Professional Documents
Culture Documents
iPods, MP3s and file-sharing networks perform a which claims about society, nature, and technology
series of actions that are often reserved for human are transformed into facts and artifacts. Despite the
agents, such as the intellectual and taste-driven rhetoric of dematerialization and virtuality, I argue
that the ANTian concept of artifacts as embodied
labor involved in selecting, sequencing, and redis- actions and knowledges remains useful for studying
covering forgotten sound recordings. At the same the politics of online artifacts. The article then turns
time, the familiar understanding of artifacts as to the controversy in the sociology of science and
stable, material, objective things “out there” is also technology concerning nonhuman agency. ANT’s
being eroded by the infinite replicability, malleabil- notion of nonhuman agency is distinguished by its
ity, and ephemeral flickering of things online. These ontological claim that the social world is constituted
by humans and artifacts and that society should
trends lead to questions regarding the ontologi- therefore be studied symmetrically with regard to
cal status of artifacts and reopen the question of human and nonhuman entities. Finally, the article
how to distinguish technical and material artifacts attends to critiques of nonhuman agency that have
from human and social relations. In this article, the important implications for the application of ANT’s
author explores actor-network theory’s (ANT) con- notion of artifacts as associations of heterogeneous
entities to the study of digital artifacts and technolo-
cept of translation, which advances an alternative
gies.
framework for understanding the role of artifacts in
everyday life. RECONSTRUCTING ANT
AROUND THE CONCEPT OF TRANSLATION
Keywords: Actor-network theory, Nonhuman agency,Politics of What is today referred to as actor-network theory or
design, Digital audio, File-sharing, Copyright law
ANT was initially called the sociology of translation.
As ANT became increasingly popular in sociology
INTRODUCTION: DO ARTIFACTS “ACT”?
and other fields in the late 1980s and 1990s, the no-
In the overlapping designs of compression algo-
tion of translation receded to the background (Latour,
rithms, online music distribution systems, MP3
1999a; Law, 1999). Both concepts were developed
players, and features like random shuffle, there is a
to stress the heterogeneity of the social world, the
new space opening up for popular, journalistic and
distribution of agential properties across the human/
academic discussions of the activeness of artifacts
nonhuman divide, and the processes through which
in social and cultural life. These trends raise impor-
socio-technical collectives extend themselves.1
tant questions concerning the ontological status
However, when divorced from the concept of trans-
of artifacts: How can we distinguish technical and
lation, the actor-network becomes at best a syno-
material artifacts from human and social relations?
nym for other kinds of social, technical, or commu-
How do we locate agency in a world where capaci-
nication networks and at worst a restatement of the
ties to act are distributed across a wide array of ma-
agency-structure debate.2 By relocating translation
terials? In order to recognize the role of artifacts in
closer to the centre of ANT, this section attempts
constituting the social world, do we need a notion of
to recover the ANTian conception of agency as a
nonhuman agency? Using industry statements, news
distribution rather than as a fixed property of cer-
reports, and technical papers on iPods, data com-
tain entities. 3 The linguistic metaphor of translation
pression codecs, and copy protection techniques,
emphasizes the manner in which entities’ interests,
this article demonstrates how actor-network theory
goals, or desires are represented, simplified, and
(ANT) provides a useful framework for engaging
transformed in the production and mobilization of
with these foundational questions regarding the role
artifacts. Reconstructing ANT as a theory of trans-
of artifacts in contemporary life. The article begins
lation puts into view the distributional processes
by reconstructing ANT around the concept of trans-
through which knowledge and action come to be
lation, a step which, I argue, is necessary in order to
embodied by a collective of humans and artifacts.4
redirect analytic attention to the processes through
40
In early ANT literature, translation refers to a To make entities accept the inter-definition of their
process through which one or a few actors become identities in the problematization, the principal actor
spokespersons for a multitude of others by defining uses various strategies and devices of interesse-
and linking their identities in increasingly simplified ment. Seduction, force, and persuasion, for example,
and fixed forms. The concept of translation was may be used to stabilize the entities’ identities, set
quickly adopted and transformed into a broader parameters for their interaction and cut them off
theory of social order and power by a group of from alternative definitions of their identities (Cal-
sociologists of science and technology, most nota- lon, 1986b). These tactics, if successful, lead to the
bly Bruno Latour, John Law, Madeleine Akrich, and enrolment of entities within a program of action. The
Michel Callon.5 These theorists offered the sociology definitions of interests, goals, and identities are then
of translation as an alternative to constructivist and mobilized through representational techniques and
realist explanations of scientific knowledge, which the physical displacement of entities. In general, the
require the analyst to take either “nature” or “soci- process of translation tends towards the association,
ety” as a given (Callon & Latour, 1992). As John Law combination, and simplification of entities and the re-
argues: duction of representatives to one or a few actors. It
transforms weak, provisional, and generally defined
[Sociologists] talk of the social. And then (if identities into durable and seemingly irreversible
they talk of it at all which most do not) they talk ties. If translation is successful, the principle actor
of the technical. And, if it appears, the techni- “speaks for others but in its own language” (Callon,
cal acts either as a kind of explanatory deux 1986a, p. 26).
