effects of this learning? In what activities does learning arise? What are these activities and how do individuals relate to each other when they are being carried out? Such questions can all be asked about the role of technology in social learning. They help create a framework within which technology can be related to the broader sweep of cultural evolution.

Technology and the Posthuman Condition.

Ingold points out a number of difficulties with the conventional distinction between biological and cultural evolution, and therefore, between evolution and history. His major point is the neglect of development by evolutionary theorists. Human beings develop within an envelope of skilled practices and the material artefacts associated with them. Together these constitute a self-replicating system that leave a permanent trace, both within the environment and with the body. For example, practices to do with posture and gait are reflected in the growth of the body. These could be seen as the impress of 'extraneous' cultural influences on a 'fundamental' biological pattern. This promotes the idea that cultural evolution is a historical superstructure built upon a relatively stable biological condition. But as recent developments in evolutionary theory show, the stability species is not merely a biological matter (Johnston & Gottlieb, 1990). This is particularly so for the human condition, where cultural and biological factors have blended in the evolutionary emergence of modern human beings (Kingdon, 1993). Accordingly, the distinction between biology and culture, while useful when considering the human condition as it now is, should not be taken as a well defined boundary between biological evolution and human history.

Ingold's critique concentrates on bodily skills, such as the use of tools or modes of walking. For example, he notes that the effects on the skeleton produced by particular traditions for carrying and squatting are considered as the imposition of a cultural practice on a distinct biological form (Molleson, 1994). He suggests that it is more appropriate to view them as a trace of both cultural and biological systems within which together stabilise the human form. Now the human skeleton is relatively resistant to change when compared with the brain, whose very nature is to be plastic and to assimilate external influences. Ingold's critique, therefore, is clearly extendible to the trace left in the mind by cultural practices. As Vygotsky recognised, physical, social and mental action rather than biology provide the form of human psychology. With the rise of intelligent machinery, cultural practices will incorporate social interaction between developing human beings and technological artefacts. This is interaction is now being represented as part of a change from the human to the posthuman condition (Pepperell, 1996).

The posthuman condition is prefigured in the work of cyberpunk writers such as William Gibson and Bruce Sterling who depict human beings blending with technology. Bodies are invaded by minute cybernetic intelligences and the senses are hugely amplified by electronic means. The body is no longer the necessary vehicle for experience and agency. Instead individuals diffuse into virtual worlds of disembodied powers and feelings (Sterling, 1986). Rather than being mere literary tropes, such writing is often considered to have anticipated the remarkable growth of global networked communications such as the Internet, of the prosthetic application of nanotechnology and the convergence of biology, psychology and computer science into what is now called artificial life (Gray, 1995). As technology becomes more organic, so the organic is technologised and increasingly brought within the sphere of human social concern and action. There are even serious discussions of how the human beings, once liberated from the body and expressed through technology, might become effectively eternal (Armstrong, 1996 ; Moravec, 1988).

The notion that the body can somehow become redundant seems unlikely and will not be pursued here. Even so, it is clear that computer technology raises Robertson et al., 1996 ; Haraway, 1991, 1995). Indeed, the scientific approach to the 'natural' is more technologised than ever before. In psychology, for example, the idea that the brain is much like a computer is no longer treated as just another metaphor. Now taken in a far more literal sense, it is advanced as the best candidate for a unified theory for all aspects of mental life (Newell, 1991; Churchland, 1986; Johnson-Laird, 1989). In this sense, the present era of psychology is the era of cognitivism, much as the middle of the century was the era of behaviourism (Valsiner, 1991).

Does this make the notion of the posthuman condition less of a literary conceit and support the idea that technology and human biology may be blendable? For the purposes of this article, the answer will be taken as "no". Common sense suggests that the posthuman cyborg is not a realistic option. However, the image helps weaken the idea that there is, or ever was, a pre-given, biologically definable human condition that can be qualitatively distinguished from a radically technologised posthuman condition. The emergence of the human condition was always a matter of the progressive integration of technology with biological and social practices. This, as Ingold points out, is especially true of those practices surrounding human development, which have an important role in both stabilising and directing cultural evolution.

Cultural evolution is the progressive, directed change in technology, tools and the practices that go with them and is thus Lamarckian. It has replaced Darwinian modes of biological evolution as the principal mode of human development, both individual and collective. This mode of evolution is not merely the consequence of chance events, as reductionists might have it (Campbell, 1974; Monod, 1971). Rather, it is based on constantly re-negotiated cultural goals and values. These, along with aesthetic sensitivities and social skills are assimilated with the language, customs, myths and artefacts, what Bourdieu calls the habitus, of a culture or community of practice (Bourdieu, 1984). What is significant about the latest phase in cultural evolution is that the artefacts themselves are developing the capacity for rudimentary psychological and social skills.

