Gavin Kendall and Mike Michael

1. Introduction

This paper is concerned with the complex interdigitation of the works of Michel Foucault and Bruno Latour (and more generally, actor-network theory).In particular, we are keen to develop, through an exploration of the commonalities, differences and tangentialities between these authors, new ways of thinking about the interrelations, mutualities and patternings of order and disorder in social life (conceptualised as heterogeneously constituted).

First, we examine the ways in which Latour and Foucault respectively deal with technology and/or technique in order to draw out the role of ‘uniqueness’ in processes of heterogeneous dis/ordering.

Second, we consider the ways in which Latour and Foucault deal with the subject, and its relations to identity.In particular, we note that underpinning both these thinkers’ treatments is a view of the human (but also, in Latour, of the non-human) as at once fixed and fixable, but also fluid and ‘fluidable’: whereas for Latour the emphasis that is placed upon fixability and fluidity is allowed de jure, for Foucault fluidity has a de facto status which ‘prompts’ the various techniques and technologies of self (fixity).We suggest that neither of these approaches is satisfactory, for the fluid-ised self/identity is never knowable, except as ‘in-process’, and the fixed subject/identity is too knowable – a ‘mere’ cipher for a network or assemblage (Latour) or discursive formation (Foucault).We suggest that self/identity is most interesting at those points of transition from fluidity to fixity and vice versa.

Third, we compare and contrast Foucault’s and Latour’s tacit models of temporality as they structure their historico-empirical studies, showing how in deploying differing time frames and different narratives of continuity (which, in turn, evoke different versions of self/identity), they produce different accounts of ordering processes.In contrast, and drawing on Michel Serres (which we do throughout), we shall suggest, in counterpoint, that temporality should be seen as emergent in the dis/ordering processes that we identify.In conclusion, we draw the three themes of technique/technology, self/identity and ordering/temporality together to show, in that they inform one another, they can serve as the basis of a revisioning of social (that is, heterogeneous) research.

2. Technique and technology

In the later Foucault, when the philosopher-historian turned his attention to the problem of the self, he uses a pair of words with increasing frequency: ‘technique’ and ‘technology’.It takes no more than a cursory glance at The Care of the Self (1986), however, to reveal that the term used more often by Foucault is ‘technique’.This seems to be in marked contrast to the post-Foucaultian literature, which almost exclusively uses the term ‘technology’ and rarely, if ever, discusses ‘techniques’.It seems to us that this is not just a linguistic quibble.Foucault does discriminate between the terms, using the French ‘technique‘ to refer to a practical instance, while the term ‘technologie‘ refers to a practical system.In a nutshell, techniques are singular and elemental, while technologies are agglomerations of techniques formed into a logical and systematic whole.When we think of this vocabulary as applied to the object ‘the self’, a technique of the self is merely a skill or procedure, possibly isolated but possibly integrated with other techniques; a technology of the self, by contrast, is something much more like a Wittgensteinian ‘form of life’ or a Weberian ‘department of existence’.This distinction, then, is important: it seems that Foucault’s endeavour was not so much to describe antique forms of subjectivity as systematised, but as made up of a variety of independent and non-systematic procedures (or techniques).In this usage, then, Foucault is closer to Marcel Mauss (e.g. 1973) than to Weber or Wittgenstein.Mauss described some of the various ‘techniques of the body’ that are used in different societies at different historical conjunctures, stressing, most importantly, their contingent form; for Mauss, there is no truly or simply human way of walking, eating or swimming, for example.Foucault’s account here is also reminiscent of Elias (1978), who dealt with the formation of the person of the Renaissance courtier: the courtier does not build up a coherent form of selfhood based on some telos, but merely takes elements from here and there, as they are pleasing.The self is a temporary agglomeration of these ‘pleasing ways’, but not especially systematic or coherent: it is, rather emergent and contingent.

We intend to leave aside the question of whether Foucault was right or wrong about the ancients.All we are interested in arguing for the moment is that he did not seem to go to any great lengths to demonstrate that antique subjectivities were especially systematic, and that the evidence for this is the rarity with which he uses the term ‘technologie‘ and the frequency with which he uses the term ‘technique‘.

Given that this is the case, it seems somewhat odd that in the Anglophone recuperation of Foucault’s work the term ‘technology of the self’ has become ubiquitous, while the term ‘technique of the self’ has become a rarity.As one may imagine, there is some reason to be concerned about this slippage because it inclines thought about the self toward the ordered and the coherent, and away from the partial, contingent and independent.If we leave aside the question of what the self ‘really’ looks like for a moment, the question of whence this lexical slippage appears must be raised.The main source of this slippage is Foucault himself.Although extremely careful in his native tongue to discriminate between instances and collections of actions upon the self, in English Foucault almost spurned ‘technique’ altogether.Further, Foucault’s initial marked use of thetechnique/technologie pair was not noticed in the Anglophone community, which has translated these terms into the singular ‘technology’.

Before proceeding with this line of argument, we should digress for a moment to mention the two most important resources Foucault made use of in his formulation of the technique/technology vocabulary. This section should also throw some light on why we maintain that the term ‘technique‘ is far more crucial to understand what Foucault is driving at.

2.1. Ancient technai

The Greeks did not use a word equivalent to ‘technology’ to describe forms of activity or creation.The Greek term ‘techne‘ (plural ‘technai‘) is closer to our term ‘technique’, but should be carefully discriminated from it.A techne refers to any skill or ability. Etymologically it is connected to the word for ‘weaving’ and, further down the road of derivation, to the word for a ‘text’, a thing that has been woven together out of words.The Greeks, then, have a very physical and, one may say, organic view of the technai.These skills are understood as practical, rather than mental, applications.A technique of the self (or techne heautou), then, is not simply a reflective sense of self but a lived and practical experience.One is reminded here of Pierre Hadot’s assertion that Greek philosophy is ultimately practical and concerns a way of life which can never be limited to the intellectual or non-corporeal (Hadot, 1995). We must bear this in mind when we read Foucault: the sense of self he is talking about is nothing like the modern idea of the reflective, intellectual self divorced from the realm of the body: it is a self formed from the playing out of ways of comporting oneself in the bios politikos, public life.

