The origins of actor-network theory (ANT) can be traced to science and technology studies (STS) and sociology of scientific knowledge (SSK) in the early 1980s. The approach has from the start been especially associated with the figures of Bruno Latour, Michel Callon, and John Law, who are much inspired by the French philosophers Michel Foucault, Michel Serres, Gilles Deluze, and Felix Guttari. ANT has spread to different areas of the social sciences and has, since the mid and late 1990s, increasingly been felt in some of the subdisciplines of human geography.
From the outset, the central theme of ANT has been the emergence of societal order. Thus many ANTstudies have revolved around questions of how order is accomplished and made stable in time and space. In general, ANT can be described as a methodological device based on a particular worldview which aims at tracing the practices through which society is assembled. ANT can thus be framed as a practice based perspective, but compared to other such approaches it stands out due to its grounding in material relationalism or material semiotics, clearly expressed by the principle of 'general symmetry'. This is a methodological principle that states that researchers should refute all pre given distinctions between classes of possible actors (natural/social, local/global, and economic/cultural) and treat these categories as symmetrical effects of relational practices. Consequently, ANT approaches the world as consisting of heterogeneous relations and practices through which humans and nonhumans alike are treated as possible actors. This basically means that we cannot take anything as given, as everything is an effect of relational practices. Actors are assembled and structures are arranged in a recursive process of networking or translation. Through the filter of ANT, the world is depicted as a mobile arrangement. It thus seeks to abandon much of the conceptualization of the modern episteme that rests on binary terms and dichotomous relations between purified categories such as nature and society. ANT has sought to highlight the frailty of the modernistic worldview and underline how the making of society demands association of diverse elements that never exist as pure categories cut off from wider fabric of relations. Two characteristic terms that cast light on this associational process are immutable mobiles and centers of calculation. The idea of immutable mobile refers to the general term of inscription that describes how an entity may be transformed into a stable object. An immutable mobile is an object that holds its shape even when it is displaced in time space. As such it is a crucial component in the mechanism of the modernistic worldview. Centers of calculation refer to the sites of transformation, that is, the location where immutable mobiles are drawn together, assembled, and calculated.
To explicate the argument and the worldview of ANT, it is useful to run through the meaning of the key concepts of ANT: actor, network, and theory. The principle of general symmetry is instrumental for a quite different understanding of each of the concepts from what is usual within the social sciences.
In ANT, an actor is a relational effect. Hence, agency is a matter of accomplishment or a collective achievement produced through the enactment of networks but not an inherent trait of particular subjects such as humans. This implodes the conventional distinction between subjects and objects. Instead of being cut off from each other, ANT views them as relational. Objects are thereby not passive components of our world, they are 'quasi objects', meaning that they are capable of agency as they affect all interaction. Most importantly, objects and materials are able to stabilize social interaction, rendering it more durable than pure social acts. Humans would only be naked bodies without their props and it is only through interaction – networking – with materials they become (human) actors. Hence, for ANT, the solid individual actor is nonexistent until it has been stabilized as such through relational ordering. This basically means that ANT treats all actors as hybrids and perceives agency as spun between different actants in networks. Thereby, there is no pure element but everything is part of bundles of heterogeneous networks. The point ANT is making is that the division that can be recognized in the world between the active human subjects and the passive material objects is not given in the order of things. It has to be worked upon, practiced, enacted, and re enacted in order not to crumble down.
The conventional view on networks depicts them as a sort of channel between nodes stretched across Euclidian space. The nodes can, for example, be people connected through social networks or places linked through a network of transportation. Within those channels, reciprocal exchange or transport of various sorts can take place, for example, of information, trust, goods, and money. The network concept undergoes changes on two fronts within ANT. First, it is not either social or material but made up of series of heterogeneous actants. This means that networks are always actor-networks. Hence, the social is not some stuff that is transported along system of pipelines but rather a part of and an outcome of the relational practices of networking. Second, it follows that the work inherent in networks, which is carried out by heterogeneous actants is highlighted. Hence, network cannot be thought as distinct from practices as they necessarily emerge through those. One of the implications of this is that one cannot automatically trust networks as lines of transport. They emerge through series of transformative practices; they are translations. Actor-networks may be ordered in ways which make them capable of transporting objects intact over distance but that demands work and effort by diverse actors, to make them immutable objects. Actor-networks are thereby more fluid and in secure but also more material than most other networks in the social sciences.
The concept of theory does also shift its meaning in relation to ANT. The ANT approach should be understood as a theory of what to study rather than an interpretive framework of the world. The concept of translation can be used to cast light on this. Within ANT, translation refers in general to the relational practices through which actors come into being, that is, the work involved in actor-networks. The concept highlights how actors must constantly work in relations for assembling an order to live by. Translation in this sense is a process of establishing communication or making connections between actants. The theory of ANT is nothing more and nothing less than this: in order to gain some insights into the different ways actors use to order their lives, it is necessary to follow their actor-networks, that is, their translations. It is through translation that actors (and networks) can grow by recruiting ever more actants under their power and thus construct an order. However, translation is a precarious process, which demands a lot of work if it is to create what appears as a solid order and it follows that even though an order may seem durable, it should not be taken as permanent or ever reaching a final level stabilization.
These key concepts are instrumental for the study policy of ANT. ANTstresses an empirical investigation – the tracing of relations – and the description of these. Explanations should emerge out of detailed descriptions and are thus local to the network under study but not transcendent of it. As such, ANT seeks to illuminate the emergent patterns of order and/or disorder. In turn, the social cannot be used as explanation but is part of what has to be explained. Early ANT studies, because of this objective, sometimes had a penchant to functionalistic descriptions of relational configuration, where everything seems to turn into networks. In recent years many students of ANT have sought to distance themselves from such inclinations and given up the hope to reach the end point of network order.
At least three critical points to ANT have been raised: First, some would argue that ANT inspired approaches are not the most obvious choice for studies of cultural imaginations, geographical imaginations, nationalistic projects, and the like. While some ANT studies do not discuss the discursive framing of practices much, this is however not the case with all ANT inspired research. It can thus not be taken as an inherent weakness of the general approach. Somehow along the same line, questions of power, domination, oppression, and colonization are not always made explicit. This is in spite of the obvious inspiration from Michael Foucault, the familiarity with much work flowing from feminist studies on these matters, and the fact that some ANT researchers have worked explicitly with power relations. Clearly, however, normative positions are usually not spelled out by ANT proponents themselves. Finally, there is the critique that ANT makes fluids a fetish, appealing to almost liberalist notions of the freedom of movement. But while there is a risk that ANT can be used this way, this does not need to be the case as a variety of studies show. Here it may be noted that two of the main advocates of the approach, Latour and Law, have distanciated themselves from the ANT label after 1999 (the publishing year of Actor-Network Theory and After) although the former has reclaimed it, most notably, in an introductory book to the approach published in 2005.