Virtual Society? logo


Virtual Society? Get Real!
Paper Marilyn Strathern


Abstraction and decontextualisation: an anthropological comment
or: e for ethnography

Marilyn Strathern (Cambridge University)

(pre-publication draft)


Short of becoming anthropologists [ethnographers]
of their own institutions, its members can no longer know each other.
Ron Barnett (2000: 94) on the dispersed nature of the 'virtual' university.


Where is this formidable body of research taking us? Let me take up a half-joke which Bruno Latour (1998) made at the end of his inaugurating lecture, the pun on virtual and virtuous. Can the virtual society produce new virtues he asks? Virtuous is not a word we care to use much these days, but I shall at the end try to show one situation to which it might apply. That refers to the future of higher education and to the future of research activities such as this one. For there is an important and complex relationship between the research process and the subject of the research which the Virtual Society? Programme, in the best social science tradition, maps out.

It would not surprise you that a social anthropologist would be particularly interested in the ethnographic dimensions of the programme. Several of the contributions specifically refer to studies which have deployed an 'ethnographic' mode in the course of research. But I mean something broader than this. I mean that the research programme as a whole seeks to put electronic technologies 'into' a social context, and in the way in which it [the research programme ] has opened up diverse channels of enquiry, spread over several domains, with multiple themes and cross-cutting vantage points, it has ventured into a huge and sprawling field not unlike an anthropological ethnographer. For the ethnographer's notebook is always open at the big question: where will the study take one? Not just where the subject of study is going, note, but where the study itself will take one.

However there is a seeming tautology here. If the programme seeks to provide contextualisation in order to understand the 'social and human dimensions' (Overview, see n.1) of the new technologies, it has to be because the electronic technologies themselves are already depicted as 'decontextualised'. In other words they are imagined as having the power to communicate and convey information stripped of the encumbrances of social relations, and physical limitation on travel. This, as Carrier (1998: 2; cf. Riles n.d.) reminds us, was Polyani's original analysis of abstraction in economics, a dis-embedding process, 'the removal of economic acitivites from the social and other relationships in which they had occurred, and carrying them out in a context in which the only important relationships are those defined by the economic activity itself'. The tautology is that the notion of a decontextualised process already implies or points to the wider context from it has been carved out. It would seem that this research [the VS? programme] is simply putting back in what , it has already been acknowledged, was misleadingly taken out in the first place. Yet there is more to it than this.

The idea that human artefacts may be abstracted from human life can only point to the missing 'context' -- it cannot specify or describe it. And when one comes to describe the missing dimensions (so to speak) one will realise of course that the technologies have also created their own. That is what the VS? programme means by referring to shifts and changes taking place in how people organise themselves.

e for effect
I do not want to get mired in the epistemological problems of 'context' (for an anthropological discussion, see Dilley 1999). They were there long before 'the virtual' acquired its present connotations. Latour pointed diverse the antecedents for forms of abstraction and decontextualisation -- maps and plans, scale models, spirit houses as the microcosm of society. So what does the contemporary idea of 'the virtual' add?

I am going to say that it adds the capacity of something to point to its own effects. The power of communication is communication. This is not as absurd as it sounds; indeed it has the virtue of resonating at once with virtual's original meaning and with the subsequent devolution of that meaning. As you know, the term itself has undergone a metamorphosis, from the concrete to the abstract. 'Virtual' started off as a reference to the physical qualities (or virtues) which things have, and to the effects of these qualities, like the virtual heat of wine or of sunshine. It then came to describe the state of being effective or potent. Not until the seventeenth century was the term first used of the essence or effect of qualities by themselves, so that things could be called virtual -- as one might refer to a ruler who was a virtual sovereign -- in reference to qualities that were not endowed in formal or 'actual' terms. The point is that virtual entities point to their own potency: the virtual sovereign did not need to crowned. Such efficacy appears not to need the props of human social relations or wider contexts of activity.

