7

The De-Scription of Technical Objects

Madeleine Akrich

Describing the Interaction between Technics and Humans

Although science and technology are often thought to go together,

they are concerned with very different subject matters. Science is
taken to go beyond the social world to a reality unfettered by human
contingency. Perhaps as a result, the sociology of science has studied
the ways in which the local and the heterogeneous are combined to
create knowledge with the status of universal and timeless truth. By
contrast, sociologists have found it difficult to come to terms with
technical objects. Machines and devices are obviously composite,
heterogeneous, and physically localized. Although they point to an
end, a use for which they have been conceived, they also form part of
a long chain of people, products, tools, machines, money, and so forth.
Even study of the technical content of devices does not produce a
focused picture because there is always a hazy context or background
with fuzzy boundaries. Thus even the most mundane objects appear to
be the product of a set of diverse forces. The strength of the materials
used to build cars is a function of predictions about the stresses they
will have to bear. These are in turn linked to the speed of the car,
which is itself the product of a complex compromise between engine
performance, legislation, law enforcement, and the values ascribed to
different kinds of behavior. As a consequence, insurance experts,
police, and passers-by can use the condition of the bodywork of a car
to judge the extent to which it has been used in ways that conform to
the norms it represents.
Technical objects thus simultaneously embody and measure a set of
relations between heterogeneous elements. However, the process of
describing everything about a car in such terms would be a mammoth
task.1 Furthermore, the end product might well be banal. The
automobile is so much a part of the world in which we live that its
sociography (a description of all the links making it up) would no
 206 Madeleine Akrich The De-Scription of Technical Objects 207

doubt look like a collection of commonplaces. It would, in other to avoid this difficulty.5 The methodological problem is that if we
words, look like a set of places where elements of the technical, the want to describe the elementary mechanisms of adjustment, we have
social, the economic, and so on were to be found together, and it to find circumstances in which the inside and the outside of objects
would leave observers free to switch between one element or register are not well matched. We need to find disagreement, negotiation, and
and another as this suited them.2 the potential for breakdown.
I am arguing, therefore, that technical objects participate in building There are several areas—for instance, in technological innovation
heterogeneous networks that bring together actants of all types and and technology transfer—where objects and their supposed functions,
sizes, whether human or nonhuman.3 But how can we describe the or the relationship between supply and demand, are poorly matched.
specific role they play within these networks? Because the answer has In what follows I describe a number of cases of "technology transfer"
to do with the way in which they build, maintain, and stabilize a to less-developed countries (LDCs) that are drawn from my own
structure of links between diverse actants, we can adopt neither simple fieldwork. These range from the simple transplantation of a piece of
technological determinism nor social constructivism. Thus technical apparatus widely used in industrial societies to the
technological determinism pays no attention to what is brought development of objects specifically intended for use in LDCs.6 In each
together, and ultimately replaced, by the structural effects of a net- case I describe the elementary mechanisms of reciprocal adjustment
work. By contrast social contructivism denies the obduracy of objects between the technical object and its environment.
and assumes that only people can have the status of actors. The I start by considering the way in which technical objects define
problem is not one of deciding whether a technology should be seen as actants and the relationships between actants. I show that the ease
an instrument of progress or a new method for subjugating people. It is with which the actants assumed in the design of the object are related
rather to find a way of studying the conditions and mechanisms under to those that exist in practice is partly a function of decisions made by
which the relations that define both our society and our knowledge of designers. The obduracy or plasticity of objects, something that is
that society are susceptible to partial reconstruction. established in the confrontation with users, is a function of the distri-
To do this we have to move constantly between the technical and bution of competences assumed when an object is conceived and
the social. We also have to move between the inside and the outside of designed.
technical objects. If we do this, two vital questions start to come into In the second part of the chapter I consider the way in which
focus. The first has to do with the extent to which the composition of a technical objects distribute causes. If most of the choices made by
technical object constrains actants in the way they relate both to the designers take the form of decisions about what should be delegated to
object and to one another. The second concerns the character of these whom or what, this means that technical objects contain and produce a
actants and their links, the extent to which they are able to reshape the specific geography of responsibilities, or more generally, of causes. To
object, and the various ways in which the object may be used. Once be sure this geography is open to question and may be resisted.
considered in this way, the boundary between the inside and the Nevertheless, it suggests that new technologies may not only lead to
outside of an object comes to be seen as a consequence of such new arrangements of people and things. They may, in addition,
interaction rather than something that determines it. The boundary is generate and "naturalize" new forms and orders of causality and,
turned into a line of demarcation traced, within a geography of indeed, new forms of knowledge about the world. I will consider this
delegation,4 between what is assumed by the technical object and the process and illustrate the way in which technologies may generate
competences of other actants. both forms of knowledge and moral judgments.
However, the description of these elementary mechanisms of ad-
justment poses two problems, one of method and the other of vocab- Subjects and Objects in the Making
ulary. The difficulty with vocabulary is the need to avoid terms that
assume a distinction between the technical and the social. Because the From Script to De-Scription
links that concern us are necessarily both technical and social, I For some time sociologists of technology have argued that when
develop and use a vocabulary drawn from semiotics that is intended technologists define the characteristics of their objects, they necessar-
ily make hypotheses about the entities that make up the world into
 208 Madeleine Akrich The De-Scription of Technical Objects 209

