The Seedbed of Functionalism

For roughly thirty years — from the 1930s to the early 1960s — one theoretical framework dominated sociology in the English-speaking world. It was called structural functionalism, it claimed Durkheim as its ancestor, and it was then attacked so comprehensively that "functionalist" became an insult.

This lesson sorts out what Durkheim actually said, what his successors built, what was wrong with it, and what survived. Part 5 gives functionalism a full treatment as a school; this lesson establishes the inheritance.

A question a doctor asks, and a question a historian asks.

Take the heart. A doctor asks: what does it do? Answer: it circulates blood, which delivers oxygen, which keeps the organism alive. That is a functional explanation, and in physiology it works beautifully. Organs are for something. Understanding what they contribute to the whole tells you a great deal.

Now take the necktie. A doctor's question — what does it do? — yields nothing useful. It does not keep you warm. It serves no protective purpose. It is a strip of cloth tied under the chin.

The historian's question is different: where did it come from? Answer: seventeenth-century Croatian mercenaries in French service wore a distinctive knotted neckcloth; the French court adopted it — cravate, from Croate — and three and a half centuries of drift produced the modern tie.

Neither question is wrong. But notice what happens if you ask only the first. You will be driven to invent a function for the necktie, because your framework requires one. You will find yourself explaining that it signals professionalism, marks status, and expresses conformity to organisational norms — all of which are effects, and none of which explains where the thing came from or why it is that shape.

The whole controversy about functionalism is contained in that trap, and Durkheim saw it coming.

The intellectual problem is real and does not go away by ridiculing functionalism.

Societies do have interrelated parts. Change one and others shift. Abolish an institution and things you did not expect start failing. The family, the school, the legal system, the labour market and the welfare system are visibly interlocked, and analysing any of them in isolation produces bad conclusions.

And practices do persist. Something keeps arrangements in place across generations, often against the preferences of many of the people living under them.

So: is there a legitimate form of functional explanation in sociology, or is the appearance of interlocking parts a trap?

Durkheim's answer was carefully hedged. His successors dropped the hedges. And the critics, in demolishing the unhedged version, came close to throwing out the legitimate question along with it.

What Durkheim actually said

Three things, and each has been misreported.

1. Function is not cause — his explicit warning

From The Rules of Sociological Method (1895):

"When, then, the explanation of a social phenomenon is undertaken, we must seek separately the efficient cause which produces it and the function it fulfils."

This is the single most ignored sentence in Durkheim, and it was ignored by the people who named their school after him. It is also, notably, the exact criticism later levelled at functionalism — issued by the man usually blamed for the error.

2. Function means contribution to the whole, not usefulness to individuals

Durkheim's definition of function is technical: the correspondence between a phenomenon and the general needs of the social organism. Not "what people get out of it." A practice can be functional for a society and unpleasant for most of its members — which is why the concept is not a justification.

3. The organic analogy, and how far he took it

Durkheim uses biological language throughout: organs, functions, the social organism, normal and pathological states. He inherits it from Comte and Spencer (3.3.1, 3.3.2).

How seriously did he mean it? More seriously than his defenders admit and less than his critics claim. He genuinely thought societies were systems whose parts stand in relations of interdependence, and that this made a certain kind of physiological reasoning appropriate. He also insisted repeatedly that social facts require social explanation, not biological reduction — that sociology is autonomous (4.2.1).

The trouble is that the analogy carries assumptions the user may not notice. Organisms have clear boundaries; societies do not. Organisms have an unambiguous survival criterion; societies do not. Organisms' parts do not have interests opposed to the whole; a society's parts do, constantly. The analogy imports harmony as a premise.

What his successors built

A. R. Radcliffe-Brown carried Durkheim into anthropology and made function the central concept of a research programme: the function of any practice is its contribution to the maintenance of the total social structure. His famous studies of joking relationships and of mortuary rites in the Andaman Islands analyse practices entirely in terms of what they do for social integration.

He was explicit and unembarrassed about the organic analogy, and he substituted "social structure" for Durkheim's more variable vocabulary. He is also where the synchronic bias enters: study the society as a working system at a moment, not as a historical product. That was a methodological choice with enormous consequences, and it removed exactly the "efficient cause" investigation Durkheim had insisted on.

Malinowski developed a rival, individualist functionalism — practices serve the biological and psychological needs of individuals — which Radcliffe-Brown rejected. The distinction matters: Malinowski's version is about persons, Durkheim's and Radcliffe-Brown's about systems.

