Culture as Constraint vs Culture as Toolkit

Everything in this topic so far has implied a particular picture: culture is a set of values, norms and dispositions that people internalise, and which then shapes what they do. Values generate norms, norms are enforced by sanctions, and habitus makes it all feel natural.

That picture is not wrong. It is incomplete, and the way it is incomplete has practical consequences — because it gives bad advice about how to change anything.

A sociologist interviews middle-class couples about love and marriage. She asks them what love is, how they knew, what marriage requires.

She gets two completely different answers from the same people, in the same conversation.

In one register, they speak the language of romance. There is one person for you. You know when you meet them. Love is a mysterious, overwhelming recognition. Marriage is the union of two souls, and if the love goes, the marriage has lost its point.

In the other register, sometimes forty seconds later, they speak the language of pragmatism. Love is something you build. It requires work, compromise and daily choice. You do not marry a feeling; you marry a person, and you keep choosing them. Compatibility matters more than chemistry.

These are not two groups of people. They are the same individuals, holding both accounts, apparently without discomfort — and, on inspection, using them for different jobs. The romantic register appears when they are explaining why they married this person rather than someone else, and when they talk about the beginning. The pragmatic register appears when they are explaining how a marriage is sustained, and when they discuss friends whose marriages are in difficulty.

That study is Ann Swidler's Talk of Love (2001), and the observation that the same people carry contradictory cultural accounts and deploy them situationally is the beginning of a genuine reorientation in how sociologists think about culture.

The two models, in plain words

Swidler's own formulation is the neatest: culture influences action less by supplying the ends people pursue than by supplying the repertoire of capacities from which they construct lines of action. People have more culture than they use. What varies is which parts get activated.

Swidler's key distinction: settled and unsettled lives

Her most useful addition, and the one that resolves much of the argument.

In settled lives, culture and practice are deeply intertwined. Institutions are stable, routines work, and culture operates largely as unnoticed habit — supporting arrangements already in place. Here culture looks like constraint, because it is invisible and effortless. People cannot easily say why they do things; they simply do them. (This is essentially Bourdieu's habitus in its natural habitat.)

In unsettled lives — during migration, conversion, revolution, rapid economic change, war, a new social movement, or a personal upheaval — the routines no longer work. Now culture becomes explicit and contested: people argue about ideology, adopt new doctrines, take instruction, join movements, and self-consciously reconstruct how to live. This is where formal ideologies do their work.

And her sharpest point: in unsettled periods, culture's causal power is high but the link between belief and action is weak, because people are trying out doctrines faster than they can live them. In settled periods, the link between culture and action is tight but the culture is invisible.

This dissolves an old dispute. Weber's Protestants were living in an unsettled period — a new doctrine reorganising conduct. Once accumulation became the ordinary way of doing business, the doctrine was no longer needed: the "iron cage" is Weber's own description of culture becoming settled practice that no longer requires belief.

The third position: dual-process

The most useful contemporary synthesis, and worth knowing because it explains something both models struggle with.

Stephen Vaisey's argument (developed from cognitive science and from the sociology of Bourdieu and DiMaggio) is that both camps are describing real but different things, because there are two kinds of cultural process at work in a person.

Practical, automatic culture — fast, intuitive, embodied dispositions that actually guide most behaviour. This is habitus, and it is largely unavailable to introspection: people cannot accurately report it.

Discursive, deliberative culture — the slower, verbal repertoire of accounts, justifications and stories people use to explain themselves to others and to themselves. This is Swidler's toolkit, and it is what interviews and surveys mostly capture.

The consequence is important and slightly deflating for researchers: when you ask people why they did something, you are largely sampling the second system, not the first. Which explains the espoused/enacted gap from 2.1.2, the couples' two registers, and the general unreliability of asking people about their motives — not because they lie, but because the account and the action come from different processes.

A family in a small city continues to arrange marriages within their community.

Asked why by a stranger, the father gives a values account: this is our tradition; families know better than young people; it has worked for generations.

Asked by a sympathetic relative, he gives a pragmatic account: the girl will have people nearby who will look after her; we know the family's finances; a marriage outside would mean nobody to call on in trouble.

Asked by his own son during an argument, he gives a third: what will people say; your sister is unmarried; do you want your mother to be talked about.

None of these is a lie. All three are in the repertoire, and each is selected for its audience. And note what the toolkit model reveals that the values model hides: if you wanted to change this practice, you would need to know which account is load-bearing — because a campaign against tradition, a campaign about risk and support networks, and a campaign about community gossip are three entirely different interventions.

The values model says: change what they believe. The toolkit model says: find out which resource is doing the work, and notice that a new one — a credible alternative source of support, or a shift in what neighbours actually say (2.1.4) — may change the behaviour without changing anybody's beliefs at all.

Go deeper

What the toolkit model gets right

It explains inconsistency without calling it hypocrisy. People hold contradictory cultural materials and use them situationally. This is normal, not a defect, and any model that treats it as a defect will misread most of what people say.

It explains why cultures change faster than beliefs. If practice depends on which repertoire elements are activated, a change in circumstances can shift behaviour rapidly while beliefs lag — the reverse of the usual assumption.

It makes culture a resource for agency rather than only a constraint. People use culture: to justify, to persuade, to construct an identity, to make a novel action intelligible. This connects directly to 1.3.1: the toolkit is the cultural side of the structure–agency synthesis.

And it gives a realistic theory of intervention. If culture supplies capacities, then providing new capacities — a script, an example, a vocabulary, a visible model of a different life — can be more effective than arguing about values.

What it gets wrong, or at least underplays

Fairness requires the other column.

Not all tools are available to everyone. A repertoire is itself unequally distributed — which is 2.1.13's cultural capital. Talking as though people select freely from a common cultural stock erases exactly the inequality Bourdieu was describing.

