On actor-network theory

Podcast

---

Central Problem

Latour addresses persistent misunderstandings of Actor-Network Theory (ANT), a theoretical framework developed by the Paris group of science and technology studies since the early 1980s. The theory has been much abused due to three fundamental confusions arising from common usages of the word “network.”

First, ANT should not be confused with technical networks (sewage, telephone, computer networks) which are merely one possible stabilized state of an actor-network. Second, ANT differs fundamentally from social network analysis, which concerns social relations among individual human actors while leaving the social and natural world untouched. Third, the word “actor” has been misread through the Anglo-Saxon tradition as an intentional human individual seeking power through “networking.”

The deeper problem is that existing social theory relies on metaphors of surfaces, spheres, levels, layers, territories, structures, and systems that cannot capture the fibrous, thread-like, wiry, stringy, capillary character of modern societies. Traditional sociology fills space with either order or contingencies, creating artificial divides between micro and macro, local and global, inside and outside, description and explanation, history and theory. ANT proposes a radically different topology where there is literally nothing but networks—no “aether” in which networks should be immersed.

Main Thesis

Latour argues that ANT represents a fundamental ontological shift—a change of metaphors to describe essences: instead of surfaces one gets filaments (or rhizomes in Deleuze‘s parlance). This is a change of topology: instead of two or three dimensions, one thinks in terms of nodes that have as many dimensions as they have connections.

Three Strands of ANT:

1. Semiotic definition of entity building: ANT builds on the “semiotic turn” of the 1960s, using semiotics as a toolbox to study meaning productions. Every entity—self, society, nature—can be understood as selections of embranchments from abstract actants to concrete actors. But ANT extends semiotics to things, dropping the “meaning bit” since semioticians only limited themselves to meaning because they studied texts instead of things.

2. Methodological framework for heterogeneity: ANT provides an “infralanguage” (not metalanguage) for recording the deployment of heterogeneous associations. It places the burden of theory on the recording, not on the specific shape recorded. ANT makes no assumptions about actors’ shapes, allowing infinite pliability and freedom—necessary conditions for observation.

3. Ontological claim on networks: The weakness of semiotics is studying meaning production away from what entities really are. ANT extends meaning productions to all productions, creating continuity between “text,” “social,” and “nature.” These categories are arbitrary cutting points on continuous tracing of action.

Key Moves:

• Counter-Copernican revolution: Instead of universal laws with local contingencies as exceptions, ANT starts from irreducible, incommensurable localities that sometimes achieve provisional commensurable connections.
• Explanation as extension: Networks become more or less explicable as they grow. Explanation is “ex-plicated”—unfolded—indistinguishable from description and deployment of the net.
• Quasi-objects as tokens: What circulates transforms and is transformed by those who do the moving. Both movers and moving objects undergo dramatic changes.

Historical Context

This 1996 paper emerges from over a decade of ANT development since the early 1980s. The “Paris school” of science and technology studies (including Callon, Law, and Latour) had been developing actor-network approaches to understand how scientific facts and technological artifacts are constructed.

The broader context includes the “linguistic turn” or “semiotic turn” of the 1960s-70s, which made meaning production the central object of study. Structuralism and post-structuralism had opened the toolbox ANT would selectively appropriate. Deleuze and Guattari‘s A Thousand Plateaus (1980) provided the rhizome metaphor. Serres’s work on quasi-objects and translations provided crucial ontological resources.

ANT developed in dialogue with ethnomethodology (particularly Garfinkel and Lynch), chaos theory (Prigogine and Stengers), and the Strong Programme in sociology of scientific knowledge. The paper responds to criticisms that ANT is either social constructivist, naturalist, or textual idealist—arguing it simultaneously rejects all “(x)-isations.”

The period also saw debates about reflexivity in science studies, with Ashmore and others arguing that relativism creates insurmountable epistemological problems. Latour here offers ANT’s solution: abandon the dream of epistemological privilege for “relationism.”

