Conflict Theory - Coser, Dahrendorf and the Functions of Conflict

If functionalism's central question was how societies hold together, conflict theory's is how they are held together — by whom, at whose cost, and against what opposition.

It is not simply Marxism. Its two founding figures were both refugees from Nazi Europe, both anti-Communist, and both explicitly arguing that a general sociology of conflict was needed because Marxism had got the specifics wrong while getting the framing right.

Two families leaving Europe.

Lewis Coser is born Ludwig Cohen in Berlin in 1913. He is Jewish; he joins a socialist student group; he flees to Paris in 1933 when the Nazis take power. In Paris he is interned in a French camp as an enemy alien in 1939, released, and escapes to the United States in 1941. He changes his name. He learns English. He completes a doctorate at Columbia under Merton, and teaches at Brandeis and Stony Brook.

Ralf Dahrendorf is born in Hamburg in 1929. His father is a Social Democratic member of the Reichstag, arrested by the Nazis. In 1944, at fifteen, Ralf is himself arrested for distributing anti-Nazi leaflets at school and sent to a camp in what is now Poland. He survives. After the war he takes doctorates at Hamburg and at the London School of Economics, becomes a German MP, then a European Commissioner, then Director of the LSE, then a member of the British House of Lords.

Both men spent the 1950s and 60s in an American and British sociology whose dominant framework held that societies cohere around shared values, that conflict is a symptom of imperfect integration, and that the normal state of a social system is equilibrium.

Both had personally watched a society tear itself apart.

Neither was a Marxist — Coser co-founded the anti-Stalinist journal Dissent; Dahrendorf was a liberal who served in a centrist German government. Their objection was not that functionalism was insufficiently revolutionary.

It was that a theory in which conflict is an anomaly is a theory that cannot describe the twentieth century.

The intellectual problem is stated most precisely by Dahrendorf, and it is not a demand to replace one framework with another.

Every society displays two faces simultaneously:

The integration face — stability, consensus, shared values, the functional interrelation of parts. This is real. Societies do persist, people do share values, institutions do interlock.

The coercion face — change, conflict, disintegration, the imposition of order by some over others. This is equally real.

Dahrendorf's claim in "Out of Utopia" (1958): these are not competing theories, one of which is correct. They are two aspects of the same object, and any adequate sociology must be able to describe both. A theory that sees only integration describes a utopia — literally, a place that does not exist — and functionalism, for all its sophistication, had produced one.

His diagnosis of why: functionalism took the problem of order as its founding question (1.2.1), and a question shapes what its answers can contain. Ask how order is possible and you will produce a theory of order. The alternative question — how is change possible, and why does it constantly occur — produces a different theory of the same society.

And the resolution he proposed is not synthesis. He argued explicitly that no single theory could hold both faces without becoming vacuous, and that sociology should therefore work with two frameworks, applied to different questions, without pretending they merge.

The Functions of Social Conflict (1956) does something unexpected: it argues for conflict's importance using functionalist logic, by systematically developing Simmel's essay on conflict (4.4.1).

Coser's central claim: conflict is not the breakdown of social relations. It is a form of sociation — a way of being in relation — and it performs identifiable functions for groups and societies.

He derives sixteen propositions from Simmel. The most important:

Conflict establishes and maintains group boundaries. A group discovers what it is by discovering what it is against. Without an outside, the inside is indistinct.

Conflict with an out-group increases internal cohesion. Well documented, and Coser adds the crucial qualification: this holds only if the group already has a basic consensus and the conflict does not concern its founding values. A group already divided is fractured further by external threat, not united.

Conflict generates the machinery of organisation. A group in conflict develops leadership, discipline, centralised decision-making and a division of labour. Many organisations exist because a fight required them.

Conflict is a stimulus to innovation. It forces the invention of new methods, technologies and arrangements.

Conflict establishes and maintains balances of power. It is the mechanism by which relative strength is ascertained — and, once ascertained, is what allows accommodation.

Conflict creates associations and coalitions. Parties who would not otherwise combine do so against a common opponent.

Realistic vs non-realistic conflict — Coser's most useful analytical tool.

Realistic conflict arises from a genuine clash of interests over a scarce good — wages, territory, a position, a decision. The conflict is a means to an end, and if the end can be obtained another way, the conflict can be abandoned.

Non-realistic conflict arises from the need for tension release rather than from a clash of interests. Its object is substitutable — any available target will serve.