which the object is to be inserted.7 Designers thus define actors with the user, between the designer's projected user and the real user,
specific tastes, competences, motives, aspirations, political prejudices, between the world inscribed in the object and the world described by its
and the rest, and they assume that morality, technology, science, and displacement. For it is in this incessant variation that we obtain access
economy will evolve in particular ways. A large part of the work of to the crucial relationships: the user's reactions that give body to the
innovators is that of "inscribing" this vision of (or prediction about) designer's project, and the way in which the user's real environment is
the world in the technical content of the new object. I will call the end in part specified by the introduction of a new piece of equipment. The
product of this work a "script" or a "scenario." notion of description proposed here has to be developed within this
The technical realization of the innovator's beliefs about the rela- framework. It is the inventory and analysis of the mechanisms that
tionships between an object and its surrounding actors is thus an allow the relation between a form and a meaning constituted by and
attempt to predetermine the settings that users are asked to imagine for constitutive of the technical object to come into being. These
a particular piece of technology and the pre-scriptions (notices, mechanisms of adjustment (or failure to adjust) between the user, as
contracts, advice, etc.) that accompany it. To be sure, it may be that no imagined by the designer, and the real user become particularly clear
actors will come forward to play the roles envisaged by the designer. when they work by exclusion, whether or not this exclusion is
Or users may define quite different roles of their own. If this happens, deliberate.8 The case of the photoelectric lighting kit is an example in
the objects remain a chimera, for it is in the confrontation between which exclusion was explicitly sought by no one.
technical objects and their users that the latter are rendered real or
unreal. The Photoelectric Lighting Kit: Or How to Produce a Non-
Thus, like a film script, technical objects define a framework of User
action together with the actors and the space in which they are The photoelectirc lighting kit was born from the wish of a government
supposed to act. Sigaut (1984) gives examples of tools whose form agency to promote new energy sources. As part of its cooperative
suggests a precise description (a la Sherlock Holmes) of their users. international activities, the agency wanted to work on and and meet
The two-handled Angolan hoe is made for women carrying children on the need for lighting—something that well-intentioned informants
their backs. The laborer's stake, with its single point, can only be said was essential for all LDCs. At the same time it wanted to help the
driven in by two people, and thus presupposes a collective user. French photoelectric cell industry to create a market.
However, once one moves away from such simple examples, it be- Caught up, as they were, in a specific network involving state
comes more difficult to uncover the links between technical choices, support with industry, those involved in its design conceived of the kit
users' representations, and the actual uses of technologies. Thus the as a function of the specific needs and constraints imposed on them by
method of content analysis, as applied to texts, adopts an individual this network. At no point, for instance, did commercial considerations
and psychological approach that has little or no relevance to our come into play. Accordingly, the shape of the lighting kit can be
problem. Indeed, because it ignores the wide range of uses to which treated as a description of the way in which this network operated—a
objects may be put, it comes close to technological determinism. It is network characterized by the circulation of certain types of resources
obvious that it cannot possibly explain the wide variety of fates and the exclusion of other actors. The "narrative" patterns and scripts
experienced by technological projects—fates that range from complete dreamed up by those who conceived the kits were quite specific, a
success to total failure. function of their position. Study of the lighting kit (or any other
One way of approaching the problem is to follow the negotiations technical object) makes it possible for us to create the "sociology" of
between the innovator and potential users and to study the way in the network defined by its circulation.
which the results of such negotiations are translated into technological When I first heard the industrialists and designers talking about the
form. Indeed, this method has been widely used in sociological and lighting kit, it appeared to be a very simple array with three functional
historical studies of technology. Thus, if we are interested in technical elements. There was a panel for producing electricity, a storage
objects and not in chimerae, we cannot be satisfied methodologically battery, and a lamp that consumed the electricity. However, once I
with the designer's or user's point of view alone. Instead we have to go arrived in Africa and started to study the ways in which such kits were
back and forth continually between the designer and actually used, the picture rapidly became more compli-
210 Madeleine Akrich The De-Scription of Technical Objects 211

