The Collapse and the Return

Between roughly 1960 and 1975, structural functionalism went from being the dominant framework in American sociology to being a term of abuse. Few theoretical collapses have been so complete or so fast.

And then, from about 1980, systemic analysis quietly returned — under other names, with better mechanisms, and largely without acknowledging where it came from.

This lesson takes both movements, because the pattern is instructive well beyond functionalism.

A conference session, some time in the late 1960s.

A senior scholar presents a paper analysing student protest as a symptom of strain in the socialisation subsystem — a failure of the mechanisms by which values are transmitted between generations, producing deviance among those imperfectly integrated.

A young sociologist stands up. She points out that the students in question are not imperfectly socialised. They are, by any measure, the most educated and most articulate cohort in the country's history. They are not confused about the society's values; they are objecting to them, in detail, with arguments. And what they are objecting to includes a war their government is prosecuting, a racial order their laws sustain, and a set of institutions in which they have no voice.

"Calling that a socialisation failure," she says, "is not an analysis. It is a way of not having one."

Something like this exchange happened repeatedly across the discipline in that decade. And the point was not that functionalism gave the wrong answers about the 1960s.

It was that functionalism could not formulate the questions. A framework built around how systems maintain themselves has no vocabulary in which "the system should not be maintained" is a coherent position. It can register it only as strain, deviance or breakdown — as a symptom of the system rather than as a judgement on it.

And that is a limitation of the theory, not of the theorists, which is why better functionalists could not fix it.

The interesting question is not "was functionalism criticised?" — every theory is.

It is: why did the criticisms suddenly become decisive, when most of them had been available for years?

The teleology objection was old. Merton had dismantled the three postulates in 1949 (5.1.2). Mills's attack came in 1959. Dahrendorf's conflict critique was published in 1958. None of these destroyed the framework at the time.

What changed between 1960 and 1970 was not the arguments. It was the world the arguments were about — and, decisively, who was in the room.

The postwar expansion of higher education produced a generation of sociologists from backgrounds that had not previously supplied them: working class, Jewish, Black, and increasingly female. They arrived with different questions, and the framework had no place for them.

That is a finding in the sociology of knowledge (5.15) about how theoretical change actually occurs — not by refutation but by the arrival of people for whom a framework's blind spots are the whole subject.

The four attacks

Each has force, and each is partly answerable.

1. It cannot explain change

2. It ignores power and conflict

C. Wright Mills (The Sociological Imagination, 1959) argued that grand theory's abstraction made domination literally unspeakable — that a vocabulary of values, norms and integration has no words for who commands whom.

Ralf Dahrendorf ("Out of Utopia," 1958) put it most precisely: every society has two faces — an integration face and a coercion face — and any adequate theory must handle both. Functionalism sees only one, and a theory that can describe a society only as a going concern held together by shared values is describing a utopia, not a society.

Alvin Gouldner (The Coming Crisis of Western Sociology, 1970) demonstrated at length that Parsons's system encoded the political assumptions of postwar American liberalism — and, being Parsons's own student, could not be dismissed as an outsider.

3. It is conservative

The charge: if everything that exists performs a function, everything that exists is justified.

The textbook case is the Davis–Moore thesis (4.2.7). Kingsley Davis and Wilbert Moore argued in 1945 that stratification is functionally necessary: societies must fill their most important positions with their most talented people, difficult and important positions require training and sacrifice, and unequal rewards are the mechanism that motivates this.

Melvin Tumin's 1953 reply is one of the great demolitions in the discipline. How do we identify "functionally important" positions except by observing which are highly rewarded — which makes the argument circular? What of inherited advantage, which prevents talent from reaching positions? What of the enormous waste of talent in the excluded population? Are unequal material rewards necessary, or would other incentives work? And is stratification not dysfunctional, since it produces hostility, restricts the pool of talent and requires ideologies to justify it?

Tumin won, and the exchange is now taught mainly as a case study in what went wrong.