Posthuman? We Were Never Human.

Lewis Mumford's early works such as Technics and Civilisation and The Culture of Cities, both written in the 1930's, chart the progression of technology and human practices on a very long timescale (Mumford, 1968). He proposes eotechnic, paleotechnic and neotechnic phases to cover, in rough approximation, tool making, power generation and information handling. These remarkably far-seeing works show how technology makes possible new conditions of cultural and social interaction. they also help to understand the technological shaping of cultural conditions within which human beings develop. Their optimistic humanism is a sobering contrast to some of his later works, such as The Pentagon of Power, written at the time of Vietnam, where he warns about the abuse of technology.

Mumford shows that technology is not merely the means by which pregiven and relatively unchanging human needs are met. Rather it helps to create fundamentally new forms of human activity from which new goals, values and desires emerge. Technology creates the cultural envelop within which human beings develop and from which they assimilate, with modification, a system of values and goals. These are not always explicitly defined or transmitted by processes of formal education. They may, for example, be carried by the built environment and the practices that go with it (Sinha, 1988).

The 'natural' home of most of human kind is thus a value laden artefact, a semiotic system (Elias, 1989). Kingdon suggests that acceleration of human evolution in the late Pliocene had a great deal to do with the development of semiotic systems for transmitting tool technology (Kingdon, 1993). He also suggests that this will have had genetic consequences. In a socially co-operative species like hominids, the genome will have evolved under the selection pressure to assimilate the collective semiotic practices that transmitted tool technology. This will have created a hybrid system in which genetic and extra genetic vehicles for inheritance were blended and which was responsible for the relatively rapid appearance of modern hominids.

This, as Kingdon suggests, means that the human genome has followed in the wake, as it were, of this larger hybrid system. This neo Lamarckian mode of evolution is similar to that proposed by Piaget and Waddington, among others. A number of recent developments in evolutionary theory do in fact propose that the human genome is not as isolated as conventional neo Darwinism suggests. For example, there are good arguments for the view that the genome is not a morphological blueprint (Johnston & Gottlieb, 1990 ; Goodwin, 1994); the development of the brain may be far more influenced by assimilation of environmental influences than was previously thought (Edelman, 1992) and the continuity of ontogenetic and phylogenetic learning is becoming more widely accepted (Plotkin, 1988 ; Maynard-Smith, 1987).

Thus, both human evolution and development are bound up with system in which biology and culture are integrated. Sociolcultural relations within this system have become increasingly mediated by technology. It is within these relations that an individuals develop and learn to participate in the practices and to recognise the values that surround them. In the human case the surroundings the developing individual are shaped by previous human beings to a far greater extent than for any other species. The mutuality of organism and environment that is true for all species has been radically amplified by the loop created by the fact that human beings develop within an environment that is created by human beings themselves.

Thus, there never was a human condition to begin with. Not, that is, in the sense of a biological species. The large overlap between the genetic constitution of apes and humans demonstrates that the rapid evolutionary appearance of human beings was a cultural, not biological event. What the environment affords a particular species is, in most cases, only marginally affected by the past actions of the species in question. However, the human environment is almost totally shaped by the actions of other humans. Thus it is unproductive to assume what makes humans different from non-humans exists 'inside' the organism. It is more important to understand that human development is far more under the influence of the environment than that of any other species. Since the human environment is itself a cultural product, this means that human beings are, as Kingdon suggests, 'self-produced'.

Maturana and Varela call this process of self production 'autopoesis' and propose that it describes the mutuality of all organisms and their environment (Maturana &Varela, 1992). In the human case however, this condition has been radically intensified by cultural practices which surround individuals as they develop and draw out and shape an innate predisposition towards action. These practices now involve artefacts which simulate human social skills. As Mumford reminds us, technological artefacts do not only match a static set of human values and goals, they also create new ones. As artefacts become social, so new values and goals will appear having to do with the skilled practices that arise in interaction with them.

Learning these practices may require another individual to act as intermediary, model or teacher, recalling Rogoff's classification. It may be, however, that an object may creates practices by itself, much as an implement instructs the user how to wield it. It is in this sense that culturally produced opportunities for action, whether provided by objects, situations or people, may be considered to be social (Costall, 1995). While this broadens the meaning of 'social' considerably, this is what is interesting about artefacts with the capacity for social interaction. This adds to Ingold's critique of the boundary between biology and culture the blending of human agency with non-human artefacts.

