Structuralism and the Turn to Language

Every school so far in Part 5 has argued about how societies work. Structuralism argues about something prior: how meaning is possible at all — and its answer turned out to have consequences for everything else.

Its central claim is counterintuitive and, once grasped, difficult to shake: things do not have meanings. Systems of difference do.

Try to say what "hot" means.

Not what it feels like. What the word means.

You will find yourself saying: not cold. Perhaps: warmer than warm.

Now try without reference to any other temperature word. You cannot. "Hot" has no content of its own. Its meaning is entirely a matter of its position in a set: freezing, cold, cool, tepid, warm, hot, scalding.

And notice that the set is not given by nature. Water at 40°C is hot in a bath, cold in a cup of tea, and warm for a swimming pool. The physical fact is one; the term applied depends on the system in play.

Now a sharper case. English has one word where Welsh has two: glas covers part of what English calls blue and part of what it calls green. Russian has two words for what English calls blue — goluboy (light) and siniy (dark) — treated as distinct colours, not shades of one.

The spectrum is continuous. There are no lines in it. The lines are in the language, and different languages cut it differently.

Ferdinand de Saussure's conclusion, delivered in lectures at Geneva between 1907 and 1911:

"In language there are only differences, without positive terms."

That sentence is the foundation of structuralism, and of most of what French thought did for the following sixty years.

The problem Saussure was solving was linguistic and became general.

The commonsense view of language is that words are names for things. A word is a label, meaning is reference, and language is a bag of labels for a world of pre-existing objects.

This does not work, for reasons that pile up quickly.

Many words name nothing: "however," "if," "justice," "unicorn." Different languages carve the world differently, and not into the same objects: kinship terms, colour terms, terms for snow, ice, rice, rain and social relations vary systematically. And meaning changes without the things changing. "Nice" once meant foolish; "silly" once meant blessed.

Saussure's alternative: language is not a collection of labels but a system, and a word's value comes from its relations to the other terms in the system — principally from what it is not.

And the generalisation that made this a theory of society: if meaning in language works this way, perhaps meaning in anything works this way. Kinship, myth, food, fashion, ritual, literature, advertising. Anything that signifies might be a system of differences, analysable by the same method.

That generalisation is structuralism, and it swept through anthropology, literary criticism, psychoanalysis, Marxism, history and sociology between 1955 and 1970.

Saussure's apparatus

The sign has two inseparable parts:

The signifier (signifiant) — the sound-image or written mark: the noise "tree," the letters t-r-e-e. The signified (signifié) — the concept: the mental idea of a tree.

Note what is not included. The actual tree in the garden — the referent — is outside the sign entirely. Saussure is describing a relation between a mark and a concept, not between a word and a thing.

And the relation between them is arbitrary. There is nothing tree-like about "tree," arbre, Baum or shajar. The connection is conventional, which is why languages differ and why the same signifier can attach to different signifieds over time.

Langue and parole. Langue — the underlying system: the rules, the vocabulary, the structure, held collectively and possessed by no individual completely. Parole — actual utterances: what people say, on particular occasions.

Saussure's decision: linguistics should study langue. This is the founding methodological move of structuralism — attend to the underlying system, not to its instances.

Synchronic and diachronic. Synchronic — the system as it stands at one moment. Diachronic — its change over time.

Saussure privileged the synchronic, arguing that a system's workings can be described without its history — as one can describe chess without knowing how the rules developed. This is the move that later attracted the most criticism, and it is the source of the standing charge that structuralism cannot handle change.

And two axes on which every element sits: Syntagmatic — relations of combination: what can follow what in a sequence. "The cat sat" — each word constrains the next. Paradigmatic — relations of substitution: what could stand in this slot instead. Cat / dog / man / idea.

Meaning is produced by both: by what is selected from the possible substitutes, and by how it is combined.

Lévi-Strauss: structural anthropology

Claude Lévi-Strauss (1908–2009) took Saussure's method out of linguistics and made structuralism a general programme.

His central claim: cultural phenomena — kinship, myth, food, ritual, art — are systems of signs, produced by unconscious structuring operations of the human mind, and the operations are universal even though their contents vary.