The reply available to functionalism: Merton's dysfunctions and "functional for whom" were designed to prevent exactly this, and Merton stated explicitly that functional analysis has no intrinsic ideological commitment.

Why the reply is insufficient: the tilt is structural. A framework whose central question is how systems maintain themselves will, on average, find reasons why arrangements persist. Merton's guards are warnings, not mechanisms.

4. Teleology

The logical objection, and the most serious (4.2.7). "X exists because it performs function Y" reverses cause and effect unless you can specify a feedback mechanism by which the consequences sustain the cause. Biology has natural selection. Functionalism had no general equivalent, and never produced one.

Set the framework against what actually happened between 1955 and 1975, and the failure is specific rather than general.

The civil rights movement. Functionalism's tools: strain, deviance, imperfect integration. What was actually occurring: a sustained, organised, strategically sophisticated campaign by people with a clear analysis of their position, mobilising churches, colleges, legal expertise and national media. The framework had no concept of a movement as a purposive collective actor, and the resource mobilisation theory that supplied one (11.2) was developed explicitly against it.

Decolonisation. Between 1945 and 1975, most of the colonised world became independent (4.5.3). The functionalist reading available at the time was modernisation theory — traditional societies developing towards modernity through differentiation, with the developed world as the destination. Its predictions failed comprehensively, and the dependency theorists' reply (4.6.3) was that underdevelopment was produced by the relation, not a stage prior to it.

Vietnam. A framework in which the polity performs goal attainment on behalf of the societal community has no way to register that a government might be pursuing goals its population opposes and lying about them.

And second-wave feminism. Parsons's own analysis of the family — the instrumental male role and expressive female role as a functional specialisation contributing to system stability — became, in about five years, the standard illustration of what was wrong with the whole framework. It described an arrangement, explained why it was necessary, and could not register that half the population was about to reject it.

The pattern in all four: the framework could describe the arrangement and could not register the objection.

What came back

And now the second movement, which is less often taught.

Luhmann

Niklas Luhmann (1927–1998), who studied with Parsons at Harvard, built the most substantial systems theory of the late twentieth century — and removed the assumptions that had sunk the original.

His central move: modern society is not an integrated whole with a value consensus. It is a set of functionally differentiated subsystems — law, economy, politics, science, religion, art, education, medicine — each of which is autopoietic: self-producing, operationally closed, and reproducing its own elements by its own operations.

Each subsystem operates on a binary code that it alone applies: legal/illegal for law, payment/non-payment for the economy, true/false for science, power/no-power for politics.

And the subsystems cannot communicate directly. Law cannot process an economic argument as law; it can only translate it into legal/illegal. Politics cannot compel science to find something true. Each observes the others only through its own code — Luhmann calls this structural coupling.

No integration, no consensus, no centre. Modern society, on this account, has no apex from which it could be steered, which is why political attempts to control the economy or the media so reliably fail. That is a systems theory whose central finding is the impossibility of the control Parsons's framework assumed.

Luhmann is enormously difficult, wrote around seventy books, and is far more influential in Germany and continental Europe than in the Anglophone world. His debate with Habermas across the 1970s is one of the major theoretical confrontations of the period. His work is not functionalism as Parsons practised it, and it is unmistakably its descendant.

Neofunctionalism

Jeffrey Alexander proposed an explicit revival in the 1980s — retaining systemic analysis and the attention to culture while incorporating conflict, contingency, agency and change. It had modest success as a movement and its more durable outcome was Alexander's strong programme in cultural sociology (5.12), which insists that collective representations have autonomous causal force.

New institutionalism

The most successful return, and the one that supplied the missing mechanism.

The question is functionalist in form: why do organisational forms persist and spread? The answer is not "because they are useful" — DiMaggio and Powell's classic 1983 paper observes that organisations in a field become more similar over time even when this makes them less efficient.