Of course, human social interaction has been mediated by artefacts such as the telephone for many decades. What is significant about recent technology is the artefacts themselves participate in the interaction, rather than acting as merely a means to put human beings in touch with each other. Of course, what use human beings make of agent like artefacts cannot be predicted exactly. The use made of an artefact is often surprising to the designer. Moreover, when the artefacts can both be modified by the user and modify themselves, they pass even further beyond the control of the designer. However, the point made by Pfaffenberger (this volume) that computer technology is creating new field of social relationships, will be amplified in a unique way when computers participate in rather than merely mediate these relationships.

Artificial intelligence and artificial life.

In a suitably ironic postmodern exchange of identity, recent developments in biology and psychology reveal that people are more like artefacts than might have been thought and that artefacts may able to become more like people than perhaps we feared.

Artificial intelligence (AI), the structuralist heart of cognitivism, has developed a murmur. A radical transplant is in progress to replace it by the post-structuralist project of artificial life (AL). This is a research programme that combines biology, computing and psychology in order to expresses the logical form of life in artefacts (Langton, 1995). The aim is build complete creatures that learn and evolve rather than to capture the formal essence of intelligence in a computer program (Brooks, 1991,a). Much of the intelligence expressed in AL systems is not pre-programmed but evolves within an arena of action (Brooks, 1991, b,c). The shift from AI to AL tracks an important change that will provide a psychological theory more appropriate to understanding the cultural assimilation of technology. The shift, broadly speaking, is from cognitivism to situated action.

Cognitivism reduced the mind to a machine in the head and created a research programme for testing theories of how this machine might work. Cultural practices were of no immediate significance and would have to wait for the fundamentals of mental life, the operating principles of the machine in the head, to be properly understood (Gardener, 1985, page 6). The operating principles themselves were taken to be formalisable, either as mental language (Fodor, 1990) or as a Turing machine (Harnad, 1992 ). Although this programme is criticised for neglecting both biology (Edelman, 1992) and culture (Bruner, 1990, ch. 1), is remains the principal paradigm for most of cognitive science (e.g. Benjafield, 1996).

This reductive enterprise expresses an implicit belief that mental life is a culturally and historically independent natural kind. The content of mental operations may be assimilated from culture but the operations themselves and the brain structures that carry them out are not. In sharp and productive contrast, approaches to mental life as an aspect of situated action, consider it to be something carried within a system of which only a part in the head. The research programme this generates is less concerned to formalise intelligence and more concerned with a ethnographic natural history of human practices (Lave & Wenger, 1991 ; Costall & Leudar, 1996).

While AL is not yet developed enough to take part in this research programme, it signifies a shift in the right direction. Psychology is moving away from cognitivism and towards the treatment of cognition as something that emerges when capacities for action are expressed in situations. This theoretical stance offers a far better fit with pragmatists and action theorists alike. Once mental life is no longer seen merely as formalisable mechanisms in the head, the way is clearer to understanding the cultural nature of human intelligence and to paying more informed attention to how mental life develops through sociocultural learning. Human psychological capacities are embodied in a form of life that has a unique evolutionary history in which biological and cultural factors are more closely integrated than for any other species.

It is during development that biology and culture interact to the most significant effect. The sociocultural learning that is the medium for this interaction is increasingly mediated by technology. This has now given rise to artefacts able to participate in social processes and to take over some of what one human being has traditionally done for another, for example, in the various types of sociocultural learning. As Rogoff shows, this transmission is often carried out through implicit social interactions between human beings that transmits sensitivities and values as well as skills. If human beings adapt themselves to machines, new, as it were machinic, sensitivities and values will enter the human arena. This process is already clear in the way that software induces skills, mental images, habits and preferences in its users, as anyone who has had to learn to use both Macintosh and Microsoft operating systems will know. Here we see a similar technological imperative to that of the Berlin Key described by Latour (this volume?). Indeed, a software object like an operating system is far more capable of far more powerful control than a piece of hardware like a key.

The participation by machines in real social action is now so far advanced that it is no longer merely a matter of technology but ethics (Lanier, 1995). Not just in the usual indirect sense of the use that is made of tools by human beings. Instead the question is whether artefacts are mere tool-like adjuncts to human life or whether they might become participants in our moral and aesthetic lives. Machines used to be built and then used. Now, once built, they learn and evolve. In doing so they may elicit in the human beings that interact with them a new type of skilled practice that expresses cybernetic rather than human values. In conforming to these practices, especially during development, human beings will change. This will continue the process by human beings have always accommodated to the social side effects of technology. This will now involve accommodation to the intrinsic social capacities of artefacts. Recalling the contrast between Mumford's early and later writing shows, the question arises as to whether this accommodation could diminish human autonomy and worth.