cated. Those who were responsible for installing and maintaining kits result of chance or negligence. Each decision actually taken made sense
were confronted with considerable difficulties. The first of these was in terms of design criteria. Direct current is cheaper than alternating
that the wires linking the different components—the panel, the current because a transformer consumes a good part of the available
batteries, and the fluorescent tubes—were fixed in length and could power. Watertight batteries and nonstandard connections were chosen to
not easily be altered because the connections were made with non- prevent people from interfering with and so potentially damaging the kit.
standard plugs. This meant that it was difficult to adapt the kits to fit The length of the wiring had to be limited or it would reduce the
rooms of different sizes. Replacing components with short lifetimes, performance of the equipment. These decisions were intended to ensure
such as lamps or batteries, represented a second set of difficulties. that the lighting kit would "work" under all circumstances—an important
Neither appropriate fluorescent tubes, nor the watertight batteries consideration in the negotiations between the industrialists and their
chosen to ensure that maintenance problems would not limit the life of clients. It should be recalled that it was not the latter who were the
the system, were available in markets outside the capital. Local ultimate users of the kit, but rather the donating agency and the
sources of supply were thus of no help to the user. As a result, despite government to which the gift was to be made. Indeed, such was the
the fact that it was a major element in his or her technical environment, concern to produce a foolproof kit that the designers decided not to have
the user lost control over the installation. Suddenly, what had a separate switch in the circuit because this might become a point of
previously been familiar started to become strange (the first question illicit entry into the system. This meant that users often found it difficult
users asked was often "When do I have to add water to the batteries?"). to turn the light on or off because the only switch available was attached
A third factor also worked to prevent the user from appropriating the directly to the light and so was normally out of reach.
installation. This was the fact that the contractor who installed the kit So it was that the technical object defined the actors with which it
forbade him or her to turn to a local electrician in case of breakdown. was to interact. The lighting kit (and behind it the designers) worked
Instead, the contractor said that he would come to the area twice a year by a process of elimination. It would tolerate only a docile user and
to repair faulty installations. The reason for this embargo on local excluded other actors such as technicians or businesspeople who
repairs was the sensitivity of the photoelectric panel. This, as the might normally have been expected to contribute to the creation of a
instructions put it, "converts solar energy directly into electrical technico-economic network. Had the users really been as docile as the
energy." However, the fact that this took the form of direct current designer intended, I would not have seen that the kit represented a
with non-equivalent poles meant, at least in the view of the contractor, large set of technically delegated prescriptions addressed by the innovator
that it would be risky to call in a local electrician who would have to the user.
experience of alternating but not of direct current. The danger was that If we are to describe technical objects, we need mediators to create
if equipment was connected the wrong way, it might be damaged. the links between technical content and user. In the case of non-
The discovery of these difficulties illustrates an important point of stabilized technologies these may be either the innovator or the user.
method. Before leaving Paris for Africa, the potential significance of The situation is quite different when we are confronted with stabilized
nonstandard plugs, direct current, or waterproof batteries had not technologies that have been "black boxed." Here the innovator is no
occurred to me. It was only in the confrontation between the real user longer present, and study of the ordinary user is not very useful
and the projected user that the importance of such items as the plugs because he or she has already taken on board the prescriptions implied
for the difference between the two came to light.9 The materialization in interaction with the machine. Under such circumstances some
and implementation of this technical object, like others, was a long prescriptions may be found in user's manuals or in contracts.
process in which both technical and social elements were Alternatively, we may study disputes, look at what happens when
simultaneously brought into being—a process that moved far beyond devices go wrong, or follow the device as it moves into countries that
the frontiers of the laboratory or the workshop. are culturally or historically distant from its place of origin. In the next
The fact that the importance of these characteristics only became
section I adopt the last of these methods to describe the use of
evident in the interaction between designers and users was not the
generators in Senegal.
 The De-Scription of Technical Objects 213
212 Madeleine Akrich