Their explanation names three mechanisms of isomorphism:

Coercive — pressure from those on whom an organisation depends: regulation, law, funders' requirements. Mimetic — under uncertainty, organisations copy others they regard as successful. Nobody knows what works, so everybody copies. Normative — professionalisation. People trained in the same programmes, belonging to the same associations, carry the same templates between organisations.

This is functional-style analysis with the feedback mechanism supplied, which is exactly what the critics demanded. And it explains legitimacy-seeking rather than efficiency: organisations adopt forms because doing so makes them recognisable as proper organisations, whether or not the forms work.

And the rest

Evolutionary and selection approaches in organisational ecology and institutional economics: practices persist because of selection among competing variants — a proper mechanism.

Social capital and civic health research — Putnam's Bowling Alone (2000) and its enormous progeny — is Durkheimian integration theory with survey data, and it attracts the same criticisms: insufficient attention to power, a tendency to treat cohesion as unambiguously good, and difficulty with who is excluded from the cohesive community.

Functional analysis in comparative welfare research. Esping-Andersen's welfare regimes (4.3.8) ask what functions a welfare state performs and how different arrangements perform them — which is Merton's functional alternatives, applied comparatively.

Go deeper

The honest ledger

Dead, and rightly:

The claim of a general theory covering all levels in one scheme. AGIL as an explanatory apparatus. Value consensus as the ordinary condition of societies. The assumption that persistence implies utility. The Davis–Moore account of stratification. Modernisation theory's stages. And the idea that conflict is a symptom of imperfect integration.

Alive, and central:

Unintended consequences. Every study showing a policy produced effects nobody intended is doing latent-function analysis (5.1.2).

Systemic interdependence. That institutions are interlocked, and that changing one changes others, is uncontroversial and correct.

Functional alternatives. Comparative policy research asks constantly what else could serve a purpose.

Dysfunctions, and "functional for whom."

And the legitimate form of functional explanation, now specified: valid if and only if a feedback mechanism can be named — selection among competing organisations, deliberate design by actors who understand the effect, or reinforcement. That is a real settlement rather than a compromise, and it is more than either the functionalists or their critics had.

The pattern, which is the lesson

This is the second time in this course a school has followed the same trajectory, and it will not be the last.

Dominance — an institutional base, funding, a trained generation, and a congenial moment. Overreach — the framework is applied to everything, including things it cannot handle. Collapse — new people arrive with questions the framework cannot formulate, and old criticisms suddenly become decisive. Overcorrection — the label becomes an insult and the valid parts are discarded with the invalid. And quiet return — the useful components come back under new names, with better mechanisms, frequently without acknowledgement.

Marxism followed it. Functionalism followed it. Behaviourism followed it in psychology. Post-structuralism has arguably been through it. And whatever framework currently feels obviously correct is somewhere on the same curve.

The practical implication: when a school is being demolished, ask which of its components are being destroyed and which are merely unfashionable. The second category is where the useful salvage is.

"Functionalism was refuted." Its central logical defect — teleology without a mechanism — was correctly identified, and it was displaced by a change in the discipline's population and preoccupations as much as by argument.

"Functionalism is dead." The label is unfashionable. Systemic interdependence, unintended consequences, dysfunctions, functional alternatives and legitimacy-seeking isomorphism are all central to contemporary work.

"Luhmann is a functionalist." He is a systems theorist descended from Parsons who removed the assumptions of integration, consensus and central steering. His central finding is that modern society cannot be steered.

"New institutionalism is unrelated to functionalism." It asks a functionalist question — why forms persist and spread — and answers it with named mechanisms, which is what the critics demanded.

"Davis and Moore showed inequality is necessary." Tumin exposed the circularity, the waste of excluded talent and the ideological character of the argument. It is taught now as an example of the framework's failure mode.

"The 1960s disproved functionalism." The decade made visible phenomena the framework could not formulate. That is a different and more serious kind of failure than being wrong about a prediction.