Some culture really does constrain. Return to 1.1.3's disgust. That is not a resource you deploy; it is a reflex you cannot switch off. Taboos (2.1.3), internalised shame (2.1.4) and habitus are constraint in the strong sense, and a purely instrumental picture of culture cannot handle them.

Justifications are not merely decorative. Which accounts are available and legitimate shapes what can be done at all. If no acceptable vocabulary exists for a course of action, it becomes very hard to take — which is Foucault's point about discourse (2.1.5) and Bourdieu's about doxa.

And "toolkit" flatters the actor. It can slide into a picture of the strategic, self-aware culture-user, which fits an educated professional better than most people, and is exactly the kind of implicit self-portrait 1.5.3 warned you to check for.

The honest synthesis

Most working cultural sociologists now hold something like this, and you should be able to state it:

Culture operates in at least two registers — automatic dispositions that guide most behaviour and are hard to report, and a discursive repertoire used for justification, persuasion and identity, from which people select according to audience and circumstance. Which register dominates depends on how settled conditions are. Both are unequally distributed. And the practical question in any case is not "which model is right?" but which register is doing the work here, and is it available to everybody?

Michèle Lamont's research programme is the model for doing this well: she studies the actual repertoires people use to draw boundaries and evaluate worth, comparing across countries and classes, and takes seriously both that the repertoires are patterned by position and that people use them actively (5.12, 2.1.5). William Sewell's theoretical formulation is the tidiest: culture is both a system of meanings and a set of practices, and neither can be reduced to the other.

"Culture as toolkit means people choose freely." No. Repertoires are unequally distributed, and 2.1.13 explains why.

"Toolkit refutes Bourdieu." They address different processes. Vaisey's dual-process account is the reconciliation, and Bourdieu's habitus is the automatic register.

"So asking people about their values is pointless." Not pointless — it samples the discursive register, which is genuinely important (justifications shape what is doable). It is just not a measurement of motives.

"Contradictory accounts mean people are confused or dishonest." They mean people hold repertoires. Treat contradiction as data about audiences, not as a failure of the respondent.

"Change the culture and behaviour follows." Frequently the reverse: change the practice, the incentives or the visible examples, and accounts adjust afterwards. This is 1.3.6's material–ideal question with a mechanism.

"Settled and unsettled is about historical periods." It is about the fit between routines and conditions, and it can apply to one person's life — a migration, a bereavement, a conversion.

Why it matters

It fixes the naive theory of social change. Most people, faced with a practice they dislike, reach for persuasion: change their values. This lesson says: find out which register the practice runs on. If it runs on automatic dispositions, argument will not reach it; if it runs on justifications, the question is which alternative accounts are available and legitimate; if it runs on what neighbours will say, the target is the audience, not the actor (2.1.4).

It makes you a better interpreter of what people tell you. In any interview, conversation or survey, you are now asking: which register is this, and who is the audience? That single question will improve your reading of evidence more than any additional theory.

And it completes the picture of culture as an object. Fourteen lessons ago culture was a vague word meaning roughly everything. You can now say what it is made of, how it is enforced, how it is ranked, how it is distributed as capital, and — as of this lesson — how it actually enters into action. One lesson remains: how it moves.

Two models of how culture works. Constraint (values model): internalised values supply people's ends, and action follows (Parsons). Toolkit (repertoire model): culture supplies a stock of capacities, symbols and justifications from which people select according to circumstance (Swidler).

Swidler's evidence: the same people hold contradictory cultural accounts — romantic and pragmatic love — and deploy them for different jobs, without discomfort.

Her key distinction: in settled lives culture is invisible habit intertwined with practice; in unsettled lives it becomes explicit and contested, with high causal power but a weak belief–action link.

Vaisey's dual-process synthesis: automatic, practical culture (habitus) guides most behaviour and is hard to report; discursive culture supplies justifications and is what interviews capture. Hence the espoused/enacted gap.

The toolkit model explains inconsistency, rapid change and cultural agency — and underplays unequal access to repertoires, genuine constraint (disgust, taboo, shame), and the power of available discourse.

The synthesis: two registers, unequally distributed, with the practical question being which register is doing the work here.

Values model (culture as constraint) — the view that internalised values supply the ends of action.

Toolkit / repertoire model — Swidler's view that culture supplies capacities, symbols and justifications used situationally.

Repertoire — the stock of cultural elements available to a person or group for constructing and justifying action.

Settled lives — conditions in which culture and practice are intertwined and culture operates as invisible habit.

Unsettled lives — conditions of upheaval in which culture becomes explicit, contested and ideological.

Dual-process model — Vaisey's account of two cultural registers: automatic-practical and discursive-deliberative.

Habitus — durable, largely unconscious dispositions acquired from social position (the automatic register).

Espoused / enacted values — what people say they value versus what their choices reveal (2.1.2).

Account (justification) — a culturally available explanation offered for a line of action.

Symbolic boundary — a distinction used to sort people and evaluate worth; Lamont's repertoire research (2.1.5).

One — two registers on yourself. Take one significant choice of your own. Write the account you would give a stranger, and the account you would give your closest friend. Compare. Neither is a lie; note what each is for.

Two — find the load-bearing account. Choose a practice in your community you would like to see change. List every justification people actually give for it. Then try to work out which one, if removed, would matter — and which are decoration.

Three — settled or unsettled? Identify one period in your own life when the ordinary routines stopped working. What did you read, join, argue about or adopt in that period? That is Swidler's unsettled culture, from the inside.

Four — test the interview problem. Ask someone why they do something habitual — how they greet people, what they eat for breakfast, how they address elders. Notice how quickly the explanation becomes a post-hoc account. This is the dual-process finding, reproduced at a kitchen table.