On kinship (The Elementary Structures of Kinship, 1949): kinship systems are not primarily about biology or descent but about exchange. The incest taboo is the founding rule not because of any biological rationale but because it compels exchange: by prohibiting marriage within the group, it forces alliance between groups. Women are exchanged, and the exchange creates society.

(The feminist critique of this — that it makes women objects of exchange between male-headed groups and treats that as the origin of social order — is central and is developed by Gayle Rubin in "The Traffic in Women," 1975. It is one of the most important critiques of structuralism and it is developed further in 8.5.)

On myth (Mythologiques, four volumes, 1964–71): myths across the world, in societies with no contact, exhibit the same underlying structures. Their function is to mediate contradictions a society cannot resolve in practice — life and death, nature and culture, self and other.

His method: break the myth into its smallest units (mythemes), identify the binary oppositions they express, and show how the narrative works to mediate between them. The individual myth is like parole; the structure is langue.

The single most productive — and most criticised — structuralist tool.

Lévi-Strauss's claim: the human mind operates by making binary distinctions, and cultural systems are built on paired oppositions.

Nature / culture. Raw / cooked. Male / female. Sacred / profane. Left / right. Inside / outside. Pure / polluted. Us / them.

And the analytical payoff is that the pairs are not symmetrical. In each, one term is unmarked and privileged and the other is defined negatively against it. Culture over nature. Male over female. Inside over outside.

Where this proved genuinely powerful:

Mary Douglas's Purity and Danger (1966) (4.2.5) is the finest application. Dirt is "matter out of place" — pollution beliefs are reactions to things that violate a classification system. Her analysis of the Levitical dietary laws argues that the prohibited creatures are those that fall between the scheme's categories: creatures of the water without fins and scales, creatures of the land that do not chew the cud and part the hoof. Anomaly, not harm, is what makes something unclean.

Edmund Leach on animal categories and verbal abuse: which animals a society eats, which it keeps as pets, which are taboo, and which supply its insults, all map onto how close they sit to the social self.

Roland Barthes's Mythologies (1957) applied it to French popular culture — wrestling, wine, steak and chips, soap powder, the Citroën DS — and to advertising and photography, showing how myth naturalises: it takes a historical, contingent arrangement and makes it appear as nature. Barthes's key move: myth is a second-order semiological system that takes an existing sign and makes it the signifier of a further, ideological signified. This is ideology critique (4.1.7) conducted with linguistic tools, and it founded modern media semiotics.

And where it fails: the objection that binaries are the analyst's imposition rather than the culture's structure, that any material can be forced into pairs, and that the method is unfalsifiable — a myth that fits confirms the theory, and a myth that does not is a transformation of one that does.

Althusser: structuralist Marxism

Louis Althusser (1918–1990) applied structuralism to Marx, and the result was one of the most influential and most contested readings of the twentieth century.

The epistemological break (4.1.4): a rupture around 1845 separating the young humanist Marx from the mature scientific Marx. Most scholars now reject this, and it was decisive at the time.

Anti-humanism. Althusser's most provocative claim: Marxism is a theoretical anti-humanism. "Man" is not the origin of history but an effect of structures. History is "a process without a subject."

Overdetermination. Borrowed from Freud: any social contradiction is shaped by multiple determinations, none reducible to the others. The economy is determinant "in the last instance," but — his famous line — "the lonely hour of the last instance never comes."

Relative autonomy. The superstructure has genuine independence and its own effectivity (4.1.2).

Ideological state apparatuses. Church, school, family, media, law, unions — distinguished from the repressive state apparatus (army, police, courts, prisons) by operating primarily through ideology rather than force. The school, he argues, is the dominant ISA in modern societies.

And ideology is material. It exists in practices, rituals and institutions, not in ideas: "ideology has a material existence." His definition — "ideology represents the imaginary relationship of individuals to their real conditions of existence" — and his claim that there is no outside to ideology, even under communism.

Interpellation. The mechanism: ideology hails individuals, and by responding to the hail they become subjects. His illustration is a policeman shouting "Hey, you there!" — the person turns, and in turning becomes the subject the hail addressed. The subject is not prior to ideology; it is produced by it.