Philosophical Lineage

flowchart TD
    Diderot --> NetworkOntology[Network Ontology]
    Serres --> Latour
    Deleuze --> Latour
    Guattari --> Latour
    Greimas --> Latour
    Garfinkel --> Latour
    Foucault --> Latour
    Prigogine --> Latour
    Stengers --> Latour
    Latour --> ANT[Actor-Network Theory]
    Callon --> ANT
    Law --> ANT
    ANT --> STS[Science and Technology Studies]

    class Diderot,NetworkOntology,Serres,Deleuze,Guattari,Greimas,Garfinkel,Foucault,Prigogine,Stengers,Latour,Callon,Law,ANT,STS internal-link;

Key Thinkers

Thinker	Dates	Movement	Main Work	Core Concept
Diderot	1713-1784	Enlightenment	Encyclopédie	Réseau (network) as ontology
Serres	1930-2019	French Philosophy	Statues	Quasi-objects, translations
Deleuze	1925-1995	Post-Structuralism	A Thousand Plateaus	Rhizomes, lines of flight
Greimas	1917-1992	Semiotics	On Meaning	Actants, generative paths
Foucault	1926-1984	Post-Structuralism	Discipline and Punish	Micro-powers
Garfinkel	1917-2011	Ethnomethodology	Studies in Ethnomethodology	Local accomplishment
Callon	1945-	Science Studies	Sociology of Translation	Translation, enrollment
Prigogine	1917-2003	Complexity Theory	Order Out of Chaos	Dissipative structures

Key Concepts

Concept	Definition	Related to
Actor-network	An entity that does tracing and inscribing; not a traced network but network-tracing activity	Latour, ANT
Actant	Semiotic definition: anything granted to be source of action; no special human motivation implied	Greimas, Semiotics
Quasi-object	A circulating token that transforms those who move it because they transform it	Serres, Ontology
Translation	The infralanguage equivalent of Lorentz transformation; moving between frames of reference	Callon, Latour
Irreduction	Not reducing actors to any single (x)-isation (naturalism, socialisation, textualisation)	Latour, Ontology
Infralanguage	Poor, limited, short language enabling travel between nets; opposite of metalanguage	Latour, Methodology
Generalized symmetry	Using words from any realm (natural, social, textual) to describe others	Callon, ANT
Mediators	Entities that transform, translate, distort what they carry; opposite of intermediaries	Latour, ANT
Punctualization	Black-boxing of networks into single points; summarizing complex associations	Callon, Law
Counter-Copernican revolution	Starting from irreducible localities, not universal laws with local exceptions	Latour, Epistemology

Authors Comparison

Theme	Latour	Deleuze
Central metaphor	Filaments, networks, fibrous	Rhizomes, lines of flight
Ontology	Flat, irreductionist	Plane of immanence
Topology	Nodes with connections as dimensions	Smooth vs. striated space
Agency	Distributed among human/nonhuman actants	Assemblages, desiring machines
Method	Follow the actors, describe associations	Schizoanalysis, cartography
Politics	Diplomatic, empirical	Revolutionary, conceptual

Influences & Connections

• Predecessors: Latour ← influenced by ← Diderot, Serres, Deleuze, Guattari, Greimas, Foucault, Garfinkel
• Contemporaries: Latour ↔ collaboration with ↔ Callon, Law, Mol, Stengers
• Followers: Latour → influenced → Science and Technology Studies, New Materialism, Object-Oriented Ontology
• Opposing views: Latour ← criticized by ← Social Constructivists (too realist), Realists (too constructivist), Critical Theory (apolitical)

Summary Formulas

• Diderot: The word “réseau” describes matter and bodies to avoid the Cartesian divide between matter and spirit—network has ontological component from the beginning.
• Serres: Quasi-objects are circulating tokens that transform both what moves and who does the moving; there is no stable mover or stable moved.
• Deleuze: Rhizomes have no beginning or end but always a middle; networks are all boundary without inside and outside.
• Latour: ANT is a change of topology—from surfaces and spheres to nodes with as many dimensions as connections; there is literally nothing but networks, no aether in between.

Timeline

Year	Event
1980	Deleuze and Guattari publish A Thousand Plateaus
1983	Hughes publishes Networks of Power
1986	Callon publishes “Sociology of Translation”
1986	Callon, Law, and Rip publish Mapping the Dynamics of Science and Technology
1987	Serres publishes Statues (quasi-objects)
1987	Latour publishes Science in Action
1988	Latour publishes The Pasteurization of France
1993	Latour publishes We Have Never Been Modern
1996	Latour publishes “On Actor-Network Theory: A Few Clarifications”

Notable Quotes

“ANT is a change of metaphors to describe essences: instead of surfaces one gets filaments… Instead of thinking in terms of surfaces or spheres, one is asked to think in terms of nodes that have as many dimensions as they have connections.” — Latour

“Literally there is nothing but networks, there is nothing in between them, or, to use a metaphor from the history of physics, there is no aether in which networks should be immersed.” — Latour

“Actor-networks do connect, and by connecting with one another provide an explanation of themselves, the only one there is for ANT.” — Latour

---

Warning

This annotation was normalised using a large language model and may contain inaccuracies. These texts serve as preliminary study resources rather than exhaustive references.