panel generates current as a function of climate and latitude. The

De-Scription in Technological Transfer: Reinventing and
"standard" relationship between production and consumption (a
Reshaping Technical Objects in Use reflection of the interdependence of two groups of actors) is replaced
In rural Senegal generators are widely used by "festive groups." An by an individual, direct, and indeed arbitary submission to natural
administration buys some small generators, which it distributes to forces.
youth groups in the villages. With the generators may come lights, a The difference between this and the generator is obvious. In the case of
record player, or a loudspeaker. The youth groups use the generators the generator, the fuel tank can be used to measure the relationship
or lend them to their members who pay for the cost of fuel and oil. between its use and the cost of that use—a relationship embodied in the
Again, they may rent them out to other villagers who are also motor as a whole. The creation of a particular kind of social link, that of
responsible for the cost of fuel and oil. The money that is made by the renting out, is conditioned by the existence of this relationship, which
rental of generators is shared, with part going to the person who delocalizes the generator by creating many groups of actors:
transports the generator and part going to the association. In this way investors/purchasers, owners/users, associate users, renters, and
a small collection of actors is involved with the generator— actors transporters. The existence of transporters makes the property even
that can be seen as so many additions to the components that make up "purer," for they free it from servitude. Their payment marks the
the generator. boundary of group solidarity, for the work of a single person cannot
The generator's metal trailer means that it is mobile, and so it plays enrich the community. At the same time the generator builds a space and
an important part in this process. This is because the field of possible a social geography. Thus the teachers in one of the villages who needed
users and the relations between the different actors is defined by the lighting for their evening classes did not even consider renting a
movement of the generator. However, the fuel tank rivals the generator. The division between the world of the "market" and the
generator for the starring role because it draws a fundamental dis- "civic"10 world may not have been brought into being in the village by
tinction between capital costs and operating costs. This distinction is the social differentiation entailed in electricity and its uses, but it was
inscribed from the outset in the social setup that brings the generator certainly modified by the latter.
to the village: there is the administration, which underwrites the The lighting kit put itself forward as a "hypothetical" object, whereas
investment, and there is the group that actually manages and runs the the generator was just another piece of equipment integrated into the
generator. The technical device reduces negotiations between the two various sectors of economic life. However, we should not overstate the
parties to a minimum because it directly suggests a pre-negotiated difference between them. This is best seen in terms of differential
agreement. Obviously things could be arranged differently. This, resistance. It would would take much more effort to (re) dismantle the
however, would mean delegating a whole series of tasks to additional generator than it would the lighting kit. But in both cases we are
(legal, human, and technical) structures external to the generator and dealing with the creation and extension of networks that simultaneously
its trailer. It might even entail new systems of measurement—in which define both the social and the technical. Thus such items as
case it is not clear whether we would still be dealing with the same nonstandard plugs and fuses become significant when the real users
object. start to displace projected users. Again, the competence of the youth
The situation would be quite different if we were faced with a group, its relations with other elements of village life, the very
device whose costs were concentrated exclusively on the side of in- definition of these elements—all of these are determined at the same
vestment—as, for instance, with the photoelectric kits. What kind of time as, and by the same process, that defines the components that
relationship can there be between the buyer and the user under such make up the generator. If we were to restrict our attention to the
circumstances? This was a question faced by those promoting the "function" fulfilled by this piece of equipment within the youth group,
development of photoelectric cells in French Polynesia. Once these we might imagine that some other technical system (for instance, solar
cells had been distributed, it was not always possible to insist that panels or connection to the national grid) would function in the same
these two classes of costs should be distinguished. Not only did the way. This, however, is not the case, for under such circumstances the
technology itself fail to discriminate between them, but it offered no relationship between the youth group and others in the village would be
method of measurement that could be translated into appropriate different and probably more
socioeconomic terms. Thus no matter how it is used, a photoelectric
214 Madeleine Akrich The De-Scription of Technical Objects 215

fluid. In this sense, then, we can say that our relationships with the To be sure, the creation of a system that allocated land permanently
"real world" are mediated by technical objects. either to individuals or the state was a function of agreement in the
village as a whole about the need for such stability. Through the new
Prescriptions as a Way of Enrolling Actors: Or How to Make property system the electricity company was thus asking the villagers
Citizens to make a pre-inscription witnessing their consent to a certain kind of
So far I have described technologies that appear to exercise relatively future. Thus, individual villagers had to undertake certain formalities
weak constraints over those who use them. If the generator and those to secure title to fixed property. From the standpoint of the electricity
who sponsor it nudge some who would otherwise be outside economic company, legal ownership could be treated as a token for a range of
relations in the direction of involvement, then this effect is relatively agreements between different bodies about the future of the village.
small. In the case of the photoelectric lighting kit, the main danger is The new system of property was also the foundation for a series of
that no one will use it at all. However, technologies are not always like projects by other utilities (the highway department, the water
this. Sometimes their designers and builders use them to obtain access authority, the medical service, the education system). It meant that
to certain actors, whom they push into specific roles. This is what electrification could be integrated into various modernization
happened in the case of the Ivory Coast and its electricity network. programs, and it established economical procedures for consultation
Here the physical extension of the network was an integral part of a and political negotiation. Finally, the construction of the network itself
vast effort to reorganize the country spatially, architecturally, and would put the agreement of the village into practice and stablize it by
legally. The object was to create such new and "modern" entities as making a durable inscription on the landscape.
the individual citizen. But why should the villagers agree to enter into a game in which
Winner (1980) has argued that certain technologies are inherently they would, or so it seems, lose a part of their independence? After all,
political—for instance, nondemocratic. If he is right about this, then by so doing they would place themselves under the influence of a
the approach I have adopted here would lead to a form of technolog- central authority that would, by virtue of this very fact, increase its
ical determinism. However, the case of electrification in the Ivory
power. There are several answers to this question. The villagers
Coast shows that even in those cases where there are marked political
wanted to have access to electricity. But there was the question of the
implications, it is first necessary to interest and persuade the actors to
way in which the company negotiated with the village. Indeed, to put
play the roles proposed for them.
it in this way is misleading. The company did not negotiate directly
Until recently village property in the Ivory Coast was collectively
with the village. Rather, it negotiated with a spokesperson—
owned and under the control of elders, who allocated tracts of land to
invariably someone who had "succeeded" and moved from the village
villagers as a function of their needs. This allocation was not
to the capital. Both this spokesperson, who negotiated with a range of
permanent, and people might move to different areas. When the
central authorities on behalf of the village, and the villagers
authorities started to think about electrification, they decided that this
themselves knew that a series of indirect benefits would follow from
should be contingent on a more stable allocation of land, and in
agreement with the electricity company. After electrification the
particular on a distinction between private and public property. Those
village could hope for better teachers, an improved health service,
developing the new electricity network (who also presented
themselves as spokespersons for the general interest) assumed that the more financial support, and an increase in the number of development
network would both contribute to this division and depend on it, as it projects. In short, electrification was a method for avoiding direct and
would be installed on public land. In other words, the electricity specific negotiations between the villagers and a series of external
network made it possible for the state to create its own space (the agencies. It was a package whose terms were fixed in advance. Those
space of common interests) that could not be appropriated by anyone in the village had a choice. They could accept those terms or they
else. At the same time, it defined those with whom it would interact. could reject them, and overall the package was attractive.
Because only the individual would legally exist in this new system, In general an individual becomes a citizen only when he or she
former collective modes of village representation were thus enters into a relationship with the state. In the Ivory Coast this was
systematically excluded. effected through the intermediary of cables, pylons, transformers,