Technology often appears to have such internal efficacy. And a technology that doubles up with the capacity to communicate, which Euro-Americans take as itself a process of abstraction, seems irresistible. Indeed one of the probes that this programme has carried out is into the hype that surrounds ICT as offering communication and information networks free of the labour of other forms of interaction. We also know that ICT carries a symbolic load here.

Now there are social precursors to imagined technologies in this auto-enabling sense: systems (or closed system, cf Dilley 1999: 21) which generate their own rationales for existence. I have been interested myself in systems which describe themselves, the self-describing practices of institutions and organisations which represent themselves as institutions or organisations, and that is partly because of the stake which ethnography also has in description. I would not have thought of referring to such systems as 'virtual', although the epithet has been used of one set of institutional practices deeply committed to a certain form of description: audit. Indeed, audit is quite usefully thought of as a producer of virtual realities, and I pursue it along these lines because it throws up some comparisons with the kind of descriptions the VS? programme may be aiming for in parts of its research.

Self-description meets the criterion of auto-enabling efficacy: it effects what it points to. Ethnographic description is the very opposite, and means nothing unless one knows the conditions of its compilation or its theoretical underpinnings: as a producer of contexts it has no validity unless contexualised itself. Audit most interestingly can move in either direction. What I briefly examine here are its virtual qualities.

It is Daniel Miller who calls audit a manifestation of virtualism. Activities are decontextualised for the purposes of quantification, output disembedded from the complexity of organisational life. He puts this in the larger context of political and economic shifts where, he argues, the consumer is silently transformed into a virtual consumer. 'The paradox is that, while consumption is the pivot upon which these developments in history spin, the concern is not the costs and benefits of actual consumers, but of what we might call virtual consumers, which are generated by management theory and models ... [And] the rise of auditing in Britain [is thus] symptomatic not of capitalism, but of a new form of abstraction that is emerging, a form more abstract than the capitalism of firms dealing with commodities' (1998: 204-5).

However, here I want to pursue the notion that it is not only abstraction which decontextualises, so too does the perception of auto-enablement or efficacy: communication communicates. Audit enrolls various social mechanisms which confirm its internal efficacy. For instance, it can evaluate the results of social processes without having to deal with the processes themselves. One can get straight to the end-result to the extent that it possible to bypass these other processes. Let me show how this works.

In the expanded sense which it has recently acquired (Power 1994, 1997), audit has come to be used for almost any investigation involving analysis ('reckoning' or 'accounting'). Girton College was visited last year by managers from a well known plc on an MBA course at at a well known university who wanted to do a study of an organisation different from their own: they spent the day asking people questions and looking at documents, a kind of TQA without the sting. They had already decided that what was going to differ was the culture, and the operation was called 'cultural audit'. They also knew the graphic form in which their results would be presented: 'a detailed Cultural Web'.

That indeed is the nub of audit, culturally speaking, and I am back into my own language now, at least as it operates through the RAE and TQA in British higher education. The form in which the outcome is to be described is known in advance. You bypass having to construct the form from the investigations themselves. The investigation -- the research if you will -- is in that sense retrospective ; that is, it works backwards from the bottom line up, from the categories by which accountability (say) can be ascertained to the evidence for it. The outcome of the audit report is a report on the outcome of the institution under scrutiny. We all know what this means in academic life: research or teaching output. In looking at outcome as output, audit looks at the effect of productive processes. It may or may not be interested in the processes themselves, as revealed say in RAE 5 & 6, and usually only in order to verify the bottom line. In short, audit produces its own effect insofar as the report on outcome / output takes a form it itself creates.

Interest in outcome leads to management accounting or policy-directed audit (cf. Harper 2000: 37). That is, you measure the accomplishments (output) in the hope that you can improve the system. And here there is an inbuilt little gadget that gets over the problem that academics in higher education traditionally accounted for their output in terms of (student) generations -- the reputation books acquired over decades or the success of former pupils in their middle years of life. The gadget means that you dont have to wait a generation or two. You can speed up the process. It is very simple: you turn the system of measurement into a device that also sets the ideal levels attainment. In short, audit measures become targets (Hoskin 1996; Macintyre 2000). They collapse the is and the ought, continuing a long process that began when examination results became aims -- when a high score is not simply how you measure up but is a level you have aimed and striven for.