2.2. Heidegger and technology

Foucault, of course, does not just make use of the antique writers in formulating the technique/technology dyad.He also makes uses of Heidegger’s discussions of technology, which are, however, themselves derived from the ancients.When Heidegger reflects on the concept of thought, he backs a form of thought that can be opposed to reason – and his target here is Kant.

In The Question Concerning Technology, Heidegger (1977a) deals with the possibility that thought (or, at least, the new type of thought he is interested in setting out here) may be the opposite of reason.Similarly, in ‘The Letter on Humanism’ (1977b), he argues that there is a type of thought which is neither theoretical not practical.Heidegger’s attempt to distinguish thought from ‘the practical’ may seem to contradict the emphasis drawn out from Hadot above, but what one needs to remember is that Heidegger is explicitly attacking Kant (1952); Heidegger is not so much counterposing thought to practicality as to Kant’s ‘practical reason’.The type of thinking that Heidegger is getting at he sometimes calls ‘poetry’.It is an aesthetic form of thought, but it also draws on the Greek conception of poiesis, an act of creation.For Heidegger, then, there exists a rigorous type of thought, a thought that has been discussed but dismissed by Kant as ‘doing nothing’.On the contrary, says Heidegger, poetry (in the Greek sense – the Greek ‘poieo‘ is the simple and ubiquitous word for ‘I do’) is highly creative and is always located in the material: ‘poetry is the act of establishing by the word and in the word…Poetry is the establishing of being by means of the word…Poetry is the supporting ground of history’ (Heidegger 1949: 280-3). The discussion of ‘poiesis‘ and ‘techne‘ in Heidegger (1977a) reveals the nostalgic nature of his thought; for the Greeks, ‘poiesis‘ was simultaneously a simple act of doing, as well as the creation of something, whether a poem or a mundane object.Heidegger believes that we have forgotten, especially through the person of Kant, the link between the aesthetic and the everyday.This is why Heidegger says in The Question Concerning Technology that art is finished after the Greeks (for a fuller discussion of these issues, see Kendall and Michael, 1998).

One further point is of interest here.Heidegger’s term for ‘technology’ is ‘technik‘, an all-purpose German word that encompasses both ‘technique’ and ‘technology’.We suspect the subtlety of meaning available to the French and English reader, who can disentangle ‘technique’ and ‘technology’, is not available in the German.On the other hand, the strength of the German is that it directs the reader to the way in which technical activity is always ‘enframing’ (das Ge-stell) – practical, potentially transformatory, but essentially and crucially orderly and ordering.It may well be this conflation of ‘technique’ and ‘technology’ in the German ‘technik‘ that led the English-speaking Foucault, and the philosophically-informed reader of Foucault, to follow a Heideggerian path whereby a distinction between practical instances and practical systems is unmarked.This may shed some light on the Anglophone preference for ‘technology’ over ‘technique’, and perhaps also for Foucault’s failure to discriminate between ‘technique’ and ‘technology’ in English.

In spite of these problems, it is worth stressing that what we wish to take from Foucault, contrary to almost all other Anglophone readings, is the notion of uniqueness: that is, every technique has its own conditions of possibility, every technique has its own specificity.As we shall see when we move on to our discussion of Latour, this emphasis is valuable for our purposes.

2.3. From technique to socio-technological systems

We have argued so far that Foucault’s concept of techniques of the self is built upon the ancient Greek technai and upon Heidegger’s notion of technology.Given that the latter is also developed from a particular reading of the Greek writers, it can be argued that Foucault contains a double dose of Hellenism.We should also like to suggest that the use of the term ‘technology’ in the post-Foucaultian literature may sometimes be a little careless; one of the reasons why we think this is the case is that it inclines the reader to ignore another realm that we may call ‘technological’.More specifically, we wish to discuss the realm described by Bruno Latour as the ‘socio-technological’ – the realm where human and nonhuman (technological) actors live and work together.

Latour’s (1992, 1993, 1999) critique of sociology (and the other social sciences) targets the way in which it proceeds by ignoring the natural and technological actors (or ‘actants’, as he terms them) that are necessary for any social situation to work. Together with his colleague Michel Callon, Latour has pointed to the role of natural actants (scallops, yeast), technological actants (car seat belts, automatic doors) and social actants (humans, corporations) – but for Latour the distinctions between these levels are problematic.Latour suggests that a more complete social analysis must include the sorts of actants we have previously been at pains to exclude from our enquiries.

It is not our intention at this point to gloss Latour’s work any further, but when we return to Foucault’s ‘technologies of the self’, what should be apparent is that there is really no technology in them.Foucault’s self is purely human and purely social, but Latour would direct us to the fact that we never see human beings in purely human/social settings.Humans are always enmeshed in a network with other actants, many of whom are nonhumans (humans driving cars, wearing clothes and spectacles, carrying mobile phones, and so forth).And while our 21st-century human being is clearly simultaneously human, natural, social, technological, etc, the same argument mutatis mutandis can be made for the ancients – and here we can say that Foucault’s emphasis on the antique does not provide an excuse for ignoring the nonhuman.The challenge here is to conceptualise a form of self that encompasses and includes the technological.For example, as we sit at our desks and write this essay, we are ordered and constrained at the same time as our thoughts are liberated by Apple Macintosh, ‘Microsoft Word’, the internet facilities that allow us to work together over distances of thousands of miles, the tables and chairs at which we sit, the layout, heating and air-conditioning of our offices, the social rules about when it is appropriate to work, university authorities’ attempts to increase published output, and so forth. A technology of the self must simultaneously apprehend all of these dimensions: social, natural, and technological and, importantly, their sometime systematicity.