^■
 216 Madeleine Akrich The De-Scription of Technical Objects 217

and meters. By contrast, in France individuals are inserted into such a wide As I have indicated, the introduction of the electricity network has
range of networks that they have little chance of avoiding citizenship. From established links between individuals in the Ivory Coast. The way in Which
the registry office, via obligatory schooling to military service and the the individual/consumer relates to the network, and via the network to the
welfare state, the mesh of the state with its different superimposed networks electricity company, is codified and quantified by means of a basic technical
draws ever tighter around them. In countries that have been created more tool, the electricity meter. This formulates the initial contract between the
recently, specific networks may come to the aid of a weak or non-existent producer and the consumer. If one or the other fails to meet its obligations,
state. The electricity network may create and maintain a relationship between the meter becomes invalid or inactive. Meters have a symmetrical effect on
an individual and a place. Thus in the Ivory Coast, where only a minority of the producer/ consumer relationship. The agreement of both is required if
salaried workers paid income tax, the electricity bill became the means by they are to tick over. Accordingly, the set of meters is a powerful instrument
which local taxes were collected in recently built towns. Here, then, it was the of control. Taken together, the set of meters measures the cohesion of the
electricity network that fostered a wider definition of the concept of sociotechnical edifice materialized by the network. Consider the following
citizenship. story, which appeared in The Kanian, the electricity company newspaper, in
its February-May 1985 issue:
From Causes to Accusations and Forms of Knowledge
OPERATION STRIKEFORCE AT "ABOBO-THE-WAR"
In the examples above I have shown how technical objects define actors, the There is a flashing red light in the DR in Abobo, a lower class suburb of
space in which they move, and ways in which they interact. Competences in Abidjan, where there are 66,854 subscribers; the network's rate of return
(the relationship between the energy put out by the producer and the
the broadest sense of the term are distributed in the script of the technical
energy billed to the clientele) has fallen from 0.93 to 0.87 in the space of one
object. Thus many of the choices made by designers can been seen as year!
decisions about what should be delegated to a machine and what should be
left to the initiative of human actors. In this way the designer expresses the Any reduction in the rate of return can be interpreted as an increase in the
scenario of the device in question—the script out of which the future history number of illicit connections, the work of corrupt employees, or a
of the object will develop. But the designer not only fixes the distribution of consequence of trafficking in meters. With both human and technical actors
actors, he or she also provides a "key" that can be used to interpret all involved, the network measures illicit behavior and determines its character.
subsequent events. Obviously, this key can be called into question—consumer The definition of social space also extends to non-electrified areas. These
organizations specialize in such skepticism. Nevertheless, although users add are characterized in terms of their degree of deviance from the norm—that is,
their own interpretations, so long as the circumstances in which the device is from electrification. Thus another suburb of Abidjan, Marcory, was split into
used do not diverge too radically from those predicted by the designer, it is
two by the network. Each was given a name, and characterized in social
likely that the script will become a major element for interpreting interaction
terms:
between the object and its users.
Unlike residential Marcory, Marcory-No-Wire is a Marcory without elec-
Abobo-the-War and Marcory-No-Wire: Where Technology Meets Morality tricity, without wires. It is well known that Abidjanis have a sense of
In this section I focus on one particular process—moral delegation—and humour. A suburb with no wires, imagine what kind of a spectacle that
discuss devices installed by designers to control the moral behavior of their offers. For if electricity is a sign of progress, its absence suggests other
users. I describe the way in which such devices may measure behavior, place absences: of hygiene in the streets, of buildings constructed to certain stan-
dards, of pharmacists, playgrounds, sportsgrounds and so on. When you add
it in a hierarchy, control it, express the fact of submission, and distribute darkness at night to these absences, then the guardians of the peace would
causal stories and sanctions. say you get a criminal haunt. (Toure 1985)