And how -- in turn -- is that device created? The mechanism is already there in the focus on outcome: by bypassing descriptive observation, or rather by restricting the output (results) of observation to data suitable for constructing indicators. So in teaching, for example, you do not have to wait for the results years hence but can create a simulacrum of what the classroom should be producing (what its outcome or effect should be) through a map or plan, a spirit house, in the form of a set of indicators.

In this folding in of processes on one another, it is no surprise that indicators come in turn to have a life or efficacy of their own; no surprise that there is much complaint about their arbitrariness. Tsoukas (1994: 4) describes the self-defeating specification of performance indicators as it might apply in the sphere of local government. In 1993 new regulations meant that local authorities in the UK had to publish indicators of output, no fewer than 152 of them. The idea was, he reports, to make Councils’ performance transparent and thus give them an incentive to improve their services. So even though elderly people might want a deep freeze and microwave rather than food delivered by home helps, if the number of home helps is the indicator for helping the elderly with their meals then an authority could only improve its recognised performance of help by providing the elderly with the very service they wanted less of, namely more home helps. Local authorities would aiming for high scores, and the language of indicators take over the language of service.

It is interesting of course that ICT is a highly visible ally of audit practices. Its speeding up of the performance of office equipment does not just facilitate the production of the audit reports and so forth, but as an entity in itself (as ICT or IT) can be used as an indicator of performance. HE audit visits frequently exhort the use of office-like technology while also managing to suggest that knowledge itself is at stake: it is all contained in the phrase 'information technology'. My point is not the benefits (or otherwise) here, but the obvious symbolic status of ICT. Using IT in teaching has been a criterion in the TQA's evaluation procedures; this is how teaching in the discipline of anthropology was chastised:

[T]he overview report [on anthropology] highlights some areas for further attention of the institutions and of anthropologists. Whilst information technology (IT) provision is generally adequate to meet current needs, there is significant variation in the use made of IT and a lack of emphasis on IT skills development (Glasner 1996: 8).

Suggestions for useful improvement and proof (measures) of improvement merge. As a self-evidently useful technology, ICT points to itself. In the same way, the self-evidently good practice of auditing emerges as a virtual activity pointing to its own effects, captured by Power in his reference to the auditing loop which creates auditees.

'Loop' is apposite. As a descriptive practice, audit cannot afford to tolerate loose ends, unpredictability or disconnections. It carves out its own domain of what is going to count as 'description'. This means that what may be brilliant accounting is bound to be poor sociology, and -- very obviously -- different from an ethnography of an institution or organisation, or of society for that matter. But what lies behind this distinction?

Itself a system, audit elicits a view of an institution or organisation as a system -- as system, not as a 'society'. Ethnographic practice, on the other hand, as we have seen at various junctures in these VS? research projects, elicits the open-endedness of institutions and organisations as 'society'. What characterises people's behaviour in 'society' is precisely their capacity to tolerate loose ends, to deal with unpredictability and revel in the disconnections which mean that they live in multiple worlds, traverse different domains. This is where ethnography, in the broad sense in which I have been using it, as a social science practice of contextualisation, comes into it own.

e for ethnography
Ethnography does not measure accomplishments in the hopes of improving the system. More to the point, its open-ended procedures refer both to the manner in which observations are made and to the process of compiling a description. Far from truncating description, it has its own search engine in the form of a question: what connections are going to be useful? That is simply because one cannot always tell in advance; more strongly, it puts one in to the situation of not necessarily wanting to tell in advance.

Indeed, as these studies attest, ethnography throws up the unplanned, the counter-intuitive, the unpredictable. It tolerates disconnections. You dont have to tie up all the loose ends; on the contrary there may be data there that will only become resource from some vantage point in the future. But how does it create this situation? The device is that of crossing domains (and thus precipitating differences 'between' domains). It refuses to be confined to a single domain of context or narrative. On the contrary, its tracks people's activities and narratives as they cross domains, and thereby create heterogeneous social worlds for themselves.