Cyborgs aren't coming, they're here

As Mumford showed, technology not only amplifies human capacities, it also creates needs, goals and values. Presently, cybernetic technology is amplifying human capacities for social interaction. This may create new needs, skills and values that will be expressed in social relations. However, is there is anything here that is qualitatively different from previous phases in the evolution of technology, for example, that of the heat engine?

Consider cars and computers. The basic technology of the car may be approaching a thermodynamic plateau. What is futuristic about the Vectra, Vauxhall's "car of the future", lies in its advertising campaign rather than in the car itself. There is nothing particularly advanced about what makes the Vectra reliable, safe and economical. It is pretty much what manufacturers have been doing since about 1960. What is happening here, in Baudrillard's terms, is the replacement of real production and distribution by the circulation of signs (Baudrillard, 1993, ch. 5). By contrast, Fordist modes of production, created by the heat-engine technology of the previous century, that were still in place in the 1960's have been radically altered. In the next century cybernetic technology will consolidate post-Fordist modes of production (Harvey, 1990). Here manufacture is dispersed and automated, assembly is multi-site and flexible, production planning is virtualised and management hierarchies are flattened. Through the convergence of stock control, customer surveys, order books and market research, consumer choice is reaching into earlier and earlier stages of design and prototyping.

Cybernetic technology has not yet come to a plateau. How far there is yet to go is impossible to know, but several orders of magnitude seems a conservative estimate (Pearson & Cochrane, 1995). An envelop of social relations mediated by this technology is developing and moving down the age scale. What used to be an activity of adults at work is rapidly becoming what children do at school and in the home while learning, playing and communicating. Cybernetics has moved from automating mundane rationality to automating mundane sociality. Hybrid intelligent action involving people and computers is increasingly part of social situations. Computer systems help people to communicate, design and decide.

For adults, such systems can be threatening, leading to technophobic attitudes. By contrast, the ease with which children get on with computers is now a common experience. Being able to operate and to co-operate with technology is not just to do with knowing how to make the video recorder work. It is about feeling at home with machines that are beginning to use language, recognise individuals, make decisions and offer advice. Being at ease with these human simulacra will have as much to do with attitudes as with skills.

Such simulacra are appearing in all areas of economic and social life. Cars now have voices to remind the driver of things and to give advice. Portable computers now accept written and spoken inputs. These, unlike typing, are the property of individuals, making the machine very much more the a personal assistant than a mere tool. Telephone systems are now computer systems which do a great deal more than put people in contact with each other. These systems are in effect quasi social agents for answering, asking questions, giving informing, holding callers, re-trying numbers that were engaged, informing users of another caller waiting, taking messages, re-routing enquiries and so on. These systems are mediating the social politesse of telephone communication. Presently, the systems we typically encounter are easy to distinguish from a real human being. With communication systems like email and the Internet in general the situation is different. These systems do not generally involve writing rather than speech and interactions do not occur in real time. Under these conditions the distinction between a human and a non-human agency is increasingly hard and unimportant to make. Computer agents sort, order, discard, answer and redirect electronic mail, field enquiries, answer questions, make recommendations and so on.

While most households have a phone and TV only about a tenth presently have a computer. However, the cost/power ratio of computers is improving and the cable is ramifying through urban culture. As it settles into its mature form, cybernetic agencies of increasing in sophistication and naturalness of use will appear. This is especially true with the rapid convergence of domestic computers, TV's, faxes, cable and satellite systems, digital audio broadcasting, telephones and other media. Soon, children will be growing up with screens, keyboards, CD players, microphones and loudspeakers that will be an integrated resource for education, communication, recreation and entertainment. By then, a great deal of control will be invested in quasi-social intelligent systems. Their forunners can already be seen in the increasingly naturalised operating systems of home computers. These prompt, instruct, ask and autonomously act to provide the user with what they want. As systems converge, more sophisticated agents will help people to use them. In particular agents will be need to handle communication between human beings, between human beings and computer systems. The complexities and subtlety of human - computer interaction and the constant changes to computer systems make this a formidable problem of knowledge engineering.