Even so, the dividing line between the permissible and the impermissible
is negotiable. Thus in their strike-force operations, elec-
218 Madeleine Akrich The De-Scription of Technical Objects 219

tricity company agents were told to replace so-called Russian meters that had proper rules of conduct. However, a flaw in the system is that there is no easy
proved defective without penalizing their owners, even though a simple tap way to measure the charge in the battery. Voltage is only a rough indication.
on the meter would block it and allow unbilled electricity to be consumed. What should be done about this? A general who is not sure of the loyalty of
Unlike the agents, the "Russian" meters found it impossible to distinguish his troops has two options. He may choose to do nothing. Or, like the
between licit and illicit behavior, between the actions of humans and designers in this case, he may redouble his precautions and disciplinary
nonhumans. Accordingly, although the contract between supplier and measures. Accordingly, as I have mentioned, a particularly inflexible system
consumer remained in force, the meter failed in its prescribed role as the with nonstandard plugs was adopted. Thus while the control device was
material inscription of that contract. telling the user not to get too big for his or her boots, the nonstandard plugs
Each individual meter intervened as referee and manager of the were imposing even more draconian limitations on conduct. No bypass of the
relationship between supplier and consumer. Taken together, the set of meters control device was permissible!
operated as police in a collective organization, uncovering irregularities. Such Even so, in French Polynesia the control device proved to be a shaky ally
irregularities appeared first as deviations in consumption curves that were for the designers, because the users felt that its sanctions were arbitrary. The
neither localized nor sanctioned. They could, however, be quickly translated result was that they denounced it and expressed their displeasure by
into "social" terms. telephoning the electrician every time the system treacherously cut off the
Some techniques move closer to "social control." They establish norms and current while they were quietly sitting watching television. The electrician,
punish those who transgress them. Thus the storage and regulation systems in who quickly became tired of doing repairs in the evening, tricked the system
photoelectric kits take the form of batteries and electronic components. The by installing a fused circuit in parallel with the control device. When the
batteries store the electricity so that it can be given out, for example, for control device shut off the current, users could bypass it with the fuse, and the
lighting when it is dark. However, the control system lies at the heart of a electrician would only be called out the following morning. The fused circuit
technical, economic, and social imbroglio. If the battery is allowed to run too thus marked the submission of electricians to the wishes of their clients and
low, its lifetime will be reduced. On the other hand, if it is overcharged, allowed them to be present by proxy instead of being summoned in person by
electricity may leak back into it and ruin the photoelectric cell. Users might, irate users.
of course, be given meters with which they could plan their electricity The precarious and makeshift character of the fuse makes it plain that
consumption while avoiding both of these dangers. In fact this solution is
some kind of intervention was necessary, even if it only took place after the
never adopted because the designers do not believe that users will allow the
event. In this particular trial it was the electricians who pleaded guilty. In
technical requirements of the system to overrule their immediate wishes.
fitting the fuse, they recognized that the control device and their clients were
Again, the designers could choose to increase the capacity of the system to
both right and moderated the judgments of the former in favor of the latter.
cope with the likely demands of the users. This, however, is a costly option.
Accordingly, the designers adopt the third option of installing a regulator that
"The Order of Things and Human Nature": The
cuts off the current to the user when the charge on the battery gets too low,
Stabilization and Naturalization of Scripts
and isolates the photoelectric panel when it gets too high.11 As a result, a
I have described several cases in which technical objects preformed their
particular mode of consumption is imposed: the user cannot be too greedy, yet
relationships with actors and vested them with what could be called "moral"
neither can he or she hope to compensate for excess consumption by
content. Because roles and responsibilities are allocated, accusations and
prolonged abstinence. The penalty for breaking the rules—rules that are both
trials tend to follow. In principle, no one and nothing is protected from such
social and technical—is immediate and abrupt: the current is cut off and is not
denunciation. In the case of the electricity network, the users were accused of
reconnected until the battery is adequately recharged.
failing to respect the contract with the meter. However, the electricity
This method of regulation is designed to "groom" the user. It offers a set of
company also accused some of the meters of failing to represent that contract.
rewards and punishments that is intended to teach In the case of the photoelectric kits, it was the electricians, and
220 Madeleine Akrich The De-Scription of Technical Objects 221