2.4. Between technique and technology

In comparing Latour and Foucault, we can make the following observations.Foucault’s sensitivity to uniqueness – the ad hoc method that arises in isolation (technique) and may or may not cohere with other techniques to produce a system (a technology) seems a particularly useful counterpoint to Latour’s emphasis upon the systemic.Of course, for Latour the emergence of systems or networks is constantly being reproduced moment by moment (an ethnomethodological given), and this entails ad hoc (or what we might term ‘technical’) processes. Nevertheless, as we have seen, Latour stresses systematicity – the tying together of actors through the movements of intermediaries which must be more or less routinised if a network is to be successfully put together and operate durably.We see this most clearly in Latour’s ascription to particular technological artefacts (that is, those technological artefacts that are embedded within material-semiotic systems) the status of ‘missing masses’ that serve in the reproduction of social order.While this emphasis upon the process of ordering seems one-sided, it does supplement Foucault by injecting technology more forcefully as an actor in the process of ‘caring for the self’, of ‘governmentality’, of discipline and so on.And the reader here should note the precise terms that we use here to convey the force of Latour’s thought – it dwells on the ‘technological’ (the systematic, routinised, enframing) rather than the technical (the singular, the discrete, the ad hoc).

If we take these two contributions together – if we squint so that we can keep both technique and technology in at least partial focus – we can embrace the relation between uniqueness and systematicity: let us call those technological artefacts involved in this double movement of uniqueness and systematicity ‘artefactual nonhumans’ (to distinguish them from, and to encompass though not ‘resolve’, the dichotomy of technique/technology). That is to say, we can begin to trace how ‘artefactual nonhumans’ are complex distributions, relationalities and assemblages, that are systematised and yet which are also moments of uniqueness, of novelty.They serve in processes of ordering (and thus heterogeneous reproduction) and disordering (and thus heterogeneous production). Let us take as an example the ‘artefactual nonhuman’ known as the television remote control (cf. Michael, 2000). Clearly, this can be regarded as a component in the socio-technological system or network that includes at minimum the television, TV companies and manufacturers, consumers, infra-red light.However, the remote control is also liable to be lost.As Michael argues, this is due in part to a coincidentalisation of technologies – in particular the consonance in the designs of the remote control and the sofa (or couch). What makes this consonance possible are structuring assumptions concerning the hand (the remote is designed to fit into one; the sofa is designed to allow the hand entry for removal of cushions, opening of sofa beds, etc.).As such, a pathway is opened whereby the remote control can get lost.The response to this occasional breakdown in the socio-technical system (in which the remote control is embedded) is the local innovation of new, more or less unique routines – that is, techniques – that guard against such loss.For instance, the remote control can be systematically located in some ‘safe’ place after use. However, these routines can also reflect more systemic conditions such as power relations within the family.To the extent that such relations are partly mediated by the father’s monopoly of the remote control (cf. Morley, 1992), it may well be the arm of father’s chair on which the remote regularly and safely settles.Further, such mishaps also come to be articulated and fixed by more ‘global’ socio-technical actors: thus Philips produces Magnavox Remote Locator colour televisions which help TV viewers swiftly re-locate their lost remote controls by pressing the TV’s Power-On button. This example suggests that certain engagements with technology as materials lead to disorder (the missing remote, the lumpy sofa), points of change (a realignment of viewers and viewing from the floor to the chair), new relations of power and disciplines (as the battle for rights to use the technology is joined), points of change (new technologies are invented to deal with the existing socio-technical problem), and so forth.This allows us to focus on the moments of transition where technique becomes technology (the individual, ad hoc problem turns into a routinised solution – from the lost remote control to the Magnavox Remote Locator), and technology technique (the forms of routinisation themselves develop ad hoc, singular problems from the useful device which saves us having to get up and go to the television to change channels to the useless device stuck down the side of the sofa).In this way, we can go beyond the lexical slippage between the two terms to suggest that when considered as a pair, technique and technology- that is as ‘artefactual nonhumans’ – allow us to apprehend uniqueness, materiality, change and ordering.All of these elements exist in the space between the practical instances of techniques and the practical systems of technologies.

3. Identity and Self

Once again, our starting point for this section is with a nuance of language, a distinction that might, at first glance, seem trivial.In the Anglophone social sciences, the term ‘identity’ is used frequently (and carelessly) to refer to some sense of self, and especially to a self-reflective sense of self.Michel Serres (1998) has reflected on a basic problem with the use of this term which serves as a useful starting point for our discussion here.Serres, as a mathematician, draws attention to the literal meaning of the term ‘identity’: an exact equivalence between two entities, statements or symbols.The identity of ‘x’ is of course ‘x’; ‘x’ may be identical with ‘y’, but we would then know that there would not be much point in using separate symbols for them; or a transformation of one object may render it identical with another (2x+c=y).When one transfers this kind of understanding to the use of the term ‘identity’ in the social sciences, one can see the problem; most of the time in the social sciences, the notion of identity is used to link two entities which are far from identical (as in ‘person x’s identity is that s/he is a y’, or, more simply, ‘x is a y’).The problem here is that we are imposing identity when person x merely belongs (and probably not exclusively) to the set of entities named y.

This is not trivial: it suggests that the way we typically understand identity (and hence afortiori identity politics) is to engage in a process of over-extension and simplification.In fact, it is almost impossible to designate that which is identical with person x (that is to say, to designate all the qualities of which person x partakes); logically, it may only be possible to say something which is a truism – person x is person x.

Interesting enough, the opposite logical problem may also point up a problem for ‘identity’: frequently the tern identity is used to fix two entities that are not identical.Person x may be said to have an identity as, say, a heterosexual black woman – yet it is clearly the case that no form of ‘identity’ is evident. However, such uses of the term identity allows the commentator to fix, as if by magic, the entity rather than engaging in a description of that entity’s relationship to its putative class or classes.In short, ‘identity’ is a troublesome term, short-circuiting thought and accurate description, and giving a false sense of the mastery of an analytical category over a material reality.