Recent work in this area has explored a solution based on artificial life (Maes, 1995 ; Ray, 1995). Software agents are being developed that can both learn and evolve. These learn to interact with particular individuals and to acquire skills in handling particular aspects of computer networks, such as the variations in traffic density with time of day and global location. They evolve by making variants of themselves which are eliminated by selection pressures from human users, the availability of resources within the network and interaction with other agents. Presently these techniques are being used to develop agents that help people locate, sort and present information found on the internet. It is important to note that the ways in which these agents work will not have been fully anticipated in the intentions of a designer. If successful, they will be applicable in other areas of human computer interaction, including communication itself. Accordingly, it is quite conceivable that computer agents able to communicate in a relatively natural way, perhaps with a particular individual, might be created by the individual themselves interacting with evolving versions of the required agent. As the result would reflect the user as well as the artefact, it could stand for the human original in some circumstances becoming some sort of a computational doppelganger

Computer systems have already become socialised and personalised to a remarkable degree. Supermarket tills not only display prices and product codes, but also provide help and supervision at a quite naturalistic level. They prompt the operator to ask the customer questions, to obtain signatures, to carry out various phases of the transactions and to watch out for errors. The till has access to a lot of information the operator no longer needs to do their job. For example, if a customer asks the price of an item, the operator is quite unlikely to know it, but will obtain it from the machine. The level of prompting from and intervention by the till varies with the skill of the operator. With the checkout closed, operators can be trained by the till itself, with only occasional interventions by a human supervisor required.

Home computers, in addition to paper manuals, now come with extensive CD-ROM tutorials. These can be on the machine and blend with the actual use of it, remaining in the background and adjusting their interventions as the user becomes more skilled. When a difficulty arises adults are likely to say "where's the manual?". Children are more inclined to say "let's ask the computer".

Children who grow up using such systems may on occasion be unconcerned whether they are communicating with a human agent. They are likely to take more rapidly than adults to new forms of cybernetically mediated social interaction. For adults it may feel as if an alien has arrived in the household and only the children can communicate with it. Unlike children, adults may be subconsciously concerned with whether or not they are dealing with another human being. Again, it is important to note that it is impossible to say exactly how this or that aspect of technology will be used in human activities. Even so, it seems clear here that something important is emerging from the way human beings are growing up with technology that simulates human social interaction.

It is becoming clear that such developments will influence how human beings communicate with each other and how they think of themselves (Lanier, 1996). The question arises: could artefacts that can evolve and learn in conjunction with human beings, generate skilled practices and thereby participate in human social learning? Such practices are not only ways to gain practical outcomes, they also help create the habitus of cultures where social relations are mediated by machines. These relations will thereby transmit technologised values and sensitivities. As well as being technological adjuncts to human life, computer systems may help in creating attitudes, tastes and modes of social interaction.

The economic and social forces produced by technology, as Walter Benjamin pointed out, produce changes in human sensitivities (Benjamin, 1979). With the advent of computer technology, these changes will now appear more directly and rapidly than ever before. The sensitivities, however, will be those of human communication and social relations

Benjamin's Warning

"The information society is on its way. A 'digital revolution' is triggering structural changes comparable to last century's industrial revolution with the corresponding high economic stakes. The process cannot be stopped and will lead eventually to a knowledge-based economy."

Such techno-enthusiasm, often accompanied by colourful depictions of cybernetic wonders to come, now fills the popular media. As the century closes, computer technology is being celebrated with the visionary fervour that greeted heat engine technology at the close of the nineteenth century. The speed with which information can pass between people, around the workplace, between the governors and the governed and around the whole globe, is the latest leitmotif in the story of progress that cultures tell themselves. In advertisements the computer screen is an sign for assured worth, technological excellence. High street windows display a glittering cascade of CD ROM's, computers and the means to bring together television, telephone, fax, and Internet. The result is a system that mediates and combines education, entertainment, work, leisure and, looking to the future, citizenship.

The quotation above, however, was not from a product brochure or cybertechnology magazine. It was the opening section of a report to the European Commission entitled Europe's way to the information society: an action plan ( European Commision, 1994; see also Bangermann, 1994). Its recommendations, in line with other recent reports were: " .... an acceleration of the liberalisation process and the achievement and the preservation of universal service and the Internal Market principles of free movement. ..... The deployment and financing of an information infrastructure will be the primarily responsibility of the private sector." The tone of the report was urgent: "The race is on at global level, notably with US and Japan. Those countries which will adapt themselves most readily will de facto set technological standards for those who follow."

Computer technology is a symbol of where our culture is heading, of economic progress, of openness and of a free market philosophy for communication and governance. Interestingly, sensational treatments of cyberculture in popular magazines are often rather more modest than the technological objectives of corporations and universities. Such magazines may carry articles on sex-games in virtual reality, but it is in internal publications by the head of British Telecom's research laboratories that we find discussions of matter transference and of electronic implants for the direct creation of pleasure (Pearson & Cochrane, 1995).