Let me refer to a specific project which included anthropologists. Along with Jon Agar, Penny Harvey & Sarah Green's (1999) account of people setting up an inter-system network of communications in order to turn Manchester into a virtual city describes a significant and absolute requirement of electronic technological systems: they must be compatible (made into a single system, inhabit the same circuit / domain) before they interact. Interaction is an outcome or effect of their compatibility. Society, to the contrary, effects interactions between people which may, but may not, lead to compatibility. In other words compatibility may be an outcome; but many other states of being could be the outcome too. Social interactions are thus inherently open-ended -- can encompass diverse aims and intentions -- because communication between persons does not require compatibility between them in advance. From this comes much of the creativity and energy of social life. Predictability and unpredictability go hand in hand.

It follows that an interaction may represent as much a potential divergence as well as convergence of people's trajectories. It also follows that if contextualisation is what ethnography does, it cannot possibly be to englobe data within 'a [single] context', make it all compatible, but instead emphasises the differentiations of social life, its own preconditions for traversing contexts and crossing domains. The little social engine that makes this possible are of course relations.. By this I refer doubly to the conceptual relations that link data and to the living relations people have with one another.

Virtual relations
I return again to the address which Latour gave at the start of the project, his evocation of the long history of virtual artefacts as bureaucratic procedure, plans, representations, as in the Papua New Guinean spirit house which lays out a map of social groupings. Evidently, making plans of what is imagined to be 'society' is not the same order of phenomenon as society as such. It is not just that plans never exhaust all there is to know but that they never replicate how people move through the society on the ground. They do so by virtue of the relations and connections they sustain.

Connections and relations, like society itself, require imagining (Battaglia 1994) and may be (virtually) imagined as virtual as well as (imagined as) actualised in interpersonal relationships. Indeed, historically speaking, the very term 'relation' referred to abstract concepts before it referred to connections between persons (Strathern 1995). It is arguable that for as often as abstraction may be regarded as taking content out of human interactions, it is also inherent to the very imagination of social relations; conversely, actualisation is inherent in the ability to inhabit as well as cross domains.

Latour emphasised the way in which ICT materialises 'society' by making its contours visible, although any of a number of technologies or practices can do this. I would underline one of the themes of the projects being discussed here and point to what is distinctive about ICT. What is distinctive is the fact that it mobilises society in a particular mode, not just as a network of communications but as the network of connections or relations which carries communications. Here the hype and rhetoric surrounding ICT is an important part of the technology itself. The fact that the rehtoric then comes adrift against 'actual' social relations is beside the point: it has already provided a measure of sorts for them. ICT points to the connecting power of connections: that is its virtue.

Virtuous relations
What we chose to describe as virtual and choose to describe as actual will depend on the purposes of the moment. But those two terms harness certain qualities about social life which we know must always work in tandem. Obviously, the point holds across numerous contexts; what makes this particular research programme important is that it addresses a field of phenomena (ICT) which in the popular imagination stands for the possibility of actualising connections otherwise beyond reach except in a virtual state. All kinds of attributions of power come in the wake of this, and with it the deflations that attend its actual actualisation. But, then, 'Realising Virtuality [could never be] actualising potential, because what is in the realisation of Virtuality is unpredictable' (Latour 1998: 6).

If one were looking for a methodology to grasp this interplay between imagination and the unpredictable outcomes of lived social relations, one could do worse that the social scientist's ethnography (and especially of course the anthropologist's ethnography). Ethnography does in particular what social science does in general: its always adds a dimension to a phenomenon and thus refuses claims to self-sufficiency. It introduces, in the simplest way, numerical complexity of the most significant kind (Macintrye 2000): there is (always) more than 'one' thing to consider (and cf. Battaglia 1999; Riles n.d.) This I see as the basis of the VS? Programme; this is what research always adds.

What is a painfully obvious as a truism about knowledge, as a truism about society also contains a germ of truth worth holding onto. The activity of description reminds us that we live in a world full of people's descriptions of what they do. And these multiply the unpredictability, quite as much as the predictability, of outcomes. This awareness I take as something approaching a virtue, although we could translate its implementation into the more humdrum language of skills.