3.1. Foucault and subjectivity

Foucault seems to agree with Serres, because he prefers not to use the term ‘identité‘ when talking about the self.Of course, in the later Foucault, we hear a lot about the self (soi), but Foucault also likes to make use of a variety of terms that are cognate with ‘subjectivity’, such as ‘sujet‘, ‘assujettir‘ and ‘assujettissement‘.For example, in The Use of Pleasure, Foucault speaks of a ‘mode of subjection’ (mode d’assujettissement) (1984: 27).Of course, to make some sense of Foucault’s choice of language here, it is necessary to understand something of structuralism and phenomenology.In particular, the term ‘subject’ captures some of the prevailing interests of French structuralism and phenomenology.The self is understood as the subject in a linguistic sense (and here Lacan’s thought is important) – that is to say, as notionally (grammatically) the one who speaks, but at the same time a function of a social system (language).The subject can simultaneously be the source of action (or the agent), without necessarily being the conscious originator of that action.On the other hand, one can be a subject while simultaneously being ‘subjected’ – governed by a series of external rules and conventions.Here we also get a hint of the omnipresence of power, in that the subject is both governor and governed, subjecting others while simultaneously him/herself subject to others.

With this kind of subtlety and flexibility, it is unsurprising that Foucault does not favour the prosaic and fixed term ‘identité‘.Yet Foucault is frequently used in ‘identity politics’ as a justificatory ‘great author’.It may be difficult to ground a left or liberatory politics on such an inflexible notion as ‘identity’ (which is maybe best left to the mathematicians and the logicians); and it seems that there is little use for such a relation in Foucault.Rather, Foucault uses the ‘subject’ to refer to the various manifestations of self, always-already located within discourses.Just as discourses are plural, so manifestations of subjectivity are plural. Subjectivity is nomadic, temporary, contradictory, and heterogeneous, while identity is stable, permanent, coherent and homogeneous.For Foucault, the character of discourse, which he regards as in flux and characterised by martial relationships, does not support something like ‘identity’.The subject, the temporary result of specific discursive combinations, is what interests Foucault, yet we can immediately see that the subject represents a kind of ‘disempowering’ of identity politics: if identity is an illusion, a straight-jacket description of something far more tenuous and subtle, if ‘the sides’ one takes are constantly reformulating and dissolving, then where is the authoritative place from which one can locate the self and be located in order to speak for or against a political position?It is for this reason that one sees in the later Foucault a description of forms of self that seem to be impossible to link to ‘politics’ in the good old-fashioned sense.In fact, it is possible that Foucault was drawn to skepticism by the dawning realisation of the impossibility of the self being fixed (being identical with anything except itself) and being able to judge (for a fuller discussion of the relationship between Foucault and skepticism, see Kendall and Wickham, 1999).However, Foucault does allow the contradictory self to act consistently: he likes to use the term ‘multiplicité‘ as a way of conveying the idea that a self, though multiple and fractured, can still act in ‘singular’ fashion.

3.2. After Foucault: technological humans

Foucault’s sense of the character of self or subject is subtle but underdeveloped.In particular, for Foucault the self is the equivalent of some kind of core human or social experience.While he is clear that this self is linked to the discursive and the non-discursive, he never fully explores what this might mean or how the self might be made up of elements which include the nonhuman (by the way, this is just a comment on Foucault, not an attack – there wasn’t really any reason why Foucault should have been interested in the nonhuman.We don’t want to suggest that Foucault was wrong or incomplete because he didn’t take an interest in everything.Nothing could be more fatuous).

Once again, Latour’s (and his co-workers’) understanding of the character of social networks (or, better, socio-technological networks) draws our attention to the various sorts of entities that can be counted as actants.Latour (see especially 1993) tries to show us that we should not be too obsessed about drawing boundaries between the various sorts of entities that can act – this, says Latour, is the characteristic of a now-crumbling modernism.So Latour is just as happy to discuss nonhumans (both living and technological – i.e. animals/bacteria and machines) as he is to discuss humans when he wishes to understand the dynamics of a problem or of a network.Latour insists on this ‘principle of symmetry’, arguing that we need to remove our modernist blinkers and see the enormous amounts of work in society that nonhumans do.Latour takes sociology (a typical modernist discourse) to task for seeking to obliterate nonhumans and to distil a pure human essence that is ‘the responsible agent’.

Once we have managed to escape the modernist demand that we keep separate the natural and the social, the human and the nonhuman, we can begin to see that the self or the subject can be formed out of a relation between consciousness and matter; and that the matter that matters does not have to be human.While this may sound quite tricky in the abstract, an example at the concrete level may help to clarify things.Humans never offer themselves to our gaze in a pure form – they are always combined with nonhumans: think of the blind person with his/her guide dog; the person in the wheelchair; the person with an artificial hip or leg or heart-valve.It surely does not take too much of a leap of imagination to think of how their subjectivities, their forms of self, are produced out of a hybrid human-nonhuman being.And while it might be objected that these are exceptions, all of us are carrying around, are fused to, bits of nonhumanity.It is the interaction (and this is probably a bad choice of word as it suggests a potential separability which may not exist) which forges what we might term ‘forms of self’.When we turn our attention to the so-called ‘primitive’ societies, we see that there also there is no possibility of separating out nature and society, human and animal (for example in totems – see the detailed discussion of how ‘primitives’ mix up humans and nonhumans in Latour, 1993).Perhaps the cyborg has already arrived – almost unnoticed, and in situations far removed from science fiction (see also Haraway, 1991).

3.3. Fluidity and fixity

For Foucault then, ‘soi‘ has an irreducible fluidity (‘multiplicité‘) that is the ‘object’ of discipline, knowledge, governmentality, and so forth.It is always on the verge of escaping the techniques and technologies which are deployed in power/knowledge networks – of becoming ‘deterritorialised’, as Deleuze and Guattari (e.g. 1988) put it.We may say that there is an ‘essential potential’ (de facto) in such fluidity.In contrast, for Latour, the actant is fixed (even as it is ‘in process’) in the nexus of a network (cf Michael, 1996).Any change is contingent upon other shifts in the network: there is a ‘contingent potential’ for fluidity – an in principle (de jure) movement into a deterritorialised or nomadic or rhizomic state.Or rather, there is a jump into another fixed state – another role within a reconfigured network.Where for Foucault there is always the fluidity of ‘soi‘ lingering beneath the surface of an ‘identity’, for Latour there are only other potential actant-roles.