The economic effects of computer technology have been clear for decades now and contribute to what Baudrillard calls the transition to an economy of signification. (Baudrillard, 1993, ch. 5). These promote the compression of space and time that Harvey presents as a central to the postmodern cultural condition (Harvey, 1990, ch. 12). These darker surmises about the role to computer technology are a much needed counter balance to the more upbeat celebrations by those who have a stake in what it can do for their own community of interest. For example, Bill Gates' book, 'The Way Ahead' on the cultural impact of computers was lauded as visionary. Now, since he is so instrumental in building it, he quite probably sees better than most what lies ahead. However, he pays little attention to how these developments fit within wider cultural patterns. For example, in an interview in a UK newspaper published roughly at the same time as, he speculated on the way the Internet might alter the way in which ideas and communities of interest form within society:

"To me, cultural homogenisation is one of the most fascinating questions. The broadcast media absolutely homogenise culture, absolutely. Take the distribution of books, some sell extremely well and then there's quite a tail-off .... the Highway is the ultimate distribution system ... more niches will grow up, people who want to read Sanskrit or whatever. Or maybe it'll make things more centred. It comes back to something we don't understand about human interest. .... But it's not really my thing. I'm not willing to spend lots of time thinking about this. Just because we're involved in building the system does not mean we know how it will be used."

Such disingenuous pursuit of technology for its own sake needs to be treated with caution. In fact, the evidence is that the internet, in concert with the media, entertainment and other aspects of screen culture, is accelerating the already rapid loss of cultural diversity. In Europe, for example, the Internet is seen as a means for resource transfer from the West into central European state where they are much needed. However, with these resources come habits and skilled practices that have to be developed if the resources are to be of use. These in turn render the users dependant on the corporations from which the resources originally came. The spread of the skilled practices that go with the Internet are a form of intellectual and economic colonialism.

In somewhat the same vein, at global, national and local levels networked systems are seen as a new resource to facilitate social and political relations. Organising, deciding and consulting are social/political processes that are now carried out via digital technology. There are great media and political celebrations of the democratisation and openness that this will bring. Other images of cyberculture are more dystopian. In "Blade Runner" malign replicants inhabit a world of disempowerment, corporate giganticism and squalor.

Between these two fictions, the fact is that cybernetic technology is now central to the flow of information, resources and power that characterise late capitalism (Harvey, 1990). Economic and political forces of a new sort are now, autonomously, running well beyond human control. On the more local scale dealt with in this article, these forces manifest themselves as non-human intelligence and communication skills. As they participate the arena of human social interaction they are creating the material conditions within which human consciousness will be formed in the coming century. The technology of the next few decades may bring the distinction between human agents and intelligent artefacts quite sharply into question. As layer after layer of interactive and responsive artefacts appear within the zone of proximal development, this distinction may become significantly more difficult and less important to make for those who develop within this envelope created by computers and the Internet.

Although the rhetoric of the Internet is that of the exotic frontier, its reality is more urban and mundane. As Walter Benjamin foresaw, as technology evolves from production and distribution into reproduction and circulation, human values and sensitivities change towards the urban and the erzatz. The simlulacrum becomes the real and the authenticity of cultural icons is challenged. Technological evolution now has an even more direct impact on sociocultural learning, and the values it transmits. While technology is a cultural product, culture is now more directly a product of technology.

This loop is energised by the power structures of late capitalism. These have passed well beyond human control and are leaving a destructive trace on the natural as well as the built environment (Pickering, 1996). This amplifies a condition about which Walter Benjamin warned. He observed that as the force of technology passes out of cultural control, human relationships become mechanised. Emotional and aesthetic life too, become separated from the forms and practices which have traditionally given them meaning. As Benjamin put it: The technique of reproduction detaches the reproduced object from the domain of tradition Although here he was referring mainly to the visual arts, his analysis applies to other aspects of culture as well. For example, in the varieties of sociocultural learning identified by Rogoff, it is becoming possible to see how many, if not all, of the roles now played by human beings could be played by non-human agents. As technology begins to both mediate and to participate in social practices so will they become detached from the human interactions that are their original domain of tradition. For example, the interaction from an early age with quasi-human artefacts may cause children to extend to them the 'theory of mind' that is proposed by a number of psychologists as the basis for human social interaction (Astington et al., 1988).