Why it matters

It shows what a theoretical framework's limits actually look like. Functionalism did not give wrong answers about protest, revolution and decolonisation. It could not ask the questions, because its central concepts had no place for a judgement on a system as opposed to a symptom within one. Recognising that kind of limit — not error but incapacity — is one of the more valuable skills in reading theory.

It supplies the settled position on functional explanation: valid only with a named feedback mechanism. That rule is transferable to any claim, in any field, that something exists because of what it does.

It documents how theoretical change actually happens. Not by refutation but by the arrival of people with different questions — which is a finding about disciplines that this course has now made four times (3.3.3, 4.5.5, 4.6.1, and here) with four different mechanisms.

And it establishes the shape of Part 5. Every school in this Part rose against something, dominated in some place for some period, and was in turn attacked. Reading them as a sequence of corrections — each seeing what the last could not — is more accurate than reading them as competitors, and considerably more useful.

The four attacks. It cannot explain change — differentiation and evolutionary universals give smooth development and no account of rupture. It ignores power and conflict — Mills on grand theory's abstraction, Dahrendorf's two faces of society (integration and coercion), Gouldner on the encoded politics. It is conservative — the Davis–Moore thesis and Tumin's demolition of its circularity. And teleology — no feedback mechanism connecting consequence to cause.

But the arguments were mostly old. What changed was the world and, decisively, who was in the room: postwar educational expansion brought a generation with questions the framework could not formulate. The civil rights movement, decolonisation, Vietnam and second-wave feminism were not phenomena functionalism got wrong — they were phenomena it could only register as strain, deviance or breakdown.

And then systemic analysis returned. Luhmann's functionally differentiated autopoietic subsystems, each operating on its own binary code, with structural coupling and no centre — a systems theory whose central finding is that modern society cannot be steered. Alexander's neofunctionalism and the strong programme. And new institutionalism, which asks why forms persist and answers with coercive, mimetic and normative isomorphism — functional analysis with the mechanism supplied, explaining legitimacy-seeking rather than efficiency.

The settled position: functional explanation is valid if and only if a feedback mechanism can be named.

And the pattern — dominance, overreach, collapse, overcorrection, quiet return — recurs across schools, which means that when a framework is being demolished, the useful question is which parts are defective and which are merely unfashionable.

Teleology — explaining a thing by its consequences; illegitimate without a feedback mechanism.

Feedback mechanism — the specified process by which consequences sustain the arrangement producing them.

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

Two faces of society — Dahrendorf's insistence that integration and coercion must both be theorised.

Modernisation theory — the developmental application of functionalism to non-Western societies; displaced by dependency theory.

Autopoiesis — Luhmann's term for a system that produces its own elements through its own operations.

Binary code — the distinction (legal/illegal, payment/non-payment, true/false) by which a Luhmannian subsystem operates.

Structural coupling — the indirect relation between operationally closed subsystems.

Neofunctionalism — Alexander's attempted revival incorporating conflict, contingency and agency.

Institutional isomorphism — DiMaggio and Powell's account of organisational convergence via coercive, mimetic and normative mechanisms.

Legitimacy-seeking — the adoption of organisational forms for recognisability rather than efficiency.

One — find the isomorphism. Take any sector — universities, hospitals, charities, startups — and identify practices that spread despite unclear evidence of benefit. Then work out which of the three mechanisms drove each.

Two — apply the mechanism rule. Take any claim of the form "X exists because it serves Y." Ask how the serving reaches back and sustains X. If no answer, the sentence is a story.

Three — locate a framework's incapacity. Take any theory you find persuasive and identify a question it cannot formulate — not one it answers badly, one it cannot pose. Every framework has some.

Four — read Tumin. The 1953 reply to Davis and Moore is short and freely available. Read it after the original. It is the best short demonstration of how to dismantle an argument.

Five — plot a current school on the curve. Dominance, overreach, collapse, overcorrection, return. Take a framework that currently feels obviously correct in some field and ask where on that curve it sits.