If these characterisations hold water, an interesting issue emerges.The fluid ‘soi‘ can never be grasped – in its essence it is ungraspable, and ungraspability is its essence.In contrast, the role that Latour unravels is only too graspable: it ends up being a mere cipher for the network as a whole – it is wholly determined.Of course, various efforts have been made to overcome this model of the actor-actant (e.g. Mol and Law, 1994), and generalising beyond the human to the nonhuman, of the entity.More particularly, the singularity and territoriality of the actants that inhabit Latourian networks are feasible only because the network is itself singular.As Star (1991; Bowker and Star, 1999) points out, we all straddle multiple networks, and engage with them to varying degrees of centrality and marginality.These networks rub up against each other and rather like two blocks of ice in close contact, the pressure yields fluid.So, rather than look for fluidity as grounded in the essence of the self, we suggest that given that humans (and nonhumans, for that matter) are always relational, always between networks, then they are always on the verge of such fluidity (and this obviously links up to those moments of innovation captured in the term ‘technique’).

Bowker and Star (1999) amply illustrate this in their analysis of the many ways in which official classification systems (which are ‘technologies’ insofar as they are materialised as particular sorts of concrete, moveable documents), while reflecting the source network that created them, must also operate in local ‘receptor’ or ‘target’ networks.For example, an artefactual nonhuman transportable system for categorising the causes of death – the International Classification of Diseases (ICD) – will often butt up against local conditions (for example, the workload of physicians, national infrastructures for statistics gathering and processing, and so on) that necessitate innovative practices.The ‘implementers’ of the ICD thus work around and re-interpret its categories (some of which display Eurocentric bias), and innovate in their handing of the documents in which the ICD is partly embodied.In such ways the ICD becomes better ‘adjusted’ to local conditions.The stable ‘singularity’ that the ICD ‘offers’ to the physician (that of ‘ICD categoriser of causes of death’) is thus ‘compromised’, or rather, rendered fluid by the sorts of work-arounds and augmentations that make the ICD locally workable semiotically and materially.In other words, new singularities – ‘hybrid’ or interstitial – emerge.But these new singularities, in turn, become routinised, contingently fixed in the context of local networks.What then becomes our point of intervention, or interpretation, should be the moment of transition from the fixed to the fluid (and vice versa).

Interestingly, then, although we started out by criticising Foucault and Latour for their over-emphases on fluidity and fixity respectively, it is possible to move to a rapprochement between these two, and such a move reveals that what is crucial is the point of transition between the fluid and the fixed.Foucault and Latour are like two pieces of a puzzle waiting to be joined.

4. Temporality: Producing Order and Disorder

So far, the answer to the question ‘What comes after Foucault?’ would appear to be ‘Latour’.Latour enables us to develop a couple of gaps in a Foucaultian programme, in particular enabling us to think seriously about technology (in case we had imagined that Foucault’s ‘technologies of the self’ were enough) and to think seriously about the hybrid character of subjectivity, enabling us to escape from an inappropriate emphasis on the pure human or the pure social.Latour, then, proposes that our analysis should proceed around the ‘network’, a loose alliance of actants oriented around certain problems.For Latour, the analytic task comes to be to understand how various actants impose their readings of the world upon others: how they manage to ‘interest’ certain actants and make sure those actants are not ‘interested’ in anyone else or any other way of understanding the world (Callon and Latour, 1981; Callon and Law, 1982).Using Latour as a supplement to Foucault also allows us to see how social orderings and disorderings emerge in the space between instances and systems, or (in the language we have developed for this essay) between techniques and technologies.We have also continued this theme of finding something in the interstices of theory when we developed the idea that it is in the analysis of the movement from fluidity to fixity and vice versa that we can apprehend the truth of the self.

Latour’s emphasis is on the analysis of a snapshot of the present.While his approach is far from anti-historical, it is at this point that we can usefully stress the importance of Foucault’s persistent methodological device of using historical analysis to understand the present.In the post-Foucaultian literature, the emphasis on history is often lost, yet it is salutary to remember that none of Foucault’s major studies ever went beyond the terrain of the nineteenth century, and few made it even that far.The main reason for this, we think, is that it is virtually impossible to analyse a series of discourses until the passage of time has allowed us to see the shape of a discourse, until a certain temporal distance has allowed us to make sense of a plethora of statements.Foucault’s method is, fundamentally, historical (Kendall and Wickham, 1999).

The emphasis on history fits in with Foucault’s concern with ‘how’ rather than ‘why’ questions.The point is to describe the conditions that allowed a certain way of conceptualising and acting upon the world to emerge. We can characterise the Foucaultian approach to historical research as oriented around two themes, problematisation and specificity.

4.1. Problematisation

Foucault’s approach to history is to select a problem rather than an historical period for investigation.The problem might be ‘how did the prison emerge as the major form of punishment?’ or ‘how did sex come to be seen as so important in terms of who we are?’But, whatever, it is crucial that we allow our investigations of a problem to surprise us. Foucault’s methods involve the generation of surprising stories.The emphasis on problematisation allows us to specify a difference from conventional historiography, which tends to be period-based.It also allows us to understand the idea that Foucaultian histories are ‘histories of the present’. This can be understood as a methodological device to link the present and the past: while the problem to be investigated will be a problem of the present, the key to understanding it lies in the analysis of its historical emergence, and of how certain ways of speaking the truth within the ‘problem-field’ came to be permissible and even compulsory.

4.2. Specificity

Foucault’s historical work is not aimed at the generation of a universal explanation, but rather at the production of an historically-limited account.The easiest way to grasp this difference is to contrast the ‘general history’ with the ‘total history’.