The practical significance: realistic conflict is negotiable, because the parties want something specific and the question is terms. Non-realistic conflict is not, because the conflict is the point.

And the diagnostic question — would this conflict end if the stated demand were met? — is one of the most useful things in this lesson. Applied to industrial disputes, political controversies, family arguments and international crises, it separates the tractable from the intractable immediately.

Coser's concept of the safety valve is his most interesting and most double-edged.

The argument: societies develop institutions that permit the discharge of hostility without threatening the arrangements that generate it. Sanctioned targets, ritualised aggression, permitted grumbling, formal complaint procedures.

His historical examples include court jesters licensed to mock the king, carnival and festivals of inversion, and duelling codes that channelled violence into regulated forms.

Modern instances: the grievance procedure that absorbs a complaint into a process; the consultation exercise whose outcome is predetermined; the protest permitted in a designated area on a designated day; the internal survey that lets people register dissatisfaction anonymously and changes nothing.

Coser's own assessment is genuinely ambivalent, and this is what makes the concept good.

The functional reading: safety valves prevent accumulation. A society with no outlet for hostility stores it, and stored hostility discharges catastrophically.

And the dysfunctional reading, which Coser states explicitly: safety valves leave the source of the hostility untouched. They dissipate the feeling without addressing the cause, and thereby protect the arrangement generating it. Worse, they can displace hostility onto substitute targets — which converts realistic conflict into non-realistic conflict and makes the original grievance harder to see.

The test he proposes: does the institution allow the grievance to reach the party who could act on it, or does it absorb the grievance and return the person to their situation unchanged?

That question, asked of any complaints procedure, consultation, ombudsman, works council or feedback form, is uncomfortably productive.

And Coser's crucial qualification, which distinguishes his position from an endorsement of conflict: the functions depend on the type of conflict and the structure of the group.

Conflict over basic values threatens a group's existence; conflict over means, allocation or terms within a shared framework strengthens it. Loosely structured, pluralistic societies tolerate and benefit from frequent conflicts, because multiple crosscutting conflicts (4.4.2) prevent any single cleavage from splitting the society. Rigidly structured societies suppress conflict, which accumulates, and when it breaks out it does so along a single line, catastrophically.

This yields a genuine and testable proposition: societies that permit frequent small conflicts are more stable than those that permit none.

Dahrendorf: authority and imperatively coordinated associations

Dahrendorf's Class and Class Conflict in Industrial Society (1959) offers a general theory of conflict, and its key move is to replace Marx's variable.

Marx located conflict in ownership of the means of production (4.1.3). Dahrendorf argues this is a special case of something more general — and that Marx's version has been overtaken by events, since ownership and control separated as the joint-stock corporation spread, and managers who own nothing exercise the authority Marx attributed to owners.

His replacement variable is authority.

Every organisation is an imperatively coordinated association (Herrschaftsverband, from Weber, 4.3.5) — an association in which some positions carry the right to issue binding commands and others carry the obligation to obey.

Authority is dichotomous: in any given association, you either hold it or you are subject to it. There is no continuum.

And authority is scarce and zero-sum: for someone to command, someone must obey.

Therefore every association generates two quasi-groups with structurally opposed latent interests: those in authority have an interest in maintaining the arrangement; those subject to it have an interest in changing it.

The structural consequences are what make the theory interesting:

Conflict is universal and permanent, because every organisation has an authority structure. There is no society without conflict, because there is no society without organisation.

Conflict is plural, not singular. A person is in authority in one association and subject to it in another — a foreman commands on the shop floor and obeys at the union branch, a bishop commands a diocese and obeys a synod. Multiple crosscutting conflicts, not a single class war, which is Simmel's affiliation argument (4.4.2) applied to power.

Class is redefined as conflict group formed from an authority position, not from property. This is a substantial departure from Marx and it is the theory's most contested element.

And the quasi-group becomes an interest group only under specifiable conditions: technical (organisation, leadership, an ideology), political (freedom to associate), and social (communication and recruitment). This is 4.1.8's problem, restated with the conditions named.

The institutionalisation of class conflict — Dahrendorf's most consequential empirical claim.

His argument: in advanced industrial societies, class conflict has not disappeared. It has been regulated.

The mechanisms: recognition of unions and employers' associations as legitimate parties; established procedures for negotiation, arbitration and industrial action; labour law; and the acceptance by both sides that the other has a right to exist and to pursue its interests.