ex machina (technological determinism). Or
it is treated as an expression of social rela- The boundaries between the various stages of trans-
tions (social reductionism). Or (with difficulty) lation are perhaps more fluid than Callon’s template
the two are treated as two classes of objects implies, but it is useful for describing the transforma-
which interact and mutually shape each other. tion of claims and projects into technological facts
(1991, p. 8) and artifacts in a way that does not take social,
natural, or technological reality as givens. Using the
The theory of translation was developed to over-
framework of translation, attention can be directed
come these problems in sociological approaches
towards the transformative processes through
to science and technology. It was introduced to
which entities are combined and linked with others.
English-speaking audiences with the translation of
As Latour (writing as “Jim Johnson”, 1988) put it: “I
Callon’s (1980) essay on scientific problematization,
use translation to mean displacement, drift, inven-
a process that he studied in the development of a
tion, mediation, the creation of a link that did not
public–private research program on electric vehicles
exist before and that to some degree modifies two
in France.6 Callon (1986a,b) later defined problema-
elements or agents” (p. 32).7 While one could use
tization as the first stage in a series of actions by
this framework to examine the various functions that
which an actor makes itself indispensable to others.
1.“In order to describe such heterogeneous worlds and their network metaphors themselves, often to exclude and define
dynamics in general terms, we introduced the notion of the actor- insides and outsides (Riles, 2000).
network. This concept is important in part because it overcomes 4.In addition to the problems mentioned here, the network meta-
the macro-micro distinction: actor-networks may either grow, phor imposes a particular shape or structure on social relations.
or decline, in size. Indeed, the strategies of scientists, as well as This is completely contrary to ANTian methodology. According
those of other actors, can be characterized as attempts to make to Latour (1999a): “ANT does not tell anyone the shape that is to
relevant actor-networks grow” (Callon, Law & Rip, 1986, p 224). be drawn – circles or cubes or lines – but only how to go about
2.One might say downplaying the role of the actor-network con- systematically recording the world-building abilities of the sites
cept in ANT does more than merely reconstruct ANTand trans- to be documented and registered” (p. 21). Indeed, in ANT litera-
forms it beyond recognition. But the actor-network concept has ture there are countless diagrams and figures that are meant to
been the weak link in ANT’s repertoire of concepts, as suggested demonstrate the variability of the shape and form that actor-
by Collins and Yearley (1992a, 1992b). Compounding the difficul- networks may take (as well as the many possible ways analysts
ties of its apparent resonance with “agency-structure”, Callon, may depict them).
Latour, and Law frequently use “actor”, “agent”, and “actant” 5.Since much of Akrich’s work has not been translated into Eng-
interchangeably, thus making it all too easy to misunderstand lish, I have concentrated primarily on the work of Callon, Latour,
the “network” in “actor-network” as a synonym for “society” or and Law. This unfortunately replicates a blind-spot in Anglo-
“social structure”. Finally, Latour points out that the application American work on ANT, which excludes Akrich from ANT’s
of the actor-network concept to digital communication networks “first wave”. In fact, her descriptive techniques and semiotics
has diminished the sense in which actor-networks are composed of machines had a significant influence on ANT (Latour as
through the transformation of the “nodes” linked together, and Johnson, 1988, p. 305-306). See especially Akrich and Latour’s
not through the instantaneous transmission of information be- (1992) repertoire of terms for describing the association and
tween nodes (Latour & Crawford, 1993). substitution of action between humans and nonhumans through
3.In contrast to translation, the actor-network seems to impose a the use of scripts in material settings.
particular topology of agency or agency effects onto the actors 6.Callon (1991) traces the concept back to work published in 1976,
(Law & Mol, 1994) and risks overlooking the way actors use although it has not yet been translated into English.
7.Latour (1988) wrote under the pseudonym Jim Johnson in this 45), setting the stage where nonhumans receive human proper-
particular article. The actor-network theorists used the term ties, and are enrolled into and mobilized within the collective.