Foucault emphasises the general history as opposed to the total history.The total history looks for overarching principles which govern the development of an epoch; by contrast, the general history gives up the ‘totalising’ theme, concentrating instead on describing differences, transformations, continuities, mutations, and so forth (Foucault, 1972: 9-10).It is worth citing Mitchell Dean to make this distinction clear:

A total history seeks a governing principle of a civilisation, epoch or society, which accounts for its coherence; it seeks to establish an homogeneous network of relations and causality across a clearly defined set of spatial and temporal co-ordinates; it imposes a totalistic form of transformation, and it is able to divide history into definite, cohesive, periods and stages … A general history, on the other hand … seeks series, divisions, differences of temporality and level, forms of continuity and mutation, particular types of transitions and events, possible relations and so on.A general history would be non-reductive, non-totalising, one which specifies its own terrain, the series it constitutes, and the relations between them: … [It] opens up an attention to detail, grain, and complexity, and the specification of form of relation which is indispensable if [we are] to move beyond caricatures of historical periodisation passing for a science of social development. (Dean, 1994: 93-4)

Foucault’s archaeological method is a general history in action.Its goal is to allow us to explore the networks of what is said, and what can be seen in a set of social arrangements: in the conduct of an archaeology, one finds out something about the visible in ‘opening up’ statements and something about the statement in ‘opening up visibilities’.This all sounds fairly obscure, but an example might make this clearer.In his (1977) discussion of the birth of the prison, Discipline and Punish, Foucault’s central task is to show how the prison as a form of visibility (a visible thing) produces statements about criminality, while statements about criminality produce forms of visibility which reinforce prison.We can take from this example the idea in Foucault that statements and visibilities mutually condition each other.

4.3. A not so simple addition: Serres’s times

It is probably not too difficult to guess where we are going next.Imagine the power and flexibility of an analysis that puts Latour’s emphasis on heterogeneous networks together with Foucault’s insistence on the historically conditioned emergence of such networks.Where Latour’s networks sometimes appear to emerge fully formed, Foucault can help us see how they were slowly, painstakingly and accidentally put together; where Foucault stresses the human social world as the dynamic mechanism behind new social arrangements, Latour can help us see how the world stretches far beyond the human or the social.In previous essays we have argued in favour of such a felicitous partnership (Kendall and Michael, 1997, 1998; Kendall and Wickham, 1999).

As we have seen, Foucault’s historical sensibility encompasses a multitude of themes: the relation of historical narrative to the present; the role of contingency, of piecemeal cohesion and change, and of disjuncture in the constitution of ‘system’ or coherence; and the shaping, structuring influence of the long haul.What he alerts us to again is the adhocery of history and the historical moment.But, in the shorter term, that is a feature of Latour’s narratives, linearity does feature – plans are made and implemented and succeed, and can be storied as such. To be sure, as various authors have noted, Latour ignores the long term (e.g. Bowker and Star, 1999; Mort and Michael, 1998) and the very short term (Singleton and Michael, 1993).But within his medium-term time frame, Latour does make the linearity of historical time seem irresistable, even as, following the ethnomethodologists, he is keen to stress that what was (the historical) must be remade each moment: the norm that structures (whether that be inter-subjective, cognitive or embodied in the nonhuman) must always be realised practically in the present and is thus always, in principle, open to revision.In contrast to Foucault’s ‘history of the present’, then, we can posit Latour’s ‘the present of history’ where the historical resonance of macrosocial structures (however regarded – discursive formations, for instance) is always reproduced on a moment by moment basis.

From each of these authors we can distil mutual correctives: Latour has linearity, and the on-going (re)production of order; Foucault has adhocery, and the structuring weight of the historical.But this is too easy.In light of the preceding discussion, we want to bring out how Foucault and Latour can be read through a concern with the movement between order and disorder.Here we draw on Serres’s view of time as flattened.

As is well known, for Serres, different resources distributed in far space and time can be of relevance in the production of order and disorder.The patterning of these dis/orders is what time is – that is, time is an emergent property.As Lash et al put it, for Serres, time is ‘a flattened and immanent scenario “of connections, fluxes, and objective intensities”‘ (1998: 4).Rather than time being stretched out linearly over space, and the present being a point on a line, or a ‘moment’ in the mind, it is emergent in the processes of connection and flux.Thus for Serres the present is singular and is seen as ‘a given assemblage of particularities’ (Lash et al, 1998: 4)).What make up the assemblage are connections wrought by what Serres (e.g. 1982a) once called ‘Hermes’, and now calls ‘angels’.For Serres (1995), angels in their multitudes are a better metaphor for capturing the circulations and connections of multifarious, heterogeneous entities: humans, knowledges, languages, objects, processes. In response to the question, ‘why should we be interested in angels nowadays?’, Serres says:

Because our universe is organised around message-bearing systems, and because, as message-bearers, they (angels) are more numerous, complex and sophisticated than Hermes, who was only one person, and a cheat and a thief to boot. Each angel is a bearer of one or more relationships; today they exist in myriad forms, and everyday we invent billions of new ones.However, we lack a philosophy for such relationships (1995: 293).

Importantly, such connections entail contact and communication between disparate entities.For example, how is it possible that the same motifs appear in science and myth?How are these connected?As Latour (1987) and Harari and Bell (1982) note, Serres’s approach to this ‘parallelism’ does not entail a hierarchy where one text or tradition dominates another – there is no critique or commentary or metalanguage exercised by one over the other (say, science over myth).Rather, as Latour frames it, there is a crossover in which the insights to be derived from myth or fables are no less valuable than those that flow from physics.One is not anachronistic, the other modern. They co-exist; they are co-present.For Serres, then, there is no linear time; rather there are movements between order and disorder.Sometimes these are entropic, sometimes negentropic, sometimes cyclical, sometimes homeostatic.It is out of these movements that, we might say, time emerges.What connects, and renders order or disorder, is, as we have seen, ‘angels’ – examples of which include not only entities such as remote controls and classification systems, but also, importantly, such minsicule material-semiotic entities as electromagnetic signals travelling through the air, wires, fibre-optic cables and so on.These traverse the globe connecting – ordering and disordering the relations between – disparate actors (human and nonhuman).