indirectly the manufacturers, who found themselves in the dock cation of causal pathways. In the case discussed above, the replace-
through the agency of the control device. Indeed, the story of the kits ment of the "Russian" meters was very much part of this process—a
can be read as a long series of reciprocal accusations. The industrial- process designed to make diagnosis automatic. Farther along the same
ists tended to argue that if it didn't work (technically), this was path lies artificial intelligence.12
because it had been misused (socially). The users, or those who
claimed to be their representatives, argued that if it didn't work Conclusion: Toward the Constitution of Knowledge
socially, this was because it had been misconceived technically. Here,
then, we see an almost perfect "reversible reaction" that reveals the Once technical objects are stablized, they become instruments of
lack of a relationship, through the kit, between designers and users. knowledge.13 Thus when an electricity company sets differential
The users did not interest the manufacturers; they were only important tariffs for high- and low-consuming domestic users, for workshops,
to the extent that they made it possible to go to the ministry of and for industrial consumers, it finds ways of characterizing and
overseas development and seek support for a product that did not yet identifying different social strata. If it also chooses categories used in
have a market. And in this interaction the kit did not actually have to other socioeconomico-political network, then the knowledge it
do anything. Rather it was the users who were treated as an instrument produces can be "exported." "Data" can thus be drawn from the
for building a relationship between the manufacturers and the network and transmitted elsewhere, for instance, to economists con-
government. cerned with the relationship between the cost of energy or GNP and
In the case of the electricity network, the situation was quite consumption. However, the conversion of sociotechnical facts into
different. It is difficult to imagine a plausible argument for illegal facts pure and simple depends on the ability to turn technical objects
connection to the network—one in which the electricity network into black boxes. In other words, as they become indispensable,
would stand in the dock. This is because the network configured a objects also have to efface themselves. I will illustrate this with an
whole range of relationships. I have already mentioned the meter and example drawn from Burkino-Faso.
the way in which it was related to the allocation of property. But Burkino-Faso is a developing country with a tiny electricity net-
relationships were structured by the network in many other ways. For work. Over the past few years it has been government policy to
example, it also tended to stabilize living space. This was because, for electrify urban centers. The first problem for the engineers and
reasons of security and as a guarantee of solvency, only "permanent" technicians was to judge potential demand and decide how large the
structures were connected to the grid. And of course, once the grid network should be. Two different approaches were adopted. The
was in place, new commercial networks for distributing electrical economic studies unit asked potential subscribers what price they
equipment quickly sprang up. Thus once it was established, the would be willing to pay for electricity. This approach assumed that
network tended to promote both physical and social stability. A wide there was a relationship between supply and demand, and that
range of elements were brought together and given substance. A small consumption would vary inversely with price. The technical unit
fringe group of "deviants" could not possibly hope to find the strength adopted a very different method. It drew maps of the towns, marked
needed to outweigh the many actors bound together by the grid. off the built-up areas, and noted the characteristics of the houses
Accordingly, the electricity company could call upon the meters to act (whether large or small, permanent or temporary, and so on). On the
as unequivocal spokespeople at will. A double irreversibility had been basis of this map they designed a network that would be legally,
established—a material irreversibility inscribed in space and practice, economically, and technically feasible—a network that would make
and a directional irreversibility where accusations and charges could use of public space and serve only permanent buildings and govern-
no longer be reversed. Obviously the two were intimately linked. ment facilities.
In this section I have argued that technical objects not only define The results obtained by the two approaches were quite different. In
actors and the relationships between them, but to continue functioning particular, the geographical and legal approach of the technical unit
must stabilize and channel these. They must establish systems of suggested the need for a far larger network than the market-led
causality that draw on mechanisms for the abstraction and simplifi- approach of the economic studies unit. The latter had acted as if
222 Madeleine Akrich The De-Scription of Technical Objects 223