Knowing how to deal with unpredictability, that is, with the effects of the unforeseen, the added elements we never thought of, is a skill which university academics, as well as citizens, are going to need in the future -- if , that is, we follow Ron Barnett's (2000: 62) chilling / stimulating vision for realising the university in an age of supercomplexity. Complexity he defines as a condition brought about by the surfeit of information, supercomplexity as that surfeit stacked and multiplied through a surfeit of frameworks for processing the information. Indeed he invites us to think of a world changing so fast it is radically unknowable. He invokes a constellation of concepts -- uncertainty, unpredictability, challengeabilty and contestability -- which renders all frameworks fragile. Or, we could say, all procedures (frameworks) for abstraction and decontextualisation become visibly contingent. One needs therefore methods of enquiry that will allow one to jump frames. And we can only practice that skill on disconnections we already have to hand.

Theoretical frameworks, like other models, can be thought of as virtual systems. Their effects are perceived in their own terms (how far does the data fit into the framework? Yet such self-realisation can only be momentary. We might, to the contrary, say that we only encounter effects when they are effected, that is, already realised (contextualised) and thus translated, as Latour noted, into something else by the contingencies and unpredictabilities of their materialisation. Research replays the essential disjunction between any imagining of our condition and social life as a fabrication of divergences and of events quite unforeseen. The research process itself introduces disconnections. And VS? does so in a field which is of course all about connection.

It is just this kind of interplay which I think Barnett is after in defining how the university of the future might realise itself.

'[T]he university is an institution that (i) contributes to our uncertainty in the world (through its research and consultancy); (ii) helps us to monitor and evaluate that uncertainty (through its work as a centre of critique); and (iii) enables us to live with that uncertainty, through both the operational capacities and the existential capacities it promotes (in its pedagogical activities)' (Barnett 2000: 69).

Note that he gets rid of the idea of knowledge as an abstraction (we do not need such a metaphysical concept he says). Research and teaching have to cut themselves loose from the axiomatic authority of knowledge and embed themselves (he more or less says) in a real world of uncertainties. At the same time the university has a role in promoting certain virtues [not his term]: to set the informed apart from the uninformed, to distinguish enlightenment from ideology, and to monitor the kinds of accounts we give, in a context that is never single, always multiple.

Barnett argues that the university is uniquely placed to help the world address the condition of supercomplexity. In other words, the condition the universities have help create they can also help us to live with. This is not a condition which to be remedied but it is one which has to be lived. The university is 'realised' in the kinds of intellectual responsibilities [this is his term] it takes on.

I am not sure I follow this argument entirely. For instance, I would take complexity not as the piling on of more of the same (uncertainty upon uncertainty) but as the intermeshing of different orders of phenomena, having to take certainties and uncertainties together. But his argument does delineate a difference between accountability, rendering an account to those to whom one is accountable, manifest in the self-evident efficacy of audit, and responsibility, which is discharged to those in one's care, whether students or colleagues or the wider public. What the Virtual Society? Programme gives me is insight into some of our choices in how to describe these dimensions. If one describes them as the difference between 'virtual' and 'social' duties, then one is taking virtualism's viewpoint; if one describes them as equally the outcomes of social process, equally social constructed, then one takes society's viewpoint. The question, among many others, which the Programme is poised to answer is what some of the consequences of these divergent views are likely to be. For sure, we always need more than one.


Battaglia, Debbora 1994 Retaining reality: Some practical problems with objects as property, Man, (n.s.) 29: 631-44.

Battaglia, Debbora 1999 Towards an ethics of the open subject: Writing culture in good conscience, in H. Moore (ed), Anthropological theory today, Cambridge: Polity Press.

Barnett, Ronald 2000 Realizing the university in an age of supercomplexity, Buckingham: The Society for Research into Higher Education and Open University Press.

Dilley, Roy 1999 The problem of context, Oxford: Berghahn books.