But there are many angels (message bearers) cross-cutting one another, moving at different rates, bearing bigger or smaller messages, reinforcing and interfering with one another.The orders we see – entropic, negentropic, cyclical, homeostatic – thus co-exist.Rather than attempt to detect the changes or continuities wrought in the long term and the short term, we should be looking to the ways in which such continuities(orderings) and changes (disorderings) map onto one another.We do not simply look at the transitions from one to the other – from orders to disorders – but examine their co-existence and consider at what (micro-, meso-, or macro-scopic) levels these patternings and non-patternings overlap.In sum, the movements of these angels describe the relations between entities – sometimes these movements reinforce relations, sometimes they disrupt them. Sometimes they enable more angels to travel thereby complexifying relations (negentropy); sometimes these angels interfere with one another – are parasites on one another (Serres, 1982b) – leading to the loss of information (entropy).Crucially, these movements and the times they mediate are ‘simultaneous’.

Foucault and Latour, then, clearly operate with different approaches to time, but what they share is a certain linearity.To give each his due, however, we should note that there are interesting hints in both that they wish to go beyond this linearity.For example, Foucault discusses the foldings of discourse, most famously in the metaphor of the ship upon the ocean in Madness and Civilization (Foucault, 1965), and here we can see how he might conceptualise discursive time as wrapped around upon itself like a moebius strip(see Kendall and Michael, 1998).In parallel vein, Latour (1988) problematises the linearity of historiographical times in his (Greimasian) semiotic retelling of the story of Pasteur and the Pasteurization of France.Here, the linearity of the narratives that historically construct Pasteur and his deeds is unravelled (although for a critique see Lynch, 1993).However, it is through a consideration of Serres’s work that we can see most clearly what we might consider the quantum moment in social theory – the realisation that time is not just a backdrop to sociology, that time is an emergent phenomenon, that time is not linear and reversible, that ‘ordered’ and ‘disordered’ time coexist.To be sure, there has been a ‘turn to temporality’ in more mainstream social theory (e.g. Adam, 1990, 1998).However, what Serres offers over and above these still positivistic readings of temporality is a textual practice which engages with different temporalities, where the analytic text traverses, brings together, compares and combines the motifs, narratives and logics of disparate epochs. It is to some of the broader implications of these various issues that we turn in the concluding section.

5. Concluding Remarks

The three themes of technology, identity and temporality cover much that is interesting in the terrain of the social sciences.However, we have to be careful with all three of these themes. To summarise, it is easy enough to place into opposition the components of the dichotomies to which we have drawn attention: technique/technology, identity/self, linearity/juncture, history of the present/the present of history (and many others we have only been able to hint at here).Or rather, it is easy enough to suggest that what is interesting is ‘only’ those moments of transition where an innovation turns into a system and vice versa, where fluidity becomes fixity and vice versa, where an ordering turns into disordering and vice versa.Of course, this is important, but it might also give a false sense that we have somehow ‘grasped’ these ‘real’ moments of transformation.We should note, following Serres, that these moments are never ‘singular’ – they contain within them other dynamics and movements.To summarise, let us draw on a very old metaphor of society (nowadays thought of as heterogeneous) – that of the body (perhaps most famously illustrated by Hobbes’s Leviathan).A body is sometimes ordered, sometimes disordered, sometimes fluid, sometimes fixed, the ‘subject’ of history (subjected) and the ‘subject’ of history (agent). But, over and above, or beneath and below, these some-times, these moments of transitions, is a multiplicity of patterns and nonpatterns.The body is at once in the process of decay (entropy), in the process of gaining experience and learning (negentropy), in the process of self-regulation (homeostasis), in the process of waking and sleeping (cyclicity).As a metaphor, the body still leaves much to be desired.But it does at least help us put on the agenda a view of social, that is, heterogeneous, order and disorder that does not see them opposed: that we might at last see them as concurrent and even mutualistic.

What then are the implications of these various points? Most obviously, we need to sensitise ourselves to the role of technologies in the processes of heterogeneous ordering and disordering: how artefactual nonhumans are moments of innovation and novelty, but also mediators of systematicity.Equally obviously, we need to expand our conceptions of self and identity so that these can accommodate the constitutive part played by technology (in both its guises).Perhaps more interestingly, the foregoing analytic sensibilities need to be complemented by a revised orientation toward temporality.The human self entails an interplay of fixity and fluidity that is partly mediated by parallel interplays in artefactual nonhumans (summed up in our distinction between technology and technique).Moreover, the range of fluidities needs to be mapped – Serres’s differentiation of orderings and disordering is particularly useful here.

The foregoing suggests that humans and technologies are implicated in one another (that is, unsurprisingly, as hybrid).Humans and technologies are conceived as emergent from their relations to one another.But these relations are temporalised – they entail different dynamics, different patternings in which the ‘identity’ (or cogency) of a particular human or nonhuman is variously, and simultaneously, stabilised and de-stabilised.

Of course, despite their emergent and heterogeneous character, we are still dealing with ‘humans’ and ‘technologies’, still working with what are traditional units of analysis.What is more, the process of emergence involves, potentially, an infinity of dynamics, of connections, of relations.The alternative, more radical option would be to derive or fashion – to invent – new units of analysis.Thus, we might construct novel combinations of humans and nonhumans (technological and ‘natural’ – cf. Kendall and Wickham, 2001; Michael, 2000, for examples). These artificial entities could then be explored in terms of their durability, of their transitoriness, of their fluidity and fixity.The result is to trace pathways, modes of relationality, routes of order and disorder that otherwise remain hidden to us.

Acknowledgements

The authors would like to express their thanks to Fiona Campbell, Alan Collins, Susan Condor, Thomas Flynn, Jeff Malpas, Clare O’Farrell, Aneirin Rees, Zlatko Skrbis and Gary Wickham.

References

Adam, B. (1990) Time and Social Theory. Cambridge: Polity Press.

Adam, B. (1998) Timescapes of Modernity.London: Routledge.

Bowker, G. and Star, S.L. (1999) Sorting Things Out: Classification and Its Consequences.Cambridge, Mass.: MIT Press.

Callon, M. and Latour, B. (1981) ‘Unscrewing the Big Leviathan: How Actors Macro-Structure Reality and How Sociologists Help Them To Do So’, in K. Knorr-Cetina and A. Cicourel (eds), Advances in Social Theory and Methodology.London: Routledge and Kegan Paul.