The consequence: conflict becomes intense but not violent, continuous but bounded, and productive of incremental change rather than rupture. It moves from the street to the negotiating table.

And Dahrendorf's claim is that this regulation, not the abolition of conflict, is what a stable industrial democracy consists of. The choice is not conflict or peace; it is regulated conflict or unregulated conflict.

How it has fared. The claim described postwar Western Europe well, and the subsequent decades have tested it. Union density has fallen sharply since 1980 (4.1.8); collective bargaining coverage has declined in most countries; and the institutions Dahrendorf described as the achievement of industrial democracy have been substantially weakened. On his own theory, that predicts a return of less regulated forms of conflict — which is a checkable claim about the present, and one worth watching.

Go deeper

The two frameworks compared

	Functionalism	Conflict theory
Society is	a system of interdependent parts	an arena of competing interests
Held together by	shared values, consensus	constraint, power, regulation
Normal state	equilibrium	change and conflict
Change comes from	differentiation, adjustment	conflict between opposed groups
Institutions	perform functions for the whole	serve the interests of the dominant
Inequality is	functionally necessary motivation	the outcome of power
Founding question	how is order possible?	how is order imposed, and by whom?

And Dahrendorf's own position on this table is worth stating precisely: he did not think one column was correct. He held that both describe real features, that no single theory can contain both without becoming empty, and that the researcher should select the framework that fits the question. That is a more sophisticated position than either partisanship or synthesis, and it is close to the toolkit approach of 4.6.4.

Randall Collins: conflict theory made micro

The most substantial development, and the one that connects this Topic to the rest of Part 5.

Randall Collins, Conflict Sociology (1975), argued that conflict theory had remained too macro and needed microfoundations (4.2.8) — an account of what people actually do, in encounters, that produces the large patterns.

His principles: everyone seeks to maximise their subjective status; people have conflicting interests because status is relational and scarce; those with resources use them to dominate encounters; and the resources are principally material means of coercion, means of emotional production (ritual), and cultural resources — networks, vocabularies, credentials.

And his most productive application is The Credential Society (1979): the argument that educational expansion is driven not by the technical requirements of jobs but by credential inflation — occupational groups using qualifications to restrict entry and defend earnings, which is Weber's social closure (4.3.6) as an engine of educational growth.

His evidence: the skill content of jobs has risen far less than the credential requirements attached to them, and the correlation between educational expansion and productivity growth is weaker than the human capital account predicts. The prediction — that credentials will inflate indefinitely because each group must acquire more to stay ahead — has held up uncomfortably well.

Collins later synthesised this with Durkheim's ritual theory into Interaction Ritual Chains (2004) (4.2.6), producing what is arguably the most complete micro-macro theory anyone has built — conflict, ritual and emotional energy in one framework.

Where conflict theory sits now

It won, in a specific sense: the assumption that societies are arenas of competing interests, that institutions distribute advantage unevenly, and that conflict is normal rather than pathological is now the discipline's default. Almost nobody argues otherwise.

Which means it stopped being a distinct school. There is no "conflict theory department." Its assumptions were absorbed into stratification research, political sociology, social movement theory, organisational sociology and the sociology of education, and the label is now largely a textbook category.

The criticisms that remain live:

Is authority really dichotomous? Dahrendorf's insistence that one either has authority or does not is hard to sustain against the enormous gradations of real organisations, and against Wright's contradictory class locations (4.1.3).

Does the plurality of conflicts explain their patterning? If everyone is in authority somewhere and subject to it elsewhere, why do conflicts nonetheless line up so consistently along class, race and gender? Dahrendorf's framework predicts more randomness than we observe.

And has conflict theory lost the analysis of what conflict is about? Coser's forms and Dahrendorf's authority relations are formal categories, and the objection is Simmel's (4.4.1): treating conflict as a form abstracts from the content, and the content — exploitation, exclusion, domination — is what people are actually fighting over.

"Conflict theory is Marxism." Coser and Dahrendorf were both anti-Communist and both argued that Marx's specifics had failed. Conflict theory is a general framework of which Marxism is one variant.

"Conflict theory says society is chaos." Dahrendorf's central empirical claim is about the institutionalisation of conflict — that regulated conflict is what stable democracies consist of.

"Coser was arguing against functionalism." He argued that conflict performs functions, using functionalist logic, and was in that sense working inside it.

"Safety valves are good for societies." Coser is deliberately ambivalent: they prevent dangerous accumulation and leave the source untouched, protecting the arrangement that generates the hostility.