Carrier, James 1998 Introduction to Virtualism: A new political economy, D Miller & J Carrier (eds), Oxford: Berg.

Glasner, Angela 1996 Teaching and learning in anthropology: the HEFCE review of anthropology provision in England and Northern Ireland, Anthropology in Action, 3: 7-9.

Harper, Richard 2000 The social organization of the IMF's mission work: an examination of international auditing, in M. Strathern (ed), Audit cultures: Anthropological studies in accountability, ethics and the academy, London & New York: Routledge.

Harvey, Penelope & Sarah Green 1999 Scaling places and networks: an ethnography of ICT 'innovation' in Manchester, presented at the Internet & Ethnography conference, Hull 1999.

Hoskin, Keith 1996 ‘The awful idea of accountability: inscribing people into the measurement of objects, in R Munro & J Mouritsen (eds), Accountability: Power, ethos and the technologies of Managing, (London, Int. Thomson Business Press).

Kogan, M and M Henkel 1998 Policy changes and the academic profession in England, European Review, 6 pp. 505-23.

Latour, Bruno 1998 Thought experiments in social science: from the social contract to virtual society. First Virtual Society? Annual Public Lecture, research/virtsoc/text/events/latour2.htm

Macintyre, Michael 2000 Audit, education and Goodhart’s law, Or, taking rigidity seriously,

 Miller, Daniel 1998 A theory of virtualism, Conclusion to Virtualism: A new political economy, D Miller & J Carrier (eds), Oxford: Berg.

Muckersie, D 1996 Continuous assessment, Managing Higher Education, pp. 13-19.

Power, Michael 1994 The Audit Explosion, London: Demos.

Power, Michael 1997 The Audit Society: Rituals of Verification, Oxford: Oxford University Press.

Riles, Annelise 1998 Infinity with the brackets, American Ethnologist 25: 378-98.

Riles, Annelise n.d. Seeing through virtualism: formalism and the propertization of time, for panel 'Temporal horizons of property', American Anthropological Association meetings, 1999.

Scott P in press The impact of the Research Assessment Exercise on the quality of British science and scholarship. Anglistik

Shore, Cris & Susan Wright 1999 Coercive accountability: the new audit culture and its impact on anthropology, Journ. Royal Anthropological Institute , 5: 557-75.

Shore, Cris & Susan Wright 2000 Coercive accountability: the rise of audit culture in higher education, in M. Strathern (ed), Audit cultures: Anthropological studies in accountability, ethics and the academy, London & New York: Routledge.

Strathern, Marilyn 1995 The relation: Issues in complexity and scale, Cambridge: Prickly Pear Pamphlet no. 6.

Strathern, Marilyn 1997 'Improving ratings': audit in the British university system, European Review, 5: 305-21.

Strathern, Marilyn 1999 The ethnographic effect, in Property, substance and effect: Anthropological essays on persons and things [collected essays], London: Athlone Press.

Tsoukas, H (1994) Introduction: from social engineering to reflective action, in organizational behaviour, in: H. Tsoukas (Ed) New Thinking in Oganizational Behaviour: from social engineering to reflective action, Oxford: Butterworth / Heinemann.

Velody I (ed) 1999 special issue of History of Human Science


I am grateful to Steven Woolgar for his invitation to deliver this talk and thus to learn something about the Virtual Society? Programme close at hand. I take my cue from his own interests in audit (eg. 'Accountability and identity in the age of UABs', CRICT Discussion paper no. 60, 1997), and from his 1999 characterisation of the Programme's initial findings: 'Many of these first findings are counter-intuitive' (Virtual Society? Profile '99, p. 5). That, it seemed to me, was itself a finding worth exploring.

My closest acquaintance with individual projects has been with Social contexts of virtual Manchester (Penny Harvey, Sarah Green and Jon Agar), and the stimulus of that work on several levels will be very evident here.


Conference Overview
Conference Programme
Conference pictures


Return to top of page
Switch to text only version

Page developed by Christine Hine
Page maintained by Marike van Harskamp
Contents current at 16th May 2000