Callon, M. and Law, J. (1982) ‘On Interests and their Transformation: Enrolment and Counter-Enrolment’, Social Studies of Science 12: 615-25.

Dean, M. (1994) Critical and Effective Histories: Foucault’s Methods and Historical Sociology.London: Routledge.

Deleuze, G. and Guattari, F. (1988) A Thousand Plateaus: Capitalism and Schizophrenia.London: Athlone Press.

Elias, N. (1978) The Civilizing Process Vol 1: The History of Manners.Oxford: Blackwell.

Foucault, M. (1965) Madness and Civilization: A History of Madness in the Age of Reason.New York: Random House.

Foucault, M. (1972) The Archaeology of Knowledge. London: Tavistock.

Foucault, M. (1973) The Birth of the Clinic: An Archaeology of Medical Perception.New York: Pantheon.

Foucault, M. (1977) Discipline and Punish: The Birth of the Prison. London: Allen Lane.

Foucault, M. (1984) The Use of Pleasure.New York: Pantheon.

Foucault, M. (1986) The Care of the Self.New York: Pantheon.

Hadot, P. (1995) Philosophy as a Way of Life: Spiritual Exercises From Socrates to Foucault.Oxford: Blackwell.

Harari, J.V. and Bell, D.F. (1982) ‘Introduction’, in M. Serres, Hermes: Literature, Science, Philosophy.Baltimore: Johns Hopkins University Press.

Haraway, D. (1991) Simians, Cyborgs and Women. London: Routledge.

Heidegger, M. (1949) Existence and Being. Chicago: Regnery-Gateway.

Heidegger, M. (1977a) The Question Concerning Technology and Other Essays.New York: Harper Torchbooks.

Heidegger, M. (1977b) ‘Letter on Humanism’, Basic Writings.New York: Harper and Row.

Kant, I. (1952) The Critique of Judgment.Oxford: Clarendon.

Kendall, G. and Michael, M. (1997) ‘Politicizing the Politics of Postmodern Social Psychology’, Theory and Psychology 7 (1): 7-30.

Kendall, G. and Michael, M. (1998) ‘Thinking the Unthought: Towards a Moebius Strip Psychology’, New Ideas in Psychology 16 (3): 141-57.

Kendall, G. and Wickham, G. (1999) Using Foucault’s Methods.London: Sage.

Kendall, G. and Wickham, G. (2001) Understanding Culture: Cultural Studies and Cultural Ordering.London: Sage.

Lash, S., Quick, A. and Roberts, R. (1998) ‘Introduction: Millenniums and Catastrophic Times’, in S. Lash, A. Quick and R. Roberts (eds), Time and Value.Blackwell: Oxford.

Latour, B. (1987) ‘Enlightenment without Critique: a Word on Michel Serres’ Philosophy’, in A. Phillips (ed.), Contemporary French Philosophy.Cambridge: Cambridge University Press.

Latour, B. (1988) The Pasteurization of France. Cambridge, Mass.: Harvard University Press.

Latour, B. (1992) ‘Where are the Missing Masses?’, in W. Bijker and J. Law (eds), Shaping Technology/Building Society.Cambridge, Mass.: MIT Press.

Latour, B. (1993) We Have Never Been Modern. Hemel Hempstead: Harvester Wheatsheaf.

Latour, B. (1999) Pandora’s Hope: Essays on the Reality of Science Studies.Cambridge, Mass.: Harvard University Press.

Lynch, M. (1993) Scientific Practice and Ordinary Action: Ethnomethodology and Social Studies of Science. Cambridge: Cambridge University Press.

Mauss, M. (1973) ‘Techniques of the Body’, Economy and Society 2 (1): 70-88.

Michael, M. (1996) Constructing Identites: The Social, the Nonhuman and Change. London: Sage.

Michael, M. (2000) Reconnecting Culture, Technology and Nature: From Society to Heterogeneity.London: Routledge.

Mol, A. and Law, J. (1994) ‘Regions, Networks and Fluids: Anaemia and Social Topology’, Social Studies of Science 24: 641-671.

Morley, D. (1992) Television, Audiences and Cultural Studies.London: Routledge.

Mort, M. and Michael, M. (1998) ‘Pain and Unemployment: Phantom Intermediaries in the Durability of a Sociotechnical Network’, Social Studies of Science 28: 355-400.

Serres, M. (1982a) Hermes: Literature, Science, Philosophy.Baltimore: Johns Hopkins University Press.

Serres, M. (1982b) The Parasite.Baltimore: Johns Hopkins University Press.

Serres, M. (1995) Angels: A Modern Myth.Paris: Flammarion.

Serres, M. (1998) Unpublished seminar, University of Queensland, Brisbane, Australia, October.

Singleton, V. and Michael, M. (1993) ‘Actor-Networks and Ambivalence: General Practitioners in the Cervical Screening Programme’, Social Studies of Science 23: 227-264.

Star,S.L. (1991) ‘Power, Technologies and the Phenomenology of Conventions: on Being Allergic to Onions’, in J. Law (ed.), A Sociology of Monsters.London: Routledge.

Biographical Notes

Gavin Kendall is Senior Lecturer in Sociology at Queensland University of Technology, Brisbane, Australia.Recent publications include Using Foucault’s Methods (Sage, 1999) and Understanding Culture: Cultural Studies and Cultural, Ordering (Sage, 2001), both co-authored with Gary Wickham.He is one of the editors of the Journal of Sociology and sits on the Executive of the Australian Sociological Association. email: g.kendall@qut.edu.au

Mike Michael is Senior Lecturer in the Sociology Department at Goldsmiths College, University of London.He has published extensively in a number of fields including critical social psychological theory, the public understanding of science, and the animal experimentation controversy.More recently his interests have turned to the role of mundane technologies in social ordering and disordering. He is author of Constructing Identities (Sage, 1996) and Reconnecting Culture, Technology and Nature: From Society to Heterogeneity (Routledge, 2000). email: M.Michael@gold.ac.uk