"External threat unites a group." Only if the group has prior basic consensus and the conflict does not touch its founding values. Otherwise it fractures.

"Conflict theory replaced functionalism." Its assumptions were absorbed so completely that it dissolved as a school — which is a more complete victory and a less visible one.

Why it matters

It supplies the discipline's default assumption, so thoroughly that it is now invisible: that societies contain groups with opposed interests, that institutions distribute advantage, and that conflict is ordinary. Almost all contemporary sociology proceeds on this basis.

It gives you two diagnostic tools you will use repeatedly. Realistic versus non-realistic conflict — would this end if the stated demand were met? And the safety valve test — does this procedure carry the grievance to someone who can act, or absorb it?

It reframes what stability is. Dahrendorf's argument that regulated conflict, not the absence of conflict, is what stable societies achieve is a genuinely useful correction to both the functionalist assumption that peace means consensus and the revolutionary assumption that regulation means suppression.

And it demonstrates Dahrendorf's methodological point, which is the most transferable thing in the lesson: two frameworks describing the same object, neither reducible to the other, selected by question. That is what having read Part 4 was supposed to equip you for, and this Topic is the first place Part 5 requires it.

Coser and Dahrendorf were both refugees from Nazi Europe, both anti-Communist, and both objecting that a theory in which conflict is anomalous cannot describe the twentieth century.

Dahrendorf's "Out of Utopia" (1958): every society has an integration face and a coercion face; both are real; no single theory holds both without becoming vacuous; therefore use two frameworks selected by question.

Coser's Functions of Social Conflict (1956) develops Simmel: conflict establishes group boundaries, increases internal cohesion (only where basic consensus already exists), generates organisation, stimulates innovation, establishes balances of power, and creates coalitions. His key distinction is realistic conflict (over a genuine scarce good, therefore negotiable) versus non-realistic (tension release, target substitutable, therefore not). And the safety valve is deliberately double-edged: it prevents dangerous accumulation and leaves the source untouched.

And his testable proposition: societies permitting frequent small crosscutting conflicts are more stable than those permitting none.

Dahrendorf replaces Marx's variable. Not ownership but authority: every imperatively coordinated association divides into those with the right to command and those with the duty to obey, generating opposed latent interests. Conflict is therefore universal, permanent and plural — one is in authority in one association and subject to it in another. And his central empirical claim is the institutionalisation of class conflict: recognition, procedure and law make conflict intense but not violent — which predicts, on his own theory, that the weakening of those institutions since 1980 returns less regulated forms.

Randall Collins supplied the microfoundations, and The Credential Society (1979) explains educational expansion by credential inflation rather than skill requirements — a prediction that has held up uncomfortably well.

Conflict theory won so completely that it dissolved as a school, its assumptions absorbed into the discipline's default.

Integration and coercion faces — Dahrendorf's two aspects of every society, requiring separate frameworks.

Realistic conflict — conflict over a genuine scarce good; negotiable.

Non-realistic conflict — conflict serving tension release, with a substitutable target; not negotiable.

Safety valve — an institution permitting the discharge of hostility without addressing its source.

Imperatively coordinated association (Herrschaftsverband) — any organisation with an authority structure.

Quasi-group — an aggregate sharing a latent interest by virtue of authority position, prior to organisation.

Interest group — a quasi-group that has organised, under specifiable technical, political and social conditions.

Institutionalisation of conflict — the regulation of conflict through recognition, procedure and law.

Credential inflation — the rising qualification requirements for occupations, driven by closure rather than by skill.

Crosscutting conflicts — multiple overlapping lines of opposition, which prevent a society splitting along one.

One — apply the realistic/non-realistic test. Take a current public controversy and ask whether it would end if the stated demand were met. If not, the stated demand is not what it is about.

Two — audit a safety valve. Find a complaints or consultation procedure you have used. Ask whether it carried your grievance to someone who could act, or absorbed it. Note how you felt afterwards.

Three — map an authority structure. Take any organisation and identify who can issue binding commands to whom. Then identify who is in authority there and subject to it elsewhere. Dahrendorf's plurality becomes visible immediately.

Four — test the cohesion claim. Find a group that faced external threat and split rather than uniting. Ask whether the threat touched its founding values. Coser predicts it did.

Five — check credential inflation. Find a job whose entry requirements have risen over thirty years. Ask whether the work itself changed correspondingly. Usually it did not.