Talcott Parsons is the towering figure. The Structure of Social Action (1937) and The Social System (1951) constructed an enormously elaborate general theory, drawing on Durkheim, Weber, Pareto and Marshall, in which societies are systems tending towards equilibrium, integrated by shared values internalised through socialisation.

His AGIL scheme holds that every system must solve four functional problems:

	Problem	Institutional locus
A	Adaptation — securing resources from the environment	Economy
G	Goal attainment — setting and pursuing collective goals	Polity
I	Integration — coordinating the parts, managing conflict	Law, community
L	Latency / pattern maintenance — transmitting and sustaining values	Family, education, religion

Parsons's prose was famously dense — his student Robert Merton once remarked on the difficulty, and C. Wright Mills devoted a chapter of The Sociological Imagination to translating a Parsons passage into a few plain sentences, with devastating effect (his verdict on the exercise: "grand theory").

Merton's repairs

Robert K. Merton was functionalism's most acute internal critic and rescued much of what was worth keeping. Social Theory and Social Structure (1949) systematically dismantled three assumptions.

Against functional unity — the assumption that every part is functional for the whole system. In differentiated societies, an item may be functional for some groups and dysfunctional for others. Whose function, for whom?

Against universal functionalism — the assumption that everything that persists must be doing something useful. Merton introduces dysfunctions: consequences that reduce adaptation or adjustment. Some things persist while doing harm.

Against indispensability — the assumption that each institution is necessary and irreplaceable. Merton introduces functional alternatives (or equivalents): the same function may be served by different arrangements. Socialisation happens in families, but also in schools, kibbutzim, boarding schools and religious orders.

And his most productive distinction: manifest functions — the intended, recognised consequences — versus latent functions — the unintended, unrecognised ones. He also insists on latent dysfunctions: unintended harms.

Merton's own illustration, and it repays attention because it is regularly misunderstood.

The Hopi rain ceremony. Manifest function: to produce rain. Assessed against that purpose, it fails — the ceremony does not affect precipitation.

But that assessment misses what the practice does. Latent function: it assembles a dispersed community, reaffirms shared identity, and reinforces group cohesion at a moment of shared anxiety. Which is Durkheim's argument (4.2.6), arriving through a different door.

And here is the point of the distinction, which is easy to lose. Merton is not saying "the participants are wrong about their own practice and I know better." He is saying that a practice's persistence often cannot be explained by whether it achieves its stated purpose, and that the sociologist's distinctive contribution is precisely to identify consequences invisible to the participants. He called this the discovery of latent functions "the chief contribution of the sociologist."

The danger is equally real, and Merton knew it: latent-function analysis licenses the analyst to attribute purposes nobody holds, and to find a hidden benefit in anything at all. The discipline that keeps it honest is evidence — you must show the consequence occurs, not merely assert that it would be convenient.

Now the harder application. Merton's own most notorious example is the political machine — the corrupt urban party organisation of American cities, manifestly a system of graft, latently a provider of welfare, jobs, and personalised help to immigrant populations that the formal state ignored. His argument: the machine persisted because it met needs no other institution met, and reformers who attacked it without replacing those functions failed.

That analysis is illuminating and it is also, in the wrong hands, a defence of corruption. Merton was explicit that identifying a function is not endorsing an arrangement. The distinction is easy to state and, as the next section shows, was not reliably observed.

The demolition

By the 1960s the attacks had become overwhelming. Four of them, and each has force.

One: the conservative bias. If everything that exists serves a function, everything that exists is justified. Kingsley Davis and Wilbert Moore's 1945 functional theory of stratification is the textbook case: inequality exists because societies must motivate the most talented people into the most important positions, and unequal rewards do that. Melvin Tumin's 1953 reply is one of the great demolitions in the discipline's literature — how do we know which positions are "functionally important" except by observing which are highly rewarded (circular)? What of inherited advantage, which prevents talent from reaching positions? What of the enormous waste of talent in the excluded population? And is the theory not simply the ideology of the beneficiaries, restated in technical language (4.1.7)? Tumin won the argument, decisively.

Two: it cannot explain change. A framework built on equilibrium, integration and pattern maintenance has no motor. Conflict appears as malfunction, deviance as failure of socialisation, and social movements as strain. This was a catastrophic weakness in a decade of civil rights struggles, decolonisation, anti-war movements and student revolt — Parsons's system had no vocabulary for any of it.

Three: it ignores power and conflict. C. Wright Mills (The Sociological Imagination, 1959) attacked "grand theory" for constructing an elaborate vocabulary that made domination invisible. Ralf Dahrendorf argued that any adequate theory must handle both the integration face and the coercion face of society, and that functionalism saw only one. Alvin Gouldner's The Coming Crisis of Western Sociology (1970) demonstrated at length that Parsons's apparently neutral system theory encoded the political assumptions of postwar American liberalism.