“translation” in a variety of ways. In some cases, it seems to At other times, translation is used to refer to this whole process
refer to a stage in a broader process of coordinating and medi- of collective formation, modification, and movement/action, as
ating action. In other cases, it is used as a kind of shorthand for in the “socio-logic of translation” (Callon, 1980).
the entire process of coordinating action. Latour (1994a) later 8.Latour & Woolgar (1979/1986) similarly found that money itself
used translation to refer to a stage in the extension of social was not the motivation for scientific activities, even though
fabric to nonhumans. Translation is “the means by which we actors consistently understood their work, associations, and
inscribe in a different matter features of our social order” (p. productions as investments.
9.Lawrence Lessig’s work (1999, 2002) provides a cautionary and cultural activities online. In Lessig’s view, the reduced
note against the utopian strand of literature on digitalization. malleability and constrained replicability of online artifacts is a
Lessig has repeatedly argued that creative uses of online consequence of the imposition of economic interests by certain
artifacts such as MP3 files are constrained and monitored in powerful human actors, in particular, “largish” entertainment
an unprecedented way through software patents and copy- conglomerates that seek to maximize their control over the
right law and through the technical codes that regulate social circulation and use of online artifacts (Lessig, 2002, chap. 11).
10.The clearest example of this point is in Latour’s (1988) study the establishment of links between humans and nonhumans.
of Louis Pasteur’s mobilization of France (and much of the rest Sickness, for example, is caused by material overlaps between
of Europe) into a program of action which, at the time, would species. Although bacteria occupy a different place in moral,
have seemed like a rather extreme measure (injecting people political, and ethical orders, they are usefully understood as
with fluids; heating all fermenting drinks, etc.). The success distributions across the human/nonhuman divide. The social
of Pasteurization cannot be reduced to the sheer force of world was redefined in terms of groups formed by that distribu-
his discovery but is a result of modification of identities and tion (the sick, the infected, the contagious, the immune, etc.).
11.In the sociological concept of “actor”, agency is unquestion- 13.“The right documents, the right devices, the right people
ably located in the human universe, not that of things (Pels et properly drilled put together they would create a structured
al., 2002, p. 2; Hetherington & Law, 2000). envelope for one another that ensured their durability and
12.“As a principle of methodological symmetry”, Alex Preda fidelity” (Law, 1986b, p. 154).
(2000) writes, “it just states that the sociologist has to 14.This is particularly evident in Karin Knorr Cetina’s notion of
analyze the human beings and artifacts embedded in such postsocial relations, which has been developed on the basis
a nexus as knots of socially sanctioned (and primarily tacit) of ANT’s arguments about the way science and technology
knowledge, and that these kinds of knowledge are contin- have been crucial sites for the socialization of nonhumans
gent upon each other” (p. 286). Preda notes that a variety of (Knorr Cetina, 1997; Knorr Cetina & Bruegger, 2002). She
approaches including ethnomethodology and the practice argues that the importance of nonhumans in sociality has
theory of Pierre Bourdieu accept this methodological princi- increased over time due to knowledge practices in science,
ple and view things as social actors insofar since they act as economics, and elsewhere, and requires a redefinition of
“knowledge-bearers”. social relations.
16. CRIA. (2005). http://www.cria.ca/freemusicmyth. what ‘holds’ everyone together. This shift from principle to prac-
php#mythsanswers. Accessed February 23, 2005. tice allows us to treat the vague notion of power not as a cause of
17.Latour (1986): “Those who are powerful are not those who ‘hold’ people’s behavior but as the consequence of an intense activity of
power in principle, but those who practically define or redefine enrolling, convincing and enlisting” (p. 273).
18.I refer here to Michel Foucault’s (1986) definition of structural- that makes them appear as juxtaposed, set off against one
ism: “the effort to establish, between elements that could have another, implicated by each other that makes them appear, in
been connected on a temporal axis, an ensemble of relations short, as a sort of configuration” (p. 22).
19.In approaching such controversies, the analyst may at- are necessary: “If sociology is the study of society it has
tempt, as I have done in this article, to trace the actions to take full account of those crowds of nonhumans mingled
attributed to individuals back to the alliances established in with humans. To take full account of this retinue of delegates,
the primary mechanisms of translation. However, pointing sociologists have to look carefully at their conflicts over who
to the heterogeneous materials that perform a particular is the most representative” (1988, p. 16).
definition of socio-technical reality does not resolve the 20. Like de Laet, Jonathan Murdoch (1998) finds the link
problem of representativity. As Latour (1998) points out, “We between immutability and mobility overly simplified when ap-
cannot simply say that ‘all of them’ count in the making of an plied to patent law. He suggests that patents perform a con-
observation. If we were stopping at that, something would nection between the standardized, classificatory schemes of
be missing from the mere deployment of heterogeneous as- law or genetic science and the more fluid and heterogeneous
sociations” (p. 434). Although ANT advocates a “distributed relations of other modes of ordering.
monism” (Barron, 2003), Latour argues that new distinctions