Althusser's fortunes, stated honestly.

The influence was enormous in the 1960s and 70s, particularly in Britain, where it shaped film theory, cultural studies, education sociology and the New Left Review.

And the reaction was equally strong. E. P. Thompson's The Poverty of Theory (1978) is a ferocious attack — a hundred and fifty pages of it — arguing that structuralist Marxism is an "ahistorical theoreticism" that eliminates human agency, cannot account for experience or struggle, and is politically quietist. Thompson's own work (4.1.8) is the counter-case: class as something people make.

And the biography cannot be omitted. In 1980 Althusser strangled his wife, the sociologist Hélène Rytmann. He was found unfit to plead by reason of insanity, was committed, and was never tried. He lived until 1990 and wrote a posthumously published memoir about it.

This course states such things. It does not settle whether his theory is right, and it is part of the record.

What survives: interpellation is genuinely useful and widely used; ideological state apparatuses is a serviceable framework; relative autonomy and overdetermination are standard corrections to base-superstructure crudity. What does not: the epistemological break, the anti-humanism, and the general claim that structures act without subjects.

Go deeper

The structuralist moment, and why it ended

The moment was roughly 1955–1968, centred in Paris, and it was a genuine intellectual revolution across disciplines: Lévi-Strauss in anthropology, Barthes in criticism, Jacques Lacan in psychoanalysis (the unconscious "structured like a language"), Althusser in Marxism, and early Foucault in the history of thought.

Its appeal: rigour, method, a way of doing "science" in the human sciences, and a decisive break from existentialism's focus on individual consciousness and choice — Sartre was the figure being displaced.

What ended it, in three parts.

May 1968. A theory in which subjects are effects of structures and history is a process without a subject was awkwardly placed to explain a spontaneous popular uprising. Lévi-Strauss's remark that structuralism had nothing to say about it was widely noticed.

The internal critique — which became post-structuralism (5.7.2). Derrida's 1966 lecture at Johns Hopkins argued that structuralism secretly relied on a fixed centre that its own principles forbade; if meaning is differential all the way down, there is no stable structure to describe.

And the empirical objections. The binary method could be applied to anything and disconfirmed by nothing. The synchronic bias could not handle change. And the universalism was challenged as European categories imposed on other societies — the standing objection this course has now met repeatedly (4.3.10, 4.6.3).

What structuralism gave sociology

Set aside the movement and the durable contributions are substantial.

The relational insight. That elements have no properties independent of their relations is now standard, and it runs into network analysis (4.4.2), relational sociology, and Bourdieu's field theory (5.8) — Bourdieu was trained in this moment and his concept of a field as a structure of positions is unmistakably structuralist in form.

Semiotics as a method. The analysis of advertising, media, fashion, food and material culture as signifying systems. Barthes's Mythologies is still the model.

Classification as a social object. Douglas on purity, Bourdieu on classification struggles, Foucault on the order of things, and the entire literature on how categories of person are produced (2.4.1, 8.4) descend from the structuralist insight that classification systems have their own logic and their own effects.

Interpellation — the best available account of how people come to occupy identities offered to them.

And the decentring of the subject as a permanent option. Whether or not one accepts it, the availability of an analysis in which the individual is a position rather than an origin is a genuine addition to the discipline's toolkit — and it is precisely what symbolic interactionism (5.3.1) denies.

"Structuralism means studying social structure." It means analysing systems of differences in which elements have no independent properties. Social structure in the ordinary sociological sense is a different concept sharing a word.

"The signifier is the word and the signified is the thing." The signified is the concept. The thing — the referent — is outside the sign entirely.

"The arbitrariness of the sign means language is meaningless." It means the link between signifier and signified is conventional rather than natural, which is why languages differ and meanings change.

"Binary oppositions are how all cultures think." The universality claim is heavily contested, and the method's unfalsifiability is a serious objection.

"Althusser's ISAs are a conspiracy theory." They are institutions reproducing the conditions of production without anyone coordinating them; his claim is structural, not intentional.

"Structuralism has been refuted." The relational method, semiotics, classification analysis and interpellation are all in use. The universalist and synchronic claims did not survive.