Four: teleology and illegitimate reasoning. The logical objection. "Institution X exists because it performs function F" reverses cause and effect unless you can specify a feedback mechanism — how do the beneficial consequences actually reach back and cause the persistence of the institution? In biology, natural selection supplies that mechanism. In sociology, no general equivalent exists. Without one, functional explanation is a story about consequences dressed as an explanation of causes.

What survived, and it is more than the rhetoric suggests.

Unintended consequences. Merton's manifest/latent distinction is now simply part of how sociologists think, usually without attribution. Every study showing that a policy produced effects nobody intended is doing latent-function analysis.

Dysfunctions. The idea that persistent arrangements can be harmful — and harmful to some while beneficial to others — is now standard.

Systemic interdependence. That institutions are interlocked and cannot be analysed in isolation is uncontroversial and correct. Change the labour market and the family changes; change the family and the school changes.

Functional alternatives. Comparative research constantly asks: what else could serve this purpose? Different countries solve pensions, childcare and healthcare in structurally different ways — an insight with real policy value.

And the legitimate form of functional explanation, which analytical philosophy of social science eventually specified. G. A. Cohen's work on Marx and Jon Elster's critique established the rule: a functional explanation is legitimate if and only if you can specify the feedback mechanism by which the consequences reinforce the cause. Selection among competing organisations, deliberate design by actors who understand the effect, and reinforcement learning all qualify. "It persists because it is useful," with no mechanism named, does not.

That is the durable settlement, and it is a genuine advance on both the functionalists and their critics.

Go deeper

Was Durkheim a functionalist?

The fair answer is: partly, and less than the label implies.

For. He used the organic analogy extensively. He explained crime, punishment and ritual by their contributions to social cohesion. He worked with normal/pathological categories that presuppose a healthy system state. His account of religion is functionalist in shape. And his political sympathies were reformist rather than revolutionary.

Against. He explicitly separated function from efficient cause and insisted both be investigated. He wrote a whole book on the abnormal forms of the division of labour (4.2.2) — pathologies produced by the system's own development, not by external disturbance. He argued for abolishing inheritance and for equality of opportunity as a condition of legitimate solidarity, which is not the position of a man who thinks what exists is functional. He gave a substantial account of anomic conflict between labour and capital. And he was a political activist in the Dreyfus Affair on behalf of an individual against the entire weight of the institutional order (4.2.5).

The verdict: Durkheim supplied concepts that functionalism used, and functionalism dropped the qualifications he attached to them. Judging him by Parsons is like judging Marx by the Soviet textbooks — and the resemblance between those two cases is worth noticing, since both are examples of a founder being simplified into an orthodoxy by successors with institutional agendas.

Neofunctionalism and the quiet return

Functionalism did not die so much as go underground and come back renamed.

Jeffrey Alexander's neofunctionalism in the 1980s attempted an explicit revival, retaining systemic analysis while incorporating conflict, contingency and change.

Niklas Luhmann built a systems theory of enormous sophistication in which modern society consists of functionally differentiated autopoietic subsystems — law, economy, science, politics, art — each operating by its own binary code (legal/illegal, payment/non-payment, true/false) and unable to communicate directly with the others. It is functionalism without the harmony assumption, and it is one of the most substantial bodies of theory produced in the late twentieth century.

New institutionalism in organisational sociology asks why organisational forms persist and spread — and answers with mechanisms: coercive, mimetic and normative isomorphism (DiMaggio and Powell). That is functional-style analysis with the feedback mechanism supplied, which is exactly what the critics demanded.

Evolutionary and cultural-selection approaches in the study of institutions do the same: practices persist because of selection among competing variants, which is a proper mechanism.

And the whole "social capital" and "civic health" literature — Putnam's Bowling Alone and its enormous progeny — is Durkheimian integration theory in modern dress, with survey data. It attracts the same criticisms: insufficient attention to power, a tendency to treat cohesion as unambiguously good, and difficulty explaining who is excluded from the cohesive community.

"Durkheim founded functionalism." He supplied its concepts and explicitly warned against its central error. The school was assembled by Radcliffe-Brown and Parsons.

"Functionalism says everything is good." It says persistent arrangements have systemic consequences. Merton added dysfunctions specifically to break the equation — though the framework does have a conservative tilt that Merton's repairs did not eliminate.

"Latent function means the real, hidden purpose." It means an unintended, unrecognised consequence. Purposes are held by people; consequences are not.

