Reassembling the Social

Bruno Latour, a French sociologist and philosopher of science, argues that the word "social" has come to mean two incompatible things: a movement of assembling heterogeneous connections, and a specific type of material or domain distinct from biology, economics, or psychology. This confusion, he contends, has rendered the social sciences unable to trace new associations. He proposes Actor-Network-Theory (ANT) as an alternative that redefines the social not as a substance but as a type of connection between things that are not themselves social.

Latour distinguishes two approaches to sociology. The dominant one, the "sociology of the social," treats the social as a ready-made domain used to explain other phenomena, positioning ordinary actors as informants whose full situation is visible only to the trained sociologist. The alternative, the "sociology of associations," denies that any distinct social domain exists and insists that actors are never merely informants. Latour traces how the Latin socius, meaning companion or associate, progressively narrowed from designating all associations to designating only what remains after politics, economics, law, and technology claimed their portions. He identifies Gabriel Tarde (1843–1904), a French social theorist, as a neglected precursor who maintained that the social was a principle of connections rather than a special domain.

The book is organized around three tasks that the sociology of the social has confused: deploying controversies about associations, stabilizing those controversies, and reassembling the social not as a fixed "society" but as an open "collective." Part I addresses the first task through five "sources of uncertainty."

The first uncertainty concerns group formation. Rather than deciding which grouping is real, ANT treats controversies over group formation as primary data. Groups always require spokespersons, define themselves against rival anti-groups, mobilize resources to make boundaries durable, and involve professionals who participate in group-making. Groups exist only as long as they are performed. Latour introduces a key distinction: An intermediary transports meaning without transformation, while a mediator transforms, translates, and modifies what it carries.

The second uncertainty addresses action. Action is never transparent or fully controlled; it is always overtaken by other agencies. But acknowledging this does not license the conclusion that a single substance called "society" is doing the overtaking. Actors must be allowed to articulate their own accounts of what makes them act. The technical term "actant," borrowed from literary semiotics (the study of signs and meaning in narratives), captures the idea that the same action can be figured through corporate bodies, individuals, or structural traits, and none of these figurations is inherently more concrete than any other.

The third uncertainty concerns objects. Latour argues that non-human entities, from hammers and speed bumps to accounting ledgers, must be recognized as participants in action because social ties alone are too weak to produce durable asymmetries of power. Drawing on primatologist Shirley Strum's research on baboons, he shows that when social connections rely on social skills alone, the resulting structure requires constant, exhausting repair. Humans differ because they enroll vastly more non-human mediators that authorize, allow, afford, encourage, suggest, influence, and block courses of action.

The fourth uncertainty addresses scientific facts. Latour recounts how the expression "social construction of scientific facts" was widely misunderstood, with "construction" taken to mean "not true" rather than something fabricated and therefore worthy of the question "how well constructed?" Science studies provided a decisive test because, for the first time, social scientists were studying informants who possessed higher cultural capital and whose objects were harder and more durable than social forces. Latour argues that the failure of social explanation when applied to science reveals it had always failed elsewhere. He proposes replacing "matters of fact" with "matters of concern": entities that are real, fabricated, disputed, and interesting all at once.

The fifth uncertainty concerns the written account itself. A text is not a transparent medium but a laboratory where the social is either successfully traced or fails to appear. A good ANT account treats every actor as a mediator rather than an intermediary. The word "network" designates not a thing shaped like interconnected points but an indicator of textual quality: the capacity to treat each entity as a full participant that transforms what it carries.

An interlude, written as a dialogue between a professor and a PhD student, dramatizes common misunderstandings. ANT is not a framework to be applied but a negative method that tells scholars what not to do. It is incompatible with structuralism because an actor is exactly what is not substitutable. Description is not opposed to explanation; if a description needs an explanation, it is a bad description.

Part II addresses the second task: rendering associations traceable. Latour argues that the 19th-century concept of "society" confused the "body politic," a virtual and fragile entity requiring continuous political action, with a permanent substance that seemed to exist whether anyone acted or not. To escape this confusion, he proposes to "flatten" the social landscape through three moves.

The first move, "Localizing the Global," replaces the notion of overarching context with specific local sites Latour calls "oligoptica," sites that see little but see it well: offices, trading rooms, statistical bureaus, and research centers where structural effects are actually produced. These sites are not larger than the interactions they shape but simply more connected. "Panoramas," such as newspaper editorials and social theory textbooks, stage totalizing views but should be treated as additional local sites rather than as the places from which everything is visible.

The second move, "Redistributing the Local," shows that face-to-face interactions are no more natural a starting point than global structures. Every local site is the assembled endpoint of actions arriving from many other places, times, and materials. Latour introduces "plug-ins," borrowing from web technology, to describe circulating entities, from legal papers to fashion magazines, that equip actors with the competence to interpret their settings. Subjectivity is not an innate property but is generated by continuous attachments from the outside.

The third move, "Connecting Sites," foregrounds the circulating connectors: forms, standards, metrological chains (systems of coordinated measurement that ensure consistency across sites), and collecting statements. Latour acknowledges the extraordinary efficacy of the sociology of the social as a force for formatting the world through calibrated standards, while insisting this formatting is only one part of the associations composing the collective. He introduces "plasma" to name the vast, unformatted background against which all standardized action takes place, a concept that accounts for both the inertia of social structures and the sudden reversals that standard theory cannot explain.

The conclusion argues that sociology must keep three duties distinct: deploying controversies, stabilizing them through formats and standards, and composing the collective through explicit political procedures. Latour contends that critical sociology, by limiting its repertoire to a few social forces and treating actors' objections as confirmation of its explanations, has rendered itself empirically empty. True political relevance lies not in revealing hidden forces but in the progressive composition of a common world out of the multiplicity of entities the sciences have helped detect. Latour closes by suggesting that if the former "West" can no longer compose a shared future using "nature" and "society" as its meta-language, the task of cohabitation demands new sensitivity to the many entities now seeking entry into the collective.