Benjamin noted celebration of violence as signifying that technology had become uncontrollable. As Baudrillard's account of the Gulf War shows, just such celebration has been amplified by cybernetic circulation (Baudrillard, 1991). While children play with combat videogames that incorporate clips from news broadcasts, news broadcasts about 'real' violence may often have been fictions, for security reasons. This characteristically postmodern confusion of the real and the simulated is now set to spread into social relations. If children develop within an envelope where social interaction with artefacts is a model, a subtle violence may be done to our capacities to interact with other human beings.

Conclusion.

The incorporation of technology into human practices will accelerate and deepen as artefacts simulate basic social capacities. Human relations will be technologised to the extent that such artefacts are able participate as agents in social interaction rather then merely to mediate it. The encounter with these artefacts will occur earlier and earlier in human development. They will thereby take part in the sociocultural learning by which skilled practices, and the values they express, are transmitted. The attribution of human like agency to artefacts will change the image of both machines and of human beings.

Theories of psychology and evolution that emphasise internal cognitive or genetic mechanisms will not be much help in understanding what is going on here. Neglecting the cultural conditions surrounding human consciousness in order to study the form of cognition leaves psychology at an impasse (del Rio & Alvarez, 1995, page 229). Fortunately, both psychology and biology are shifting away from such reductionism. Both are now paying more attention to the role of action in development and in evolution.

Moving beyond these restrictions makes it easier to bring these sciences together with anthropology and ethnography. This blend of disciplines is needed to understand how technology supports thought and creates human values. In the longer sweep of cultural evolution, and the broader perspective of anthropological and ethnographic observation, it is clear the use made of artefacts by human beings shows that technology is assimilated to human practices rather than the other way round. However, the advent of intelligent artefacts is bringing to light the neglected role of both technology and value within the foundations of Western philosophical traditions (Collins, 1996, page 3). Technology shapes the cultural conditions within which people develop their ability to live together. These conditions now include agent like artefacts with which human beings will need to co-exist. Given the destructiveness of contemporary geopolitical forces, a better ethnographic understanding of the impact this will have on human social relations is urgently needed.

References:

Armstrong, R. (1996) Cryonics. Artifice, Issue 4, pages 28 - 39.

Astington, J. W. et al. (1988) Developing Theories of Mind. Cambridge University Press.

Bangerman, H. (1994) Europe and the global information society. Report to the June 1994 meeting of the European Council. European Commission, Brussels. (At http://www2.echo.lu/eudocs/en/bangemann.html)

Baudrillard, J. (1991) The Reality Gulf, The Guardian, January 11, page 25.

Baudrillard, J. (1993) Symbolic Exchange & Death. Sage, London.

Benjafield, J. (1996) Cognition. Second Edition. Prentice Hall, New Jersey.

Benjamin, W. (1979) The Work of Art in the Age of Mechanical Reproduction. In Illuminations, translated by Zohn, H., pages 219 - 253. Fontana, London.

Bourdieu, P. (1984) Homo Academicus. Translated by Peter Collier, Polity Press, Cambridge.

Brooks (1991,a) How to Build Complete Creatures Rather than Isolated Cognitive Simulators. In Architectures for Intelligence, edited by Van Lehn, K.. Erlbaum, New Jersey.

Brooks (1991,b) Intelligence Without Representation. Artificial Intelligence, 47: 139 - 160.

Brooks, R. (1991,c) Intelligence Without Reason. Proceedings of the 12th. International Joint Conference on Artificial Intelligence, (IJCAI 12), pages 569 - 595.

Bruner, J. (1990) Acts of Meaning. Harvard University Press, London.

Campbell, D. T. (1974) Evolutionary Epistemology. Chapter 12 in The Philosophy of Karl Popper, edited by Schlipp, P., Open Court Press, La Salle, Illinois.

Churchland, P.S. (1986) Neurophilosophy: Toward a Unified Science of the Mind-Brain. MIT Press, Cambridge.

Collins, G. (1996) Editorial: A Touch of Memory. Technema, Issue, 3. Pro Universitate Verlag, Baden-Baden.

Costall, A . & Leudar, I. (1996) Special Issue: Situating Action. Ecological Psychology, vol. 8, no. 2.

Costall, A. (1995) Socialising Affordances. Theory and Psychology, vol. 5, pages 467 - 482.

del Rio, P. & Alvarez, A. (1995) Tossing, Praying and Thinking: the changing architectures of mind and agency. In Sociocultural Studies of Mind, edited by Wertsch, J. et al., Cambridge University Press, Cambridge, UK.

Edelman, G. (1992) Bright Air, Brilliant Fire. Basic Books, New York.