Why it matters

It supplies the most rigorous account of how meaning works, and the account has consequences: if meaning is differential, then categories are systems, and changing one term changes all the others. That explains why classification struggles are so bitter and why adding a category to a scheme alters what every other category means.

It founded the analysis of culture as a system rather than a reflection. Barthes's Mythologies showed that a photograph, a car or a plate of steak and chips could be read for its ideological work — which is the foundation of media and cultural analysis (9.6).

It gives a name to the mechanism by which people take up identities. Interpellation is the best short account anyone has of why a person turns when hailed, and becomes the thing they are called.

And it poses the question the rest of Part 5 has to answer. If subjects are effects of structures, what becomes of agency? Symbolic interactionism says structures are made in encounters; structuralism says encounters are made by structures. Bourdieu and Giddens are both attempts to hold both at once, and they are what Topic 5.8 takes up.

Saussure: "In language there are only differences, without positive terms." The sign = signifier (sound-image) + signified (concept); the referent is outside it; the relation is arbitrary. Langue (the system) versus parole (utterances), with structuralism attending to the former; synchronic over diachronic, which is the source of its trouble with change; and the syntagmatic (combination) and paradigmatic (substitution) axes.

Structure = a system of relations in which elements have no properties independent of their relations. Hence the decentring of the subject: the individual is a position, not an origin.

Lévi-Strauss generalised the method: kinship as exchange founded on the incest taboo (and Rubin's feminist critique of women as objects of that exchange); myth as the mediation of unresolvable contradictions, analysed through mythemes and binary oppositions.

The binary method's best results: Mary Douglas on dirt as "matter out of place" and dietary prohibition as a response to anomaly; Leach on animal categories; and Barthes's Mythologies, where myth is a second-order system that naturalises the historical — ideology critique with linguistic tools.

Althusser: the epistemological break, theoretical anti-humanism, overdetermination ("the lonely hour of the last instance never comes"), relative autonomy, ideological state apparatuses, ideology as material and as "the imaginary relationship of individuals to their real conditions of existence," and interpellation — the hail that produces the subject who answers it. Attacked by Thompson's The Poverty of Theory; and he killed his wife in 1980 and was never tried.

The moment ended with May 1968, Derrida's internal critique, and the objections to unfalsifiability, the synchronic bias and the universalism.

What survives: the relational method (into field theory and network analysis), semiotics, classification as a social object, and interpellation.

Sign / signifier / signified — the unit of meaning, its material form, and its concept.

Referent — the thing in the world; outside the sign in Saussure's scheme.

Arbitrariness of the sign — the conventional rather than natural link between signifier and signified.

Langue / parole — the underlying system versus actual utterances.

Synchronic / diachronic — the system at a moment versus its change over time.

Syntagmatic / paradigmatic — relations of combination versus relations of substitution.

Decentring of the subject — the removal of the conscious individual from the position of origin of meaning.

Mytheme — the smallest unit of a myth, in Lévi-Strauss's analysis.

Binary opposition — a paired distinction, usually asymmetric, structuring a classification system.

Myth (Barthes) — a second-order signifying system that naturalises the historical and contingent.

Overdetermination — a contradiction shaped by multiple irreducible determinations.

Ideological state apparatus — an institution reproducing the conditions of production primarily through ideology.

Interpellation — the hailing by which ideology constitutes individuals as subjects.

One — find a term that is only a difference. Take any everyday category — "casual," "professional," "local," "adult" — and try to define it without reference to what it is not. You cannot.

Two — do a Barthes. Take one advertisement and identify the first-order sign and the second-order myth it carries. Ask what historical arrangement is being made to look natural.

Three — test Douglas. Find something your culture treats as disgusting that is not harmful. Ask which classification it violates. Anomaly, not danger, is usually the answer.

Four — catch an interpellation. Notice the next time you respond to a category being called — a job title, a nationality, a customer number, a group name. Note that answering to it is what makes you it, for that moment.

Five — carve a spectrum. Ask speakers of two languages to sort colour chips, or to name degrees of relatedness in a family. The lines fall differently, and the spectrum has none.