"Functionalism is dead." The label is unfashionable; systemic analysis, unintended consequences, functional alternatives and institutional persistence are all central to contemporary work. Luhmann and new institutionalism are functionalism with mechanisms.

"Functional explanation is always invalid." It is valid when a feedback mechanism can be specified. That is the settled position, and it is a real one, not a compromise.

"Davis and Moore proved inequality is necessary." Tumin's reply exposed the circularity (importance measured by reward), the waste of excluded talent, and the ideological character of the argument. It is taught now mainly as an example of what went wrong.

Why it matters

It shows what happens to a founder's ideas after they die. Durkheim's careful separation of cause and function was dropped by a school that then claimed him. Understanding this protects you from two errors: attacking Durkheim for functionalism's sins, and taking any school's account of its own ancestry at face value.

It teaches the difference between a consequence and an explanation. This is one of the most transferable pieces of reasoning in the discipline. "X exists because it does Y" requires a mechanism. Demanding one, every time, will improve your thinking about a great deal more than sociology.

It supplies the vocabulary sociology still uses. Manifest and latent, function and dysfunction, functional alternatives, systemic interdependence — these are working tools, and their origin is here.

And it explains the shape of Part 5. Functionalism's dominance and collapse is why conflict theory (5.2), symbolic interactionism (5.4) and the critical schools (5.5) took the forms they did. Each was built partly as an answer to it.

Durkheim's explicit warning — "we must seek separately the efficient cause which produces it and the function it fulfils" — is the sentence functionalism ignored while claiming him as founder.

Radcliffe-Brown made function the central concept of British social anthropology and introduced the synchronic bias. Parsons built the general theory: systems tending to equilibrium, integrated by internalised shared values, with four functional imperatives — AGIL (adaptation, goal attainment, integration, latency).

Merton's repairs: against functional unity (functional for whom?), against universal functionalism (dysfunctions), and against indispensability (functional alternatives) — plus manifest and latent functions, whose discovery he called the sociologist's chief contribution.

The demolition: the conservative bias (Davis–Moore on stratification, destroyed by Tumin); the inability to explain change; the invisibility of power and conflict (Mills, Dahrendorf, Gouldner); and teleology — functional explanation is illegitimate unless a feedback mechanism is specified.

What survived: unintended consequences, dysfunctions, systemic interdependence, functional alternatives, and the rule that functional explanation is valid only with a named mechanism — satisfied by selection among organisations, deliberate design, or reinforcement.

Was Durkheim a functionalist? Partly. But he separated cause from function, wrote a book on the system's own pathologies, argued for abolishing inheritance, and campaigned for one man against the institutional order. Judging him by Parsons is like judging Marx by Soviet textbooks.

The return: Alexander's neofunctionalism, Luhmann's functionally differentiated autopoietic subsystems, new institutionalism's isomorphism mechanisms, and the social capital literature.

Efficient cause — what brought a phenomenon into existence, as distinct from what it contributes.

Function — the contribution a phenomenon makes to the working of a larger system.

Structural functionalism — the school treating societies as systems of interdependent parts tending towards equilibrium.

AGIL — Parsons's four functional imperatives: adaptation, goal attainment, integration, latency.

Manifest function — an intended and recognised consequence.

Latent function — an unintended and unrecognised consequence.

Dysfunction — a consequence that reduces the adaptation or adjustment of a system or group.

Functional alternative (equivalent) — a different arrangement serving the same function.

Teleology — explaining something by its purpose or end; illegitimate in social explanation without a feedback mechanism.

Feedback mechanism — the specified process by which consequences reinforce the arrangement producing them; the condition of a valid functional explanation.

Davis–Moore thesis — the functional theory of stratification, refuted by Tumin.

Autopoietic subsystem — Luhmann's term for a self-reproducing functional system operating on its own binary code.

One — run the necktie test. Take any practice you find puzzling and ask both questions: where did it come from, and what does it do now? Notice how often the two answers have nothing to do with each other.

Two — find a latent function. Take an institution with a clear stated purpose — exams, job interviews, office attendance, professional licensing — and list consequences nobody officially intends. Then ask what evidence would show those consequences actually occur.

Three — demand a mechanism. Next time you read "this practice persists because it serves the interests of X," ask exactly how those interests reach back and sustain the practice. Sometimes there is a clear answer; often there is not, and the sentence collapses.

Four — apply Merton's three challenges. For one institution, ask: functional for whom? Dysfunctional for whom? What could serve the same function instead? Three questions that improve almost any analysis.

Five — read the Tumin reply. Davis and Moore (1945) and Tumin (1953) are both short and both available. Reading them in sequence is the best short course in argument the discipline offers.