Elias, N. (1989) The Symbol Theory: An Introduction. Theory, Culture & Society, vol. 6, pages 169 - 217.

European Commision (1994) Europe's way to the information society: an action plan. Final report to the European commision. No. 347, Brussels, July 19th. 1994. (At http://www2.echo.lu/eudocs/en/com-asc.html)

Fodor, J. (1990) Why there still has to be a Language of Thought, in The Foundations of Artificial Intelligence, edited by Partridge, D. & Wilks, Y., Cambridge, UK: Cambridge University Press.

Gardner, H. (1985) The Mind's New Science: a History of the Cognitive Revolution. Basic Books, New York.

Goodwin, B. (1994). How the Leopard Changed its Spots: the Evolution of Complexity, Weidenfeld & Nicolson, London.

Gray, C. (1995) The Cyborg Handbook. Routledge, London.

Haraway, D. (1995) Cyborgs and Symbionts: Living Together in the New World Order. In The Cyborg Handbook., edited by Gray, C.. Routledge, London.

Haraway, D. (1991) Simians, cyborgs and women, the reinvention of nature. London, Free Association.

Harnad, S. (1992) The Turning Test is not a Trick. SIGART Bulletin, vol. 3, no. 4, pages 9 - 10.

Harvey, D. (1990) The Condition of Postmodernity. Blackwell, Oxford.

Ingold, T. (1996) Situating Action V: The History and Evolution of Bodily Skills. Ecological Psychology, vol. 8, no. 2, pages 171 - 182.

Johnson-Laird, P. (1989) A Computational Analysis Of Consciousness. In Consciousness In Contemporary Science, edited by Marcel, A. and Bisiach, E., Oxford, UK: Oxford University Press.

Johnston, T. & Gottlieb, G. (1990) Neophenogenesis. Journal of Theoretical Biology. Vol 147, pages 471 - 495.

Kingdon, J. (1993) Self-made Man and his Undoing. Simon & Schuster, London.

Langton, C. (Ed.) (1995) Artificial Life: an Overview. MIT Press, London.

Lanier, J. (1995) Agents of Alienation. Journal of Consciousness Studies, Vol. 2 (1): 76 - 81.

Latour, B. (this volume)

Lave, J. & Wenger, E. (1991) Situated Learning: Legitimate Peripheral Participation. Cambridge University Press.

Maes, P. (1995) Modeling Adaptive Autonomous Agents. In Artificial Life: An Overview, edited by Langton, C., MIT Press, London.

Maturana, H. & Varela, F. (1992) The Tree of Knowledge: The Biological Roots of Human Understanding. Revised edition. Shambala Press, London & Boston.

Maynard-Smith, J. (1987) When Learning Guides Evolution. Nature, vol. 329, pages 761 - 762.

Molleson, T. (1994) The Eloquent Bones of Abu Hureyra. Scientific American, vol. 271, pages 60 - 65.

Monod, J. (1971) Chance and Necessity. Knopf, New York.

Moravec, H. (1988) Mind Children: The Future Of Robot And Human Intelligence. Harvard University Press.

Mumford, L. (1968) The Future of Technics and Civilisation. Freedom Press, London.

Newell, A. (1991) Unified Theories of Cognition. Harvard University Press.

Pearson, I. & Cochrane, P. (1995) 200 Futures for 2020. British Telecommunications Engineering, vol. 13, pages 312 - 318.

Pepperell, R. (1996) The Post-Human Condition. Intellect, Oxford.

Pfaffenberger, B. (this volume)

Pickering. J. (1996) Cyberspace and the Architecture of Power. Architectural Design, Issue 117, April. Academy Editions, London.

Plotkin, H. (1988) The Role of Behaviour in Evolution. MIT Press, Cambridge.

Ray, T. (1995) An Evolutionary Aproach to Synthetic Biology. In Artificial Life: an overview, edited by Langton, C., MIT Press, London.

Roberston, G. et al. (1996) Future Natural: Nature, Science, Culture. Routledge, London.

Rogoff, B. (1995) Observing Sociocultural Activity on Three Planes. In Sociocultural Studies of Mind, edited by Wertsch, J. et al., Cambridge University Press, Cambridge, UK.

Sinha, C. (1988) Language and Representation, a Socio-naturalistic Approach to Human Development. Harvester, Hemel Hempstead.

Sterling, B. (1986) Mirrorshades: the Cyberpunk Anthology. Arbor House, New York.

Valsiner, J. (1991) The Construction of the Mental. Theory & Psychology, 1991, 1(4): 477 - 494.