there were no need for technical mediation between price and consumption. more conceptualized form of the conclusions I drew from the various field studies
They assumed, that is, that this relationship was a fact of nature that would be discussed here.
given concrete form by the electricity network. In a sense they were led astray 1. Doubtless it could be satisfying to paint on a broad canvas, starting with nuts and
by the naturalization effect, which occurs when technical systems are bolts, pistons and cracks, cogs and fan belts, and moving on to voting systems, the
strategies of large industrial groups, the definition of the family, and the physics of
completely integrated into the social fabric. It is only when the script set out by
solids. In the case of such an inquiry we would no doubt find a mass of guides
the designer is acted out—whether in conformity with the intentions of the (people, texts, objects) ready to suggest ways in which we could extend our network.
designer or not—that an integrated network of technical objects and (human and But such suggestions would be endless. On what grounds would the analyst stop—
nonhuman) actors is stabilized. And it is only at this point that this network can apart from the arbitrary one of lassitude? Quite apart from the indefinite amount
be characterized by the circulation of a finite number of elements—objects, of time such a study would take, there is also the question as to whether it would be
physical components, or monetary tokens. Disciplines such as economics and interesting.
technology studies depend on the presence of a self-effacing apparatus that lies 10. Here we are concerned with what might be called the consensual zone of the
automobile, which is defined simultaneously by the major technical elements com-
outside their domains. Economists extract one kind of information from
mon to most vehicles and by their generally recognized uses. As is obvious, there are
technical objects, technologists another. They are able to do this because such highly controversial zones around the margins, and it is around these points of
objects function in stable situations. The introduction of a new device can thus friction that the battles leading to the establishment of supremacy of such and such a
be assimilated, for example by economists, into the price/consumption manufacturer or such and such a car are waged.
relationship. The economy is not cut off from technology; there is no radical 11. This term is used only as a convenient but imprecise shorthand. Depending on
disjunction. circumstances, the actor (a more general term to be prefered) may be a citizen, a
This is why it makes sense to say that technical objects have political member of a particular social class, a member of a profession, or even a finger or a
strength. They may change social relations, but they also stabilize, naturalize, body with a particular temperature as measured by a system of detection.
depoliticize, and translate these into other media. After the event, the processes 12. See Bruno Latour's text (this volume) for further discussion of delegation.
involved in building up technical objects are concealed. The causal links they 13. This vocabulary is further discussed in Latour's text in this volume and in the joint
established are naturalized. There was, or so it seems, never any possibility that appendix to our papers.
it could have been otherwise.14 6.1 am aware that the reader may be frustrated by the way in which these examples are
used. Within a short article it is not possible to give full details. But as they are intended
We are ourselves no more innocent in this respect than anyone else. For we to exemplify an argument, I hope that the reader will agree that the benefit of using them
are able to say that technical objects changed, stabilized, naturalized, or in this way outweighs the costs.
depoliticized social relations only with the benefit of hindsight. The burden of 7. For a striking example of the interrelationship between the definition of technical
this essay is that technical objects and people are brought into being in a process parameters and the definition of a "world" for which the object is destined, see
of reciprocal definition in which objects are defined by subjects and subjects by Callon's article on the electric vehicle in Bijker, Hughes, and Pinch 1987.
objects. It is only after the event that causes are stabilized. And it is only after 8. See, for example, Winner 1980 and Latour 1988a. Winner describes how the
the event that we are able to say that objects do this, while human beings do height of overpasses on the Long Island Parkway was chosen to prevent the passage
of buses, the mode of transport most used by blacks, so that the use of leisure zones
that. It is in this sense, and only in this sense, that technical objects build our
was effectively limited to whites. Latour, reinterpreting the example described by
history for us and "impose" certain frameworks. And it is for this reason that an Daumas (1977), tells how, in exactly the same way, the radical Paris city council
anthropology of technology is both possible and necessary. at the end of the nineteenth century decided to build metro tunnels too narrow for
standard railway company trains. The objective, which succeeded for seventy years,
Notes was to prevent the private railway companies (supported by the right) from getting
I would like to thank Geoffrey Bowker, who translated this text, John Law, who their hands on the Paris metro, whatever party happened to be in power. Multiple
carefully reviewed the entire text, and Bruno Latour, who helped me arrive at the translations are necessary in order to arrive at such results. In Winner's case we need
to move from the white/black to the car/bus distinction, and then on to the height
of the overpasses. This is only possible because the black/white distinction is already
pre-inscribed in unequal access to economic resources and, as a consequence, to
expensive products such as cars. In Latour's case it is the width of the tunnels that
allows the railway (and so the different companies and political parties) to be
kept at arm's length from the metro.
224 Madeleine Akrich
8
9. In the French there is a play of words on dessein (design in the sense of plan) and
dessin (design in the sense of drawing). The two have the same etymology. Where Are the Missing Masses? The
14. I am drawing here on the distinction between "marchand" and "civique" discussed
by Boltanski and Thevenot (1987).
Sociology of a Few Mundane Artifacts
15. Naturally, the different parts of the system are reconnected automatically once Bruno Latour
conditions change.
16. The question of "breakdown" is relevant to this issue and deserves further
consideration. A "breakdown" relates closely to the definition I have offered of a
technical object. This is because it can only be understood as a part of practice—
that is, as the collapse of the relationship between a piece of apparatus and its use.
A breakdown is thus a test of the solidity of the sociotechnical network To Robert Fox
materialized by a technical object. The rapidity with which the search for the
causes of breakdown can be completed is a measure of this solidity. Again, might not the glory of the machines consist in their being without this same
boasted gift of language? "Silence," it has been said by one writer, "is a virtue which
17. Perhaps it would be better to say that the stablization of a technical object is
render us agreeable to our fellow-creatures."
inseparable from the constitution of a form of knowledge of greater or lesser
significance. This hypothesis is powerfully supported by the case described by Samuel Butler {Erewhon, chap. 23)
Misa (this volume): there an industry, a market, and the notion about what was to
count as "steel" were all constructed simultaneously.
18. As is well known, Foucault (1975) has described the links between the technology Early this morning, I was in a bad mood and decided to break a law
of the penitentiary, power relations, and new forms of knowledge.
and start my car without buckling my seat belt. My car usually does not
want to start before I buckle the belt. It first flashes a red light
"FASTEN YOUR SEAT BELT!", then an alarm sounds; it is so high
pitched, so relentless, so repetitive, that I cannot stand it. After ten
seconds I swear and put on the belt. This time, I stood the alarm for
twenty seconds and then gave in. My mood had worsened quite a bit,
but I was at peace with the law—at least with that law. I wished to
break it, but I could not. Where is the morality? In me, a human driver,
dominated by the mindless power of an artifact? Or in the artifact
forcing me, a mindless human, to obey the law that I freely accepted
when I get my driver's license? Of course, I could have put on my seat
belt before the light flashed and the alarm sounded, incorporating in
my own self the good behavior that everyone—the car, the law, the
police—expected of me. Or else, some devious engineer could have
linked the engine ignition to an electric sensor in the seat belt, so that I
could not even have started the car before having put it on. Where
would the morality be in those two extreme cases? In the electric
currents flowing in the machine between the switch and the sensor? Or
in the electric currents flowing down my spine in the automatism of my
routinized behavior? In both cases the result would be the same from
an outside observer— say a watchful policeman: this assembly of a
driver and a car obeys

I^W11